PHARMACEUTICAL MANUFACTURING HANDBOOK
Regulations and Quality

SHAYNE COX GAD, PH.D., D.A.B.T. 
Gad Consulting Services 
Cary, North Carolina 
A JOHN WILEY & SONS, INC., PUBLICATION




CONTENTS 
SECTION 1 GOOD MANUFACTURING PRACTICES (GMP) AND 
OTHER FDA GUIDELINES 1 
1.1 Good Manufacturing Practices (GMPs) and Related FDA 
Guidelines 3 
James R. Harris 
1.2 Enforcement of Current Good Manufacturing Practices 45 
Kenneth J. Nolan 
1.3 Scale-Up and Postapproval Changes (SUPAC) Regulations 67 
Puneet Sharma, Srinivas Ganta, and Sanjay Garg 
1.4 GMP-Compliant Propagation of Human Multipotent Mesenchymal 
Stromal Cells 97 
Eva Rohde, Katharina Schallmoser, Christina Bartmann, Andreas Reinisch, 
and Dirk Strunk 
SECTION 2 INTERNATIONAL REGULATIONS OF GOOD 
MANUFACTURING PRACTICES 117 
2.1 National GMP Regulations and Codes and International GMP 
Guides and Guildelines: Correspondences and Differences 119 
Marko Narhi and Katrina Nordstrom

x CONTENTS 
SECTION 3 QUALITY 163 
3.1 Analytical and Computational Methods and Examples for Designing 
and Controlling Total Quality Management Pharmaceutical 
Manufacturing Systems 165 
Paul G. Ranky, Gregory N. Ranky, Richard G. Ranky, and Ashley John 
3.2 Role of Quality Systems and Audits in Phatmaceutical 
Manufacturing Environment 201 
Evan B. Siegel and James M. Barquest 
3.3 Creating and Managing a Quality Management System 239 
Edward R. Arling, Michelle E. Dowling, and Paul A. Frankel 
3.4 Quality Process Improvement 287 
Jyh-hone Wang 
SECTION 4 PROCESS ANALYTICAL TECHNOLOGY (PAT) 311 
4.1 Case for Process Analytical Technology: Regulatory and Industrial 
Perspectives 313 
Robert P. Cogdill 
4.2 Process Analytical Technology 353 
Michel Ulmschneider and Yves Roggo 
4.3 Chemical Imaging and Chemometrics: Useful Tools for Process 
Analytical Technology 411 
Yves Roggo and Michel Ulmschneider 
SECTION 5 PERSONNEL 433 
5.1 Personnel Training in Pharmaceutical Manufacturing 435 
David A. Gallup, Katherine V. Domenick, and Marge Gillis 
SECTION 6 CONTAMINATION AND CONTAMINATION 
CONTROL 455 
6.1 Origin of Contamination 457 
Denise Bohrer 
6.2 Quantitation of Markers for Gram-Negative and Gram-Positive 
Endotoxins in Work Environment and as Contaminants in 
Pharmaceutical Products Using Gas Chromatography–Tandem 
Mass Spectrometry 533 
Alvin Fox 
6.3 Microbiology of Nonsterile Pharmaceutical Manufacturing 543 
Ranga Velagaleti

CONTENTS xi 
SECTION 7 DRUG STABILITY 557 
7.1 Stability and Shelf Life of Pharmaceutical Products 559 
Ranga Velagaleti 
7.2 Drug Stability 583 
Nazario D. Ramirez-Beltran, Harry Rodriguez, and L. Antonio Estevez 
7.3 Effect of Packaging on Stability of Drugs and Drug Products 641 
Emmanuel O. Akala 
7.4 Pharmaceutical Product Stability 687 
Andrew A. Webster 
7.5 Alternative Accelerated Methods for Studying Drug Stability: 
Variable-Parameter Kinetics 701 
Giuseppe Alibrandi 
SECTION 8 VALIDATION 725 
8.1 Analytical Method Validation: Principles and Practices 727 
Chung Chow Chan 
8.2 Analytical Method Validation and Quality Assurance 743 
Isabel Taverniers, Erik Van Bockstaele, and Marc De Loose 
8.3 Validation of Laboratory Instruments 791 
Herman Lam 
8.4 Pharmaceutical Manufacturing Validation Principles 811 
E. B. Souto T. Vasconcelos D. C. Ferreira, and B. Sarmento 
INDEX 839



PREFACE 
This Handbook of Manufacturing: Regulations and Quality focuses on all regulatory 
aspects and requirements that govern how drugs are produced for evaluation (and, 
later, sale to and use in) humans. The coverage ranges from what the issues are at 
the early stages (when the amounts are small and the materials of limited sophistication) 
up to until the issue is reproducibly and continuously making large volumes 
of a highly sophisticated manufactured product. These 25 chapters cover the full 
range from preformulation of a product (the early exploratory work that allows us 
to understand how to formulate and deliver the drug) to identifi cation of sources 
of contamination and assessment of stability. 
The Handbook of Manufacturing: Regulations and Quality seeks to cover the 
entire range of available approaches to satisfying the wide range of regulatory 
requirements for making a highly defi ned product that constitutes a successful new 
drug and how to do so in as effective and as effi cient a manner as possible. 
Thanks to the persistent efforts of Michael Leventhal, these 25 chapters, which 
are written by leading practitioners in each of these areas, provide coverage of the 
primary approaches to the fundamental regulatory challenges that must be overcome 
to manufacture successfully a deliverable and stable new drug. 


GOOD MANUFACTURING PRACTICES 
( GMP ) AND OTHER FDA GUIDELINES 
SECTION 1


3 
1.1 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc. 
GOOD MANUFACTURING 
PRACTICES ( GMP ) AND RELATED 
FDA GUIDELINES 
James R. Harris 
James Harris Associates, Inc., Durham, North Carolina 
Contents 
1.1.1 FDA Regulations: Real and Imagined 
1.1.2 21 CFR 210 and 211: Current Good Manufacturing Practice for Finished 
Pharmaceuticals 
1.1.3 Guidance for Industry: Quality Systems Approach to Pharmaceutical Current Good 
Manufacturing Practice Regulations 
1.1.3.1 CGMPS and the Concepts of Modern Quality Systems 
1.1.3.2 Quality Systems Model 
1.1.4 Guidance for Industry: PAT — Framework for Innovative Pharmaceutical Development, 
Manufacturing, and Quality Assurance 
1.1.4.1 PAT Framework 
1.1.5 Guidance for Industry: Part 11. Electronic Records; Electronic Signatures — Scope and 
Application 
1.1.6 Guidance for Industry and FDA: Current Good Manufacturing Practice for Combination 
Products 
1.1.7 Guidance for Industry: Powder Blends and Finished Dosage Units — Stratifi ed In - 
Process Dosage Unit Sampling and Assessment 
1.1.7.1 Validation of Batch Powder Mix Homogeneity 
1.1.7.2 Verifi cation of Manufacturing Criteria 
1.1.8 Guidance for Industry: Immediate - Release Solid Oral Dosage Forms Scale - Up and 
Postapproval Changes (SUPAC) — Chemistry, Manufacturing and Controls, In Vitro 
Dissolution Testing, and In Vivo Bioequivalence Documentation 
1.1.9 Other GMP - Related Guidance Documents 

4 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
1.1.1 FDA REGULATIONS: REAL AND IMAGINED 
A regulation is a law. In the United States, all federal laws have been arranged or 
codifi ed in a manner that makes it easier to fi nd a specifi c law. The Code of Federal 
Regulations (CFR) is a compilation of all federal laws published in the Federal 
Register by the executive departments and agencies of the federal government. This 
code is divided into 50 titles which represent broad areas of federal regulation. Each 
title is further divided into chapters. The chapters are then subdivided into parts 
covering specifi c regulatory areas. Changes and additions are fi rst published in the 
Federal Register . Both the coded law and the Federal Register must be used to determine 
the latest version of any rule. All food - and drug - related laws are contained 
in Title 21 of the CFR. Each title of the CFR is updated annually. Title 21 is updated 
as of April 1 of each year. 
Because virtually all of the drug regulations are written to state what should be 
done but do not tell how to do it, the Food and Drug Administration (FDA) also 
publishes guidance documents. These documents are intended to provide precisely 
what the name implies — guidance. In this context, guidance documents are not law 
and do not bind the FDA or the public . Manufacturers are not required to use the 
techniques or approaches appearing in the guidance document. In fact, FDA representatives 
have repeatedly stated that the regulations were not written to suggest 
how something should be done in order to encourage innovation. While following 
the recommendations contained in the guidance documents will probably assure 
acceptance (agency philosophy and interpretation may have changed since the guidance 
document was published), other approaches are encouraged. No matter how 
they choose to proceed, manufacturers should be prepared to show that their 
methods achieve the desired results. 
A method used by the FDA to “ fl oat ” new ideas is to discuss them at industry 
gatherings such as FDA - sponsored seminars or meetings of industry groups such as 
the Pharmaceutical Manufacturers Association (PMA), the Parenteral Drug Association 
(PDA), and the International Society of Pharmaceutical Engineering (ISPE). 
Again, it must be remembered that while these comments refl ect current FDA 
thinking, they are simply thoughts and recommendations. They are not law. 
Several industry groups also publish comments, guidelines, and so on, that put 
forth current thinking of the group writing the document. These publications are 
interesting and often bring out valuable information. However, it is important to 
remember that these publications are not regulations or even offi cial guidance documents. 
If a fi rm chooses to follow the recommendations of such documents, they are 
probably following good advice. However, since the advice comes from a nonoffi cial 
source, fi rms should still be prepared to defend their actions with good scientifi c 
reasoning. 
1.1.2 21 CFR 210 AND 211: CURRENT GOOD MANUFACTURING 
PRACTICE FOR FINISHED PHARMACEUTICALS 
Parts 210 and 211 of CFR Title 21 are the laws defi ning good manufacturing practices 
for fi nished pharmaceutical products. All manufacturers must follow these 
regulations in order to market their products in the United States. When a fi rm fi les 
an application to market a product in the United States through a New Drug Application 
(NDA), abbreviated NDA, (ANDA), Biological License Application (BLA), 

CURRENT GOOD MANUFACTURING PRACTICE 5 
or other product application, one of the last steps in approving the application is a 
preapproval inspection of the manufacturing facility. A major purpose of this inspection 
is to assure adherence to the GMP regulations. Preapproval inspections are a 
part of every application approval. Thus, if a fi rm has 10 applications pending, it 
should expect 10 inspections. The fact that the manufacturing facility has already 
been inspected will not alter the need for another inspection. 
The FDA also has the right to visit and inspect any manufacturing facility that 
produces a product or products sold in the United States. Such inspections are unannounced. 
A manufacturer must admit an inspector when he or she appears at that 
facility and must do so without undue delay. 
GMP requirements for manufacturers of pharmaceutical dosage forms are discussed 
below. This information should not be considered to be an exact statement 
of the law. We have attempted to show intent and, occasionally, add some comments 
that will clarify how that particular regulation is interpreted. For precise wording of 
a regulation, refer to the CFR and then check the Federal Register to determine if 
there have been any changes since the last update. 
General Provisions 
1. This section pertains to the manufacture of drug products for humans or 
animals. 
2. These requirements will not be enforced for over - the - counter (OTC) drug products 
if the products and all their ingredients are ordinarily marketed and considered 
as human foods and which products may also fall within the legal defi nition 
of drugs by virtue of their intended use. 
Organization and Personnel 
1. Responsibilities of quality control unit 
(a) A quality control unit must be a part of the facility organization. 
(b) This unit must be given responsibility and authority to approve or reject all 
components, drug product containers, closures, process materials, packaging 
material, labeling, and drug products, and the authority to review production 
records. 
(c) Adequate laboratory facilities for testing and approval or rejection of the 
above listed materials must be available. 
(d) The quality control unit is responsible for approving or rejecting all procedures 
or specifi cations that impact on the identity, strength, quality, and purity 
of the drug product. 
(e) Responsibilities and procedures applicable to the quality control unit must 
be written and these procedures must be followed. 
2. Personnel qualifi cations 
(a) Every person involved in the manufacture, processing, packing, or holding of 
a drug product must have education, training, and experience that enable 
that individual to perform their duties. Employees must be trained in the 
particular operations that they perform and in Current GMPs (CGMPs). The 
GMP training must be conducted by qualifi ed individuals and with suffi cient 
frequency to assure that workers remain familiar with the requirements 
applicable to them. 

6 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
(b) Persons responsible for supervision must have the education, training, and 
experience to perform their assigned functions in such a manner as to assure 
that the drug product has the safety, identity, strength, quality, and potency 
that it is represented to possess. 
(c) There must be an adequate number of qualifi ed personnel to perform the 
needed tasks. 
3. Personnel responsibilities 
(a) Personnel shall wear clean clothing appropriate for the duties they perform. 
Protective apparel must be worn as necessary. 
(b) Personnel shall practice good sanitation and health habits. 
(c) Only personnel authorized by supervisory personnel shall enter those areas 
designated as limited - access areas. 
(d) Any worker considered to have an apparent illness or open lesions that may 
adversely affect safety or quality of drug products shall be excluded from 
direct contact with product, components, or containers. 
4. Consultants that advise on the manufacture, processing, packing, or holding of 
drug products must have suffi cient education, training, and experience to advise 
on the subject for which they are retained. The manufacturer must maintain 
records of name, address, and qualifi cations of any consultants and the type of 
service they provide. 
Buildings and Facilities 
1. Design and construction features 
(a) Buildings should be of suitable size, construction location to facilitate cleaning, 
maintenance, and proper operations. 
(b) Space should be adequate for the orderly placement of equipment and materials 
to prevent mix - ups between different components, drug product containers 
and closures, labeling, in - process materials, or drug products and to 
prevent contamination. 
(c) The movement of components and product through the building must be 
designed to prevent contamination. 
(d) Operations should be performed within specifi cally defi ned areas having 
adequate control systems to prevent contamination or mix - ups during each 
of the following procedures: 
(i) Receipt, identifi cation, storage, and withholding from use of components, 
drug product containers, closures, and labeling, pending the 
appropriate sampling, testing, and release for manufacturing or 
packaging. 
(ii) Holding rejected materials listed in (a) above. 
(iii) Storage of released components, drug product containers, closures, and 
labeling. 
(iv) Storage of in - process materials. 
(v) Manufacturing and processing operations. 
(vi) Packaging and labeling operations. 
(vii) Quarantine storage before release of drug products. 
(viii) Storage of drug products after release. 
(ix) Control and laboratory operations. 

CURRENT GOOD MANUFACTURING PRACTICE 7 
(x) Aseptic processing, which includes: 
(1) Floors, walls, and ceilings of smooth, hard surfaces that are easily 
cleanable. 
(2) Temperature and humidity controls. 
(3) An air supply fi ltered through High - Effi ciency Particulate Air 
(HEPA) fi lters under positive pressure regardless of whether fl ow 
is laminar or nonlaminar. 
(4) A system for monitoring environmental conditions. 
(5) A system for cleaning and disinfecting the room and equipment to 
produce aseptic conditions. 
(6) A system for maintaining any equipment used to control the aseptic 
conditions. 
(e) Operations relating to the manufacture, processing, and packing of penicillin 
must be performed in facilities separate from those used for other drug 
products for humans. Note : For all purposes of these GMP regulations, the 
FDA considers cephalosporins to be penicillin. 
2. Adequate lighting should be provided in all areas. 
3. Heating, ventilation, and air conditioning (HVAC) 
(a) Adequate ventilation is required in all areas. 
(b) Equipment for adequate control over air pressure, microorganisms, dust, 
humidity, and temperature must be provided when appropriate for the manufacture, 
processing, packing, or holding of a drug product. 
(c) When appropriate, air supplied to production areas should be fi ltered to 
avoid any possibility of contamination or cross - contamination. 
(d) Air - handling systems for the manufacture, processing, and packing of penicillin 
shall be completely separate from those for other drug products for 
humans. 
4. Plumbing 
(a) Potable water should be supplied in a continuous positive - pressure system 
free from defects that could contribute to contamination of any drug 
product. 
(b) Potable water must meet the standards prescribed in the Environmental 
Protection Agency (EPA) Primary Drinking Water Regulations defi ned in 
40 CFR Part 141. 
(c) Drainage must be of adequate size. Where connected directly to a sewer, an 
air break or other suitable mechanical device must be provided to prevent 
back - siphonage. 
5. Sewage, trash, and other refuse in and from the building and immediate premises 
must be disposed of in a safe and sanitary manner. 
6. Adequate washing facilities should be provided. This is to include hot and cold 
water, soap or detergent, air driers or single - service towels, and clean toilet facilities 
easily accessible to all work areas. 
7. Sanitation 
(a) Any building used for manufacture, processing, packing, or holding of a drug 
product should be maintained in a clean and sanitary condition. Such buildings 
should be free of infestation by rodents, birds, insects, and other vermin. 
(b) Trash and organic waste matter should be held and disposed of in a timely 
and sanitary manner. 

8 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
(c) Written procedures assigning responsibility for sanitation and describing in 
suffi cient detail the cleaning schedules, methods, equipment, and materials 
to be used in cleaning the buildings and facilities are required. Such procedures 
must be followed. 
(d) Written procedures for use of suitable rodenticides, insecticides, fungicides, 
fumigating agents, and cleaning and sanitizing agents are required and must 
be followed. These written procedures should be designed to prevent the 
contamination of equipment, components, product containers, closures, packaging, 
labeling materials, or drug products. Agent may not be used unless 
registered and used in accordance with the Federal Insecticide, Fungicide, 
and Rodenticide Act (7 U.S.C. 135). 
(e) All sanitation procedures apply equally to contractors or temporary employees 
as to regular employees. 
8. All buildings used for GMP - related purposes must be maintained in a good state 
of repair. 
Equipment 
1. Equipment should be of appropriate design, adequate size, and suitably located 
to facilitate operations for its intended use and for cleaning and maintenance. 
2. Equipment construction 
(a) Equipment should be constructed so that surfaces that contact components, 
in - process materials, or drug products should not be reactive, additive, or 
absorptive so as to alter the safety, identity, strength, quality, or purity of the 
drug product beyond offi cial or other established requirements. 
(b) Any substance required for operation such as lubricants or coolants shall not 
come into contact with drug products, containers, and so on, so as to alter 
the safety, identity, strength, quality, or purity of the drug product beyond 
established requirements. 
3. Equipment cleaning and maintenance 
(a) Equipment and utensils should be cleaned, maintained, and sanitized at 
appropriate intervals to prevent malfunctions or contamination that would 
alter the drug product beyond the offi cial requirements. 
(b) Written procedures must be established and followed for cleaning and 
maintenance of equipment and utensils used in the processing of a drug 
product. These procedures must include but are not limited to the 
following: 
(i) Assignment of responsibility for cleaning and maintaining equipment. 
(ii) Maintenance and cleaning schedules, including sanitizing schedules if 
appropriate. 
(iii) A suffi ciently detailed description of the methods, equipment, and 
materials used in cleaning and maintenance operations and the methods 
of disassembling and reassembling equipment as a part of cleaning and 
maintenance. 
(iv) Removal or obliteration of previous batch identifi cation. 
(v) Protection of clean equipment from contamination prior to use. 
(vi) Inspection of equipment for cleanliness immediately before use. 

CURRENT GOOD MANUFACTURING PRACTICE 9 
(vii) Records should be kept of maintenance, cleaning, sanitizing, and inspection 
of all processing equipment. 
4. Automatic, mechanical, and electronic equipment 
(a) All such equipment, including computers or related systems that will perform 
a function to be used in any GMP - related activity, must be routinely calibrated, 
inspected, or checked according to a written program designed to 
assure proper performance. Written records must be maintained for all such 
activities. 
(b) Appropriate controls should be exercised to assure that changes in master 
production and control records or other similar records are made only by 
authorized personnel. Input to and output from such systems should be 
checked for accuracy. 
A backup fi le of data entered into a computer - related system must be 
maintained except where certain data such as calculations performed in connection 
with laboratory analysis are eliminated by computerization or other 
automated processes. In this situation, a written record of the program should 
be maintained along with validation data. 
5. Filters for liquid fi ltration used as a part of the manufacture, processing, or 
packing of injectable drug products intended for human use must not release 
fi bers into such products. Fiber - releasing fi lters may not be used unless it is not 
possible to manufacture the product without the use of such a fi lter. In this situation, 
an additional non - fi ber - releasing fi lter of 0.22 . m maximum must be used 
after the fi ber - releasing fi ltration. Use of an asbestos - containing fi lter is permissible 
only upon submission of proof to the appropriate FDA bureau that use of 
a non - fi ber - releasing fi lter will compromise the safety or effectiveness of the drug 
product. 
Control of Components and Drug Product Containers and Closures 
1. General requirements 
(a) There must be written procedures describing in suffi cient detail the receipt, 
identifi cation, storage, handling, sampling, testing, and approval or rejection 
of product components, containers, and closures. Of course, all such procedures 
must be followed. It is quite common and even more embarrassing to 
be cited for not following your own written procedures. Note: For the rest of 
this discussion, the term components will mean product ingredients, containers, 
closures, and so on. 
(b) All components listed above must be handled and stored in a manner that 
will prevent contamination. 
(c) Bagged or boxed components should be stored off the fl oor. Spacing should 
allow cleaning and inspection. 
(d) Every container of components must be identifi ed with a distinctive code or 
lot number for each receival of that product. Even if the next receival is the 
same vendor lot number, it must be a new identifying number by the pharmaceutical 
manufacturer. Each lot must be appropriately identifi ed as to its 
status (quarantined, approved, or rejected). 

10 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
2. Receipt and storage of untested components 
(a) Upon receipt each container of components must be visually examined for 
appropriate labeling and any damage or contamination to the component 
container. 
(b) Components must be stored under quarantine until they have been tested as 
appropriate and released for use. 
3. Testing and approval or rejection of components 
(a) Each lot of components shall be withheld from use until it has been sampled, 
tested, and released by the quality control unit. 
(b) Representative samples must be taken from every receival of every component. 
The number or amount of component to be sampled should be based 
on component appearance, statistical confi dence levels, the past history of 
the supplier, and the quantity needed to analyze and reserve samples if 
required. 
(c) Sampling procedures 
(i) The component containers should be cleaned where necessary. 
(ii) The containers should be opened, sampled, and resealed in a manner 
designed to prevent contamination of the sample and remaining contents 
of the container. 
(iii) If appropriate, sterile equipment and aseptic sampling techniques should 
be used. 
(iv) Where sampling is done from various parts of a container, samples 
should not be composited for testing. 
(v) Containers from which samples have been taken must be marked to 
show that samples have been removed. 
(d) Examination and testing of samples 
(i) At least one test should be conducted on each lot of component drug 
product to verify identity. 
(ii) Each component must be tested for conformity with all appropriate 
written specifi cations for purity, strength, and quality if an ingredient or 
for conformity with written specifi cations for containers or closures. 
(iii) In lieu of the above testing by the manufacturer, a report of analysis 
may be accepted from the supplier provided that at least one specifi c 
identity test is conducted on the component by the manufacturer and 
provided that the manufacturer has established the reliability of the 
supplier ’ s analyses through appropriate validation. 
(iv) When appropriate, components should be examined microscopically. 
(v) Each lot of a component that is liable to contamination with dirt, insect 
infestation, or other extraneous adulterant should be examined against 
established specifi cations for such contamination. 
(vi) Each lot of a component that is subject to microbial contamination that 
is contrary to its intended use should be subjected to microbiological 
tests before use. 
(e) If a lot of components meets the written specifi cations, it may be approved 
and released for use. Any lot of such material that does not meet such speci- 
fi cations must be rejected. 
4. Use of approved components (including drug product containers and closures) 
must be rotated to assure that the oldest approved stock is used fi rst. 

CURRENT GOOD MANUFACTURING PRACTICE 11 
5. Components must be retested and/or reexamined after storage for a long period 
of time or after exposure to the atmosphere, heat, or other condition that might 
adversely affect the component. 
6. Rejected components should be identifi ed and controlled under a quarantine 
system designed to prevent their use in manufacturing or processing. 
7. Containers and closures 
(a) Containers and closures must not be reactive, additive, or absorbent so as to 
alter the drug beyond established acceptance criteria. 
(b) Container closure systems must provide adequate protection against foreseeable 
external factors in storage that can cause deterioration or contamination 
of the product. 
(c) Containers and closures should be clean and, if necessary, sterile and processed 
to remove pyrogens. 
(d) Standards or specifi cation, methods of testing, and, if appropriate, sterilization 
and depyrogenation must be written and followed. 
Production and Process Controls 
1. Written procedures and procedure deviations 
(a) Written procedures for production and process control must be written and 
followed. These procedures should be designed to assure that the drug products 
have the identity, strength, quality, and purity they are represented to 
possess. These procedures must include all requirements given below and 
must be drafted, reviewed, and approved by the affected organizational units 
and reviewed and approved by the quality control unit. 
(b) When following the above identifi ed procedures, all actions must be documented 
at the time of performance. Any deviations from the written procedure 
must be recorded and justifi ed. 
2. Charge - in of components — Written production and control procedures must 
include the following, which are designed to assure that the drug products produced 
meet all specifi cations and standards. 
(a) The batch must be formulated with the intent to provide not less than 100% 
of the labeled amount of active ingredient. 
(b) Components used must be weighed, measured, or subdivided appropriately. 
If a component is removed from its original container and placed in 
another, the new container should be identifi ed with the following 
information: 
(i) Component name and/or item code. 
(ii) Receiving or control number. 
(iii) Weight or measure of material in the new container. 
(iv) Batch or lot number for which the component was dispensed, including 
its product name, strength, and lot number. 
(c) Weighing, measuring, or subdividing operations for all components must be 
adequately supervised. Each container of component dispensed to manufacturing 
must be examined by a second person to assure that: 
(i) The component was released by the quality control unit. 
(ii) The weight or measure is correct as stated in the batch production 
records. 

12 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
(iii) The containers are properly identifi ed and contain the quantity stated 
on the label. 
(d) Addition of each component must be performed by one person and verifi ed 
by a second person. 
3. Actual yield and percentage of theoretical yield should be determined at the 
completion of each appropriate phase of manufacturing, processing, packaging, 
or holding. These calculations should be performed by one person and independently 
verifi ed by a second individual. 
4. Equipment identifi cation 
(a) All compounding and storage containers, processing lines, and major equipment 
used during the production of a batch of a drug product must be properly 
identifi ed at all times to indicate their contents and the phase of processing 
of the batch. 
(b) Major equipment should be identifi ed by a distinctive identifi cation that shall 
be recorded in the batch production record to indicate the specifi c equipment 
used. In cases where only one of a particular type of equipment exists in a 
given manufacturing facility, the name of the equipment may be used instead 
of creating a distinctive identifi cation. 
5. Sampling and testing of in - process materials and drug products 
(a) To assure batch uniformity and integrity, it is necessary to write and follow 
procedures that describe the in - process controls and tests or examinations 
that will be conducted on samples taken according to procedure. Procedures 
should be written to monitor the output and to validate the performance of 
those manufacturing processes that may be responsible for causing variability 
in the product being manufactured. These control procedures should 
include but are not limited to the following: 
(i) Tablet or capsule weight variation. 
(ii) Disintegraton time. 
(iii) Adequacy of mixing or blending to assure uniformity and 
homogeneity. 
(iv) Dissolution time and rate. 
(v) Clarity of solutions. 
(vi) pH of solutions. 
(b) In - process specifi cations for all characteristics must be consistent with the 
drug product fi nal specifi cations and must be developed from previous 
acceptable product average and process variability data. 
(c) In - process materials should be tested for identity, strength, quality, and purity 
as appropriate. As a part of the production process, they must be approved 
for continued use or rejected by the quality control unit before production 
continues. 
(d) Rejected in - process materials must be identifi ed and controlled under a 
quarantine system designed to prevent their use in manufacturing operations 
for which they have been found to be unsuitable. 
6. When appropriate, time limits should be established for the completion of each 
phase of production. The purpose of this is to assure the quality of the drug 
product. Deviation from the established time limits may be acceptable if this 
deviation does not compromise the quality of the product. Any deviation must 
be documented, including the justifi cation for such deviation. 

CURRENT GOOD MANUFACTURING PRACTICE 13 
7. Control of microbial contamination 
(a) To prevent the growth of objectionable microorganisms in products not 
required to be sterile, appropriate written procedures designed to prevent 
such growth should be written and followed. 
(b) If sterilization is a part of any procedure described in (a) above, this procedure 
must be validated. 
8. Reprocessing 
(a) Written procedures describing any system used to reprocess batches that do 
not conform to the established standards must be written and followed. 
(b) Reprocessing must not be performed without the review and approval of the 
quality control unit. 
Packaging and Labeling Control 
1. Materials examination and usage criteria 
(a) Written procedures describing in detail the receipt, identifi cation, storage, 
handling, sampling, examination, and/or testing of labeling and packaging 
materials must be developed, approved, and followed. These materials must 
be representatively sampled, examined, or tested on receipt and accepted by 
the quality control unit before use. 
(b) Any materials that do not fully meet acceptance criteria must be rejected to 
prevent their use. 
(c) Records of each receival of each different label and packaging material must 
be maintained indicating receipt, examination or testing, and whether 
accepted or rejected. 
(d) Labels and other labeling materials for each different drug product, strength, 
dosage form, or quantity of contents must be stored separately with suitable 
identifi cation. Access to the storage area must be limited to authorized 
personnel. 
(e) Obsolete and outdated labels, labeling, and other packaging materials must 
be quarantined and destroyed. 
(f) The use of gang - printed labels for different drug products or different 
strengths or different net contents is prohibited. The only exception to this 
rule is if labels from gang - printed sheets are adequately differentiated by 
size, shape, or color that will prevent mixing of labels. 
(g) If cut labeling is used, packaging and labeling operations must include one 
or more of the following special control procedures: 
(i) Dedication of a labeling and packaging line to each different strength 
of each different drug product. 
(ii) Use of appropriate electronic or electromechanical equipment to 
conduct a 100% examination for correct labeling during or after completion 
of the fi nishing operation. 
(iii) Use of visual inspection to conduct a 100% examination for correct 
labeling. If visual inspection is used, the inspection should be performed 
by one person and independently verifi ed by a second individual. 
(h) Printing devices on or associated with the manufacturing line used to imprint 
labeling upon the drug product unit label or case must be monitored to assure 

14 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
that the printing conforms to the print specifi ed in the batch production 
record. 
2. Issuance of labeling 
(a) Strict control should be exercised over the issuance of labeling for use in 
drug product labeling operations. 
(b) Labeling materials issued for a batch must be carefully examined for identity 
and conformity to the labeling specifi ed in the batch production record. 
(c) Procedures should be written and followed for reconciliation of the quantities 
of labeling issued, used, destroyed, and returned. Procedures should 
require evaluation of discrepancies found between the number of packages 
fi nished and the amount of labeling issued if discrepancies outside narrow 
preset limits occur. Limits should be established on the basis of historical 
operating data. Labeling reconciliation is waived for either cut or roll labeling 
if a 100% examination for correct labeling is performed. 
(d) All excess labeling bearing a lot or control number must be destroyed. 
(e) Returned labeling should be maintained and stored in a manner to prevent 
mix - ups. 
(f) Written procedures should describe the control procedures used for the issuance 
of labeling. 
3. There must be written procedures designed to assure that correct labels, labeling, 
and packaging materials are used. These procedures should incorporate the following 
features: 
(a) Prevention of mix - ups and cross - contamination by physical or spatial separation 
of operations on other drug products. 
(b) Identifi cation and handling of fi lled drug product containers that are set 
aside and held in unlabeled condition for future labeling operations. Such 
procedures should be designed to prevent mislabeling individual containers, 
lots, or portions of lots. It is not necessary to apply identifi cation to each 
individual container, but the procedure should be adequate to determine the 
name, strength, quantity of contents, and lot or control number of each 
container. 
(c) Identifi cation of the drug product with a lot or control number that permits 
determination of the history of the manufacture and control of the batch. 
(d) Examination of packaging and labeling materials for suitability and correctness 
before issuing for use and before packaging operations. These examinations 
must be documented in the batch production record. 
(e) Inspection of the packaging and labeling facility immediately before use to 
assure that all drug products and labeling materials from the previous operation 
have been removed. Inspection results must be documented in the batch 
production record. 
4. Tamper - evident packaging requirements for OTC human drug products 
(a) An OTC product (with the exception of a dermatological, dentifrice, insulin, 
or lozenge product) intended for retail sale is considered adulterated or 
misbranded or both if it is not packaged in a tamper - resistant package. 
(b) Requirements for a tamper - evident package 
(i) With the exceptions listed above, all OTC products must be packaged 
in a tamper - evident package if the product is accessible to the 
public while being held for sale. A tamper - evident package must have 

CURRENT GOOD MANUFACTURING PRACTICE 15 
one or more indicators or barriers to entry which, if breached or missing, 
can reasonably be expected to provide visible evidence to consumers 
that tampering has occurred: A tamper - evident package may involve an 
immediate container and closure system or a secondary container or 
carton system or a combination of systems intended to provide a visual 
indication of package integrity. The tamper - evident feature must be 
designed to and shall remain intact when handled in a reasonable 
manner during manufacture, distribution, and retail display. 
(ii) In addition to the tamper - evident packaging feature described above, 
any two - piece hard gelatin capsule covered by this regulation must be 
produced using an acceptable tamper - evident technology. 
(c) Labeling 
(i) In order to alert consumers to the specifi c tamper - evident features used, 
each retained package of an OTC drug product covered by this regulation 
is required to bear a statement that: 
(1) Identifi es all tamper - evident features and any capsule - sealing 
technologies. 
(2) Is prominently placed on the package. 
(3) Is so placed that it will be unaffected if the tamper - evident feature 
of the package is breached or missing. 
(ii) If the tamper - evident feature chosen to meet the requirement uses an 
identifying characteristic, that characteristic is required to be referred 
to in the labeling statement. For example, the labeling statement on a 
bottle with a shrink band could say For your protection, this bottle has 
an imprinted seal around the neck . 
(d) A manufacturer or packer may request an exemption from the tamper - 
evident requirement. A request for exemption is required to be submitted in 
the form of a petition and should be clearly identifi ed on the envelope as a 
“ Request for Exemption from the Tamper - Evident Packaging Rule. ” This 
petition is required to contain the following: 
(i) The name of the drug product or, if the petition seeks an exemption for 
a drug class, the name of the drug class and a list of products within that 
class. 
(ii) The reasons that the drug product ’ s compliance with the tamper - evident 
packaging and labeling requirements is unnecessary or cannot be 
achieved. 
(iii) A description of alternative steps that are available or that the petitioner 
has already taken to reduce the likelihood that the product or 
drug class will be the subject of malicious adulteration. 
(iv) Other information justifying an exemption. 
(e) Holders of approved new drug applications for OTC drug products are 
required to provide the FDA with notifi cation of changes in packaging and 
labeling to comply with the requirements of this section. Changes in packaging 
and labeling required by the regulation may be made before FDA 
approval. Manufacturing changes by which capsules are to be sealed require 
prior FDA approval. 
(f) This section does not affect any requirements for “ special packaging ” as 
required under the Poison Prevention Packaging Act of 1970. 

16 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
5. Drug product inspection 
(a) Packaged and labeled products must be examined during fi nishing operations 
to provide assurance that containers and packages in the lot have the 
correct label. 
(b) A representative sample of units should be collected at the completion of 
fi nishing operations and should be visually examined for correct labeling. 
(c) Results of these examinations must be recorded in the batch production 
records. 
6. Expiration dating 
(a) All packaged drug products must carry an expiration date that has been 
determined from appropriate stability testing. 
(b) Expiration dates must be related to the recommended storage conditions 
stated on the label as determined by stability studies. 
(c) If the drug product is to be reconstituted at the time of dispensing, its label 
must carry expiration information for both the reconstituted and unreconstituted 
forms. 
(d) Expiration dates must appear on labeling in accordance with the requirements 
stated elsewhere in this regulation. 
(e) Homeopathic drug products are exempt from the requirements of this 
section. 
(f) Allergenic extracts that are labeled “ No U.S. Standard of Potency ” are 
exempt. 
(g) New drug products for investigational use are exempt provided that they 
meet appropriate standards or specifi cations as demonstrated by stability 
studies during their use in clinical investigations. If new drug products for 
investigational use are to be reconstituted at the time of dispensing, their 
labeling must bear expiration information for the reconstituted product. 
(h) Pending consideration of a proposed exemption published in the Federal 
Register , September 29, 1978, the requirements in this section will not be 
enforced for human drug products if their labeling does not bear dosage 
limitations and they are stable at least three years as supported by stability 
data. 
Holding and Distribution 
1. Warehousing procedures 
(a) Written procedures describing the warehousing of drug products must be 
written and followed. These procedures should include: 
(i) Quarantine of drug products before release by the quality control 
unit. 
(ii) Storage of drug products under appropriate conditions of temperature, 
humidity, and light so that the quality of the drug products is not affected. 
2. Distribution procedures 
(a) Written procedures concerning the distribution of drug products must be 
established and followed. These procedures should include: 
(i) A procedure that assures the distribution of the oldest approved stock 
fi rst. Deviation from this procedure is acceptable if it is temporary and 
appropriate. 

CURRENT GOOD MANUFACTURING PRACTICE 17 
(ii) A system for documenting distribution so that distribution of each lot 
of drug product can be readily determined to facilitate its recall if 
required. 
Laboratory Controls 
1. General requirements 
(a) The establishment of any specifi cations, standards, sampling plans, test processes, 
or other laboratory control mechanism required by this part of the 
regulation, including any changes to the above must be drafted by the appropriate 
organizational unit and reviewed and approved by the quality control 
unit. All actions must be documented at the time of performance and any 
deviation must be recorded and justifi ed. 
(b) Laboratory controls must include the establishment of scientifi cally 
sound and appropriate specifi cations, standards, sampling plans, and test 
procedures designed to assure that all materials conform to appropriate 
standards of identity, strength, quality, and purity. Laboratory controls should 
include: 
(i) Determination of conformance to written specifi cations for the acceptance 
of each lot within each shipment of raw materials. The specifi cations 
should include a description of the sampling and testing procedures 
used. Samples must be representative and adequately identifi ed. These 
procedures must also require appropriate retesting of any material that 
is subject to deterioration. 
(ii) Determination of conformance to written specifi cations and a description 
of sampling and testing procedures for in - process materials. 
(iii) The calibration of instruments, apparatus, gauges, and recording devices 
at specifi ed intervals in accordance with an established written program 
containing specifi c directions, schedules, limits for accuracy and precision, 
and provisions for remedial action in the event that the limits are 
not met. Any such devices that do not meet the established specifi cations 
must not be used. 
2. Testing and release for distribution 
(a) Laboratory testing of each lot of drug product must be conducted to establish 
conformance to fi nal specifi cations for the product. Testing must include 
identity and strength of each active ingredient. Where sterility and/or pyrogen 
testing are required on short - lived radiopharmaceuticals, batches may be 
released prior to completion of this testing provided that such testing is 
completed as soon as possible. 
(b) Each batch of product required to be free of objectionable microorganisms 
must be tested appropriately. 
(c) All sampling and testing plans must be described in written procedures that 
include the method of sampling and the number of units to be tested. 
(d) Acceptance criteria for the sampling and testing conducted by the quality 
control unit must be adequate to assure that the batch being tested meets all 
specifi cations. Appropriate statistical quality control criteria should be used. 
The statistical quality control criteria must include acceptance levels and/or 
rejection levels. 

18 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
(e) The accuracy, sensitivity, specifi city, and reproducibility of test methods used 
must be established and documented. Validation and documentation must 
be accomplished in accordance with this regulation. 
(f) Drug products failing to meet established standards or specifi cations and any 
relevant quality control criteria must be rejected. Reprocessing may be performed, 
however, prior to acceptance and use, and reprocessed material must 
meet all standards, specifi cations, and other relevant criteria. 
3. Stability testing 
(a) There must be a written testing program designed to assess the stability 
characteristics of every drug product. The results of such testing must be used 
to determine appropriate storage conditions and expiration dates. The written 
program must include: 
(i) Sample size and test intervals based on statistical criteria for each attribute 
examined. 
(ii) Storage conditions for sampled retained for testing. 
(iii) Reliable, meaningful and specifi c test methods. 
(iv) Testing of the product in the same container - closure system as the one 
in which the product is to be marketed. 
(v) Testing of drug products for reconstitution at the time of dispensing as 
well as after they are reconstituted. 
(b) An adequate number of batches of each drug product must be tested to 
determine appropriate expiration date. A record of such data must be maintained. 
Accelerated studies, combined with basic stability information on the 
components and drug product in its container - closure system may be used 
to project a tentative expiration date that is beyond the date supported by 
shelf life studies. However, there must be stability studies conducted including 
drug product testing at appropriate intervals until the tentative expiration 
date is verifi ed. 
(c) The requirements for homeopathic drug products are as follows: 
(i) There must be a written assessment of stability based on testing or 
examination of the drug product for compatibility of the ingredients, 
and based on marketing experience with the drug product to indicate 
that there is no degradation of the product for the normal or expected 
period of use. 
(ii) Evaluation of stability must be based on the same container - closure 
system as the one in which the drug product is to be marketed. 
(d) Allergenic extracts that are labeled “ No U.S. Standard of Potency ” are exempt 
from the requirements of this section. 
4. Special testing requirements 
(a) For each batch of drug product claimed to be sterile and/or pyrogen 
free, there must be appropriate laboratory testing to establish conformance 
to this claim. The test procedures must be in writing and must be followed. 
(b) For each batch of ophthalmic ointment, there must be appropriate testing to 
determine conformance to specifi cations regarding the presence of foreign 
particles and harsh or abrasive substances. The test procedures must be in 
writing and must be followed. 
(c) For each batch of controlled - release dosage form, there must be appropriate 
laboratory testing to determine conformance to the specifi cations for the rate 

CURRENT GOOD MANUFACTURING PRACTICE 19 
of release of each active ingredient. The test procedures must be in writing 
and must be followed. 
5. Reserve samples 
(a) An identifi ed reserve sample that is representative of each lot or of each 
shipment of each active ingredient must be retained. This reserve sample 
should contain at least twice the quantity needed for all tests required to 
determine whether the active ingredient meets its established specifi cations 
with the exception of sterility and pyrogen testing. The required retention 
time is as follows: 
(i) For an active ingredient in a drug product other than those described 
in paragraphs (b) and (c) below, the reserve sample must be retained 
for one year after the expiration date of the last lot of drug product 
containing that lot of active ingredient. 
(b) For an active ingredient in a radioactive drug product except for nonradioactive 
reagent kits, the reserve sample must be retained for: 
(i) Three months after the expiration date of the last lot of the drug product 
containing that lot of active ingredient if the expiration dating period 
of the drug product is 30 days or less. 
(ii) Six months after the expiration date of the last lot of the drug product 
containing that lot of active ingredient if the expiration dating period 
of the drug product is more than 30 days. 
(c) For an active ingredient in an OTC drug product that is exempt from bearing 
an expiration date, the reserve sample must be retained for three years after 
distribution of the last lot of drug product containing that lot of active 
ingredient. 
(d) A properly identifi ed reserve sample that is representative of each batch 
of drug product must be retained and stored under conditions consistent 
with the product labeling. The reserve sample must be stored in the 
same immediate container closure system in which the drug product is 
marketed or in one that has essentially the same characteristics. The 
reserve sample consists of at least twice the quantity needed to perform 
all the required tests except those for sterility and pyrogens. Reserve samples 
from representative sample lots or batches selected by acceptable statistical 
procedures must be examined visually at least once a year for evidence of 
deterioration unless visual examination would affect the integrity of 
the reserve sample. Any evidence of reserve sample deterioration must be 
investigated. The results of the examination must be recorded and maintained 
with stability data concerning that drug product. Retention times are 
as follows: 
(i) For a drug product other than the exceptions noted above, the reserve 
sample must be retained for one year after the expiration date of the 
drug product. 
(ii) For a radioactive drug product, except for nonradioactive reagent kits, 
the retention sample must be retained for: 
(1) three months after the expiration date of the drug product if the 
expiration date is 30 days or less or 
(2) six months after the expiration date of the drug product if the expiration 
date is more than 30 days. 

20 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
(iii) For an OTC drug product that is exempt from bearing an expiration 
date, the reserve sample must be retained for three years after the batch 
of drug product is fully distributed. 
6. Animals used in testing components, in - process materials, or drug products for 
compliance with established specifi cations must be maintained and controlled in 
a manner that assures their suitability for their intended use. They must be identi- 
fi ed and adequate records must be maintained showing the history of their use. 
7. If a reasonable possibility exists that a nonpenicillin drug product has been 
exposed to cross - contamination with penicillin, the nonpenicillin drug product 
must be tested for the presence of penicillin. The drug product may not be 
marketed if a detectable level of penicillin is found when tested according to 
procedures specifi ed in “ Procedures for Detecting and Measuring Penicillin contamination 
in Drugs ” which is incorporated in the regulation by reference. 
Records and Reports 
1. General Requirements 
(a) Any production, control, or distribution record that is associated with a batch 
of a drug must be retained for at least one year after the expiration date of 
the batch OR, for OTC drug products that do not have expiration dates, three 
years after complete distribution of the batch. 
(b) Records must be retained for all components, containers, closures, and labeling 
for the same time periods shown in (a) above. 
(c) All retained records or copies of these records must be readily available for 
authorized inspection at any time in the required retention period. Records 
must be available for inspection where the activities described therein 
occurred. Photocopying or similar reproduction by investigators must be 
permitted. 
(d) Retained records may be original records or true copies such as photocopies, 
microfi lm, microfi che, or other accurate reproduction of the original. 
(e) Written records that must be retained must be maintained so that data contained 
therein can be used for evaluating the quality standards of each drug 
product to determine the need for changes in drug product specifi cations or 
manufacturing or control procedures. Such reviews should be conducted at 
least annually. Written procedures must be established and followed for these 
evaluations and must include provisions for: 
(i) A review of a representative number of batches, whether approved or 
rejected, and records associated with the batch. 
(ii) A review of complaints, recalls, returned or salvaged drug products, and 
investigations conducted under Section 211.192 of the GMP regulations 
for each drug product. 
(f) Procedures must be established to assure that the responsible offi cials of the 
fi rm are notifi ed in writing of any investigations conducted under Sections 
211.198, 211.204, or 211.208 of any recalls, reports of inspectional observations 
issued by the FDA, or any regulatory actions relating to GMP brought 
by the FDA. 
2. A written record of major equipment cleaning, maintenance (except routine 
maintenance), and use must be included in individual equipment logs that show 

CURRENT GOOD MANUFACTURING PRACTICE 21 
the date, time, product, and lot number of each batch processed. The persons 
performing and double checking the cleaning and maintenance should date and 
sign or initial the log indicating that the work was performed. Entries in the log 
must be in chronological order. 
3. Component, drug product container, closure, and labeling records must include 
the following: 
(a) The identity and quantity of each shipment of each lot of components, drug 
product containers, closures, and labeling. Also required are the identity of 
the supplier, the supplier ’ s lot number(s), the receiving code, the date of 
receipt, and name and location of the prime manufacturer if different from 
the supplier. 
(b) The results of any test or examination performed and any conclusions derived 
from these results. 
(c) An individual inventory record of each component and a reconciliation of 
the use of each lot of such component. The inventory record must contain 
suffi cient information to allow determination of any batch or lot of drug 
product associated with the use of each component. 
(d) Documentation of the examination and review of labels and labeling for 
conformance with established specifi cations. 
(e) The disposition of rejected materials. 
4. Master production and control records 
Batch production and control records should be prepared for each batch of 
drug product produced and must include complete information about the production 
and control of that batch. These records must include: 
(a) A full and complete reproduction of the appropriate master production or 
control record. The copy must be checked for accuracy, dated, and signed. 
(b) Documentation that each signifi cant step in the manufacture, processing, 
packaging, and holding of the batch was accomplished as prescribed, 
including: 
(i) Dates. 
(ii) Identity of individual major equipment used. This includes packaging 
lines. 
(iii) Complete and specifi c identifi cation of each batch of component or 
in - process material used. 
(iv) Weight and measures of components used in the course of 
processing. 
(v) In - process and laboratory control results. 
(vi) Inspection of the packaging and labeling area before and after use. 
(vii) Documentation of the actual yield and the percentage of theoretical 
yield that this represents at critical stages of processing. 
(viii) Complete labeling control records, including specimens or copies of all 
labeling used. 
(ix) A description of drug product containers and closures. 
(x) Any sampling performed. 
(xi) Identifi cation of the persons performing and directly supervising or 
checking signifi cant steps in the operation. 
(xii) Any investigations conducted. 
(xiii) Results of examinations made. 

22 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
5. All drug product production and control records, including those for packaging 
and labeling, must be reviewed and approved by the quality control unit to determine 
compliance with all established written procedures before a batch is released 
or distributed. Any unexplained discrepancy or the failure of a batch or any of 
its components to meet any of the established specifi cations must be thoroughly 
investigated. The investigation must be extended to other batches of the same 
drug product and other drug products that may have been associated with the 
specifi c fault or discrepancy. A written record of the investigation must be made 
and include the conclusions and any required follow - up. 
6. Laboratory records 
(a) Laboratory records must include complete data derived from all tests needed 
to assure compliance with established specifi cations and standards. This 
includes examinations and assays as follows: 
(i) A description of the sample received for testing with identifi cation of 
source. For example, location where the sample was obtained, quantity, 
lot number or other distinctive code, date the sample was taken, and 
the date that it was received for testing. 
(ii) A statement of each method used in the testing of the sample. The 
statement must indicate the location of data that establish that the 
methods used in the testing of the sample meet proper standards of 
accuracy and reliability as applied to the product tested. (If the method 
used is in the current revision of the U.S. Pharmacopeia (USP), National 
Formulary (NF), or other recognized standard reference or if it is 
detailed in an approved NDA, this statement will not be required.) 
(iii) A statement of the weight or measure of sample used for each test. 
(iv) A complete record of all data secured in the course of each test, including 
all graphs, charts, and spectra from laboratory instrumentation 
properly identifi ed to the specifi c component and lot tested. 
(v) A record of all calculations performed in connection with the test, 
including units of measure, conversion factors, and equivalency 
factors. 
(vi) A statement of the results of tests and how the results compare with 
established standards of identity, strength, quality, and purity for the 
component tested. 
(vii) The initials or signature of the person who performed each test and 
the date the tests were performed. 
(viii) The initials or signature of a second person showing that the original 
records have been reviewed for accuracy, completeness, and compliance 
with established standards. 
(b) Complete records must be maintained of any modifi cation of an established 
method employed in testing. These records must include the reason for the 
modifi cation and verify that the modifi cation produced results that are at 
least as accurate and reliable for the material being tested as the established 
method. 
(c) Complete records must be maintained of any testing and standardization of 
laboratory reference standards, reagents, and standard solutions. 
(d) Complete records must be maintained of the periodic calibration of laboratory 
instruments, apparatus, gauges, and recording devices. 

CURRENT GOOD MANUFACTURING PRACTICE 23 
(e) Complete records must be maintained of all stability testing performed in 
accordance with Section 211.166 of the regulation. 
7. Distribution records must contain the name and strength of the product and 
description of the dosage form, name and address of the consignee, date and 
quantity shipped, and lot or control number of drug product. For compressed 
medical gas products, distribution records are not required to contain lot or 
control numbers. 
8. Complaint fi les 
(a) Written procedures describing the handling of all written and oral complaints 
regarding a drug product must be established and followed. These procedures 
must include provisions for review by the quality control unit of any complaint 
involving the possible failure of a drug product to meet any of its 
specifi cations and a determination as to the need for an investigation. These 
procedures must include provisions for review to determine whether the 
complaint represents a serious and unexpected adverse drug experience 
which is required to be reported to the FDA. 
(b) A written record of each complaint must be maintained in a fi le designated 
for product complaints. The fi le may be maintained at another facility if the 
written records of such fi les are readily available for inspection at that other 
facility. Written reports involving a drug product must be maintained until 
at least one year after the expiration date of the drug product or one year 
after the date that the complaint was received, whichever is longer. In the 
case of certain OTC drug products lacking expiration dating because they 
meet the criteria for exemption, such written records must be maintained for 
three years after distribution of the drug product. 
(i) The written record must include the following information 
where known: the name and strength of the drug product, lot number, 
name of complainant, nature or complaint, and reply to the 
complainant. 
(ii) Where an investigation is conducted, the written record must include 
the fi ndings of the investigation and follow - up. The record or a copy 
of the record of investigation must be maintained at the location where 
the investigation occurred. 
(iii) Where an investigation is not conducted, the written record must include 
the reason that an investigation was not considered to be necessary and 
the name of the responsible person making the determination. 
Returned and Salvaged Drug Products 
1. Returned drug products — Returned drug products must be identifi ed as such 
and held. If the conditions under which returned drug products have been held, 
stored, or shipped before or during the return or the condition of the drug product, 
its container, carton, or labeling is a result of storage or shipping casts doubt on the 
safety, identity, strength, quality, or purity of the drug product, the returned drug 
product must be destroyed unless examination testing or other investigation proves 
the drug product meets appropriate standards. Records of returned drug products 
must be maintained and must include the name and label potency of the drug 

24 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
product dosage lot number, reason for the return, quantity returned, date of disposition, 
and ultimate disposition of the returned product. If the reason for a drug 
product being returned implicates associated batches, an investigation must be 
conducted. Procedures for the holding, testing, and reprocessing of returned drug 
products must be in writing and must be followed. 
2. Drug product salvaging — Drug products that have been subjected to improper 
storage conditions, including extremes in temperature, humidity, smoke, fumes, pressure, 
age, or radiation due to natural disasters, fi res, accidents, or equipment failures, 
must not be salvaged and returned to the marketplace. Whenever there is a question 
whether drug products have been subjected to such conditions, salvaging operations 
may be conducted only if there is (a) evidence from laboratory tests and assays that 
the drug products meet all applicable standards of identity, strength, quality, and 
purity and (b) evidence from inspection of the premises that the drug products and 
associated packaging were not subjected to improper storage conditions as a result 
of the disaster or accident. Organoleptic examinations are acceptable only as supplemental 
evidence that the drug products meet appropriate standards of identity, 
strength, quality, and purity. Records including name, lot number, and disposition 
must be maintained for drug products subject to this section. 
1.1.3 GUIDANCE FOR INDUSTRY: QUALITY SYSTEMS APPROACH 
TO PHARMACEUTICAL CURRENT GOOD MANUFACTURING 
PRACTICE REGULATIONS 
This guidance document was written by the FDA to help manufacturers implement 
what they consider to be modern quality systems and risk management approaches 
that will meet the requirements of the FDA ’ s GMP regulations. The guidance 
describes what the FDA considers a comprehensive quality systems (QS) model. It 
also explains how manufacturers can be in full compliance with the GMP regulations 
by implementing such quality systems. The FDA does not intend this guidance 
to place new expectations on manufacturers nor does this replace the GMPs. 
As is true with all guidance documents, this document does not establish legally 
enforceable responsibilities, but rather it describes the FDA ’ s current thinking. Thus, 
this guidance should be viewed as a set of recommendations unless a regulation is 
cited. 
The objective of this guidance is to describe a quality systems model and demonstrate 
how and where the elements of this model can fi t within the requirements 
of the CGMP regulations. The philosophy being put forward is that quality should 
be build into the product, and testing alone cannot be relied on to ensure product 
quality . 
1.1.3.1 CGMPS and the Concepts of Modern Quality Systems 
The FDA believes that several key concepts are critical for any discussion of modern 
quality systems. The following concepts are used throughout this guidance as they 
relate to the manufacture of pharmaceutical dosage forms: 

CURRENT GOOD MANUFACTURING PRACTICE 25 
Quality For the purposes of this guidance, the phrase achieving quality means 
achieving the identity, strength, purity, and other quality characteristics 
designed to ensure safety and effectiveness. 
Quality by Design and Product Development This means designing and developing 
a product and its associated manufacturing processes that will be used 
to ensure that the product consistently attains a predefi ned quality at the end 
of the manufacturing process. 
Quality Risk Management This component of a quality systems framework can 
help guide the setting of specifi cations and process parameters for dosage form 
manufacturing, assess and mitigate the risk of changing a process or specifi cation, 
and determine the extent of discrepancy investigations and corrective 
actions. 
Corrective and Preventative Action (CAPA) This is a regulatory concept that 
focuses on investigating, understanding, and correcting discrepancies while 
attempting to prevent their recurrence. This model separates CAPA into three 
separate concepts: 
• Remedial corrections of an identifi ed problem 
• Root cause analysis with corrective action to help understand the cause of 
the deviation and prevent recurrence of a similar problem 
• Preventative action to prevent recurrence of similar problems 
Change Control This process focuses on managing change to prevent unintended 
consequences. 
Quality Unit While the GMPs refer to a quality unit, current industry practice 
is to divide the responsibilities of this unit between two groups: 
• Quality control (QC) usually involves (a) assessing the suitability of incoming 
components and the fi nished products, (b) evaluating the performance 
of the manufacturing process, and (c) determining the acceptability of each 
batch for release and distribution 
• Quality assurance (QA) involves (a) review and approval of all procedures 
related to manufacturing and maintenance, (b) review of records, and 
(c) auditing and performing/evaluating trend analyses. 
Six - System Inspection Model The FDA ’ s instruction manual for its investigators 
is a systems - based approach to inspection consistent with this guidance. The 
FDA defi nes six interlocked systems: (1) the quality system which encompasses 
all the other systems, (2) a materials system, (3) a production system, 
(4) a packaging and labeling system, (5) a facilities and equipment system, and 
(6) a laboratory controls system. The agency believes that use of this overall 
system approach will help fi rms achieve better control. 
1.1.3.2 Quality Systems Model 
This section was written to describe a model for use in pharmaceutical manufacturing 
that can supply the controls to consistently produce a product of acceptable 
quality. The model is described by four major factors: 

26 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
• Management responsibilities 
• Resources 
• Manufacturing operations 
• Evaluation 
Management Responsibilities The FDA feels that a robust quality system model 
calls for management to play a key role in the design, implementation, and 
management of the quality system. 
Resources Suffi cient resources should be provided to create a robust quality 
system that complies with the GMP regulations. Senior management or a 
designee should be responsible for providing adequate resources. 
Facilities and Equipment The technical experts who have an understanding of 
pharmaceutical science, risk factors, and manufacturing processes related to 
the product are responsible for defi ning specifi c facility and equipment requirements. 
The equipment must be qualifi ed, calibrated, cleaned, and maintained 
to prevent contamination and product mix - ups. It is important to remember 
that the GMPs place as much emphasis on process equipment as on testing 
equipment while most quality systems focus only on testing equipment. 
Control Outsourced Operations Quality systems call for contracts with outside 
suppliers that clearly describe the materials or service, quality specifi cation 
responsibilities, and communication mechanisms. 
Manufacturing There is an overlap between the elements of a quality system 
and the GMP regulation requirements for manufacturing operations. One 
should always remember that the FDA ’ s enforcement programs and inspectional 
coverage are based on the GMPs. The FDA feels that the following 
factors are essential in a manufacturing quality system: 
1. Design, develop, and document product and processes 
2. Examine inputs 
3. Perform and monitor operations 
4. Address nonconformities 
Evaluation Activities This includes the following activities: 
1. Analyze data for trends 
2. Conduct internal audits 
3. Quality risk management 
4. Corrective action 
5. Preventative action 
6. Promote improvements 
1.1.4 GUIDANCE FOR INDUSTRY: PAT — FRAMEWORK FOR 
INNOVATIVE PHARMACEUTICAL DEVELOPMENT, 
MANUFACTURING, AND QUALITY ASSURANCE 
This guidance is intended to describe a regulatory framework that the FDA chooses 
to call process analytical technology , or PAT. It is the FDA ’ s hope that this will 

encourage the voluntary development and implementation of innovative pharmaceutical 
development, manufacturing, and quality assurance. The FDA intended this 
guidance for a broad audience in different organizational units. To a large extent, 
the guidance discusses principles with the goal of highlighting opportunities and 
developing regulatory processes that encourage innovation. 
Conventional pharmaceutical manufacturing is usually accomplished using batch 
processing with laboratory testing of samples at various stages of manufacturing to 
evaluate quality. The FDA believes that opportunities exist for improving the development, 
manufacturing, and quality assurance steps through innovation in product 
and process development, process control, and analysis. 
Typically, the pharmaceutical industry has been reluctant to try something new 
due to the fear that the new approach will not fi nd favor with the FDA. An FDA 
rejection would result in costly delays and processing revisions that industry is 
unwilling to risk. The FDA now says that this hesitancy is undesirable from a public 
health perspective and it would like to see more innovation introduced. According 
to the FDA, pharmaceutical manufacturing should be based on: 
• The design of effective and effi cient manufacturing manufacturing processes 
• Product and process specifi cations based on an understanding of how formulation 
and process factors affect product performance 
• Continuous real - time quality assurance 
• Relevant regulatory policies and procedures tailored to accommodate the most 
current level of scientifi c knowledge 
• Risk - based regulatory approaches that recognize: 
The level of scientifi c understanding of how formulation and manufacturing 
process factors affect product quality and performance 
The capability of process control strategies to prevent or mitigate the risk of 
producing a poor - quality product 
It is the intent of this guidance to facilitate progress to this state. So far, the FDA ’ s 
stated goal is not being met. FDA representatives have stated the agency ’ s concern 
about the failure of industry to rush to implement change. However, the economies 
of change continue to favor the status quo. 
1.1.4.1 PAT Framework 
Quality should be built into pharmaceutical products through a comprehensive 
understanding of: 
• Intended therapeutic objectives, patient population, route of administration, 
and pharmacokinetic characteristics of a drug 
• Chemical, physical, and biopharmaceutic characteristics of a drug 
• Design of a product and selection of product components and packaging based 
on drug attributes 
• Design of manufacturing processes using principles of engineering, material 
science, and quality assurance to ensure acceptable and reproducible product 
quality and performance throughout a product ’ s shelf life 
GUIDANCE FOR INDUSTRY 27 
0
0

28 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
Process Understanding A process is considered to be well understood when all 
critical sources of variability are identifi ed and explained, variability is managed by 
the process, and product quality attributes can be accurately and reliably predicted. 
Principles and Tools Pharmaceutical manufacturing often consists of a series of 
unit operations, each of which is intended to change certain properties of the materials 
being processed. To assure these changes are acceptable and reproducible, consideration 
should be given to the quality attributes of incoming materials and their 
acceptability for the given unit operation. Most current pharmaceutical processes 
are based on time - defi ned endpoints such as “ blend for ten minutes. ” In some cases, 
these time - defi ned endpoints do not consider the effects of physical differences in 
raw materials. Processing diffi culties can arise that result in the failure of a product 
to meet specifi cations even if the raw materials conform to established specifi cations. 
Use of PAT tools and principles can provide relevant information relating to 
physical, chemical, and biological attributes. The process understanding gained from 
this information will enable process control and optimization, address the limitation 
of the time - defi ned endpoints, and improve effi ciency. 
PAT Tools There are many tools available that enable process understanding. 
These tools, when used within a system, can provide effective and effi cient means 
for acquiring information to facilitate process understanding, continuous improvement, 
and development of risk mitigation strategies. Such tools are categorized as 
follows: 
• Multivariate tools for design, data acquisition, and analysis 
• Process analyzers 
• Process control tools 
• Continuous improvement and knowledge management tools 
Strategy for Implementation To enable successful implementation of PAT, fl exibility, 
coordination, and communication with manufacturers are critical. The FDA 
believes that current regulations are suffi ciently broad to accommodate these strategies. 
In the course of implementing the PAT framework, manufacturers may want 
to evaluate the suitability of a tool on experimental and/or production equipment 
and processes. It is recommended that risk analysis of the impact on product quality 
be conducted before installation. This can be accomplished within the facility ’ s 
quality system without prior notifi cation to the agency. Data collected using an 
experimental tool should be considered research data. If conducted in a production 
facility, it should be done under the facility ’ s quality system. The FDA does not 
intend to inspect research data collected on an existing product for the purpose of 
evaluating the suitability of an experimental PAT tool. Its routine inspection of a 
fi rm ’ s manufacturing process that incorporates a PAT tool for research purposes 
will be based on current regulatory standards. 
The FDA has posted much of the information that fi rms will need in order to 
implement a PAT program on the Web at http://www.fda.gov/cder/ops/pat.htm . 
All marketing applications, amendments, or supplements to an application should 
be submitted to the appropriate Center for Drug Evaluation and Research (CDER) 
or Center for Veterinary Medicine (CVM) division in the usual manner. In general, 

PAT implementation plans should be risk based. The FDA has suggested the following 
possible implementation plans, where appropriate: 
• PAT can be implemented under the facility ’ s own quality system. CGMP inspections 
by the PAT team or PAT - certifi ed investigator can precede or follow PAT 
implementation. 
• A supplement [Changes Being Expected (CBE), Changes Being Expected in 30 
Days (CBE - 30), or Prior Approval Supplement (PAS)] can be submitted to the 
agency prior to implementation, and, if necessary, an inspection can be performed 
by a PAT team or PAT certifi ed investigator before implementation. 
• A comparability protocol can be submitted to the agency outlining PAT research, 
validation and implementation strategies, and time lines. Following approval of 
this comparability protocol by the agency, one or a combination of the above 
regulatory pathways can be adopted for implementation. 
To facilitate adoption or approval of a PAT process, manufacturers may request 
a preoperational review of a PAT manufacturing facility and process by the PAT 
team by contacting the FDA Process Analytical Technology Team at PAT@cder.fda. 
gov . It should be noted that when certain PAT implementation plans neither affect 
the current process nor require a change in specifi cations, several options can be 
considered. Manufacturers should evaluate and discuss with the agency the most 
appropriate option for their situation. 
1.1.5 GUIDANCE FOR INDUSTRY: PART 11. ELECTRONIC RECORDS; 
ELECTRONIC SIGNATURES — SCOPE AND APPLICATION 
Of the many regulations written by the FDA, the least understood is undoubtedly 21 
CFR Part 11. Rather than review the regulation itself, which is under review and possible 
revision, we will review the guidance for industry that FDA published in August 
2003 to “ aid ” industry in their puzzlement. Depending on the source, it appears to be 
questionable as to whether this guidance document aids or confuses. It exists, however, 
and like it or not, understand it or not, the regulation must be followed. 
The guidance indicates that the FDA ’ s approach is based on three main 
components: 
• The regulation will be interpreted narrowly. Fewer records will be considered 
subject to Part 11. 
• Those records that are considered subject to Part 11 will be subject to enforcement 
discretion with regard to the requirements for validation, audit trails, 
record retention, and record copying in the manner described and with regard 
to all Part 11 requirements for systems that were operational before the effective 
date of this regulation. 
• All predicate rule requirements will be enforced. This includes record and 
record - keeping requirements. 
The FDA does intend to enforce all other provisions of Part 11, including certain controls 
for closed systems. The following controls and requirements will be enforced: 
GUIDANCE FOR INDUSTRY 29

30 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
• Limiting system access to authorized individuals 
• Use of operational system checks 
• Use of authority checks 
• Use of device checks 
• Determination that persons who develop, maintain, or use electronic systems 
have the education, training, and experience to perform their assigned tasks 
• Establishment of and adherence to written policies that hold individuals 
accountable for actions initiated under their electronic signatures 
• Appropriate controls over systems documentation 
• Controls for open systems corresponding to controls for closed systems 
• Requirements related to electronic signatures 
Part 11 Records Under the narrow interpretation, the FDA considers Part 11 to 
be applicable to the following records or signatures in electronic format: 
1. Records that are required to be maintained under predicate rule requirements 
and that are maintained in electronic format in place of paper format. 
2. Records that are required to be maintainer under predicate rules, that are 
maintained in electronic format in addition to paper format, and that are relied 
on to perform regulated activities. 
3. Records submitted to the FDA under predicate rules in electronic format. 
However, a record that is not itself submitted but is used in generating a submission 
is not a Part 11 record. 
4. Electronic signatures that are intended to be the equivalent of handwritten 
signatures, initials, and other general signings required. 
FDA ’s Approach to Specifi c Part 11 Requirements 
1. Validation With respect to validation, the agency intends to exercise enforcement 
discretion regarding specifi c Part 11 requirements. However, compliance 
with all applicable predicate rules for validation is still expected. The FDA 
suggests an approach to validation be based on a justifi ed and documented 
risk assessment and a determination of the potential of the system to affect 
product quality, safety, and record integrity. 
2. Audit Trail The agency also intends to exercise enforcement discretion 
regarding specifi c requirements related to computer - generated, time - stamped 
audit trails and any corresponding requirements in Part 11. Compliance with 
all applicable predicate rule requirements related to documentation of date, 
time, or sequencing of events is still expected. It is also required to comply 
with rules for ensuring that changes to records do not obscure previous 
entries. 
3. Legacy Systems The FDA intends to exercise enforcement discretion with 
respect to all Part 11 requirements for systems that otherwise were operational 
prior to August 20, 1997. Thus they do not intend to take enforcement action 
to enforce compliance with any Part 11 requirements if all of the following 
criteria are met for a specifi c system: 

• The system was operational before the effective date. 
• The system met all applicable predicate rule requirements before the effective 
date. 
• The system currently meets all applicable predicate rule requirements. 
• There is documented evidence and justifi cation that the system is fi t for its 
intended use. 
4. Copies of Records Enforcement discretion will be applied with respect to 
specifi c Part 11 requirements for generating copies of records and any corresponding 
requirements in this part. An investigator should be provided with 
reasonable and useful access to records during an inspection. All records held 
by a manufacturer are subject to inspection. 
5. Record Retention The FDA intends to exercise enforcement discretion with 
regard to the Part 11 requirements for the protection of records to enable their 
accurate and ready retrieval at any time throughout the records retention 
period. 
1.1.6 GUIDANCE FOR INDUSTRY AND FDA : CURRENT GOOD 
MANUFACTURING PRACTICE FOR COMBINATION PRODUCTS 
This document discusses the applicability of GMPs to combination products as 
defi ned under 21 CFR 3.2(e). Manufacturers must ensure that the product is not 
adulterated; the product possesses adequate strength, quality, identity, and purity; 
and the product complies with performance standards as appropriate. This guidance 
does not address technical manufacturing methods or make recommendations for 
manufacturers ’ selection of facilities used in manufacturing. 
A combination product is a product composed of a drug and a device, a biological 
product and a device, a drug and a biological product, or a drug, a device, and a 
biological product. For the purposes of this document, a constituent part of a combination 
product is an article in a combination product that can be distinguished by 
its regulatory identity as a drug, device, or biological product. 
For regulatory purposes, a combination product is assigned to an agency center 
or alternative organizational component that will have primary jurisdiction for its 
premarket review and regulation. Manufacturers will be required to use the applicable 
GMP for their products. Regulations that may apply are: 
• GMP regulations for fi nished pharmaceuticals (21 CFR Parts 210 and 211). 
• Quality system regulations for devices (21 CFR Part 820). 
• The biological product regulations (21 CFR Parts 600 – 680) may also apply to 
the manufacture of drugs that are also biological products along with the drug 
provisions. 
There are no GMP regulations specifi cally for combination products. Until such 
regulations are promulgated, the manufacture of each constituent part is governed 
by the regulations for that component. 
The Offi ce of Combination Products is available as a resource to sponsors 
throughout the lifecycle of a combination product. This offi ce can be reached at 
GUIDANCE FOR INDUSTRY AND FDA 31

32 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
(301) 427 - 1934 or by E - mail at combination@fda.cov . Updated guidance documents 
are available at the offi ce ’ s Internet website, http://www.fda/gov/oc/combination . 
1.1.7 GUIDANCE FOR INDUSTRY: POWDER BLENDS AND FINISHED 
DOSAGE UNITS — STRATIFIED IN - PROCESS DOSAGE UNIT SAMPLING 
AND ASSESSMENT 
This guidance is intended to assist manufacturers in meeting the GMP requirements 
for demonstrating the adequacy of mixing to ensure uniformity of in - process powder 
blends and fi nished dosage units. 
Stratifi ed Sampling In this process dosage units are sampled at predefi ned intervals 
and representative samples collected from specifi cally targeted locations in the 
compression/fi lling operations that have the greatest potential to yield extremes of 
drug concentration. 
This guidance describes methods of sampling that might be used to demonstrate 
active ingredient homogeneity. These methods are put forward as suggestions and 
are not intended to be the only methods for meeting FDA requirements for demonstration 
of the adequacy of a powder mix. 
Assessment of Powder Mix Uniformity The following procedures are 
recommended: 
1. Conduct blend analysis on batches by extensively sampling the mix in the 
blender and/or intermediate bulk containers. 
2. Identify appropriate blending time and speed ranges, dead spots in blenders, 
and locations of segregation in intermediate bulk containers (IBCs). 
3. Defi ne the effects of sample size (1 – 10 times the dosage unit range) while 
developing a technique capable of measuring the true uniformity of the blend. 
Sample quantities larger than 3 times the dosage size can be used with adequate 
scientifi c justifi cation. 
4. Design blend - sampling plans and evaluate them using appropriate statistical 
analyses. 
5. Quantitatively measure any variability that is present among the samples. 
Attribute the sample variability to either lack of uniformity of the blend or 
sampling error. Signifi cant variances in the blend data within a given location 
can be an indication of one factor or a combination of factors such as inadequacy 
of blend mix, sampling error, or agglomeration. Signifi cant between - 
location variance can indicate that the blending operation is inadequate. 
Correlation of Powder Mix Uniformity with Stratifi ed In -Process Dosage Unit 
Data The following steps are recommended for correlation: 
1. Conduct periodic sampling and testing of the in - process dosage units by sampling 
them at defi ned intervals and locations throughout the compression or 
fi lling process. Use a minimum of 20 appropriately spaced in - process dosage 

unit sampling points. There should be at least 7 samples taken from each of 
these locations for a total minimum of at least 140 samples. 
2. Take 7 samples from each additional location to further assess each signifi cant 
event, such as fi lling or emptying of hoppers and IBCs, start and end of the 
compression or fi lling process, and equipment shutdown. This may be accomplished 
by using process development batches, validation batches, or routine 
manufacturing batches for approved products. 
3. Signifi cant events may also include observations or changes from one batch 
to another (e.g., batch scale - up and observations of undesirable trends in previous 
batch data). 
4. Prepare a summary of the data and analysis used to correlate the stratifi ed 
sampling locations with signifi cant events in the blending process. 
5. Compare the powder mix uniformity with the in - process dosage unit data 
described above. 
6. Investigate any discrepancies observed between powder mix and dosage 
unit data and establish root causes. At least one troubleshooting guide is 
available that may be helpful with this task. Possible corrections may range 
from going back to formulation development to improve powder characteristics 
to process optimization. Sampling problems may also be negated by use 
of alternate state - of - the - art methods of in situ real - time sampling and 
analysis. 
Correlation of Stratifi ed In -Process Samples with Finished Product The following 
steps are recommended: 
1. Conduct testing for uniform content of the fi nished product using an appropriate 
procedure or as specifi ed in the ANDA or the NDA for approved 
products. 
2. Compare the results of stratifi ed in - process dosage unit analysis with uniform 
content of the fi nished dosage units from the previous step. This analysis 
should be done without weight correction. 
3. Prepare a summary of the data and analysis used to conclude that the stratifi ed 
in - process sampling provides assurance of uniform content of the fi nished 
product. 
1.1.7.1 Validation of Batch Powder Mix Homogeneity 
This section describes sampling and testing the powder mix of demonstration and 
process validation batches used to support implementing the stratifi ed sampling 
method described in this guidance. 
The guidance document recommends that during the manufacture of demonstration 
and process validation batches, the following uniformity characteristics be 
assessed: (1) the powder blend, (2) the in - process dosage units, and (3) the fi nished 
product. Each attribute should be determined independently. It is further recommended 
that the following steps be used to identify sampling locations and acceptance 
criteria prior to the manufacture of the exhibit and/or validation batches: 
GUIDANCE FOR INDUSTRY AND FDA 33

34 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
1. Carefully identify at least 10 sampling locations in the blender to represent 
potential areas of poor blending. For example, in tumbling blenders (such as 
V - blenders, double cones, or drum mixers), samples should be selected from 
at least two depths along the axis of the blender. For convective blenders (such 
as a ribbon blender), a special effort should be made to implement uniform 
volumetric sampling to include the corners and discharge area (at least 20 
locations are recommended to adequately validate convective blenders). 
2. Collect at least three replicate samples from each location. Samples should 
meet the following criteria: 
• Assay one sample per location (number of samples n = 10, or n = 20 for 
ribbon blender). 
• RSD (relative standard deviation) of all individual results is 5.0%. 
• All individual results are within 10.0% (absolute) of the mean of the 
results. 
It is also recommended that you not proceed any further with implementation 
of the methods described in this guidance until the criteria are met. 
Sampling errors may occur in some powder blends, sampling devices, and techniques 
that make it impractical to evaluate adequacy of mix using only the blend 
data. In such cases, it is recommended that in - process dosage unit data be used in 
conjunction with blend sample data to evaluate blend uniformity. 
Some powder blends may present an unacceptable safety risk when directly 
sampled. The safety risk, once described, may justify an alternate procedure. In such 
cases, process knowledge and data from indirect sampling combined with additional 
in - process dosage unit data may be adequate to demonstrate the adequacy of the 
powder mix. Data analysis used to justify using these alternate procedures 
should be described in a summary report that is maintained at the manufacturing 
facility. 
1.1.7.2 Verifi cation of Manufacturing Criteria 
The assessment of powder mix uniformity and correlation of stratifi ed in - process 
dosage unit sampling development procedures should be completed before establishing 
the criteria and controls for routine manufacturing. It is also recommend 
that the normality be assessed and that the RSD be determined from the results of 
stratifi ed in - process dosage unit sampling and testing that were developed. The RSD 
value should be used to classify the testing results as either readily pass (RSD 4.0%) , 
marginally pass (RSD 6.0%), or inappropriate for demonstration of batch homogeneity 
when RSD > 6.0%. 
The FDA recommends that routine manufacturing batches be evaluated 
against the following criteria after completing the procedures described above to 
assess the adequacy of the powder mix and uniform content in the fi nished dosage 
form: 
1. Standard criteria method (SCM) — This method is recommended when either of 
the following conditions is met: 
1.1. Results of establishing initial criteria are classifi ed as readily pass . 

1.2. Results of testing to the marginal criteria method (MCM) pass the criteria 
for switching to the SCM. 
1.2.1. Stage 1 Test To perform the stage 1 test, collect at least three dosage 
units from each sampling location, assay one dosage unit from each 
location, weight correct the results, and compare the results with the 
following criteria: 
1.2.1.1. RSD of all individual results is less than 5%. 
1.2.1.2. Mean of all results is 90 – 110% of target assay. 
If the results pass these criteria and the adequacy of mix and uniformity 
of dosage unit content for the batch are adequate, the SCM 
can be used for the next batch. If test results fail stage 1 criteria, 
extended testing to stage 2 is required. 
1.2.2. Stage 2 Test To perform the stage 2 test, assay the remaining two 
dosage units from stage 1 for each sampling location and compute the 
mean and RSD of data combined from both stage 1 and stage 2. 
Compare the results with the following criteria: 
1.2.2.1. For all individual results, the RSD should be less than 5.0%. 
1.2.2.2. Mean of all results is 90 – 110% of target assay. 
If results pass the above criteria, the adequacy of mix and uniformity 
of content for the batch are adequate and stage 1 can be used for 
the next batch. If test results fail the criteria, use the MCM described 
in the section below. 
2. Marginal criteria method — The MCM can be used when either of the following 
conditions is met: 
2.1. Results of initial criteria establishment qualifi ed as marginally pass . 
2.2. Results of initial criteria establishment qualifi ed as readily pass or a batch 
was tested according to SCM and the test results failed both stage 1 and 
stage 2 criteria. 
2.3. If either of the above two criteria apply, use the weight corrected 
results from the stage 2 SCM analysis and compare this with the MCM 
criteria: 
2.3.1. For al individual results, the RSD is less than 6.0% . 
2.3.2. The mean of all results is 90.0 – 110.0% of target assay. 
2.4. It is acceptable to switch to the SCM when fi ve consecutive batches pass the 
MCM criteria and result in RSD of less than 5.0%. 
1.1.8 GUIDANCE FOR INDUSTRY: IMMEDIATE - RELEASE SOLID 
ORAL DOSAGE FORMS SCALE - UP AND POSTAPPROVAL CHANGES 
( SUPAC ) — CHEMISTRY, MANUFACTURING AND CONTROLS, 
IN VITRO DISSOLUTION TESTING, AND IN VIVO 
BIOEQUIVALENCE DOCUMENTATION 
This guidance provides recommendations to NDA and ANDA sponsors who intend 
to make changes to the product during the postapproval period. Changes include 
any change in components or composition of the product, the site of manufacture, 
the scale - up/scale - down of batch size, and/or the manufacturing process and/or 
equipment of an immediate - release oral formulation. 
GUIDANCE FOR INDUSTRY 35

36 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
Changes in Components (Excipients) and Composition Changes in the amount 
or source of drug substance are not addressed by this guidance. Changes in components 
or composition that have the effect of adding a new excipient or deleting an 
excipient are defi ned at level 3 except as described below: 
1. Level 1 changes 
1.1. Level 1 changes are those that are unlikely to have any detectable impact 
on formulation quality and performance. 
1.2. Allowed changes (changes that can be made without prior FDA 
approval) are shown below. This is based on the assumption that the drug 
substance in the product is formulated to 100% of label potency. To be considered 
a level 1 change, the total additive effect of all excipient changes 
should not be more than 5% relative to the target dosage form weight. 
Excipient 
Percentage of Excipient (W/W) 
Out of Total Target Dosage 
Form Weight 
Filler ± 5 
Disintegrant 
Starch ± 3 
Other ± 1 
Binder ± 0.5 
Lubricant 
Calcium or magnesium Stearate ± 0.25 
Other ± 1 
Glidant 
Talc ± 1 
Other ± 0.1 
Film coat ± 1 
1.3. Test documentation 
1.3.1. Chemistry — Application/compendial release requirements and stability 
testing. For stability testing, one batch should be on long - term stability 
testing with data being reported in the annual report. 
1.3.2. Filing documentation — All information must be included in the annual 
report (including long - term stability data). 
2. Level 2 changes 
2.1. Level 2 changes are those that could have a signifi cant impact on formulation 
quality and performance. Tests and fi ling documentation for a level 2 change 
depend on three factors: (1) therapeutic range, (2) solubility, and (3) permeability. 
Therapeutic range is defi ned as either narrow or nonnarrow. Drug 
solubility and drug permeability are defi ned as either low or high. Changes 
in excipients, expressed as percent (w/w) of total formulation, greater than 
those listed for a level 1 change but less than or equal to the following 
percent ranges are acceptable level 2 changes: 

Excipient 
Percentage of Excipient (w/w) of Total Target 
Dosage Form Weight 
Filler ± 10 
Disintegrant 
Starch ± 6 
Other ± 2 
Binder ± 1 
Lubricant 
Ca or Mg stearate ± 0.5 
Other ± 2 
Glidant 
Talc ± 2 
Other ± 0.2 
Film coat ± 2 
These percentages are based on the assumption that the drug substance in 
the fi nished product is formulated to 100% of labeled potency. The total 
additive effect of all excipient changes should not change by more than 
10%. 
All components in the formulation should have numerical targets that 
represent the nominal composition of the product on which any future 
changes in the composition of the product are based. Allowable changes in 
the composition should be based on the approved target composition and 
not on the composition based on previous level 1 or level 2 changes. 
2.2. Test documentation 
2.2.1. Chemistry 
2.2.1.1. Application/compendial release requirements and batch 
records. 
2.2.1.2. Stability testing — Test one batch with three months of accelerated 
stability data in supplement and on batch on long - term 
stability. 
2.2.2. dissolution 
2.2.2.1. High - permeability, high - solubility drugs — Dissolution of 85% 
in 15 min in 900 mL of 0/1 N HCl. If a drug product fails to 
meet this criterion, tests in 2.2.2.2 or 2.2.2.3 below should be 
performed. 
2.2.2.2. Low - permeability, high - solubility drugs — Multipoint dissolution 
profi le should be performed in the application/compendial 
medium at 15, 30, 45, 60, and 120 min or until an asymptote 
is reached. The dissolution profi le of the proposed and currently 
used product formulations should be similar. 
2.2.2.3. High - permeability, low - solubility drugs — Multipoint dissolution 
profi les should be performed in water, 0.1 N HCl, and 
USP buffer media at pH 4.5, 6.5, and 7.5 (fi ve different pro- 
fi les) for the proposed and currently accepted formulations. 
Adequate sampling should be performed at 15, 30, 45, 60, and 
120 min until either 90% of drug from the drug product is 
GUIDANCE FOR INDUSTRY 37

38 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
dissolved or an asymptote is reached. A surfactant may be 
used, but only with appropriate justifi cation. The dissolution 
profi le of the proposed and currently used product formulations 
should be similar. 
2.2.3. In vivo bioequivalence documentation is not required for level 2. If 
the product does not meet any of the level 1 cases above, refer to level 
3 changes. 
2.2.4. Filing documentation — A prior approval supplement with all data 
including the accelerated stability data is required. This change should 
also be documented in the annual report along with the long - term 
stability data. 
2.3. Level 3 changes 
2.3.1. Level 3 changes are those that are likely to have a signifi cant impact 
on formulation quality and performance. Tests and fi ling documentation 
vary depending on the following three factors: therapeutic range, 
solubility, and permeability. For example: 
2.3.1.1. Any qualitative and quantitative excipient changes to a narrow 
therapeutic drug beyond the ranges specifi ed in the level 1 
table. 
2.3.1.2. All other drugs not meeting the dissolution criteria under level 
2. 
2.3.1.3. Changes in the excipient ranges of low - solubility, low - 
permeability drugs beyond those listed in level 1. 
2.3.1.4. Changes in the excipient ranges of all drugs beyond those 
listed in the level 2 table. 
2.3.2. Test documentation 
2.3.2.1. Chemical 
(a) Application/compendial release requirements and batch 
records: 
• Information available — One batch with three months 
accelerated stability data reported in a supplement and 
one batch on long - term stability reported in the annual 
report. 
• Information NOT available — Up to three batches with 
three months accelerated stability data reported in the 
supplement and one batch on long - term stability data 
reported in annual report. 
(b) Dissolution documentation — Case B dissolution profi le as 
described in the table for level 2. 
(c) In vivo bioequivalence documentation — Full bioequivalence 
study. This requirement may be waived with a veri- 
fi ed acceptable in vivo/in vitro correlation. 
2.3.2.2. Filing documentation — Prior approval supplement including 
accelerated stability data plus an annual report showing long - 
term stability data. 
Site Changes Site changes are changes in the location of manufacture for both 
company - owned and contract manufacturing facilities. A site change does not 
include, for example, scale - up changes, changes in manufacturing equipment or a 
manufacturing process, and changes in Standard Operating Procedures (SOPs) or 
environmental changes. Each change must be considered separately. 
1. Level 1 changes — A level 1 change consists of a site change within a single facility 
where the same equipment, SOPs, environmental conditions, and personnel are 
used and where no changes are made to the manufacturing batch records other 
than location of the facility and administrative changes. 
1.1. Required documentation — No documentation is required beyond the usual 
application/compendial requirements. No in vivo bioequivalence documentation 
is required. 
1.2. Filing requirements — Annual report. 
2. Level 2 changes — A level 2 change is a site change within a contiguous campus 
or between facilities in adjacent city blocks where the same equipment, SOPs, 
environmental conditions and controls, and personnel common to both manufacturing 
sites are used. There must be no changes to the manufacturing batch 
records except for administrative information and the location of the facility. 
2.1. Required documentation 
2.1.1. Chemistry—Identify location of new site and updated batch records. 
No other documentation is required beyond application/compendial 
release requirements, although one batch produced at the new site 
should be placed on long - term stability and the data should be reported 
in the annual report. Dissolution data other than normal release requirements 
are not required nor is in vivo bioequivalence testing required. 
2.1.2. Filing documentation — A supplement should be fi led showing the 
changes being effected. Long - term stability test data should be included 
in the annual report. 
3. Level 3 changes — A level 3 change is a change in manufacturing site to a different 
campus. However, the same equipment, SOPs, environmental conditions, and 
controls should be used in the manufacturing process at the new site. No changes 
may be made to the manufacturing batch records except for administrative information, 
location, and language translation if needed. 
3.1. Documentation 
3.1.1. Chemistry — Location of new site and updated batch records. 
3.1.2. Stability 
3.1.2.1. If a signifi cant body of data is available, one batch with three 
months accelerated stability data must be reported in a supplement. 
One batch should be on long - term stability with the 
stability data reported in the annual report. 
3.1.2.2. If a signifi cant body of data is not available, up to three batches 
with three months accelerated stability data should be reported 
in the supplement. Up to three batches should be on long - term 
stability with these data being reported in the annual report. 
3.1.3. Dissolution — A multipoint dissolution profi le should be performed in 
the application/compendial medium at 15, 30, 45, 60, and 120 min or 
until an asymptote is reached. The dissolution profi le of the drug 
product at the current and proposed site should be similar. 
3.1.4. In vivo bioequivalence — None required. 
GUIDANCE FOR INDUSTRY 39

40 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
3.2. Filing documentation required — Changes being effected should be identifi ed 
in a supplement. Long - term stability data are reported in the annual 
report. 
Changes in Batch Size Postapproval changes in the size of a batch from the pilot 
scale used to manufacture product for clinical trials to larger or smaller commercial 
batch sizes require submission of additional information in the application. Scale - 
down below 100,000 dosage units is not covered by this guidance. All scale - up 
changes should be properly validated and, where needed, inspected by appropriate 
FDA personnel. 
1. Level 1 changes — A change in batch size, up to and including a factor of 10 times 
the size of the pilot batch, is considered a level 1 change. However, (1) the equipment 
used must be of the same design and operating principles, (2) the product 
is manufactured in full compliance with the prevailing GMPs, and (3) the same 
formulation and manufacturing procedures are used as well as the same SOPs 
and controls. 
1.1. Chemistry documentation — (1) Application/compendial release requirements, 
(2) notifi cation of change to the FDA and submission of updated 
batch records in the annual report, and (3) one batch should be on long - term 
stability with results being provided in the annual report. 
1.2. Dissolution documentation — None beyond application/compendial release 
requirements. 
1.3. In vivo bioequivalence — None. 
1.4. Filing documentation — Annual report with long - term stability data. 
2. Level 2 changes — Level 2 consists of changes in batch size beyond a factor of 10 
times the size of the pilot batch where (1) the equipment used to produce the 
pilot batches is of the same design and operating principles, (2) the product is 
manufactured in full compliance with the prevailing GMPs, and (3) the same 
formulation and manufacturing procedures are used as well as the same SOPs 
and controls. 
2.1. Chemistry — Application/compendial release requirements. Notifi cation of 
change in batch size and submission of updated batch records to the FDA. 
One batch must be placed on accelerated stability testing and one on long - 
term stability. 
2.2. Dissolution — None beyond application/compendial release requirements. 
2.3. In vivo bioequivalence — None. 
2.4. Filing requirements — Must submit changes being effected in the supplement. 
Long - term stability data are reported in the annual report. 
Manufacturing Manufacturing changes may be either the equipment used in the 
manufacturing process or the process itself: 
1. Equipment 
1.1. Level 1 equipment changes — This category includes change from the use of 
nonautomated or nonmechanical equipment to automated or mechanical 
equipment to move ingredients and a change to alternative equipment of 
the same design and operating principles of the same or different capacity. 

1.1.1. Chemistry documentation — Application/compendial release requirements, 
notifi cation of change, and submission of updated batch records. 
One batch should be placed on long - term stability. 
1.1.2. Dissolution documentation — None other than application/compendial 
release requirements. 
1.1.3. In vivo bioequivalence documentation — None. 
1.1.4. Filing documentation—Annual report with long-term stability data. 
1.2. Level 2 equipment changes — This type of change involves a change in equipment 
to a different design and different operating principles. 
1.2.1. Chemistry documentation — Application/compendial release requirements, 
notifi cation of change, and submission of updated batch 
records. 
1.2.1.1. If a signifi cant body of data are available, one batch with three 
months of accelerated stability data reported in the supplement 
and one batch on long - term stability with data reported 
in the annual report. 
1.2.1.2. If a signifi cant body of data are not available, submit up to 
three batches with three months accelerated stability data in 
the supplement and up to three batches on long - term stability 
with data reported in the annual report. 
1.2.2. Dissolution documentation — A multipoint dissolution profi le should 
be performed in the application/compendial medium at 15, 30, 45, 60, 
and 120 min or until an asymptote is reached. The dissolution profi le 
of the drug product at the current and proposed site should be 
similar. 
1.2.3. In vivo bioequivalence documentation — None. 
1.2.4. Filing documentation — Prior approval supplement with justifi cation 
for change; long - term stability data must be reported in the annual 
report. 
2. Process changes 
2.1. Level 1 process changes — This includes process changes such as changes in 
mixing times and operating speeds within application/validation ranges. 
2.1.1. Chemistry documentation — None beyond application/compendial 
release requirements. 
2.1.2. Dissolution documentation — None beyond application/compendial 
release requirements. 
2.1.3. In vivo bioequivalence documentation — None. 
2.1.4. Filing documentation — Annual report. 
2.2. Level 2 process changes — Level 2 changes include process changes such as 
mixing times and operating speeds outside of application/validation ranges. 
2.2.1. Chemistry documentation — Application/compendial release requirements; 
notifi cation of change and submission of updated batch records. 
One batch on long - term stability. 
2.2.2. Dissolution documentation — A multipoint dissolution profi le should 
be performed in the application/compendial medium at 15, 30, 45, 60, 
and 120 min or until an asymptote is reached. The dissolution profi le 
of the drug product at the current and proposed site should be 
similar. 
GUIDANCE FOR INDUSTRY 41

42 GOOD MANUFACTURING PRACTICES & RELATED FDA GUIDELINES 
2.2.3. In vivo bioequivalence documentation — None. 
2.2.4. Filing documentation — A supplement with changes being effected. 
Long - term stability data should be reported in the annual report. 
2.3. Level 3 process changes — Level 3 includes change in the type of process used 
in the manufacture of the product, such as a change from wet granulation to 
direct compression. 
2.3.1. Chemistry documentation — Application/compendial release requirements. 
Notifi cation of change and submission of updated batch records. 
Stability testing varies depending on the amount of data available: 
2.3.1.1. Signifi cant body of data available — One batch with three 
months accelerated stability data should be reported in the 
supplement; one batch should also be put on long - term stability 
with data being reported in the annual report. 
2.3.1.2. No signifi cant body of data available — Up to three batches 
with three months accelerated stability data should be reported 
in the supplement. Up to three batches should be on long - term 
stability with data being reported in the annual report. 
2.3.2. Dissolution documentation — A multipoint dissolution profi le should 
be performed in the application/compendial medium at 15, 30, 45, 60, 
and 120 min or until an asymptote is reached. The dissolution profi le 
of the drug product at the current and proposed site should be 
similar. 
2.3.3. In vivo bioequivalence documentation — An in vivo bioequivalence 
study should be performed. This may be waived if a suitable in vivo/in 
vitro correlation has been verifi ed. 
2.3.4. Filing documentation — A prior approval supplement must be fi led 
with justifi cation for the change. Long - term stability data should be 
submitted in the annual report. 
1.1.9 OTHER GMP - RELATED GUIDANCE DOCUMENTS 
This chapter has discussed the CGMP regulations and some of the more important 
guidances. There have been a number of additional guidance documents related to 
GMPs published by the FDA. These documents are all posted on the FDA website. 
They are listed below along with their URL: 
• Current good manufacturing practice for combination products: http://www. 
fda.gov/cder/guidance/OCLove1dft.pdf 
• Questions and answers on current good manufacturing practices (cGMP) for 
drugs: http://www.fda.gov/cder/guidance/cGMPs/default.htm 
• Powder blends and fi nished dosage units — Stratifi ed in - process dosage unit 
sampling and assessment 
• Sterile drug products produced by aseptic processing — Current good manufacturing 
practice: http

• Current good manufacturing practice for medical gases: http://www.fda.gov/ 
cder/guidance/3823dft.pdf 
• General principles of process validation
• SUPAC - IR: Immediate - release solid oral dosage forms: Scale - up and post - 
approval changes: Chemistry, manufacturing and controls, in vitro dissolution 
testing, and in vivo bioequivalence documentation: http://www.fda.gov/cder/ 
guidance/cmc5.pdf 
• SUPAC - IR/MR: Immediate release and modifi ed release solid oral dosage 
forms manufacturing equipment addendum
• SUPAC - MR: Modifi ed release solid oral dosage forms scale - up and postapproval 
changes: Chemistry, manufacturing, and controls; in vitro dissolution 
testing and in vivo bioequivalence documentation: http://www.fda.gov/cder/ 
guidance/1214fnl.pdf 
• SUPAC - SS: Nonsterile semisolid dosage forms; scale - up and post - approval 
changes: Chemistry, manufacturing and controls; in vitro release testing and in 
vivo bioequivalence documentation
• SUPAC - SS: Nonsterile semisolid dosage forms manufacturing equipment 
addendum 
OTHER GMP-RELATED GUIDANCE DOCUMENTS 43


45 
1.2 
ENFORCEMENT OF CURRENT GOOD 
MANUFACTURING PRACTICES 
Kenneth J. Nolan 
Nolan & Auerbach, P. A., Fort Lauderdale, Florida 
Contents 
1.2.1 Introduction and Background 
1.2.2 Enforcement Players 
1.2.3 FDA Enforcement Techniques 
1.2.3.1 Inspections 
1.2.3.2 After the Inspection: Form 483 
1.2.3.3 Recalls 
1.2.3.4 Warning Letter 
1.2.4 Judicial Enforcement: Beyond the Warning Letter 
1.2.4.1 Introduction 
1.2.4.2 Civil Proceedings 
1.2.4.3 Criminal Proceedings 
1.2.5 Conclusion 
1.2.1 INTRODUCTION AND BACKGROUND 
The legal authority for the Food and Drug Administration (FDA) to impose 
minimum manufacturing standards is set forth in the federal Food and Drug and 
Cosmetic Act (FDCA), 21 U.S.C. sec. 301 et seq. Section 351(a)(2)(B) of 21 U.S.C. 
requires manufacturers of drugs to operate in conformance with manufacturing 
regulations established by the FDA. The regulations are primarily contained in Title 
21 of the U.S. Code of Federal Regulations (CFR), Parts 210 and 211, and are called 
the current good manufacturing practice (cGMP) regulations. 
The cGMP regulations stem from congressional concern that impure and otherwise 
adulterated drugs might escape detection under a system predicated only on 
seizure of drugs shown to be in fact adulterated. That is, the U.S. Congress desired 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

46 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
to require manufacturers to utilize manufacturing practices designed to prevent 
pharmaceuticals from such defects as contamination, nonconforming bioavailability, 
or potency defects. 
Congress stated the rationale for imposing cGMP on the pharmaceutical industry 
this way 1 : 
The manufacturing of drugs is a business that requires highly qualifi ed and trained 
personnel, and special laboratory and other facilities and most careful internal manufacturing, 
packaging, and labeling controls. These requirements are necessary to the 
assurance that the drugs will be safe for the user and will have, and so far as possible 
retain, the identity, strength, quality, purity, and effectiveness that they purport to 
have. 
The purpose of the cGMP requirement is to prevent injury and death “ by building 
quality into the design and production of pharmaceuticals, ” 2 so that substandard 
prescription drugs do not jeopardize the health and safety of the patients. 
The cGMPs require manufacturers to have adequately equipped manufacturing 
facilities, adequately trained personnel, precisely controlled manufacturing processes, 
appropriate laboratory controls, complete and accurate records and reports, 
appropriate fi nished product examination, and so on. Current GMPs are not “ best 
practices ” ; rather, they establish threshold or minimum standards which must be 
satisfi ed in order for a pharmaceutical manufacturing operation to be compliant. 
The cGMPs were modifi ed only once between 1963 and 2002 with changes made 
in 1978 to update them in light of the current technology and also to describe the 
requirements more explicitly and with more specifi city. Meanwhile, the intervening 
decades saw myriad advances in manufacturing science, engineering, and technology, 
including the development of better quality systems. These advances, combined with 
the desire to harmonize manufacturing standards in an increasingly globalized production 
environment, created the impetus to revamp the cGMPs again. 
In August 2002, the FDA announced a comprehensive review of the pharmaceutical 
cGMPs. The agency identifi ed its cGMP initiative “ Pharmaceutical cGMPs for 
the 21st Century: A Risk - Based Approach. ” The FDA ’ s articulated goals for the 
initiative, relevant to enforcement, were: 
• The submission review program and the inspection program operate in a coordinated 
and synergistic manner. 
• Regulation and manufacturing standards are applied consistently. 
• FDA resources are used most effectively and effi ciently to address the most 
signifi cant health risks. 
One of the major products of the cGMP initiative was issued by the FDA in 
September 2006 in a document entitled “ Guidance for Industry — Quality Systems 
Approach to Pharmaceutical cGMP Regulations. ” 3 The FDA described the guid- 
1 H. R. Rep. No. 2464, 87th Cong., 2d Sess. 2 (1962). See also 1962 U.S. Cong. and Admin. News , p. 2884. 
2 FDA, Pharmaceutical cGMPs for the 21st century: A risk - based approach, Rockville, MD, August 21, 
2002. 
3 The FDA ’ s guidance documents advise the reader that they “ do not establish legally enforceable responsibilities. 
Instead, guidances describe the [the FDA ’ s] current thinking on a topic and should be viewed 
only as recommendations, unless specifi c regulatory or statutory requirements are cited. ” 

ance as a comprehensive quality systems model which, if followed, would improve 
quality control and satisfy the requirements of the cGMP regulations. Quality 
systems and quality assurance are important parts of the cGMP modernization 
process because quality assurance problems have been the cGMP issues most frequently 
cited by FDA investigators in recent years. 
Drugs which are manufactured not in accordance with any cGMP requirement, 
including the quality control and quality process mandates, are “ adulterated ” under 
the FDCA. Section 351 of 21 U.S.C. defi nes a drug as adulterated 
[if] the methods used in, or the facilities or controls used for, its manufacture, processing, 
packing, or holding do not conform to or are not operated or administered in 
conformity with current good manufacturing practice to assure that such drug meets 
the requirement of the act as to safety and has the identity and strength, and meets the 
quality and purity characteristics, which it purports or is represented to possess. 
1.2.2 ENFORCEMENT PLAYERS 
The FDA is obviously one of the most important regulatory agencies in the United 
States. It may also be characterized as the most important consumer protection 
agency in the world. Its decisions involving approval of drugs have a direct effect 
on testing, approval, access, and distribution of prescription drugs worldwide. As a 
regulatory agency in a largely scientifi c role, it is involved in shaping pharmaceutical 
science and drug access throughout the world. As a scientifi c agency, the FDA 
employs physicians, pharmacists, biologists, biochemists, engineers, biostatisticians, 
and other highly educated and specialized professionals. 
But the FDA also has very important law enforcement responsibilities. The 
agency employs civil and criminal investigators, auditors, attorneys, and other 
enforcement professionals. One of the FDA ’ s many enforcement functions is investigation, 
remediation, and prosecution of cGMP violations. 
The FDA district offi ces operate under the auspices of the agency ’ s Offi ce of 
Regulatory Affairs (ORA). The ORA fi eld organization is divided into fi ve regional 
offi ces (northeast, central, southeast, southwest, Pacifi c). Each region includes district 
offi ces, of which there are 20 nationwide. Most district offi ces have three or 
four branches, including either a compliance branch or an enforcement branch. The 
branch offi ces are the primary regulatory contacts within the districts and act as the 
“ eyes and ears ” for FDA headquarters. 
The FDA ’ s Offi ce of Criminal Investigations (OCI) is responsible for reviewing 
allegations which if proven would violate the U.S. criminal code, including potential 
violations of the cGMPs. The OCI investigators conduct such investigations as is 
deemed appropriate, sometimes in connection with other federal investigative 
agencies, including the FBI and the Offi ce of Inspector General of the Department 
of Health and Human Services. If the OCI chooses not to recommend to the Department 
of Justice (DOJ) 4 that criminal indictment be pursued, then the district offi ce 
is at liberty to pursue the matter through administrative or civil proceedings. 
4 The DOJ is under the direction of the attorney general of the United States. Its mission, relevant to this 
chapter, is to enforce federal statutes and uphold the rule of the law. It pursues violations brought to its 
attention by the FDA as well as other federal agencies. 
ENFORCEMENT PLAYERS 47

48 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
Although the FDA ’ s Offi ce of General Counsel is involved with enforcement of 
both civil and criminal matters, cases involving court enforcement are handled by 
assistant U.S. attorneys (AUSAs), who are located in U.S. attorneys ’ offi ces located 
across the United States. U.S. attorneys are the local representatives of the DOJ; 
they are appointed by and serve at the discretion of the president, with advice 
and consent of the Senate. There are 93 U.S. attorneys, and they are located (by 
district) across the United States and its territories. Each U.S. attorney is the chief 
federal law enforcement offi cer of the United States within his or her particular 
district. 
The AUSAs are the principal trial attorneys for the U.S. government. Each U.S. 
attorney exercises wide discretion in the use of his or her resources to further the 
priorities of the local jurisdiction. Discretion and expertise are big factors in case 
decisions. There may be signifi cant disparity in the experience, interest, and capability 
of U.S. attorneys ’ offi ces with respect to their pursuit of cGMP violations. 
The impact of this disparity is mitigated or eliminated by the expertise of the 
DOJ ’ s Offi ce of Consumer Litigation (OCL), which is charged with coordinating 
and supporting FDCA prosecutions nationwide. The OCR ’ s attorneys exercise considerable 
infl uence over and discretion in deciding what to and what not to prosecute, 
thus fostering consistent prosecutive decision making. Many civil actions, 
particularly those seeking injunctive relief, cannot be brought by a U.S. attorney 
without OCL approval, minimizing the risk that an inconsistent policy position is 
taken by a U.S. attorney ’ s offi ce. 
1.2.3 FDA ENFORCEMENT TECHNIQUES 
1.2.3.1 Inspections 
The FDA has the right to conduct surveillance inspections of manufacturing facilities 
for the purpose of enforcement. The goal of inspections is “ to minimize consumers 
exposure to adulterated products. ” 5 
The FDCA, 21 U.S.C. 374, provides that the FDA is authorized to enter and “ to 
inspect, at reasonable times and within reasonable limits and in a reasonable manner 
… all pertinent equipment, fi nished and unfi nished materials, containers, and labeling 
” in the manufacturing or related facility. This statute further authorizes the 
inspection to “ extend to all things therein (including records, fi les, papers, processes, 
controls, and facilities) ” as long as the records, for example, are relevant to any 
potential adulteration or misbranding 6 or other FDCA violations. The statute denies 
the agency the right to review “ fi nancial data, sales data, pricing data, personnel data 
(other than data as to qualifi cations of technical and professional personnel performing 
functions) ” and certain other types of documents. 
Inspectors are required to notify the company that the inspection is occurring 
but need not provide their reasons. They may take samples and photographs related 
5 Compliance Program Guidance Manual for FDA Staff: Drug Manufacturing Inspection Program , 
7356.002, available: www.fda.gov . 
6 Misbranding involves labeling a pharmaceutical product in a misleading way. See 21 U.S.C. 331(k). 

FDA ENFORCEMENT TECHNIQUES 49 
to the subject of the inspection. It is a criminal offense to deny entry to FDA inspectors 
or other offi cials who have appropriately made attempts to conduct an inspection. 
[21 U.S.C. 331(f)] 
In addition to the for - cause inspections, the FDCA mandates that the FDA routinely 
inspect a manufacturer ’ s facilities for cGMP compliance every two years. 
This applies to domestic and foreign facilities which manufacture drugs for sale 
within the United States. 7 Unfortunately, this two - year mandate is rarely satisfi ed 
because the FDA ’ s district offi ces, which are charged with the responsibility for the 
inspections, lack suffi cient resources to conduct regular cGMP compliance 
inspections. 
The FDA conducts two categories of facility inspections — surveillance inspections 
and compliance inspections. Surveillance inspections are periodic. Whether 
and when to inspect a particular manufacturing facility is decided in part by application 
of an analytical model to determine high risk sites. In late 2004, the FDA issued 
a report entitled “ Risk - Based Method for Prioritizing cGMP Inspections of Pharmaceutical 
Manufacturing Sites — A Pilot Risk Ranking Model, ” which allows the 
agency to rank manufacturing plants ’ risk of noncompliance by using an analytical 
process to (1) pose a risk question, (2) identify potential hazards and risks, (3) 
characterize factors that can be used as variables for quantifying risk, and (4) mathematically 
combine the variables to yield an overall risk score. Since the publication 
of the report, the FDA has added adverse events reports data to the model. Surveillance 
inspections are supposed to involve audit coverage of two or more systems, 8 
with mandatory coverage of the quality system. 9 
Compliance inspections are for the purpose of evaluating or verifying compliance 
corrective actions after a problem has been identifi ed and regulatory action has 
been taken. Compliance inspections cover the areas found defi cient and subjected 
to corrective actions. One type of compliance inspection is a “ for - cause ” inspection, 
which is conducted to investigate a specifi c problem that has come to the attention 
of the FDA. The sources that trigger a compliance inspection include fi eld alert 
reports, industry complaints, and recalls. 
In fi scal year 2005, the FDA fi eld offi ce conducted 1437 cGMP inspections, resulting 
in 15 warning letters, six injunctions, and one seizure. These enforcement actions 
are discussed later in this chapter. Data for the years 2000 – 2005 are set forth in 
Figures 1 and 2 . 
1.2.3.2 After the Inspection: Form 483 
If the inspector determines that there are deviations from cGMP, he will complete 
a form FDA - 483 (Inspectional Observations) detailing the violations. The fi ndings 
are presented to the manufacturer, which is given an opportunity to respond. The 
FDA - 483 advises: 
7 The other cGMP basic enforcement strategy is collection and analysis of drug samples during factory 
inspections as well as collecting and analyzing drug products in distribution. 
8 The FDA has separated the cGMP regulation into six systems: quality, facilities and equipment, production, 
materials, packaging and labeling, and laboratory controls. 
9 Compliance Program Guidance Manual , 7356.002, February 1, 2002. 

50 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
This document lists observations made by the FDA representative(s) during the inspection 
of your facility. They are inspectional observations, and do not represent a fi nal 
Agency determination regarding your compliance. If you have an objection regarding 
an observation, or have implemented, or plan to implement, corrective action in 
response to an observation, you may discuss the objection or action with the FDA 
representative(s) during the inspection or submit this information to FDA at the 
address above. If you have any questions, please contact FDA at the phone number 
and address above. 
Most manufacturers provide a written response to the FDA - 483, either disputing 
the fi ndings or addressing how they will correct the issues and how problems are to 
be corrected. Negotiations typically proceed for months or years until the inspectional 
problems and issues are resolved or the FDA elects to pursue elevated 
enforcement. The agency retains discretion to pursue elevated enforcement if it 
concludes that there is a signifi cant risk of harm to patients, with such action being 
more likely where patient harm is more likely or more serious. 
In addition to providing a form FDA - 483, FDA investigators prepare an establishment 
inspection report (EIR), which is sent to FDA headquarters, which then 
evaluates the report and determines the corrective action, if any. The FDA then 
classifi es the inspection as “ no action indicated, ” “ voluntary action indicated, ” or 
“ offi cial action indicated. ” The EIR contains much greater detail than contained in 
the 483 and is not provided to the manufacturer until after the inspection is deemed 
closed. 
FIGURE 1 CDER fi ve - year Inspection data. ( Source : FDA .) 
2610 
2529 
2585 
2627 
2682 
2450 
2500 
2550 
2600 
2650 
2700 
2001 2002 2003 2004 2005 
Inspections (Foreign & Domestic) 
FIGURE 2 Surveillance activity. ( Source : FDA .) 
2529 2585 2627 2600 2682 
1982 
1712 
2087 
1434 
1548 
174 
362 
180 260 
1183 
0 
500 
1000 
1500 
2000 
2500 
3000 
2001 2002 2003 2004 2005 
Inspections 
Domestic Samples 
Import Samples

FDA ENFORCEMENT TECHNIQUES 51 
When the FDA conducted an analysis of past FDA - 483 reports, 10 the two most 
reported violations were: 
1. Violations of 21 CFR 211.100(b) (failure to follow and/or document production 
and process control procedures), occurring in over half of all of the 
483 ’ s 
2. Violations of 21 CFR 122d (failure to create adequate, written responsibilities 
and procedures for the quality control unit or failure to follow them), occurring 
in 42% of the 483 ’ s 
The next eight violations, in order of prevalence, were as follows: 
• Failure to have written procedures for production and process controls. 
• Failure to have testing and release of drug product for distribution for determination 
of satisfactory conformance to the fi nal specifi cations/identity and 
strength of each active ingredient prior to release. 
• Batch production and control records were not prepared or are incomplete. 
• Control procedures are not established to monitor the output/validate the 
performance of manufacturing processes that may be responsible for causing 
variability in the characteristics of the drug product. 
• Employees were not given appropriate training. 
• Laboratory controls do not include the establishment of scientifi cally sound 
and appropriate specifi cations/standards/sampling plans/test procedures. 
• Drug product production and control records are not certifi ed by the quality 
control unit to assure compliance with all established, approved written procedures 
before a batch is released or distributed. 
• Procedures describing the handling of all written and oral complaints regarding 
a drug product either not established or not followed. 
1.2.3.3 Recalls 
Chapter 7 of the Regulatory Procedures Manual (March 2007, available at www.fda. 
gov ) provides detailed instructions to FDA personnel regarding recalls. The FDCA 
does not authorize the FDA to “ order ” a manufacturer to recall a drug product. 11 
In practice, however, the manufacturers or distributors of the drug products are 
encouraged to implement and carry out recalls voluntarily to fulfi ll their responsibility 
to protect the public. It is not uncommon for a company to discover that one of 
its products is defective and recall it entirely on its own; or the FDA informs a 
company of its fi ndings that one of its products is defective and suggests or requests 
10 The data for the analysis were compiled by the FDA and derived from 614 Turbo EIR reports completed 
from 2001 to 2003. (FDA investigators enter their inspections observations on the FDA ’ s Turbo 
EIR system. The Turbo ’ s electronic format prompts investigators to select the specifi c cGMP violation 
in question and then to explain their fi ndings uncovered during the inspection.) 
11 The FDCA gives authority to the FDA to order a recall in some cases involving infant formulas, 
biological products, and devices that present a “ serious hazard to health, ” but not involving 
pharmaceuticals. 

52 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
a recall. 12 Once a voluntary recall is initiated, the FDA generally follows the following 
protocol (Figures 3 – 5 ): 
1. Classify the Recall The FDA reviews relevant information and then assigns 
a recall classifi cation according to the level of health risk involved: 
Class I recalls involve drug products in which the reason for recall predictably 
could cause serious health problems or death. 
Class II recalls involve drug products which defect might cause a temporary 
health problem or pose only a slight threat of a serious nature. 
Class III recalls involve products that are unlikely to cause any adverse 
health reaction but that violate FDA labeling or manufacturing 
regulations. 
2. Monitor and Audit the Recall The FDA oversees a recall depending upon the 
health risk involved. For a class I recall, the FDA checks to make sure that the 
defective product has been recalled in full. In contrast, for a class III recall, 
FDA oversight may be to simply spot - check. 
FIGURE 3 2005 recall by class. [ Source : Centre for Drug Evaluation and Research (CDER) 
2005 Report to the Nation. ] 
Class I: 18 
Class II: 314 
Class III 170 
TOTAL: 502 
FIGURE 4 Drug recalls. One fi rm had over 100 recalls in 2005, which caused a spike in the 
2005 recall fi gures. ( Source : CDER 2005 Report to the Nation .) 
191 
226 
248 
176 
352 
316 
248 
354 
254 
215 
401 
60 53 
34 
88 
72 
156 
72 
83 88 
71 
101 
0 
50 
100 
150 
200 
250 
300 
350 
400 
450 
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 
Number of Recalls 
Rx 
OTC 
12 If the company does not comply, then FDA can seek judicial enforcement under the FDCA. 

FDA ENFORCEMENT TECHNIQUES 53 
3. Notifi cation and Public Warning Class I recalls almost always warrant a 
press release to the media. Classes II and III are not necessarily announced 
in the media, but all of them are included in the FDA ’ s weekly enforcement 
report, posted at www.fda.gov/opacom/Enforce.html on the FDA ’ s 
website. 
4. Termination The FDA provides written notice to the recalling manufacturer 
on when the recall should be terminated. 
5. Noncompliance If applicable, the FDA will take appropriate legal action if 
a manufacturer fails or refuses to timely complete a recall. 
1.2.3.4 Warning Letter 
A warning letter is intended to notify manufacturers about violations that the FDA 
has documented during its inspections or investigations. A warning letter will notify 
a responsible individual and/or fi rm that the FDA considers one or more products, 
practices, processes, or other activities to be in violation of the cGMPs. Warning 
letters should only be issued for violations of regulatory signifi cance, that is, those 
that may actually lead to an enforcement action if the documented violations are 
not promptly and adequately corrected. A warning letter is one of the FDA ’ s principal 
means of achieving prompt voluntary compliance. 
Examples of situations in which the FDA may be expected to issue a warning 
letter include: 
• An active pharmaceutical ingredient (API) batch fails to conform to established 
specifi cations and yet the manufacturer distributed it anyway. 
• Deliberately blending API batches to dilute or hide noxious contaminant or 
fi lth or failing to determine actual yield and percentages of expected yields. 
• Contamination of drugs with toxic chemicals, drug residues, airborne contaminants, 
or fi lth. 
• Failing to comply with commitments in drug applications. 
• Combining a batch that does not conform with critical attributes with a batch 
that does. 
• Failing to demonstrate water used in the manufacturing process is suitable. 
• Failing to validate water systems. 
FIGURE 5 Top 10 reasons for drug recalls in fi scal year 2005. ( Source : FDA .) 
• Miscellaneous cGMP deviations (other than below) 
• Failed USP dissolution test requirements 
• Microbial contamination of non-sterile products 
• Lack of efficacy 
• Impurities/degradation products 
• Lack of assurance of sterility 
• Lack of product stability 
• Labeling: Label error on declared strength 
• Misbranded: Promotional literature with unapproved therapeutic 
claims 
• Labeling: Correctly labeled product in incorrect carton or package

54 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
• Lacking a formal written program to validate an API validation process. 
• Failing to demonstrate homogeneity of fi nal blending operations. 
• Failing to keep adequate batch records. 
• Failing to have a formal process change control system in place. 
• Using inadequate or unvalidated laboratory test methods. 
• Packaging and labeling processes that could introduce a signifi cant risk of 
mislabeling. 
• Failing to test for residues of organic or inorganic solvents that may carry over 
to the API. 
• Using incomplete stability studies to establish API stability for the intended 
period of use. 
Warning letters detailing cGMP violations typically conclude with the following: 
“ The article(s), (DRUG NAME), is (are) adulterated within the meaning of Section 
501(a)(2)(B) of the Act, 21 U.S.C. 351(a)(2)(B), in that the methods used in, or the 
facilities or controls used for, its manufacture, processing, packing, or holding fails 
to conform to, or is not operated or administered in conformity with, cGMP regulations 
[21 CFR 210, 211]. ” The number of warning letters issued by the FDA concerning 
prescription and over - the - counter drugs has ranged from 130 letters in 2000 to 
79 letters in 2005. 
A warning letter is distinguishable from a notice of violation, also called an 
untitled letter. An untitled letter cites violations that do not meet the threshold of 
regulatory signifi cance for a warning letter, but the FDA has a need nevertheless to 
communicate. Unlike a warning letter, an untitled letter does not include a warning 
statement that failure to take prompt correction may result in enforcement action 
and does not evoke a mandated FDA follow - up. Further, the untitled letter requests 
(rather than requires) a written response (from the manufacturer) within a reasonable 
amount of time (e.g., “ Please respond within 45 days ” ). 
1.2.4 JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 
1.2.4.1 Introduction 
The FDA is likely to bypass sending a Warning Letter in certain circumstances. 
According to Chapter 4 of the FDA Regulatory Procedures Manual , the following 
violations are likely to result in an enforcement action without necessarily issuing 
a warning letter: 
1. The violation refl ects a history of repeated or continual conduct of a similar 
or substantially similar nature during which time the individual and/or fi rm 
has been notifi ed of a similar or substantially similar violation. 
2. The violation is intentional or fl agrant. 
3. The violation presents a reasonable possibility of injury or death. 
4. Adequate notice has been given by other means and the violations have not 
been corrected or are continuing. 

5. The violations, under Title 18 U.S.C. 1001, are intentional and willful acts that 
once having occurred cannot be retracted. Also, such a felony violation does 
not require prior notice. Therefore, Title 18 U.S.C. 1001 violations are not suitable 
for inclusion in warning letters. 
In addition, actively deceiving the FDA is almost guaranteed to bring judicial 
enforcement actions. This includes false representations in the written record - 
keeping requirements or in written communication with the FDA. Manufacturing 
record - keeping requirements which give exposure to fraud liability are summarized 
in Figure 6 . Potential violations include the following: 
FIGURE 6 Written record highlights. 
Written records are required to be kept as set forth in 211.180 to 211.208. Highlights are as follows: 
§ 211.182 Equipment cleaning and use log. 
A written record of major equipment cleaning, maintenance (except routine 
maintenance such as lubrication and adjustments), and use shall be included in 
individual equipment logs that show the date, time, product, and lot number of each 
batch processed. 
§ 211.184 Component, drug product container, closure, and labeling records. 
These records shall include the following: 
(a) The identity and quantity of each shipment of each lot of components, drug 
product containers, closures, and labeling; the name of the supplier; the supplier's 
lot number(s) if known; the receiving code as specified in § 211.80; and the date 
of receipt. The name and location of the prime manufacturer, if different from the 
supplier, shall be listed if known. 
(b) The results of any test or examination performed (including those performed 
as required by § 211.82(a), § 211.84(d), or §211.122(a)) and the conclusions 
derived therefrom. 
§ 211.186 Master production and control records. 
To assure uniformity from batch to batch, master production and control records 
for each drug product, including each batch size thereof, shall be prepared, dated, 
and signed (full signature, handwritten) by one person and independently checked, 
dated, and signed by a second person. The preparation of master production and 
control records shall be described in a written procedure and such written 
procedure shall be followed. 
§ 211.188 Batch production and control records. 
Batch production and control records shall be prepared for each batch of drug 
product produced and shall include complete information relating to the production 
and control of each batch. 
§ 211.194 Laboratory records. 
Laboratory records shall include complete data derived from all tests necessary to 
assure compliance with established specifications and standards, including 
examinations and assays ... 
§ 211.198 Complaint files. 
A written record of each complaint shall be maintained in a file designated for drug 
product complaints 
JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 55

56 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
(a) Accepting and validating drug products that failed to meet established standards 
or specifi cations and any other relevant quality control criteria (i.e., 
dissolution rates, content uniformity, purity, potency) and then falsely recording 
the untruthful data as if the drug products did not fail 
(b) Accepting and validating the stability characteristics of drug products and 
then falsely recording the untruthful data as if the drug products did not 
fail 
(c) Documenting the examination and review of labels, when in truth and fact 
no review occurred (which results in inaccurate labels distributed with 
drugs) 
(d) Falsely documenting any components of master production and central 
records 
(e) Falsely documenting any component of the batch production and control 
records 
(f) Falsely describing testing methods when no (or inadequate) testing methods 
were performed 
(g) Failing to accurately make a written record of all written and oral complaints 
regarding a drug product and/or certifying that investigations were performed 
when they were not, falsely certifying that the fi ndings were negative 
when they were not, and so on 
(h) Falsifying records which would indicate manufacturing changes which require 
approval by the FDA 
(i) False representations that contain statements of fact in correspondence sent 
to the FDA addressing violations in an inspector ’ s form 483 
The FDA typically initiates progressive enforcement, as described in Figure 7 . 
Once the FDA and DOJ decide to bring enforcement action, the U.S. courts have 
held that the FDA ’ s interpretation of its cGMPs is entitled to substantial deference. 
As long as the FDA ’ s interpretation of its regulations are “ reasonable ” and “ sensibly 
conforms to the purpose and wording of the regulations, ” courts are required to 
follow the FDA ’ s interpretations. 
1.2.4.2 Civil Proceedings 
Seizures If during an inspection of a facility the FDA inspector or employee 
making the inspection has reason to believe that a drug found in such facility is 
adulterated, such inspector or employee may order the drug detained for a reasonable 
period which may not exceed 20 days (unless the FDA institutes an action 
under Subsection 334(a) or an injunction, in which case a longer detention period 
may be authorized). 
The FDCA expressly permits administrative seizure on the basis of an ex parte 
showing of reasonable belief [21 U.S.C. 334 (g)]. Seizure of a company ’ s inventory 
deprives the company of both capital investment and potential profi t. 
If the FDA pursues relief beyond detainment, the United States can fi le a complaint 
for forfeiture directing the U.S. marshall to “ seize ” the pharmaceuticals (or 
take possession or place in constructive custody of the court). The theory in a complaint 
for forfeiture is that there is a violation of the law by the pharmaceutical 

product itself. Accordingly, the government asks the court to condemn the article 
and declare forfeiture. Upon fi ling of the complaint, the clerk automatically issues 
a warrant. Thus, the FDA is able to obtain a warrant without review by a judicial 
offi cer or even a fi nding of probable cause. 
There are three types of seizures: mass, open ended, and lot specifi c. A mass 
seizure is the seizure of all FDA - regulated products at an establishment/facility. 
Mass seizures might be conducted when all of the products are produced under the 
same conditions (e.g., nonconformance with cGMPs). An open - ended seizure is the 
seizure of all units of a specifi c product or products, regardless of lot or batch 
number, when the violation is expected to be continuous. An open - ended seizure 
may be conducted when a specifi c product extends to all lots or batches of a product 
but not to all of the products in the facility. 
Following seizure of its drugs a manufacturer has three courses of action. First, 
it may do nothing, in which case the drug will be disposed of. Second, it can enter 
into a consent decree, admitting the violation, agreeing to pay costs, and seeking to 
destroy or rehabilitate the article. The consent decree will typically provide for 
(1) condemnation of the article as being in violation of the law; (2) a penal bond in 
approximately twice the retail value of the article under seizure; (3) provisions for 
payment of costs for storage and handling by the U.S. marshall and for supervision 
by the FDA before release of the product; and (4) a provision that the manufacturer 
will attempt to bring the article into compliance under the supervision of and to the 
satisfaction of the FDA. 13 
Third, it can contest the action. If the manufacturer contests the action, the case 
is then treated like any other civil case under the federal rules of civil procedure, 
and the government must prove its case by a preponderance of the evidence. The 
government must produce evidence, in support of its allegations, including proof of 
interstate shipment of the drug or its components. FDA employees may testify, but 
FIGURE 7 Progressive enforcement. 
FDA enforcement mechanisms are often utilized progressively. A good example is the 
enforcement action against Glaxo SmithKline (“GSK”), which began in July 2002, identifying 
numerous significant cGMP violations found during a February/April 2002 inspection. A Warning 
Letter requested that the violations be corrected and stated that failure to correct the violations may 
result in regulatory action, including seizure and/or injunction. Although a limited follow-up FDA 
inspection in October 2002, found that some specific corrections were acceptable, the subsequent 
FDA inspections in November/December 2003 and September/November 2004, revealed continuing 
significant cGMP violations. FDA concluded that the firm’s data and corrective plans were not 
adequate to correct the cGMP violations. GSK also initiated recall of some, but not all, lots of the 
two products. On March 4, 2005, in response to ongoing concerns about manufacturing quality, 
FDA and the DOJ initiated seizures of two GSK pharmaceuticals. The Agency initiated these 
seizures actions based on concerns that GSK’s violation of manufacturing standards may have 
resulted in the production of poor quality drug products that could potentially pose risks to 
consumers. On April 28, 2005, FDA announced that GSK had signed a Consent Decree with FDA 
to correct manufacturing deficiencies at its Cidra, Puerto Rico, facility. The Consent Decree was 
initiated based on FDA’s continued concerns that GSK’s violation of manufacturing standards may 
have resulted in the production of drug products that could potentially pose risks to consumers. 
13 Regulatory Procedures Manual , Chapter 6 - 1 - 11, March 2007. This contemplates that seizure of a specifi c 
product(s) is the sole issue. More complex consent decrees are described hereinafter. 
JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 57

58 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
also outside experts testify such as to the signifi cance of failure to comply with 
cGMP requirements. If a decree of condemnation is entered (either after trial or by 
consent), the court may direct disposition of the article by destruction. 
Injunctions The FDCA expressly authorizes the courts to restrain and enjoin acts 
that are in violation of 21 U.S.C. 331, which includes prohibition of adulterated 
products. FDA policy provides that an injunction action is appropriate where: 
(a) there is a current and defi nite health hazard or a gross consumer deception 
requiring immediate action to stop the violative practice; 
(b) there are signifi cant amounts of violative products owned by the same person 
in many locations, voluntary recall by the fi rm was refused or is signifi cantly 
inadequate to protect the public, and seizures are impractical or uneconomical; 
or 
(c) there are long - standing (chronic) violative practices that have not produced 
a health hazard or gross consumer fraud, but which have not been corrected 
through use of voluntary or other regulatory approaches. 14 
A complaint for injunction is typically accompanied by a motion for preliminary 
injunction. 15 The court schedules a court hearing to determine whether to grant a 
preliminary injunction, often very quickly and on short notice. The government ’ s 
main focus at this preliminary stage will be to prove that there is a “ substantial 
likelihood ” that the defendant has been producing adulterated drugs in violation 
of 21 U.S.C. 331, by substantial noncompliance with the cGMPs. The government 
will also typically present evidence, if applicable, that the defendant has had 
a history of prior noncompliance with the FDCA and implementing regulations. 
No specifi c fi nding of irreparable harm is necessary as is required in the typical 
injunction, because the passage of the statute proscribing adulterated products has 
15 The government may also apply for a temporary restraining order (TRO) seeking immediate, temporary 
relief (for a period of 10 days, which may be extended for 10 additional days) prior to the hearing 
for preliminary injunction. The FDA will typically recommend a TRO when it believes that the violation 
is so serious that it must be controlled immediately . 
14 Regulatory Procedures Manual , March 2007. 
FIGURE 8 Disgorgement. 
Major recent consent decrees are United States v. Abbott Labs., Consent Decree of 
Permanent Injunction filed Nov. 2, 1999; United States v. Various Articles of Drug Identified in 
Attachment A & Wyeth-Ayerst Labs., Consent Decree of Condemnation and Permanent Injunction 
filed Oct. 4, 2000; and United States v. Schering-Plough Corp., Consent Decree of Permanent 
Injunction filed May 20, 2002. To avoid giving manufacturers the wrong message by allowing them 
to keep on the market what FDA had determined to be produced in violation of the cGMPs, the FDA 
included three separate types of “disgorgement” payments in the Abbott, Wyeth and Schering 
consent decrees: (1) a lump sum payment (Abbott, $100 million; Wyeth, $30 million; and Schering, 
$500 million) (2) if the remedial work was not achieved by the deadline established in the decree, 
(i) a percentage of sales (Abbott, 16%; Wyeth, 18.5%; and Schering, 24.6%) and (ii) daily payments 
of a certain flat amount. Both were to be paid until compliance was achieved.

been held itself to be an implied fi nding by Congress that violations will harm the 
public. 
United States courts are imbued with authority to enjoin present and future 
violations of Section 331 based upon proof by the FDA that such violations have 
occurred and could recur. Factors that courts consider when determining whether 
there is a reasonable chance of future infractions include (1) the degree of scienter 
involved on the part of the defendant; (2) the isolated or recurrent nature of the 
infraction; (3) the defendant ’ s recognition of the wrongful nature of his or her 
conduct; (4) the sincerity of the defendant ’ s assurances against future violations; 
and (5) the nature of the defendant ’ s violation. The court also considers whether 
the defendant voluntary ceased the challenged conduct, the genuineness of the 
defendant ’ s efforts to conform to the law, the defendant ’ s progress toward improvement, 
and the defendant ’ s compliance with any recommendations made by the 
government. 
Good faith is not a defense to the issuance of an injunction. Nor may a defendant 
successfully defend against the issuance of an injunction by asserting that the injunction 
would drive it out of business. 
Consent Decrees and Disgorgement A consent decree is a judgment (legal order) 
issued by the court that has been agreed to by the parties whereby the defendant 
agrees to stop illegal or improper activity as alleged by the government. Once court 
approval is obtained, the seizure or injunctive lawsuit, for instance, is dropped, and 
the government ’ s remedy is then based upon any breach of the consent decree, itself 
which is enforceable by the court. 
Consent decrees typically involve a defendant agreeing to address the areas of 
noncompliance in a manner satisfactory to the FDA within a certain amount of time. 
It can also provide for the hiring of an expert consultant to certify in detailed reports 
that the manufacturing facility, at periodic dates, is in full compliance with the 
cGMPs, and has adequate adverse - event controls, adequate training, and adequate 
recall procedures. It may also require the payment of money to the U.S. Treasury 
such as under the equitable remedy of “ disgorgement, ” as described in Figure 8 . 
As part of a court action, the FDA will sometimes pursue “ disgorgement. ” The 
purpose of disgorgement is to deprive the wrongdoer of ill - gotten gains as well as 
provide deterrence. The amount of disgorgement is not necessarily directly tied to 
restitution. In practice, the amount the FDA exacts is supposed to be enough to 
send a message but certainly does not provide for full disgorgement of profi ts of 
the drug product(s) at issue. 
False Claims Act The U.S. Civil False Claims Act, 31 U.S.C. 3729 et seq., is the 
government ’ s principal means of redressing fraud by government contractors. The 
act has implications for cGMP violations because the United States (funding as it 
does the Medicare program, the state Medicaid programs, the Veterans Administration, 
the TRICARE program, and others) is the world ’ s largest purchaser of prescription 
medications. 
Nonetheless, the government has yet to bring a False Claims Act case which seeks 
damages for cGMP. One reason may be that the government has multiple other 
remedies within which to recover damages from noncompliant manufacturers, such 
as criminal fi nes and penalties and disgorgement. 
JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 59

60 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
Qui tam whistleblowers, 16 however, have already begun bringing such cases. 
Because the False Claims Act imposes liability on any government contractor which 
knowingly submits false claims to the United States or which uses false documents 
to get a false claim paid, a pharmaceutical manufacturer which knew or was recklessly 
indifferent to the fact that the manufacturing process was compromised by 
cGMP violations is in the same position as any other contractor which is required 
to conform to contractual or regulatory standards. The basis of liability under the 
False Claims Act is that false records have been generated which caused (false) 
claims for drugs to be paid by the United States. 17 The monetary damages result 
because the payor (in this case, the United States) is potentially paying for substandard 
drugs due to the cGMP violations — later covered up by false statements in 
documents required to be completed under the cGMP. 
It makes sense, too: The cGMPs are a set of regulations which, by their very 
nature, are designed to ensure that drugs are manufactured in such a way that they 
meet the requirements of the federal Food, Drug and Cosmetic Act as to safety and 
have the identity and strength and meet the purity characteristics that they purport 
or are represented to possess. The major federally funded government health care 
programs, Medicare and Medicaid, operate under the express provisions that they 
will only pay for medical services and products that are “ reasonable and necessary. ” 
Unsafe or ineffective drug products are neither reasonable nor necessary. Accordingly, 
as the theory goes, the United States suffers monetary damages if Medicare 
and Medicaid programs pay for unsafe or less effective products. These and other 
federally funded health care programs spend billions of dollars every year on 
pharmaceuticals. 
False representations concerning minor or technical violations will not be the 
basis for FCA liability. Distribution of products that are not totally cGMP compliant 
(but have been falsely documented to be) does not necessarily result in unsafe (or 
subpotent) products. Substantial violations of the cGMP, later covered up in writing, 
however, could very well be the basis for FCA liability. The common thread through 
each violation is that the violation is severe enough so that the drug product that 
16 Qui tam is shorthand for the Latin phrase, qui tam pro domino rege quam pro seipso , meaning “ He 
who is as much for the king as for himself. ” Qui tam statutes date back to thirteenth - century England. 
The actions were a means of enabling private parties to allege the king ’ s interest and therefore gain 
access to the royal courts. 
The qui tam provisions of the federal False Claims Act allow any citizen who has knowledge of fraud 
that has taken place against the government to bring a civil action in federal court in the name of the 
United States. In return for his or her efforts, the citizen is entitled to share in the proceeds of the recovery. 
The qui tam provisions raise the incentive for insiders to put the spotlight on the criminals, thereby 
providing the government with tangible and detailed evidence upon which to base an investigation and 
prosecution. 
In 1986, Congress enacted amendments to the False Claims Act which strengthened the law and 
increased monetary awards. When hearings were held in 1985 and 1986, the climate was favorable for 
strengthened antifraud legislation, and Congress expected that most qui tam cases would involve defense 
contractor fraud. In the last decade, the majority of cases have instead been against the health care 
industry. 
17 Even so, factual questions will be raised, including: (1) Even with the false representations, was a false 
claim “ caused ” to be submitted? (2) Had the FDA known about the falsities, would it have enjoined the 
manufacturer from any further production, etc? (3) What about the false record or statement made the 
claims for such drugs false? 

fi nally reaches the public is foreseeably and substantially less safe or less effective 
than if the cGMPs were not violated. 
1.2.4.3 Criminal Proceedings 
Introduction Criminal prosecutions of violations of the FDCA are intended to 
further the goal of protecting the health and safety of the public. The FDA historically 
has not pursued criminal charges unless the defendant shows a continuous or 
repetitive course of violative conduct, with the exception of intentional violations, 
fraud, or danger to health. 18 
While the FDCA contains various prohibitions and restrictions which a drug 
company could violate, the most common FDCA violation arising out of cGMPs 
is charged by using 21 U.S.C. 331(a), which specifi cally prohibits introducing an 
adulterated 19 drug into interstate commerce. In addition to introducing an adulterated 
drug into interstate commerce, some other acts prohibited by Section 331(a) 
which could be involved in manufacturing violations include 331(e), which prohibits 
the refusal to allow access to records mandated elsewhere in the act and 331(f), 
which prohibits the refusal to allow inspection of production facilities (21 U.S.C. 
374). 
Commission of any act prohibited by Section 331 is a federal misdemeanor (21 
U.S.C. 333). However, violations of Section 331(a) may be charged as felonies where 
there is intent to defraud or mislead or where the defendant previously has been 
convicted of a misdemeanor under the FDCA [21 U.S.C. 333(b)]. Federal misdemeanor 
charges are typically resolved in proceedings before U.S. magistrate judges, 
and federal felonies are resolved by U.S. district judges. 
Individual versus Corporate Liability Introducing an adulterated product into 
interstate commerce is a strict liability crime that can be enforced against individuals 
in positions of suffi cient authority and responsibility as well as their company. 
Persons at risk are those who, at minimum, 20 fail to take adequate measures to 
prevent the cGMP violations. As such, warning letters and other communications 
are often directed at presidents and CEOs as well as their companies. As stated by 
the U.S. Supreme Court 21 in 1964, just two years after the FDCA as we know it was 
passed: 
Food and drug legislation, concerned as it is with protecting the lives and health of 
human beings, under circumstances in which they might be unable to protect themselves, 
often “ dispenses with the conventional requirement for criminal conduct — 
awareness of some wrongdoing. In the interest of the larger good it puts the burden of 
acting at hazard upon a person otherwise innocent but standing in responsible relation 
to a public danger. . . . ” 
18 A government review of recent FDA enforcement has suggested that adequate FDA enforcement 
activity is lacking. See “ Prescription for Harm: The Decline in FDA Enforcement Activity, ” House Committee 
on Government Reform, June 2006. 
19 Failure to follow cGMP is the most common form of violating the prohibition against introducing an 
adulterated drug into interstate commerce. 
20 They may also be directly implicated in fraud and cover - ups. 
21 United States v. Wiesenfeld Warehouse Co. , 376 U.S. 86, 91 (1964). 
JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 61

62 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
In 1975, the Supreme Court made clear that individual responsibility is very 
important 22 : 
The [FDCA] imposes not only a positive duty to seek out and remedy violations when 
they occur but also, and primarily, a duty to implement measures that will insure that 
violations will not occur. The requirements of foresight and vigilance imposed on 
responsible corporate agents are beyond question demanding, and perhaps onerous, 
but they are no more stringent than the public has a right to expect of those who voluntarily 
assume positions of authority in business enterprises whose services and products 
affect the health and well - being of the public that supports them. 
Manufacturing executives therefore carry a great liability burden. The government 
only need establish that the individual defendant failed to act on his or 
her own authority and that such an action could have prevented or corrected 
the violation. The individual need not have formed any intent to break any laws in 
order to be found guilty. What is relevant is, did the executive have the power to 
prevent the acts or omissions complained of? This includes a consideration of 
whether the executive could have prevented the acts or omission by the systems 
and processes alone. The job is not made easier to the extent that the cGMP regulations 
are open to varying interpretations or that the technology is constantly 
changing. 
Section 305 Proceedings Due to the nature of the inspection process, a company 
that the FDA deems is in violation of the cGMPs should not be surprised when a 
warning letter or more elevated enforcement techniques are implemented. Even so, 
the FDA sometimes issues a formal form of notice that criminal charges will be 
brought by what is called a Section 305 notice. 
Section 305 of 21 U.S.C., the statutory basis for a 305 notice, seemingly requires 
that before any violation of the FDCA is reported to the DOJ for institution of a 
criminal proceeding, the target defendant must “ be given appropriate notice and an 
opportunity to present his views, either orally or in writing, with regard to such 
contemplated proceeding. ” The U.S. Supreme Court has watered down this provision, 
holding that a notice under Section 305 is not a legal prerequisite to government 
prosecution. 
In practice, then, the FDA only sometimes issues a 305 notice and conducts a 305 
hearing when it is considering a misdemeanor prosecution. A very informal process, 
the manufacturer can approach it with as much or as little of a defense as counsel 
deems appropriate, as there are pros and cons to providing the government with 
the company ’ s full defense at that juncture. 
A prototype Section 305 notice appears in Figure 9 . 
Grand Jury Proceedings If the government will be pursuing felony criminal 
charges against a manufacturing facility or persons associated with such facility, it 
will proceed by grand jury. The Fifth Amendment to the U.S. constitution requires 
that charges for all capital or “ infamous ” crimes be brought by an indictment 
22 United States v. Park , 421 U.S. 658 (1975). 

returned by a grand jury. This has been interpreted by the U.S. courts to require that 
an indictment be used to charge federal felonies. 
The activities, deliberations, or matters occurring before a grand jury are secret. 23 
Strict adherence to grand jury secrecy is important to the integrity of the investigative 
process and ensures that the grand jury will be able to deliberate without 
outside pressure, to encourage people with information about a crime to come 
forward without fear of disclosure, and to protect the rights of the accused, specifi - 
cally the innocent accused, from disclosure of the fact that he or she or it was 
investigated. Other than attorneys for the government, only the witness, interpreters 
when needed, and a court reporter are authorized persons permitted to be present 
while a grand jury is in session. 
A grand jury ’ s function is to determine whether there is probable cause to believe 
that a certain person(s) or company(ies) have committed a federal offense. 24 Prosecutors 
are permitted to appear before the grand jury and, in practice, conduct the 
grand jury proceedings. In general, the prosecutor is the one who makes the decision 
FIGURE 9 Section 305 notice. 
In reply refer to: 
Sample No. 
Product 
Firm Name and Individual Date 
Street Address 
City, State, Zip 
Investigation by this Administration indicates your responsibility for violations of the 
Federal Food, Drug, and Cosmetic Act, and other Federal Laws, as described in the 
attached Charge Sheet, with respect to the following: 
[describes specifics of cGMP violations] 
A meeting has been scheduled for (day, date, time) at (location), to give you an 
opportunity to present your views on this matter. The enclosed INFORMATION SHEET 
explains the purpose and nature of the meeting, and how you may reply. If no response 
is received on or before the date set, our decision on whether to refer the matter to the 
Department of Justice for prosecution will be based on the evidence in hand. 
By direction of the Secretary of the Department of Health and Human Services: 
Compliance Officer 
Enclosures: 
Legal Status Sheet (3) 
Charge Sheet 
Information Sheet 
Regulations 
23 See Rule 6(e) Fed. R. Crim. Pro. 
24 The grand jury system is not presently used by countries outside the United States. The United 
Kingdom, New Zealand, Canada, and Australia, for instance, all have abolished the use of grand juries. 
See http://enwikipedia.org/wiki/Grand_jury . 
JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 63

64 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
as to which witnesses to call and what evidence should come before the grand jury. 
The prosecutor asks the witness questions and subsequently members of the grand 
jury may also question witnesses directly or through the prosecutor. 
During the course of a grand jury investigation regarding cGMP violations/adulterated 
product allegations, for instance, the grand jury may hear testimony from 
not only federal investigators, federal agents, and federal inspectors but also former 
employees of the company (or current if a custodian of records) and/or experts in 
the pharmaceutical manufacturing fi eld. These persons are considered witnesses. 
Witnesses are typically subpoenaed and may not refuse to appear before the grand 
jury or be subject to contempt charges. 
In the federal grand jury system, a witness is not permitted to bring an attorney 
into the grand jury room. However, a witness is permitted to consult with his attorney 
outside the grand jury room even interrupting his own testimony. 25 It is typical 
for corporations such as pharmaceutical manufacturing companies to provide an 
attorney for any and all employees subpoenaed by a grand jury of which the manufacturer 
is a target. 
A target is the person who is the focus of the grand jury investigation and is likely 
to be indicted. This company or person may receive a “ target letter ” from the grand 
jury which offi cially advises them of their jeopardy and serves as a formal warning 
of their status. 
In practice, if a manufacturer is the target, the government will likely attempt 
to develop evidence by subpoena of persons and materials which will help 
prove culpability. It is likely the subpoenas will ask for correspondence, notes, 
and memos during a particular time period and involving a particular subject 
matter. 
The grand jury may also issue a subpoena to the manufacturer ’ s designated 
“ custodian of records ” for specifi c document production. The description of subpoenaed 
documents can include statements or charts of an organization, 
announcements, statements of policy and procedure, diaries, records of email, 
manufacturing logs, emails, travel vouchers, fi nancial records and statements, 
correspondence, notes of conversations, and any other documents that relate to the 
manufacturing of certain drugs. A document subpoena may also request every 
writing or record of whatever type and description in the possession, custody, or 
control of the company that relates to a particular element of the criminal violation 
the grand jury is investigating. The request typically includes all handwritten, typed, 
printed, recorded, or transcribed records, including computer records tapes or 
disks. 
The burden is on the government to prove that the crime was committed in the 
district in which the prosecution is brought. The grand jury should not consider a 
case unless venue lies in the district where the grand jury is sitting. 26 In the case of 
adulterated drugs, courts have generally held that it is proper to have venue in a 
district from which the defendant caused the unlawful introduction of goods into 
commerce, even though the physical shipment commenced from a different 
district. 
25 See Rule 6(d) Fed. R. Crim. Pro.; 28 U.S.C. secs. 515, 542, 547. 
26 See Rule 18 Fed. R. Crim. Pro. 

Form of Charges and Penalties A grand jury investigation may culminate in the 
return of an indictment. This means that the grand jury found probable cause to 
believe that a violation of law occurred. While the focus of the initial inquiry can 
surround adulterated drugs by virtue of failure to abide by the cGMPs, additional 
criminal violations may be charged, as, for instance, where there are actions to evade 
or mislead a grand jury. The end result could, for instance, include accusations of 
making false statements to the FDA and obstructing the FDA ’ s or DOJ ’ s investigation, 
in addition to the “ adulterated ” drug charges. 
An indictment consists of a statement describing the time, place, and manner 
through which the defendant violated the law. Each violation of the law is set out 
in a separate count. A defendant charged by an indictment is entitled to a trial by 
jury, although this right can be waived. A defendant has the right to a trial by 
jury for any criminal offense punishable by imprisonment for more than six 
months. 27 
If the matter only involves a misdemeanor violation, the prosecutor charges 
“ by information. ” The information is often referred to as a complaint. An information, 
like an indictment, is simply a pleading that accuses the defendant of committing 
crimes. The distinction between an information and an indictment is that a 
prosecutor can issue and fi le an information without the grand jury ’ s participation 
or fi nding of probable cause but a grand jury must approve and return an 
indictment. 28 
The penalty for a violation of 21 U.S.C. 331(b) by violating the cGMPs resulting 
in the adulteration of drug products in interstate commerce is set forth in 21 U.S.C. 
sec. 333(a). Each separate count for violating the cGMPs (where a misdemeanor is 
charged) carries with it a possible imprisonment of not more than one year or a fi ne 
not to exceed $ 1000 or both. If the government charges a felony for violation of the 
cGMPs, then the penalties are imprisonment of not more than three years or a fi ne 
not to exceed $ 10,000 or both. Of course, there may be other charges with greater 
or lesser penalties which are not related to the adulterated drug charges. 
The criminal fi ne amounts (but not the imprisonment durations), however, are 
superceded by the criminal fi ne amounts contained in a different federal statute 
enacted later. Section 3571 of 18 U.S.C. provides for much greater fi nes than those 
provided for within the FDCA itself. For a manufacturer convicted of a felony, the 
fi ne can be as much as $ 500,000; for a misdemeanor (not resulting in death), it could 
be $ 200,000. Fines for individuals include a maximum up to $ 150,000 for a felony 
and an amount up to $ 100,000 for a conviction of a misdemeanor not resulting in 
death. The statute also provides for a multiplier of 2 based upon a fi nding that the 
defendant derived a pecuniary gain from the offense. The FDA and DOJ are able 
to elevate the monetary recoveries against the manufacturers for violations of the 
cGMP under a civil disgorgement theory explained infra. Recoveries have been in 
the hundreds of millions in recent years, typically agreed to in a negotiated consent 
decree. 
27 See Sixth Amendment, U.S. Constitution. 
28 Sometimes, prosecutors are in communication with defendants and their counsel during the investigatory 
stage. If there are negotiations concerning a plea to a felony and they are successful, a defendant 
can waive his or her right to be indicted by a grand jury, and the prosecutor can charge them by 
information. 
JUDICIAL ENFORCEMENT: BEYOND THE WARNING LETTER 65

66 ENFORCEMENT OF CURRENT GOOD MANUFACTURING PRACTICES 
1.2.5 CONCLUSION 
The diligent enforcement of good manufacturing practices is a cornerstone of the 
safety net for drugs in the United States. Congress, the courts and, manufacturers, 
most importantly, expect a degree of consistency and responsibility in enforcement 
policy over a statute as powerful and central to public health and safety as the 
FDCA. To the extent there is consistency and effective and evenhanded enforcement, 
it not only protects the public, but it provides a level playing fi eld for those 
manufacturers who operate in accordance with the cGMPs. 

67 
1.3 
SCALE - UP AND POSTAPPROVAL 
CHANGES (SUPAC) REGULATIONS 
Puneet Sharma , Srinivas Ganta , and Sanjay Garg 
University of Auckland, Auckland, New Zealand 
Contents 
1.3.1 Introduction 
1.3.2 Scientifi c and Regulatory Rationale for SUPAC 
1.3.2.1 Supporting Documents and Extent of Change 
1.3.2.2 Supporting Documents for Change in Specifi cations 
1.3.2.3 Comparability Protocols 
1.3.2.4 In Vitro – In Vivo Requirements 
1.3.3 Regulatory Agencies and Guidelines 
1.3.3.1 FDA SUPAC Regulations 
1.3.3.2 Regulatory Guidance on SUPAC by Pharmaceutical Unit of EU 
1.3.3.3 Regulatory Guidance on SUPAC by Agencia Nacional de Vigilancia 
Sanitaria 
1.3.4 Harmonization 
1.3.5 GMP Issues: Change Control and Process Validation 
1.3.5.1 Change Control 
1.3.5.2 Process Validation 
1.3.6 Conclusion 
1.3.1 INTRODUCTION 
Product development aims at formulating active drug ingredient in a palatable form. 
Technology transfer of a pharmaceutical product from research to the production 
fl oor (referred to as “ shop fl oor ” ) with simultaneous increase in production outputs 
is commonly known as scale - up. In simple terms, the process of increasing batch size 
is termed as scale - up. Conversely, scale - down refers to decrease in batch size in 
response to reduced market requirements. 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

68 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
Often, changing of scale from the research lab to the shop fl oor is fraught with 
problems. The basic reason for such problems is the usage of different processing 
equipment in research and on the shop fl oor. Moreover, insuffi cient information 
about the equipment, various requirements of process control, complexity of a particular 
pharmaceutical process which may have a several different unit operations, 
limited information about the behavior of ingredients at different scales, and adoption 
of trial - and - error methodology also add signifi cantly to scale - up issues. Every 
product coming from research should be manufacturable and the process should be 
capable to demonstrate its ruggedness at the shop fl oor level. This statement points 
toward the criticality and signifi cance of scale - up and technology transfer in a pharmaceutical 
development process. After successful accomplishment of technology 
transfer and validation activity, a product usually has a smooth run on large - scale 
production machines. Changes are being made in the manufacturing process and 
chemistry of a drug product following approval and continue throughout its life. 
Depending upon foreseen (or unforeseen) requirements, there can be changes in 
the raw materials, process, equipment or manufacturing site, and batch size which 
ultimately affect quality attributes of a drug or fi nished product. Therefore, there is 
a need to anticipate and fully evaluate the impact of any kind of change on the 
quality of a drug or fi nished product. There can be several reasons for these changes, 
such as changed market requirement affecting batch size, new source of raw material, 
change in manufacturing process, upgrades of packaging material, or shifting 
to a new analytical methodology. 
The intensity of the adverse effect produced by a particular change depends on 
the type of dosage form. For example, a change in the inactive ingredient beyond a 
certain range will have more effect on a modifi ed - release (MR) dosage form than 
it would on an immediate - release (IR) dosage form, where bioavailability is not rate 
limiting. Likewise, a change in the primary packaging of liquid parenteral may have 
more pronounced effect on its effectiveness than it would have on a solid dosage 
form. Hence, depending upon the intensity of change or the adverse effect it may 
have on the critical parameters of a dosage form, reporting requirements to regulatory 
authorities also vary. 
A drug or drug product may experience many changes during its life cycle. These 
changes may have an adverse effect on the overall safety and effectiveness of the 
drug or drug product. After a number of changes over a long time period, the 
product coming to market may be completely different from the one that was 
approved. Hence, data submitted to regulatory authorities in support of a change 
must have a comparison record of the drug or drug product to the one that was 
approved initially. Documentation generated in support of any change to the 
approved drug or drug product is submitted to regulatory authority for review, and 
based on the benefi t - to - risk ratio, the drug or drug product is approved. Depending 
upon the intensity of change, supporting documents are provided to the regulatory 
agency. 
Regulatory authorities such as the U.S. Food and Drug Administration (FDA), 
the European Commission, the Agencia Nacional de Vigilancia Sanitaria (ANVISA) 
(in english the National Health Surveillance Agency — Brazil, and others require the 
pharmaceutical industries in respective countries to follow guidelines on scale - up 
and postapproval changes (SUPAC) to maintain the quality of the pharmaceutical 
produced. From time to time these guidelines are assessed so as to keep pace with 

SCIENTIFIC AND REGULATORY RATIONALE FOR SUPAC 69 
the technological advances and new guidelines are developed to reduce the burden 
on the pharmaceutical industry and regulatory authorities. Apart from these guidelines, 
there are other checkpoints within an industry to assure production of quality 
products, such as change control and validation exercises, which will be discussed in 
detail in this chapter. These operations are controlled through the principles of good 
manufacturing practices issued by regulatory authorities. 
This chapter describes the regulations imposed by different regulatory authorities 
and measures taken by a pharmaceutical industry to assure quality and performance 
of pharmaceuticals. The FDA guidelines, being most descriptive, have been 
discussed at length. Other guidelines have been described in general terms and the 
interested reader is referred to the references or the regulatory websites for more 
specifi c details. 
1.3.2 SCIENTIFIC AND REGULATORY RATIONALE FOR SUPAC 
Guidelines pertaining to postapproval changes classify these changes in various 
categories depending upon the effect a particular change may have on the quality 
and performance of a drug or drug product. Irrespective of the terminologies used 
by regulatory agencies, in general terms, changes can be described as mild, moderate , 
and major and the extent of supporting document varies with the nature of the 
change. For example, U.S. FDA guideline “ Changes to an Approved NDA or ANDA ” 
describe these changes as mild changes that can be implemented immediately and 
fi led in the next periodic report, moderate changes that can be implemented immediately, 
moderate changes that require 30 days notice before implementation, and 
major changes that require FDA approval before implementation [1] . Similarly, any 
changes in an approved drug or drug product under European Union (EU) domain 
type I (type IA and type IB) and II variation are fi led prior to marketing products 
[2] . The therapeutics Good Administration — Australia (TGA) describes postapproval 
changes in three categories: nonassessable, self - assessable, and changes 
requiring prior approval [3] . 
1.3.2.1 Supporting Documents and Extent of Change 
As per FDA guidelines, changes in excipients (%w/w) of total formulation not 
greater than 5% are considered minor and all information is provided in the annual 
report. However, changes likely to have signifi cant effect on the quality and performance 
of a drug product calls for submission of a prior approval supplement on all 
information (in vitro dissolution and in vivo dissolution), including accelerated stability 
and long - term stability testing in the annual report [4] . Similarly, in EU guidelines, 
a change in the batch size of the fi nished product up to10 - fold compared to 
the original batch size approved at the grant of the marketing authorization (or 
downscaling to 10 - fold) has been defi ned as type IA and requires batch analysis 
data (in a comparative tabulated format) on a minimum of one production batch 
manufactured to both the currently approved and the proposed sizes. Batch data 
on the next two full production batches should be made available upon request and 
reported by the marketing authorization holder if outside specifi cations (with proposed 
action). However, for type IB (more than 10 - fold), in addition to the above 

70 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
data, a copy of an approved release and end - of - shelf - life specifi cations as well as 
the batch numbers ( . 3) used in the validation study should be indicated or a validation 
protocol (scheme) be submitted and the number of batches used in the stability 
studies should be indicated. 
1.3.2.2 Supporting Documents for Change in Specifi cations 
Changes in any type of specifi cation also need to be supported by documentation. 
In all the guidelines, relaxing an acceptance criterion or deleting any part of the 
specifi cation is classifi ed as a major change and hence extensive documentation is 
required, for example, submission of a prior approval supplement to the FDA or 
comparative table of current and proposed specifi cations and details of any new 
analytical method and validation data and batch analysis data on two production 
batches of the fi nished product for all tests in the new specifi cation to EU. The 
specifi cations are benchmarks for comparison of performance of any product. For 
example, content uniformity specifi cation of 90 – 110% assay limit of a 20 - mg (average 
weight) tablet of a potent drug signifi es the challenge in maintaining the uniformity 
of such a low - dose drug during the blending operation. Any relaxation in specifi cation 
of this potent drug should be justifi ed with extensive documentation to assure 
the performance. However, tightening of an acceptance criterion is considered as a 
minor level change and to have minimal potential for an adverse effect on the 
identity, quality, purity, or potency of a product. 
1.3.2.3 Comparability Protocols 
The FDA has introduced the concept of comparability protocols to expedite the 
process of approval after submission of supporting document for a particular change 
[5] . The protocol covers anticipated changes a product may experience during 
it shelf life. Its recently published draft guidance “ Comparability Protocols — 
Chemistry, Manufacturing, and Controls (CMC) Information ” describes the general 
principles and procedures to prepare comparability protocols. The FDA suggests a 
less stringent reporting category for any future change, where appropriate. Additionally, 
if a detailed comparability protocol is provided, the FDA is less likely to 
request additional supporting documents while comparing pre - and postapproval 
change, and this could also help in implementing a particular CMC change, thereby 
moving the product in the distribution line sooner. According to the FDA: 
A comparability protocol is a well - defi ned, detailed, written plan for assessing the effect 
of specifi c CMC changes in the identity, strength, quality, purity, and potency of a specifi 
c drug product as these factors relate to the safety and effectiveness of the product. 
A comparability protocol describes the changes that are covered under the protocol 
and specifi es the tests and studies that will be performed, including the analytical procedures 
that will be used, and acceptance criteria that will be achieved to demonstrate 
that specifi ed CMC changes do not adversely affect the product. The submission of a 
comparability protocol is optional. 
A comparability protocol may be submitted with a new drug application (NDA), 
abbreviated new drug application (ANDA), or supplements to these applications. 

SCIENTIFIC AND REGULATORY RATIONALE FOR SUPAC 71 
Comparability protocols can have single or multiple changes provided that each 
change is discrete and specifi cation of the acceptance criteria for a change is well 
defi ned. 
1.3.2.4 In Vitro – In Vivo Requirements 
Stability of a drug product, in vitro dissolution, and in vivo bioequivalence are prerequisites 
for performance of a drug product and play a key role in establishing the 
quality of a drug product after a postapproval change has been implemented. Any 
type of major change, for example, in the manufacturing process from dry granulation 
to wet granulation could affect the bioavailability and stability of a drug 
product. Careful selection of the dissolution condition can obviate the need for a 
costly bioequivalence study. Guidelines by the FDA [4] and ANVISA [6] take into 
consideration the solubility and permeability of a drug substance for selection of 
dissolution criteria for a particular drug product (immediate release or modifi ed 
release) whereas guidelines by the EU and TGA recommend submitting comparison 
records between a particular number of manufacturing batches pre - and 
postapproval. 
While categorizing a change or variation for its effect, suffi cient consideration 
should be given to those parameters of a drug product which could affect its bioavailability. 
Critical parameters like the particle size of active ingredient or excipients, 
solid - state characteristics, and surface wettability may change during the 
process variation and could adversely affect product performance resulting in an 
altered dissolution profi le. The effect would be more pronounced in drug products 
containing poorly soluble potent drugs and could have a deleterious effect on bioavailability. 
The FDA guideline considers recommendation of the Biopharmaceutic 
Classifi cation System (BCS) regarding solubility and permeability characteristics to 
see whether any in vivo bioequivalence study is needed along with an in vitro dissolution 
study. In the same pattern, ANVISA places drugs in three categories for 
solid TM dosage form: case A, active substances with high permeability and high 
solubility; case B, active substances with low permeability and high solubility; and 
case C, active substances with high permeability and low solubility. As per the 
guideline, for alteration of registration due to excipient change, for level 2 alterations 
(that could cause signifi cant impact on quality and performance) the following 
requirements should be met: 
Case A “ The required documentation must include the undertaking of the 
technical report and assessment of the results of the dissolution test, carried 
out as described in the Brazilian Pharmacopoeia and, in its absence, other 
codes authorized by the legislation in force. There must be dissolution of at 
least 85% of the active substance in up to 15 minutes, using 900 ml of HCl 
0.1 M . In case this criterion is not complied with, the tests described for Cases 
B or C must be carried out. ” 
Case B “ The required documentation must include the undertaking of the 
technical report and assessment of the results of the dissolution profi le employing 
Pharmacopeial conditions and removing samples from the medium at 
appropriate time points until the plateau is reached. The dissolution profi le 
obtained must be similar to the profi le of the unaltered formulation. ” 

72 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
Case C “ The required documentation must include the undertaking of the 
technical report and assessment of the results of the dissolution profi le in fi ve 
different conditions: distilled water, HCl 0.1 M and phosphate buffer pH 4.5, 
6.5 and 7.5 for the proposed formulation and the previous formulation, without 
change. Samples of the dissolution medium must be removed at appropriate 
time points until 90% of the active substance is dissolved or the plateau is 
reached. A tensoactive may be used only when appropriately justifi ed. The 
profi le obtained must be similar to the profi le of the unaltered formulation. ” 
In addition, for level 2 change, no additional bioequivalence study is required if 
the proposed alteration matches with the situation for cases A, B, and C. However, 
if there is any deviation, then documentation containing the results and assessment 
of a new bioequivalence and/or bioavailability study [if proper in vitro/in vivo correlation 
( ivivc ) has not been established] should be submitted as per the conditions 
mentioned in the level 3 alteration. 
1.3.3 REGULATORY AGENCIES AND GUIDELINES 
1.3.3.1 FDA SUPAC Regulations 
The Food and Drug Administration Modernization Act (FDAMA) of 1997 (the 
Modernization Act) was passed on November 21. With FDAMA in effect, another 
Section 506A was added to the federal Food, Drug, and Cosmetic Act (the act) and 
Section 314.70 (21 CFR 314.70) and the section included recommendations for 
reporting categories (in terms of defi ned words) for any type of manufacturing 
changes to an approved application (NDA or ANDA). In accordance with the act, 
the FDA issued “ Guidance for Industry: Changes to an Approved NDA or ANDA ” 
(fi nalized in 2004). This guidance is a current standard for pharmaceutical manufacturers 
for making and reporting manufacturing changes to an approved application 
and for distributing a drug product made with such changes. 
“ SUPAC - IR: Immediate - Release Solid Oral Dosage Forms: Scale - Up and Post - 
Approval Changes: Chemistry, Manufacturing and Controls, In vitro Dissolution 
Testing, and In vivo Bioequivalence Documentation ” (issued 1995) was the fi rst 
attempt to provide the pharmaceutical industry with a clear - cut guideline covering 
the requirements for notifi cation and submission of documentation to regulatory 
authorities pertaining to postapproval changes. This guideline was an outcome of 
(a) a workshop on the scale - up of IR products conducted by the American Association 
of Pharmaceutical Scientists with the U.S. Pharmacopoeia (USP) Convention 
and the FDA; (b) research conducted by the University of Maryland at Baltimore 
on the CMC of IR products; (c) drug categorization research on the permeability 
of drug substances at the University of Michigan and the University of Uppsala; 
and (d) SUPAC task force set up by the Centre for Drug Evaluation and Research 
(CDER) CMC coordination committee. Following the issuance, it became a benchmark 
for the industry. Two more guidances have been published on the same format 
(level of changes as defi ned by SUPAC IR) as of SUPAC for MR drug products 
(issued in 1997) [7] and nonsterile semisolid drug products (issued in 1997) [8] . 
“ Guideline for Changes to Approved NDA or ANDA ” supersedes any previous 

guidelines which have information on reporting categories that is inconsistent with 
this guideline. 
Guideline to Industry: Changes to Approved NDA or ANDA “ Guideline for 
Industry: Changes to Approved NDA or ANDA ” provided reporting categories for 
various postapproval changes and relaxed certain requirements that were considered 
to have minimal or no impact on the drug product [1] . Moreover, it lessened 
the burden on regulatory authorities and companies as well. Four reporting categories 
provided in this guideline are as follows: 
1. Prior Approval Supplement For a major change (substantial potential to 
have effect on quality and performance), a supplement has to be submitted to 
the FDA for approval before a product made with the change is distributed. 
There is also a provision for “ Prior Approval Supplement: Expedite Review 
Requested ” for public health reasons and if the delay in approval may cause 
any substantial concerns for the applicant. 
2. Supplement: Changes Being Effected (CBE) in 30 Days For a moderate 
change (moderate potential to have effect on quality and performance), a 
supplement has to be submitted to the FDA for approval 30 days before a 
product made with the change is distributed. 
3. Supplement: Changes Being Effected in 0 Days For some changes a supplement 
has to be submitted to the FDA and simultaneously the product made 
with the change can be distributed. 
4. Annual Report For a minor change (minimal potential to have effect on 
quality and performance), all information has to be submitted to the FDA in 
the next annual review and the product made with the change can be 
distributed. 
All three types of changes under this guidance have been categorized as 
follows: 
(a) Changes in Components and Composition Any qualitative or quantitative 
changes in the components and composition of a drug product is considered 
as major changes. The current Guideline to Industry: Changes to Approved 
NDA or ANDA does not mention these in detail because of the complexity 
involved in the recommendations and therefore the SUPAC guideline has to 
be followed for any such type of changes and regarding documentation 
requirements for regulatory submission. 
(b) Changes in Manufacturing Sites A change in a manufacturing site (for 
manufacturing, packaging, labeling of drug products, testing components, 
drug product containers, closures, packaging materials), either owned or 
contract site, of drug products from the one that is approved requires prior 
approval from the CDER. A prior approval supplement has to be submitted 
for a change to a site that does not have a satisfactory CGMP inspection for 
the type of operation to be performed. Further, changes in sites related to 
operations like labeling, secondary packaging, and testing are considered to 
have effect independent of drug product dosage form and therefore the 
REGULATORY AGENCIES AND GUIDELINES 73

74 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
reporting categories for any of type of manufacturing site changes will be the 
same. However, changes in sites related to operations like manufacturing and 
primary packaging are considered to have effect that is dependent on dosage 
form and hence reporting categories may be different. 
(c) Changes in Manufacturing Process Changes in the manufacturing process 
can have substantial effect on the identity, strength, quality, purity, or potency 
of a drug product and there may be a change in the effi cacy of the drug 
product regardless of the testing of drug product for conformance for the 
approved specifi cation. 
(d) Changes in Specifi cations Specifi cation, acceptance criteria, and regulatory 
analytical procedure are a part of every dossier submitted to regulatory 
agencies. Specifi cations are the standards, acceptance criteria are the limits 
for specifi cations, and the regulatory analytical procedure is used for testing 
a specifi cation ’ s acceptance criteria for the test substance that is approved 
by the regulatory authority. Alternative analytical procedure may be 
included in the application simultaneously with the main analytical 
procedure. 
(e) Changes in the Container Closure System Effects related to changes in the 
container closure system are largely dependent on route on administration, 
the operation in which the container closure system is involved, and contact 
with the drug product. In some cases, there may be an effect in spite of the 
conformance of drug product with the approved specifi cation. 
(f) Changes in Labeling Changes in the package insert and package container 
label are included in the labeling changes and applicant must immediately 
revise all promotional labeling and drug advertisement in accordance with 
the change in the approved labeling. 
(g) Miscellaneous Changes Apart from categories mentioned above, changes 
like stability protocol, expiration period, and addition of stability protocol 
or comparability protocol have been included in the miscellaneous 
category. 
(h) Multiple Related Changes One change may lead to advertent or inadvertent 
incorporation of another change, for example, a change in the manufacturing 
site may lead to a change in the manufacturing equipment and manufacturing 
process or changes in packaging material may cause changes in stability 
protocol. For such combination changes, the CDER recommends submitting 
documents in accordance with the most stringent reporting category for the 
individual change. 
Scale - up and Postapproval Changes: Immediate - Release and Modifi ed - Release 
Dosage Forms SUPAC guidelines categorized postapproval changes in terms of 
“ levels ” [4] . Three levels were defi ned depending upon the intensity of the adverse 
effect on the formulation. Level 1 signifi es that the resulting effect on the quality 
would be minimal and less extensive documentation should be presented to the 
FDA in an annual review. Changes in accordance with level 2 could have signifi cant 
effect on the quality and performance of the dosage form. Level 3 changes are most 
likely to affect the quality and performance of the dosage form and hence extensive 

documentation justifying those changes should be submitted to the FDA prior to 
distribution of the products made with these changes. Apart from describing these 
levels, recommendations were also made on the extent of CMC documentation, in 
vitro dissolution, and in vivo bioequivalence tests that need to be submitted. Each 
section in the guideline [(a) components and compositions, (b) site change, (c) scale - 
up/scale - down; and (d) manufacturing equipment and process] was categorized in 
terms of these three levels. Further, SUPAC IR also takes into consideration the 
therapeutic range, solubility, and permeability of the drug for defi ning any particular 
change. As per the guideline, three cases have been defi ned for the dissolution 
testing (as mentioned in Table 1 ). Moreover, changes in excipient limits for a narrow 
therapeutic range drug beyond that mentioned in level 1 have been recommended 
as level 3 changes and extensive documentation is required for justifi cation. In the 
SUPAC guideline for MR dosage forms, changes have been described at the same 
three levels [7] . 
However, dissolution conditions have been distinguished quite reasonably 
between extended - and delayed - release dosage form (Table 2 ). For reporting any 
level 3 change, three - month accelerated stability data of three batches (signifi cant 
body of information not available) or three - month accelerated stability data for one 
batch (signifi cant body of information available) have to be submitted in a supplement 
along with long - term stability data for one batch in an annual review. A signifi 
cant body of information has been defi ned in the guideline as availability of 
suffi cient stability information of the product (stability data of fi ve commercial 
batches). To provide a comparative outline, the guidelines for MR and IR dosage 
forms are described in Tables 3 – 8 : 
TABLE 1 Different Cases and Respective Dissolution Conditions for Immediate - Release 
Solid Dosage Form 
Case A a Case B b Case B c 
Dissolution of 85% in 
15 min in 900 mL of 0.1 N 
HCl. If a drug product fails 
to meet this criterion, the 
applicant should perform 
the tests described for case 
B or C. 
Multipoint dissolution 
profi le should be 
performed in the 
application/compendial 
medium at 15, 30, 45, 60, 
and 120 min or until an 
asymptote is reached. 
Multipoint dissolution profi les 
should be performed in water, 
0.1 N HCl, and USP buffer 
media at pH 4.5, 6.5, and 7.5 
(fi ve separate profi les) for the 
proposed and currently accepted 
formulations. Adequate sampling 
should be performed at 15, 30, 
45, 60, and 120 min until either 
90% of drug from the drug 
product is dissolved or an 
asymptote is reached. A 
surfactant may be used, but only 
with appropriate justifi cation. 
a High - permeability, high - solubility drugs. 
b Low - permeability, high - solubility drugs. 
c High - permeability, low - solubility drugs. 
REGULATORY AGENCIES AND GUIDELINES 75

76 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
TABLE 2 Dissolution Conditions for Modifi ed - Release Dosage Form 
Extended Release Delayed Release 
In addition to application/compendial 
release requirements, multipoint 
dissolution profi les should be 
obtained in three other media, for 
example, in water, 0.1 N HCl, and 
USP buffer media at pH 4.5 and 6.8 
for the changed drug product and the 
biobatch or marketed batch 
(unchanged drug product). Adequate 
sampling should be performed, for 
example, at 1, 2, and 4 h and every 2 
hours thereafter until either 80% of 
the drug from the drug product is 
released or an asymptote is reached. 
A surfactant may be used with 
appropriate justifi cation. 
In addition to application/compendial release 
requirements, dissolution tests should be 
performed in 0.1 N HCl for 2 h (acid stage) 
followed by testing in USP buffer media, in 
the range of pH 4.5 – 7.5 (buffer stage) under 
standard (application/compendial) test 
conditions and two additional agitation speeds 
using the application/compendial test apparatus 
(three additional test conditions). Multipoint 
dissolution profi les should be obtained during 
the buffer stage of testing. Adequate sampling 
should be performed, for example, at 15, 30, 45, 
60, and 120 min (following the time from which 
the dosage form is placed in the buffer) until 
either 80% of the drug from the drug product is 
released or an asymptote is reached. The above 
dissolution testing should be performed using 
the changed drug product and the biobatch or 
marketed batch (unchanged drug product). 
(a) Changes in Components and Compositions (Table 3 ) The guideline for 
changes to approved NDA or ANDA does not defi ne these changes in detail, and 
thus the SUPAC guideline has to be followed for reference and reporting. Changes 
in excipient levels are submitted as a prior approval supplement (with accelerated 
stability data) whereas any changes in the levels of colors or fl avors are submitted 
in an annual review (long - term stability data). In MR dosage forms these changes 
have been logically categorized as (a) changes in excipient levels not affecting the 
release profi le and (b) changes in excipient levels affecting the release profi le. In 
level 2 changes for IR product and MR dosage forms for a non - narrow therapeutic 
drugs, three - month accelerated stability data of one batch (in MR dosage form for 
narrow therapeutic drugs three - month accelerated stability data of three batches) 
in a supplement and long - term stability data of one batch in an annual review should 
be submitted. Additionally, for delayed - release MR dosage forms of a narrow therapeutic 
range drug, the multipoint dissolution profi le in the buffer stage of testing 
should be generated for changed and commercial product using the medium that is 
approved or in pharmacopeia. For extended - release MR dosage forms of a narrow 
therapeutic range drug, the multipoint dissolution profi le should be generated for 
changed and commercial product using the medium that is approved or in 
pharmacopeia. 
(b) Changes in Manufacturing Site (Table 5 ) A change in the manufacturing 
or packaging site (or a contract manufacturing location) that has been approved by 
the FDA in the original application has to be evaluated for its effect on the product 
quality and performance. These changes have been described in detail in current 
guideline changes to approved NDA or ANDA. 

TABLE 3 Changes in Nonrelease Controlling Components and Composition 
Level Classifi cation 
Therapeutic 
Range/Type of 
Drug Test Documnetation 
Filing 
Documentation 
I 
Complete or partial 
deletion 
of color/fl avor 
Change in inks, 
imprints 
SUPAC - IR level 1 
excipient ranges 
No other changes 
All drugs Stability 
Application/compendial requirements 
No biostudy 
Annual report 
II 
Change in technical 
grade and/or 
specifi cations 
Higher than 
SUPAC - IR level 
1 but less than 
level 2 excipient 
ranges 
No other changes 
All drugs for 
MR 
Depending upon 
therapeutic 
range 
solubility and 
permeability 
(as per BCS) 
for IR 
MR (ER): 
Notifi cation and updated 
batch record 
Stability 
Application/ 
compendial 
requirements plus 
multipoint dissolution 
profi les in three other 
media (e.g., water, 
0.1 
N HCl, and USP 
buffer media at pH 4.5 
and 6.8) until . 80% of 
drug released or an 
asymptote is reached 
Apply some statistical 
test (f2 test) for 
comparing dissolution 
profi les 
No biostudy 
MR (DR): 
Notifi cation and updated 
batch record 
Stability 
Application/compendial 
requirements 
plus multipoint 
dissolution profi les in 
additional buffer stage 
testing (e.g., USP 
buffer media at pH 
4.5 – 7.5) under 
standard and increased 
agitation conditions 
until . 80% of drug 
released or an 
asymptote is reached 
Apply some statistical 
test (f2 test) for 
comparing dissolution 
profi les 
No biostudy 
IR: 
Notifi cation and 
updated batch 
record 
Stability 
Dissolution 
requirements: 
case A, case B, or 
case C 
Apply some 
statistical test (f2 
test) for comparing 
dissolution profi les 
No biostudy 
Prior approval 
supplement 
Change in technical 
grade and/or 
specifi cations 
Higher than 
SUPAC - IR 
level 1 
No other changes 
77

Level Classifi cation 
Therapeutic 
Range/Type of 
Drug Test Documnetation 
Filing 
Documentation 
III 
Higher than 
SUPAC - IR level 
2 excipient ranges 
for MR and IR, 
change in 
excipient range 
for low solubility 
and low 
permeability 
drugs beyond 
level 1 
All drugs for 
MR and all 
drugs failing 
dissolution 
criteria for 
level 2 for IR 
Updated batch record 
Application/compendial (profi le) requirements and as 
mentioned for level II 
Stability 
Biostudy or ivivc 
Updated batch 
record 
Dissolution profi le as 
for level II 
Stability 
Biostudy or ivivc 
Prior approval 
supplement 
Note : 
MR, 
Modifi ed - release dosage form; ER, extended - release dosage form; DR, delayed - release dosage form; IR, immediate - release dosage form. 
TABLE 3 Continued 
78

TABLE 4 Changes in Release Controlling Components and Composition 
Level Classifi 
cation Therapeutic Range Test Documentation 
Filing Documentation 
I 
. 5% w/w change based 
on total release 
controlling excipient 
(e.g., controlled - 
release polymer, 
plasticizer) content 
No other changes 
All drugs Stability 
Application/compendial requirements 
No biostudy 
Annual report 
II 
Change in technical 
grade and/or 
specifi cations 
. 10% w/w change based 
on total release 
controlling excipient 
(e.g., controlled - 
release polymer, 
plasticizer) content 
No other changes 
Nonnarrow 
MR (ER): 
MR (DR): 
Prior approval 
supplement Notifi cation and updated 
batch record 
Stability 
Application/compendial 
requirements plus 
multipoint dissolution 
profi les in three other 
media (e.g., water, 0.1 N 
HCl, and USP buffer media 
at pH 4.5 and 6.8) until 
. 80% of drug released or 
an asymptote is reached 
Apply some statistical test (f2 
test) for comparing 
dissolution profi les 
No biostudy 
Notifi cation and updated 
batch record 
Stability 
Application/compendial 
requirements plus 
multipoint dissolution 
profi les in additional 
buffer stage testing (e.g., 
USP buffer media at pH 
4.5 – 7.5) under standard 
and increased agitation 
conditions until . 80% of 
drug released or an 
asymptote is reached 
Apply some statistical test 
(f2 test) for comparing 
dissolution profi les 
No biostudy 
Narrow Updated batch record 
Stability 
Application/compendial (profi le) requirements 
Biostudy or ivivc 
Prior approval 
supplement 
III 
> 10% w/w change 
based on total release 
controlling excipient 
(e.g., controlled - 
release polymer, 
plasticizer) content 
All drugs Updated batch record and stability 
Application/compendial (profi le) requirements 
Biostudy or ivivc 
Prior approval 
supplement 
79

TABLE 5 Site Changes 
Level Classifi cation 
Therapeutic 
Range Test Documentation Filing Documentation 
I Single facility 
Common 
Personnel 
No other changes 
All drugs 
Application/compendial requirements 
No biostudy 
Annual report 
II 
Same contiguous 
campus 
Common personnel 
No other changes 
All drugs MR (ER): 
MR (DR): 
IR: 
Changes being effected 
supplement 
(accelerated stability 
data for MR and no 
stability data for IR) 
Annual report (long - 
term stability data for 
MR and IR) 
Identifi cation and 
description of site 
change and updated 
batch record 
Notifi cation of site 
change 
Stability 
Application/ 
compendial 
requirements plus 
multipoint 
dissolution profi les 
in three other media 
(e.g., water, 0.1 N 
HCl, and USP buffer 
media at pH 4.5 and 
6.8) until . 80% of 
drug released or an 
asymptote is reached 
Apply some statistical 
test (f2 test) for 
comparing 
dissolution profi les 
No biostudy 
Identifi cation and 
description of site 
change, and updated 
batch record 
Notifi cation of site 
change 
Stability 
Application/ 
compendial 
requirements plus 
multipoint 
dissolution profi les 
in additional buffer 
stage testing (e.g., 
USP buffer media at 
pH 4.5 – 7.5) under 
standard and 
increased agitation 
conditions until 
. 80% of drug 
released or an 
asymptote is reached 
Apply some statistical 
test (f2 test) for 
comparing 
dissolution profi les 
No biostudy 
Identifi cation and 
description of site 
change and 
updated batch 
record 
Notifi cation of site 
change 
Stability 
Application/ 
compendial 
requirements 
No biostudy 
III 
Different campus 
Different personnel 
All drugs Notifi cation of site change 
Updated batch record 
Application/compendial (profi le) requirements 
(as for level II) 
Stability 
Biostudy or ivivc 
Notifi cation of site 
change 
Updated batch 
record 
Case B dissolution as 
for excipient 
change (level II) 
Stability 
No biostudy 
Prior approval 
supplement 
(accelerated stability 
data) for MR and 
changes being 
effected supplement 
for IR 
Annual report 
80

TABLE 6 Changes in Batch Size: 
Scale - Up/Scale - Down 
Level Classifi 
cation Change Test Documentation 
Filing 
Documentation 
I 
Scale - up of biobatch(s) 
or pivotal clinical 
batch(s) 
No other changes 
. 10 . (all drugs) Updated batch record 
Stability 
Application/compendial requirements 
No biostudy 
Annual report 
II 
Scale - up of biobatch(s) 
or pivotal clinical 
batch(s) 
No other changes 
> 10 . (all drugs) 
MR (ER): 
MR (DR): 
IR: 
Changes being 
effected 
supplement 
(accelerated 
stability data) 
Annual report 
(long - term 
stability data) 
Updated batch record 
Stability 
Application/ 
compendial 
requirements plus 
multipoint 
dissolution profi les 
in three other 
media (e.g., water, 
0.1 
N 
HCl, 
and 
USP buffer media 
at pH 4.5 and 6.8) 
until . 80% of drug 
released or an 
asymptote is 
reached 
Apply some statistical 
test (f2 test) for 
comparing 
dissolution profi les 
No biostudy 
Updated batch record 
Stability 
Application/compendial 
multipoint dissolution 
profi les in additional 
buffer stage testing 
(e.g., USP buffer 
media at p H 
4.5 – 7.5) 
under standard and 
increased agitation 
conditions until . 80% 
of drug released or an 
asymptote is reached 
Apply some statistical 
test (f2 test) for 
comparing dissolution 
profi les 
No biostudy 
Updated batch 
record 
Stability 
Case B dissolution 
as for excipient 
change (level II) 
No biostudy 
81

TABLE 7 Changes in Manufacturing: equipment 
Level Classifi cation Change Test Documentation Filing Documentation 
I Equipment 
changes 
No other changes 
(all drugs) 
Alternate 
equipment of 
same design 
and principle 
Automated 
equipment 
Updated batch record 
Stability 
Application/compendial requirements 
No biostudy 
Annual report 
II 
Equipment 
changes 
No other changes 
(all drugs) 
Change to 
equipment of 
a different 
design and 
operating 
principle 
MR (ER): 
MR (DR): 
IR: 
Prior approval 
supplement 
(accelerated 
stability data) 
Annual report (long - 
term stability data) 
Updated batch record 
Stability 
Application/ 
compendial 
requirements plus 
multipoint 
dissolution profi les 
in three other media 
(e.g., water, 0.1 N 
HCl, and USP 
buffer media at pH 
4.5 and 6.8) until 
. 80% of drug 
released or an 
asymptote is 
reached 
Apply some statistical 
test (f2 test) for 
comparing 
dissolution profi les 
No biostudy 
Updated batch 
record 
Stability 
Application/ 
compendial 
requirements plus 
multipoint 
dissolution 
profi les in 
additional buffer 
stage testing (e.g., 
USP buffer media 
at ph 4.5 – 7.5) 
under standard 
and increased 
agitation 
conditions until 
. 80% of drug 
released or an 
asymptote is 
reached 
Apply some 
statistical test 
(f2 test) for 
comparing 
dissolution 
profi les 
No biostudy 
Updated batch 
record 
Stability 
Case C dissolution 
as for excipient 
change (level II) 
No biostudy 
82

TABLE 8 Changes in Manufacturing: Processes 
Level Classifi cation Change Test Documentation 
Filing 
Documentation 
I 
Processing changes 
affecting the 
nonrelease/release 
controlling excipients 
for MR 
Changes within 
validation ranges (IR) 
No other changes 
Adjustment of 
equipment 
operating 
conditions (mixing 
times, operating 
speeds) — within 
approved 
application ranges 
Updated batch record 
Application/compendial requirements 
No biostudy 
Annual report 
II 
Processing changes 
affecting the 
nonrelease controlling 
excipients and/or the 
release controlling 
excipients 
Processing changes 
outside validation 
ranges for IR 
No other changes 
Adjustment of 
equipment 
operating 
conditions (e.g. 
mixing times, 
operating speeds, 
etc.) 
Beyond approved 
application ranges 
MR (ER): 
MR (DR): 
IR: 
Changes being 
effected 
supplement 
(accelerated 
stability data 
for MR) 
Annual report 
(long - term 
stability data 
for MR and 
IR) 
Updated batch record 
Stability 
Application/compendial 
requirements plus 
multipoint dissolution 
profi les in three other 
media (e.g. water, 
0.1 
N HCl, and USP 
buffer media at pH 
4.5 and 6.8) until 
. 80% of drug 
released or an 
asymptote is reached 
Apply some statistical 
test (f2 test) for 
comparing dissolution 
profi les 
No biostudy 
Updated batch record 
Stability 
Application/compendial 
requirements plus 
multipoint dissolution 
profi les in additional 
buffer stage testing (e. 
g., USP buffer media at 
pH 4.5 – 7.5) under 
standard and increased 
agitation conditions 
until . 80% of drug 
released or an 
asymptote is reached 
Apply some statistical test 
(f2 test) for comparing 
dissolution profi les 
No biostudy 
Notifi cation of change 
Updated batch record 
Stability 
Case B dissolution as 
for excipient 
change (level II) 
No biostudy 
III 
Processing changes 
affecting the 
nonrelease controlling 
excipients and/or the 
release controlling 
excipients 
Change in the type 
of process used 
(e.g. from wet 
granulation to 
dry) 
Updated batch record 
Stability 
Application/compendial (profi le) requirements 
Biostudy or ivivc 
Updated batch 
record — stability 
Case B dissolution as 
for excipient 
change (level II) 
No biostudy 
Prior approval 
supplement 
(accelerated 
stability data 
for MR) 
Annual report 
(long - term 
stability data 
for MR and 
IR) 
83

84 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
The FDA should be notifi ed of the new location. For any type of moderate 
changes, accelerated stability data of one batch should be submitted with the CBE 
supplement for MR dosage forms and long - term stability data of one batch should 
be submitted in the annual review for IR as well as MR dosage forms. Additionally, 
the CBE should be submitted to the FDA in case of any moderate change. Stability 
requirements for any major change are the same as those mentioned in the previous 
section, that is, three - month accelerated stability data of three batches (signifi cant 
body of information not available) or three - month accelerated stability data for one 
batch (signifi cant body of information available) have to be submitted in a prior 
approval supplement along with long - term stability data for one batch in the annual 
review. 
(c) Scale - Up/Scale-Down (Table 6 ) A change in the batch size of a drug product, 
either scale - up or scale - down, is likely to induce some changes in the operation 
parameters. This in turn can adversely affect product quality. 
(d) Changes in Manufacturing Equipment (Table 7 ) and Process (Table 8 ) Any 
manufacturing changes in equipment and process are included in this section. For 
example, a change in the blending equipment from octagonal blender to double 
cone blender or a change in the granulation process from wet to dry granulation 
calls for submission of proper validation documentation for FDA approval. All 
these changes along with reporting categories have been described in current guidelines 
for changes to approved NDA or ANDA. 
Biowaivers In vitro and in vivo approaches are commonly used for establishment 
of bioavailability and bioequivalence. Dissolution studies are used as in vitro 
approaches and also serve as quality control tools for pharmaceuticals. Under 
certain circumstances, in vitro dissolution may also act as a surrogate marker 
for in vivo biostudy and enable the establishment of in vitro and in vivo 
bioequivalence. 
“ CDER Guidance for Industry: Waiver of In Vivo Bioavailability and Bioequivalence 
Studies for IR Solid Dosage Form Based on Biopharmaceutic Classifi cation 
System (BCS) ” recommends waiving an in vivo biostudy under specifi c circumstances. 
For example, a waiver of the in vivo biostudy of one or more lower strengths 
is acceptable based on the correlation data and in vivo bioequivalence of the higher 
strength, provided all strengths are proportionally equivalent in terms of active and 
inactive ingredients. A biostudy on a lower strength may also be requested based 
on safety reasons (as for mitrazapine tablets) and a biowaiver for highest strength 
is acceptable provided elimination kinetics is linear over a dose range, strengths are 
proportional, and comparative dissolution data of all strengths are acceptable. 
The BCS classifi es drugs in four classes: 
Class I: high solubility, high permeability 
Class II: low solubility, high permeability 
Class III: high solubility, low permeability 
Class IV: low solubility, low permeability 
Dissolution, solubility, and permeability are the three fractors that control the 
bioavailability of a drug for an IR drug product. Provided the inactive excipient 

does not control or modify the release and absorption of the active ingredient, the 
biostudy may be waived. According to the guideline, the solubility class is determined 
for the highest dose strength of a drug product for which a biowaiver has 
been requested. When the highest dose strength of a solid dosage form is soluble in 
250 mL of water or less across a pH range of 1 – 7.5, it is considered as highly soluble. 
For determination of permeability class various in vivo methods like mass balance, 
absolute bioavailability, and intestinal perfusion approaches and in vitro methods 
like permeation studies using excised tissue or monolayer of cultured epithelial cells 
are used. When extent of absorption is greater than 90% of the administered dose 
in humans, it is considered as highly permeable. For a dissolution study, drug release 
should be evaluated in three media that are 0.1 N HCl or USP - simulated gastric 
fl uid without enzymes, pH 4.5 buffer, and pH 6.8 buffer or USP - simulated intestinal 
fl uid without enzymes. Rapidly dissolving drug products are those that dissolve 
more than 80% in 900 mL of the above - mentioned media in less than 30 min using 
USP apparatus at 100 rpm (or USP II apparatus of 50 rpm). 
A biowaiver can be requested for the postchange products if it falls under class 
I of the BCS and displays a rapidly dissolving profi le and there is a similarity (as 
determined by f2 test) between the pre - and postchanged drug product in all three 
media. For BCS class II drugs, a meaningful correlation (level A, B, or C correlation) 
between in vitro drug release and in vivo absorption also may be used for requesting 
the biowaiver. Deconvolution techniques are used for prediction of in vivo dissolution 
and absorption. 
1.3.3.2 Regulations Guidance on SUPAC by Pharmaceutical Unit of EU 
The pharmaceutical market in European countries is one of the largest in the world. 
To ensure that the EU promotes pharmaceutical trade and ensures safety, effi cacy, 
and quality of medicinal products within the European member states, the pharmaceutical 
unit of the EU runs a series of information and communication projects, 
collectively called EUDRA projects. Out of these projects, the EUDRALEX pharmaceutical 
unit is responsible for making community pharmaceutical legislation, 
guidelines, and notices for applicants [9] . Under Volume 2, Section C, of Regulatory 
Guidelines (Pharmaceutical Legislation: Notice to Applicants) of Eudralex, “ Guideline 
on Dossier Requirements for Type IA and Type IB Notifi cations ” has been 
provided [2] . 
Regulations were introduced to lessen the administrative load on the authority 
and to simplify the procedure for granting a postapproval variation without negotiating 
any quality attribute of drug product [10] . Under these regulations, type IA 
and type IB were defi ned; also, clearcut terms were introduced for extension application, 
parallel/consequential notifi cation/variation, and urgent safety restriction. 
For streamline operation of these regulations, four documents have been 
prepared: 
(a) A procedural guidance for the member states (reference or concerned) and 
the applicant for notifi cations/variations in the mutual recognition procedure 
(b) A procedural guidance for the applicant for notifi cations/variations in the 
centralized procedure 
REGULATORY AGENCIES AND GUIDELINES 85

86 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
(c) A common application form which may be used for type IA and type IB 
notifi cations or type II variations in both the centralized and mutual recognition 
procedures 
(d) A guideline on the documentation to be submitted for type IA and type IB 
notifi cations 
All member states of the EU follow the same regulations for a change or “ variation 
” in an already approved medical product. As per the guidance, three types of 
variations or changes have been identifi ed — type I variation, which is further classi- 
fi ed into types IA and IB and type II variations. The guidelines classify some specifi c 
changes in type IA or IB. It also provides specifi c data analysis required for variation 
and the types of document that need to be submitted to the regulatory authority. 
Any change that is not listed in this section is classifi ed as type II variation. 
According to Commission Regulation (EC, No. 1084/2003), type I variation has 
been defi ned as “ A ‘ minor variation ’ of type IA or type IB means a variation listed 
in Annex I, which fulfi ls the conditions, set out therein. ” Annex I of the regulation 
provides a list of changes and conditions (to be satisfi ed) to be classifi ed as type IA 
or type IB variation and Annex II provides changes falling under the extension 
application category. Type II variations in proposed documentation are not type I 
or extension application. There is also a provision for “ urgent safety restrictions. ” 
These are any temporary or provisional changes in the product summary characteristics, 
such as indications, posology, contraindications, warnings, target species, and 
withdrawal periods, as result of a new information that may cause signifi cant safety 
concerns about the medicinal product [11] . 
Any change arising from the primary change has to be notifi ed separately. Consequential 
changes form part of the same notifi cation whereas parallel changes do 
not. A consequential change to type IA can only be another type IA whereas a 
consequential change to type IB can be type IA or type IB. All other variations 
should be submitted as Type II variations. “ Guideline on Dossier Requirements for 
Type IA and Type IB Notifi cations ” provides a complete list of all changes, conditions 
required to be met for the particular change, and documentation required by 
the regulatory authority [10] . 
1.3.3.3 Regulatory Guidance on SUPAC by Agencia Nacional de 
Vigilancia Sanitaria 
Agencia Nacional de Vigilancia Sanitara (ANVISA) issues Brazil ’ s generic drug 
policy. Under legislation for industry, Resolution RE N ° 893, of May 29, 2003, is 
described in “ Guide for Making Post - Registration Alterations, Inclusions and Noti- 
fi cations of Drug Products ” [6] . This guideline describes postregistration changes as 
“ alterations ” and “ inclusions ” and also tells about the documentation and assays 
that need to be submitted in support of any type of change. As per the guideline, 
each type of alteration or inclusion has to be submitted separately and approved 
by the ANVISA before it can be implemented. Table 9 presents some examples for 
each category. 
Under each category, certain requirements have to be met before its implementation. 
For example, for inclusion in the batch size, the company should notify, in 
alteration, if the included batch size is more than 10 times. The documentation that 
needs to be submitted includes the original proof of payment of fee or of exemption; 

a copy of the certifi cate of good manufacturing and control practices (CBPFC) 
issued by ANVISA; technical justifi cation; production and quality control records 
of one batch of each strength of the product; a technical report; and a technical 
report and assessment of the dissolution profi le. 
1.3.4 HARMONIZATION 
It is essential to evaluate the safety and quality of new or changed medical products 
before they reach the market. However, the need to set specifi c guidelines has been 
recognized at different times in different countries. For example, in the United 
States a tragic incident with a junior paracetamol formulation was the alarm to initiate 
guidelines for authorization of medical products. European countries followed 
this trend in the 1960s after the thalidomide incident. Since then there have been a 
large number of guidelines that have been put into place to evaluate medical products 
in terms of their quality, safety, and effi cacy [12] . However, with the pharmaceutical 
industries becoming international and aiming for a worldwide market, there 
is a move toward internationally accepted guidelines and approval systems. In order 
for medical products to be marketed internationally, companies have found it necessary 
to duplicate many tests and studies that are time consuming and broad. 
TABLE 9 Examples of Different Categories 
Postregistration 
Alterations 
Postregistration 
Inclusions 
Postregistration 
Notifi cations 
Postregistration 
Cancellation 
Labeling alteration Inclusion of new 
commercial 
presentation 
Temporary 
suspension of 
manufacture 
Cancellation upon 
request of 
registration of 
drug 
presentation 
Alteration of corporate 
name 
Inclusion of new 
packing 
Resumption 
of drug 
manufacture 
Cancellation of 
drug registration 
Alteration of date of 
expiry 
Inclusion of new 
concentration 
already approved in 
country 
Alteration of 
preservation 
conditions 
Inclusion of new 
dosage form already 
approved in country 
Alteration of synthesis 
path of drug 
Inclusion of new 
therapeutic 
indication in country 
Alteration of 
manufacturer of 
drug 
Inclusion of 
manufacture site 
Alteration of 
manufacturing site 
Inclusion of 
manufacturer of 
drug 
Alteration of excipient Inclusion in the batch 
size 
HARMONIZATION 87

88 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
Table 10 shows examples of documentation required by different countries when 
a postapproval change is made during the manufacturing process of medical products. 
When there are changes in the specifi cation of an excipient, the documents 
required by the TGA and European Agency for Evaluation of Medicinal Products 
(EMEA) are variable. Furthermore this would indicate that the benefi t of a patent/ 
medical product might not reach globally. There are also many chances of making 
an error. For example, the 2003 recall of clotrihexal 100 - mg vaginal tablets in New 
Zealand pharmacies was due to the fact that clotrihexal was packed according to 
TGA guidelines and thus its sale was prohibited in New Zealand. This resulted in 
much confusion and problems among patients and medical professionals. Harmonization 
is the process by which the pharmaceutical industries worldwide adopt 
the same laws and regulations. Harmonization is intended to assure the safety, 
quality, and effi cacy of a medical product globally. The main goal of harmonization 
is to recognize and minimize the differences in the scientifi c requirements for 
medical product development within different regulatory agencies in different 
countries. 
Harmonization activities focus on reducing and simplifying the types of studies 
that the pharmaceutical industries need to carry out in order to register a medical 
product in another country, protocols to be followed when performing these studies, 
techniques used to validate supporting data, and techniques used to perform risk 
assessment. 
Harmonization reduces replications and unnecessary production and registration 
of new and changed products. The concept of harmonization was explored by 
European countries in the 1980s. The success of harmonization in these countries 
has demonstrated that it is practical and possible. Following harmonization in 
Europe the International Conference on Harmonization (ICH) was set up in 1990 
[13] . Table 11 shows some of the harmonized rules that have been successfully 
developed by ICH. 
TABLE 10 Changes in Specifi cation of Excipients (Addition of New Test Limit): 
Comparison between Guidelines 
Guidelines Documentation Type of Change 
TGA Details of the test method must be provided. 
Appropriate validation data have been generated for 
the test method. 
The limits proposed are based on batch analytical data 
and are in compliance with offi cial standard and/or 
relevant accepted guidelines if applicable. 
Self - assessable 
changes 
EMEA Comparative table of current and proposed 
specifi cations. 
Batch analysis data on two production batches for all 
tests in the new specifi cation. 
Where appropriate, comparative dissolution profi le 
data for the fi nished product on at least one pilot 
batch containing the excipient complying with the 
current and proposed specifi cation. For herbal 
medicinal products, comparative disintegration data 
may be acceptable. 
Minor change 
type IB 
requires 
approval 

1.3.5 GMP ISSUES: CHANGE CONTROL AND PROCESS VALIDATION 
Changes are unavoidable in a manufacturing setup. Manufacturers make changes 
at some stage of manufacturing during and after approval of a product. However, 
consistent quality of a drug product can only be assured through well - defi ned validation 
procedures. When a change is made in the manufacturing process of a drug 
product, sponsors are responsible for evaluating the effect of any change on the 
safety, effi cacy, quality, stability, and potency of a drug product and ensuring that 
these properties are not infl uenced by the change. In a manufacturing setup, 
various disciplines like sales, marketing, medical, regulatory affairs, manufacturing, 
electrical, and technical services work together. Hence, any kind of change in one 
discipline will have direct consequences on other disciplines. Each company should 
have a procedure with regard to handling a change. Quality control and quality 
assurance departments usually keep track of various changes occurring in a 
GMP environment. Therefore, it is required that personnel performing the job are 
trained enough to assess the effect of any kind of change or variation and take 
appropriate action for its evaluation or control. Supporting data should be generated 
and once evaluated can confi rm whether further clinical or nonclinical studies 
are required. 
1.3.5.1 Change Control 
When a change is made in a manufacturing setup, it is important to assess its impact. 
As a change can have impact on regulatory fi ling, manufacturing parameters, speci- 
fi cations, and technical services, it is important to consider the concerns and objections 
of various disciplines involved and only through well - defi ned standard 
operating procedures should it be properly validated, evaluated, and fi nally implemented. 
A properly defi ned order of evaluation of a change with strategic input of 
trained personnel is key to delivering a consistent quality product (Figure 1 ). 
When a change is the processed, the manufacturer should have protocols in place 
with regard to assessing the change. Therefore, “ control of change ” is important. 
Control can be implemented effectively only through well - defi ned standard operating 
procedures. The main purpose of “ change control ” exercise is to have a 
TABLE 11 Example of Quality Guidelines Harmonized by ICH 
Quality Topic Example of Guideline 
Q1: Stability Q1B: Photostability testing 
Q2: Validation of analytical procedure Q2A: Methodology 
Q3: Impurity testing Q3A: Impurities in new drug 
substances 
Q4: Pharmacopoeias Q4: Pharmacopoeial harmonization 
Q5: Quality of biotechnological products Q5A: Viral safety evaluation of 
biotechnological products 
Q6: Specifi cations for new drug substance and 
products 
Q6A: Acceptance criteria for new drug 
substances 
Q7: GMP for pharmaceutical ingredients Q7A: GMP for active pharmaceutical 
ingredients 
GMP ISSUES: CHANGE CONTROL AND PROCESS VALIDATION 89

90 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
systematic process in place to accurately evaluate a change using specifi c tests. 
Moreover, it aims to measure the effects on quality safety and effi cacy before a 
change is implanted. Change control and its evaluation through proper documentation 
should include [14] : 
(a) Description and purpose of change 
(b) Inputs from research and development (R & D) department 
(c) Evaluation steps for impact assessment, such as evaluation of stability, validation 
requirements, and in vivo bioequivalence requirement 
(d) Need and extent of regulatory documentation and approval 
(e) Implementation schedule 
(f) Clear defi nition of personnel authorized for change approval 
(g) Monitoring protocol for change implementation and periodic review of 
impact 
Following the informal proposal of a change, it should be reviewed by the responsible 
initiator, who will then generate a formal proposal [15] . The proposal should 
describe accurately what the change is concerned with, how to validate the change, 
and the time frame within which the change should be implemented. The fi nal proposal 
should be reviewed and assessed by all functional groups involved. Once the 
change is approved, it can be implemented and the change cycle is completed. Figure 
2 describes responsibilities of different departments of a pharmaceutical company, 
FIGURE 1 Change control cycle for change in manufacturing process. 
Comments 
And Signature 
Comments 
And Signature 
Comments 
And Signature 
Comments 
And Signature 
Change 
Implement 
Initiator 
Approved by 
Quality 
Assurance 
Approved by 
Production 
Department 
Approved by 
Regulatory 
Department 
Approved by 
Head of 
Department 
Change Control Cycle 

in the change control procedure. Standard operating procedures (SOPs) for change 
control are an important part of any GMP audit. Hence it is important that it is 
implemented by trained and qualifi ed personnel from appropriate disciplines. 
After a change has been approved by all functional groups within the manufacturing 
setup and if it has no regulatory concerns, it can be implemented immediately. 
However, if the impact comes under any regulatory domain, the company may have 
to wait for regulatory approval. 
1.3.5.2 Process Validation 
Process validation is an important part in the implementation of a postapproval 
change. It establishes the documented evidence of conformance of a pharmaceutical 
operation in accordance with specifi cations. FDA “ Guideline on General Principles 
of Process Validation ” describes in detail the principles and practices of process 
validation and documentation required by the regulatory authority [13] . In general 
terms, process validation may be defi ned as the procedure which generates suffi cient 
assurance and documented evidence that a particular operation is operating 
and producing drug products in accordance with the specifi cations and process 
controls. 
FIGURE 2 Responsibilities of different disciplines of a pharmaceutical company in a 
change control procedure ( modifi ed from ref. 15 ). 
INITIATION OF 
CHANGE CONTROL 
PROOFREADING 
CONFORMANCE TO CGMP AND 
APPLICABILITY TO OTHER SYSTEMS 
REVIEW AND APPROVAL 
REGULATORY IMPACT AND 
WORLDWIDE FILING STRATEGY 
VALIDATION 
Quality assurance, Quality control, 
Manufacturing, Process Engineering, 
Technical services, Regulatory affairs, 
Owner of system or procedure being 
changes 
Quality assurance 
Quality assurance 
Regulatory affairs 
Quality assurance, Quality control, 
Manufacturing, Process Engineering, 
Technical services, Regulatory affairs, 
Owner of system or procedure being 
changes 
Quality assurance 
GMP ISSUES: CHANGE CONTROL AND PROCESS VALIDATION 91

92 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
Prospective validation, retrospective validation, concurrent validation, and revalidation 
are the four validation components. Prospective validation is performed 
before the distribution of drug products in the market or after the manufacturing 
of a drug product using revised changes that can affect product quality and characteristics. 
Retrospective validation is conducted for an established drug product 
whose manufacturing process is stable to ensure that the current pharmaceutical 
operation is performing as per the protocols and specifi cation and yielding satisfactory 
product. Concurrent validation is conducted by monitoring in - process critical 
manufacturing parameters and end - product testing to ensure that the current manufacturing 
process is per the in - process control specifi cations. Revalidation is performed 
after changes to an approved drug product are implemented to ascertain 
that there is no adverse effect on the quality and performance of a drug product 
[16] . 
During a validation process, the products and processes are subjected to testing 
at extreme conditions of in - process limits and their performance is evaluated against 
the acceptance criteria. The parameters of different pharmaceutical operations are 
varied and product properties are recorded and evaluated (Figure 3 ). When it is 
found that adjustment is required, necessary actions are taken in consultation with 
R & D personnel. Generally, validation data of three production scale batches are 
compared to generate a high level of quality assurance. 
Systematic documentation of the effect on the product attributes by varying 
various process parameters is very important in the validation process. The product 
development team, engineering and technical services, and production and regulatory 
departments are also consulted while making any process change or before 
fi nalizing any validation protocol or report. Depending on the “ level ” of change or 
degree of effect to be produced, the extent of the validation is determined. 
Based on the validation requirements, samples are collected at different stages 
and submitted for analysis per the validation protocol. The data are fi nally compiled 
in the form of a validation report. A systematic validation protocol and validation 
report are the backbone of the validation process. Table 12 gives key components 
of any validation activity [16] . These protocols and reports should be verifi ed and 
approved by the relevant functions. 
Some changes are often made in the manufacturing process without prior notifi - 
cation, and hence it is advisable to consider revalidation at predetermined frequencies 
(or whenever an unusual behavior is noted). 
When new equipment is purchased or there is a change in the manufacturing site, 
qualifi cation exercises are performed as part of the validation process. Qualifi cation 
(installation qualifi cation, operation qualifi cation, and performance qualifi cation) 
for any equipment or facility is an extreme process which involves testing, verifi cation, 
and documentation to assure that the particular equipment or facility is per 
the specifi cation and meets the appropriate standards as defi ned by vendor and 
required by manufacturing and engineering personnel [14] . 
1.3.6 CONCLUSION 
The global pharmaceutical industry is continuously growing in a rapidly changing 
and dynamic environment of the health care sector. New drugs and delivery systems 

FIGURE 3 Various process parameters and product characteristics associated with validation 
activity of typical coated tablet. 
Wet Granulation 
Drying 
Milling 
Blending 
Tabletting 
Coating 
Process Parameters Product Characteristics 
Premix time, Binder addition 
time, Impeller/Chopper Speed, 
Inlet air temperature, Bed Temperature, 
Airflow rate, Raking frequency 
Screen size, Hammer/knives 
direction and speed 
Compression speed, force (hardness 
and thickness), tablet weight 
Pan capacity, inlet/exhaust 
temperatures, Pan Speed, spray 
rate, Air flow rate, Bed temperature 
Capacity of Blender, Mixing 
time, mixing speed 
Granule hardness and size 
distribution 
Moisture content of Granules, 
Amount of residual solvent 
Size distribution, Bulk/ Tapped 
density of granules 
Content uniformity in blender 
and drum, Bulk/ Tapped density 
of final blend 
Tablet weight, hardness, thickness, 
content uniformity, friability, 
dissolution, disintegration, Assay/ 
potency 
Dissolution, disintegration, coating 
weight gain, mottling, assay/ potency 
CONCLUSION 93 
TABLE 12 Key Components of Validation Activity 
Validation Protocol Validation Report 
Purpose of study Aim of study 
Personnel responsibility List of raw material used in study 
Critical process steps List of manufacturing equipment 
Critical process parameters Critical steps studied 
Critical product parameters Collected data and its analysis 
Sampling plan Acceptance criteria evaluation 
Testing plan Statistical analysis 
Acceptance criteria Recommendations by validation department 

94 SCALE-UP AND POSTAPPROVAL CHANGES (SUPAC) REGULATIONS 
surface each year in the market. To maintain the quality of new and existing drugs 
and delivery technologies, pharmaceutical operations are controlled by regulatory 
guidelines. The purpose of developing guidelines is to keep the health and safety of 
a person on the highest priority by delivering quality pharmaceuticals. Implementation 
of these guidelines and systematic follow - up of the effect of postapproval 
changes in the form of documentation are essential to safeguard against any possible 
failure of the whole system. Change control and validation ensure that there is no 
deleterious impact on the drug product characteristics. Anticipated changes incorporated 
in comparability protocols reduce signifi cant risk of experiencing unpredictable 
adverse effects and help to introduce the product in less time. When an impact 
is anticipated, it should be properly discussed with R & D, process development, 
and other concerned departments for appropriate regulatory fi ling by following 
regulatory guidelines. Provided that these guidelines are followed properly, quality 
and performance of a drug product can be ensured. 
REFERENCES 
1. U.S. Department of Health and Human Services, Food and Drug Administration, Centre 
for Drug Evaluation and Research (CDER) , Guidance for industry: Changes to an 
approved NDA or ANDA, available, 
accessed Apr. 15, 2006 . 
2. European Commission , Guideline on dossier requirements for type IA and type IB 
notifi cations: Pharmaceuticals: Regulatory framework and market authorizations, available: 
http://ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol - 2/c/gdvartypiab_rev0_ 
200307.pdf , accessed Apr. 20, 2006 . 
3. Department of Health and Ageing , Therapeutic Goods Administration, Australian regulatory 
guidelines for prescription medicines. Appendix 12: Changes to the quality information 
of registered medicines: Notifi cation. Self - assessment and prior approval, available: 
http://www.tga.gov.au/pmeds/argpmap12.pdf , accessed Apr. 12, 2006 . 
4. Food and Drug Administration, Centre for Drug Evaluation and Research (CDER) , 
Guidance for industry: SUPAC - IR: Immediate - release solid oral dosage forms: Scale - up 
and post - approval changes: Chemistry, manufacturing and controls, in vitro dissolution 
testing, and in vivo bioequivalence documentation, available: http://www.fda.gov/cder/ 
guidance/cmc5.pdf , accessed May 11, 2006 . 
5. Food and Drug Administration, Centre for Drug Evaluation and Research (CDER) , 
Guidance for industry: Comparability protocols — Chemistry, manufacturing, and controls 
information (draft), available, accessed Apr. 15, 2006 . 
6. Brazilian Sanitary Surveillance Agency (ANVISA) , Resolution: Guide for making post - 
registration alterations, inclusions and notifi cations of drug products, Brazil ’ s generic drug 
policy, industry legislation, available: http://www.anvisa.gov.br/hotsite/genericos/legis/ 
resolucoes/893_03re_e.htm , accessed Apr. 11, 2006 . 
7. U.S. Department of Health and Human Services, Food and Drug Administration, Centre 
for Drug Evaluation and Research (CDER) , Guidance for industry: SUPAC - MR: Modi- 
fi ed release solid oral dosage forms scale - up and postapproval changes: Chemistry, 
manufacturing, and controls; in vitro dissolution testing and in vivo bioequivalence documentation, 
, accessed May 5, 
2006 . 

8. U.S. Department of Health and Human Services, Food and Drug Administration, Centre 
for Drug Evaluation and Research (CDER) , Guidance for industry: SUPAC - SS: Nonsterile 
semisolid dosage forms; scale - up and post - approval changes: Chemistry, manufacturing 
and controls; in vitro release testing and in vivo bioequivalence documentation. 
9. Pharmaceutical Unit of European Commission at EUROPA, available: http://ec.europa. 
eu/enterprise/pharmaceuticals/pharmacos/docs/brochure/pharmaeu.pdf , accessed May 
21, 2006 . 
10. Variations. Pharmaceuticals: Regulatory framework and market authorizations, Chapter 
5, in Procedures for Marketing Authorisation , Vol. 2A, European Commission, available: 
http://ec.europa.eu/enterprise/pharmaceuticals/eudralex/vol - 2/a/v2a_chap5_r1_2004 - 02. 
pdf , accessed Apr. 16, 2006 . 
11. Commission regulation (EC) No. 1084/2003 of June 3, 2003, concerning the examination 
of variations to the terms of a marketing authorisation for medicinal products for human 
use and veterinary medicinal products granted by a competent authority of a member 
state (Offi cial Journal L 159, 27/6/2003, pp. 1 – 23), available: http://ec.europa.eu/ 
enterprise/pharmaceuticals/eudralex/homev1.htm , accessed Apr. 10, 2006 . 
12. The ICH process for harmonisation of guidelines, available: http://www.ich.org/cache/ 
compo/276 - 254 - 1.html , accessed May 15, 2006 . 
13. Food and Drug Administration, Centre for Drug Evaluation and Research (CDER) , 
Guideline on general principles of process validation, available: http://www.fda.gov/cder/ 
guidance/pv.htm , accessed Apr. 26, 2006 . 
14. Willig , S. H. ( 2001 ), Production and process controls , in Swarbrick , J. , Ed., Good Manufacturing 
Practices for Pharmaceuticals: A Plan for Total Quality Control from Manufacturer 
to Consumer , Marcel Dekker , New York , pp. 99 – 138 . 
15. Waterland , N. H. , and Kowtna , C. C. ( 2003 ), Change control and SUPAC , in Nash , R. A. , 
and Wachter , A. H. , Eds., Pharmaceutical Process Validation , Marcel Dekker , New York , 
pp. 699 – 748 . 
16. Ahmed , S. U. , Naini , V. , and Wadgaonkar , D. ( 2005 ), Scale - up, process validation and 
technology transfer , in Shargel , L. , and Kanfer , I. , Eds., Generic Drug Product Development: 
Solid Oral Dosage Form , Marcel Dekker , New York , pp. 95 – 136 . 
REFERENCES 95


97 
1.4 
GMP - COMPLIANT PROPAGATION 
OF HUMAN MULTIPOTENT 
MESENCHYMAL STROMAL CELLS 
Eva Rohde , Katharina Schallmoser , Christina Bartmann , 
Andreas Reinisch , and Dirk Strunk 
Medical University of Graz, Graz, Austria 
Contents 
1.4.1 Introduction 
1.4.2 Acronyms and Defi nitions 
1.4.2.1 Mesenchymal Stromal Cells 
1.4.2.2 Somatic Stem Cell Therapy 
1.4.2.3 Good Manufacturing Practice 
1.4.2.4 Cell - Based Medicinal Products 
1.4.2.5 Human Platelet Lysate 
1.4.3 Approaches 
1.4.3.1 Adherence to Principles of GMP in a Preclinical Developmental Process 
1.4.3.2 Effi cient Standardized MSC Propagation Using Low Cell Seeding Density 
1.4.3.3 Superior MSC Proliferation Resulting from HPL - Driven as Compared to 
FBS - Driven Cultures 
1.4.3.4 Contamination Risks Can Be Minimized in Rational MSC Propagation 
Procedures 
1.4.4 Testing Methods 
1.4.4.1 Safety and Effi cacy of CBMP in Preclinical Stage 
1.4.4.2 Quality Controls During Cell Culture (In - Process Controls) and Final Product 
Release Criteria 
1.4.4.3 MSC Functionality and Potency Assays 
1.4.5 Conclusion 
References 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

98 GMP-COMPLIANT PROPAGATION 
1.4.1 INTRODUCTION 
Somatic stem cell therapy (SCT) is a rapidly growing fi eld that opens a broad spectrum 
of therapeutic options. The concept of regenerative SCT is based on the 
assumption that transplantation of adult human stem cells may support organ regeneration, 
modulate immunity, and regulate hematopoiesis. Transplantation of bone 
marrow (BM) – derived hematopoietic stem cells (SCs) for blood and immune system 
regeneration has been a clinical reality for almost 40 years. The existence of detectable 
numbers of mesenchymal and endothelial progenitors within blood and BM 
has promoted the readily harvestable hematopoietic tissue as a source of SCs for 
nonhematopoietic regenerative SCT (Figure 1 ). 
Multipotent mesenchymal stromal cells (MSCs) are currently undergoing evaluation 
in a number of clinical trials ( www.clinicaltrials.gov ). These nonhematopoietic 
cells have been fi rst described by Friedenstein et al. in a fi broblast colony - forming 
unit assay (CFU - F) based on low - density culture of adherent BM - derived cells 
[1 – 3] . Alternative sources for MSCs have been identifi ed in a number of studies 
showing the successful isolation of fi broblast precursors from umbilical cord blood, 
placenta, umbilical cord, amniotic fl uid, and adipose tissue [4 – 13] . To date, most 
experimental and clinical experience has been accumulated with BM - MSC [14 – 21] . 
Ex vivo expansion of these rare BM constituents (representing less than 1% of 
aspirated BM nucleated cells) is a prerequisite to achieve a reasonable MSC application 
dose of at least 2 . 10 6 MSCs/kg of the recipients ’ body weight. The majority 
of expansion procedures are currently based on the use of fetal bovine serum (FBS), 
which carries the risk of xenoimmunization and transmission of known (e.g., prions 
transmitting bovine spongiforme encephalopathia, BSE) and unknown pathogens. 
These risks could be avoided by developing MSC expansion protocols that use 
human alternatives which replace FBS. 
The preclinical development of medicinal products in general bears high complexity 
due to the lack of fi xed routines. Long term manipulations of cell - based 
medicinal products (CBMPs) may enhance the risks for undesirable effects in the 
course of ex vivo cell expansions. Safety concerns regarding the clinical application 
of ex vivo generated MSCs require a logistic environment providing an established 
good manufacturing practice (GMP) background embedded in a highly effective 
quality system. Demonstration of manufacturing and product consistency is achieved 
by applying rational in - process controls. Release criteria should ideally emerge from 
successful product development and optionally include the sterility, safety, purity, 
identity, and potency. 1 They must to be rapid, sensitive, and reliable and should 
retain some fl exibility in type and timing of testing. The complexity and function of 
different CBMPs require an array of analytical procedures to adequately characterize 
the particular product (potency assays). Personalized (patient - specifi c) CBMPs 
differ from large drug batches in the pharmaceutical industry in terms of practicability 
in fi nal product release in that they may require process - oriented rather than 
single - product potency testing. U.S. regulations demand that “ tests for potency 
shall consist of either in vitro or in vivo tests, or both, which have been specifi cally 
1 U.S. legislation: 21 CFR 610, General biological products standards, CFR 610.10 Potency, CFR 
600.3(s). 

FIGURE 1 Hematopoietic tissue - derived SC and progenitors. Hematopoietic tissue contains 
( a ) mesenchymal and ( b ) endothelial in addition to ( c ) hematopoietic progenitor cells. 
( a ) Adult human BM - derived MSCs were stained to visualize the actin cytoskeleton, mitochondria, 
and nuclei. ( b ) The periphery of an umbilical cord blood – derived endothelial progenitor 
cell (EPC) colony is depicted demonstrating typical cobble stone – like morphology. 
The entire colony was derived from a single UCB - EPC indicating impressive proliferation 
potential (more than 70,000 cells were obtained by harvesting single EPC - derived colonies 
indicating the completion of at least 16 population doublings). Less than 10 mL of adult BM 
[( a ) and ( c )] but at least 40 mL of UCB ( b ) were suffi cient to generate appropriate numbers 
of cells for therapeutic purposes. 
(c) 
(b) 
(a) 
INTRODUCTION 99

100 GMP-COMPLIANT PROPAGATION 
designed for each product so as to indicate its potency in a manner adequate to 
satisfy the interpretation of potency given by the defi nition in 21 CFR 600.3(s). ” 
Functional analyses accompanying the expansion process development leading to 
a full product characterization and optimization of manufacturing steps are prerequisites 
that allow for the creation of a safe and effective CBMP. 
This chapter demonstrates that rapid and standardized expansion of human 
MSCs to achieve a reasonable cell dose (i.e., . 2 . 10 6 /kg body weight of a 75 - kg 
person corresponding to . 1.5 . 10 8 MSCs) is feasible within less than four weeks. 
Replacing FBS with human platelet lysate (HPL) provides one strategy toward a 
safer CBMP (Figure 2 ). Appropriate preclinical development adhering to GMP 
principles will enhance safety in the course of a consecutive clinical evaluation of 
MSCs as a therapeutic agent. 
1.4.2 ACRONYMS AND DEFINITIONS 
1.4.2.1 Mesenchymal Stromal Cells 
Adhesion of mononuclear cells from human bone marrow aspirates (BM - MNC) to 
tissue culture plastic and removal of nonadherent cells during the fi rst days of 
culture selects for a population of proliferating spindle - shaped fi broblast - like non- 
FIGURE 2 GMP - compliant propagation of human MSCs. The summary of a 
two - step MSC production procedure shows seeding and harvest numbers of BM - MNC and 
resulting numbers of MSC HPL as compared to MSC FBS . ( Reproduced with permission from 
ref. 23 .) 
4 x 2.5mL heparinized BM Aspiration diluted 
immediately (without density gradient) in 
.-MEM / 10% FBS .-MEM / 10% HPL 
1 x 107 MNC / 60mL / 225cm 2 in 
10 - 20 x T225 or 1 -2 CF-4 ( < 105 BM-MNC/cm 2) 
8 x 106 MSC / 225cm 2 1 x 106 MSC / 225cm 2 
STORE: n x 3x10 5 
MSCHPL aliquots 
STORE: n x 1x10 6 
MSCFBS aliquots 
.-MEM / 10% HPL 
3 x 105 MSC / 1m 2 
.-MEM / 10% FBS 
3 x 105 MSC / 1m 2 
STEP II 
STEP I 
1° SEEDING 
day 0 
1° HARVEST 
. 10 - 16 days 
BM aspiration 
day 0 
2° HARVEST 
. 11 - 15 days 
2° SEEDING 
day 0 
3.0 - 5.4 x 10 8 MSCHPL 0.5 – 1.1 x 10 8 MSCFBS

hematopoietic multipotent MSCs. Mesenchymal stromal cells can also be obtained 
from umbilical cord blood, umbilical cord, placenta, adipose tissue, and several fetal 
tissues. The minimum criteria for MSCs are defi ned in an ISCT (International 
Society for Cellular Therapy) position paper published in 2006 [22] . The MSCs have 
a high self - renewal potential and the capacity to be differentiated in vitro into 
progeny displaying an osteo - , chondro - , or adipogenic phenotype. 
1.4.2.2 Somatic Stem Cell Therapy 
The concept of regenerative SCT is based on experimental and early clinical observations 
indicating that the application of adult stem cells can improve organ regeneration 
after ischemic, toxic, or metabolic injury. Bone marrow harbors hematopoietic 
and mesenchymal stem cells and endothelial progenitor cells and is an easily accessible 
but not the sole, source of candidate cells to promote organ repair after systemic 
or local application. Regulation of hematopoiesis and immune modulation 
are the two established applications in the broad fi eld of SCT with autologous and 
allogeneic stem and progenitor cells. 
1.4.2.3 Good Manufacturing Practice 
Good manufacturing practice is that part of the quality management system (QMS) 
that is concerned with the production and quality control of medicinal products 
(drugs) for human and veterinary use. It includes documentation, personnel training, 
facility, equipment, and process controls for the manufacture of pharmaceuticals. 
1.4.2.4 Cell - Based Medicinal Products 
Medicinal products containing viable cells are summarized under the umbrella term 
cell - based medicinal products . The term CBMP does not cover products containing 
nonviable cells or cellular fragments. The CBMPs may have much potential in the 
treatment of various diseases that to date have no cure. They are heterogeneous in 
terms of origin and type of cells and with regard to the complexity of the product. 
Cells may be self - renewing stem cells, more committed progenitors, or terminally 
differentiated cells exerting a specifi c regenerative function. Cells may be of autologous 
or allogeneic origin. Cells may be used alone or in combination with biomolecules, 
chemical substances, or structural materials that possibly potentiate their 
desired effects. 
1.4.2.5 Human Platelet Lysate 
Human platelet lysate can be obtained from buffy coat – derived platelet rich plasma. 
The platelet fraction is separated from the plasma and the white and red blood cell 
fraction by centrifugation steps and concentrated to a density of at least 1 . 10 9 platelets/
mL. Platelets can either be activated with thrombin or lysed by repeated freeze – 
thaw cycles. Both mechanisms result in the release of growth factors and mitogens 
that are stored in intact platelets. Mediators released from platelets include, among 
others, epidermal growth factor (EGF), basic fi broblast growth factor (bFGF), platelet 
- derived growth factors (PDGFs), transforming growth factor (TGF - . 1), and 
insulinlike growth factor (IGF) [23, 24] . Perhaps HPL may replace FBS in many 
ACRONYMS AND DEFINITIONS 101

102 GMP-COMPLIANT PROPAGATION 
cell culture systems that have previously been thought to strictly depend on the 
presence of FBS. 
1.4.3 APPROACHES 
1.4.3.1 Adherence to Principles of GMP in a Preclinical Developmental Process 
The standardized MSC propagation should be conducted as a well - planned, consistently 
documented, and optimized procedure that also minimizes risks of microbiological, 
particulate and pyrogen contamination by reducing manipulation steps and 
manipulation time. According to current European legislation, 2 the principles of 
GMP should be applied to CBMP when they are manufactured for use in human 
subjects in phase 1 studies. These requirements do not apply to cellular or tissue - 
based medicinal products used in phase 1 studies according to U.S. legislation 3 or 
to products in the preclinical developmental phase. If it is expected that preclinical 
fi ndings are to be translated into clinical use rather rapidly, it may be recommended 
to establish GMP - compliant technology during the preclinical developmental 
phase of any cell product. As a result, this ensures that products are consistently 
produced and controlled to meet the quality standards appropriate for their intended 
use or product specifi cation. The GMP requirements are well described in “ PIC/S 
Guide to Good Manufacturing Practice for Medicinal Products ” and include 
the implementation of an effi ciently running quality management system, dedi - 
cated areas for manufacture of sterile medicinal products complying with GMP, 
appropriately qualifi ed and trained personnel, suitable equipment, correct materials, 
containers and labels, approved procedures and instructions, suitable storage and 
transport facilities, and a record - keeping system that allows the complete history of 
a medicinal product to be traced (See http://www.picscheme.org ). It is a challenge 
to conduct preclinical research and development complying to GMP as procedures 
routinely turn out to be much more time and cost intensive than common laboratory 
- scale research. These circumstances can advance either the developmental 
progress at the expense of quality standards or vice versa. It should therefore be 
decided on a case - by - case basis how closely to adhere to GMP standards depending 
on the more or less stringent time schedule for the considered clinical use of a 
CBMP. 
1.4.3.2 Effi cient Standardized MSC Propagation Using Low Cell 
Seeding Density 
The future use of MSCs in clinical studies may require very high absolute MSC 
numbers to gain appropriate cell doses ( > 5 . 10 6 /kg body weight) per patient compared 
to in vivo experimental models with small animals [20] . It is consequently 
advantageous to develop large - scale MSC expansion protocols that allow for the 
2 European legislation: Directive 65/65/EEC, Directive 75/318/EEC, Directive 75/318/EEC, Commission 
Communication on the Community marketing authorisation procedures for medicinal products 
(98/C229/03); Directive 2001/20/EC, EMEA/CHMP/410869/2006. 
3 U.S. legislation: 21 CFR 210; 21 CFR 211; 21 CFR 312.21; 21 CFR 312.22(a) and 21 CFR 
312.23(a)(7)(i). 

APPROACHES 103 
generation of up to 5 . 10 8 – 10 . 10 8 MSCs from the limited starting volume of 
primary material. 
The cell seeding density is of critical importance for the expansion rate of MSCs 
and must be defi ned for the primary seeding and the following passaging steps. Most 
experimental and clinical expansions described to date were started with a 
high seeding density of more than 1 . 10 5 BM - MNC/cm 2 [2, 14, 16] . For further passages 
pioneering studies showed that a very low seeding density between 0.5 and 
10 MSCs/cm 2 selects for the expansion of a rapidly proliferating subpopulation of 
recycling SCs, termed RS cells [25 – 28] . This seeding density, referred to as “ clonal 
density, ” would necessitate a theoretical growth area of from 2,000,000 to 100,000 cm 2 
(from 200 to 10 m 2 ) to obtain a clinical quantity of > 1 . 10 8 MSCs from 1 . 10 6 starting 
MSCs within one passage. Plating 30 – 100 MSCs/cm 2 therefore is a reasonable 
compromise density requiring a more realistic growth area between 10,000 and 
25,000 cm 2 (1 and 2.5 m 2 ). We have recently shown that the primary seeding of only 
10 mL bone marrow aspirates on approximately 0.2 m 2 culture area for two weeks 
(culture step 1; BM diluted immediately after aspiration in culture medium without 
density gradient separation; removal of nonadherent cells at day 3) followed by an 
expansion on 2.5 m 2 (step 2) is suffi cient to consistently generate at least 1.5 . 
10 8 MSCs in FBS - supplemented medium within less than four weeks (Figure 2 ) [29] . 
This study furthermore corroborated earlier data on the inverse correlation of the 
seeding density to MSC proliferation (Figure 3 ) [25 – 28] . 
1.4.3.3 Superior MSC Proliferation Resulting from HPL - Driven as Compared 
to FBS - Driven Cultures 
The most commonly used basic cell culture medium compositions for MSC propagation 
are minimum essential medium alpha ( . - MEM) and low - glucose (1 g/L) Dulbecco 
’ s modifi ed Eagle medium (DMEM - LG) supplemented with l - glutamin, 
antibiotics, and 5 – 20% FBS [14, 16, 19, 24, 25, 30] . Our experience with MSC propagation 
relates to the use of . - MEM supplemented with either FBS or HPL. In 
contrast to HPL that has been recognized only recently as a potent culture medium 
supplement [24] , FBS is a well - known key medium supplement for cell culture and 
its role has been unchallenged for more than 50 years [31] . The common use of FBS 
in MSC cultures as a source of growth factors and mitogens bears the risk of transmission 
of known and unknown pathogens as well as xenoimmunization against 
bovine pathogens and should therefore be avoided for clinical use [32, 33] . 
In a recent study we analyzed the capacity of HPL to replace FBS in large - scale 
(clinical) MSC expansions and were able to demonstrate a superior propagation of 
MSC cultured with HPL (MSC HPL ) as compared to MSC derived from FBS - driven 
cultures (MSC FBS ) [23] . Figure 4 illustrates superior MSC proliferation at low plating 
density and higher population doublings (PDs) with HPL after a culture period of 
less than 14 days. 
1.4.3.4 Contamination Risks Can Be Minimized in 
Rational MSC Propagation Procedures 
Cell expansion is mainly performed according to labor - intensive time - consuming 
protocols using open systems that increase the risks of microbiological or particulate 
contamination and supplementation with potent antibiotics to control these prob

104 GMP-COMPLIANT PROPAGATION 
lems. Avoiding the use of penicillin during clinical - scale cell propagation follows the 
rationale to reduce the risk of sensitization as well as anaphylactic precipitation. 
Thus, it may be worthwhile not using other antibiotics for the GMP - compliant MSC 
propagation. One approach to minimize potential contamination risks is to rigorously 
reduce handling in the course of MSC propagation to the absolute minimum 
of necessary steps. In our experience, the commonly used density gradient centrifugation 
step can be skipped prior to the primary cell seeding of the bone marrow 
aspirate. Immediate dilution of limited volumes (e.g., 10 – 20 mL) of heparinized BM 
aspirate into supplemented . - MEM medium for direct cell seeding does not result 
in a loss of MSC recovery [23] . Furthermore, the aforementioned low cell seeding 
density and the employment of an increased growth area in a simplifi ed procedure 
together with the use of HPL in fact allow for an effi cient production of high MSC 
numbers within one to two harvest - replating cycles. The relatively short ex vivo 
expansion time of less than three to four weeks may be helpful in reducing the 
cumulative risk of contamination. 
1.4.4 TESTING METHODS 
1.4.4.1 Safety and Effi cacy of CBMP in Preclinical Stage 
The preclinical developmental period should be used for the extensive characterization 
of the CBMP. Release criteria have to be defi ned and reasonable time frames 
1000/cm2 
day 1 
day 3 
day 5 
day 10 
100/cm2 10/cm2 1/cm2 
P+1 P+1 
FIGURE 3 Inverse correlation of seeding density to MSC proliferation. BM - derived MSCs 
derived from passage 2 were seeded at log fold deescalated density of 1000, 100, 10, and 1 cm . 2 . 
Photographs were taken after 1, 3, 5, and 10 days of culture in . - MEM/10% FBS (original 
magnifi cation 40 . ). In the case of MSC seeded at 100 and 1000 cells/cm 2 confl uence necessitated 
trypsinisation between days 5 and 10 followed by reseeding at 100 and 1000 cells/cm 2 , 
respectively, and is therefore indicated as P + 1. 

must be set to allow for a high safety and quality standard of the fi nal cellular 
product. On the other hand, the logistic background should allow for a rapid release 
of the CBMP within a few hours due to the potential short shelf life of many cellular 
products. Ranges of cell purity, sterility, and absence of pyrogens and endotoxins 
are factors of utmost importance which must be determined. It is an inherent feature 
of CBMPs that product specifi cations must be adapted to the individual application. 
The challenge in the preclinical developmental phase is to fi nd satisfactory answers 
to unresolved questions in terms of cell type, source, dose, and mode of application 
according to the particular target disease. Thus, in - process controls and defi nitive 
release criteria must be met by each CBMP. Since many CBMPs are personalized 
medicine, potency assays must be performed for selected representative products 
(e.g., before initiating a study and consecutively once per year). 
1.4.4.2 Quality Controls During Cell Culture (In - Process Controls) and Final 
Product Release Criteria 
General Safety According the Food and Drug Administration (FDA), cellular 
therapy products are exempt from general safety testing [21 CFR 610.11(g)(1)]. 
Cell Dose The preclinical stage can be used to determine the specifi cations for the 
minimum effective and maximum tolerable number of viable and functional cells. 
The optimum dose of cells to be administered still needs to be established [20] . 
FIGURE 4 MSC proliferation capacity depends on seeding density in xenogeneic FBS and 
HPL - supplemented cultures. The inverse correlation of MSC proliferation to their seeding 
density resulted in the formation of a confl uent MSC layer in cultures starting with 1 MSC/cm 2 
in . - MEM/10% FBS and cultures starting with 1 – 10 MSCs/cm 2 in . - MEM/10% HPL but not 
when initiating cultures with the respective higher seeding densities within less than two 
weeks. The calculated fold increase of the cell number and corresponding population doublings 
from a representative experiment harvested at day 13 are shown. 
0
2
4
6
8 
10 
1/cm2 10/cm2 100/cm2 
FBS HPL 
FBS HPL 
1/cm2 1/cm2 
10/cm2 10/cm2 
100/cm2 100/cm2 
0 
200 
400 
600 
800 
1/cm2 10/cm2 100/cm2 
FBS HPL 
PD (d 13) Fold increase 
MSC cultured for 13 days 
TESTING METHODS 105

106 GMP-COMPLIANT PROPAGATION 
Viability Viability of MSCs can easily be determined immediately after trypsinization 
via trypan blue or 7 - amino - actinomycin D (7 - AAD) exclusion. According 
to the specifi cations developed from our cell culture studies, viability should be 
> 90%. In selected exceptional cases a lower limit of 70% viability of total harvested 
cells may be acceptable. 
Microbiological Testing Sterility testing that detects fungal, anaerobic, and aerobic 
bacterial and mycoplasma contamination should be performed after each critical 
manipulation step during MSC culture that is prone to microbiological contamination 
[34] . The crucial bacterial sterility check at the end of last harvesting step cannot 
be evaluated prior to in vivo application if MSCs need to be applied immediately 
after propagation due to the duration of the cultures. Mycoplasma polymerase chain 
reaction (PCR) results can be obtained at the day of harvest within less than 6 h. 
MycoAlert ® results are available within less than 1 h at the day of harvest. Defi nitive 
culture results to exclude mycoplasma contamination are available within two to 
three weeks and therefore are not applicable for CBMPs with a short shelf life that 
are planned to be administered immediately after production. 
Endotoxin and Pyrogenicity Testing Endotoxin measurement using the Limulus 
amebocyte lysate (LAL) assay is typically done as an alternative to pyrogenicity 
testing for early phase trials. For any parenteral drugs, except those administered 
intrathecally, the FDA recommends that the upper limit for endotoxin be 5 EU/kg 
body weight/dose. The LAL assay method can be applied to the safety evaluation 
of biological preparations according to existing regulations. 4 We use the LAL assay 
to substitute for the lengthy delay in microbiological data availability to obtain 
results prior to the clinical application of the fi nal product within less than 2 h after 
harvest. 
Phenotypic Identity of MSC In addition to morphological identifi cation by microscopy, 
the immunophenotypic characterization of MSCs can be done using a broad 
panel of fl uorescence - conjugated antibodies directed against surface molecules. To 
date there is no specifi c marker uniquely defi ning MSCs. Therefore a profi le is used 
to show the expression of certain markers and to exclude the contamination by cells 
expressing other marker profi les. Flow cytometry is recommended by the ISCT to 
reveal that MSCs stain positive for CD73, CD90, and CD105 and negative for HLA - 
DR, CD14, CD31, CD34, and CD45 (Figure 5 ) [22, 23] . Much more extensive phenotypic 
analyses have been performed without retrieving additional information 
about MSC type or function [35] . Gene expression profi ling will hopefully result in 
a better defi nition of human MSCs [35 – 42] . 
1.4.4.3 MSC Functionality and Potency Assays 
Clonogenicity The self - renewal capacity of cells and the proportion of proliferating 
cells within a heterogeneous cell mixture can be evaluated using the CFU assay. 
4 Endotoxin testing, LAL, according to Eur. Pharm. 2.6.14 and Guideline on Validation of the Limulus 
Amebocyte Lysate Test as an End - Product Endotoxin Test for Human and Animal Parenteral Drug, 
Biological Products and Medical Devices, 1987, Sections I – IV, http://www.fda.gov/cber/gdlns/lal.pdf . 

A tissue culture method allowing for the clone counting of cells was fi rst describend 
in 1956 [43] . The introduction of bone marrow CFU assays led to the discovery of 
hematopoietic stem cells [44] . Fibroblast precursors existing within the hematopoietic 
system also have been evaluated with another specifi c CFU assay method 
introduced by Friedenstein in 1974 (CFU - F) [2] . We analyzed the clonal expansion 
capacity of MSC with the CFU - F method. Figure 6 shows differences in CFU - F 
appearance between MSC HPL and MSC FBS . In the case of primary BM appropriate 
dilution is necessary to determine the CFU - F frequency (Figure 7 ). Once MSCs are 
enriched, the appropriate MSC seeding density recommended for CFU - F enumeration 
may range from 1 to 5 MSCs/cm 2 [2, 29] . 
Osteo - , Chondro - , and Adipogenic Differentiation Isolated BM - derived MSCs 
were shown to differentiate along multiple mesechymal lineages in 1999 [45] . Evidence 
suggests MSCs can also express phenotypic characteristics of endothelial, 
neural, smooth muscle, skeletal myoblast, and cardiac myocyte cells [46] . The prototype 
pathways of MSC differentiation occur along osteogenic, chondrogenic, and 
adipogenic lineages and have been extensively demonstrated in a large number of 
publications [47] . This kind of potency assay may be performed regularly if bone or 
connective tissue repair is intended, although time limits do not enable immediate 
product release. 
FIGURE 5 Immune phenotype of human MSCs. Flow cytometric analysis of at least 10,000 
viable MSCs was used to determine antibody reactivity (gray - fi lled histograms) compared to 
appropriately diluted isotype controls (black line). Phenotypic criteria require positivity 
( . 90%) for CD73, CD90, and CD105 and negativity ( . 2%) for HLA - DR, CD14 (or CD11b), 
CD19 (or CD79 . ), CD34, and CD45. Absence of CD3+ T cells may be desirable in the case 
of GvHD treatment. Depending on the culture conditions, MSCs share reactivity with the 
anti-disialoganglioside antibody GD2 with neuroblastoma cells, melanoma, and small - cell 
lung cancer cells. 
HLA-AB CD 13 CD 29 CD 73 CD 90 CD105 CD 146 
CD 45 CD 3 HLA-DR CD 31 CD 14 CD 34 
FBS 
FBS 
HPL 
HPL 
GD-2 
CD 19 CD 133 
TESTING METHODS 107

108 GMP-COMPLIANT PROPAGATION 
Immune Modulatory Effects Mesenchymal stromal cells inhibit T - cell alloreactivity 
in mixed lymphocyte cultures (MLCs) or lymphocyte proliferation induced by 
mitogens, such as phytohemaglutinin (PHA) or concanavalin A [29, 48 – 51] . It is of 
note that high concentrations of MSCs (representing 10 – 40 MSCs per 100 responder 
lymphocytes) have an inhibitory effect while low MSC concentrations (0.1 – 1%) 
may stimulate lymphocyte proliferation in mixed lymphocyte cultures [50] . If MSCs 
are used for immunosuppressive therapies, these fi ndings may imply that high doses 
of MSCs are needed to inhibit T - cell proliferation in patients with graft - versus - host 
disease following allogeneic bone marrow transplantation. The application of low 
MSC numbers could stimulate lymphocyte proliferation in vivo and hence result in 
an undesirable boost to graft - versus - host disease as an adverse reaction of the MSC 
therapy. It is not clear so far whether the precise number of T cells in a 
given MSC transplant needs to be determined to exclude a potential boost to 
alloreactivity. Immune modulation can be measured with carboxyfl uorescein diacetate 
N - succinimidyl ester (CFSE) labeling of cells to quantify proliferation in 
response to allogeneic or mitogenic stimuli [52] . We analyzed the loss of CFSE fl uorescent 
intensity indicating cell proliferation by fl ow cytometry after culturing 
CFSE - labeled MNCs in the absence or presence of different numbers of MSCs [53] . 
The immune regulatory capacity of MSC HPL and MSC FBS was studied by measurement 
of allogenic MNC proliferation after co - culturing pairs of MNC from three 
different donors with two independent MSC HPL and two other MSC FBS (Figure 8 ). 
Hematopoiesis Regulation Regulation of the behavior of early hematopoietic 
progenitor cells (HPCs) can be analyzed by MSC - HPC cocultures in vitro [54] . 
FIGURE 6 Morphological evaluation of MSCs. CFU - F of MSC HPL compared to MSC FBS 
differ in size, morphology, and density (scale bar identifi es magnifi cation in the upper panel; 
colony photographs taken on day 12, 40 . original magnifi cation). 
HPL 
500.m 
FBS 
500.m

Liquid cultures of purifi ed CD34 + (HPC) with a preestablished MSC feeder layer 
result in the expansion of CD34 + /38 + HPC and CD34 + /38 . hematopoietic SCs and 
support the growth of mature hematopoietic total nucleated cell (TNC) progeny 
(Figure 9 ). 
Genetic Stability and Potential Tumorigenicity Genetic analysis of human MSCs 
is not well established. The signifi cance of standard metaphase chromosome G 
banding is limited due to the low number of metaphases recovered during standard 
analyses. Advances in multicolor fl uorescence in situ hybridization (FISH) and high - 
resolution array - based techniques may also soon be translated into practicable 
diagnostic tools in relation to CBMP safety in regenerative medicine [55] . 
Genetic instability can occur as a rare event after extended culture of mouse and 
human MSCs in FBS - supplemented medium [56, 57] . To test for potential in vivo 
tumor formation, MSCs derived from short - term clinical - scale expansions in FBS - 
or HPL - supplemented media were injected into immunocompromised athymic 
nude mice subcutaneously. Putative tumor formation was evaluated by histological 
analyses three months after injection of 2 . 10 6 and 2 . 10 4 MSCs and compared to 
controls that were injected 48 h prior to euthanasia. A primary cell deposit was 
visible immediately and 48 h after injection and MSCs could be recovered by conventional 
microscopic evaluation. However, none of 12 animals tested developed a 
macroscopic or microscopic detectable tumor over the 90 - day observation period 
[23] . In this situation, genetic testing may be encouraged for prospective data acqui- 
FIGURE 7 CFU - F Evaluation of MSCs depends on BM seeding density. An appropriate 
dilution of the heparinised BM aspiration is needed for accurate enumeration of the primary 
CFU - F frequency as indicated in this representative experiment where whole heparinized 
BM was seeded corresponding to the respective measured BM - MNC number per square 
centimeters of growth area, cultured for 11 days at 37 ° C/humidifi ed atmosphere/3% O 2 /5% 
CO2 . Nonadherent cells were removed at day 3. CFU - F are visualized by Harris hematoxylin 
staining. 
2.23 x 104 4.46 x 103 0.89 x 103 0.18 x 103 
Seeded MNC / cm2 
Day 11 
TESTING METHODS 109

110 GMP-COMPLIANT PROPAGATION 
FIGURE 8 MSC-mediated Immune Modulation. Allogeneic MNC proliferation (mean 
cell number ± SEM) was measured after co - culturing pairs of MNC from three different 
donors with two independent MSC HPL and two other MSC FBS as shown in the bar chart. MSC 
were added in a 1 : 10 (3 . 10 4 MSC to 3 . 10 5 MNC/well; MNC[+PHA]:MSC = 10 : 1) or 1 : 100 
(MNC[+PHA]:MSC = 10 : 1) ratio to test their infl uence on PHA - driven proliferation of 
MNC. MNC numbers were measured by fl ow cytometric MNC count using BD Truecount TM 
tubes. As a control numbers of mitogen stimulated MNC without additional MSC 
(MNC[+PHA] only) and background proliferation without PHA stimulation (MNC w/o 
PHA) are shown. MSC did not induce MNC proliferation (MNC + MSC w/o PHA). Signifi - 
cant differences are marked by asterisks ( *p < 0.05 and * * p < 0.01). (Figure reproduced with 
permission from reference 53 ) 
MNC[+PHA]:MSC = 100:1 
MNC[+PHA] only 
MNC[+PHA]:MSC = 10:1 
MNC + MSC w/o PHA 
MNC w/o PHA 
0 
2x105 
4x105 
6x105 
8x105 
1x106 
1.2x106 
1.4x106 
CELL NUMBER 
HPL FBS

200 
CD34+ No 
0 
4x105 
8x105 
1,2x106 
1,6x106 
2x106 
50 
100 
150 
250 
0 
(b) 
(c) 
CD34+ FOLD INCREASE 
0 
3x106 
6x106 
8x106 
1,2x107 
1,5x107 
300 
1,8x107 
2,1x107 
TNC N o 
FOLD INCREASE 
20 
5
10 
15 
25 
0 
CD34 only CD34 
+ 
MSCFBS 
(a) 
FOLD INCREASE 
CELL NUMBER 
CD34 
+ 
MSCHPL 
CD34 only CD34 
+ 
MSCFBS 
CD34 
+ 
MSCHPL 
CD34+/CD38- No 
0 
1x105 
2x105 
3x105 
4x105 
12 
6
8 
10 
14 
0
4
2 
CD34 only CD34 
+ 
MSCFBS 
CD34 
+ 
MSCHPL 
CD34+/CD38- FI vs. CTRL 
FIGURE 9 MSC - mediated hematopoiesis regulation. ( a ) Umbilical cord blood (UCB) - 
derived sorted CD34 + cells were expanded in cytokine - supplemented medium [Roswell Park 
Memorial Institue (RMPI) - 1640/10% Fetal Bovine Serum (FBS)/Granulocyte and Macrophage 
Colony Stimulating Factor (GM - CSF)/Interleukin 3(IL - 3)/Stem Cell Factor (SCF)/ 
FMS-like tyrosin kinase 3 ligand (Flt - 3L)] in the absence or presence of clinical - scale 
expanded MSCs. Gray bars show harvested total nucleated cell number (TNC N ° .) and black 
bars show the fold increase (FI) of the TNC N ° . compared to the starting CD34 + cell number. 
( b ) Harvested number of CD34 + cells (gray bars) and fold increase (black bars) of CD34 + 
cells after liquid culture with or without MSC support. ( c ) Harvested number (gray bars) and 
fold increase (black bars) of CD34 + /CD38 . hematopoietic stem cells after liquid culture of 
CD34 + cells with MSC FBS or MSC HPL support compared to cytokine - supplemented liquid 
cultures in the absence of MSCs. (Mean ± Standard Error of the Mean (SEM) of two independent 
expansions.) ( Reproduced with permission from ref. 53 .) 
TESTING METHODS 111

112 GMP-COMPLIANT PROPAGATION 
sition but is not considered as mandatory for product release in current MSC clinical 
trials. 
1.4.5 CONCLUSION 
There are considerable limitations of common pharmacological techniques used in 
determining the safety and effi cacy of CBMPs at the preclinical stage. Conventional 
methods used in the pharmaceutical industry to develop pharmacological profi les 
and to determine the acute toxicity of drugs in animals as well as toxicity studies 
may not directly be translated to ex vivo generated cellular products. Nevertheless, 
it is inevitable that preclinical research and development of cellular products will 
be conducted under the guidance of either individual or consensus specifi cations 
and defi nitions that will continuously be improved. This approach will be helpful in 
developing successful therapeutic cellular agents. 
ACKNOWLEDGMENTS 
This work was supported in part by The Adult Stem Cell Research Foundation 
(TASC RF ; C.B. and A.R.) and a young investigator fellowship of the Austrian Federal 
Ministry for Education, Science and Culture, bm:bwk (A.R.). The Austrian Nano - 
Initiative co - fi nanced this work as part of the Nano - Health project (no. 0200), the 
sub-project NANO - STEM being fi nanced by the Austrian Science Fund (FWF 
Project no. N211 - NAN). 
REFERENCES 
1. Luria , E. A. , Panasyuk , A. F. , and Friendenstein , A. Y. ( 1971 ), Fibroblast colony formation 
from monolayer cultures of blood cells , Transfusion , 11 ( 6 ), 345 – 349 . 
2. Friedenstein , A. J. , Deriglasova , U. F. , Kulagina , N. N. , et al. ( 1974 ), Precursors for fi broblasts 
in different populations of hematopoietic cells as detected by the in vitro colony 
assay method , Exp. Hematol. , 2 ( 2 ), 83 – 92 . 
3. Kuznetsov , S. A. , Friedenstein , A. J. , and Robey , P. G. ( 1997 ), Factors required for bone 
marrow stromal fi broblast colony formation in vitro , Br. J. Haematol. , 97 ( 3 ), 561 – 570 . 
4. Erices , A. , Conget , P. , and Minguell , J. J. ( 2000 ), Mesenchymal progenitor cells in human 
umbilical cord blood , Br. J. Haematol. , 109 ( 1 ), 235 – 242 . 
5. Kogler , G. , Sensken , S. , Airey , J. A. , et al. ( 2004 ), A new human somatic stem cell from 
placental cord blood with intrinsic pluripotent differentiation potential , J. Exp. Med. , 
200 ( 2 ), 123 – 135 . 
6. Bieback , K. , Kern , S. , Kluter , H. , and Eichler , H. ( 2004 ), Critical parameters for the isolation 
of mesenchymal stem cells from umbilical cord blood , Stem Cells , 22 ( 4 ), 625 – 634 . 
7. Romanov , Y. A. , Svintsitskaya , V. A. , and Smirnov , V. N. ( 2003 ), Searching for alternative 
sources of postnatal human mesenchymal stem cells: Candidate MSC - like cells from 
umbilical cord , Stem Cells , 21 ( 1 ), 105 – 110 . 
8. Miao , Z. , Jin , J. , Chen , L. , et al. ( 2006 ), Isolation of mesenchymal stem cells from human 
placenta: Comparison with human bone marrow mesenchymal stem cells , Cell. Biol. Int. , 
30 ( 9 ), 681 – 687 . 
9. In ‘ t Anker , P. S. , Scherjon , S. A. , Kleijburg - van der Keur , C. , et al. ( 2004 ), Isolation of 
mesenchymal stem cells of fetal or maternal origin from human placenta , Stem Cells , 
22 ( 7 ), 1338 – 1345 . 

10. In ‘ t Anker , P. S. , Scherjon , S. A. , Kleijburg - van der Keur , C. , et al. ( 2003 ), Amniotic fl uid 
as a novel source of mesenchymal stem cells for therapeutic transplantation , Blood , 
102 ( 4 ), 1548 – 1549 . 
11. Gronthos , S. , Franklin , D. M. , Leddy , H. A. , Robey , P. G. , Storms , R. W. , and Gimble , J. M. 
( 2001 ), Surface protein characterization of human adipose tissue - derived stromal cells , 
J. Cell Physiol. , 189 ( 1 ), 54 – 63 . 
12. Zuk , P. A. , Zhu , M. , Ashjian , P. , et al. ( 2002 ), Human adipose tissue is a source of multipotent 
stem cells , Mol. Biol. Cell , 13 ( 12 ), 4279 – 4295 . 
13. Kern , S. , Eichler , H. , Stoeve , J. , Kluter , H. , and Bieback , K. ( 2006 ), Comparative analysis 
of mesenchymal stem cells from bone marrow, umbilical cord blood, or adipose tissue , 
Stem Cells , 24 ( 5 ), 1294 – 1301 . 
14. Lazarus , H. M. , Haynesworth , S. E. , Gerson , S. L. , Rosenthal , N. S. , and Caplan , A. I. ( 1995 ), 
Ex vivo expansion and subsequent infusion of human bone marrow - derived stromal 
progenitor cells (mesenchymal progenitor cells): Implications for therapeutic use , Bone 
Marrow Transplant. , 16 ( 4 ), 557 – 564 . 
15. Koc , O. N. , Peters , C. , Aubourg , P. , et al. ( 1995 ), Bone marrow - derived mesenchymal stem 
cells remain host - derived despite successful hematopoietic engraftment after allogeneic 
transplantation in patients with lysosomal and peroxisomal storage diseases . Exp. 
Hematol. , 27 ( 11 ), 1675 – 1681 . 
16. Koc , O. N. , Gerson , S. L. , Cooper , B. W. , et al. ( 2000 ), Rapid hematopoietic recovery after 
coinfusion of autologous - blood stem cells and culture - expanded marrow mesenchymal 
stem cells in advanced breast cancer patients receiving high - dose chemotherapy , J. Clin. 
Oncol. , 18 ( 2 ), 307 – 316 . 
17. Lee , S. T. , Jang , J. H. , Cheong , J. W. , et al. ( 2002 ), Treatment of high - risk acute myelogenous 
leukaemia by myeloablative chemoradiotherapy followed by co - infusion of T cell - 
depleted haematopoietic stem cells and culture - expanded marrow mesenchymal stem 
cells from a related donor with one fully mismatched human leucocyte antigen haplotype , 
Br. J. Haematol. , 118 ( 4 ), 1128 – 1131 . 
18. Koc , O. N. Day , J. , Nieder , M. , Gerson , S. L. , Lazarus , H. M. , and Krivit , W. ( 2002 ), 
Allogeneic mesenchymal stem cell infusion for treatment of metachromatic leukodystrophy 
(MLD) and Hurler syndrome (MPS - IH) , Bone Marrow Transplant. , 30 ( 4 ), 
215 – 222 . 
19. Le Blanc , K. , Rasmusson, I. , Sundberg , B. , et al. (2004), Treatment of severe acute graft- 
versus - host disease with third party haploidentical mesenchymal stem cells , Lancet , 1, 
363 ( 9419 ), 1439 – 1441 . 
20. Lazarus , H. M. , Koc , O. N. , Devine , S. M. , et al. ( 2005 ), Cotransplantation of HLA - identical 
sibling culture - expanded mesenchymal stem cells and hematopoietic stem cells in hematologic 
malignancy patients , Biol. Blood Marrow Transplant. , 11 ( 5 ), 389 – 398 . 
21. Ringden , O. , Uzunel , M. , Rasmusson , I. , et al. ( 2006 ), Mesenchymal stem cells for 
treatment of therapy - resistant graft - versus - host disease , Transplantation , 81 ( 10 ), 1390 – 
1397 . 
22. Dominici , M. , Le Blanc , K. , Mueler , I. , et al. ( 2006 ), Minimal criteria for defi ning multipotent 
mesenchymal stromal cells. The International Society for Cellular Therapy position 
statement , Cytotherapy , 8 ( 4 ), 315 – 317 . 
23. Schallmoser , K. , Bartmann , C. , Rohde , E. , et al. ( 2007 ), Human platelet lysate can replace 
fetal bovine serum for clinical scale expansion of functional MSC , Transfusion . 2007 Aug; 
47 ( 8 ), 1436 – 1446 . 
24. Doucet , C. , Ernou , I. , Zhang , Y. Z. , et al. ( 2005 ), Platelet lysates promote mesenchymal 
stem cell expansion: A safety substitute for animal serum in cell - based therapy applications 
. J. Cell. Physiol. , 205 ( 2 ), 228 – 236 . 
REFERENCES 113

114 GMP-COMPLIANT PROPAGATION 
25. Digirolamo , C. M. , Stokes , D. , Colter , D. , Phinney , D. G. , Class , R. , and Prockop , D. J. 
( 1999 ), Propagation and senescence of human marro stromal cells in culture: A simple 
colony - forming assay identifi es samples with the greatest potential to propagate and differentiate 
, Br. J. Haematol. , 107 ( 2 ), 275 – 281 . 
26. Colter , D. C. , Class , R. , DiGirolamo , C. M. , Prockop , D. J. ( 2000 ), Rapid expansion of 
recycling stem cells in cultures of plastic - adherent cells from human bone marrow , Proc. 
Nat. Acad. Sci. USA , 97 ( 7 ), 3213 – 3218 . 
27. Colter , D. C. , Sekiya , I. , and Prockop , D. J. ( 2001 ), Identifi cation of a subpopulation of 
rapidly self - renewing and multipotential adult stem cells in colonies of human marrow 
stromal cells , Proc. Nat. Acad. Sci. USA , 98 ( 14 ), 7841 – 7845 . 
28. Sekiya , I. , Larson , B. L. , Smith , J. R. , Pochampally , R. , Cui , J. G. , and Prockop , D. J. ( 2002 ), 
Expansion of human adult stem cells from bone marrow stroma: Conditions that maximie 
the yields of early progenitors and evaluate their quality , Stem Cells , 20 ( 6 ), 530 – 541 . 
29. Bartmann , C. , Rohde , E. , Schallmoser , K. , et al. ( 2007 ), Two steps to functional MSC for 
clinical application , Transfusion . 2007 Aug; 47 ( 8 ), 1426 – 1435 . 
30. Horwitz , E. M. , Gordon , P. L. , Koo , W. K. , et al. ( 2002 ), Isolated allogeneic bone marrow - 
derived mesenchymal cells engraft and stimulate growth in children with osteogenesis 
imperfecta: Implications for cells therapy of bone , Proc. Natl. Acad. Sci. USA , 99 ( 13 ), 
8932 – 8937 . 
31. Pollard , J. W. ( 1989 ), Basic cell culture , in Pollard , J. W. , Ed., Animal Cell Culture , Human , 
Clifton, NJ , pp. 1 – 2 . 
32. WHO ( 1997 ), Medicinal and other products and human and animal transmissible spongiform 
encephalopathies: Memorandum from a WHO meeting , Bull. World Health Org. , 
75 ( 6 ), 505 – 513 . 
33. European Union ( 2004 ), Note for guidance on minimising the risk of transmitting animal 
spongiform encephalopathy agents via human and veterinary medicinal products 
(EMEA/410/01 Rev. 2 — October 2003) adopted by the Committee for Proprietary Medicinal 
Products (CPMP) and by the Committee for Veterinary Medicinal Products (CVMP) , 
Off. J. Europ. Union , C24/26 – C24/19 . 
34. Schallmoser , K. , Rosin , C. , Vormittag , R. , et al. ( 2006 ), Specifi cities of platelet autoantibodies 
and platelet activation in lupus anticoagulant patients: A relation to their history 
of thromboembolic disease , Lupus , 15 ( 8 ), 507 – 514 . 
35. Wagner , W. , Wein , F. , Seckinger , A. , et al. ( 2005 ), Comparative characteristics of mesenchymal 
stem cells from human bone marrow, adipose tissue, and umbilical cord blood , 
Exp. Hematol. , 33 ( 11 ), 1402 – 1416 . 
36. Ramalho - Santos , M. , Yoon , S. , Matsuzaki , Y. , Mulligan , R. C. , and Melton , D. A. ( 2002 ), 
“ Stemness ” : Transcriptional profi ling of embryonic and adult stem cells , Science , 298 ( 5593 ), 
597 – 600 . 
37. Boquest , A. C. , Shahdadfar , A. , Fronsdal , K. , et al. ( 2005 ), Isolation and transcription 
profi ling of purifi ed uncultured human stromal stem cells: Alteration of gene expression 
after in vitro cells culture , Mol. Biol. Cell , 16 ( 3 ), 1131 – 1141 . 
38. Shahdadfar , A. , Fronsdal , K. , Haug , T. , Reinholt , F. P. , and Brinchmann , J. E. ( 2005 ), In 
vitro expansion of human mesenchymal stem cells: Choice of serum is a determinant of 
cell proliferation, differentiation, gene expression, and transcriptome stability , Stem Cells , 
23 ( 9 ), 1357 – 1366 . 
39. Silva , W. A. , Jr. , Covas , D. T. , Panepucci , R. A. , et al. ( 2003 ), The profi le of gene expression 
of human marrow mesenchymal stem cells , Stem Cells , 21 ( 6 ), 661 – 669 . 
40. Panepucci , R. A. , Siufi , J. L. , Silva , W. A. , Jr. , et al. ( 2004 ), Comparison of gene expression 
of umbilical cord vein and bone marrow - derived mesenchymal stem cells , Stem Cells , 
22 ( 7 ), 1263 – 1278 . 

41. Jeong , J. A. , Hong , S. H. , Gang , E. J. , et al. ( 2005 ), Differential gene expression profi ling 
of human umbilical cord blood - derived mesenchymal stem cells by DNA microarray , 
Stem Cells , 23 ( 4 ), 584 – 593 . 
42. Monticone , M. , Liu , Y. , Tonachini , L. , et al. ( 2004 ), Gene expression profi le of human bone 
marrow stromal cells determined by restriction fragment differential display analysis , 
J. Cell. Biochem , 92 ( 4 ), 733 – 744 . 
43. Puck , T. T. , and Marcus , P. I. ( 1956 ), Action of x - rays on mammalian cells . J. Exp. Med. , 
103 ( 5 ), 653 – 666 . 
44. Tepperman , A. D. , Curtis , J. E. , and McCulloch , E. A. ( 1974 ), Erythropietic colonies in 
cultures of human marrow , Blood , 44 ( 5 ), 659 – 669 . 
45. Pittenger , M. F. , Mackay , A. M. , Beck , S. C. , et al. ( 1999 ), Multilneage potential of adult 
human mesenchymal stem cells , Science , 284 ( 5411 ), 143 – 147 . 
46. Pittenger , M. F. , and Martin , B. J. ( 2004 ), Mesenchymal stem cells and their potential as 
cardiac therapeutics , Circ. Res. , 95 ( 1 ), 9 – 20 . 
47. Delorme , B. C. S. , and Charbord , P. ( 2006 ), The concept of mesenchymal stem cells , Regenerative 
Med. , 1 ( 4 ), 497 – 509 . 
48. Di Nicola , M. , Carlo - Stella , C. , Magni , M. , et al. ( 2002 ), Human bone marrow stromal 
cells suppress T - lymphocyte proliferation induced by cellular or nonspecifi c mitogenic 
stimuli , Blood , 99 ( 10 ), 3838 – 3843 . 
49. Bartholomew , A. , Sturgeon , C. , Siatskas , M. , et al. ( 2002 ), Mesenchymal stem cells suppress 
lymphocyte proliferation in vitro and prolong skin graft survival in vivo , Exp. 
Hematol. , 30 ( 1 ), 42 – 48 . 
50. Le Blanc , K. , Tammik , L. , Sundberg , B. , Haynesworth , S. E. , and Ringden , O. ( 2003 ), 
Mesenchymal stem cells inhibit and stimulate mixed lymphocyte cultures and mitogenic 
responses independently of the major histocompatibility complex , Scand. J. Immunol. , 
57 ( 1 ), 11 – 20 . 
51. Tse , W. T. , Pendleton , J. D. , Beyer , W. M. , Egalka , M. C. , and Guinan , E. C. ( 2003 ), Suppression 
of allogeneic T - cell proliferation by human marrow stromal cells: Implications 
in transplantation , Transplantation , 75 ( 3 ), 389 – 397 . 
52. Muller , I. , Kordowich , S. , Holzwarth , C. , et al. ( 2006 ), Animal serum - free culture conditions 
for isolation and expansion of multipotent mesenchymal stromal cells from human 
BM , Cytotherapy , 8 ( 5 ), 437 – 444 . 
53. Reinisch , A. , Bartmann , C. , Rohde , E. S. K. , Bjelic - Radisic , V. , Lanzer , G. , Linkesch , W. , 
and Strunk , D. ( 2007 ), A humanized system to propagate cord blood - derived mesenchymal 
stem cells for clinical application , Regen. Med. 2007 Jul; 2 ( 4 ), 371 – 382 . 
54. Robinson , S. N. , Ng , J. , Niu , T. , et al. ( 2006 ), Superior ex vivo cord blood expansion following 
co - culture with bone marrow - derived mesenchymal stem cells , Bone Marrow 
Transplant. , 37 ( 4 ), 359 – 366 . 
55. Speicher , M. R. , and Carter , N. P. ( 2005 ), The new cytogenetics: Blurring the boundaries 
with molecular biology , Nat. Rev. Genet. , 6 ( 10 ), 782 – 792 . 
56. Rubio , D. , Garcia - Castro , J. , Martin , M. C. , et al. ( 2005 ), Spontaneous human adult stem 
cell transformation , Cancer Res. , 65 ( 8 ), 3035 – 3039 . 
57. Tolar , J. , Nauta , A. J. , Osborn , M. J. , et al. ( 2007 ), Sarcoma derived from cultured mesenchymal 
stem cells , Stem Cells , 25 ( 2 ), 371 – 379 . 
REFERENCES 115


INTERNATIONAL REGULATIONS OF 
GOOD MANUFACTURING PRACTICES 
SECTION 2


119 
2.1 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc. 
NATIONAL GMP REGULATIONS AND 
CODES AND INTERNATIONAL GMP 
GUIDES AND GUIDELINES: 
CORRESPONDENCES AND 
DIFFERENCES 
Marko N a rhi and Katrina Nordstr o m 
Helsinki University of Technology, Helsinki, Finland 
Contents 
2.1.1 Introduction 
2.1.2 National GMP Regulations and Codes 
2.1.2.1 United States 
2.1.2.2 Canada 
2.1.2.3 European Union 
2.1.2.4 East Asian Countries 
2.1.2.5 India 
2.1.2.6 Australia 
2.1.2.7 New Zealand 
2.1.2.8 South Africa 
2.1.3 International GMP Guides and Harmonization 
2.1.3.1 World Health Organization 
2.1.3.2 Pharmaceutical Inspection Cooperation Scheme 
2.1.3.3 International Conference on Harmonization 
2.1.3.4 Association of Southeast Asian Nations (ASEAN) 
2.1.3.5 Mercado Comun del Sur (MERCOSUR) 
2.1.4 Correspondences of the U.S. GMP Regulations with GMP Codes and Guidelines 
2.1.4.1 General Issues 
2.1.4.2 Organization and Personnel 
2.1.4.3 Buildings and Facilities 
2.1.4.4 Equipment 
2.1.4.5 Control of Components and Drug Product Containers and Closures 
2.1.4.6 Production and Process Controls 

120 CORRESPONDENCES AND DIFFERENCES 
2.1.4.7 Packaging and Labeling Control 
2.1.4.8 Holding and Distribution 
2.1.4.9 Laboratory Controls 
2.1.4.10 Records and Reports 
2.1.4.11 Returned and Salvaged Drug Products 
References 
2.1.1 INTRODUCTION 
The fi rst predecessors of manufacturing and quality requirements, which later 
evolved into good manufacturing practices (GMPs), were issued in the 1940s in the 
United States by the Food and Drug Administration (FDA) [1] . In the general 
meeting of the World Health Organization (WHO) held in 1969, the World Health 
Assembly issued a recommendation for the introduction of GMPs [2] . Since then, 
most industrialized countries have passed laws on control procedures essential for 
the manufacture of drug products. In some countries GMPs are integrated into 
national legislation as a part of laws or regulations on production, distribution, 
marketing, and use of drug products (GMP regulations). In other countries, GMPs 
are separate guidelines outside the national drug legislation (GMP codes). In addition 
to national GMPs, also some international organizations and trade blocks have 
issued their own international GMP guidelines to harmonize the requirements for 
drug production in different countries. However, regardless of their origin, the main 
purpose of GMPs is to ensure that manufactured drug products have the safety, 
identity, potency, purity, and quality that they are presented to have [3] . To fulfi ll 
this aim, most GMPs usually cover quality management, personnel, premises, equipment, 
documentation, materials management, production and in - process controls, 
packaging and labeling of intermediate and fi nished products, laboratory controls, 
validation, and change controls [4] . 
2.1.2 NATIONAL GMP REGULATIONS AND CODES 
2.1.2.1 United States 
In the United States the production of drug products is controlled under the federal 
Food, Drug and Cosmetic Act, which states that a drug product will be deemed to 
be adulterated unless the methods used in or the facilities or controls used for its 
manufacture, processing, packaging, or holding conform to or are operated or 
administered in conformity with current GMP [5] . The actual GMP regulations are 
issued as a part of the Code of Federal Regulations and as such they are a federal 
law. The current set of GMP regulations is based on the 1978 revision [6, 7] of the 
original GMP regulations, which were fi rst promulgated in 1963. The GMP regulations 
are updated every year in April [8] ; however, no major changes have been 
implemented since 1978. As an addition to GMP regulations, the FDA also publishes 
other GMP - related guidance documents covering various issues of drug manufacturing 
[9] . On the other hand, although these documents refl ect current views and 

NATIONAL GMP REGULATIONS AND CODES 121 
expectations of the agency, they only provide guidance on principles and practices 
that are not legal requirements [1] . As a member of the International Conference 
on Harmonization of Technical Requirements for Registration of Pharmaceutical 
for Human Use (ICH), the United States has adopted the ICH guidance document 
Q7, Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients, and 
published it as a guidance for industry document [10] . 
The U.S. GMP regulations are divided into two parts: 210 [6] and 211 [7] . Part 
210, “ Current Good Manufacturing Practice in Manufacturing, Processing, Packing 
or Holding of Drugs — General, ” provides the framework for the regulations [6] , and 
Part 211, “ Current Good Manufacturing Practice for Finished Pharmaceuticals, ” 
states the actual requirements. Part 211 is further divided into 11 subparts, which 
cover the requirements for personnel, premises, equipment, control of materials, 
production and process controls, packaging and labeling control, holding and distribution, 
laboratory controls, documentation, and returned and salvaged products [7] . 
The contents of Part 211 are presented in Table 1 . 
2.1.2.2 Canada 
The production of drug products (drugs) in Canada is controlled under the Food 
and Drugs Act, which states that distributors and importers are not allowed to sell 
a drug product unless it has been manufactured according to the requirements of 
GMP. The principles of GMP are laid down by Division 2 in Part C of the Food and 
Drug Regulations, which is a part of the Food and Drugs Act [11] . The Health 
Products and Food Branch Inspectorate has also issued a guidance document (GMP 
code), which has been prepared to assist in the interpretation of GMP regulations. 
The current set of the Canadian GMP code was issued in 2002 and has not been 
revised since. It has been written with a view to harmonization with GMP standards 
of other countries and international organizations [WHO, Pharmaceutical Inspection 
Cooperation Scheme (PIC/S), ICH]. Canadian Healthcare authorities have also 
published several annexes to the basic GMP code covering topics such as GMP for 
medical gases, biological drug products, blood products, and production of investigational 
new drugs. In addition to the GMP code and its annexes, the Canadian 
TABLE 1 Contents of Part 211 of U . S . GMP Regulations [7] 
Section Subject 
Subpart A General provisions 
Subpart B Organization and personnel 
Subpart C Buildings and facilities 
Subpart D Equipment 
Subpart E Control of components and drug product containers and closures 
Subpart F Production and process controls 
Subpart G Packaging and labeling control 
Subpart H Holding and distribution 
Subpart I Laboratory controls 
Subpart J Records and reports 
Subpart K Returned and salvaged drug products 

122 CORRESPONDENCES AND DIFFERENCES 
authorities have also issued several other specifi c guidelines dealing with issues 
related to GMP and manufacturing methods [12] . 
As shown in Table 2 the Canadian GMP code can be divided into four chapters 
and three annexes. The fi rst three chapters cover general issues such as scope and 
applicability of the code, defi nitions of used terms, and issues concerning quality 
management and GMP in general. GMP regulations and their application are presented 
in the fourth chapter ( “ Regulation ” ), which is divided into 14 subchapters 
covering the requirements for premises, equipment, personnel, sanitation, testing of 
components and packaging materials, testing of fi nished product, production control, 
quality control department, documentation, reserve samples, stability testing, and 
manufacture of sterile drug products and medical gases. Each subchapter contains 
the corresponding regulation according to regulations in Division 2 [11] issued with 
a rationale and interpretation to assist in their application. The annexes include 
requirements for batch certifi cation, application form for alternate sample retention 
site, and references such as hyperlinks to Canadian laws concerning drug products 
and other GMP - related national and international guidelines [12] . 
2.1.2.3 European Union 
The production of drug products (medicinal products) in the European Union (EU) 
is controlled under Directive 2001/83/EC of the European parliament and of the 
Council, which states that the holder of a manufacturing authorization for medicinal 
products is obliged to comply with good manufacturing practices as laid down by 
European Community law [13] . The principles and guidelines of GMP for medicinal 
products are stated by the Commission directive 2003/94/EC, which provides the 
TABLE 2 Contents of Canadian GMP Code [12] 
Introduction 
Quality management 
Glossary of terms 
Regulation 
Premises 
Equipment 
Personnel 
Sanitation 
Raw material testing 
Manufacturing control 
Quality control department 
Packaging material testing 
Finished product testing 
Records 
Samples 
Stability 
Sterile products 
Medical gases 
Annex A: Internationally Harmonized Requirements for 
Batch Certifi cation 
Annex B: Application for Alternate Sample Retention 
Annex C: References 

NATIONAL GMP REGULATIONS AND CODES 123 
legal basis for GMP in the EU [14] . The actual GMP code with detailed written 
procedures is published in The Rules Governing Medicinal Products in the European 
Union , volume 4. The current set of the EU GMP code was fi rst introduced in 1989 
consisting of nine chapters covering the general requirements of GMP and one 
annex on the manufacture of sterile drug products. Since then the EU GMP code 
has been revised many times and several new annexes have been issued [15] . In 
addition to the GMP code, the EU has also published several other guidelines concerning 
the quality issues of drug production in The Rules Governing Medicinal 
Products in the European Union , volume 3A [16] . 
As shown in Tables 3 – 5 , the EU GMP code is presented in two parts of basic 
requirements and 18 annexes. Part I, “ Basic Requirements for Medicinal Products, ” 
covers GMP principles for the manufacture of drug products. It consists of nine 
chapters covering the requirements for quality management and control, personnel, 
premises, equipment, documentation, production, contract services, complaints, 
product recall, and self - inspection. Part II, “ Basic Requirements for Active Substances 
Used as Starting Materials, ” covers GMPs for active substances used as 
starting materials. It is based on the ICH document Q7, Good Manufacturing Practice 
Guide for Active Pharmaceutical Ingredients , and was originally introduced in 
2001 as Annex 18 of the EU GMP code. In the restructured revision of the EU 
GMP code issued in October 2005, Annex 18 was replaced with Part II. It consists 
of 19 chapters, which cover basic GMP issues related to quality management, personnel, 
premises, equipment, documentation, materials, production and process controls, 
packaging and labeling, storage and distribution, laboratory controls, validation, 
change control, complaints, recalls, contract services, co - operators, active pharmaceutical 
ingredients (APIs) manufactured by cell culture/fermentation, and APIs 
used in clinical trials. The annexes give more detailed specifi c guidance on the 
manufacture of sterile drug products, biological drug products, radiopharmaceuticals, 
veterinary drug products, medical gases, herbal drug products, oral liquids, 
external preparations (creams, ointments), aerosols, investigational new drugs, and 
blood and blood products. They also cover sampling of materials, computerized 
systems, use of ionizing radiation, qualifi cation and validation, batch release, parametric 
release, reference, and retention samples [15] . 
TABLE 3 Contents of Part I of EU GMP Code Covering 
Basic Requirements for Manufacture of Drug Products [15] 
Section Subject 
Introduction 
Chapter 1 Quality management 
Chapter 2 Personnel 
Chapter 3 Premises and equipment 
Chapter 4 Documentation 
Chapter 5 Production 
Chapter 6 Quality control 
Chapter 7 Contract manufacture and analysis 
Chapter 8 Complaints and product recall 
Chapter 9 Self - inspection 
Glossary 

124 CORRESPONDENCES AND DIFFERENCES 
TABLE 4 Contents of Part II of EU GMP Code Covering Basic Requirements for 
Manufacture of Active Substances Used as Starting Materials [15] 
Section Subject 
1 Introduction 
2 Quality management 
3 Personnel 
4 Buildings and facilities 
5 Process equipment 
6 Documentation and records 
7 Materials management 
8 Production and in - process controls 
9 Packaging and identifi cation labeling of APIs and intermediates 
10 Storage and distribution 
11 Laboratory controls 
12 Validation 
13 Change control 
14 Rejection and reuse of materials 
15 Complaints and recalls 
16 Contract manufacturers (including laboratories) 
17 Agents, brokers, traders, distributors, repackers, and relabelers 
18 Specifi c guidance for APIs manufactured by cell culture/fermentation 
19 APIs for use in clinical trials 
20 Glossary 
TABLE 5 Annexes of EU GMP Code Covering Specifi c Guidance [15] 
Section Subject 
Annex 1 Manufacture of sterile medicinal products 
Annex 2 Manufacture of biological medicinal products for human use 
Annex 3 Manufacture of radiopharmaceuticals 
Annex 4 Manufacture of veterinary medicinal products other than immunological 
veterinary medicinal products 
Annex 5 Manufacture of immunological veterinary medicinal products 
Annex 6 Manufacture of medicinal gases 
Annex 7 Manufacture of herbal medicinal products 
Annex 8 Sampling of starting and packaging materials 
Annex 9 Manufacture of liquids, creams, and ointments 
Annex 10 Manufacture of pressurised metered - dose aerosol preparations for inhalation 
Annex 11 Computerized systems 
Annex 12 Use of ionizing radiation in manufacture of medicinal products 
Annex 13 Manufacture of investigational medicinal products 
Annex 14 Manufacture of products derived from human blood or human plasma 
Annex 15 Qualifi cation and validation 
Annex 16 Certifi cation by a qualifi ed person and batch release 
Annex 17 Parametric release 
Annex 19 Reference and retention samples 

NATIONAL GMP REGULATIONS AND CODES 125 
2.1.2.4 East Asian Countries 
Japan In Japan the production of drug products (drugs) is regulated under the 
Pharmaceuticals Affairs Law (PAL), which states that any drug manufacturer who 
plans to manufacture a drug product for sale in Japan must have a Japanese drug 
manufacturing license and comply with Japanese GMP requirements. The fi rst regulations 
of Japanese GMP were introduced in 1974 as The Standards for Manufacturing 
Control and Quality Control . In 1979 PAL was partially revised and GMPs 
became legally binding [2] . 
PAL is managed and enforced via ministerial ordinances and notices, which are 
detailed regulations prepared by the Japanese government. The requirements for 
premises for drug manufacture are given in Ministry of Health, Labor and Welfare 
(MHLW) Ministerial Ordinance No. 73, 2005 Regulations for Buildings and Facilities 
for Pharmacies, etc. [originally Ministry of Health and Welfare (MHW) Ministerial 
Ordinance No. 2, 1961] [17] , and the requirements for manufacturing and quality 
controls in MHLW Ministerial Ordinance No. 95, 2003 Regulations for Manufacturing 
Control and Quality Control of Drugs (originally MHW Ministerial Ordinance 
No. 3, 1994). As a member of the ICH Japan has adopted the ICH guidance 
document Q7, Good Manufacturing Practice Guide for Active Pharmaceutical 
Ingredients , and published it as Pharmaceutical and Food Safety Bureau (PFSB) 
Director - General Notifi cation No. 1200, 2001 Guidelines on GMP for Drug Substances 
, which states the requirements for the manufacture of APIs. The requirements 
concerning imported drug products are given in MHLW Ministerial Ordinance 
No. 97, 2003 Regulations for Importing/Retail Management and Quality Control of 
Drugs and Quasi - Drugs (originally MHW Ministerial Ordinance No. 62, 1999). The 
requirements specifying manufacture of investigational products are given in PAB 
Notifi cation No. 480, 1997 Products and Standards for the Buildings and Facilities of 
Manufacturing Plants for Investigational Products (Investigational Product GMP) 
[2] . 
South Korea The production of drug products (drugs) in South Korea is regulated 
under the Pharmaceutical Affairs Law, which was fi rst enacted in 1953 and has since 
been revised several times [18] . New drug approval and related activities are regulated 
in much the same way as in the United States and Japan. Korean GMP, which 
is often called KGMP, was initiated in 1984 and became mandatory in 1995 [19] . A 
drug manufacturer who intends to manufacture a drug product for sale in Korea 
must have approval from the Commissioner of the Korea Food and Drug Administration 
(KFDA). In order to require the license for manufacturing business the 
manufacturer has to prove the compliance of facility standards with KGMP [20] . 
China China regulates the production of drug products (drugs) under the Drug 
Administration Law of the People ’ s Republic of China, which states that a drug 
manufacturer has to conduct drug manufacture according to the GMP for pharmaceutical 
products formulated by the Drug Regulatory Department under the State 
Council on the basis of the Drug Administration Law [21] . In June 2004 GMP 
became mandatory in China and the State Drug Administration announced that 
local drug manufacturing establishments lacking approved GMP certifi cation would 
not be allowed to continue the production of pharmaceuticals [22] . 

126 CORRESPONDENCES AND DIFFERENCES 
2.1.2.5 India 
The production of drug products (drugs) in India is controlled under the Drugs and 
Cosmetics Rules (1945, last amended in 2005), which states that the holder of the 
license to manufacture drugs has to comply with the requirements of GMP as laid 
down in Schedule M [23] . Schedule M is a part of the Drugs and Cosmetics Rules 
and embodies the Indian GMP regulations [24] , which are based on the 1982 version 
of WHO GMP guidelines [25] . 
As shown in Tables 6 – 8 the Indian GMP regulations consists of eight parts: I, IA, 
IB, IC, ID, IE, IF, and II. Part I covers the general requirements of GMP. It is divided 
into 29 chapters, which deal with the requirements for personnel, premises, equipment, 
sanitation, production and process controls, materials, documentation, quality 
management, validation, reserve samples, recalls, complaints, and self - inspection. 
Parts IA to IE cover specifi c requirements for the manufacture of different dosage 
forms regarding premises, equipment, and methods. Part IA deals with the require- 
TABLE 6 Contents of Part I of Indian GMP Regulations 
Covering Good Manufacturing Practices for Premises and 
Materials [24] 
Section Subject 
1 General requirements 
2 Warehousing area 
3 Production area 
4 Ancillary areas 
5 Quality control area 
6 Personnel 
7 Health, clothing, and sanitation of workers 
8 Manufacturing operations and control 
9 Sanitation in the manufacturing premises 
10 Raw materials 
11 Equipment 
12 Documentation and records 
13 Labels and other printed materials 
14 Quality assurance 
15 Self - inspection and quality audit 
16 Quality control system 
17 Specifi cation 
18 Master formula records 
19 Packing records 
20 Batch packaging records 
21 Batch processing records 
22 Standard operating procedures (SOPs) and records 
23 Reference samples 
24 Reprocessing and recoveries 
25 Distribution records 
26 Validation and process validation 
27 Product recalls 
28 Complaints and adverse reactions 
29 Site master fi le 

NATIONAL GMP REGULATIONS AND CODES 127 
ments for the manufacture of parenteral preparations; Part IB with the requirements 
for the manufacture of oral solid dosage forms such as tablets and capsules; 
Part IC with the requirements for the manufacture of oral liquids such as syrups, 
elixirs, emulsions, and suspensions; Part ID with the requirements for the manufacture 
of external preparations such as creams, ointments, pastes, emulsions, and 
lotions; and Part 1E with the requirements for the manufacture of inhalers. Part 1F 
covers specifi c requirements for the manufacture of APIs regarding buildings and 
facilities, utilities, equipment, controls, and containers. Part II of the Indian GMP 
regulations consist of detailed recommendations for the process equipment to be 
used in the manufacture of different dosage forms and requirements for the partition 
of the production area [24] . 
TABLE 7 Contents of Parts IA , IB , IC , ID , IE , and IF of Indian GMP Regulations 
Covering Specifi c Guidance [24] 
Section Subject 
Part IA Specifi c requirements for manufacture of sterile products, parenteral 
preparations (small - volume injectables and large - volume parenterals) and 
sterile ophthalmic preparations 
Part IB Specifi c requirements for manufacture of oral solid dosage forms (tablets and 
capsules) 
Part IC Specifi c requirements for manufacture of oral liquids (syrups, elixirs, 
emulsions, and suspensions) 
Part ID Specifi c requirements for manufacture of topical products, i.e., external 
preparations (creams, ointments, pastes, emulsions, lotions, solutions, dusting 
powders, and identical products) 
Part IE Specifi c requirements for manufacture of metered - dose inhalers (MDIs) 
Part IF Specifi c requirements of premises, plant, and materials for manufacture of 
active pharmaceutical ingredients (bulk drugs) 
TABLE 8 Contents of Part II of Indian GMP Regulations 
Covering Requirements of Plant and Equipment [24] 
Section Subject 
1 External preparations 
2 Oral liquid preparations 
3 Tablets 
4 Powders 
5 Capsules 
6 Surgical dressing 
7 Ophthalmic preparations 
8 Pessaries and suppositories 
9 Inhalers and vitrallae 
10 Repacking of drugs and pharmaceutical chemicals 
11 Parenteral preparations 

128 CORRESPONDENCES AND DIFFERENCES 
2.1.2.6 Australia 
In Australia the production of drug products (medicinal products) is controlled 
under the Therapeutics Goods Act, which provides the Minister for Health and 
Aged Care the right to determine written principles including codes of GMP to be 
observed in the production of drug products for use in humans [26] . The Therapeutic 
Goods (Manufacturing Principles) Determination No. 2 of 2002 given by the 
minister states that drug products must be manufactured in compliance with the 
Australian Code of Good Manufacturing Practice for Medicinal Products , dated 
August 16, 2002 [27] . The current set of the Australian GMP code is based entirely 
on the PIC/S GMP guide version PH 1/97 (Rev. 3) published in 2002 with some 
minor modifi cations [28] . 
As shown in Table 9 , the Australian GMP code consists of 9 chapters and 13 
annexes. The chapters present the general requirements of GMP for the manufacture 
of drug products, the requirements for quality management and control, 
personnel, premises, equipment, documentation, production, contract services, complaints, 
product recall, and self - inspection. The annexes give specifi c guidance on 
the manufacture of sterile drug products, biological drug products, radiopharmaceuticals, 
medical gases, herbal drug products, oral liquids, external preparations (creams, 
ointments), aerosols, investigational new drugs, blood, and blood products. They also 
TABLE 9 Contents of Australian GMP Code [28] 
Section Subject 
Introduction 
Interpretation 
Chapter 1 Quality management 
Chapter 2 Personnel 
Chapter 3 Premises and equipment 
Chapter 4 Documentation 
Chapter 5 Production 
Chapter 6 Quality control 
Chapter 7 Contract manufacture and analysis 
Chapter 8 Complaints and product recall 
Chapter 9 Self - inspection 
Annex 1 Manufacture of sterile medicinal products 
Annex 2 Manufacture of biological medicinal products for human use 
Annex 3 Manufacture of radiopharmaceuticals 
Annex 6 Manufacture of medicinal gases 
Annex 7 Manufacture of herbal medicinal products 
Annex 8 Sampling of starting and packaging materials 
Annex 9 Manufacture of liquids, creams, and ointments 
Annex 10 Manufacture of pressurised metered - dose aerosol preparations for 
inhalation 
Annex 11 Computerized systems 
Annex 12 Use of ionizing radiation in the manufacture of medicinal products 
Annex 13 Manufacture of investigational medicinal products 
Annex 15 Qualifi cation and validation 
Annex 17 Parametric release 
Glossary 

NATIONAL GMP REGULATIONS AND CODES 129 
cover sampling of materials, computerized systems, use of ionizing radiation, quali- 
fi cation, and validation and parametric release [28] . 
Australia has not adopted Annexes 4, 5, 14, 16, and 18 of the PIC/S GMP guide. 
Annexes 4 and 5 cover the manufacture of veterinary drug products. Annex 14 
covers the manufacture of products derived from human blood or human plasma, 
which is excluded from the Australian GMP code. Annex 16 is specifi c to the EU 
GMP code and Annex 18 is the ICH GMP guide for the manufacture of APIs, which 
Australia has adopted separately as a manufacturing principle [28] . 
2.1.2.7 New Zealand 
The production of drug products (medicines) in New Zealand is controlled under 
the Medicines Act 1981, which states that a drug manufacturer is not allowed to 
manufacture drug products without a manufacturing license issued by the licensing 
authority. In order to obtain a manufacturing license the applicant must satisfy the 
licensing authority with respect to the proposed manufacturing premises and equipment, 
which must be suitable and adequate for the manufacture of drugs. Moreover, 
the applicant must show that adequate arrangements have been made or are to be 
made for the making, maintaining, and safekeeping of adequate records with reference 
to the drug products that are to be manufactured [29] . The authorities (Medsafe) 
require that any drug manufacturer who plans to manufacture drug products for 
sale in New Zealand must deliver evidence of GMP compliance for the manufacturing 
site. Copies of appropriate certifi cates, manufacturing licenses, or reports issued 
by a regulatory authority whose competence is recognized by Medsafe are accepted 
as proof of GMP compliance [30] . 
As shown in Table 10 New Zealand ’ s own GMP code consists of fi ve parts. The 
fi rst part covers the manufacture of drug products and the second part the manufacture 
of blood products. Part 3 covers compounding and dispensing, including 
compounding of sterile drug products. Part 4 deals with wholesaling and Part 5 with 
product recalls. Parts 4 and 5 are combined in one document [31] . 
2.1.2.8 South Africa 
South Africa controls the production of drug products (medicines) under the Medicines 
and Related Substances Control Act (Act 101 of 1965), which states that the 
Medicines Control Council may issue to a drug manufacturer a license to manufacture 
a drug product upon such conditions as to the application of such acceptable 
TABLE 10 Contents of New Zealand ’ s GMP Code [31] 
Section Subject 
Part 1 Manufacture of pharmaceutical products 
Part 2 Manufacture of blood and blood products 
Part 3 Compounding and dispensing 
Part 4 Wholesaling of medicines and medical devices 
Part 5 Uniform recall procedure for medicines and 
medical devices 

130 CORRESPONDENCES AND DIFFERENCES 
quality assurance principles and GMPs as the council may determine [32] . As a part 
of the license application the manufacturer must provide acceptable documentary 
proof of the ability to comply with GMP as determined by the council [33] . The 
current set of South African GMP code determined by the council is entirely based 
on the PIC/S GMP guide version PE 009 - 2, published in 2004 with some minor 
modifi cations [34] . 
As shown in Table 11 the South African GMP code consists of 9 chapters and 17 
annexes. The chapters present the general requirements of GMP for the production 
of drug products covering the requirements for quality management and control, 
personnel, premises, equipment, documentation, production, contract services, complaints, 
product recall, and self - inspection. The annexes give specifi c guidance on 
the manufacture of sterile drug products, biological drug products, radiopharmaceuticals, 
veterinary drug products, medical gases, herbal drug products, oral liquids, 
external preparations (creams, ointments), aerosols, investigational new drugs, and 
blood and blood products. They also cover sampling of materials, computerized 
systems, use of ionizing radiation, qualifi cation and validation, organization, and 
TABLE 11 Contents of South African GMP Code [34] 
Section Subject 
Introduction 
Chapter 1 Quality management 
Chapter 2 Personnel 
Chapter 3 Premises and equipment 
Chapter 4 Documentation 
Chapter 5 Production 
Chapter 6 Quality control 
Chapter 7 Contract manufacture and analysis 
Chapter 8 Complaints and product recall 
Chapter 9 Self - inspection 
Annex 1 Manufacture of sterile medicinal products 
Annex 2 Manufacture of biological medicinal products for human use 
Annex 3 Manufacture of radiopharmaceuticals 
Annex 4 Manufacture of veterinary medicinal products other than 
immunologicals 
Annex 5 Manufacture of immunological veterinary medical products 
Annex 6 Manufacture of medicinal gases 
Annex 7 Manufacture of herbal medicinal products 
Annex 8 Sampling of starting and packaging materials 
Annex 9 Manufacture of liquids, creams, and ointments 
Annex 10 Manufacture of pressurized metered - dose aerosol preparations for 
inhalation 
Annex 11 Computerized systems 
Annex 12 Use of ionizing radiation in the manufacture of medicinal products 
Annex 13 Manufacture of investigational medicinal products 
Annex 14 Manufacture of products derived from human blood or human plasma 
Annex 15 Qualifi cation and validation 
Annex 16 Organisation and personnel 
Annex 17 Parametric release 
Glossary 

personnel and parametric release. The original Annex 16, which is specifi c to the 
EU GMP code, has been replaced in South African GMP code with an annex covering 
organization and personnel. Nor has South Africa adopted Annex 18, which 
covers the ICH GMP guide for the manufacture of APIs, as it has been adopted 
separately as a manufacturing principle [34] . 
2.1.3 INTERNATIONAL GMP GUIDES AND HARMONIZATION 
2.1.3.1 World Health Organization 
The WHO was established in 1948 as a specialized agency of the United Nations 
(UN). Its purpose is to serve as the directing and coordinating authority for international 
health matters and public health. One of the main functions of the WHO 
is to provide objective and reliable information and advice in the fi eld of human 
health, a task that it partly fulfi lls through WHO publications [35] . The fi rst WHO 
draft text on GMP was prepared in 1967 and a revised version was published in 
1968 as an annex of the twenty - second report of the WHO expert committee on 
specifi cations for pharmaceutical preparations. Over the years the WHO has issued 
several versions of its GMP guidelines as well as other guidelines related to the 
GMP and quality issues of the production of therapeutic products. The latest version 
of the WHO GMP guideline was published in 2003 as an annex of the WHO Technical 
Report 908 [36] . 
As shown in Table 12 the WHO GMP guideline is divided into fi ve parts: introduction, 
general considerations, glossary, quality management in the drug industry, 
and references. The actual GMP guidelines are presented in the fourth part, which 
consists of 17 chapters covering the requirements for quality assurance and control, 
personnel, premises, equipment, sanitation, materials, validation, documentation, 
production, contract services, complaints, recalls, and self - inspection [36] . In addition 
to this guideline laying down the main principles of GMP, the WHO has also published 
several other guidelines covering specifi c requirements for components, 
quality of water for pharmaceutical use, APIs, excipients, sterile drug products, 
biological drug products, investigational drug products, herbal drug products, and 
radiopharmaceuticals (Table 13 ). 
2.1.3.2 Pharmaceutical Inspection Cooperation Scheme 
The Pharmaceutical Inspection Convention (PIC), which is the predecessor of 
PIC/S, was founded in 1970 by the European Free Trade Area (EFTA). The initial 
members comprised of the 10 EFTA member countries at that time. From the beginning 
one of the main goals has been the harmonization of GMP requirements as 
well as the promotion of mutual recognition of inspections and uniformity of inspection 
systems by training the inspectors, improving the exchange of information, and 
mutual confi dence [46] . Originally PIC was a formal treaty between member countries 
and as such it also had a legal status. When countries outside Europe were 
seeking to join PIC, it became evident that, according to European law, individual 
EU countries that were members of PIC were not permitted to sign agreements 
with countries outside Europe. Only the European Commission, which itself was 
INTERNATIONAL GMP GUIDES AND HARMONIZATION 131

132 CORRESPONDENCES AND DIFFERENCES 
not a member of PIC, was permitted to sign agreements. Consequently, a less formal 
and more fl exible PIC/S was developed to continue the work of PIC. The PIC/S, 
which became operational in November 1995, is an informal arrangement without 
legal status between regulatory authorities instead of countries. The PIC and the 
PIC Scheme, operating together as PIC/S, provide an active and constructive cooperation 
in the fi eld of GMP [47] . 
The current members of PIC/S are Australia, Austria, Belgium, Canada, Czech 
Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, 
TABLE 12 Contents of WHO GMP Guideline Covering General Requirements of GMP 
for Manufacture of Drug Products [36] 
Introduction 
General considerations 
Glossary 
Quality management in the drug industry: philosophy and essential elements 
Section Subject 
1 Quality assurance 
2 Good manufacturing practices for pharmaceutical products (GMP) 
3 Sanitation and hygiene 
4 Qualifi cation and validation 
5 Complaints 
6 Product recalls 
7 Contract production and analysis 
8 Self - inspection and quality audits 
9 Personnel 
10 Training 
11 Personal hygiene 
12 Premises 
13 Equipment 
14 Materials 
15 Documentation 
16 Good practices in production 
17 Good practices in quality control 
References 
TABLE 13 GMP -Related WHO Documents Covering Specifi c Guidance 
Document Subject 
TRS 929, Annex 2 [37] Requirement for the sampling of starting materials 
TRS 823, Annex 1 [38] Active pharmaceutical ingredients (bulk drug substances) 
TRS 885, Annex 5 [39] Pharmaceutical excipients 
TRS 902, Annex 6 [40] Sterile pharmaceutical products 
TRS 834, Annex 3 [41] Biological products 
TRS 863, Annex 7 [42] Investigational pharmaceutical products for clinical trials 
in humans 
TRS 863, Annex 8 [43] Herbal medicinal products 
TRS 908, Annex 3 [44] Radiopharmaceutical products 
TRS 929, Annex 3 [45] Water for pharmaceutical use 

Italy, Latvia, Liechtenstein, Malaysia, Netherlands, Norway, Poland, Portugal, 
Romania, Singapore, Slovak Republic, Spain, Sweden, Switzerland, and the United 
Kingdom. In addition, Estonia, the European Agency for the Evaluation of Medicinal 
Products (EMEA), UNICEF, and the WHO participate in PIC/S activities as 
observers [48] . Also many other regulatory authorities have shown interest in joining 
PIC/S, in particular Argentina, Brazil, Cyprus, Indonesia, Israel, Philippines, 
Slovenia, Thailand, the United States, Bulgaria, Estonia, Lithuania, Oman, Russia, 
South Africa, and the Ukraine [49] . 
To become a PIC/S member, a joining regulatory authority is required to go 
through a detailed assessment to prove that the authority has the arrangements and 
competence necessary to apply an inspection system equivalent to inspection 
systems of existing PIC/S members. To ensure that both new applicants and older 
members fulfi ll the same requirements, also existing members are reassessed on a 
regular basis. One of the main functions of PIC/S is to develop GMP guidance documents, 
which it carries out in close cooperation with the EU and relevant agencies 
thereof. Under this cooperation both parties have been able to adopt each others ’ 
documents, thus minimizing the duplication of effort in development of GMP - 
related documents. Among other highly informative guides on various aspects of 
GMP and quality issues [49] , PIC/S has also published its own GMP guide ( Guide 
to Good Manufacturing Practice for Medicinal Products ), which is harmonized with 
the EU GMP code [50] . 
The latest revision of the PIC/S GMP guide (version PE 009 - 3) was issued in 
January 2006. As shown in Table 14 , it consists of 9 chapters and 16 annexes. Chapters 
present the general requirements of GMP for the production of drug products 
covering the requirements for quality management and control, personnel, premises, 
equipment, documentation, production, contract services, complaints, product recall, 
and self - inspection. The annexes give specifi c guidance on the manufacture of sterile 
drug products, biological drug products, radiopharmaceuticals, veterinary drug products, 
medical gases, herbal drug products, oral liquids, external preparations (creams, 
ointments), aerosols, investigational new drugs, and blood and blood products. In 
addition, there are annexes covering the sampling of materials, computerized 
systems, use of ionizing radiation, qualifi cation and validation, and parametric 
release [50] . 
Although the PIC/S GMP guide is harmonized with the EU GMP code and their 
contents are similar, there are some minor differences between them. Instead of the 
term qualifi ed person , the PIC/S GMP guide uses the term authorized person . Furthermore, 
all references to EU directives have been deleted from the PIC/S GMP 
guide. Moreover, PIC/S has not adopted Annexes 16 and 18 of the EU GMP code. 
Annex 16 is specifi c to the EU GMP code covering the status of a qualifi ed person 
in batch release and Annex 18 is the ICH GMP guide for the manufacture of APIs, 
which the PIC/S Committee has adopted as a stand - alone document (PE 007) 
[50] . 
2.1.3.3 International Conference on Harmonization 
The ICH was established in 1990. Its main aim is to improve the effi ciency of the 
drug development process and the registration of new drug products in its member 
countries through harmonization of national guidelines. This is a joint initiative 
INTERNATIONAL GMP GUIDES AND HARMONIZATION 133

134 CORRESPONDENCES AND DIFFERENCES 
involving both regulators and industry as equal partners. The founders and current 
members of ICH, which represent the regulatory bodies and the research - based 
industry in the member countries, are the EU, European Federation of Pharmaceutical 
Industries and Associations (EFPIA), MHLW, Japan Pharmaceutical Manufacturers 
Association (JPMA), FDA, and Pharmaceutical Research and Manufactures 
of America (PhRMA). In addition to the actual member countries there are also 
observers who act as a link between ICH and non - ICH countries and regions. 
Current observers are the WHO, EFTA, Swissmedic (representing Switzerland), and 
Health Canada (representing Canada) [51] . 
Among other guidelines, ICH has also published a guide on GMP for APIs (Q7: 
Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients ). It is 
intended to provide guidance regarding GMP for the manufacture of APIs and to 
help ensure that APIs meet the quality and purity requirements that they are presented 
to possess. This covers APIs that are manufactured by chemical synthesis, 
extraction, cell culture/fermentation, recovery from natural sources, or any combination 
of these processes. Excluded are vaccines, medical gases, bulk - packaged drug 
TABLE 14 Contents of PIC / S GMP Guide [50] 
Section Subject 
Introduction 
Chapter 1 Quality management 
Chapter 2 Personnel 
Chapter 3 Premises and equipment 
Chapter 4 Documentation 
Chapter 5 Production 
Chapter 6 Quality control 
Chapter 7 Contract manufacture and analysis 
Chapter 8 Complaints and product recall 
Chapter 9 Self - inspection 
Annex 1 Manufacture of sterile medicinal products 
Annex 2 Manufacture of biological medicinal products for human use 
Annex 3 Manufacture of radiopharmaceuticals 
Annex 4 Manufacture of veterinary medicinal products other than 
immunologicals 
Annex 5 Manufacture of immunological veterinary medical products 
Annex 6 Manufacture of medicinal gases 
Annex 7 Manufacture of herbal medicinal products 
Annex 8 Sampling of starting and packaging materials 
Annex 9 Manufacture of liquids, creams, and ointments 
Annex 10 Manufacture of pressurised metered - dose aerosol preparations for 
inhalation 
Annex 11 Computerized systems 
Annex 12 Use of ionizing radiation in manufacture of medicinal products 
Annex 13 Manufacture of investigational medicinal products 
Annex 14 Manufacture of products derived from human blood or human plasma 
Annex 15 Qualifi cation and validation 
Annex 17 Parametric release 
Glossary 

products, radiopharmaceuticals, whole cells, whole blood and plasma, blood and 
plasma derivatives, and gene therapy APIs. However, APIs that are produced 
using blood or plasma as raw materials are included [52] . All ICH member countries 
have adopted this guideline: the EU in November 2000, Japan in November 
2001, and the United States in September 2001 [53] . In addition, it has also been 
adopted by several other non - ICH countries such as Australia [28] and South Africa 
[34] . 
The basic structure of the ICH GMP guideline for API production is shown in 
Table 15 . It consists of 19 chapters, which cover the requirements for quality management, 
personnel, premises, equipment, documentation, materials, production and 
process controls, packaging and labeling, storage and distribution, laboratory controls, 
validation, change control, complaints, recalls, contract services, cooperators, 
APIs manufactured by cell culture/fermentation, and APIs used in clinical trials 
[52] . 
2.1.3.4 Association of Southeast Asian Nations ( ASEAN ) 
ASEAN was established in 1967 by Indonesia, Malaysia, Philippines, Singapore, and 
Thailand. Current members include also Brunei and Darussalam (joined in 1984), 
Vietnam (joined in 1995), Laos and Myanmar (joined in 1997), and Cambodia 
(joined in 1999). The aims and purposes of ASEAN involve cooperation in the 
economic, social, cultural, technical, educational, and other fi elds [54] . Among other 
cooperation schemes the ASEAN countries have also developed their own GMP 
guidelines, which were issued in 1984 [55] . 
TABLE 15 Contents of ICH GMP Guideline for API Production [52] 
Section Subject 
1 Introduction 
2 Quality management 
3 Personnel 
4 Buildings and facilities 
5 Process equipment 
6 Documentation and records 
7 Materials management 
8 Production and in - process controls 
9 Packaging and identifi cation labeling of APIs and intermediates 
10 Storage and distribution 
11 Laboratory controls 
12 Validation 
13 Change control 
14 Rejection and reuse of materials 
15 Complaints and recalls 
16 Contract manufacturers (including laboratories) 
17 Agents, brokers, traders, distributors, repackers, and relabelers 
18 Specifi c guidance for APIs manufactured by cell culture/fermentation 
19 APIs for use in clinical trials 
20 Glossary 
INTERNATIONAL GMP GUIDES AND HARMONIZATION 135

136 CORRESPONDENCES AND DIFFERENCES 
2.1.3.5 Mercado Comun del Sur ( MERCOSUR ) 
MERCOSUR was established in 1991 by Argentina, Brazil, Paraguay, and Uruguay 
to develop a common market between its member countries. Current members 
include also Bolivia and Chile (joined in 1996). One of the original aims was to 
harmonize the pharmaceutical legislation of the member countries. As a part of 
these harmonization activities MERCOSUR has developed its own GMP guidelines, 
which are based on WHO recommendations. In addition to the GMP guideline, 
MERCOSUR has also issued other GMP - related guides covering inspections, 
requirements for facilities, and quality control [56] . 
2.1.4 CORRESPONDENCES OF THE U . S . GMP REGULATIONS WITH 
GMP CODES AND GUIDELINES 
The following sections deal with the correspondences and differences between the 
U.S. GMP regulations and the Canadian and EU GMP codes and the WHO GMP 
guideline. As the EU GMP code is harmonized with the PIC/S GMP guide, the 
correspondences between the EU GMP code and the U.S. GMP regulations cover 
also the correspondences between the U.S. GMP regulations and the PIC/S GMP 
guide as well as all other national GMPs that are based on the PIC/S GMP guide. 
Differences between the EU GMP code and the PIC/S GMP guide have been presented 
in Section 2.1.3.2 . 
2.1.4.1 General Issues 
In the U.S. GMP regulations general issues related to the use and applicability of 
GMP regulations are presented in Part 210 [6] , which consists of regulations 210.1, 
210.2, and 210.3 and in Subpart A of Part 211 [7] , which consists of regulations 211.1 
and 211.3. Contents of Part 210 and Subpart A of Part 211 are presented in Table 
16 . Regulation 210.1 defi nes the status, 210.2 deals with the applicability, and 211.1 
states the scope of the regulations. Defi nitions of terms used in the regulations are 
provided in regulation 210.3 and in regulation 211.3, which states that the defi nitions 
provided in regulation 210.3 apply also in Part 211. 
Correspondences in Canadian GMP Code In the Canadian legislation general 
issues related to the use and applicability of the GMP regulations and code are 
covered in the introduction of the GMP code [12] and in Divisions 1A [57] and 2 
TABLE 16 Contents of Part 210 and Subpart A of Part 211 of US GMP Regulations 
Covering General Issues Related to Use and Applicability of Regulations [6, 7] 
Section Subject 
CFR 210.1 Status of current good manufacturing practice regulations 
CFR 210.2 Applicability of current good manufacturing practice regulations 
CFR 210.3 Defi nitions 
CFR 211.1 Scope 
CFR 211.3 Defi nitions 

of the Part C of the Food and Drug Regulations [11] . Defi nitions for the GMP regulations 
are covered in regulation C.01A.001 of Division 1A [57] and in regulation 
C.02.002 of Division 2 [11] . Defi nitions for the GMP code are covered in the glossary 
of terms of the code [12] . 
Correspondences in EU GMP Code In the EU legislation general issues related 
to the use and applicability of the GMP regulations and the code are covered in the 
Commission Directive 2003/94/EC [14] and in the introduction of the GMP code 
[15] . Defi nitions for the directive are covered in Article 2 of the directive [14] and 
defi nitions for the code in the glossary of the GMP code [15] . 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
general issues related to the use and applicability of the GMP guide are covered in 
section “ General Considerations. ” Defi nitions for the GMP guide are covered in the 
glossary of the guideline. 
2.1.4.2 Organization and Personnel 
For GMP regulations in the United States issues related to organization and personnel 
are covered in Subpart B [7] , which consists of regulations 211.22, 211.25, 211.28, 
and 211.34. The contents of Subpart B is presented in Table 17 . Regulation 211.22 
states the responsibilities and authorities of the quality control unit, including 
requirements for the resources. Regulation 211.25 deals with personnel qualifi cations 
covering the requirements for their education and experience and it also states 
the requirements for the training of the personnel. Regulation 211.28 states the 
responsibilities of personnel covering the requirements for the clothing and other 
protective apparel, personal sanitation and health habits, as well as personal health 
conditions. Furthermore, it states the requirements for the authorization for limited 
access. Regulation 211.34 deals with consultants and lays down the requirements 
for their education, training, and experience, including the requirements for 
documentation. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to organization and personnel are mainly covered in the interpretation of 
regulation C.02.006 (Personnel) and partly in the interpretations of regulations 
C.02.004 (Premises), C.02.008 (Sanitation), C.02.011 (Manufacturing Control), 
C.02.013 (Quality Control Department), C.02.015 (Quality Control Department), 
and C.02.024 (Records). Correspondences to regulation 211.22 are covered in 
TABLE 17 Contents of Subpart B of Part 211 of U. S . 
GMP Regulations Covering Organization and Personnel [7] 
Section Subject 
CFR 211.22 Responsibilities of quality control unit 
CFR 211.25 Personnel qualifi cations 
CFR 211.28 Personnel responsibilities 
CFR 211.34 Consultants 
CORRESPONDENCES 137

138 CORRESPONDENCES AND DIFFERENCES 
Sections 1 – 5 of the interpretation of regulation C.02.015 and in Section 2 of the 
interpretation of regulation C.02.013. Sections 1 – 5 of the interpretation of regulation 
C.02.015 state the responsibilities of the quality control unit (quality control 
department), and Section 2 of the interpretation of regulation C.02.013 covers the 
requirements for resources. Correspondences to regulation 211.25 stating the 
requirements for the education, training, and experience of the personnel are 
covered in Sections 1 – 5 of the interpretation of regulation C.02.006. Correspondences 
to regulation 211.28 are covered in Section 6.3 of the interpretation of regulation 
C.02.004, Sections 1 – 2 of the interpretation of regulation C.02.008, Section 8 
of the interpretation of regulation C.02.011, and Section 4 of the interpretation of 
regulation C.02.013. Section 1 of the interpretation of regulation C.02.008 states the 
health requirements and Section 2 the requirements for clothing, other protective 
apparel, and personal hygiene. Section 6.3 of the interpretation of regulation 
C.02.004, Section 8 of the interpretation of regulation C.02.011, and Section 4 of the 
interpretation of regulation C.02.013 cover the requirements regarding limited 
access. Correspondences to regulation 211.34 are covered in Section 6 of the interpretation 
of regulation C.02.006 and in Subsection 1.3.2 of the interpretation of 
regulation C.02.024. Section 6 of the interpretation of regulation C.02.006 states the 
requirements for the education, training, and experience of consultants and contractors 
and Subsection 1.2.3 of the interpretation of regulation C.02.024 the requirements 
for documentation. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
organization and personnel are mainly covered in Chapter 2 (Personnel) and partly 
in Chapters 3 (Premises and Equipment), 5 (Production), and 6 (Quality Control). 
Correspondences to regulation 211.22 are covered in Subchapters 2.6, 2.7, 6.1, and 
6.2. Subchapters 2.6 and 2.7 deal with the responsibilities of the head of the quality 
control unit (quality control department) and 6.2 with the responsibilities of 
the quality control unit as a whole. Requirements for resources are covered in 
Subchapter 6.1. Correspondences to regulation 211.25 are covered in Subchapters 
2.1, 2.4, and 2.8 – 2.12. Subchapters 2.1 and 2.4 deal with the requirements for personnel 
and Subchapters 2.8 – 2.12 with the requirements for their training. Correspondences 
to regulation 211.28 are covered in Subchapters 2.15, 2.16, 3.5, 3.21, 5.16, and 
6.4. Subchapter 2.15 deals with requirements for personal health conditions and 2.16 
with requirements for clothing and protection. Access limitations are covered in 
Subchapters 3.5, 3.21, 5.16, and 6.4. In the EU GMP code there is no correspondence 
to regulation 211.34, which covers the requirements for the use of consultants. 
However, Chapter 7 of the code deals with the requirements for the contract services 
in general. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to organization and personnel are mainly covered in Chapter 9 (Personnel) 
and partly in Chapters 10 (Training), 11 (Personal Hygiene), 16 (Good 
Practices in Production), and 17 (Good Practices in Quality Control). Correspondences 
to regulation 211.22 are covered in Subchapters 9.8, 9.10, 17.3, and 17.4. 
Subchapters 9.8 and 9.10 state the responsibilities of the head of the quality control 
unit and Subchapter 17.4 the responsibilities of the quality control unit as a whole. 
Subchapter 17.3 covers the requirements for resources. Correspondences to regula

tion 211.25 are covered in Subchapters 9.2, 9.4, 9.7, and 10.1 – 10.4. Subchapters 9.2, 
9.4, and 9.7 state the requirements for the personnel covering their education and 
experience and Subchapters 10.1 – 10.4 the requirements for the training. Correspondences 
to regulation 211.28 are covered in Subchapters 11.1 – 11.8, 9.5, and 16.7. 
Subchapters 11.1 – 11.5 state the requirements for the health conditions and personal 
hygiene and Subchapters 11.6 – 11.8 the requirements for the clothing and other 
protective apparel. Subchapters 9.5 and 16.7 cover the requirements for the limited 
access. Correspondences to regulation 211.34 are covered in Subchapter 10.6, which 
covers the requirements for the use of consultants. 
2.1.4.3 Buildings and Facilities 
In the United States GMP regulations on issues related to buildings and facilities 
are covered in Subpart C [7] , which consists of regulations 211.42, 211.44, 211.46, 
211.48, 211.50, 211.52, 211.56, and 211.58. Contents of Subpart C are presented in 
Table 18 . Regulation 211.42 deals with design and construction features covering 
the requirements for the size, construction, and location of buildings used in the 
production. Furthermore, it states the requirements for the placement of equipment 
as well as the fl ow of materials and products and specifi es operations, which have 
to be performed in separate or defi ned areas to prevent contamination or mix - ups. 
It also covers the special requirements for the facilities used in aseptic processing 
and facilities used in the production of penicillin. Regulation 211.44 states the 
requirements for lighting and 211.46 for ventilation, including the requirements for 
controls and air - handling systems. Furthermore, it states the special requirements 
for ventilation in the production of penicillin. Regulation 211.48 deals with plumbing 
covering requirements for the plumbing system, drains, and the quality of potable 
water. Regulation 211.50 deals with sewage, trash, and other refuse stating the 
requirements for their disposal. Regulation 211.52 covers the requirements for 
washing and toilet facilities. Regulation 211.56 deals with sanitation stating the 
requirements for the conditions to be maintained in the manufacturing facilities. It 
also states the requirements for handling of trash and organic waste. Furthermore, 
it states the requirements for the written procedures for sanitation operations 
and use of biocides, fumigating, cleaning, and sanitizing agents. It also states the 
TABLE 18 Contents of Subpart C of Part 211 of U. S . 
GMP Regulations Covering Buildings and Facilities [7] 
Section Subject 
CFR 211.42 Design and construction features 
CFR 211.44 Lighting 
CFR 211.46 Ventilation, air fi ltration, air heating and 
cooling 
CFR 211.48 Plumbing 
CFR 211.50 Sewage and refuse 
CFR 211.52 Washing and toilet facilities 
CFR 211.56 Sanitation 
CFR 211.58 Maintenance 
CORRESPONDENCES 139

140 CORRESPONDENCES AND DIFFERENCES 
requirements for the use of biocides and the scope of sanitation procedures. Regulation 
211.58 states the requirements for the maintenance of the buildings used in the 
production. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to buildings and facilities are mainly covered in the interpretation of regulation 
C.02.004 (Premises) and partly in the interpretations of regulations C.02.005 
(Equipment), C.02.007 (Sanitation), C.02.009 (Raw Material Testing), C.02.011 
(Manufacturing Control), and C.02.029 (Sterile Products). Correspondences to regulation 
211.42 are covered in Sections 1, 2, 2.3, 6, 6.2, and 6.4 of the interpretation 
of regulation C.02.004, in Section 15 of the interpretation of regulation C.02.011, 
and in section “ Premises ” of the interpretation of regulation C.02.029. Sections 1 
and 2 of the interpretation of regulation C.02.004 cover the requirements for the 
size, construction, and location of buildings used in the production. Section 6.2 of 
the interpretation of regulation C.02.004 and Section 15 of the interpretation of 
regulation C.02.011 state the requirements for the placement of equipment. Requirements 
for the fl ow of materials and products are covered in Section 6 of the interpretation 
of regulation C.02.004 and operations, which have to be performed in 
separate or defi ned areas in Sections 2.3 and 6.4 of the interpretation of regulation 
C.02.004. The requirements for the facilities used in aseptic processing are covered 
in section “ Premises ” of the interpretation of regulation C.02.029 and the requirements 
for the facilities used in the production of penicillin in Section 11.1 of the 
interpretation of regulation C.02.004. Correspondences to regulation 211.44 stating 
the requirements for the lighting are covered in Section 6.5 of the interpretation of 
regulation C.02.004. Correspondences to regulation 211.46 are covered in Sections 
3.6 and 4 of the interpretation of regulation C.02.004. Section 3.6 of the interpretation 
of regulation C.02.004 states the requirements for the air - handling systems and 
Section 4 the requirements for the control of temperature and humidity. The specifi c 
requirements regarding the production of penicillin are covered in Section 11.1. 
Correspondences to regulation 211.48 are covered in Sections 3.5 and 7 of the 
interpretation of regulation C.02.004, Section 3.7 of the interpretation of regulation 
C.02.005, Section 4 of the interpretation of regulation C.02.009, and section “ Water 
Treatment Systems ” of the interpretation of regulation C.02.029. Section 7 of the 
interpretation of regulation C.02.004 states the requirements for the utilities and 
support systems, including supplies of purifi ed water. Section 3.7 of the interpretation 
of regulation C.02.005 states the requirements for the operation of water puri- 
fi cation, storage, and distribution equipment. Requirements for the quality of water 
are covered in Section 4 of the interpretation of regulation C.02.009 and in section 
“ Water Treatment Systems ” of the interpretation of regulation C.02.029. The requirements 
for drains are covered in Section 3.5 of the interpretation of regulation 
C.02.004. Correspondences to regulation 211.50 stating the requirements for the 
handling of sewage and refuse are covered in Section 2.6 of the interpretation of 
regulation C.02.007. Correspondences to regulation 211.52 stating the requirements 
for washing and toilet facilities are covered in Section 5 of the interpretation of 
regulation C.02.004. Correspondences to regulation 211.56 stating the requirements 
for sanitation are covered in Sections 1 and 2 of the interpretation of regulation 
C.02.007. The Canadian GMP code does not state any separate requirements for 
the handling of organic waste. General requirements for the handling of waste 

materials are covered in Section 2.6 of the interpretation of regulation C.02.007. 
Correspondences to regulation 211.58 stating the requirements for the maintenance 
of the premises are covered in Section 9 of the interpretation of regulation 
C.02.004. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
buildings and facilities are mainly covered in Chapter 3 (Premises and Equipment) 
and partly in Annex 1 (Manufacture of Sterile Medicinal Products). Correspondences 
to regulation 211.42 are covered in the foreword of Chapter 3 and in Subchapters 
3.6 – 3.8, 3.13, 3.22, 3.23, 3.26, and 3.33. The requirements for the size, 
construction and location of buildings used in the production are covered in the 
foreword of Chapter 3. Subchapter 3.8 states the requirements for the placement 
of equipment and Subchapter 3.7 for the fl ow of materials and products. Operations, 
which have to be performed in separate or defi ned areas, are specifi ed in Subchapters 
3.6, 3.13, 3.22, 3.23, 3.26, and 3.33. Annex 1 covers the requirements for facilities 
used in aseptic processing and Subchapter 3.6 the requirements for the facilities 
used in the production of penicillin. Correspondences to regulation 211.44 are 
covered in Subchapters 3.3 and 3.16, which state the requirements for lighting. Correspondences 
to regulation 211.46 are covered in Subchapters 3.3 and 3.12, which 
state the requirements for ventilation. The specifi c requirements for the production 
of penicillin are covered in Subchapter 3.6. Correspondences to regulation 211.48 
are covered in Subchapters 3.10 and 3.11 and in Subsections 35 and 44 of Annex 1. 
Subchapter 3.10 states the requirements for the plumbing and Subchapter 3.11 the 
requirements for drains. Section 35 of Annex 1 covers the requirements for water 
treatment plants and distribution systems and Section 44 the requirements for the 
monitoring of water sources and water treatment equipment. More guidance on the 
quality of water is given in the EU guidance document Note for Guidance on Quality 
of Water for Pharmaceutical Use [58] . The EU GMP code does not have correspondence 
to regulation 211.50, which covers the requirements for the handling of 
sewage and other refuse. Correspondences to regulation 211.52 are covered in Subchapter 
3.31, which covers the requirements for the facilities for washing and toilet 
purposes. Correspondences to regulation 211.56 are covered in Subchapters 3.2, 3.4, 
3.43, and 4.26. Subchapters 3.2 and 3.4 cover the requirements for the conditions to 
be maintained in the manufacturing facilities. Subchapter 4.26 covers the procedures 
for cleaning and sanitization and Subchapter 3.43 the requirements for the sanitization 
of water pipes. The EU GMP code does not cover any separate requirements 
for the handling of organic waste. Correspondences to regulation 211.58 are covered 
in Subsection 3.2, which covers the requirements for the maintenance of the buildings 
used in the production. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to buildings and facilities are mainly covered in Chapter 12 (Premises) 
and partly in Chapters 3 (Sanitation and Hygiene), 14 (Materials), and 15 (Documentation). 
Correspondences to regulation 211.42 are covered in Subchapters 12.1, 
12.2, 12.4, 12.5, 12.10, 12.14, 12.17, 12.19, 12.22 – 12.26, and 12.33. The requirements 
for the size, construction, and location of buildings are stated in Subchapters 12.1, 
12.4, and 12.5. Subchapters 12.2 and 12.26 cover the requirements for the placement 
of equipment and Subchapters 12.10 and 12.25 the requirements for the fl ow of 
CORRESPONDENCES 141

142 CORRESPONDENCES AND DIFFERENCES 
materials and products. Operations, which have to be performed in separate or 
defi ned areas, are specifi ed in Subchapters 12.14, 12.17, 12.19, 12.22 – 12.24, and 12.33. 
The requirements for the facilities used in aseptic processing are covered in Chapter 
9 of Annex 6 of the WHO TRS 902 [40] and the requirements for the facilities used 
in the manufacture of penicillin are in Subchapter 12.24. Correspondences to regulation 
211.44 are covered in Subchapters 12.8 and 12.32, which state the requirements 
for lighting. Correspondences to regulation 211.46 are covered in Subchapters 12.8 
and 12.30, which state the requirements for ventilation. The specifi c requirements 
for the production of penicillin are covered in Subchapter 12.24. Correspondences 
to regulation 211.48 are covered in Subchapters 12.28, 12.29, and 14.6 and in Annex 
3 of the WHO TRS 929 [45] . Subchapter 12.28 states the requirements for the 
plumbing and Subchapter 14.6 for the quality of water used in the production of 
drug products. More guidance on the quality of water is given in Annex 3 of the 
WHO TRS 929 [45] . The requirements for the drains are stated in Subchapter 12.29. 
Correspondences to regulation 211.50 are covered in Subchapters 14.44 and 14.45, 
which state the requirements for the handling of sewage and other refuse. Correspondences 
to regulation 211.52 are covered in Subchapter 12.12, which states the 
requirements for the facilities for washing and toilet purposes. Correspondences to 
regulation 211.56 are covered in Subchapters 3.1, 12.7, 12.9, 14.44 – 14.46, and 15.48. 
Subchapters 12.7 and 12.9 state the requirements for the conditions to be maintained 
in the manufacturing facilities and Subchapter 3.1 the general requirements 
for sanitation and hygiene. In the WHO GMP guideline there is no separate guidance 
on the handling of organic waste. General requirements for the handling of 
waste materials are stated in Subchapters 14.44 and 14.45. Subchapter 15.48 states 
the requirements for the written procedures for sanitation operations and Subchapter 
14.46 for the use of rodenticides, insecticides, fumigating agents, and sanitizing 
materials. Correspondences to regulation 211.58 are covered in Subchapter 12.6, 
which states the requirements for the maintenance of the buildings used in drug 
production. 
2.1.4.4 Equipment 
For GMP regulations in the United States issues related to equipment are covered 
in Subpart D [7] , which consists of regulations 211.63, 211.65, 211.67, 211.68, and 
211.72. Contents of Subpart D are presented in Table 19 . Regulation 211.63 states 
the requirements for the production equipment covering design, size, and location. 
Regulation 211.65 states the requirements for the construction of equipment cover- 
TABLE 19 Contents of Subpart D of Part 211 of U. S . 
GMP Regulations Covering Equipment [7] 
Section Subject 
CFR 211.63 Equipment design, size, and location 
CFR 211.65 Equipment construction 
CFR 211.67 Equipment cleaning and maintenance 
CFR 211.68 Automatic, mechanical, and electronic 
equipment 
CFR 211.72 Filters 

ing the characteristics of used materials and special requirements for the structure 
of the equipment. Regulation 211.67 deals with cleaning, maintenance, and sanitizing 
of equipment and utensils covering the requirements for the procedures for 
cleaning and maintenance operations. Regulation 211.68 deals with automatic, 
mechanical, and electronic equipment covering requirements for their calibration 
and inspection, including the requirements for the documentation of checks and 
inspections. Furthermore, it covers the requirements for the controls for computer 
or related systems, including the requirements for the maintenance of backup data. 
Regulation 211.72 covers the requirements for the fi lters for liquid fi ltration used 
in the manufacture of injectable products, including the specifi c requirements for 
the use of fi ber - releasing and asbestos - containing fi lters. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to equipment are mainly covered in the interpretation of regulation C.02.005 
(Equipment) and partly in the interpretation of regulation C.02.007 (Sanitation) 
and C.02.024 (Records). Correspondences to regulation 211.63 stating the requirements 
for the design, construction, and location of equipment used in the manufacture 
of drug products are covered in Sections 1 and 5 of the interpretation of 
regulation C.02.005. Correspondences to regulation 211.65 stating the requirements 
for the construction of equipment are covered in Sections 2.1 – 2.3 of the interpretation 
of regulation C.02.005. Correspondences to regulation 211.67 stating the 
requirements for the sanitation are covered in Sections 1, 2, and 3 of the interpretation 
of regulation C.02.007. Correspondences to regulation 211.68 are covered in 
Section 5.4 of the interpretation of regulation C.02.005 and in the foreword of the 
interpretation of regulation C.02.024. Section 5.4 of the interpretation of regulation 
C.02.005 states the requirements for the use of automatic, mechanical, and electronic 
equipment, including computerized systems, and the foreword of the interpretation 
of regulation C.02.024 the requirements for the maintenance of backup data. The 
Canadian GMP code does not have correspondence to regulation 211.72, which 
states the requirements for the fi lters for liquid fi ltration used in the manufacture 
of injectable products. Nor does it cover requirements for the use of fi ber - releasing 
or asbestos - containing fi lters. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
equipment are mainly covered in Chapter 3 (Premises and Equipment) and partly 
in Chapter 4 (Documentation) and Annexes 1 (Manufacture of Sterile Medicinal 
Products) and 11 (Computerised Systems). Correspondences to regulation 211.63 
are covered in Subchapter 3.34, which states the requirements for the design and 
location of equipment used in the manufacture of drug products. Correspondences 
to regulation 211.65 are covered in Subchapters 3.38 and 3.39, which state the 
requirements for the construction of equipment. Correspondences to regulation 
211.67 are covered in Subchapters 3.36, 3.37, and 3.43, which cover the requirements 
for cleaning and sanitizing the manufacturing equipment. Correspondences to regulation 
211.68 are covered in Subchapters 3.41 and 4.9 and Annex 11. Subchapter 
3.41 states the requirements for the maintenance of measuring, weighing, recording, 
and control equipment and Subchapter 4.9 the requirements for the use of electronic 
data processing systems and the maintenance of backup data. Additional 
guidance on the use of computerized systems is given in Annex 11. Correspondences 
CORRESPONDENCES 143

144 CORRESPONDENCES AND DIFFERENCES 
to regulation 211.72 stating the requirements for fi lters for liquid fi ltration used in 
the sterile fi ltration are covered in Sections 84 – 87 of Annex 1. The EU GMP code 
does not have any separate guidance for the use of fi ber - releasing or asbestos - 
containing fi lters. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to equipment are mainly covered in Chapter 13 (Equipment) and 
partly in Chapters 14 (Materials), 15 (Documentation), and 16 (Good Practices in 
Production). Correspondences to regulation 211.63 are covered in Subchapters 13.1 
and 13.2, which state the requirements for the design, location, and installation of 
equipment used in the manufacture of drug products. Correspondences to regulation 
211.65 are covered in Subchapters 13.9 and 14.3, which state the requirements 
for the construction of equipment. Correspondences to regulation 211.67 are covered 
in Subchapters 13.6, 13.8, 13.12, 16.17, 16.18, and 16.22, which state the requirements 
for cleaning and sanitizing the equipment. Correspondences to regulation 211.68 
are covered in Subchapters 16.23 and 15.9. The requirements for the maintenance 
of measuring, weighing, recording, and control equipment and instruments 
are covered in Subchapter 16.23. Subchapter 15.9 states the requirements for the 
use of electronic data - processing systems, including the requirements for the maintenance 
of backup data. Correspondences to regulation 211.72 stating the requirements 
for the use of fi lters are covered in Subchapters 7.6 – 7.9 of Annex 6 of the 
WHO TRS 902 [40] . Subchapter 7.6 covers the requirements for asbestos - containing 
fi lters. 
2.1.4.5 Control of Components and Drug Product Containers and Closures 
In the United States issues related to control of components and drug product 
containers and closures are covered in Subpart E [7] , which consists of regulations 
211.80, 211.82, 211.84, 211.86, 211.87, 211.89, and 211.94. Contents of Subpart E are 
presented in Table 20 . Regulation 211.80 defi nes the requirements for the procedures 
for the control of components, containers, and closures. It also states the 
requirements for their handling, storing, and identifi cation. Regulation 211.82 covers 
the requirements for receipt and storage of untested components, containers, and 
TABLE 20 Contents of Subpart E of Part 211 of U . S . GMP Regulations Covering 
Control of Components and Drug Product Containers and Closures [7] 
Section Subject 
CFR 211.80 General requirements 
CFR 211.82 Receipt and storage of untested components, drug product containers, 
and closures 
CFR 211.84 Testing and approval or rejection of components, drug product 
containers, and closures 
CFR 211.86 Use of approved components, drug product containers, and closures 
CFR 211.87 Retesting of approved components, drug product containers, and 
closures 
CFR 211.89 Rejected components, drug product containers, and closures 
CFR 211.94 Drug product containers and closures 

closures. Regulation 211.84 deals with testing and approval or rejection of components, 
containers, and closures covering the requirements for sampling, testing, and 
release. Regulation 211.86 deals with the use of approved components, containers, 
and closures stating the requirements for the rotation of the storage. Regulation 
211.87 states the requirements for the retesting of approved components, containers, 
and closures. Regulation 211.89 covers the requirements for the handling of rejected 
components, containers, and closures. Regulation 211.94 deals with drug product 
containers and closures covering the requirements for materials and the cleanliness 
of containers and closures. Furthermore, it states the requirements for container 
closure systems, standards and methods. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to control of components and drug product containers and closures are 
covered in interpretations of regulations C.02.009 (Raw Material Testing), C.02.010 
(Raw Material Testing), C.02.011 (Manufacturing Control), C.02.014 (Quality 
Control Department), C.02.016 (Packaging Material Testing), and C.02.017 (Packaging 
Material Testing). Correspondences to regulation 211.80 stating the general 
requirements for the handling, storing, and identifi cation of components (raw materials) 
and drug product containers and closures (packaging materials) are covered 
in Sections 1, 20, and 21 of the interpretation of regulation C.02.011. Correspondences 
to regulation 211.82 stating the requirements for the receipt, testing, and 
storage of untested components and drug product containers and closures are 
covered in Sections 16, 18, and 19 of the interpretation of regulation C.02.011. Correspondences 
to regulation 211.84 stating the requirements for testing and approval 
of components and drug product containers and closures are covered in Sections 6 
and 7 of the interpretation of regulation C.02.009, Sections 1 – 8 of the interpretation 
of regulation C.02.010, Sections 1 and 2 of regulation C.02.016, Section 4 of its 
interpretation, and Section 1 of the interpretation of regulation C.02.017. Interpretations 
6 and 7 of regulation C.02.009 and interpretations 1 – 8 of regulation C.02.010 
cover the requirements for components. Sections 1 and 2 and interpretation 4 of 
regulation C.02.016 and interpretation 1 of regulation C.02.017 state the requirements 
for drug product containers and closures. In the Canadian GMP code there 
is no correspondence to regulation 211.86, which covers the requirements for the 
rotation of the storage. Correspondences to regulation 211.87 stating the requirements 
for the retesting of approved components are covered in Sections 8 – 10 of the 
interpretation of regulation C.02.009. For the retesting of drug product containers 
and closures the Canadian GMP code has no guidance. Correspondences to regulation 
211.89 stating the requirements for the handling of rejected components and 
drug product containers and closures are covered in Section 14 of the interpretation 
of regulation C.02.011 and in Section 5 of the interpretation of regulation C.02.014. 
The Canadian GMP code does not have correspondence to regulation 211.94, which 
covers the requirements for containers and closure systems. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
control of components and drug product containers and closures are mainly covered 
in Chapter 5 (Production) and partly in Chapter 6 (Quality Control). Correspondences 
to regulation 211.80 are covered in Subchapters 5.2, 5.7, 5.10, 5.29, and 
5.40 – 5.42, which cover the requirements for the handling, storing, and identifi cation 
CORRESPONDENCES 145

146 CORRESPONDENCES AND DIFFERENCES 
of components (starting materials) and drug product containers and closures 
(primary packaging materials). Correspondences to regulation 211.82 are covered 
in Subchapters 5.5, 5.27, and 5.40, which state the requirements for the receipt, 
testing, and storage of untested components and drug product containers and closures. 
Correspondences to regulation 211.84 are covered in Subchapters 5.31, 5.40 
and 6.11 – 6.22 and Annex 8. The general requirements for sampling and testing are 
covered in Subchapters 6.11 – 6.22. More guidance on sampling is given in Annex 8. 
The requirements for the approved use of components and drug product containers 
and closures are stated in Subchapters 5.31 and 5.40. Correspondences to regulation 
211.86 are covered in Subchapter 5.7, which states the requirements for the storage 
conditions and rotation. Correspondences to regulation 211.87 are covered in 
Subchapters 5.29 and 5.40, which deal with the retesting of components and drug 
product containers and closures. Correspondences to regulation 211.89 are covered 
in Subchapter 5.61, which states the requirements for the handling of rejected components 
and drug product containers and closures. Correspondences to regulation 
211.94 are covered in Subchapter 5.48, which states the requirements for drug 
product containers and closures. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to control of components and drug product containers and closures 
are covered in Chapters 14 (Materials), 16 (Good Practices in Production), and 17 
(Good Practices in Quality Control). Correspondences to regulation 211.80 are 
covered in Subchapters, 14.5, 14.13, 14.14, 14.19 – 14.21, and 16.2, which state the 
requirements for the handling, storing, and identifi cation of components (starting 
materials) and drug product containers and closures (primary packaging materials). 
Correspondences to regulation 211.82 are covered in Subchapters 14.4, 14.9 – 14.11, 
and 14.19, which state the requirements for receipt, testing, identifi cation, and 
storage of untested components and drug product containers and closures. Correspondences 
to regulation 211.84 are covered in Subchapters 14.12, 14.15, and 17.7 – 
17.17. The requirements for sampling and testing are covered in Subchapters 
17.7 – 17.17 and 14.12. Subchapter 14.15 states the requirements for the approved 
use of components and drug product containers and closures. Correspondences to 
regulation 211.86 are covered in Subchapter 14.5, which states the requirements for 
the storage conditions and the rotation of the storage. Correspondences to regulation 
211.87 are covered in Subchapter 14.13, which states the requirements for the 
retesting of approved components. The WHO GMP guideline does not cover the 
requirements for the retesting of drug product containers and closures. Correspondences 
to regulation 211.89 are covered in Subchapter 14.28, which states the 
requirements for the handling of rejected components and drug product containers 
and closures. Correspondences to regulation 211.94 are covered in Subchapter 16.19, 
which states the requirements for the drug product containers and closures. 
2.1.4.6 Production and Process Controls 
In the United States GMP regulations on issues related to production and process 
controls are covered in Subpart F [7] , which consists of regulations 211.100, 211.101, 
211.103, 211.105, 211.110, 211.111, 211.113, and 211.115. Contents of Subpart F are 
presented in Table 21 . Regulation 211.100 states the requirements for procedures 

regarding production and process controls, including the requirements for the documentation 
and handling of deviations. Regulation 211.101 deals with the requirements 
for the charge - in of components. Regulation 211.103 states the requirements 
for the determination of yields. Regulation 211.105 covers requirements for the 
identifi cation of processing equipment such as containers, processing lines, and 
major equipment used during manufacture. Regulation 211.110 states the requirements 
for in - process controls, including the testing and approval of in - process 
materials and handling of rejected in - process materials. Regulation 211.111 covers 
the requirements for the time limitations on production, including the handling of 
deviations from established limits. Regulation 211.113 covers the control of microbiological 
contaminations. Regulation 211.115 states the requirements for the reprocessing 
of batches that do not conform to standards or specifi cations. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to production and process controls are mainly covered in the interpretation 
of regulation C.02.011 (Manufacturing Control) and partly in the interpretations of 
regulations C.02.005 (Equipment), C.02.014 (Quality Control Department), and 
C.02.029 (Sterile Products). Correspondences to regulation 211.100 are covered in 
Sections 1 – 5 of the interpretation of regulation C.02.011. Interpretations 1 – 4 state 
the requirements for manufacturing processes and interpretation 5 for the handling 
of deviations. Correspondences to regulation 211.101 stating the requirements for 
charge - in of components are covered in Section 22 of the interpretation of regulation 
C.02.011. Correspondences to regulation 211.103 stating the requirements for 
the determination of yields including the handling of deviations from the expected 
yield are covered in Sections 6 and 7 of the interpretation of regulation C.02.011. 
Correspondences to regulation 211.105 stating the requirements for the identifi cation 
of piping, containers, equipment, and rooms used in the manufacturing of drug 
products are covered in Section 3.5 of the interpretation of regulation C.02.005 and 
in Section 13 of the interpretation of regulation C.02.011. Correspondences to regulation 
211.110 are covered in Sections 11 and 14 of the interpretation of regulation 
C.02.011 and in Section 5 of the interpretation of regulation C.02.014. Section 11 of 
the interpretation of regulation C.02.011 states the requirements for the in - process 
controls. The Canadian GMP code does not cover requirements for the testing of 
in - process materials. The handling of rejected materials is covered in Section 14 
of the interpretation of regulation C.02.011 and in Section 5 of the interpretation 
TABLE 21 Contents of Subpart F of Part 211 of U . S . GMP Regulations Covering 
Production and Process Controls [7] 
Section Subject 
CFR 211.100 Written procedures, deviations 
CFR 211.101 Charge - in of components 
CFR 211.103 Calculation of yield 
CFR 211.105 Equipment identifi cation 
CFR 211.110 Sampling and testing of in - process materials and drug products 
CFR 211.111 Time limitations on production 
CFR 211.113 Control of microbiological contamination 
CFR 211.115 Reprocessing 
CORRESPONDENCES 147

148 CORRESPONDENCES AND DIFFERENCES 
of regulation C.02.014. Correspondences to regulation 211.111 dealing with the 
requirements for the time limitations on production are covered in Section 24.7 of 
the interpretation of regulation C.02.011. Correspondences to regulation 211.113 
are covered in the interpretation of regulation C.02.029, which deals with the manufacture 
of sterile products. Correspondences to regulation 211.115 stating the 
requirements for the reprocessing of batches that do not conform to specifi cations 
are covered in Sections 7 – 9 of the interpretation of regulation C.02.014. 
Correspondences in EU GMP In the EU GMP code [15] issues related to production 
and process controls are mainly covered in Chapter 5 (Production) and partly 
in Chapters 3 (Premises and Equipment), 4 (Documentation), and 6 (Quality 
Control). Correspondences to regulation 211.100 are covered in Subchapters 5.2, 
5.15, and 5.22 – 5.24. The requirements for manufacturing processes are covered in 
Subchapters 5.2 and 5.22 – 5.24. Subchapter 5.15 states the requirements for handling 
of deviations from instructions or procedures. Correspondences to regulation 
211.101 are covered in Subchapters 5.28 – 5.34, which state the requirements for the 
charge - in of components. Correspondences to regulation 211.103 are covered in 
Subchapters 5.8 and 5.39, which state the requirements for determination of yields, 
including the handling of deviations from the expected yield. Correspondences to 
regulation 211.105 are covered in Subchapters 3.42 and 5.12, which state the requirements 
for identifi cation of piping, containers, equipment, and rooms used in the 
manufacture of drug products. Correspondences to regulation 211.110 are covered 
in Subchapters 3.17, 4.10, 4.12, 5.38, 5.61, and 6.18. The requirements for in - process 
controls are covered in Subchapters 3.17, 5.38, and 6.18. Subchapters 4.10 and 4.12 
state the requirements for the specifi cations for in - process materials (intermediate 
products) and Subchapter 5.61 for handling of rejected materials. The EU GMP 
code does not cover separate guidance on testing and approval of in - process materials. 
General guidance on sampling and testing is given in Subchapters 6.11 – 6.22. 
Correspondences to regulation 211.111, which deals with the time limitations on 
production, are covered in Chapter 4.15. Correspondences to regulation 211.113 are 
covered in Subchapter 5.10 and in Annex 1, which cover the requirements for the 
control of microbiological contaminations. Correspondences to regulation 211.115 
are covered in Subchapters 5.62 and 5.64, which state the requirements for the 
reprocessing of rejected batches. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to production and process controls are covered in Chapters 13 (Equipment), 
14 (Materials), 15 (Documentation), 16 (Good Practices in Production), and 
17 (Good Practices in Quality Control). Correspondences to regulation 211.100 are 
covered in Subchapters 16.1 – 16.3. Subchapters 16.1 and 16.2 state the requirements 
for the manufacturing operations and Subchapter 16.3 for the handling of deviations 
from instructions or procedures. Correspondences to regulation 211.101 are covered 
in Subchapters 14.12 – 14.18, which state the requirements for the charge - in of components. 
Correspondences to regulation 211.103 are covered in Subchapters 16.4 
and 16.20, which state the requirements for the determination of yields, including 
the handling of deviations from the expected yield. Correspondences to regulation 
211.105 are covered in Subchapters 13.3, 13.4, and 16.6, which state the requirements 
for the identifi cation of piping, containers, equipment, and rooms used during pro

duction. Correspondences to regulation 211.110 are covered in Subchapters 14.28, 
15.20, 16.9, 16.16, and 17.8. Subchapters 16.9, 16.16, and 17.8 cover the requirements 
for the in - process controls and Subchapter 15.20 for the specifi cations for in - process 
materials (intermediate products). The requirements for the handling of rejected 
materials are stated in Subchapter 14.28. Correspondences to regulation 211.111, 
which deals with the time limitations on production, are covered in Chapter 15.23. 
Correspondences to regulation 211.113 are covered in Subchapters 16.10 – 16.14 and 
in Annex 6 of the WHO TRS 902 [40] . Subchapters 16.10 – 16.14 cover general 
requirements for the prevention of cross - contamination and bacterial contamination 
during production and Annex 6 general requirements for the manufacture of 
sterile drug products. Correspondences to regulation 211.115 are covered in Subchapters 
14.29, 14.31, and 15.40, which state the requirements for the reprocessing 
of rejected batches. 
2.1.4.7 Packaging and Labeling Control 
For GMP regulations in the United States issues related to packaging and labeling 
control are covered in Subpart G [7] , which consists of regulations 211.122, 211.125, 
211.130, 211.132, 211.134, and 211.137. The contents of Subpart G is presented in 
Table 22 . Regulation 211.122 deals with materials examination and usage criteria 
covering the requirements for the receipt, identifi cation, storage, handling, sampling, 
testing, and approval of labeling and packaging materials, including documentation. 
Furthermore, it covers the requirements for the control of labeling, handling of 
obsolete and outdated labeling and packaging materials, and special requirements 
for different labeling methods. Regulation 211.125 states the requirements for the 
labeling issuance covering the testing of labeling materials, the control of discrepancy 
between the quantities of labeling issued, used, and returned, and the handling 
of excess and returned labeling. Regulation 211.130 states the requirements for the 
packaging and labeling operations covering the written procedures. Regulation 
211.132 states the requirements for the tamper - evident packaging. Regulation 
211.134 states the requirements for the inspections of packaged and labeled 
products covering sampling, examination, and documentation. Regulation 211.137 
states the requirements for the expiration dates, including exemptions from the 
requirements. 
TABLE 22 Contents of Subpart G of Part 211 of U . S . GMP Regulations Covering 
Packaging and Labeling Control [7] 
Section Subject 
CFR 211.122 Materials examination and usage criteria 
CFR 211.125 Labeling issuance 
CFR 211.130 Packaging and labeling operations 
CFR 211.132 Tamper - evident packaging requirements for over - the - counter (OTC) 
human drug products 
CFR 211.134 Drug product inspection 
CFR 211.137 Expiration dating 
CORRESPONDENCES 149

150 CORRESPONDENCES AND DIFFERENCES 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to packaging and labeling control are covered in the interpretations of regulations 
C.02.011 (Manufacturing Control), C.02.017 (Packaging Material Testing), 
C.02.016 (Packaging Material Testing), C.02.019 (Finished Product Testing), and 
C.02.027 (Stability). Correspondences to regulation 211.122 are covered in Sections 
1, 16, 40, and 43 – 48 of the interpretation of regulation C.02.011, Sections 1, 8, and 9 
of the interpretation of regulation C.02.017, and Sections 1 and 4 – 7 of the interpretation 
of regulation C.02.016. Sections 1, 16, 43, and 48 of the interpretation of regulation 
C.02.011 state the general requirements for the handling of packaging and 
labeling materials covering receipt and storage. Section 8 of the interpretation of 
regulation C.02.017 states the requirements for the identifi cation of the packaging 
and labeling materials. The requirements for the testing of the packaging and labeling 
materials are covered in Sections 1 and 9 of the interpretation of regulation 
C.02.017. Sections 1 and 4 of the interpretation of regulation C.02.016 and Sections 
6 and 7 of the interpretation of regulation C.02.017 state the requirements for the 
approval of packaging and labeling materials. Sections 44 – 47 of the interpretation 
of regulation C.02.011 cover requirements for the use of roll - fed labels, cut labels, 
gang printing, and the monitoring of the performance of printing. The requirements 
for the handling of obsolete and outdated packaging and labeling materials are 
covered in Section 40 of the interpretation of regulation C.02.011 and in Section 5 
of the interpretation of regulation C.02.016. Correspondences to regulation 211.125 
are covered in Sections 39 and 42 of the interpretation of regulation C.02.011 
and in Section 8 of the interpretation of regulation C.02.017. Section 8 of the interpretation 
of regulation C.02.017 states the requirements for the examination of 
packaging and labeling materials. The requirements for the control and handling of 
discrepancy between the quantities of labeling issued, used, and returned are covered 
in Section 42 and the requirements for the handling of unused batch - coded packaging 
and labeling materials in Section 39 of the interpretation of regulation C.02.011. 
Correspondences to regulation 211.130 stating the requirements for the packaging 
and labeling operations are covered in Sections 29 – 38 of the interpretation of regulation 
C.02.011. In Canadian GMP code there is no correspondence to regulation 
211.132 stating the requirements for the tamper - evident packaging. Correspondences 
to regulation 211.134 stating the requirements for the inspections of packaged 
and labeled products are covered in Section 1 of the interpretation of regulation 
C.02.019. Correspondences to regulation 211.137 stating the requirements for the 
expiration dates are covered in regulation C.02.027 and in Section 1 of its 
interpretation. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
packaging and labeling control are mainly covered in Chapter 5 (Production) and 
partly in Chapters 4 (Documentation) and 6 (Quality Control). Correspondences 
to regulation 211.122 are covered in Subchapters 4.11, 4.19, 4.21 – 4.23, 5.2, 5.40 – 5.43, 
and 5.50 – 5.52. Subchapters 4.19, 4.21 – 4.23, 5.2, and 5.40 – 5.42 cover the requirements 
for purchase, handling, control, storage, and identifi cation of packaging and labeling 
materials. Specifi cations for packaging and labeling materials are stated in Subchapter 
4.11 and the requirements for handling of outdated or obsolete packaging and 
labeling materials in Subchapter 5.43. Subchapter 5.51 covers the requirements for 
the use of cut - labels, off - line overprinting, and roll - feed labels. The requirements for 

the control of the printing and labeling operations are stated in Subchapters 5.50 
and 5.52. Correspondences to regulation 211.125 are covered in Subchapters 5.2, 
5.56, and 5.57. Subchapter 5.2 states the general requirements for the handling of 
packaging and labeling materials. The requirements for the control of discrepancy 
between the quantities of labeling issued, used, and returned are covered in Subchapter 
5.56 and the requirements for the handling of unused batch - coded packaging 
and labeling materials in Subchapter 5.57. Correspondences to regulation 211.130 
are covered in Subchapters 5.2 and 5.44 – 5.49. Subchapter 5.2 states the general 
requirements for the handling of packaging and labeling materials and Subchapters 
5.44 – 5.49 cover the requirements for the packaging and labeling operations. The 
European Community GMP code does not have correspondence to regulation 
211.132, which covers the requirements for the tamper - evident packaging. Correspondences 
to regulation 211.134 are covered in Subchapters 5.54 and 6.3, which 
state the requirements for the control of packaged and labeled products. The EU 
GMP code does not have correspondence to regulation 211.137, which covers the 
requirements for expiration dates. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to packaging and labeling control are covered in Chapters 6 (Product 
Recalls), 12 (Premises), 14 (Materials), 15 (Documentation), 16 (Good Practices in 
Production), and 17 (Good Practices in Quality Control). Correspondences to regulation 
211.122 are covered in Subchapters 12.21, 14.19 – 14.23, 15.18, 16.2, 17.14, and 
17.16. Subchapters 6.2, 12.21, 14.19 – 14.21, 14.23, and 17.16 state the requirements 
for the purchase, handling, control, storage, and identifi cation of packaging and 
labeling materials. The requirements for the approval of packaging and labeling 
materials are covered in Subchapters 17.14 and 15.18. Subchapter 14.20 states the 
requirements for the use of roll - feed and cut labels and Subchapter 14.22 the 
requirements for the handling of outdated and obsolete packaging and labeling 
materials. Correspondences to regulation 211.125 are covered in Subchapters 16.2, 
16.34, and 16.35. Subchapter 16.2 states the general requirements for the handling 
of packaging and labeling materials and Subchapter 16.34 the requirements for the 
handling of discrepancy between the quantities of labeling issued, used, and returned. 
The requirements for the handling of unused batch - coded packaging and labeling 
materials are covered in Subchapter 16.35. Correspondences to regulation 211.130 
are covered in Subchapters 16.25 – 16.30, which state the requirements for the packaging 
and labeling operations. The WHO GMP guideline does not cover correspondence 
to regulation 211.132, which covers the requirements for tamper - evident 
packaging. Correspondences to regulation 211.134 are covered in Subchapter 16.32, 
which states the requirements for the control of packaged and labeled products. 
Correspondences to regulation 211.137 are covered in Subchapter 17.24, which 
states the requirements for the determination of expiration dates and shelf - life 
specifi cations. 
2.1.4.8 Holding and Distribution 
In the United States GMP regulations on issues related to holding and distribution 
are covered in Subpart H [7] , which consists of regulations 211.142 and 211.150. The 
contents of Subpart H is presented in Table 23 . Regulation 211.142 states the 
CORRESPONDENCES 151

152 CORRESPONDENCES AND DIFFERENCES 
requirements for the warehousing procedures covering quarantine and storage and 
regulation 211.150 for the distribution procedures covering distribution order and 
recalls. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to holding and distribution are covered in the interpretations of regulations 
C.02.004 (Premises), C.02.011 (Manufacturing Control), C.02.012 (Manufacturing 
Control), and C.02.019 (Finished Product Testing). Correspondences to regulation 
211.142 stating the requirements for the quarantine and storage of products are 
covered in Sections 1 and 49 of the interpretation of regulation C.02.011, Section 
11.4 of the interpretation of regulation C.02.004, and Section 2 of the interpretation 
of regulation C.02.019. Correspondences to regulation 211.150 stating the requirements 
for distribution and recalls are covered in Section 1 of the interpretation of 
regulation C.02.011 and Section 1 of the interpretation of regulation C.02.012. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
holding and distribution are covered in Chapters 4 (Documentation), 5 (Production), 
and 8 (Complaints and Product Recall). Correspondences to regulation 
211.142 are covered in Subchapters 5.2, 5.58, and 5.60, which state the requirements 
for the storage and quarantine of products. Correspondences to regulation 211.150 
are covered in Subchapters 4.25, 5.2, and 8.8 – 8.15, which state the requirements for 
distribution and recalls. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to holding and distribution are covered in Chapters 6 (Product 
Recalls), 14 (Materials), 15 (Documentation), and 16 (Good Practices in Production). 
Correspondences to regulation 211.142 are covered in Subchapters 14.4, 14.26, 
and 16.2, which state the requirements for the storage and quarantine of products. 
Correspondences to regulation 211.150 are covered in Subchapters 6.1 – 6.8, 15.45, 
and 16.2, which state the requirements for distribution and recalls. 
2.1.4.9 Laboratory Controls 
In the United States GMP regulations [7] issues related to laboratory controls are 
covered in Subpart I, which consists of regulations 211.160, 211.165, 211.166, 211.167, 
211.170, 211.173, and 211.176. The contents of Subpart I is presented in Table 24 . 
Regulation 211.160 states the requirements for the establishment of laboratory 
controls such as specifi cations, standards, sampling plans, and test procedures. 
Furthermore, it covers the requirements stated for the calibration of instruments, 
apparatus, gauges, and recording devices. Regulation 211.165 states the require- 
TABLE 23 Contents of Subpart H of Part 211 of U . S . 
GMP Regulations Covering Holding and Distribution [7] 
Section Subject 
CFR 211.142 Warehousing procedures 
CFR 211.150 Distribution procedures 

ments for the laboratory testing of batches prior to release covering the requirements 
for sampling, testing, and approval. Furthermore, it states the requirements 
for the handling of rejected drug products. Regulation 211.166 states the requirements 
for stability testing, including the requirements for the determination of 
expiration dates and the requirements for stability testing of homeopathic drug 
products. Regulation 211.167 deals with special testing requirements covering sterile 
products, ophthalmic ointments, and controlled - release dosage forms. Regulation 
211.170 states the requirements for reserve samples covering identifi cation, quantity, 
retention time, and storage. Furthermore it covers the requirements for the deterioration 
investigations. Regulation 211.173 deals with laboratory animals covering the 
requirements for their maintenance and control. Regulation 211.176 states the 
requirements for the testing of penicillin contamination and the handling of penicillin 
contaminated drug product. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to laboratory controls are covered in the interpretations of regulations 
C.02.004 (Premises), C.02.009 (Raw Material Testing), C.02.011 (Manufacturing 
Control), C.02.014 (Quality Control Department), C.02.015 (Quality Control 
Department), C.02.016 (Packaging Material Testing), C.02.017 (Packaging Material 
Testing), C.02.018 (Finished Product Testing), C.02.025 (Samples), C.02.026 
(Samples), C.02.027 (Stability), and C.02.028 (Stability). Correspondences to regulation 
211.160 stating the general requirements for laboratory controls are covered in 
regulation C.02.009 and Sections 1 – 3 and 5 – 6 of its interpretation, regulation 
C.02.016 and Sections 1 – 3 of its interpretation, Section 1 of the interpretation of 
regulation C.02.017, regulation C.02.018 and Sections 1 – 5 of its interpretation, and 
Section 6.4 of the interpretation of regulation C.02.015. Correspondences to regulation 
211.165 stating the requirements for the release for distribution including the 
testing of fi nished drug products and the handling of rejected drug products are 
covered in Sections 7 and 14 of the interpretation of regulation C.02.011, Sections 
2 and 5 of the interpretation of regulation C.02.014, Section 3 of the interpretation 
of regulation C.02.015, and Section 2 of regulation C.02.018 and Sections 1 and 4 of 
its interpretation. Correspondences to regulation 211.166 stating the requirements 
for stability testing are covered in Section 1 of the interpretation of regulation 
C.02.027 and Sections 1 and 2 of the interpretation of regulation C.02.028. The 
Canadian GMP code does not cover separate requirements for the stability testing 
TABLE 24 Contents of Subpart I of Part 211 of U . S . GMP 
Regulations Covering Laboratory Controls [7] 
Section Subject 
CFR 211.160 General requirements 
CFR 211.165 Testing and release for distribution 
CFR 211.166 Stability testing 
CFR 211.167 Special testing requirements 
CFR 211.170 Reserve samples 
CFR 211.173 Laboratory animals 
CFR 211.176 Penicillin contamination 
CORRESPONDENCES 153

154 CORRESPONDENCES AND DIFFERENCES 
of homeopathic drug products. Correspondences to regulation 211.167 stating the 
requirements for sterility testing are covered in Sections 1 – 4 of the interpretation 
of regulation C.02.029 (Sterile Products). The Canadian GMP code does not have 
any guidance covering the testing of ophthalmic ointments and controlled - release 
dosage forms. Correspondences to regulation 211.170 stating the requirements for 
reserve samples are covered in Section 1 of regulation C.02.025 and in regulation 
C.02.026 and Sections 1 and 3 – 5 of their interpretation. Correspondences to regulation 
211.173 stating the requirements for laboratory animals are covered in Section 
2.4 of the interpretation of regulation C.02.004. The Canadian GMP code does not 
have correspondence to regulation 211.176, which covers the requirements for the 
testing and handling of penicillin contamination. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
laboratory controls are covered in Chapters 1 (Quality Management), 4 (Documentation), 
5 (Production), and 6 (Quality Control) and in Annexes 1 (Manufacture of 
Sterile Medicinal Products), 9 (Manufacture of Liquids, Creams, and Ointments), 
and 19 (Reference and Retention Samples). Correspondences to regulation 211.160 
are covered in Subchapters 1.4, 4.2, 4.3, 4.10 – 4.13, 5.15, 6.7, and 6.18, which cover 
the general requirements for laboratory controls. Correspondences to regulation 
211.165 are covered in Subchapters 4.22, 4.23, 5.61, 5.62, 6.3, 6.11, and 6.15, which 
state the requirements for the release for distribution, the testing of fi nished drug 
products, and the handling of rejected drug products. The EU GMP code does not 
have correspondence to regulation 211.166, which states the requirements for stability 
testing. However, there is a separate guideline, Stability Testing on Active Ingredients 
and Finished Products [59] , which provides guidance on issues related to 
stability testing. Furthermore, Subchapters 6.23 – 6.33 cover the requirements for the 
on - going stability program. Correspondences to regulation 211.167 are covered in 
Annexes 1 and 9. Section 93 of Annex 1 covers the requirements for sterility testing 
and Annex 9 the requirements for ointments. In the EU GMP code there is no 
guidance on the testing of the controlled - release dosage forms. Correspondences to 
regulation 211.170 are covered in Subchapters 1.4 and 6.12 and Annex 19, which 
state the requirements for reserve samples. Correspondences to regulation 211.173 
are covered in Subchapters 3.33 and 6.22, which state the requirements for the 
maintenance of animals. The EU GMP code does not have correspondence to regulation 
211.176, which covers the requirements for the testing and handling of penicillin 
contaminations. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to laboratory controls are covered in Chapters 14 (Materials), 15 
(Documentation), 16 (Good Practices in Production), and 17 (Good Practices in 
Quality Control). Correspondences to regulation 211.160 are covered in Subchapters 
15.14 – 15.16, 15.18 – 15.21, 16.3, and 16.23, which state the general requirements 
for laboratory controls. Correspondences to regulation 211.165 are covered in 
Subchapters 14.28, 14.29, 15.13, 15.42, 17.7 – 17.13, 17.19, and 17.20, which state the 
requirements for the release for distribution covering the testing of fi nished drug 
products and the handling of rejected drug products. Correspondences to regulation 
211.166 are covered in Subchapters 17.23 – 17.26, which state the requirements for 
stability testing. The WHO GMP guideline does not cover separate requirements 

for the stability testing of homeopathic drug products. Correspondences to regulation 
211.167 are covered in Annex 6 of the WHO TRS 902 [40] , which states the 
requirements for sterility testing. The WHO GMP guideline does not cover any 
requirements for the testing of ophthalmic ointments or controlled - release dosage 
forms. Correspondences to regulation 211.170 are covered in Subchapter 17.22, 
which states the requirements for reserve samples. The WHO GMP guideline does 
not have correspondence to regulation 211.173, which covers the requirements 
for the maintenance of laboratory animals. Nor does it have correspondence to 
regulation 211.176, which covers the requirements for the testing of penicillin 
contaminations. 
2.1.4.10 Records and Reports 
In the United States issues related to records and reports are covered in Subpart J 
[7] , which consists of regulations 211.180, 211.182, 211.184, 211.186, 211.188, 211.192, 
211.194, 211.196, and 211.198. The contents of Subpart J is presented in Table 25 . 
Regulation 211.180 states the general requirements for documentation covering 
maintenance, retention times, and availability of the records. Furthermore, it states 
the requirements for the annual quality standards evaluation. Regulation 211.182 
states the requirements for individual equipment logs. Regulation 211.184 states the 
requirements for component, drug product container, closure, and labeling records. 
Regulation 211.186 states the requirements for master production and control 
records. Regulation 211.188 states the requirements for batch production and control 
records. Regulation 211.192 states the requirements for the review and approval of 
production and control records, including the requirements for the investigation of 
any unexplained discrepancies. Regulation 211.194 states the requirements for laboratory 
records, including the requirements for the documentation of modifi cations. 
Furthermore, it covers the requirements for the documentation of the testing and 
standardization of reference standards, reagents, and standard solutions; calibration 
of laboratory instruments and recording devices; and stability tests. Regulation 
211.196 states the requirements for the distribution records. Regulation 211.198 
states the requirements for the handling of complaints, including the maintenance 
and retention times of complaint fi les. 
TABLE 25 Contents of Subpart J of Part 211 of U . S . GMP Regulations Covering 
Records and Reports [7] 
Section Subject 
CFR 211.180 General requirements 
CFR 211.182 Equipment cleaning and use log 
CFR 211.184 Component, drug product container, closure, and labeling records 
CFR 211.186 Master production and control records 
CFR 211.188 Batch production and control records 
CFR 211.192 Production record review 
CFR 211.194 Laboratory records 
CFR 211.196 Distribution records 
CFR 211.198 Complaint fi les 
CORRESPONDENCES 155

156 CORRESPONDENCES AND DIFFERENCES 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to records and reports are mainly covered in regulations C.02.021, C.02.022, 
C.02.023, and C.02.024 (Records) and in their interpretations and partly in the 
interpretations of regulations C.02.005 (Equipment), C.02.010 (Raw Material 
Testing), C.02.011 (Manufacturing Control), C.02.012 (Manufacturing Control), 
C.02.014 (Quality Control Department), C.02.015 (Quality Control Department), 
and C.02.017 (Packaging Material Testing). Correspondences to regulation 211.180 
stating the general requirements for the maintenance of records, including periodic 
quality evaluation (self - inspection) and the retention time of the records are covered 
in regulations C.02.021, C.02.022, C.02.023, and C.02.024 and their interpretations 
and in Section 2 of the interpretation of regulation C.02.012. Correspondences 
to regulation 211.182 stating the requirements for individual equipment logs are 
covered in Section 5.5 of the interpretation of regulation C.02.005. Correspondences 
to regulation 211.184 stating the requirements for records to be kept on components, 
drug product containers, closures, and labeling are covered in Sections 4 and 5 of 
the interpretation of regulations C.02.020 – 24, Section 5 of the interpretation of 
regulation C.02.010, and Section 7 of the interpretation of regulation C.02.017. 
Correspondences to regulation 211.186 stating the requirements for the master 
production and control records (manufacturing and packaging master formulas) are 
covered in Sections 23 – 25 of the interpretation of regulation C.02.011 and Section 
1.1 of the interpretation of regulations C.02.020 – 24. Correspondences to regulation 
211.188 stating the requirements for the batch production and control records 
(manufacturing and packaging batch document) are covered in Sections 26, 27, 29, 
and 30 of the interpretation of regulation C.02.011 and Section 1.2 of the interpretation 
of regulations C.02.020 – 24. Correspondences to regulation 211.192 stating the 
requirements for review and approval of production and control records including 
investigation of batch deviations are covered in Section 2 of the interpretation of 
regulation C.02.014. Correspondences to regulation 211.194 stating the requirements 
for laboratory records are covered in Sections 6.4, 6.6, and 6.7 of the interpretation 
of regulation C.02.015. Correspondences to regulation 211.196 stating the 
requirements for distribution records are covered in Section 1.6 of the interpretation 
of regulation C.02.012 and Section 2.1 of the interpretation of regulations 
C.02.020 – 24. Correspondences to regulation 211.198 stating the requirements for 
the maintenance of complaint fi les including retention times are covered in Section 
4 of the interpretation of regulation C.02.015, Section 3.1 of the interpretation of 
regulations C.02.020 – 24, and regulation C.02.023. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
records and reports are mainly covered in Chapter 4 (Documentation) and partly 
in Chapters 1 (Quality Management), 5 (Production), 6 (Quality Control), 8 (Complaints 
and Product Recall), and 9 (Self Inspection). Correspondences to regulation 
211.180 are covered in Subchapters 4.1 – 4.9, 6.8, and 9.1 – 9.3, which state the general 
requirements for the maintenance of the records, including periodic quality evaluation 
(self - inspection) and retention times. Correspondences to regulation 211.182 
are covered in Subchapters 4.28 and 4.29, which state the requirements for individual 
equipment logs. Correspondences to regulation 211.184 are covered in Subchapters 
4.19 and 4.20, which state the requirements for the records to be kept on the 
receipt of components, drug product containers, closures, and labeling. Correspon

dences to regulation 211.186 are covered in Subchapters 4.14 – 4.16, which state the 
requirements for the master production and control records (manufacturing formula, 
processing, and packaging instructions). Correspondences to regulation 211.188 are 
covered in Subchapters 4.17 and 4.18, which state the requirements for the batch 
production and control records (batch processing and packaging record). Correspondences 
to regulation 211.192 are covered in Subchapters 1.4, 4.3, 4.24, 5.8, and 
5.39, which state the requirements for review and approval of production and 
control records, including the investigation of unexplained discrepancies. Correspondences 
to regulation 211.194 are covered in Subchapters 3.41, 6.7, 6.17, 6.20, 
and 6.21, which state the requirements for laboratory records. Correspondences to 
regulation 211.196 are covered in Subchapter 4.25, which states the requirements 
for distribution records. Correspondences to regulation 211.198 are covered in 
Subchapters 4.26 and 8.1 – 8.8, which state the requirements for the handling of 
complaints. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to records and reports are covered in Chapters 5 (Complaints), 
8 (Self - Inspection and Quality Audits), 13 (Equipment), 14 (Materials), 15 
(Documentation), 16 (Good Practices in Production), and 17 (Good Practices in 
Quality Control). Correspondences to regulation 211.180 are covered in Subchapters 
8.1 – 8.6 and 15.1 – 15.9, which state the general requirements for the maintenance 
of the records, including periodic quality evaluation (self - inspection) and retention 
times. Correspondences to regulation 211.182 are covered in Subchapters 15.46 
and 15.47, which state the requirements for individual equipment logs. Correspondences 
to regulation 211.184 are covered in Subchapters 15.32 and 15.33, which state 
the requirements for the records to be kept on the receipt of components, 
drug product containers, closures, and labeling. Correspondences to regulation 
211.186 are covered in Subchapters 15.22 – 15.24, which state the requirements for 
the master production and control records (master formula and packaging instructions). 
Correspondences to regulation 211.188 are covered in Subchapters 15.25 – 
15.30, which state the requirements for the batch production and control records 
(batch processing and packaging records). Correspondences to regulation 211.192 
are covered in Subchapters 16.4, 16.20, and 17.21, which state the requirements for 
review and approval of production and control records covering also the requirements 
for the investigation of unexplained discrepancies. Correspondences to regulation 
211.194 are covered in Subchapters 13.5, 14.34, 14.35, 14.41, 15.12, 15.42, 15.43, 
and 16.23, which states the requirements for laboratory records. Correspondences 
to regulation 211.196 are covered in Subchapter 15.45, which states the requirements 
for the distribution records. Correspondences to regulation 211.198 are 
covered in Subchapters 5.1 – 5.10, which state the requirements for the handling of 
complaints. 
2.1.4.11 Returned and Salvaged Drug Products 
In the United States GMP regulation [7] issues related to returned and salvaged 
drug products are covered in Subpart K, which consists of regulations 211.204 and 
211.208. The contents of Subpart K is presented in Table 26 . Regulation 211.204 
states the requirements for the handling of returned drug products, including repro- 
CORRESPONDENCES 157

158 CORRESPONDENCES AND DIFFERENCES 
cessing and documentation. Regulation 211.208 states the requirements for drug 
product salvaging. 
Correspondences in Canadian GMP Code In the Canadian GMP code [12] issues 
related to returned and salvaged drug products are covered in the interpretation of 
regulation C.02.014 (Quality Control Department). Correspondences to regulation 
211.204 stating the requirements for the handling of returned drug products are 
covered in Section 4 of the interpretation of regulation C.02.014. The Canadian 
GMP code does not have correspondence to regulation 211.208, which covers the 
requirements for drug product salvaging. 
Correspondences in EU GMP Code In the EU GMP code [15] issues related to 
returned and salvaged drug products are covered in Chapters 4 (Documentation) 
and 5 (Production). Correspondences to regulation 211.204 are covered in Subchapters 
4.26 and 5.26, which state the requirements for the handling of returned drug 
products. The EU GMP code does not have correspondence to regulation 211.208, 
which covers the requirements for drug product salvaging. 
Correspondences in WHO GMP Guideline In the WHO GMP guideline [36] 
issues related to returned and salvaged drug products are covered in Chapter 14 
(Materials). Correspondences to regulation 211.204 are covered in Subchapter 
14.33, which states the requirements for the handling of returned drug products. The 
WHO GMP guideline does not have correspondence to regulation 211.208, which 
covers the requirements for drug product salvaging. 
REFERENCES 
1. Immel , B. K. ( 2001 ), A brief history of the GMPs for pharmaceuticals , Pharm. Technol. 
No. Am. , 25 ( 7 ), 44 – 48 . 
2. Anonymous Pharmaceutical Administration and Regulations in Japan , Japan Pharmaceutical 
Manufacturers Association, available: http://www.jpma.or.jp/english/library/pdf/2005. 
pdf . 
3. Vesper , J. L. ( 2003 ), So what are GMPs, anyway? BioProcess Int. , 1 ( 2 ), 24 – 29. 
4. Rosin , L. J. ( 2006 ), Regulatory affairs: If you didn ’ t write it down, it didn ’ t happen , 
BioProcess Int. , 4 ( 3 , Suppl), 16 – 23 . 
5. Anonymous ( 2002 ), Sec. 351: Adulterated drugs and devices, in United States Code , 
Title 21, Chapter 9, Subchapter V, Part A, U.S. Government Printing Offi ce, Washington, 
DC, available: http://frwebgate.access.gpo.gov/cgi - bin/getdoc.cgi?dbname=browse_ 
usc&docid=Cite:+21USC351 . 
TABLE 26 Contents of Subpart K of Part 211 of U . S . 
GMP Regulations Covering Returned and Salvaged Drug 
Products [7] 
Section Subject 
CFR 211.204 Returned drug products 
CFR 211.208 Drug product salvaging 

6. Anonymous ( 2005 ), Part 210: Current good manufacturing practice in manufacturing, 
processing, packing, or holding of drugs: General, in Code of Federal Regulations , Title 
21, Chapter I, U.S. Government Printing Offi ce, Washington, DC, pp. 118 – 119, available: 
http://www.access.gpo.gov/nara/cfr/waisidx_05/21cfr210_05.html . 
7. Anonymous ( 2005 ), Part 211: Current good manufacturing practice for fi nished pharmaceuticals, 
in Code of Federal Regulations , Title 21, Chapter I, U.S. Government 
Printing Offi ce, Washington, DC, pp. 120 – 141, available: http://www.access.gpo.govnara/ 
cfr/waisidx_05/21cfr211_05.html . 
8. Grazal , J. G. , and Earl , D. S. ( 1997 ), EU and FDA regulations: Overview and comparison , 
Qual. Assur. J. , 2 , 55 – 60 . 
9. Anonymous ( 2006 ), Guidance page, available: http://www.fda.gov/cder/guidance/index. 
htm . 
10. Anonymous ( 2001 ), Guidance for industry: Q7A Good manufacturing practice guidance 
for active pharmaceutical ingredients, Offi ce of Training and Communications, available: 
http://www.fda.gov/cber/gdlns/ichactive.pdf . 
11. Anonymous ( 2005 ), Division 2: Good manufacturing practices, in Consolidated 
Statutes and Regulations , Food and Drugs Act, Food and Drug Regulations, Part C, 
Department of Justice, Canada, available: http://laws.justice.gc.ca/en/f - 27/c.r.c. - c.870/ 
230079.html . 
12. Anonymous ( 2002 ), Good manufacturing practices guidelines, Version 2, Health Products 
and Food Branch Inspectorate, available: http://www.hc - sc.gc.ca/dhp - mps/alt_formats/ 
hpfb - dgpsa/pdf/compli - conform/2002v2_e.pdf . 
13. Anonymous ( 2001 ), Directive 2001/83/EC of the European Parliament and of the Council , 
Off. J. Eur. Union , 44 ( L311 ), 67 – 128 , available: http://europa.eu.int/eur - lex/pri/en/oj/ 
dat/2001/l_311/l_31120011128en00670128.pdf . 
14. Anonymous ( 2003 ), Commission directive 2003/94/EC , Off. J. Eur. Union , 
46 ( L262 ), 22 – 26 , available: http://europa.eu.int/eur - lex/pri/en/oj/dat/2003/l_262/ 
l_26220031014en00220026.pdf . 
15. Anonymous ( 2005 ), EU guidelines to good manufacturing practice, in The Rules Governing 
Medicinal Products in the European Union , Vol. 4, European Commission Enterprise 
and Industry Directorate - General, available: http://pharmacos.eudra.org/F2/eudralex/ 
vol - 4/home.htm . 
16. Anonymous ( 1998 ), Quality and biotechnology, in The Rules Governing Medicinal Products 
in the European Union , Vol. 3A, European Commission Directorate General III, 
available: http://pharmacos.eudra.org/F2/eudralex/vol - 3/home.htm . 
17. Anonymous ( 2005 ), Regulations for Buildings and Facilities of Pharmacies , etc., MHLW 
Ministerial Ordinance, Yakuji Nippon Ltd. , Tokyo , No.73. 
18. Kim , Y. - O. , Ha , K. - W. , and Choi , K. - S. ( 2001 ), Safety evaluation for new drug approval in 
Korea , Drug Info. J. , 35 ( 1 ), 285 – 291 . 
19. Shin , S. - G. ( 1998 ), Current status of clinical trials in the Republic of Korea , Drug Info. J. , 
32 (Suppl), 1217S – 1222S . 
20. Anonymous ( 2000 ), Competition in the Pharmaceutical Industry — Republic of Korea , 
Working Party No. 2 on Competition and Regulation, Committee on Competition Law 
and Policy, OECD Publishing , Paris . 
21. Anonymous ( 2001 ), Drug Administration Law of the People ’ s Republic of China , Order 
of the President of the People ’ s Republic of China No. 45, available: http://www.sfda.gov. 
cn/cmsweb/webportal/W45649037/A48335975.html . 
22. Deng , R. , and Kaitin , K. I. ( 2004 ), The regulation and approval of new drugs in China , 
Drug Info. J. , 38 ( 1 ), 29 – 39 . 
REFERENCES 159

160 CORRESPONDENCES AND DIFFERENCES 
23. Anonymous ( 2005 ), The Drugs and Cosmetics Act and Rules, Ministry of Health 
and Family Welfare, Department of Health, available: http://cdsco.nic.in/html/ 
Drugs&CosmeticAct.pdf . 
24. Anonymous ( 2005 ), Schedule M: Good manufacturing practices and requirements of 
premises, plant and equipment for pharmaceutical products, in The Drugs and Cosmetics 
Act and Rules, Ministry of Health and Family Welfare, Department of Health, pp. 386 – 
436, available: http://cdsco.nic.in/html/Drugs&CosmeticAct.pdf . 
25. Venkateswarlu , M. ( 2006 ), Why do we need revision of schedule M, available: http://www. 
pharmabiz.com , 2006. 
26. Anonymous ( 2006 ), Section 36: Manufacturing principles, in Therapeutic Goods Act 1989 , 
Offi ce of Legislative Drafting and Publishing, pp. 96 – 97, available: http://www.tga.gov.au/ 
legis/index.htm#instruments . 
27. Slater , T. ( 2002 ), Therapeutic goods (manufacturing principles) determination no 2 of 
2002 , Commonwealth Austral. Gaz. , GN 34, 2306 – 2307 . 
28. Anonymous ( 2002 ), Australian code of good manufacturing practice for medicinal 
products, Therapeutic Goods Administration, available: http://www.tga.gov.au/docs/pdf/ 
gmpcodau.pdf . 
29. Anonymous ( 2005 ), Medicines Act 1981, Parliamentary Counsel Offi ce, available: http:// 
www.legislation.govt.nz/browse_vw.asp?content - set=pal_statutes . 
30. Anonymous ( 2001 ), Guidance notes for applicants for consent to distribute new and 
changed medicines and related products, in New Zealand Regulatory Guidelines for 
Medicines , Vol. 1, 5th ed., MedSafe, available: http://www.medsafe.govt.nz/downloads/ 
vol1.doc . 
31. Anonymous ( 2005 ), New Zealand code of good manufacturing practice for manufacture 
and distribution of therapeutic goods, available: http://www.medsafe.govt.nz/Regulatory/ 
Guideline/code.htm . 
32. Anonymous ( 2002 ), Medicines and Related Substances Control Act 101 of 1965, 
Medicines Control Council, available: http://www.mccza.com/showdocument. 
asp?Cat=27&Desc=Acts%20and%20Regulations . 
33. Anonymous ( 2003 ), General regulations made in terms of the Medicines and Related 
Substances Act 1965 (Act no. 101 of 1965) as Amended, Government Notice, Department 
of Health, available: http://www.mccza.com/showdocument.asp?Cat=27&Desc= 
Acts%20and%20Regulations . 
34. Anonymous ( 2005 ), Guide to Good Manufacturing Practice for Medicines in South 
Africa, Medicines Control Council, available: http://www.mccza.com/showdocument. 
asp?Cat=21&Desc=Guidelines%20 - %20Good%20Manufacturing%20Practices . 
35. Anonymous ( 2002 ), WHO Expert Committee on Specifi cations for Pharmaceutical Preparations: 
37th Report, WHO Technical Report Series 908, World Health Organization, 
Singapore, available: http://whqlibdoc.who.int/trs/WHO_TRS_908.pdf . 
36. Anonymous ( 2002 ), Annex 4: Good manufacturing practices for pharmaceutical products: 
Main principles, in WHO Expert Committee on Specifi cations for Pharmaceutical 
Preparations: 37th Report, WHO Technical Report Series 908, World Health Organization, 
Singapore, pp. 36 – 89, available: http://whqlibdoc.who.int/trs/WHO_TRS_908.pdf . 
37. Anonymous ( 2005 ), Annex 2: Good manufacturing practices: Requirement for the sampling 
of starting materials (Amendment), in WHO Expert Committee on Specifi cations 
for Pharmaceutical Preparations: 39th Report, WHO Technical Report Series 929, 
World Health Organization, Singapore, pp. 38 – 39, available: http://whqlibdoc.who.int/trs/ 
WHO_TRS_929_eng.pdf . 
38. Anonymous ( 1992 ), Annex 1: Good manufacturing practices for pharmaceutical products: 
Good manufacturing practices for active pharmaceutical ingredients (bulk drug 

substances), in WHO Expert Committee on Specifi cations for Pharmaceutical Preparations: 
32th Report, WHO Technical Report Series 823, World Health Organization, 
Geneva, pp. 72 – 79, available: http://whqlibdoc.who.int/trs/WHO_TRS_823.pdf . 
39. Anonymous ( 1998 ), Annex 5: Good manufacturing practices: Supplementary guidelines 
for the manufacture of pharmaceutical excipients, in WHO Expert Committee on Speci- 
fi cations for Pharmaceutical Preparations: 35th Report, WHO Technical Report Series 
885, World Health Organization, Madrid, Spain, pp. 50 – 71, available: http://whqlibdoc. 
who.int/trs/WHO_TRS_885.pdf . 
40. Anonymous ( 2002 ), Annex 6: Good manufacturing practices for sterile pharmaceutical 
products, in WHO Expert Committee on Specifi cations for Pharmaceutical Preparations: 
36th Report, WHO Technical Report Series 902, World Health Organization, Singapore, 
pp. 76 – 93, available: http://whqlibdoc.who.int/trs/WHO_TRS_902.pdf . 
41. Anonymous ( 1993 ), Annex 3: Good manufacturing practices for biological products, in 
WHO Expert Committee on Specifi cations for Pharmaceutical Preparations: 33th Report, 
WHO Technical Report Series 834, World Health Organization, Geneva, pp. 20 – 30, 
available: http://whqlibdoc.who.int/trs/WHO_TRS_834.pdf . 
42. Anonymous ( 1995 ), Annex 7: Good manufacturing practices: Supplementary guidelines 
for the manufacture of investigational pharmaceutical products for clinical trials in 
humans, in WHO Expert Committee on Specifi cations for Pharmaceutical Preparations: 
34th Report, WHO Technical Report Series 863, World Health Organization, Geneva, pp. 
97 – 108, available: http://whqlibdoc.who.int/trs/WHO_TRS_863_(p1 - p98).pdf (pp. 97 – 98); 
http://whqlibdoc.who.int/trs/WHO_TRS_863_(p99 - p194).pdf (pp. 99 – 108). 
43. Anonymous ( 1995 ), Annex 8: Good manufacturing practices: Supplementary guidelines 
for the manufacture of herbal medicinal products, in WHO Expert Committee on Speci- 
fi cations for Pharmaceutical Preparations: 34th Report, WHO Technical Report Series 
863, World Health Organization, Geneva, pp. 109 – 113, available: http://whqlibdoc.who. 
int/trs/WHO_TRS_863_(p99 - p194).pdf . 
44. Anonymous ( 2002 ), Annex 3: Guidelines on good manufacturing practices for radiopharmaceutical 
products, in WHO Expert Committee on Specifi cations for Pharmaceutical 
Preparations: 37th Report, WHO Technical Report Series 908, World Health 
Organization, Singapore, pp. 26 – 35, available: http://whqlibdoc.who.int/trs/WHO_TRS_ 
908.pdf . 
45. Anonymous ( 2005 ), Annex 3: WHO Good manufacturing practices: Water for pharmaceutical 
use, in WHO Expert Committee on Specifi cations for Pharmaceutical Preparations: 
39th Report, WHO Technical Report Series 929, World Health Organization, 
Singapore, pp. 40 – 58, available: http://whqlibdoc.who.int/trs/WHO_TRS_929_eng.pdf . 
46. Anonymous ( 2006 ), Background to PIC, available: http://www.picscheme.org/ 
indexnofl ash.php?p=backg . 
47. Anonymous ( 2006 ), Introduction, available: http://www.picscheme.org/indexnofl ash. 
php?p=intro . 
48. Anonymous ( 2006 ), List of PIC/S participating authorities ( & observers), available: http:// 
www.picscheme.org/indexnofl ash.php?p=members . 
49. Brunner , D. ( 2004 ), Pharmaceutical inspection co - operation scheme (PIC/S) , Qual. Assur. 
J. , 8 , 207 – 211 . 
50. Anonymous ( 2006 ), Guide to good manufacturing practice for medicinal products, PE 
009 – 3, Pharmaceutical inspection co - operation scheme, available: http://www.picscheme. 
org/guides.php# . 
51. Anonymous ( 2006 ), Structure of ICH, available: http://www.ich.org/cache/html/ 
510 - 272 - 1.html . 
REFERENCES 161

162 CORRESPONDENCES AND DIFFERENCES 
52. Anonymous ( 2000 ), Good manufacturing practice guide for active pharmaceutical ingredients 
Q7, ICH harmonised tripartite guideline, ICH Steering Committee, available: 
http://www.ich.org/LOB/media/MEDIA433.pdf . 
53. Anonymous ( 2006 ), Quality guidelines, available: http://www.ich.org/cache/compo/ 
363 - 272 - 1.html . 
54. Anonymous ( 2006 ), The founding of ASEAN, available: http://www.aseansec.org/ 
7069.htm . 
55. Anonymous ( 2006 ), Pharmaceuticals, available: http://www.aseansec.org/8657.htm . 
56. Vernengo , M. J. ( 1998 ), Advances in pharmaceutical market integration in MERCOSUR 
and other Latin American countries , Drug Info. J. , 32 ( 3 ), 831 – 839 . 
57. Anonymous ( 2005 ), Division 1A establishment licences, in Consolidated Statutes and 
Regulations , Food and Drugs Act, Food and Drug Regulations, Part C, Department of 
Justice Canada, available: http://laws.justice.gc.ca/en/f - 27/c.r.c. - c.870/230049.html . 
58. Anonymous ( 2002 ), Note for guidance on quality of water for pharmaceutical use, CPMP/ 
QWP/158/01, Committee for Proprietary Medicinal Products, Quality Working Party, 
London, available: http://www.emea.eu.int/pdfs/human/qwp/015801en.pdf . 
59. Anonymous ( 1998 ), Stability testing on active ingredients and fi nished products, in The 
Rules Governing Medicinal Products in the European Union , Vol. 3A, European Commission 
Directorate General III, pp. 143 – 151, available: http://pharmacos.eudra.org/F2/ 
eudralex/vol - 3/pdfs - en/3aq16aen.pdf . 

QUALITY 
SECTION 3


165 
3.1 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc. 
ANALYTICAL AND COMPUTATIONAL 
METHODS AND EXAMPLES FOR 
DESIGNING AND CONTROLLING 
TOTAL QUALITY MANAGEMENT 
PHARMACEUTICAL 
MANUFACTURING SYSTEMS 
Paul G. Ranky, 1 Gregory N. Ranky, 2 Richard G. Ranky, 1 and 
Ashley John 1 
1 New Jersey Institute of Technology, Newark, New Jersey 
2 Public Research University of New Jersey, Newark, New Jersey 
Contents 
3.1.1 Introduction 
3.1.2 Flexible Pharmaceutical Manufacturing and Assembly System Design 
3.1.3 Flexible Manufacturing Model Integrated with Design 
3.1.4 Real - Time Operation Control 
3.1.5 Innovative Design 
3.1.6 Open Innovation Architecture 
3.1.7 Generic, Object - Oriented Innovation Process Modeling Method and Sample 
Model 
3.1.8 Systems Approach to Pharmaceutical Manufacturing Systems Management 
3.1.9 Requirements Analysis for System Product, Process, and Service Design 
Innovation 
3.1.10 Innovation Risk Analysis and Opportunity Method and Tool with Pharmaceutical 
Manufacturing System Applications 
3.1.11 Open - Source Computational Statistical and Three - Dimensional Multimedia for 
Pharmaceutical Manufacturing System Innovation and Project Communication 
3.1.12 RFID Applications 
3.1.13 RFID Examples 

166 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
3.1.14 RFID Integration Models for Digital Pharmaceutical Manufacturing and Assembly 
Supply Chains 
3.1.15 Evaluation of Network Simulation Results 
3.1.16 Summary 
3.1.17 Complimentary Video on DVD 
References 
3.1.1 INTRODUCTION 
Total quality management (TQM) and operation control in pharmaceutical manufacturing 
system design engineering is essential. TQM - focused pharmaceutical 
manufacturing system engineering involves the continual satisfaction of customer 
requirements at lowest cost by harnessing the efforts of everybody in the company. 
Quality assurance means sustaining a system that prevents defects. This includes 
quality control and quality engineering. Quality control means establishing and 
maintaining specifi ed quality standards of products; quality engineering is the establishment 
and execution of tests to measure product quality and adherence to acceptance 
criteria. 
This chapter explains the importance of reducing variation for the purpose of 
implementing total quality in every process of the pharmaceutical design and manufacturing 
enterprise. Furthermore, it represents a modular product, process, service 
design, implementation, and management approach to the introduction of various 
TQM methods, tools, technologies, and their management issues within a variety of 
small, medium, and large enterprises for the purpose of designing and controlling 
pharmaceutical manufacturing systems. 
These aspects are very important, clearly illustrated by the fact that the U.S. Food 
and Drug Administration (FDA) has three classifi cation levels for medical 
products: 
• Class I products are passive devices that do not enter the patient ’ s body or 
contact only the skin. 
• Class II products are active devices or devices that are used to administer fl uids 
to the patient ’ s body. 
• Class III products are implanted inside the patient ’ s body. 
The FDA is familiar with the complexity of designing pharmaceutical systems. To 
support this activity, there are several software tools that help product/process and 
system designers to achieve the above. 
It should also be noted that the FDA expects design validation results to accompany 
some submissions. This is particularly true of class II and III devices. The 
agency expects such analysis results to match those obtained with established experimental 
methods. A number of software tools, including fi nite element analysis 
(FEA), motion and actuation simulation, computational fl uid dynamics (CFD), in 
conjunction with the computer - aided design (CAD) used for the designs themselves 
and other solutions are available that help today ’ s pharmaceutical/medical designer/ 
medical manufacturing/assembly system designer to meet the complex requirements 
of the industry as well as the FDA. (The key, here, is to accept the important 

principle that pharmaceutical design and manufacturing/assembly and even packaging 
must be an integrated approach.) 
The main problems when applying a traditional quality management philosophy 
to any pharmaceutical design/manufacturing/assembly challenge include the 
following: 
• This philosophy focuses on correcting mistakes after they have been made, 
rather than preventing them in the fi rst place. 
• It allows mistakes to be made. It actually builds them into every aspect of the 
system, typically costing around 20% of the turnover. 
• It accepts that quality has to be sacrifi ced as the volume and the productivity 
go up. 
• As viewed by accountants, it is an expensive add on item of the value chain. 
However, modern thinking claims that, because TQM involves every person, 
aspect, and machine of the organization, it requires a total commitment. It is not a 
“ test - and - fi x ” approach. It is a preventive system designed into every aspect of the 
world - class design, manufacturing, and service enterprise, including product design, 
manufacture, and management (and even in accounting terms costing somewhat 
less than conventional quality systems, i.e., typically around 10% of the turnover). 
The fundamental goal of TQM and TQC (total quality management and control) 
is to program, measure, and keep process variability under control. Some of these 
methods discussed in this chapter are as follows: 
• Pharmaceutical manufacturing system design methods and tools with 
examples 
• Process modeling for designing and running pharmaceutical manufacturing 
systems 
• Requirements analysis modeling for pharmaceutical manufacturing systems 
• Risk analysis modeling for pharmaceutical manufacturing systems 
• Dynamic modeling and network simulation for globally distributed pharmaceutical 
manufacturing systems and other methods and tools 
3.1.2 FLEXIBLE PHARMACEUTICAL MANUFACTURING AND 
ASSEMBLY SYSTEM DESIGN 
A fl exible pharmaceutical manufacturing/assembly system, (FMS) is a highly automated, 
distributed feedback - controlled system of data, information, and physical 
processors, such as computer and manually controlled machines, cells, workstations, 
and robots, in which decisions have to be made often in real time. This is only possible 
if all information processors (including the human resources of such systems) 
are “ well informed ” and lean/fl exible, meaning that they have the exact information 
at the exact time, format, and mode they need to allow responsible decision making 
within given time constraints. Note that this is a fundamentally different system 
design concept than that of the transfer line, operating on a fi xed cycle time, and 
designed for large batch production [1 – 8] . 
When designing a fl exible manufacturing/assembly system (FMS/FAS), the design 
team should consider the following steps: 
FLEXIBLE PHARMACEUTICAL MANUFACTURING 167

168 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
1. Collect all current and possible future user and system requirements. 
2. Analyze the system (i.e., the data processing and the FMS/ FAS hardware and 
software constraints). 
3. Design an appropriate data structure and database for describing processors 
and their resources, such as machines, robots, and tools (and/or robot hands, 
probes, sensory - based inspection and assembly tools, etc.). 
4. Specify and design programs and query routines and dialogues that are capable 
of accessing this database as well as communicating with the real - time production 
planning and control system of the FMS/FAS. 
5. Design and integrate the system with the rest of the hardware and software, 
including on - line manuals, education, and training packages, preferably in 
interactive, engineering multimedia format. 
6. Maintain the system and continuously learn for the benefi t of the existing as 
well as future system designs. 
Probably the most important questions to be answered before starting to design 
such a system are: Who is going to use it? For what purposes? With what data? How 
will it be used? 
As an example, consider that tooling data in FMSs will typically be used by 
several subsystems as well as by human beings are as follows: 
• Production planning subsystem 
• Process control 
• Part programming 
• Tool preset and tool maintenance 
• Tool assembly (manual or robotized) 
• Stock control and material storage 
By employing the above subsystems, the production planning system has to be 
informed in real time about the availability of tools in stock as well as about the 
current contents of the tool magazines of the machine tools (in the case of FASs 
the robot hands in the end - of - arm - tool magazines); otherwise it will not be able to 
generate a proper production schedule. 
It must be noted that the real - time aspect is important because tools are changed 
in the magazines of machines (or cells), not only because they wear, but also because 
different part programs may need different sets of tools. (The actual tool - changing 
operation is done in most cases by manipulators or by robots. The tool magazine 
loading/unloading procedure is performed mostly by human operators, sometimes 
by robots or special - purpose mechanisms, such as a tool shuttle.) 
Both the process control and the production planning systems have to update 
any changes and act in real time; otherwise the operation of the system can be 
disrupted. 
From the FMS/FAS tooling and tool management points of view one must 
emphasize the links between the CAD system, in which the parts are designed 
(using design for manufacturing principles), and the computer - aided manufacturing 
(CAM) system, where the FMS part programs are written. Typically, an FMS part 
programmer analyzes the CAD output (i.e., the design drawings of the pharmaceuti

cal products to be manufactured/assembled on the FMS), the fi xturing, the different 
setup (i.e., work - mounting) tasks, as well as the necessary operations, their alternatives, 
the required tools, and fi nally a precedence list of the resources (i.e., the possible 
candidates of processing stations, or cells, or machines). 
Real - time databases and software systems are also important, since they provide 
the reports and status information that are needed for the smooth operation of the 
FMS (in particular, its dynamic scheduler and other subsystems such as maintenance 
should be emphasized here) [4, 9 – 14] . 
3.1.3 A FLEXIBLE MANUFACTURING MODEL 
INTEGRATED WITH DESIGN 
The output of the CAM system is a production rule base. This is the knowledge the 
FMS needs to produce each pharmaceutical product. In this production rule base, 
among others, tools are assigned to each operation. The tool codes are selected by 
the FMS process planner or automatically assigned by a process planning system 
and are obtained from the tool database. 
On the basis of the requested tools a list is sent via the network to the tool 
preparation facility, or station, where the actual tools are prepared (i.e., assembled 
and preset) and stored in an appropriate way such that the material - handling system 
of the FMS can pick them up [12 – 21] . 
The tool preparation station also deals with other activities, among which the 
most important are as follows: 
• Tool service and maintenance 
• Tool assembly to orders (as it is necessary to replace worn tools) 
• Tool preset, tool inspection and adjustment 
• Real - time tool pickup and tool transportation organized to serve the needs of 
the real - time FMS 
The tool preparation station receives its orders, initially originated by the CAD 
data processing system, via the FMS network and technically specifi ed by the CAM 
system in the form of a production rule base. Order data arriving at the tool preparation 
station include the following: 
• Part orders (consisting of part codes and quantities). Note that this is a very 
important data set for the real - time FMS dynamic scheduler too. 
• Notifi cation of when the parts are physically available for FMS processing, 
representing a due date for tool preparation. 
• A priority order (note that this can change because of some real - time changes 
in the system, and thus this station must be able to cope with this task too). 
• The portion of the production rule base describing the requirements regarding 
tool preparation. 
The tool preparation station keeps in touch with the real - time FMS system, as well 
as with the rest of the system, by feeding back important tooling system - related 
data: 
A FLEXIBLE MANUFACTURING MODEL INTEGRATED WITH DESIGN 169

170 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
• Stock reviews (regarding tools) 
• FMS status report (regarding tools) 
• Part priority status reports (in case dynamic changes must be performed in the 
FMS which have an effect on tooling needs and tool preparation due dates) 
3.1.4 REAL - TIME OPERATION CONTROL 
The real - time part of the FMS operation control and management system must deal 
with the following tasks: 
• It must handle the application of tools for a variety of processes as defi ned in 
the production rule base and assigned in real time to the FMS/FAS resources 
by the dynamic scheduler. 
• It must provide data to control the transportation of tools and tool magazines 
within the FMS. 
• It must provide information to perform and supervise tool changes and tool 
magazine changes at all levels. 
• It must be notifi ed of tool inspection results (e.g., if it fi nds a worn - out tool as 
a result of an inspection procedure, it must generate a command that instructs 
the tool magazine update system to change the tool in question in the appropriate 
tool magazine). 
• It must provide information in the case of emergency. 
• It must provide the necessary interfaces and data to perform diagnostic/recovery 
operations, preferably using diagnostic expert systems. 
Finally, let us underline an important feedback loop starting at the real - time 
system and ending at the tool preparation station, which contains the real - time tool 
status, wear, and part priority information. These data are often useful to those 
people and/or system software systems that deal with the generation of the production 
rule base. It is also a very useful data set for FMS designers, since a lot of data 
which would previously have been lost will be saved in this way. 
The most important operation control activities in FMS/FAS identify three levels 
at which simulation and optimization are required prior to or during FMS/FAS part 
manufacturing: 
1. The factory level or business level handled by the business system of the 
computer integrated manufacturing (CIM) or, even broader, the enterprise 
resource management system 
2. The FMS off - line level representing scheduling, simulation, and optimization 
activities prior to loading a batch or a single component on the FMS (handled 
sometimes by the CAM system, sometimes by the FMS part programming 
computer) 
3. The real - time controlled level handled by the FMS/FAS operation control 
system, a dynamic scheduler with integrated tool management and multimedia 
support, representing a situation where the parts are already physically as well 
as logically in the real - time controlled environment 

Due to its complexity, a truly integrated approach is required in designing a 
production rule base to provide the job description for the FMS dynamic scheduler. 
This is because the dynamic system relies heavily on the knowledge base as represented 
by the rule base, and an overly restrictive rule base will lead to ineffi cient, 
at times even wrong, decisions. In other words, such a structure should represent all 
the multilevel interactions and their possible precedence rules that relate to the 
manufacturing process planning and processing decisions in an FMS. This turns out 
to be a diffi cult task. 
It should be underlined that the application of multimedia at this level is extremely 
benefi cial in terms of part program preparation, teaching/training operators on 
setting up parts, fi xtures, tools, machines, for troubleshooting, for regular maintenance, 
at the computer numerical control (CNC) level programming, robot programming, 
placement machine programming, programmable logic controller (PLC) 
programming, quality control, maintenance, and other tasks. 
Most FMSs have some part - buffering capability. This may be not for scheduling 
reasons, but for technological, that is, process - planning, reasons (e.g., the part must 
cool before an accurate inspection procedure is performed). Some level of buffering 
is useful and necessary because of reliability reasons. (The actual number of buffer 
store locations should be established on the basis of simulation and experience.) 
Cells often have some buffers too. The reason for this is that, by providing a part 
in the input queue of the cell just before the currently processed part is fi nished at 
the particular cell, the cell is kept running at its highest effi ciency level, since time 
is only “ wasted ” for part changing. The other important point to note is that well - 
designed part buffers offer a direct access pickup/load facility, making the rescheduling 
process in the queues short, simple, and dynamic [18, 19, 21 – 27] . 
3.1.5 INNOVATIVE DESIGN 
The key objective of this chapter is to describe a generic and systematic pharmaceutical 
manufacturing/assembly system design method that includes product, 
process, service systems, and even innovation project management architecture 
aspects of such systems. 
This architecture must be simultaneously novel as well as compliant with set 
guidelines by the product/process design industry and the PMI (Project Management 
Institute), following International Organization for Standardization (ISO) 
9000:2000 quality standards. Our tested pharmaceutical manufacturing system 
design solution integrates object - oriented process modeling, requirements and risk 
analysis, statistical methods, design of experiments, and three - dimensional (3D) 
interactive multimedia methods and tools which are 100% Web compatible. 
Furthermore, our methods and software tools are generic in that they can be 
applied not only to systems such as the pharmaceutical industries or automobile 
manufacturing but also to processes such as the oil business or services such as 
education. 
A pharmaceutical manufacturing system design requires signifi cant level of innovation. 
The broadest defi nition of innovation is the act of introducing something 
new to a society or community, whether a product or process. This is often confused 
with invention, which focuses more on specifi c objects. Within pharmaceuticals 
innovation can therefore include new business structures within the company, 
INNOVATIVE DESIGN 171

172 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
manufacturing processes and quality control for the medications, and product materials. 
Process and service improvements can also qualify as innovation, but note that 
in this case services are usually counted as processes [4, 13, 21, 23, 28 – 32] . 
Discoveries such as the charting of new planets, land masses, or forms of life are 
not classifi ed as innovations as they had existed before being observed by humans. 
When new species are introduced into a society and fi nd a specifi c use, it can be 
classifi ed as innovation. A pharmaceutical example of this is the antibiotic penicillin. 
Although it had existed as a fungal secretion, it was only within the past century 
that it was used to actively eliminate infectious bacteria. In initial analysis, however, 
it was not thought to survive long enough within the human body to be effective. 
This brings a vital aspect of innovation, namely the ability to recognize alternative 
uses for existing processes or tools. This is diffi cult as unexpected changes within a 
system are usually labeled as mistakes or anomalies. The development of the Post - It 
note is an example of this. The original goal was to create a high - strength adhesive, 
and an extremely weak one was created by accident. Nonetheless, instead of simply 
disposing of it, the possibilities of this new substance were examined by technicians 
and managers alike, allowing the use of easily placed reminders for everyday usage. 
Possessing an ultraweak adhesive allows Post - It notes to be removed without damaging 
the surfaces that they are placed on, and they are available in a variety of 
colors and sizes. 
The former example is a radical innovation, not only because it allowed signifi - 
cant changes in message reminders and adhesives, but also because it was completely 
unexpected. The diversifi cation of Post - It notes into different sizes and colors 
is an example of incremental innovation, which involves step - by - step changes and 
improvements to existing products or processes. Radical innovations are far less 
common, though their effects are farther reaching over both society and history. 
The general trend through human history has been one of learning to consciously 
recognize and direct innovation, particularly through combining science and technology. 
Human societies have often worked with certain processes even without 
fully understanding their effects or underlying ideas. Metallurgy shows this clearly, 
as iron, bronze, and gold have been used for millennia before the molecular structure 
could be seen and analyzed. Note also that, although our ancestors could not 
describe their chemical composition, these metals served a great many successful 
purposes. In these cases, the goals of innovation are highly pragmatic, as successful 
solutions are passed down and taught to future generations. 
Those who innovate can therefore learn from working, viable solutions to begin 
their own practices. Those who continuously work with a fi xed set of designs must 
be willing to experiment, test, and diversify their practices to avoid stratifi cation, as 
innovation not only allows survival but also encourages prosperity. 
The ability to innovate also involves learning from past mistakes, not just one ’ s 
own. Mistakes and errors in practices can be both costly and dangerous but can be 
prevented from occurring successively if their causes are determined. This can be 
diffi cult because a near - miss scenario can be seen either as an infrequent event or 
as an averted disaster. The fi rst reaction to this is usually to continue without changing 
current practices, allowing for similar mistakes to occur. Learning requires all 
levels of an organization to participate and create channels of communication to 
innovate effectively, as the inability to share experience denies new opportunities 
[13, 29 – 40] . 

3.1.6 OPEN INNOVATION ARCHITECTURE 
Innovation as a process and the related research - and - development (R & D) project 
management are considered to be two of the most complex information systems 
and engineering architectures due to the large number of attributes, processes, and 
dynamic changes projects go through during their life cycle. 
Following our integrated and simultaneously open architectural approach, we 
look at every innovation process and project as a system built of objects and classes 
of objects. 
Then we look at the way the components of these systems interact with each 
other. Once we understand these behaviors, we follow our integrated system 
approach in terms of looking at the project management system as processes, trying 
to satisfy customer requirements and also representing risks. 
We then embed this system model into a statistical analysis and 3D interactive 
multimedia framework (Figure 1 ). We use statistical methods to capture processes 
before they go out of control as well as to perform trend analysis, a great opportunity 
for innovation, and use 3D interactive multimedia and 3D visualization methods 
over the Web for communication purposes with global innovation team members. 
The emphasis on collaboration in today ’ s competitive medical drug fi eld requires 
these virtual environments to streamline team interaction. (Note that the active 
FIGURE 1 When designing lean and fl exible pharmaceutical manufacturing/assembly/ 
packaging systems, one needs to analyze the required processes, customer, user, maintenance, 
quality, reliability, fl exibility, lean, design requirements, and risks involved with any of the 
listed processes, all in a statistical framework. (Note that our 3D interactive multimedia and 
simulation framework supports integrated digital design and digital manufacturing system 
design principles, meaning that one should test all designs and systems fi rst on the screen, 
and only if everything looks fi ne, in the real world.) 
Process analysis 
Risk analysis 
statistical analysis, design of experiments, 
Web-base 3D interactive multimedia, DVD full-screen 
vidoeos and just-in-time iPod videos for 
knowledge managemet 
Requirements analysis 
OPEN INNOVATION ARCHITECTURE 173

174 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
code spreadsheets and 3D objects referred to in this presentation are all part of 
Ranky ’ s eLibrary and are available at http://www.cimwareukandusa.com .) 
To illustrate the importance of the “ openness ” of our architecture, consider 
modern simulation/analysis tools by Parametric Technology Corporation (PTC) 
(Figure 2 ), and PLM (product life - cycle management) tools, such as the IBM/Dassault 
Systemes Delmia tools for pharmaceutical manufacturing system modeling 
and design (Figure 3 ) with sensory feedback processing (Figure 4 ). 
Since these models can be designed, edited, run, and driven even over the limits, 
they can be extremely valuable sources for modeling in the digital domain, process 
analysis, requirements modeling, risk analysis, and even collecting statistical data 
and modeling breakdowns of complex systems. 
Observe the FEA in Figure 2 a . This is a torsional test of a pharmaceutical manufacturing 
machine element on the assembly line of a new medication packaging line. 
The line is still being tested and improved in the virtual environment, which greatly 
streamlines the refi nement process. As can be seen by the von Mises stress distribution, 
the sharp edges of the shaft will need to be rounded with a fi llet. These would 
also increase the distribution of the same stress and thus reduce the majority of the 
red zones (high stress) to blue or even green (low stress). Without using the virtual 
assembly line to test ideas before for the physical, the unexpected failure of this 
part could create delays or contamination of product or even harm human 
operators. 
As can be seen, digital pharmaceutical manufacturing/assembly/packaging and 
factory design tools include not only machines but also advanced sensors, actuators, 
controls, material - handling systems, labeling machines, and even ergonomically realistic 
human models and operators performing real - world tasks in extremely realistic 
model factories. Simulations like these are not just pretty models; they actually save 
huge investments because the factories are not built until the models are satisfactory. 
Keep in mind that making changes in a physical factory costs time, money, and 
possibly production effi ciency, even to just check a possible improvement. Virtual 
models can be simultaneously run thousands of times over a period of days, with 
hundreds of variables being optimized until the appropriate combination is chosen 
[30 – 36, 38, 40 – 44] . 
3.1.7 GENERIC, OBJECT - ORIENTED INNOVATION PROCESS 
MODELING METHOD AND SAMPLE MODEL 
Understanding, modeling, and then following processes, procedures, and best practice 
reusable processes are essential for every business to stay at the top. The pharmaceutical 
manufacturing system “ innovation business ” is not exception. 
Major international product/process design standards written and reviewed by 
thousands of leading researchers and companies around the world always help to 
create a model for complex problem - solving challenges such as innovation. Therefore 
this section discusses two of the eight quality management principles of the 
ISO 9000:2000 international quality standard and the way these rules should be 
applied to pharmaceutical manufacturing system designs. 
We do this for the purpose of developing systematic innovation (with related 
project modeling skills) and reusable, tested pharmaceutical system design 

FIGURE 2 Finite element torsional test of pharmaceutical manufacturing machine element 
on assembly line of new medication packaging line. The line is still being tested and improved 
in the virtual environment, which greatly streamlines the refi nement process. As can be seen 
by the von mises stress distribution, the sharp edges of the shaft will need to be rounded with 
a fi llet. These would also increase the distribution of the same stress and thus reduce the 
majority of the red zones (high stress) to blue or even green (low stress). 
Von Mises stress distribution plot Displacement distribution plot 
(a) 
max_disp_mag 
(mm) 
P_Pass 
Scale 1.0000E + 00 
Loadset: LoadSet1 
max_disp_mag 
max_disp_mag 
0.00 
0.25 
0.20 
0.15 
0.10 
0.05
0 2 3
PLoop Pass
4 5 6 
strain_energy 
(mm N) 
P_Pass 
Scale 1.0000E + 00 
Loadset: LoadSet1 
strain_energy 
strain_energy 
2.00 
4.00 
6.00 
10.00 
8.00 
12.00 
14.00 
16.00 
18.00
0 2 3
PLoop Pass
4 5 6 
max_strees_vm 
(N/mm^2) 
P_Pass 
Scale 1.0000E + 00 
Loadset: LoadSet1 
max_strees_vm 
max_strees_vm 
40.00 
50.00 
60.00 
70.00 
80.00 
90.00 
100.00 
110.00 
120.00 
130.00 
140.00
0 2 3
PLoop Pass 
4 5 
(b)

176 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
FIGURE 3 Modern simulation/analysis and PLM tool: IBM/Dassault Systemes Delmia for 
pharmaceutical manufacturing system modeling and design. The benefi ts are huge, since the 
system can be built and tested in the digital domain. (Courtesy of IBM/Dassault Systemes 
Delmia, Inc.) 
processes. We use our own object - oriented process modeling method, called CIMpgr. 
The nature of this method tansforms into UML models [the Unifi ed Modeling 
Language of information technology (IT)] and also complies with international 
process modeling standards used in complex system modeling environments. 
First, we will discuss a few important defi nitions that closely relate to ISO 
9000:2000 (quality process modeling) standard principle 4: 
• A process, or activity, can be defi ned as a transfer function with one or more 
inputs, outputs, controls, and resources that together all enable the variables to 
gain data and then fi re. 
• Transfer functions, when fi red, create a transformation process. A transformation 
process in a project is made up of methods, steps, tasks, and various algorithms 
and processes that acquire and manipulate data and then turn it into 
system output(s). Note that the input data can describe material, human knowledge, 
technological standing, fi scal information, and others. 
• The output of the process is a product that consists of specifi c technical and/or 
social products and services that conform to the sponsor ’ s requirements. 
• Processes, in terms of quality project management, have visibility, documentation, 
and traceability. 
• In this context visibility, relates to whether we know and transparently (or 
graphically) see what methods and techniques, system process steps, and technologies 
are involved when creating the desired output. Do we know the 

FIGURE 4 Advanced sensors working in pharmaceutical assembly systems help real - time 
operation control and quality assurance system to test every product. (This is often referred 
to as the zero - defect policy designed into a system.) The luminescence sensor illustrated will 
detect a wide variety of invisible targets. This STEALTH - UV sensor was designed to sense 
the presence of invisible fl uorescent materials contained in or added to many products. Users 
can detect the most diffi cult targets, including clear tamper - proof seals, clear labels, and invisible 
registration marks. This unique sensor is also ideal for solving many of today ’ s toughest 
problems in product orientation, inspection, and verifi cation. (Courtesy of TRI - TRONICS 
Co., Inc., www.ttco.com .) 
sequence of these steps and the possible parallel process relationships? How 
does one process affect the other? 
• Documentation means that the methods, steps, processes, and technologies are 
well specifi ed and recorded according to agreed - upon standard specifi cations. 
• Traceability means that the process steps as well as the output(s) can be traced 
back to actual customer requirements. 
• Process capability can be defi ned as the ability of the production process to 
meet certain specifi cations and tolerances. 
• Process discrepancy is the deviation of process settings from specifi cations. 
• Process variability is the variation in dimensional or other measurable characteristics 
of output from a production process. (Note that in any project the 
ultimate goal is to stay within the predefi ned limits of process variability and, 
if possible and feasible, to reduce process variability, because this typically 
reduces risk too.) 
• Variability can be expressed in terms of average range of standard deviation. 
• A process variable is a process parameter that fl uctuates in the manner of a 
random variable and hence requires surveillance. 
• Process management means getting the activities and procedures that highly 
skilled and experienced managers carry in their heads into the open by means 
of a well - documented model, often referred to as the process model [40 – 46] . 
GENERIC, OBJECT-ORIENTED INNOVATION PROCESS MODELING METHOD 177

178 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
3.1.8 SYSTEMS APPROACH TO PHARMACEUTICAL 
MANUFACTURING SYSTEMS MANAGEMENT 
Identifying, understanding, and managing interrelated processes as a system contribute 
to the organization ’ s effectiveness and effi ciency in achieving its objectives. 
(Note that every one of the key drivers listed below embed one or more innovation 
opportunities!) 
Key drivers and achievable gains include the following: 
• Processes that will achieve the desired results will become better integrated 
and aligned. 
• Management and process owners will have the ability to focus their efforts on 
the key processes. 
• Since the consistency, effectiveness, and effi ciency of the organization will grow, 
the confi dence of interested parties and collaborators in the organization will 
grow too. 
• System structuring and fi ne - tuning will become possible to achieve the organization 
’ s objectives in the most effective and effi cient way. 
• Understanding the interdependencies between the processes of the system will 
yield good results. 
• Structured (and object/component - oriented) modeling approaches that harmonize 
and integrate processes will become reality. Employees will understand 
them, follow them, and therefore reduce waste and increase quality in every 
process. 
• The resistance created by cross - functional barriers will be reduced, providing 
a better understanding of the roles and responsibilities necessary for achieving 
common objectives. 
• Organizational capabilities and the establishment of resource constraints prior 
to action will be better understood by all involved (and mostly by all those who 
have created the models). 
• Targeting and defi ning how specifi c activities within a system should operate 
will become reality. 
• Continually improving the system through measurement and feedback - 
controlled evaluation becomes possible due to the analytical and quantifi able 
approach of the process models. (Note that at its ultimate level this will lead 
to a real - time, feedback - controlled enterprise capable of reacting to dynamically 
changing market needs.) 
After this introduction, let us show our object - oriented system components, 
following the above described ISO 9000:2000 principles, and how we can 
model complex innovation and related project management processes using them 
[40 – 54] . 
As a simple example, consider, that you are packaging a pharmaceutical product 
using a line that performs various process steps. Figure 5 a illustrates one of 
these steps. It has input(s), output(s), control(s), and resource(s). These data 

types help to identify under what conditions the process should be executed by 
the pharmaceutical manufacturing system. (Defi nitions are offered in the diagrams.) 
We can also see the way the CIMpgr process maps into a UML diagram. 
This is important, since UML is the modeling language of the IT professionals 
who will program the PLCs and control systems for the lines. Figure 5 b shows 
how the CIMpgr process maps into a UML diagram. We can see in Figure 5 a how 
multiple processes have to interact as we design a pharmaceutical assembly 
system. 
FIGURE 5 Object - oriented process modeling method (CIMpgr) as applicable to pharmaceutical 
manufacturing/assembly/packaging system design. 
This box represents the process. We can 
identify a process by naming it A0 (as a 
parent), and it’s children (a A1, A2, etc.) 
CIMpgr Process Model A0 
This is the control side to our process. This is where data 
somehow limits, or controls the process. (As an example for 
controls, imagine the international emission control 
regulations that automotive designers must follow). We can 
identify each control by a variable name, such as C1A0. 
A0 
DBI_A0: 
This identifies a data 
storage, a file, or a 
database for the process. 
This is the resource side to our process. This is where 
data describes the available manpower, hardware, 
software, and other available resources for executing the 
process. We can identify each resource by a variable 
name, such as R1A0. 
This is the administrative 
section of out model. 
Purpose: Why are we doing this? What is the fundamental purpose of this model? 
Viewpoint: ‘As is’ System Analyst, System Designer; ‘To be’ System Analyst, System Designer 
Authoring Team Members: Ranky, with 
Key Contact: Paul G. Ranky, Email: cimware@earthink.net, USA Tel: (201) 493 0521 
Client Ref: Company ABC Inc. 
Data: January 21, 2004, Version: ver. 1.0 
Confidential! Public Release: OK; Object / Class inheritance: ON 
R1A0: 
R2A0:
C3A0: 
C4A0: 
C2A0: 
C1A0: 
I1A0: 
I2A0: 
I3A0: 
I4A0: 
I5A0: 
I6A0: 
C5A0: 
C6A0: 
R3A0: 
R4A0: 
R5A0: 
O5A0: 
O4A0: 
O2A0: 
O1A0: 
This is the 
output side to 
our process. 
This is where 
data leaves 
the process. 
We can 
identify each 
output by a 
variable 
name, such 
as O1A0. 
This is the 
input side 
to our 
process. 
This is 
where data 
enters the 
process. 
We can 
identify input 
by a 
variable 
name, such 
as I1A0. 
(a) 
PHARMACEUTICAL MANUFACTURING SYSTEMS MANAGEMENT 179

180 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
Class NameL ClassA0 
Attributes in detail (also in CIMpgr model): 
Input(s): 
I1A0= , 
I2A0= , 
Output(s): 
O1A0= , 
O2A0= , 
Control(s): 
C1A0= , 
C2A0= , 
Resource(s): 
R1A0= , 
R2A0= , 
Link to 
Requirements Analysis Model Filename: 
Risk Analysis Model Filename: 
Attributes 
Attributes 
Operations 
Operations 
Operations (Mathematical TR functions, pseudo code for 
proc. descr., etc., using the attributes, as above=CIMpgr 
model attributes) 
Class Name: 
ClassA1 
Attributes 
Attributes 
Operations 
Operations 
Class Name: 
ClassA2 
(b) 
I1A1: New 
need from 
the customer 
/project 
sponsor 
C1A1: 
I2A1: 
I1A2: 
I3A1: 
I4A1: 
I5A1: 
I6A1: 
I7A1: 
C6A1: 
C1A3: 
I1A3: 
C2A3: 
C2A4: 
C1A4: 
C1A5: 
C2A5: 
I1A4: 
R2A4: 
C2A4: 
I1A4: 
C2A2: 
C1A2: 
C5A1: 
C2A1: 
Project time, budget and quality control. (Note that there are many other types of control, that relate to environmental 
issues, specific design, material, manufacturing, assembly, test, service, IT, and other processes and sub-systems.) 
Conception, or Conceptual, 
or Requirements Analysis 
Phase, or Process A1
A1 
A2 
A3 
A4 
O1A1: The results of the requirements analysis study. (This data-set offers valuable information for this future projects, 
for data mining, and for knowledge management.) O2A1: Project 
specification: What? 
When? How much? 
etc. 
R1A1: CORA 
software tool, 
and 
experienced 
CORA 
consultant R2A1: 
Requirements 
analysis team 
R3A1: PFRA 
risk analysis 
tool,project 
time and cost 
management 
software and 
consultant Definition, or Planning, or 
System Analysis Phase, or 
Process A2 
O1A2: Project specification results (This is n important data-set in case of multiple projects with precedence constraints. 
Also, this data-set offers valuable information for data mining, and for knowledge management.) O2A2: 
Detailed 
project 
specification 
DBI_A1: 
Process A1 
related 
requirements 
analysis 
results are 
stored here 
(e.g. results 
of a CORA 
study) 
DBI_A2: 
The 
system 
analysis 
documents 
and data 
are stored 
here 
R1A2: CIMpgr 
process 
model 
drawing tools, 
optional 
dynamic 
simulation 
software, and 
experienced 
consultant 
R2A2: Project 
time 
management 
and budget 
management 
software, and 
experienced 
planning team Design, or Acquisition, or 
System Design Phase, or 
Process A3 
O1A3 Project design result. (This data-set offers valuable information for 
data mining, and for knowledge management.) O2A3: 
Detailed 
project 
design 
DBI_A3: 
The 
system 
design 
blueprints 
are stored 
here 
R1A3: Project 
time 
management 
and budget 
management 
software, and 
experienced 
project design 
team 
R2A3: 
Specific 
CORA 
requirements 
analysis and 
PFRA risk 
analysis, and 
product/ 
process 
experts 
Operation, or Integration, 
or System Implementation 
and Test Phase, or 
Process A4 
A5 
O1A4: Project implementation results. (This data-set 
offers valuable information for data mining, and 
for knowledge management.) 
O2A4: 
Operational 
parameters of 
the 
implemented 
project 
Design-To-Analysis Feedback Loop 
Design-To-Requirements Feedback Loop 
Generic Project Management Model in CIMpgr for A1 to A5 Processes 
Key Contact: Paul G. Ranky, Email: cimware@earthlink.net, USA Tel: (201)493 0521 
Client Ref: Company ABC Inc., Date: June 09, 2004, Version: Ver. 5.0 
Detailed Project Management eBook Template Build-up. 
Object/Class Inheritance: ON 
Notation: Example: ‘Conception, or Conceptual, or Requirements Analysis Phase, or 
Process A1’, meaning, that this is a process, typically called by one or more of the listed 
names, for our purposes meaning exactly the same. 
DBI_A4: 
The system 
integration, 
implementation 
and test data 
are stored here 
R1A4: Project 
time 
management 
and budget 
management 
software, and 
experienced 
project 
implementation 
team 
DBI_A5: The 
post project 
review data is 
stored here 
based on longterm 
test and 
maintenance 
results 
Post Project Review, or 
Long Term System Test, or 
Maintenance phase, or 
Process A5 
O1A5: 
Test 
results 
O2A5: 
R2A5: 
R1A5: 
Continuous 
improvement 
team 
Implementation-To-Design Feedback Loop 
Long Term Test/Maintenance-To-Design Feedback Loop 
(c) 
FIGURE 5 Continued

3.1.9 REQUIREMENTS ANALYSIS FOR SYSTEM PRODUCT, PROCESS, 
AND SERVICE DESIGN INNOVATION 
Processes in a successful innovation project must satisfy requirements set by the 
market, the sponsors, and/or the inventor ’ s own dreams. Requirements analysis is 
considered to be one of the most important features of any innovative pharmaceutical 
manufacturing system project, because if done professionally, it helps to specify, 
research, and develop appropriate features and processes that customers need. 
In our innovation project examples we have focused on generic needs and 
requirements, and our associated “ customers ” are the pharmaceutical R & D team 
members, managers, and operators in various industries. 
In terms of our research approach, we have followed a proven method: Analyze 
the needs and the requirements, the demonstrated processes, and the methods and 
systems they try to or have to satisfy, and if you fi nd a “ gap ” , you have found an 
innovation opportunity. Note that when we search for this gap, it will simultaneously 
appear as a missing process in our CIMpgr model or as an existing process but 
missing attributes as well as a requirement in our CORA model (component - 
oriented requirements analysis) model: 
• Analyze the actual methods presented. Find the core methodologies, the 
mathematical models, and the underlying engineering and/or other science 
foundation. 
• Analyze the technologies involved. (How is science turned into a practical 
solution/engineering and/or computing technology?) Is there a need for a new, 
novel technology that has not been invented yet or applied in this fi eld? 
• Analyze and review the actual processes and the way the process fl ow is integrated. 
(Follow an object - oriented process analysis method, i.e., from concept 
to product.) Focus on the attributes of the processes. Note that by adding a new 
attribute you create new data types, with new information, and if your process 
can reason over this in a new way, new knowledge; therefore your combined 
CIMpgr and UML model becomes a new knowledge representation model too. 
This is important because innovation is formalized this way and can be communicated 
among global teams. 
• Analyze potential alternative solutions. (A pharmaceutical manufacturing/ 
assembly/packaging system must be very fl exible these days, due to dynamically 
changing customer requirements and even operating conditions.) 
• Analyze the benefi ts and the disadvantages of each process/solution. 
• Design alternative methods, processes based on what you have experienced/ 
seen and learned. 
• Design an integrated system, based on what you have analyzed in this case. 
• Work in a multidisciplinary team and exchange ideas. 
• Understand the boundaries as well as the tremendous potential of new ideas 
and developments by working on this case (realize that in order to survive and 
win, you must add value) [54 – 57] . 
After this short introduction, we demonstrate our CORA spreadsheet solution 
with a real - world example (Figure 6 ). 
REQUIREMENTS ANALYSIS 181

182 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
Object / Component Oriented Requirements Analysis Program for 
Networked Lean Manufacturing by Paul G. Ranky © 1992–2006 
© Paul G. Ranky. 2000–2002 
Engineering / Software Solutions 
Responding to Customer 
Requirements 
Lean Manufacturing Manager: 
Customer Requirements 
( Reflecting component object 
behavior related to customer needs) 
S.No Describe the Requirement 
Reliability of data transfer for realtime 
access should be high 
Reliability of reporting process failure 
to the line manager’s computer 
Ease of integration into a system (plug 
and play networking): important! 
Ease of machine programming (CNC 
machining / inspection) 
Ease of changng CNC part programs 
(locally, and via the network) 
Ease of adding new sensors to a 
workstation, CNC, or cell: high 
safety of operation: critical! 
Cost of change/extenaion/system 
expansion should be low 
Operator training needs and costs 
should be low 
Network installation complexity and 
cost should be low 
1
2
3 
3 
3 3 3 3 
3
3 3 
3 
4 
4
4
4
4 
5 
5
5 
6
7
8
9 
9 9 9 9 3 9 
9 9 9 9 
9 9 9 9 
9 
9 9 9 9 9 9 
9 9 9 9 
9 9 9 
9 
3 9 
3 
3
9 
3 3 
3 3 
3 9 
9 
9 9 
9 
9 9 9 9 
9 3 3 
9 3 3 
9 9 9 
9 9 9 
9 9 9 
9 3 3 
9 9 9 9 9 
9 9 9 
9 9 9 
9 3 3 10 
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
Importance Rating (1–5) 
Fieldbus Network 
Profibus Network 
DeviceNet Network 
Ethemet 
Graphical CNC Progr. 
On-site maint. Support 
Off-site maint. Support 
On-site Redundant Server 
Off-site Redundant Server 
Link to Factory prod. contr. 
Link to Factory TQM/TQC 
2D videos/3D multimedia 
Cell-based web Camera 
PC-based CNC Controller 
PC based workkstation 
contr. 
PC based Cell controller 
(a) 
FIGURE 6 CORA method. This is a spreadsheet - based tool, designed to analyze customer 
requirements. (A “ customer ” here can mean a pharmaceutical manufacturing line vendor, 
user, operators, maintenance engineers, and many others.) The key approach is that we create 
a correlation matrix and then evaluate the results using a quantitative, computational 
approach. This is much more accurate than just a simple structured list. Our method offers 
a list of all key requirements as well as the priorities for the pharmaceutical manufacturing 
system design team they should follow during the design process. (For more about this software 
tool, see http://www.cimwareukandusa.com . 
0 
Competitor C’s product 
Competitor B’s product 
Competitor A’s product 
Our product 
0
1
2
3
4
5
6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 
Enter Competitor C’s ratings (1-5); Graph 
Enter Competitor B’s ratings (1-5); Graph 
Enter Competitor A’s ratings (1-5); Graph 
Enter Our Product ratings 
(1 = low, 5 = High) 
Relative Importance Rating 
Absolute Importance Rating 
3
3 3
3 3 
3 3 
3 3
3 
3 
3 3 3 
3 
3 
5 5 5 
5 
5 
5 5 5 5 
2 3 3 2 2 2 
2 2 
2 2 
2 1 1 4 4 
4 4
4 4 4 
4 
4 
4 
4 
4 4 4 4 4 4 4 
4 
4 4 
4 4 
4 4 5 
416 416 416 489 207 147 111 186 107 144 135 87 87 567 0 0 0 0
0 0 0 0 13 8.5 13 1.9 1.9 3 3.2 2.4 4.1 2.5 3.3 4.6 11 9.3 9.3 9.3 
381 588 
Target Values (List here the parameters 
that specify engineering solutions 
accurately. If you don’t know the range of 
the acceptable values, use our Taguchi 
Calculator Program for Designing an 
Experiment) 
The system history database should 
be on the network 20 5 3 3 3 3 3 3 3 3 9 9 9 9 9 
Within 27 m sec 
Within 27 m sec 
Within 27 m sec 
Within 27 m sec 
GUI, iconized, 
multimedia 
Less than 3 
minutes 
Less than 24 hrsr 
0 sec switch 
Less than 30 sec 
switch 
Retresh every 2 
minutes 
Refresh every 2 
minutes 
320 240 pxels 
or better 
320 240 pxels 
or better 
Win, Linux, 
Solans, or OSX 
Response within 
12 m sec 
Response within 
24 m sec 
4 0 
(b)

3.1.10 INNOVATION RISK ANALYSIS AND OPPORTUNITY METHOD 
AND TOOL WITH PHARMACEUTICAL MANUFACTURING 
SYSTEM APPLICATIONS 
Our failure risk analysis and opportunity method and iterative software tool, as part 
of our New Product & Process Innovation (NPPI) Tool Library, promotes systematic 
collaboration and team - oriented engineering thinking when a new pharmaceutical 
manufacturing system process and/or product are developed. (We call it “ opportunity 
method ” too, since most risks, if not all, offer new opportunities for innovation.) 
It is based on our generic process failure risk analysis method that could be applied 
to literally any process that involves risk — and innovation is a very risky process. 
We follow a rule - based method when we analyze risk objects and components 
and their attributes. These plug - and - play rules can be different for different subjects, 
research areas, and industries. They can be designed and standardized for different 
industry sectors, enabling an analytical approach, systematic standardization, and 
accurate and predictable results. 
Our risk analysis method and tools help the engineering management team to 
understand some of the following concerns: 
• What could go wrong with the processes involved during the innovation 
project? 
• How badly might it go wrong and what could the fi nancial loss be? 
• Which are the highest risk processes/operations when working on the product/ 
process/service - related innovative design and project? 
• What needs to be done to prevent failures? 
• Which processes must be changed to reduce the risk of failure? 
• What tools and fi xtures are required to prevent failures and reduce the risk? 
• What education is needed for participants, innovators, engineers, and process 
owners, such as line management and operators, to reduce or prevent 
failures? 
After this introduction, we show the risk analysis system components, following 
the already described ISO 9000:2000 principles, and how we can model complex 
project management risks using them (Figure 7 ) (note that the active code spreadsheets 
and 3D objects are part of Ranky ’ s eLibrary) [50 – 55] . 
3.1.11 OPEN - SOURCE COMPUTATIONAL STATISTICAL 
AND THREE - DIMENSIONAL MULTIMEDIA FOR 
PHARMACEUTICAL MANUFACTURING SYSTEM INNOVATION 
AND PROJECT COMMUNICATION 
Since we follow an analytical, quantitative, and open - source computational approach, 
our pharmaceutical product/process and project management method and software 
toolset are implemented as (Internet browser readable) MS - Excel spreadsheets, 
integrated with several hyperlinks to the rule base and to optional 2D video and 
3D virtual - reality objects for visualization. 
OPEN-SOURCE COMPUTATIONAL STATISTICAL 183

184 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
FIGURE 7 The process failure risk analysis (PFRA) tool is an analytical and computational 
tool using rule bases for evaluating process risks. It is an ideal method and tool for reducing 
costly failures. (For more about this software tool, see http://www.cimwareukandusa.com . 
Rev.2.1.3. by Ranky 
9/19/01 
11/16/01 
Ranky 111601/DFRA_Ver.5 
List/Identify the Parts/Components Retrieved 
in Each Disassembly Process Step 
Painted metal PC cover (File: 3DMetalCover. 
mov) 
Floppy drive, hard drive, mounted in a solid 
sheet metal bracket inside the PC 
(File: 3DFloppyHDassy. mov) 
Floppy drive (File: 3DFloppyDrive. mov) 
Hard drive (File: 3DHDobj. mov) 
Metal bracket holding the floppy and the hard 
drive (File: 3DFloppyHDBracket. mov) 
Disassembly Process Code 
Engineering Rerease Date or Process Methodology 
Type of Product Disassembled 
Product Group Classifier 
Engineering Release Date of the Product 
Process 
Time 
Process 
Cost 
Accumulated 
Process Cost 
Ranky PC DisassyCode: 05/07/97 
5/7/97 
Electro-mechanical 
Desktop PC 
Estimated: 1993 
The DFRA Team Describes/Illustrates the Potential Disassembly 
Failure Mode and the Effect; the Risk of Failure 
Failure Mode(s) and Effect(s) 
Metal Cover scratched by slipped screwdriver 
As PC Metal Cover is removed, internal parts are cratched 
Floppy drive assy. screw removal can damage nother board 
Movie, illustrating floppy/HD assy. removal risks 
Can damage Floppy Drive if assy. Is dropped 
Can damage Hard Drive if assembly is dropped 
Can damage Hard Drive if assembly is dropped 
Proc. ID 
ID 5.1 
ID 5.2 
ID 5.3 
ID 4.1 
ID 4.2 
ID 4.3 
ID 3.1 
ID 3.2 
ID 3.3 
ID 2.1 
ID 2.2 
ID 2.3 
ID 1.1 
ID 1.2 
ID 1.3 
[sec] [USD] [USD] 
16.40 
45 0.21 
137 0.62 
35 0.16 
65 0.30 
12 0.05 
0.21 
0.83 
0.99 
0.28 
1.34 
(a) 
Ranky PC DisassyCode: 05/07/97 
5/7/97 ss 
Electro-mechanical 
Desktop PC 
Estimated: 1993 oduct 
ed 
ost 
This DFRA Study Prepared By 
DFRA Team 
Responsible Organization/Department 
Paul 
NJIT 
NJIT 
RPN 
Pr 
Nu 
Comments 
The DFRA Team Describes/Illustrates the Potential Disassembly 
Failure Mode and the Effect; the Risk of Failure 
Proc.ID 
ID 1.1 
ID 1.2 
ID 1.3 
ID 2.1 
ID 2.2 
ID 2.3 
ID 3.1 
ID 3.2 
ID 3.3 
ID 4.1 
ID 4.2 
ID 4.3 
ID 5.1 
ID 5.2 
ID 5.3 
Metal Cover scratched by slipped screwdriver 
As PC Metal Cover is removed, internal parts are cratched 
Floppy drive assy, screw removal can damage mother board 
Movie, illustrating floppy/HD assy. removal risks 
Can damage Floppy Drive if assy. Is dropped 
Can damage Hard Drive if assembly is dropped 
Can damage Hard Drive if assembly is dropped 
Severity 
Rating 
Detection 
Rating 
Occurrence 
Rating 
(1-10) (1-10) (1-10) 
3
3
3 
3
5
5 
4
2 
2 
2
2
2
9 
9
8
8 
1
1 
Failure Mode(s) and Effect(s) 
(b) 
The reason for this is because we would like to offer our users the opportunity 
not just to understand the method and the coded logic, but also to be able to enjoy 
the 3D interactive graphics, the digital videos, the color images, and most importantly 
the active code spreadsheets. Along with any other imaginable visualization, 
this can be executed and experimented with using their own data. 
In terms of statistical methods, our NPPI Tool Library has several statistical 
analysis tools to capture innovation opportunities at processes that are likely to drift 
and become out of control or processes that execute with random failure. 

This DFRA Study Prepared By 
OFRA Team 
Responsible Organization/Department 
Comments 
Paul G Ranky NJIT/MERC 
NJIT/MERC CFRA Team 
NJIT/MERC 
Severity 
Rating 
Detection 
Rating 
Occurrence 
Rating 
RPN (Risk 
Priority 
Number) 
Max. RPN Tooling 
Factor 
Clamping/ 
Fixturing 
Factor 
Skill Factor 
Any Other 
Factor You 
Define 
Accumulated 
RPN 
Risk 
Associated 
(1-10) (1-10) (1-10) 0.1-2.1=100% 0.1-2.1=100% 0.1-2.1=100% 0.1-2.1=100% 
3 
3
3
3 
2
2
2
2 
2 
5
5
8
8
9 
9
4 1
1 
18 
20 20 
0
0
0
0
0
0 
0 
0
0
0 
90 
90 
16 
48 48 
54 
54 
1.40 
1.40 
1.40 
1.40 
1.40 
1.40 
1.60 
1.20 1.20 
1.20 
1.20 
1.00 1.00 
1.00 
1.00 
1.00 
1.00 1.00 1.00 1.00 
33.60 
282.24 
96.77 
127.01 
0.00 
Low 
Low 
Low 
Low 
HIGH 
Mc 
Ha 
(c) 
(d) 
FIGURE 7 Continued 
OPEN-SOURCE COMPUTATIONAL STATISTICAL 185

186 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
For capturing such critical opportunities for innovation and process improvement, 
we use a range of control charts for drifting data analysis, Taguchi DOE 
(design of experiments) methods for developing the desired list of parameters for 
our engineering solutions in our requirements analysis method, and Weibull methods 
for process reliability analysis. As we progress, we plan to introduce further statistical 
and other tools to our NPPI Tool Library [56 – 70] . 
3.1.12 RFID APPLICATIONS 
Radio - frequency identifi cation (RFID) technologies are being adopted in the 
United States at a fast pace in pharmaceutical/assembly and packaging, in general 
manufacturing, warehousing, distribution, and global supply chain management. The 
market size for this technology is expected to rise from around $ 500 million in 2005 
to about $ 4 billion in 2010. In this section we outline some of the main application 
areas with a focus on the pharmaceutical applications. 
We also deal with the R & D opportunities and some digital pharmaceutical 
manufacturing systems with RFID information system modeling results . Furthermore 
we offer a generic factory assembly and tracking digital model for RFID 
integration, the most complicated task manufacturing systems engineers, industrial 
engineers, and IT experts have faced due to the mixed real - time as well as global 
traceability and messaging challenges one faces with RFID - tagged parts and shipments. 
RFID opportunities are great since with the appropriate IT infrastructure 
they help both major distributors and manufacturers as well as other logistics operations, 
such as in the health care system, defense industries, and others, dealing with 
complex, global supply chains in which products and product shipments must be 
traced and identifi ed in a noncontact, wireless fashion using a computer network, 
because of cost, security, or safety or because parts are subject to corrosion or medicine 
is subject to quality degradation. 
All of these requirements point to an automated, wireless - readable sensory - 
based identifi cation method and network that offers more functionalities and is 
signifi cantly “ smarter ” than the well - known bar code or the unifi ed product code 
(UPC). 
RFIDs are available as passive and/or active radio read/write sensor packages 
with active read (and often write) capabilities in relatively large areas (e.g., a large 
distribution center warehouse or a containership), all performed automatically, 
supervised by computers, and communicated in a wireless fashion over secure 
intranets. The attraction to a pharmaceutical assembly factory or a supply chain 
manager is that when the RFID network is integrated with the factories ’ material 
resource IT management systems, accurate information can be obtained on all 
tagged parts in close to real time throughout the entire supply chain. This can 
include the globally distributed factories as well as information about parts and 
assemblies during shipment, including in transit. This is why RFID represents great 
research and technology as well as huge business opportunities. 
We introduce here some of the most important engineering and information 
systems management principles and challenges that RFID researchers, implementers, 
and users should keep in mind when developing such systems and/or planning 

for such applications as well as offer an RFID digital factory integration model in 
UML [60 – 64, 70, 73] . 
3.1.13 RFID EXAMPLES 
To set the scene, consider a large storage house for a variety of medications or their 
distribution center with thousands of boxes, parts, and assemblies that range from 
low cost to high value, on occasion even highly sensitive technology or perishable 
drugs that must be kept in certain environmental conditions, such as temperature, 
humidity, or pressure, for the entire period of the shipment and/or production/packaging 
operations. 
Just - in - time delivery in an environment like this means that in order to build an 
order with variety or a new combination of treatments in a medical drug, every 
component must be in place on time and in good condition, which is a very diffi cult 
criterion to satisfy. 
Obviously supply chains are global these days, and shipments are typically made 
by a variety of means, including cargo ships, air, rail, and trucks; all of these can be 
late or can get in trouble because of the weather, traffi c, industrial disputes, or other 
reasons. Supply chain systems like this are very complex, because of the uncertainties 
in deliveries, parts and shipments are lost and/or stolen, goods get damaged 
during shipment, or the number of international ports and customs often take 
unpredictable time to check shipments with different levels of safety/security, and 
many other reasons. 
There are many valid reasons why wireless, computer - networked, sensory - based 
part identifi cation methods, tools, and technologies are being researched and 
deployed in industry. The application fi elds and opportunities are vast. The key 
driver is that even in chaotic, largely distributed, more stochastic than deterministic 
business environments, adaptive organizations and enterprises must react to 
demands quickly, else a competitor will take the business. Therefore they must 
reduce waste and improve effi ciency at all fronts. The most important aspect of this 
strategy is to know exactly what parts they have in stock, exactly where these parts 
are, and in what condition/state of assembly or preparedness they are. Furthermore, 
major distributors dealing with complex, global supply chains must be able to trace 
their shipments in detail because of cost, security, safety, quality degradation (as in 
the case of temperature - , humidity - , and/or shock - sensitive components or drugs), 
or other reasons. 
RFID technologies with the appropriate IT infrastructure help major distributors 
and manufacturers as well as other logistics operations such as the health care 
system, defense industries, and others deal with complex, global supply chains in 
which products and product shipments must be traced and identifi ed in a noncontact, 
wireless fashion using a computer network. 
All of the above - listed requirements point to an automated, wireless - readable, 
sensory - based identifi cation method and network that offer more functionalities, 
and are signifi cantly “ smarter ” than the well - known bar code or the UPC — hence 
the new popularity of RFID technology. 
RFID tags carry a serialized tag data construct. As an example, a 64 - bit class 0 
tag offered by a supplier includes 64 bits of total user memory on the tag itself, 
RFID EXAMPLES 187

188 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
including a unique serial number. This number is encoded by the manufacturer and 
uniquely identifi es up to 264 = 18,446,744,073,709,551,616 tagged items. 
RFIDs are available as passive and/or active radio read/write sensor packages 
with active read (and often write) capabilities in relatively large areas (like a large 
distribution center warehouse or a containership), all performed automatically, 
supervised by computers, and communicated in a wireless fashion. The attraction to 
an assembly factory or a supply chain manager here is that when the RFID network 
is integrated with the factories ’ material resource IT management systems, accurate 
information can be obtained on all tagged parts at close to real time, all throughout 
the entire supply chain. This can include the globally distributed factories as well as 
information about parts and assemblies in shipment/in transit. 
This is why RFID represents excellent research, technology, as well as big business 
opportunities. To illustrate this, consider research challenges such as remotely 
scanning and tracing products and parts in boxes on a cargo ship as it approaches 
national waters from international waters; tracing parts that are subject to corrosion 
and being used in agricultural or military equipment; medical drugs that are being 
counterfeited and repackaged and then shipped and imported illegally; or laptop 
computers that are dropped and damaged by accident. As a clear sign of the business 
opportunity, consider that according to a U.S. Department of Defense published 
presentation, RFID - enabled supply chain savings reached over U.S. $ 460 
million in 2004 and the projections for 2010 are in excess of $ 4 billion! 
3.1.14 RFID SYSTEM INTEGRATION MODELS FOR DIGITAL 
PHARMACEUTICAL MANUFACTURING AND ASSEMBLY 
SUPPLY CHAINS 
In the U.S. manufacturing and assembly industry, many of the RFID pilot projects 
focus on achieving 100% read rates at speeds set by the widely used bar code technology. 
The focus for these projects is to achieve proper tag placement on cases and 
pallets as well as the proper confi guration of pallets to enable 100% RFID tag 
read rates. This is a huge issue in pharmaceutical manufacturing and assembly 
in the fi ght to eliminate fake products and packages reaching the market! (See 
Figures 8 – 15 .) 
Based on the above - described requirements analysis, network planning, and 
server balancing reasoning, for our purposes, we have decided to follow a simple 
but powerful network architecture. In this architecture, we have included subnetworks. 
In terms of the way OPNET IT - Guru handles subnetworks, a subnetwork 
contains other network objects and abstracts them into a single object. A subnetwork 
can encompass a set of nodes and links to represent a physical grouping of 
objects (this can be a local - area network of CNC machines or robot PC controllers) 
or it can contain other subnetworks (e.g., including the material - handling system 
control of the line) [32, 34, 63, 66, 69 – 77] . 
Subnetworks within other subnetworks form the hierarchy of the network model. 
This hierarchy can then be extended as required to model the structure of the 
network. A subnetwork is considered the parent of the objects inside of it, and 
the objects are the children of the subnetwork. The highest level subnetwork in the 
network hierarchy does not have a parent, and therefore it is the top subnetwork, 

FIGURE 8 UML model segment illustrating the way the stock fi le is integrated with the 
routing and tooling fi les, assuming that all parts and all tools are RFID tagged. UML models 
like this should be used prior to any implementation work to assess requirements, technology 
needs, and RFID integration challenges with the rest of the factory ’ s IT infrastructure. 
FIGURE 9 Simulation network for distributed pharmaceutical manufacturing systems and 
their warehouses in U.S., Europe, India, and Asia. Model focuses on information and data 
management, the way the servers can cope with the task of tracking pharmaceutical product, 
and RFID data on a world wide basis. As a modeling tool we use OPNET, a professional 
network simulation tool. 
RFID SYSTEM INTEGRATION MODELS FOR DIGITAL PHARMACEUTICAL 189

FIGURE 10 Segment of simulation model illustrating corporate headquarters in New York. 
This is where we have our main servers in our distributed system. Modeling tool is OPNET. 
FIGURE 11 Segment of simulation model illustrating New Delhi campus network. This is 
where the business process outsourcing team and related servers in our distributed system 
are located. Modeling tool is OPNET. 

FIGURE 12 Pharmaceutical company portals as a wireless network of a pharmaceutical 
manufacturing system. The power of the model is that we can simulate a shop - fl oor request, 
comment, or warning throughout the entire international network of globally distributed 
pharmaceutical companies, with all important functions and processes. This means that before 
any pharmaceutical manufacturing system is actually built, we can simulate the entire system 
in the digital domain, saving huge expense and time. Modeling tool is OPNET. 
FIGURE 13 Simulation diagram illustrating and confi rming that the network system design 
from an ATM variation – response time point of view can cope with the demand. Modeling 
tool is OPNET. 

192 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
or global subnetwork. Subnetworks can be created and interconnected within this 
top level or within other subnetworks. Subnetworks provide a powerful mechanism 
for manipulating complex networks by breaking down the system ’ s complexity 
through abstraction. 
Since in our pharmaceutical network simulation models we deal with packets, let 
us explain a few aspects of packet formats. Packets carry information and can be 
sent between transmitters and receivers. In our example, packets can carry robot 
programs when uploaded from the design/programming offi ce servers to the robot 
lines and then to the individual CNCs, or robots, or parts of them if there is a need 
for an update, edit, quality control, production control, maintenance, and other data. 
(Packets can include mission - critical, “ panic ” related real - time data between the 
robot controller PCs and the line servers.) 
Packets are data structures consisting of storage areas called fi elds and can either 
be formatted or unformatted. Formatted packets have fi elds designed according to 
a packet format which specifi es the packets ’ fi eld names, data types, sizes, and default 
values. Formatted packets can be read by corresponding communication protocols 
only. Unformatted packets have no predefi ned fi elds. In IT - Guru, packet formats 
are predefi ned and typically named according to the model in which they are 
intended to be used. 
FIGURE 14 Simulation diagram illustrating and confi rming that the server balancing 
aspects of the network system design can cope with the demand. Modeling tool is OPNET. 

3.1.15 EVALUATION OF NETWORK SIMULATION RESULTS 
The goal of most simulation scenarios is to evaluate some aspect of a system ’ s 
behavior or performance and to quantify, typically in terms of statistics, the results 
and then use the results for decisions. This requires a simulation environment with 
software tools that provide insight into a model ’ s dynamic operation. 
Based on IT - Guru ’ s in - depth analysis, the pharmaceutical manufacturing system 
network engineering analyst can collect object, scenariowide, and global statistics 
as follows: 
• Object statistics are collected from individual objects. They allow the network 
engineering analyst to evaluate the performance of specifi c network nodes or 
links (a single hub ’ s Ethernet delay or a server balancing change, as in our 
example). 
• Scenariowide object statistics are collected from all relevant objects in a network 
(e.g., Ethernet delay for every node). They allow the network engineering 
analyst to easily monitor the performance of all objects of a specifi c type. 
EVALUATION OF NETWORK SIMULATION RESULTS 193 
FIGURE 15 Simulation diagram illustrating and confi rming that the network system design 
from an object variation – response time point of view can cope with the demand. As an 
example, this is important if a pharmaceutical manufacturing system line manager in India 
wants to notify a manager in New York by sending an image object, a sound object, or a 
multimedia object of a machine in the line for quality evaluation. Modeling tool is OPNET. 

194 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
• Global statistics are collected from the entire network. They represent results 
that apply to the network as a whole (such as global end - to - end delay) and let 
the designers and management analyze aspects of the network ’ s overall 
performance. 
• More specifi cally, IT - Guru offers the following types of statistics when analyzing 
networks: 
Queue size 
Available space 
Overfl ow occurrences 
Delay 
Interarrival times 
Packet sizes 
Throughput 
Utilization 
Error rates 
Collisions 
Application - specifi c statistics defi ned by a model developer 
Because there are many possible statistics to collect, the data fi les would quickly 
grow past practical use if the simulation program recorded them all. Therefore, the 
analyst must specifi cally select the statistics that are valuable for the particular study 
before running a simulation [71 – 79]. 
3.1.16 SUMMARY 
In this chapter, we have presented the foundations of an analytical and simultaneously 
computational lean and fl exible pharmaceutical manufacturing system design 
approach based on total quality standards. We have discussed why this approach is 
essential for pharmaceutical product, process, and manufacturing system designs. 
As illustrated, based on simulation results, using the plotted graphs and screens, 
management can easily evaluate different design alternatives, machine and human 
behavior models, control systems, sensory feedback processing, and the need of a 
balanced server architecture, and even investigate “ what if ” scenarios further, 
without committing to major upfront investment. 
We can clearly state that the time has come when pharmaceutical manufacturing 
systems can be designed and built in an entirely digital domain, saving huge amounts 
of capital and other related cost, and simultaneously increasing quality. 
3.1.17 COMPLIMENTARY VIDEO ON DVD 
To show real - world high - technology examples of pharmaceutical product, process, 
manufacturing, assembly, and packaging system designs, in action, something we 
cannot do in static, printed books, we have created a supplementary video, in high 
defi nition, and compressed onto a DVD. This professionally edited DVD supports 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 

this chapter as an independent, self - contained publication illustrating advanced 
pharmaceutical and medical product, process, and manufacturing system designs, 
related quality assurance processes and solutions, and others, explained by industry 
experts. To fi nd out more about this DVD, refer to Ranky, P. G., Ranky, G. N., and 
Ranky, R. G. (2006), Design Principles and Examples of Pharmaceutical Manufacturing 
Systems (Product, Process, Lean and Flexible Manufacturing, Assembly and 
Packaging System Designs) , Video on DVD, available: www.cimwareukandusa. 
com . 
REFERENCES 
1. Ranky , P. G. , A generic, analytical method to assess process - related risk with case studies, 
The Project Management Institute (PMI) Risk SIG and the Institute for International 
Research (IIR), paper presented at the Annual US National Project Risk Symposium, 
Houston, TX, May 22 – May 26, 2006. 
2. Ashley , S. ( 2004 ), Penny - wise smart labels , Sci. Am . 291 ( 2 ), 30 – 31 . 
3. Ashton , P. , and Ranky , P. G. , ( 1999, Feb. ), An advanced concurrent engineering research 
toolset and its application at Rolls - Royce motor cars, ADAM (Adv. Des. Manufacturing), 
available: http://www.cimwareukandusa.com , listed and indexed by the Association of 
Research Libraries , Washington, DC, and the Edinburgh Engineering Virtual Library, 
United Kingdom, Vol. 1. 
4. Bradbrook , R. ( 2004 ), Wal - Mart and RFID, folding carton industry , Printing News , 31 ( 4 ), 
30 – 33 . 
5. Bradbrook , R. ( 2004 ), Procter and Gamble aim to be among the world leaders in RFID 
implementation , Int. Paper Board Ind ., 47 ( 8 ), 20 – 23 . 
6. Brzozowski , C. ( 2004 ), Tags, tickets & labels: New technologies emerge , Printing News , 
153 ( 3 ). 
7. Ranky , P. G. , An integrated PM approach, including: Process modeling, requirements 
analysis, risk analysis, statistical tools and 3D multimedia, presented at the Project Management 
Institute (PMI), New Jersey Chapter. Jan. 2006. 
8. Ranky , P. G. , Focus on RFID (radio frequency identifi cation) methods, technologies and 
education, presented as part of the NCME Mission (National Center for Manufacturing 
Education), sponsored by NSF (National Science Foundation, USA) and industry, Jan. 
2006. 
9. Flaherty , M. , Ranky , P. G. , Ranky , M. F. , Sands , S. , and Stratful , S. ( 1999, Mar. ), An 
engineering multimedia approach to servo pneumatic positioning, ADAM (Adv. Des. 
Manufacturing), available: http://www.cimwareukandusa.com , listed and indexed by the 
Association of Research Libraries, Washington DC, and the Edinburgh Engineering 
Virtual Library, United Kingdom, Vol. 1. 
10. Glidden , R. , Bockorick , C. , etal . ( 2004 ), Design of ultra - low - cost UHF RFID tags for 
supply chain applications , IEEE Commun. Mag . 42 ( 8 ), 140 – 151 . 
11. Graham-Rowe , D. (2004), Tags to banish forgetfulness , New Scientist , 183 ( 2460 ), 19 . 
12. Graham - Rowe , D. ( 2004 ), Who ’ s keeping tabs on your tags? New Scientist , 183 ( 2462 ), 
22 . 
13. Ho , K. L. , and Ranky , P. G. ( 1999, Mar. ), An object oriented approach to fl exible 
conveyor system design, ADAM (Adv. Des. Manufacturing), available: http://www. 
cimwareukandusa.com , listed and indexed by the Association of Research Libraries, Washington 
DC, and the Edinburgh Engineering Virtual Library, United Kingdom, Vol. 1. 
REFERENCES 195

196 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
14. Knights , P. F. , Henderson , E. , and Daneshmend , L. K. ( 2004 ), Drawpoint control using 
radio frequency identifi cation systems , CIM Bull ., 89 ( 1003 ), 53 – 58 . 
15. Loose , D. C. , and Ranky , P. G. ( 2007 ), A Case - Based Introduction to IBM ’ s Telematics 
Solutions , interactive multimedia eBook with 3D objects, text, and videos in a browser - 
readable format on CD - ROM/intranet, available: http://www.cimwareukandusa.com , 
CIMware USA, Inc., and CIMware Ltd., United Kingdom; multimedia design and programming 
by P. G. Ranky and M. F. Ranky. 
16. Mongeon , D. G. ( 2005, Feb. ), RFID: A tool for the 21st century distribution, USA Department 
of Defense Conference Presentation, RFID Media Briefi ng, Washington DC. 
17. Nadler , S. F. , Ranky , P. G. , and Ranky , M. ( 2002 – 2003 ), A 3D multimedia approach to the 
diagnosis of low back pain (Vol. 1, 18 - and 40 - year - old males), interactive 3D multimedia 
presentation on CD - ROM with off - line Internet support (650 Mbytes, approx. 150 interactive 
screens, 50 minutes of digital videos, 3D internal and external body tour, animation, 
and 3DVR objects), by CIMware (IEE and IMechE Approved Professional Developer); 
also in Multimedia design and programming by P. G. Ranky and M. F. Ranky. 
18. Ranky , G. N. , and Ranky , P.G. ( 2005 ), Japanese service robot R & D trends and examples , 
Ind. Robot , 32 ( 6 ), 460 – 467 . 
19. Ranky , P. G. , Caudill , R. J. , Limaye , K. , Alli , N. , ChamyVelumani , S. , Bhatia , A. , and 
Lonkar , M. ( 2002, May ), A Web - enabled virtual disassembly manager (web - VDM) for 
electronic product/process designers, disassembly line managers and operators, a UML 
(Unifi ed Modeling Language) model of our generic digital factory, and some of our 
electronic support system analysis tools, ADAM with IT (Adv. Des. Manufacturing), 
available: http://www.cimwareukandusa.com , listed and indexed by the Association of 
Research Libraries, Washington DC, and the Edinburgh Engineering Virtual Library, 
United Kingdom, Vol. 3. 
20. Ranky , P. G. ( 2006 ), Introduction to RFID — Radio frequency identifi cation methods and 
solutions , Assembly Automation , 26 ( 1 ), 28 – 33 . 
21. Ranky , P. G. ( 1999, Apr. ), New trends in fl exible, lean and agile manufacturing cells and 
systems, ADAM (Adv. Des. Manufacturing) , available: http://www.cimwareukandusa.com , 
listed and indexed by the Association of Research Libraries, Washington DC, and the 
Edinburgh Engineering Virtual Library, United Kingdom, Vol. 1. 
22. Ranky , P. G. ( 2002 ), A method for planning industrial robot networks for automotive 
welding and assembly lines , Ind. Robot: Int. J ., 29 ( 6 ), 530 – 537 . 
23. Ranky , P. G. ( 2003 ), A real - time manufacturing/assembly system performance evaluation 
and control model with integrated sensory feedback processing and visualization , Assembly 
Automation . 
24. Ranky , P. G. ( 2003 ), A simulation method and distributed server balancing results of networked 
industrial robots for automotive welding and assembly lines , Ind. Robot: Int. J ., 
30 ( 2 ), 192 – 197 . 
25. Ranky , P. G. ( 2002 ), Advanced digital automotive sensor applications , Sensor Rev.: Int. J ., 
22 ( 3 ), 213 – 217 . 
26. Ranky , P. G. ( 2003 ), Advanced machine vision systems and application examples , Sensor 
Rev.: Int. J ., 23 ( 3 ), 242 – 245 . 
27. Ranky , P. G. ( 2003 ), Collaborative synchronous robots serving machines and cells , Ind. 
Robot: Int. J ., 30 ( 3 ), 213 – 217 . 
28. Ranky , P. G. ( 2004 ), Digital, Internet - enabled assembly line and factory modeling , Assembly 
Automation , 24 ( 3 ), 247 – 253 . 
29. Ranky , P. G. ( 2004 ), Novel automated inspection methods, tools and technologies , Assembly 
Automation: Int. J ., 23 ( 3 ), 252 – 257 . 

30. Ranky , P. G. ( 2003 ), Reconfi gurable robot tool designs and integration applications , Ind. 
Robot: Int. J ., 30 ( 4 ), 338 – 344 . 
31. Ranky , P. G. ( 2002 ), Smart sensors , Sensor Rev.: Int. J ., 22 ( 4 ), 312 – 318 . 
32. Ranky , P. G. ( 1999, Jan. ,) Some generic algorithmic solutions to the problem of dynamic 
scheduling in fl exible manufacturing systems that operate globally, ADAM (Adv. Des. 
Manufacturing), available: http://www.cimwareukandusa.com , listed and indexed by the 
Association of Research Libraries, Washington DC, and the Edinburgh Engineering 
Virtual Library, United Kingdom, Vol. 1. 
33. Ranky , P. G. ( 2001 ), Trends and R & D in virtual and robotized product disassembly , Ind. 
Robot , 28 ( 6 ), 454 – 456 . 
34. Ranky , P. G. ( 2000, May ), Some analytical considerations of engineering multimedia 
system design within an object oriented architecture , Int. J. CIM , 13 ( 2 ), 204 – 214 . 
35. Ranky , P. G. , and ChamyVelumani , S. ( 2003 ), A method, a tool (CORA), and application 
examples for analyzing disassembly user interface design criteria , Int. J. CIM , 16 ( 4 – 5 ), 
317 – 325 . 
36. Ranky , P. G. , and ChamyVelumani , S. ( 2003 ), An analytical approach, a tool (DFRA) and 
application examples for assessing process - related failure risks , Int. J. CIM , 16 ( 4 – 5 ), 
326 – 333 . 
37. Ranky , P. G. , and Nadler , S. F. , A novel multimedia approach to low back pain diagnosis 
with internal and external 3D interactive body tours, paper presented at the 29th Annual 
Northeast Bioengineering Conference, New Jersey Institute of Technology, University 
Heights, Newark, NJ, Mar. 2003. 
38. Ranky , P. G. , and Nadler , S. F. , A new, Web-enabled multimedia approach with 3D virtual 
reality internal and external body tours to support low back pain diagnosis, paper presented 
at the 4th Annual Faculty Best Practices Showcase in Kean University, NJ, Mar. 2003. 
39. Ranky , P. G. , and Ranky , M. F. ( 2000 ), A Dynamic Operation control algorithm with 
multimedia objects for fl exible manufacturing systems , Int. J. CIM , 13 ( 2 ), 245 – 263 . 
40. Ranky , P. G. , Lonkar , M. , and ChamyVelumani , S. ( 2003 ), eTransition models of collaborating 
design and manufacturing enterprises , Int. J. CIM , 16 ( 4 – 5 ), 255 – 266 . 
41. Ranky , P. G. , Morales , C. , and Caudill , R. J. , Lean Disassembly line layout, process and 
network simulation models and cases, based on real - world data, paper presented at the 
IEEE (USA) International Symposium on Electronics and the Environment and the 
IAER Electronics Recycling Summit, Boston, MA, May 19 – 22, 2003. 
42. Ranky , P. G. , Ranky , G. N. , and Ranky , R.G. ( 2006 ), Examples of pharmaceutical product/ 
process/manufacturing/assembly and packaging system designs, video on DVD, available: 
www.cimwareukandusa.com . 
43. Ranky , P. G. , Subramanyam , M. , Caudill , R. J. , Limaye , K. , and Alli , N. , A dynamic scheduling 
and balancing method and software tool for lean and reconfi gurable disassembly 
lines, paper presented at the IEEE (USA) International Symposium on Electronics and 
the Environment and the IAER Electronics Recycling Summit, Boston, MA, May 19 – 22, 
2003. 
44. Ranky , P. G. , 3D engineering multimedia cases. A customizable 3D Web - enabled library 
with reusable objects, paper presented at the ASEE (American Society of Engineering 
Educators) Mid - Atlantic Conference, Kean University, NJ, Apr. 2003. 
45. Ranky , P. G. , A 3D multimedia approach to biomedical engineering: Low back analysis, 
paper presented at the ASEE, American Society of Engineering Educators, U.S. National 
Meeting, Biomedical Engineering Division, Nashville, TN, June 2003. 
46. Ranky , P. G. , Ranky , G. N. , and Ranky , R. G. ( 2006 ), Design principles and examples of 
pharmaceutical manufacturing systems (product, process, lean & fl exible manufacturing, 
REFERENCES 197

198 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
assembly and packaging system designs), video on DVD, available: www.cimwareukandusa.
com . 
47. Ranky , P. G. ( 2001 – 2006 ), A 3D multimedia case: Component oriented disassembly failure 
risk analysis, an interactive multimedia publication with 3D objects, text and videos in a 
browser - readable format on CD - ROM/intranet available: http://www.cimwareukandusa. 
com , CIMware USA, Inc., and CIMware Ltd., United Kingdom; Multimedia design and 
programming by P. G. Ranky and M. F. Ranky (published 6 volumes of this main title 
with different risk analysis challenges explained). 
48. Ranky , P. G. ( 2001 – 2005 ), A 3D multimedia case: component oriented disassembly user 
requirements analysis, an interactive multimedia eBook publication with 3D objects, text 
and videos in a browser - readable format on CD - ROM/intranet available: http://www. 
cimwareukandusa.com , CIMware USA, Inc., and CIMware Ltd., United Kingdom, Multimedia 
design and programming by P. G. Ranky and M. F. Ranky (published 7 volumes 
of this main title with different requirements analysis challenges explained). 
49. Ranky , P. G. , A 3D Web collaborative concurrent automotive engineering Method Based 
on our “ distributed digital factory ” and “ digital car ” models, paper presented at the 
Society of Automotive Engineers World Congress, Detroit, MI, Mar. 2003. 
50. Ranky , P. G. ( 2003 ), A 3D Web - enabled, case based learning architecture and knowledge 
documentation method for engineering, information technology, management, and 
medical science/biomedical engineering , Int. J. CIM , 16 ( 4 – 5 ). 346 – 356 . 
51. Ranky , P. G. , A Biomedical Engineering case with 3D lower back interactive virtual 
anatomy tours inside and outside the human body with automated post - test student 
assessment, paper presented at the ASEE (American Society of Engineering Educators) 
Mid - Atlantic Conference, Kean University, NJ, Apr. 2003. 
52. Ranky , P. G. , A new approach for teaching and learning about engineering process failure 
risk analysis with IE (industrial engineering) case studies, paper presented at the ASEE, 
American Society of Engineering Educators, US National Meeting, Industrial Engineering 
Division, Nashville, TN, June 2003. 
53. Ranky , P. G. , A novel 3D Internet - based multimedia method for teaching and learning 
about engineering management requirements analysis, paper presented at the ASEE, 
American Society of Engineering Educators, US National Meeting, Engineering Management 
Education Division, Nashville, TN, June 2003. 
54. Ranky , P. G. , An interactive 3D multimedia problem - based library for manufacturing 
engineering technology education with Internet support, paper presented at the ASEE, 
American Society of Engineering Educators, US National Meeting, Engineering Technology 
Division, Nashville, TN, June 2003. 
55. Ranky , P. G. ( 2003 – 2005 ), An introduction to alternative energy sources: Hybrid & fuel 
cell vehicles; an interactive multimedia eBook publication with 3D objects, text, and 
videos in a browser - readable format on CD - ROM/intranet, available: http://www. 
cimwareukandusa.com , CIMware USA, Inc., and CIMware Ltd.; United Kingdom, Multimedia 
design and programming by P. G. Ranky and M. F. Ranky, (2003 – 2005), Customer 
needs, wants & requirements analysis: Automotive exterior rearview mirror, an interactive 
multimedia eBook publication with 3D objects, text, and videos in a browser - readable 
format on CD - ROM/intranet, available: http://www.cimwareukandusa.com , CIMware 
USA, Inc., and CIMware Ltd., United Kingdom. 
56. Ranky , P. G. ( 2003 – 2005 ), An introduction to digital factory & digital telematic car modeling 
with R & D and industrial case studies, an interactive multimedia eBook publication 
with 3D objects, text, and videos in a browser - readable format on CD - ROM/intranet, 
available: http://www.cimwareukandusa.com , CIMware USA, Inc. and CIMware Ltd., 
United Kingdom, Multimedia design and programming by P. G. Ranky and M. F. Ranky. 

57. Ranky , P. G. (2005), An introduction to RFID, radio frequency identifi cation methods and 
applications, DVD video, available: www.cimwareukandusa.com (approximately 30 min). 
58. Ranky , P. G. ( 2005 – 2006 ), An introduction to RFID, radio frequency identifi cation 
methods and applications with a total quality management and control focus, interactive 
browser - readable 3D eBook, available: www.cimwareukandusa.com , (approximately 
30 min). 
59. Ranky , P. G. ( 1999, Apr. ), An object oriented system analysis and design method (CIMpgr) 
and an R & D case study, Adv. Des. Manufacturing, available: http://www.cimwareukan 
dusa.com , listed and indexed by the Association of Research Libraries, Washington DC, 
and the Edinburgh Engineering Virtual Library, United Kingdom, Vol. 1. 
60. Ranky , P. G. , Computerized engineering assessment method based on 3D interactive 
multimedia, That students enjoy, paper presented at the ASEE, American Society of 
Engineering Educators, US National Meeting, Continuing Professional Development 
Division, Nashville, TN, June 2003. 
61. Ranky , P. G. ( 1999 ), Design, manufacturing and assembly automation trends and strategies 
in China , Assembly Automation , 19 ( 4 ), 301 – 305 . 
62. Ranky , P. G. ( 2003 , Feb.), Designing a lean infrastructure; advanced machining cell design 
concepts, methods, architectures and cases , Manuf. Eng., J. IEE , 22 – 24 . 
63. Ranky , P. G. ( 2000 ), Engineering multimedia in CIM (computer integrated manufacturing) 
, Int. J. CIM , 13 ( 2 ), 169 – 171 . 
64. Ranky , P. G. ( 2003 ), eTransition in the multi - lifecycle CIM (computer integrated manufacturing) 
context , Int. J. CIM , 16 ( 4 – 5 ), 229 – 234 . 
65. Ranky , P. G. , Interactive 3D multimedia cases for engineering education with Internet 
support, ASEE, American Society of Engineering Educators, paper presented at the 
U.S. National Meeting, Computers in Education Division, Nashville, TN, June 2003. 
66. Ranky , P. G. , Interactive 3D multimedia cases for manufacturing engineering education 
with Internet support, paper presented at the ASEE, American Society of Engineering 
Educators, US National Meeting, Manufacturing Engineering Education Division, 
Nashville, TN, June 2003. 
67. Ranky , P. G. ( 2002, Dec. ), Introduction to concurrent engineering, an NSF (National 
Science Foundation, USA) sponsored Gateway Coalition streamed multimedia narrated 
web presentation, New Jersey Intitute of Technology, Public Research University, Newark, 
New Jersey . 
68. Ranky , P. G. ( 2003 – 2005 ), Key R & D and eTransition trends in US and international collaborative 
design & manufacturing enterprises, an interactive multimedia eBook publication 
with 3D objects, text, and videos in a browser - readable format on CD - ROM/intranet, 
available: http://www.cimwareukandusa.com , CIMware USA, Inc., and CIMware Ltd., 
United Kingdom; Multimedia design and programming by P. G. Ranky and M. F. Ranky . 
69. Ranky , P. G. ( 2000, Jan. ), Modular fi eldbus designs and applications , Assembly Automation 
, 20 ( 1 ), 40 – 45 . 
70. Ranky , P. G. ( 2003 ), Network simulation models of lean manufacturing systems in digital 
factories and an intranet server balancing algorithm , Int. J. CIM , 16 ( 4 – 5 ), 267 – 282 . 
71. Ranky , P. G. , Rapid prototyping cases for integrated design and manufacturing engineering 
education with 3D Internet support, paper presented at the ASEE, American Society 
of Engineering Educators, US National Meeting, Design in Engineering Education Division, 
Nashville, TN, June 2003. 
72. Roman , H. T. , and Ranky , P. G. ( 2003 – 2005 ), A case - based Introduction to Service robotics, 
an interactive multimedia eBook publication with 3D objects, text, and videos in a 
browser - readable format on CD - ROM/intranet, available: http://www.cimwareukandusa. 
REFERENCES 199

200 ANALYTICAL AND COMPUTATIONAL METHODS AND EXAMPLES 
com , CIMware USA, Inc., and CIMware Ltd., United Kingdom. Multimedia design and 
programming by P. G. Ranky and M. F. Ranky. 
73. Romero , C. , Department of logistics passive RFID initial implementation, paper presented 
at the USA Department of Defense Conference, RFID Media Briefi ng, 
Washington DC, Feb. 2005. 
74. Sangoi , R. , Smith , C. G. , et al . ( 2004 ), Printing radio frequency identifi cation (RFID) tag 
antennas using inks containing silver dispersions , J. Dispersion Sci. Technol . 25 ( 4 ), 
513 – 521 . 
75. Smith , K. , Enabling the supply chain, paper presented at the USA Department of Defense 
Conference, RFID Media Briefi ng, Washington, DC, Feb. 2005. 
76. Sugimoto , M. , Kusunoki , F. , Inagaki , S. , Takatoki , K. , and Yoshikawa , A. ( 2004 ), A system 
for supporting collaborative learning with networked sensing boards , Syst. Comput. Jpn ., 
35 ( 9 ), 39 – 50 . 
77. Wilke , P. , and Braunl , T. ( 2001 ), Flexible wireless communication network for mobile 
robot agents , Ind. Robot , 28 ( 3 ), 220 – 233 . 

201 
3.2 
ROLE OF QUALITY SYSTEMS AND 
AUDITS IN PHARMACEUTICAL 
MANUFACTURING ENVIRONMENT 
Evan B. Siegel and James M. Barquest 
Ground Zero Pharmaceuticals, Inc., Irvine, California 
Contents 
3.2.1 cGMP Regulations 
3.2.1.1 Duties of Quality Control Unit under cGMP Regulations 
3.2.2 Quality Assurance Function 
3.2.3 Quality Systems Approach 
3.2.4 Management Responsibilities 
3.2.5 Resources 
3.2.6 Manufacturing Operations 
3.2.6.1 Design, Develop, and Document Product and Processes 
3.2.6.2 Inputs 
3.2.6.3 Perform and Monitor Operations 
3.2.6.4 Address Nonconformities 
3.2.7 Evaluation Activities 
3.2.7.1 Trend Analysis 
3.2.7.2 Conduct Internal Audits 
3.2.7.3 Quality Risk Management 
3.2.7.4 Corrective and Preventive Actions 
3.2.7.5 Promote Improvement 
3.2.8 Transitioning to Quality Systems Approach 
3.2.9 Audit Checklist for Drug Industry 
3.2.9.1 Instructions for Using Audit Checklist 
References 
By regulation, appropriate practice, and common sense, quality assurance (QA) is 
a critical function in the pharmaceutical manufacturing environment. The need for 
an independent unit to audit and comment on the appropriate application of 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

202 ROLE OF QUALITY SYSTEMS AND AUDITS 
standard operating procedures, master batch records, procedures approved in 
product applications, and the proper functioning of the quality control (QC) unit is 
paramount. This helps assure that products are manufactured reliably, with adherence 
to approved specifi cations, and that current good manufacturing practices 
(cGMP) are maintained in conformance to regulation, both in the facility in general 
and the microenvironment of each product ’ s manufacturing sequence. 
Quality assurance personnel must have the appropriate training, experience, 
familiarization with the manufacturing facility and products, enforced independence 
from the production chain of command, and the ability to review adherence to 
procedures, policies, and agreed - upon approaches to manufacturing quality pharmaceuticals. 
This helps to provide both an environment and a manufactured product 
that can withstand Food and Drug Administration (FDA) inspection and support 
a fi rm ’ s reputation for quality products. 
The cGMP regulations establish requirements that are intended to provide a high 
level of assurance that the pharmaceutical products produced satisfy the strength, 
purity, potency, and other quality requirements established for the fi nished product 
to assure that it is fi t for its intended use. Manufacturers must establish a quality 
control unit that is responsible for many of the quality - related activities required 
by the regulations. These regulations have not been substantially updated since 1978. 
Since then, the science and practice of quality assurance have substantially evolved 
to include the development of quality management systems and risk management 
approaches to better assure product quality and fi tness for use. Pharmaceutical 
product manufacturers are increasingly interested in implementing a comprehensive 
quality management system (QMS) and employing risk management approaches 
because they allow them to apply newer quality management principles that they 
believe enable them to more effectively assure product quality and better allow 
harmonization with evolving international regulatory quality system requirements. 
The FDA has not changed the cGMP regulations but, as part of its Pharmaceutical 
CGMPs for the 21st Century Initiative, encourages this quality systems approach to 
cGMP compliance. 
This Chapter describes outlines and discusses the regulations applicable to the 
QA function and unit, structure, function, charter, and application of the unit in the 
pharmaceutical manufacturing environment. In addition, it discusses additional 
quality - related responsibilities that may result when manufacturers move toward a 
quality systems approach to quality that incorporates current quality system models 
to further improve quality and harmonize with international quality system 
requirements. 
The justifi cation for, and execution of, the QA audit are also described, including 
preparation, key items of interest, a typical checklist of the audit itself, corrective 
and preventive actions following the audit, and suggested measures for assuring 
successful operation of the unit. 
3.2.1 c GMP REGULATIONS 
The cGMP regulations for the manufacture of pharmaceutical products are contained 
in Parts 210 and 211 of Title 21 of the Code of Federal Regulations (CFR) 
[1] . These regulations, as well as guidance documents and other FDA documents 

pertaining to the regulation and FDA inspection of pharmaceutical product manufacturers, 
may be accessed on the FDA website at www.fda.gov . Part 210 specifi es 
the scope and applicability of the cGMP regulations and defi nes terms used in the 
regulations. Part 210 also indicates that the regulations establish “ minimum ” cGMP 
requirements and that products that are not manufactured under cGMP are adulterated. 
Adulterated products and the persons responsible for the adulteration are 
subject to regulatory action by the FDA. 
Part 211 contains specifi c good manufacturing practice requirements for fi nished 
pharmaceuticals and is divided into Subparts A – K as follows: 
A. Scope 
B. Organization and Personnel 
C. Buildings and Facilities 
D. Equipment 
E. Control of Components and Drug Product Containers and Closures 
F. Production and Process Controls 
G. Packaging and Labeling Control 
H. Holding and Distribution 
I. Laboratory Controls 
J. Records and Reports 
K. Returned and Salvaged Drug Products 
The cGMP regulations are written to address the primary potential sources of 
product variability. Subpart B establishes the quality control unit and the duties of 
that unit, establishes personnel requirements and addresses personnel practices 
(e.g. sanitation) intended to reduce the likelihood of product contamination. Subparts 
C and D establish requirements for buildings and facilities and equipment 
used in the manufacture, processing, packing, or holding of a drug product. Subparts 
E through H establish controls over the major processes associated with the production 
of a fi nished and packaged drug product that is ready to be shipped for distribution 
to users. Controls are established for incoming raw materials and components 
and continue through manufacturing, packaging, labeling, holding, and distribution 
of fi nished, packaged, labeled, and released drug product. Subpart I requires the 
establishment of scientifi cally sound and appropriate specifi cations, standards, sampling 
plans, and test procedures; requires instrument specifi cations and calibration; 
and establishes lot or batch testing and release requirements. Subpart J establishes 
documentation requirements including master and batch records, and Subpart K 
addresses the control and disposition of returned drug products and places limitations 
on the salvage of drug products that have been subjected to improper storage 
conditions (e.g., smoke, heat, fi re, moisture). 
3.2.1.1 Duties of Quality Control Unit under c GMP Regulations 
The cGMP regulations assign specifi c duties to the quality control unit. The unit is 
required to have the responsibility and authority to approve or reject all components, 
drug product containers, closures, in - process materials, packaging material, 
cGMP REGULATIONS 203

204 ROLE OF QUALITY SYSTEMS AND AUDITS 
labeling, and drug products and the authority to review production records to assure 
that no errors have occurred or, if errors have occurred, that they have been fully 
investigated. The responsibilities of the unit extend to approving or rejecting drug 
products manufactured, processed, packed, or held by contract manufacturers. The 
organization must assure that the quality control unit has adequate laboratory 
facilities for the testing and approval (or rejection) of components, drug product 
containers, closures, packaging materials, in - process materials, and drug products. 
In addition to duties associated with the approval of materials and fi nished products, 
the unit is also responsible for approving or rejecting all procedures or speci- 
fi cations impacting on the identity, strength, quality, and purity of the drug product. 
This includes review and approval of procedures for production and process control, 
including any changes to these procedures. These procedures, and the responsibilities 
and procedures applicable to the quality control unit within the organization, 
must be written and followed. 
All specifi cations, standards, sampling plans, test procedures, or other laboratory 
control mechanisms, including any changes, must be in writing and reviewed and 
approved by the quality control unit. 
Written procedures describing the handling of all written and oral complaints 
regarding a drug product are required. The quality control unit is responsible for 
reviewing any complaint involving the possible failure of a drug product to meet 
any of its specifi cations and, for such drug products, making a determination as to 
the need for an investigation in accordance with cGMP requirements. The review 
should include a determination if the complaint represents a serious and unexpected 
adverse drug experience, which is required to be reported to the FDA. A written 
record of each complaint must be maintained in a complaint fi le. 
3.2.2 QUALITY ASSURANCE FUNCTION 
The term quality is used in many industries and in everyday life and can have various 
meanings depending on context. For the purposes of discussion here quality means 
the product requirements or attributes that have a bearing on the product ’ s specifi ed 
requirements. Quality assurance activities are those processes and activities conducted 
to assure that a product or service consistently satisfi es its requirements and 
is fi t for its intended use. In the pharmaceutical manufacturing environment, this 
means the activities conducted to assure that the pharmaceutical product ’ s identity, 
strength, purity, potency, and other quality attributes conform to approved 
specifi cations. 
In the United States, cGMP requirements for the manufacture of drugs were 
established by regulation in 1978 and have not been substantially updated since 
then. The science and practice of quality assurance has substantially evolved since 
then to include the development of quality systems [2, 3] and risk management 
approaches [4] to better assure product quality and fi tness for use. Pharmaceutical 
product manufacturers are increasingly interested in implementing these approaches 
because they allow the manufactures to apply newer quality management principles 
that they believe enable them to more effectively assure product quality and 
better allow harmonization with evolving international regulatory quality system 
requirements. 

QUALITY SYSTEMS APPROACH 205 
3.2.3 QUALITY SYSTEMS APPROACH 
The systems approach to quality involves a coordinated approach to the management 
of quality - related activities as processes that work in conjunction with one 
another to provide assurance that the product meets its specifi ed requirements. It 
involves: 
• A management commitment to quality that is communicated throughout the 
organization 
• Identifying quality requirements using risk management and other methods as 
appropriate 
• Developing a quality policy, plan, objectives 
• Establishing an organizational structure with identifi ed responsibilities and 
authorities that allows quality objectives to be met 
• Providing the resources needed to meet quality objectives 
• Developing the required systems and processes 
• Establishing methods for the ongoing objective evaluation of the performance 
of systems and processes including quality auditing 
• Initiating corrective and preventive actions as needed to assure that quality 
objectives are consistently and reliably met 
The use of risk management techniques in identifying product requirements, establishing 
processes and process control and monitoring methods, evaluating quality 
data, identifying appropriate corrective and preventive actions to address quality 
problems, and for other quality - related activities can increase the overall effi ciency 
and effectiveness of the quality system. 
The FDA has recognized the value of and encourages a risk based quality systems 
approach for the manufacture of pharmaceutical products. This is refl ected in its 
Pharmaceutical CGMPs for the 21st Century Initiative. In association with this initiative 
the FDA has published reports and guidance documents that collectively 
provide information that can be used by pharmaceutical product manufacturers in 
implementing a quality systems and risk management approach to pharmaceutical 
cGMP regulations compliance [5 – 8] . In implementing this initiative, the FDA has 
made it clear that it does not impose new regulatory requirements on manufacturers. 
The FDA has provided information and guidance that is intended to serve as a 
bridge between the 1978 regulations and current quality systems by explaining how 
manufacturers implementing such systems can do so in full compliance with the 
cGMP regulations. This approach differs from that used by the FDA when it updated 
the cGMP regulations for medical devices to employ a quality systems approach. 
The 1996 Quality System Regulation updated the GMP requirements for fi nished 
medical device manufacturers to reduce the risk of inadequate device design and 
to harmonize them with international quality system standards that were in effect 
at that time [9] . These international standards have since been updated [10] ; however 
the device quality system regulation remains consistent with modern quality system 
models. 
In a modern quality system, the organizational unit responsible for quality - 
related activities within the organization generally has a central role in the 

206 ROLE OF QUALITY SYSTEMS AND AUDITS 
development and management of the overall quality system. These activities can 
include quality control, quality assurance, quality planning, and quality improvement. 
The cGMP regulations do not defi ne or employ these terms, but the activities 
the regulations assign to the quality control unit fall within these defi nitions as currently 
defi ned [2, 8, 11] . 
Current quality system models involve quality - related activities and terms that 
are not included in the cGMP regulations. Further, quality as a professional discipline 
is evolving. It is, therefore, important for organizations adopting a quality 
systems approach to unambiguously defi ne the terms and quality concepts they will 
be using and to include these defi nitions as appropriate in training all staff in the 
organization who will be involved in quality - related activities. This will help assure 
effective communication throughout the organization and with vendors and others 
(e.g., regulatory agencies, third - party auditors) who interact with the organization 
on quality - related matters. Regulatory defi nitions should be recognized, and the use 
of nonstandard or outdated terminology should be avoided to the extent possible. 
Incorporating terms and defi nitions by reference from pertinent standards and FDA 
guidance documents may be helpful. FDA guidance on the quality systems approach 
to pharmaceutical cGMP regulations (pharmaceutical QS guidance) includes the 
following defi nitions: 
Quality Assurance (QA) Proactive and retrospective activities that provide con- 
fi dence that requirements are fulfi lled. 
Quality Control (QC) The steps taken during the generation of a product or 
service to ensure that it meets requirements and that the product or service is 
reproducible. 
Quality Management (QM) Accountability for the successful implementation 
of the quality system. 
Quality System (QS) Formalized business practices that defi ne management 
responsibilities for organizational structure, processes, procedures, and 
resources needed to fulfi ll product/service requirements, customer satisfaction, 
and continual improvement. 
Quality Unit (QU) A group organized within an organization to promote quality 
in general practice. 
The FDA notes in its pharmaceutical QS guidance document that many current 
quality system concepts correlate very closely with the cGMP regulations and that 
the activities required by the regulations are generally consistent with a quality 
systems approach. In this and other guidance documents, the FDA uses the 
term quality unit rather than quality control unit as defi ned in the cGMP regulations 
to refer to the organizational unit with responsibility for quality - related 
activities. In a modern quality systems model these quality - related activities may go 
beyond, but are not necessarily inconsistent with, those required by the cGMP 
regulation. 
Use of the term quality unit is consistent with current quality management system 
models [2, 10] , which are intended to assure that the various operations associated 
with all systems are appropriately planned, approved, conducted, and monitored, 
and because the cGMP regulations specifi cally assign the QU the authority to create, 

QUALITY SYSTEMS APPROACH 207 
monitor, and implement a quality system. The FDA cautions that such activities do 
not substitute for, or preclude, the daily responsibility of manufacturing personnel 
to build quality into the product. The FDA has specifi cally indicated that the overarching 
philosophy articulated in both the cGMP regulations and in robust modern 
quality systems is that quality should be built into the product, and testing alone 
cannot be relied on to ensure product quality. 
Other cGMP - assigned responsibilities of the QU that are consistent with modern 
quality system approaches include the following: 
• Ensuring that controls are implemented and completed satisfactorily during 
manufacturing operations 
• Ensuring that developed procedures and specifi cations are appropriate and 
followed, including those used by a fi rm under contract to the manufacturer 
• Approving or rejecting incoming materials, in - process materials, and drug 
products 
• Reviewing production records and investigating any unexplained 
discrepancies 
The FDA has stressed that the release of the pharmaceutical QS guidance document 
does not impose new regulatory requirements on manufacturers but encourages 
manufactures to adopt a quality systems approach to cGMP compliance 
because of the potential benefi ts. An appropriately designed and implemented 
quality system can do the following: 
• Reduce the number of (or prevent) recalls, returned or salvaged products, and 
defective products entering the marketplace 
• Harmonize the cGMP regulations to the extent possible with other widely used 
quality management systems, which is desirable because of the globalization of 
pharmaceutical manufacturing, and the increasing prevalence of drug – device 
and biologic – device combination products 
• When coupled with manufacturing process and product knowledge and the use 
of effective risk management practices, handle many types of changes to facilities, 
equipment, and processes without the need for prior approval regulatory 
submissions 
• Potentially result in shorter and fewer FDA inspections by lowering the risk of 
manufacturing problems 
• Provide the necessary framework for implementing quality by design (building 
in quality from the product development phase and throughout a product ’ s life 
cycle), continual improvement, and risk management in the drug manufacturing 
process 
This suggests that even without making changes in the cGMP regulations, the 
FDA may be looking at them from a “ new ” quality systems perspective. The regulations 
include terms such as adequate and appropriate that may be subject to interpretation 
based on relevant technical or scientifi c capabilities and state - of - the - art 
knowledge. As these improve, the interpretation of what is adequate or appropriate 

208 ROLE OF QUALITY SYSTEMS AND AUDITS 
can change as well. Practically, most manufacturers are more than willing to adopt 
methods that can improve the quality and safety of their pharmaceutical products 
because it is cost effective in the long run [11] but may be reluctant to do so for 
fear of being considered out of compliance with the cGMP regulations. Current 
FDA efforts in this regard should serve to allay manufacturers ’ concerns in this 
area. 
The major elements of the quality system model described in the FDA ’ s 
pharmaceutical QS guidance document are consistent with existing quality 
system standards. These elements are as follows: 
• Management responsibilities 
• Resources 
• Manufacturing operations 
• Evaluation activities 
3.2.4 MANAGEMENT RESPONSIBILITIES 
Current quality system models assign management a major role in the deployment 
and operation of a successful quality system. In such systems, major management 
responsibilities include the following: 
• Provide leadership by establishing a commitment to quality that is supported 
by all levels of management and is communicated throughout the 
organization 
• Create an organizational structure with clearly defi ned responsibilities and 
authorities to perform quality functions associated with achieving quality 
objectives 
• Building and documenting a quality system to meet specifi ed quality and 
regulatory requirements and achieve quality objectives 
• Establishing a quality policy and objectives, and quality plans that are aligned 
with the organization ’ s strategic plans and communicate this throughout the 
organization 
• Reviewing the system by establishing appropriate accountability systems within 
the organization to monitor and report quality data and system status to management 
and assure that appropriate corrective and preventive actions are 
taken in response to quality problems using effective change control procedures 
and documented 
The cGMP does not specifi cally assign management responsibility for these 
actions, although actions of this nature are required by the regulation. Table 1 from 
the pharmaceutical QS guidance document shows this relationship. 
Under a comprehensive quality system the QU can expect an expanded and more 
visible role within the organization with greater accountability to and interaction 
with upper management. The QU should ideally be independent of the other organizational 
units to assure clear delineation of responsibility and authority and avoid 
confl icts. In certain instances, such as auditing, independence or objectivity is central 

to the effectiveness of the audit process, and auditors therefore should not have 
direct responsibility over the areas being audited. 
The cGMP regulations do not specify how the QU should be integrated into the 
overall organization but, in general, the QU should be structured to refl ect management 
’ s strong commitment to quality and to facilitate achieving quality objectives. 
The structure (e.g., organizational relationship to other organizational units, reporting 
relationships) should provide clear lines of responsibility and authority that 
support the production, quality, and management activities necessary to achieve 
quality objectives. Different organizations may accomplish this in different ways; 
however, experience has been that placement of the quality function on the same 
level within the organizational hierarchy as other major organizational units (e.g., 
production) sends a clear message both within and outside the organization that 
top management has a strong commitment to quality. 
The cGMP regulations require quality - related activities to be conducted during 
all phases of manufacturing from the acceptance of raw materials through batch 
release, packaging, and labeling. The regulations also require that all personnel, 
including those engaged in quality - related activities, have suffi cient education, training, 
and experience or any combination thereof to enable them to perform their 
assigned functions. In a quality systems approach to cGMP compliance, the role of 
quality personnel can be signifi cantly expanded to include internal quality auditing, 
expanded review and analysis of quality data, investigation of nonconformance, root 
cause analysis, risk analysis, and other quality - related activities. Many of these activities 
are likely to be conducted with personnel from other organizational elements 
such as manufacturing, material control, facilities, product development, or engineering 
staff. Quality staff should have suffi cient scientifi c and technical knowledge 
and training (e.g., statistical methods, risk analysis) and knowledge of the product 
and manufacturing processes to effectively perform their assigned functions 
TABLE 1 21 CFR cGMP Regulations Related to Management Responsibilities 
Quality System Element Regulatory Citations 
1. Leadership 
2. Structure Establish quality function: § 211.22(a) [see defi nition 
§ 210.3(b)(15)[ 
Notifi cation: § 211.180(f) 
3. Build QS QU procedures: § 211.22(d) 
QU procedures, specifi cations: § 211.22(c), with reinforcement 
in: § § 211.100(a), 211.160(a) 
QU control steps: § 211.22(a), with reinforcement in 
§ § 211.42(c), 211.84(a), 211.87, 211.101(c)(1), 211.110(c), 
211.115(b), 211.142, 211.165(d), 211.192 
QU quality assurance; review/investigate: § § 211.22(a), 
211.100(a – b) 211.180(f), 211.192, 211.198(a) 
Record control: § § 211.180(a – d), 211.180(c), 211.180(d), 
211.180(e), 211.186, 211.192, 211.194, 211.198(b) 
4. Establish policies, 
objectives, and plans 
Procedures: § § 211.22(c – d), 211.100(a) 
5. System review Record review: § § 211.100, 211.180(e), 211.192, 211.198(b)(2) 
MANAGEMENT RESPONSIBILITIES 209

210 ROLE OF QUALITY SYSTEMS AND AUDITS 
and competently interact with personnel from other organizational elements as 
necessary. 
3.2.5 RESOURCES 
The appropriate assignment of resources is essential to the success of any endeavor, 
and this is particularly critical in a pharmaceutical manufacturing environment. 
Inadequate staffi ng, training, manufacturing equipment and facilities, environmental 
controls, analytical equipment, and other resources can be sources of variability 
leading to the production of product that does not meet specifi ed requirements. 
Modern quality system standards specifi cally address the issue of resources by 
requiring the organization to determine and provide the human, infrastructure, and 
work environment resources necessary for the quality system. The cGMP regulations 
address the resource issue in provisions that are intended to assure the 
adequacy of personnel (including consultants), manufacturing facilities including 
contract facilities, equipment, and laboratory facilities. The QU has signifi cant 
responsibility in this regard. 
The FDA, in its pharmaceutical QS guidance document, discusses the need for 
adequate resources in developing, implementing, and managing a quality system 
that complies with the cGMP regulations. Management is responsible for identifying 
resource requirements and providing resources accordingly, including providing 
training that is appropriate to the assigned activities. Personnel should understand 
the impact of their activities on their assigned duties and be familiar with cGMP 
requirements and the organization ’ s quality system. This is consistent with the 
generally accepted idea that a culture of quality within an organization requires 
personnel to understand quality concepts, the organization ’ s quality and regulatory 
objectives, and how their assigned activities contribute to the achievement of these 
objectives and fi t into the overall quality system. Management should establish a 
working environment that encourages problem solving and communication in 
identifying and acting upon quality - related issues. While the provision of resources 
is generally considered a management function, the QU and other organizational 
units should be involved in the identifi cation of the resources required to achieve 
quality objectives, including regulatory compliance, the assessment of the adequacy 
of existing resources, evaluating the effect of personnel, facility, product, process, 
regulatory, and other changes on resource needs, and generally providing management 
the information needed to make necessary and appropriate resource 
decisions. 
Current quality system models employ a risk - based and data - driven approach to 
the development of QS system requirements to assure their adequacy. The FDA 
notes that the cGMP regulations place as much emphasis on processing equipment 
as testing equipment and contain specifi c requirements for the qualifi cation, calibration, 
cleaning, and maintenance of production equipment that may be a higher 
standard than most nonpharmaceutical quality system models. Organizations should 
always keep in mind that, while the FDA may be encouraging the adoption of a 
comprehensive quality system, any system developed must satisfy the requirements 
of the cGMP regulations. 

Under a quality system model, the specifi cation of facility and equipment requirements 
may be performed by technical experts (e.g., engineers, development scientists) 
who have an understanding of the pharmaceutical science, manufacturing 
processes, and risk factors associated with the product and its manufacture. The 
cGMP regulations require the QU to be responsible for reviewing and approving 
all initial design criteria and procedures pertaining to facilities and equipment and 
any subsequent changes. These requirements are not mutually exclusive; while they 
place ultimate responsibility for review and approval of these activities with the QU, 
the regulations do not preclude a cross - functional review involving persons with 
relevant expertise from multiple areas of the organization. A requirement of both 
the cGMP and current quality system models is that such review and approval be 
conducted by persons who are qualifi ed by education, training and experience to 
do so. 
In the control of outsourced operations, the cGMP regulations require that the 
QU approve or reject products or services provided under a contract. Under current 
quality system models, the organization must follow a formal vendor qualifi cation 
process to qualify outsource providers and verify through inspection or other appropriate 
means that the provider is capable of meeting the requirements of the organization. 
To comply with the regulation, these operations should be conducted by 
the QU. 
Table 2 compares the major elements of a quality systems approach to addressing 
resource issues with corresponding requirements in the CGMP regulations. 
3.2.6 MANUFACTURING OPERATIONS 
There is signifi cant commonality between the requirements contained in current 
quality system models such as ISO 9001 - 2000 and the cGMP regulation requirements 
for manufacturing operations. The FDA has identifi ed four major elements 
of a QS approach to manufacturing operations. These are identifi ed and compared 
to the cGMP requirements in Table 3 . 
TABLE 2 21 CFR cGMP Regulations Related to Resources 
Quality System Element Regulatory Citation 
1. General arrangements 
2. Develop personnel Qualifi cations: § 211.25(a) 
Staff number: § 211.25(c) 
Staff training: § 211.25(a – b) 
3. Facilities and equipment Buildings and facilities: § § 211.22(b), 211.28(c), 
211.42 – 211.58, 211.173 
Equipment: § § 211.63 – 211.72, 211.105, 211.160(b)(4), 
211.182 
Lab facilities: § 211.22(b) 
4. Control outsourced operations Consultants: § 211.34 
Outsourcing: § 211.22(a) 
MANUFACTURING OPERATIONS 211

212 ROLE OF QUALITY SYSTEMS AND AUDITS 
3.2.6.1 Design, Develop, and Document Product and Processes 
In a modern quality systems manufacturing environment, the signifi cant characteristics 
of the product being manufactured should be defi ned and verifi ed as meeting 
requirements from design to delivery, and control should be exercised over all 
changes. This is consistent with the requirements of the cGMP regulation that 
require quality and manufacturing processes and procedures, and changes to them, 
to be defi ned, approved, and controlled. The idea of controlling the design of both 
product and process is consistent with concepts included in the FDA Pharmaceutical 
cGMPs for the 21st Century Initiative to assure product safety that focus on the 
entire product life cycle. No amount of “ downstream ” control and testing can compensate 
for a design that results in a product or production process that is incapable 
of meeting the requirements necessary to assure that the product is safe and 
effective for its intended use. Documentation is required and can include 
the following: 
• Resources and facilities used 
• Procedures to carry out the process 
• Identifi cation of the process owner who will maintain and update the process 
as needed 
• Identifi cation and control of important variables 
• Quality control measures, necessary data collection, monitoring, and appropriate 
controls for the product and process 
• Any validation activities, including operating ranges and acceptance criteria 
• Effects on related process, functions, or personnel 
The cGMP regulations include specifi c packaging and labeling controls, so packaging 
and labeling requirement, processes, and controls should be included in a QS - 
based approach to product and process design and development. 
TABLE 3 21 CFR cGMP Regulations Related to Manufacturing Operations 
Quality System Element Regulatory Citation 
1. Design and develop product and 
processes 
Production: § 211.100(a) 
2. Examine Inputs Materials: § § 210.3(b), 211.80 – 211.94, 211.101, 
211.122, 211.125 
3. Perform and monitor operations Production: § § 211.100, 211.103, 211.110, 
211.111, 211.113 
QC criteria: § § 211.22(a – c), 211.115(b), 
211.160(a), 211.165(d), 211.188 
QC checkpoints: § § 211.22 (a), 211.84(a), 
211.87, 211.110(c) 
4. Address nonconformities Discrepancy investigation: § § 211.22(a), 
211.100, 211.115, 211.192, 211.198Recalls: 
21 CFR Part 7 

Manufacturers and the FDA have expressed concern that existing regulatory 
requirements (e.g., the need to effect manufacturing process changes through 
the regulatory submission process) may be excessively rigid and not conducive 
to innovation regardless of the potential benefi ts. The FDA acknowledges that 
the reluctance to pursue potentially innovative changes in pharmaceutical manufacturing 
can be undesirable from a public health perspective and has published a 
process analytical technology (PAT) guidance document that is intended to address 
this by promoting the use of analytical tools to gain process understanding 
and meet regulatory requirements for validating and controlling manufacturing 
processes [7] . 
The PAT guidance document describes a voluntary approach to the design, 
analysis, and control of manufacturing processes that involves the timely (e.g., in - 
process) measurement of critical quality and performance attributes of raw and 
in - process materials and processes, with the goal of ensuring fi nal product quality. 
The term analytical in PAT is broadly interpreted to include the integrated application 
of chemical, physical, microbiological, mathematical, and risk analysis as 
appropriate. One goal of PAT is to design and develop well - understood processes 
that will consistently ensure predefi ned quality at the end of the manufacturing 
process. This is consistent with a quality systems approach. PAT should ideally 
be initiated during the development stage and is intended to be integrated into 
existing regulatory processes with timely communication with the FDA a key 
element. The FDA has published the guidance document and other pertinent PAT 
information on its website at www.fda.gov . Companies interested in PAT methods 
should contact the FDA. FDA internal implementation of PAT includes the 
following: 
• A PAT team approach of CMC review and cGMP inspections 
• Joint training and certifi cation of FDA PAT review, inspection, and compliance 
staff 
• Scientifi c and technical support for the PAT review, inspection, and compliance 
staff 
Process analytical technology is consistent with the quality systems approach in 
that it is based on science and engineering principles for assessing and mitigating 
risks related to poor product and process quality. In the PAT guidance, the FDA 
indicates that the desired state for pharmaceutical manufacturing may be characterized 
as follows: 
• Product quality and performance are ensured through the design of effective 
and effi cient manufacturing processes 
• Product and process specifi cations are based on a mechanistic understanding 
of how formulation and process factors affect product performance 
• Continuous real - time quality assurance 
• Relevant regulatory policies and procedures are tailored to accommodate the 
most current level of scientifi c knowledge 
• Risk - based regulatory approaches recognize the following: 
MANUFACTURING OPERATIONS 213

214 ROLE OF QUALITY SYSTEMS AND AUDITS 
The level of scientifi c understanding of how formulation and manufacturing 
process factors affect product quality and performance 
The capability of process control strategies to prevent or mitigate the risk of 
producing a poor - quality product 
Process analytical technology is consistent with a modern risk - based data - driven 
quality systems approach to cGMP compliance. 
3.2.6.2 Inputs 
Current QMS models adopt a process - oriented approach to the design and operation 
of a QMS as a system of interrelated processes, each with inputs and outputs, 
which are designed to function in a defi ned way. Some process outputs are inputs 
to other processes. This concept is easily applied and understood within the manufacturing 
environment because it is process oriented. Inputs to manufacturing processes 
include any material that goes into the fi nal product, including materials 
purchased from vendors for use in manufacturing and in - process materials. Manufacturing 
operations generally involve multiple processes conducted in a defi ned 
manner to produce the fi nished product. Each process has a set of inputs and produces 
one or more outputs that may, in turn, be an input to a subsequent process. 
Each process has an input – output relationship such that changes or variation in one 
or more inputs will produce an attendant change in the output. Input specifi cations 
are established to assure that the fi nal product meets its requirements. A robust 
quality system will ensure that all inputs to the manufacturing process are suitable 
for use by establishing quality controls for the receipt and acceptance from qualifi ed 
vendors, production, storage, and use of all inputs. 
The cGMP regulations require either testing or use of a certifi cate of analysis 
(COA) plus an identity analysis for the release of materials for manufacturing. The 
quality systems approach additionally calls for initial supplier qualifi cation based 
on an objective evaluation and periodic auditing of suppliers based on risk assessment 
to verify the adequacy of suppliers ’ quality systems. During the audit, a manufacturer 
can observe the testing or examinations conducted by the supplier to help 
determine the reliability of the supplier ’ s COA. Under a QMS model, the QU would 
normally be responsible for auditing suppliers as part of its overall responsibility 
for materials acceptance. 
Change control involves the evaluation of proposed changes in a systematic way 
to determine how they would affect process outputs and ultimately the fi nished 
product and is an important element of current quality system models. The cGMP 
regulations require the QU to approve specifi cations, and certain changes require 
review and approval by the QU. Under a quality system model, changes to materials 
(e.g., specifi cation, supplier, or materials handling) should be implemented through 
a formal change control system involving the documented competent review and 
approval of the proposed change prior to implementation and communication of 
changes as appropriate throughout the organization. Manufacturers should also 
consider how best to assure that changes made by suppliers in supplied materials 
that may affect the quality of the fi nished product can be identifi ed and appropriately 
evaluated by the manufacturer. Such provisions should be included in supplier 
agreements where possible. 
0 
0 

3.2.6.3 Perform and Monitor Operations 
Both the cGMP regulations and quality system models call for the monitoring of 
critical processes that may be responsible for causing variability during production. 
The cGMP regulations require written production and process control procedures 
and specify process control activities that must be performed and documented. 
Current quality system models also require written procedures, process verifi cation 
and validation as appropriate, the establishment of appropriate process control 
measures and documentation. Risk analysis methods and design and development 
data may be used to establish process control and monitoring requirements. A 
quality systems approach allows the manufacturer to more effi ciently and effectively 
validate, perform, and monitor operations and ensure that the controls are scientifi - 
cally sound and appropriate. Production and process controls should be designed 
to ensure that the fi nished products have the identity, strength, quality, and purity 
they purport or are represented to possess. A systems approach will consider all 
sources of variability from inputs, through manufacturing processes, packaging, 
labeling, and shipping to assure that the product that is delivered to the user meets 
quality requirements. 
One important aspect of the quality systems approach is the ongoing collection 
and analysis of quality data to continuously evaluate quality system effectiveness. 
Historical data, process knowledge, and risk analysis methods can be applied to 
identify specifi c data requirements. Trending and other data analysis methods can 
allow identifi cation of actual and potential sources of nonconformity so that appropriate 
corrective and preventive actions can be taken in accordance with established 
change control procedures. 
The entire product life cycle should be addressed by the establishment of monitoring 
and continual improvement mechanisms in the quality system. Even well - 
defi ned or mature manufacturing processes may “ drift ” due to a host of factors 
including equipment and facility aging, changes or variation in raw materials, electrical 
power fl uctuations, and environmental changes. Thus, process validation is not a 
one - time event but an activity that continues throughout a product ’ s life. One 
major quality system objective should be to identify emerging quality problems 
before nonconformities occur. Trending of periodically collected environmental 
monitoring data may, for example, identify a slow but steady increase in airborne 
particulate levels that, if left unaddressed and the trend continues, could exceed a 
fi rm ’ s internal environmental standards and adversely affect the product. Early 
identifi cation of such problems allows an investigation to be initiated to identify the 
cause so that appropriate corrective and preventive actions can be taken in accordance 
with established change control procedures. After a change is implemented, 
its effectiveness should be objectively verifi ed and affected processes revalidated if 
necessary. 
3.2.6.4 Address Nonconformities 
A key component in any quality system is appropriately responding to nonconformities 
(i.e., deviations from requirements established under the quality system for 
in - process material or fi nal product quality attributes, process control parameters, 
records, procedures, etc.). Nonconformities may be detected during any stage of the 
MANUFACTURING OPERATIONS 215

216 ROLE OF QUALITY SYSTEMS AND AUDITS 
manufacturing process or during quality control activities. The cGMP regulations 
require an investigation to be initiated and that the investigation, conclusion, and 
follow - up be documented. A primary objective of any manufacturing quality system 
is to prevent nonconforming product from being produced and distributed. The 
complete response to nonconformities should be risk based and can include the 
following components: 
• Assessment of how the nonconformity will affect the quality of the fi nished 
product (i.e., determination if the nonconformity has resulted, or could result, 
in product that does not meet its specifi ed purity, potency, and quality 
characteristics). 
• Determine any actions necessary to assure that product that does not meet its 
specifi ed requirements is not produced and that appropriate steps are taken 
with regard to any nonconforming product that has been produced to assure 
that consumers are not harmed and that regulatory requirements are 
satisfi ed. 
• Determine the cause of the nonconformity. 
• Identify any actions needed to correct the cause and to prevent recurrence. 
• Document the investigation, fi ndings, and follow - up actions. 
• Assess the effectiveness of follow - up actions. 
• Repeat the cycle as needed. 
A nonconformity may not result in the fi nished product failing to meets its 
requirements; however, investigation of the nonconformity may identify process or 
quality system defi ciencies that require attention. For example, a small but unexpected 
deviation from a process control requirement (e.g., temperature, blending 
time) may not exceed the limit for which the process was initially validated and 
thus not be expected to adversely effect the fi nished product but could suggest an 
emerging process control or equipment issue that if not corrected could result in 
future product nonconformities. Similarly, nonconformities in the form of errors or 
omissions in production records or deviations from written procedures may not 
always result in product nonconformity but could suggest training, process design, 
or other issues that ought to be addressed. Thus the response to nonconformities 
should not be limited to a determination of the immediate impact on the fi nished 
product, but also consider its implications regarding overall quality system 
performance. 
3.2.7 EVALUATION ACTIVITIES 
The evaluation component of a QMS is intended to provide objective information 
and data that allow the organization to assess the conformity of the product, 
evaluate the performance of its quality system, and maintain and improve its effectiveness 
[10] . The cGMP regulations similarly require evaluation activities as shown 
in Table 4 . 

3.2.7.1 Trend Analysis 
The cGMP regulations require review and analysis of certain quality data annually 
at least. Current quality system models emphasize data - based decision making and 
the use of appropriate statistical analysis methods [2, 11] . Trend analysis is one statistical 
tool specifi cally recommended by the FDA in its pharmaceutical QS guidance 
document that can be very valuable in monitoring processes and quality system 
performance to identify emerging problems and to assess the effectiveness of 
improvement efforts. Traditional statistical process control and other methods also 
provide valuable support in the objective and ongoing analysis of quality data and 
can be helpful in implementing real - time quality assurance practices as recommended 
by the FDA [7] . 
3.2.7.2 Conduct Internal Audits 
Internal auditing is not specifi cally required by the cGMP regulations, but manufacturers 
have traditionally used internal audits as a self - assessment tool and to 
prepare for FDA inspections. The FDA has for some time recognized the value of 
internal auditing and encourages fi rms to conduct audits by, as a matter of policy, 
not reviewing internal audit results during inspections [12] . 
Current quality system models call for audits to be conducted at planned intervals 
to evaluate effective implementation and maintenance of the quality system and to 
determine if processes and products meet established parameters and specifi cations. 
International standards provide guidance on auditing [13] . Audit procedures should 
be developed and documented to ensure that the planned audit schedule takes into 
account the relative risks of the various quality system activities. Factors that can 
be incorporated into a risk - based approach to planning audit frequency and scope 
include the following [6] : 
• Existing legal requirements (e.g., cGMPs) 
• Overall compliance status and history of the company or facility 
• Robustness of a company ’ s quality risk management activities 
• Complexity of the site 
• Complexity of the manufacturing process 
• Complexity of the product and its therapeutic signifi cance 
• Number and signifi cance of quality defects (e.g., recall) 
TABLE 4 21 CFR cGMP Regulations Related to 
Evaluation Activities 
Quality System Element Regulatory Citation 
1. Analyze data for trends Annual review: § 211.180(e) 
2. Conduct internal audits 
3. Risk assessment 
4. Corrective action Discrepancy investigation: 
§ § 211.22(a), 211.192 
5. Preventive action 
6. Promote improvement § 211.110 
EVALUATION ACTIVITIES 217

218 ROLE OF QUALITY SYSTEMS AND AUDITS 
• Results of previous audits/inspections that can include prior internal audit 
results as well as regulatory (e.g., state, federal, or other regulatory agencies) 
and third - party audits 
• Major changes of building, equipment, processes, and key personnel 
• Experience with manufacturing of a product (e.g., frequency, volume, number 
of batches) 
• Test results of offi cial control laboratories 
In general, auditors should not have direct responsibility over the matters being 
audited. Auditors should be trained in auditing methods and have suffi cient technical 
knowledge to be able to evaluate the systems being audited using objective audit 
criteria [14] . Audit criteria may be based on applicable regulatory requirements, 
standards to which the quality system is intended to conform (e.g., ISO 9001 - 2000), 
and the specifi c requirements of the quality system being audited as indicated in 
quality system documents. Auditing criteria should be defi ned prior to the initiation 
of the audit. 
Different audit approaches may be applied depending on the intended purpose 
and scope of the audit. A top - down approach fi rst evaluates the overall structure of 
the quality system and its subsystems. Selected subsystems may be chosen for review. 
Systems identifi ed and developed by the FDA in a six - system inspection model for 
the inspection of drug manufacturers [15] include the following: 
• Overall quality system 
• Facilities and equipment 
• Materials system 
• Production system 
• Packaging and labeling 
• Laboratory controls 
Subsystems must be pertinent to the specifi c quality system being audited and may 
coincide with major elements of a standard to which the quality system is intended 
to conform or the major elements identifi ed in the FDA pharmaceutical QS guidance. 
When using the top - down approach, the auditor will fi rst review each subsystem 
to determine if the requirements that apply to that subsystem (e.g., 
regulatory requirements, the requirements of the standard) are met by defi ning, 
documenting, and implementing appropriate procedures. Once the auditor has 
verifi ed that the requisite procedures are in place, he or she will review the associated 
records and other documents to verify that the procedures have been followed 
and documented and that the quality system is functioning effectively as designed 
and conforms to applicable regulatory requirements and standards. This approach 
allows for a systematic evaluation of each subsystem and can be as detailed as 
needed. 
A bottom - up approach may be used to follow up on a specifi c quality problem 
identifi ed from trend analyses, product nonconformities, adverse experiences, customer 
complaints, or other sources of quality data. Starting with quality records 
associated with the problem, the auditor will work his or her way up through the 

quality system, examining the quality processes having a bearing on the quality 
problem. This approach is helpful in identifying quality system issues that may be 
associated with specifi c quality problems but does not readily allow evaluation of 
the entire quality system. 
A combination approach may also be used that employs elements of top - down 
and bottom - up audits. This allows some level of assessment of the effectiveness of 
the overall quality system while evaluating the cause of specifi c quality problems. 
Auditors should select the audit method most appropriate for their intended 
audit purpose. Initial quality system audits or regularly scheduled audits are likely 
candidates for the top - down approach, while audits conducted as part of a root 
cause analysis, for example, may best employ a bottom - up approach. The FDA 
employs a similar approach to inspections. Regular scheduled biennial inspections 
are more likely to employ a top - down methodology. For cause inspections conducted 
in response to a specifi c product issue such as a recall are more likely to 
employ a bottom - up approach. FDA investigators may employ a combination 
approach during biennial inspections if investigators are aware of specifi c quality 
problems that they wish to include in the inspection. 
Auditing as described in QMS models is intended to assess the effectiveness of 
the overall quality system as designed and conformance to applicable standards. The 
overall quality system does not have to be covered in a single audit. Manufacturers 
may choose to employ a rolling audit approach in which specifi cally identifi ed subsystems 
are chosen for evaluation in accordance with an approved audit schedule. 
Audit plans should be designed to effectively perform this assessment. 
Compliance with cGMP requirements is also a major concern, and audit planning 
should include assessment of conformance to cGMP requirements and readiness 
for FDA inspections. Existing FDA guidance documents and compliance policy 
guides describe FDA inspectional approaches and policy and can be used for reference 
in developing audit plans [15 – 17] . It can be helpful to include mock FDA audits 
as part of an overall auditing regimen. Some fi rms prefer to use outside auditors for 
mock audits to better simulate the FDA inspection process. Mock audits are also 
useful for training purposes to prepare the organization for FDA inspections. 
The audit plan should be consistent with written quality auditing procedures 
included in the quality manual or other quality system documentation. The plan 
should include or refer to the objective criteria to be used to evaluate conformance 
to requirements. The plan should include or refer to other documents that will be 
used during the audit, including previous audit reports. If the audit is to include the 
review of batch or production records, such review should be conducted in accordance 
with a specifi ed sampling plan or other appropriate statistical rationale as 
specifi ed in a fi rm ’ s quality system procedures. 
Manufacturers implementing a quality system that conforms to an existing standard 
may fi nd it helpful to create a table or some other document that shows the 
relationship between cGMP requirements, requirements of the standard, and the 
element(s) of the manufacturer ’ s quality system. Such a tool can help assure that 
all pertinent requirements are covered in the quality system design and that audit 
plans designs include assessment of all pertinent requirements. 
Since current quality system models employ a systems approach, an audit checklist 
that is organized by subsystem may be helpful, as described in Table 5 . The 
form would include appropriate document control information such as form 
EVALUATION ACTIVITIES 219

220 ROLE OF QUALITY SYSTEMS AND AUDITS 
identifi cation, revision, and approval information. Companies may also wish to 
include reference information used in planning the audit such as previous audit 
reports, completed FDA Form 483 Inspectional Observations, third - party audit 
reports, and pertinent internal QS documents (e.g., audit procedures). Depending 
on the purpose of the audit, the subsystems may correspond to the six subsystems 
identifi ed by the FDA for use by investigators in conducting cGMP inspections (i.e., 
quality, production, facilities and equipment, laboratory controls, materials, packaging 
and labeling) or the major elements of a quality system standard. Cross references 
between elements of the standard being used and the pertinent sections of 
the cGMP regulations may be included as appropriate. The audit form should allow 
entry of information regarding conformance or nonconformance to each requirement 
and have space for a description of pertinent fi ndings. 
The QMS models require periodic audits but do not specify audit frequency. 
Audit frequency must be determined based on the risk associated with the matters 
to be audited and other factors including results of previous audits and other quality 
data. Periodic audits should be conducted over the entire product life cycle and 
follow - up audits conducted as appropriate to verify that previously identifi ed quality 
problems have been corrected in accordance with applicable quality system and 
regulatory requirements. 
3.2.7.3 Quality Risk Management 
The FDA has endorsed quality risk management as part of an overall quality 
systems approach to compliance with the cGMP regulations and achieving overall 
TABLE 5 Example Audit Checklist 
[Company Name] Quality System Audit Checklist 
Form: Rev: Date: Approved: 
Audit Date(s): Refs: 
Auditor: Title Signature: 
Requirement cGMP Section Cross 
Reference 
Conforms (Y/N/NA) Objective Evidence 
and Comments 
Subsystem 1 
Requirement 1.1 
Requirement 1.2 
Subsystem 2 
Requirement 2.1 
Requirement 2.2 
Subsystem 3 
Requirement 3.1 
Requirement 3.2 
Subsystem N 

quality system objectives [6] . Risk management methodologies permit management 
to assign priorities to activities or actions based on an assessment of the risk including 
both the probability of occurrence of harm and the severity of that harm. 
Implementation of quality risk management includes assessing the risks, selecting 
and implementing risk management controls commensurate with the level of risk, 
and evaluating the results of the risk management efforts. In a manufacturing quality 
systems environment, risk management is used as a tool in the development of 
product specifi cations and critical process parameters. Used in conjunction with 
process understanding, quality risk management helps manufacturers effectively 
manage and control change. 
A formal risk management process consists of several components: 
• Risk assessment 
Risk identifi cation 
Risk analysis 
Risk evaluation 
• Risk control 
Risk reduction 
Risk acceptance 
• Risk communication 
• Risk review 
Risk assessment starts with risk identifi cation , a systematic use of available information 
to identify hazards (i.e., events or other conditions that have the potential 
to cause harm). Information can be from a variety of sources including stakeholders, 
historical data, information from the literature, and mathematical or scientifi c analyses. 
Risk analysis is then conducted to estimate the degree of risk associated with 
the identifi ed hazards. This is estimated based on the likelihood of occurrence and 
resultant severity of harm. In some risk management tools, the ability to detect the 
hazard may also be considered. If the hazard is readily detectable, this may be considered 
a factor in the overall risk assessment. Risk evaluation determines if the risk 
is acceptable based on specifi ed criteria. In a quality system environment, criteria 
would include impact on the overall performance of the quality system and the 
quality attributes of the fi nished product. The value of the risk assessment depends 
on how robust the data used in the assessment process is judged to be. The risk 
assessment process should take into account assumptions and reasonable sources 
of uncertainty. Risk assessment activities should be documented. 
Risk control starts with risk reduction, which includes any actions taken to eliminate 
or reduce the risk. Actions taken should be commensurate with the signifi cance 
of the risk. If the risk has been reduced to an acceptable level, an affi rmative decision 
can be made to accept the risk (risk acceptance). One question to ask is if new 
risks have been introduced as a result of the identifi ed risks being controlled. Risk 
control measures should generally be conducted in accordance with change control 
procedures and documented. 
Risk communication involves the communication of appropriate information 
about the risk to stakeholders (e.g., others involved in or affected by the quality 
EVALUATION ACTIVITIES 221 
0 
0 
0 
0 
0 

222 ROLE OF QUALITY SYSTEMS AND AUDITS 
system including management, users, regulatory agencies). Risk communication 
should be documented. The included information might relate to the existence, 
nature, form, probability, severity, acceptability, control, treatment, detectability, or 
other aspects of risks to quality. Communication should be as appropriate and does 
not necessarily need to be carried out for each and every risk acceptance. 
Risk review should be conducted to evaluate the outputs of the risk management 
process and repeated as necessary, based on new quality data or if there are process 
or product changes. 
The Q9 Quality Risk Management guidance document [6] identifi es a number 
of risk management tools that manufacturers can apply, including failure mode 
effects and criticality analysis (FMECA), hazard analysis and critical control 
points (HAACP), and preliminary hazard analysis (PHA), and provides examples 
of how quality risk management might be applied to quality management, development, 
materials management, production, and other operations within the 
organization. 
3.2.7.4 Corrective and Preventive Actions 
Corrective and preventive action (CAPA) is the term commonly used to describe 
the subsystem of a comprehensive quality system that deals with the systematic 
investigation, understanding, and response to quality issues including nonconformities. 
A corrective or preventive action may be initiated based on review and analysis 
of quality data from a variety of sources including adverse experiences, product 
complaints, quality audits, FDA inspections, third - party inspections, nonconforming 
materials reports, process control information, trend analyses, and other sources. 
A corrective action is initiated to correct the cause of an identifi ed nonconformity 
and to prevent it or similar problems from reoccurring. It may include initial and 
follow - up actions (e.g., conducted after root cause analysis). Current quality system 
models and the cGMP regulations emphasize corrective actions and require that 
actions be documented. Under current quality system models, preventive actions 
include actions taken in response to quality data to address the cause of potential 
nonconformities to prevent their occurrence. An effective CAPA system therefore 
includes both reactive and proactive components. The effectiveness of corrective 
and preventive actions should be evaluated using objective criteria when possible 
and the evaluation documented. 
A fi rm ’ s CAPA system and processes should be designed to analyze and respond 
to quality issues in a systematic way that is commensurate with the risk. The system 
should provide for the verifi cation or validation of corrective and preventive actions 
to assure their effectiveness and to assure that actions do not adversely affect the 
fi nished product. The system should also assure that pertinent CAPA information 
is appropriately disseminated throughout the organization as necessary to assure 
the effective operation of the quality system and for management review. 
3.2.7.5 Promote Improvement 
Continual improvement is a requirement of existing quality system models such as 
ISO 9001 - 2000 in which the organization is required to continually improve the 
effectiveness of the quality management system through the use of the quality 

policy, quality objectives, audit results, analysis of data, corrective and preventive 
actions, and management review. In adapting the ISO 9001 - 2000 standard to serve 
as a regulatory standard for medical device quality management systems, drafters 
of the ISO 13485 standard altered the requirement slightly to require the organization 
to “ identify and implement any changes necessary to ensure and maintain the 
continued suitability and effectiveness of the quality management system through 
the use of the quality policy, quality objectives, audit results, analysis of data, corrective 
and preventive actions, and management review. ” The word improvement 
was deleted as not an objective of current regulatory standards, but the concept of 
continually monitoring the performance of the quality system and appropriately 
responding to quality data was retained. 
The cGMP regulation does not specifi cally require continual improvement; 
however, the regulations are specifi c with regard to the sampling and testing of in - 
process materials and drug products, and failure to take reasonable action to reduce 
identifi ed sources of variability may be of concern to FDA investigators. The FDA 
in its pharmaceutical QS guidance document encourages organizations to promote 
improvement through quality system activities and notes that it is critical for senior 
management to be involved. Process improvement, along with improvement of in - 
process controls, can render a manufacturing process more effi cient and more 
robust. The end result can reduce costs and further prevent product failures and 
defects from occurring. 
3.2.8 TRANSITIONING TO QUALITY SYSTEMS APPROACH 
The cGMP regulations assign signifi cant responsibilities to the organizational unit 
responsible for quality - related activities. Organizations implementing a quality 
system model will be responsible for additional quality - related activities including, 
but not necessarily limited to, conducting quality audits, analysis of quality data, 
risk assessment, and preventive actions based on review and analysis of quality data 
to prevent the occurrence of product nonconformities. In addition, management 
is required to provide requisite leadership by actively participating in the 
quality system and assuring that the quality system functions as intended. This is 
accomplished by establishing a quality policy and associated objectives, planning 
for quality, establishing an appropriate organization structure with designated 
responsibilities and authorities to appropriately carry out quality system requirements, 
providing appropriate resources and training, and periodically reviewing 
quality information and data, and assuring that the organization responds 
appropriately. 
The organizational unit responsible for quality - related activities will in all likelihood 
have an even greater role within the organization, and roles and responsibilities 
throughout the organization are likely to change. Careful planning will be 
required to assure that the transition is effected smoothly with no adverse impact 
on product quality. Following are some points to consider in planning the 
transition: 
• Create a transition team: A cross - functional team should be developed involving 
key managers and staff from throughout the organization to plan and 
TRANSITIONING TO QUALITY SYSTEMS APPROACH 223

224 ROLE OF QUALITY SYSTEMS AND AUDITS 
execute the transition. The transition team should have a clear understanding 
of its mission and the organizational objectives associated with the transition. 
• Train the transition team: The decision to make the transition must come from 
management and management should assure that all individuals on the transition 
team receive proper training on quality systems requirements, risk management, 
and FDA ’ s recommended approach to quality systems. 
• Develop a transition plan: A transition plan, based on clearly defi ned objectives, 
should be developed by the transition team. 
• Identify staffi ng requirements: The transition will likely affect individual job 
descriptions and create additional duties that will have to be addressed through 
the reassignment of staff, hiring new staff, and providing necessary training to 
all affected staff. 
• Identify other resource needs: The plan should include a defi nition of resource 
requirements for planning and executing the plan. 
• Defi ne roles and responsibilities: the plan should clearly defi ne the roles and 
responsibilities of those responsible for development and execution of the plan 
for quality system implementation as well as staff roles and responsibilities 
under the quality system. 
• Consider organizational structure requirements: In order to function properly, 
persons responsible for quality - related activities must have the responsibility 
and associated authority defi ned and appropriately communicated within the 
organization. 
• Conduct a gap analysis: The plan should conduct a gap analysis that identifi es 
how the quality system model chosen can be effectively integrated with existing 
processes to create a quality system that conforms to the organization ’ s quality 
objectives, meets regulatory requirements, and is consistent with other organizational 
requirements. The quality systems approach is intended to be somewhat 
fl exible in application and can be tailored to specifi c organizational 
requirements. In order to function properly the quality system must be effectively 
integrated into the organization so that it is not viewed as an “ add - on ” 
or a set of extra requirements that prevent the “ real ” work from getting 
done. 
• Consider benchmarking: If possible, arrange with other organizations that have 
successfully made the transition to meet with them, review their system, and 
discuss transition issues and how they were solved. 
• Consult with experts: In addition to benchmarking, seeking assistance from 
persons familiar with quality systems can be very helpful, particularly when 
existing staff are relatively inexperienced with quality systems. It may be useful 
for one or more outside experts to work with the transition team on a regular 
basis as a coach or facilitator. 
• Communicate regularly: Clear and ongoing communication within the transition 
team and with management is essential to effectively coordinate plan 
activities, report progress, resolve issues, and identify evolving resource 
needs. 
• Sell the system: Successful implementation of a QS requires the active and 
informed participation of many individuals within the organization. Manage

ment commitment should be clearly communicated and training provided so 
that affected staff understand basic quality system concepts and their role in 
the quality system. 
• Validate the system. 
• Maintain regulatory compliance. 
3.2.9 AUDIT CHECKLIST FOR DRUG INDUSTRY 
The checklist provided in Table 6 [15] is intended to aid in the systematic GMP audit 
of a facility that manufactures drug components or fi nished products. 
The adequacy of any procedures is subject to the interpretation of the auditor. 
Therefore, the author accepts no liability for any subsequent regulatory observations 
or actions stemming from the use of this audit checklist. 
3.2.9.1 Instructions for Using Audit Checklist 
Before starting an on - site audit, plan the audit. Review past audits, note indications 
of possible problem areas and items, if any, that were identifi ed for corrective action 
in a previous audit. If you are not already familiar with this facility, learn the type 
of product produced and how it is organized by personnel and function. What does 
your “ customer, ” that is, your superior or senior facility management, expect to learn 
from this audit? 
1. The checklist is to be used with a notebook into which detailed entries can be 
made during the audit. 
2. While the checklist is to guide the auditor, it is not intended to be a substitute 
for knowledge of the GMP regulations. 
3. Although a single question may be included about any requirement, the 
answer will usually be a multipart one since the auditor should determine the 
audit trail for several products that may use many different components. Enter 
details in you notebook and cross reference your comments with the 
questions. 
4. At least three production batches should be selected for thorough analysis to 
include: (a) traceability of all components or materials used in the subject 
batches, (b) documentation of raw material or component, in - process, and 
fi nished goods testing for the subject product batches, and (c) warehousing 
and distribution records as they would relate to a possible recall. 
5. Responses entered on the checklist should be consistent. “ X ” is recommended 
for “ No ” ; a checkmark for “ Yes ” ; “ N/A ” for not applicable to questions that 
do not apply. An asterisk and notebook page number should be entered on 
the checklist to identify where relevant comments or questions are recorded 
in your notebook. 
6. The notebook used should be a laboratory - type notebook with bound pages. 
The notebook should be clearly labeled as to the audit type, date, and auditor(s). 
Many auditors prefer to use a notebook for a single audit so it may be fi led 
with the checklist and the fi nal report. 
AUDIT CHECKLIST FOR DRUG INDUSTRY 225

226 ROLE OF QUALITY SYSTEMS AND AUDITS 
TABLE 6 Audit Checklist 
Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
1.0 General Controls 
Does the facility and its departments (organizational units) 
operate in a state of control as defi ned by the GMP 
regulations? 
1.1 Organizational & Management Responsibilities 
1.101 Does this facility/business unit operate under a facility or 
corporate quality policy? 
1.102 § 211.22(a) Does a Quality Assurance unit (department) 
exist as a separate organizational entity? 
1.103 § 211.22(a) Does the Quality Assurance unit alone have 
both the authority and responsibility to approve or reject 
all components, drug product containers and closures, in - 
process materials, packaging materials, labeling, and drug 
products? 
1.104 § 211.22 Does the QA department or unit routinely review 
production records to ensure that procedures were 
followed and properly documented? 
1.105 § 211.22(b) Are adequate laboratory space, equipment, and 
qualifi ed personnel available for required testing? 
1.106 If any portion of testing is performed by a contractor, has 
the Quality Assurance unit inspected the contractor ’ s site 
and verifi ed that the laboratory space, equipment, 
qualifi ed personnel, and procedures are adequate? 
1.107 Date of last inspection: — 
1.108 § 211.22(c) Are all QA procedures in writing? 
1.109 § 211.22(c) Are all QA responsibilities in writing? 
1.110 Are all written QA procedures current and approved? 
(Review log of procedures) 
1.111 Are the procedures followed? (Examine records to ensure 
consistent record - keeping that adequately documents 
testing.) 
1.112 § 211.25 Are QA supervisory personnel qualifi ed by way of 
training and experience? 
1.113 § 211.25 Are other QA personnel (e.g., chemists, analysts, 
laboratory technicians) qualifi ed by way of training and 
experience? 
1.2 Document Control Program 
1.201 § 211.22(a) Does the QA unit have a person or department 
specifi cally charged with the responsibility of designing, 
revising, and obtaining approval for production and 
testing procedures, forms, and records? 
1.202 § 211.22(d) Does a written SOP, which identifi es how the 
form is to be completed and who signs and countersigns, 
exist for each record or form? 
1.203 § 211.165(a)(b)(c) Is the production batch record and 
release test results reviewed for accuracy and 
completeness before a batch/lot of fi nished product is 
released?

Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
1.3 Employee Orientation, Quality Awareness, and Job 
Training 
1.301 Circle the types of orientation provided to each new 
employee: (1) Company brochure. (2) Literature 
describing GMP regulations and stressing importance of 
following instructions. (3) On - the - job training for each 
function to be performed ( before the employee is allowed 
to perform such tasks). (4) Other: enter in notebook. 
1.302 § 211.25(a) Does each employee receive retraining on an 
SOP (procedures) if critical changes have been made in 
the procedure? 
1.303 Indicate how ongoing, periodic GMP training is 
accomplished. 
1.304 § 211.25 is all training documented in writing that indicates 
the date of the training, the type of training, and the 
signature of both the employee and the trainer? 
1.305 § 211.25 Are training records readily retrievable in a manner 
that enables one to determine what training an employee 
has received, which employees have been trained on a 
particular procedure, or have attended a particular 
training program? 
1.306 Are GMP trainers qualifi ed through experience and 
training? 
1.307 § 211.25(a) Are supervisory personnel instructed to prohibit 
any employee who, because of any physical condition 
(as determined by medical examination or supervisory 
observation) that may adversely affect the safety or 
quality of drug products, from coming into direct contact 
with any drug component or immediate containers for 
fi nished product? 
1.308 § 211.28(d) Are employees required to report to supervisory 
personnel any health or physical condition that may have 
an adverse effect on drug product safety and purity? 
1.309 § 211.25(a) Are temporary employees given the same 
orientation as permanent employees? 
1.310 § 211.34 Are consultants, who are hired to advise on any 
aspect of manufacture, processing, packing or holding, of 
approval for release of drug products, asked to provide 
evidence of their education, training, and experience? 
1.311 § 211.34 Are written records maintained stating the name, 
address, qualifi cations, and date of service for any 
consultants and the type of service they provide? 
1.4 Plant Safety and Security 
1.401 Does this facility have a facility or corporate safety 
program? 
1.402 Are safety procedures written? 
1.403 Are safety procedures current? 
TABLE 6 Continued 
AUDIT CHECKLIST FOR DRUG INDUSTRY 227

228 ROLE OF QUALITY SYSTEMS AND AUDITS 
Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
1.404 Do employees receive safety orientation before working in 
the plant area? 
1.405 Is safety training documented in a readily retrievable 
manner that states the name of the employee, the type of 
training, the date of the training, and the name of the 
trainer and the signature of the trainer and the 
participant? 
1.406 Does this facility have a formal, written security policy? 
1.407 Is access to the facility restricted? 
1.408 Describe how entry is monitored/restricted: 
1.409 Is a security person available 24 hours per day? 
1.5 Internal Quality/GMP Audit Program 
1.501 Does this business unit/facility have a written quality 
policy? 
1.502 Is a copy of this quality policy furnished to all employees? 
1.503 If “ yes ” to above, when provided? — 
1.504 Is training provided in quality improvement? 
1.505 Does a formal auditing function exist in the Quality 
Assurance department? 
1.506 Does a written SOP specify who shall conduct audits and 
qualifi cations (education, training, and experience) for 
those who conduct audits? 
1.507 Does a written SOP specify the scope and frequency of 
audits and how such audits are to be documented? 
1.508 Does a written SOP specify the distribution of the audit 
report? 
1.6 Quality Cost Program 
1.601 Does this facility have a periodic and formal review of the 
cost of quality? 
1.602 Does this facility have the ability, through personnel, 
software, and accounting records, to identify and capture 
quality costs? 
1.603 Does this facility make a conscious effort to reduce quality 
costs? 
2.0 Design control 
Not directly related to the drug regulation 
3.0 Facility control 
3.1 Facility Design and Layout 
3.101 § 211.42(a) Are all parts of the facility constructed in a way 
that makes them suitable for the manufacture, testing, 
and holding of drug products? 
3.102 § 211.42(b) Is there suffi cient space in the facility for the 
type of work and typical volume of production? 
3.103 Does the layout and organization of the facility prevent 
contamination? 
3.2 Environmental Control Program 
3.201 The facility is NOT situated in a location that potentially 
subjects workers or product to particulate matter, fumes, 
or infestations? 
TABLE 6 Continued

Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
3.202 Are grounds free of standing water? 
3.203 § 211.44 Is lighting adequate in all areas? 
3.204 § 211.46 Is adequate ventilation provided? 
3.205 § 211.46 Is control of air pressure, dust, humidity, and 
temperature adequate for the manufacture, processing, 
storage, or testing of drug products? 
3.206 § 211.46 If air fi lters are used, is there a written procedure 
specifying the frequency of inspection and replacement? 
3.207 Are drains and routine cleaning procedures suffi cient to 
prevent standing water inside the facility? 
3.208 § 211.42(d) Does the facility have separate air - handling 
systems, if required, to prevent contamination? 
(MANDATORY IF PENICILLIN IS PRESENT!) 
3.3 Facility Maintenance and Good Housekeeping Program 
3.301 § 211.56(a) Is this facility free from infestation by rodents, 
birds, insects, and vermin? 
3.302 § 211.56(c) Does this facility have written procedures for 
the safe use of suitable (e.g., those that are properly 
registered) rodenticides, insecticides, fungicides, and 
fumigating agents? 
3.303 Is this facility maintained in a clean and sanitary condition? 
3.304 Does this facility have written procedures that describe 
in suffi cient detail the cleaning schedule, methods, 
equipment, and material? 
3.305 Does this facility have written procedures for the safe and 
correct use of cleaning and sanitizing agents? 
3.306 § 211.58 Are all parts of the facility maintained in a good 
state of repair? 
3.307 § 211.52 Is sewage, trash, and other refuse disposed of in a 
safe and sanitary manner (and with suffi cient frequency)? 
3.4 Outside Contractor Control Program 
3.401 § 211.56(d) Are contractors and temporary employees 
required to perform their work under sanitary 
conditions? 
3.402 Are contractors qualifi ed by experience or training to 
perform tasks that may infl uence the production, 
packaging, or holding of drug products? 
4.0 Equipment control 
4.1 Equipment Design and Placement 
4.101 § 211.63 Is all equipment used to manufacture, process, or 
hold a drug product of appropriate design and size for its 
intended use? 
4.102 Are the following pieces of equipment suitable for their 
purpose: blender(s), conveyor(s), tablet, presses, capsule 
fi llers, bottle fi llers, other (specify)? 
4.103 Are the following pieces of equipment suitable in their size/ 
capacity: blender(s), conveyor(s), tablet, presses, capsule 
fi llers, bottle fi llers, other (specify)? 
TABLE 6 Continued 
AUDIT CHECKLIST FOR DRUG INDUSTRY 229

230 ROLE OF QUALITY SYSTEMS AND AUDITS 
Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
4.104 Are the following pieces of equipment suitable in their 
design: blender(s), conveyor(s), tablet, presses, capsule 
fi llers, bottle fi llers, other (specify)? 
4.105 Are the locations in the facility of the following pieces of 
equipment acceptable: blender(s), conveyor(s), tablet, 
presses, capsule fi llers, bottle fi llers, other (specify)? 
4.106 Are the following pieces of equipment properly installed: 
blender(s), conveyor(s), tablet, presses, capsule fi llers, 
bottle fi llers, other (specify)? 
4.107 Is there adequate space for the following pieces of 
equipment: blender(s), conveyor(s), tablet, presses, 
capsule fi llers, bottle fi llers, other (specify)? 
4.108 § 211.65(a) Are machine surfaces that contact materials 
or fi nished goods nonreactive, nonabsorptive, and 
nonadditive so as not to affect the product? 
4.109 § 211.65(b) Are design and operating precautions taken to 
ensure that lubricants or coolants or other operating 
substances do NOT come into contact with drug 
components or fi nished product? 
4.110 § 211.72 Fiber - releasing fi lters are NOT used in the 
production of injectable products. 
4.111 § 211.72 Asbestos fi lters are NOT used in the production of 
products. 
4.112 Is each idle piece of equipment clearly marked “ needs 
cleaning ” or “ cleaned; ready for service ” ? 
4.113 Is equipment cleaned promptly after use? 
4.114 Is idle equipment stored in a designated area? 
4.115 § 211.67(a)(b) Are written procedures available for each 
piece of equipment used in the manufacturing, processing, 
or holding of components, in - process material, or fi nished 
product? 
4.116 Do cleaning instructions include disassembly and drainage 
procedure, if required, to ensure that no cleaning solution 
or rinse remains in the equipment? 
4.117 Does the cleaning procedure or startup procedure ensure 
that the equipment is systematically and thoroughly 
cleaned? 
4.2 Equipment Identifi cation 
4.201 § 211.105 Are all pieces of equipment clearly identifi ed with 
easily visible markings? 
4.202 § 211.105(b) Are all pieces of equipment also marked with 
an identifi cation number that corresponds with an entry 
in an equipment log? 
4.203 Does each piece of equipment have written instructions for 
maintenance that includes a schedule for maintenance? 
4.204 Is the maintenance log for each piece of equipment kept on 
or near the equipment? 
TABLE 6 Continued

Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
4.3 Equipment Maintenance & Cleaning 
4.301 § 211.67(b) Are written procedures established for the 
cleaning and maintenance of equipment and utensils? 
4.302 Are these procedures followed? 
4.303 § 211.67(b)(1) Does a written procedure assign responsibility 
for the cleaning and maintenance of equipment? 
4.304 § 211.67(b)(2) Has a written schedule been established 
and is it followed for the maintenance and cleaning of 
equipment? 
4.305 Has the cleaning procedure been properly validated? 
4.306 § 211.67(b)(2) If appropriate, is the equipment sanitized 
using a procedure written for this task? 
4.307 § 211.67(b)(3) Has a suffi ciently detailed cleaning and 
maintenance procedure been written for each different 
piece of equipment to identify any necessary disassembly 
and reassembly required to provide cleaning and 
maintenance? 
4.308 § 211.67(b)(3) Does the procedure specify the removal or 
obliteration of production batch information from each 
piece of equipment during its cleaning? 
4.309 Is equipment cleaned promptly after use? 
4.310 Is clean equipment clearly identifi ed as “ clean ” with a 
cleaning date shown on the equipment? 
4.311 § 211.67(b)(5) Is clean equipment adequately protected 
against contamination prior to use? 
4.312 § 211.67(b) Is equipment inspected immediately prior to 
use? 
4.313 § 211.67(c) Are written records maintained on equipment 
cleaning, sanitizing, and maintenance on or near each 
piece of equipment? 
4.4 Measurement Equipment Calibration Program 
4.401 § 211.68(a) Does the facility have approved written 
procedures for checking and calibration of each piece of 
measurement equipment? (Verify procedure and log for 
each piece of equipment and note exceptions in notebook 
with cross reference.) 
4.402 § 211.68(a) Are records of calibration checks and inspections 
maintained in a readily retrievable manner? 
4.5 Equipment Qualifi cation Program 
4.501 § 211.63 Verify that all pieces of equipment used in 
production, packaging, and quality assurance are capable 
of producing valid results. 
4.502 § 211.68(a) When computers are used to automate 
production or quality testing, have the computer and 
software been validated? 
4.503 Have on - site tests of successive production runs or tests 
been used to qualify equipment? 
4.504 Were tests repeated a suffi cient number of times to ensure 
reliable results? 
TABLE 6 Continued 
AUDIT CHECKLIST FOR DRUG INDUSTRY 231

232 ROLE OF QUALITY SYSTEMS AND AUDITS 
Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
4.505 § 211.63 Is each piece of equipment identifi ed to its 
minimum and maximum capacities and minimum and 
maximum operating speeds for valid results? 
4.506 Have performance characteristics been identifi ed for each 
piece of equipment? (May be provided by the 
manufacturer but must be verifi ed under typical 
operations conditions.) 
4.507 Have operating limits and tolerances for performance been 
established from performance characteristics? 
5.0 Material/component control 
5.1 Material/Component Specifi cation and Purchasing Control 
Although purchasing is not specifi cally addressed in the 
current GMP regulation, incumbent upon user of 
components and materials to ensure quality of product, 
material, or component. 
5.101 Has each supplier/vendor of material or component been 
inspected/audited for proper manufacturing controls? 
(Review suppliers and audits and enter names, material 
supplied, and date last audited in notebook.) 
5.2 Material/Component Receipt, Inspection, Sampling, and 
Laboratory Testing 
5.201 § 211.80(a) Does the facility have current written procedures 
for acceptance/rejections of drug products, containers, 
closures, labeling, and packaging materials? (List selected 
materials and components in notebook and verify 
procedures.) 
5.202 § 211.80(d) Is each lot within each shipment of material or 
components assigned a distinctive code so material or 
component can be traced through manufacturing and 
distribution? 
5.203 § 211.82(a) Does inspection start with visual examination of 
each shipping container for appropriate labeling, signs of 
damage, or contamination? 
5.204 § 211.82(b) Is the number of representative samples taken 
from a container or lot based on statistical criteria and 
experience with each type of material or component? 
5.205 § 211.160(b) Is the sampling technique written and followed 
for each type of sample collected? 
5.206 Is the quantity of sample collected suffi cient for analysis 
and reserve in case retesting or verifi cation is required? 
Verify that the following steps are included in written 
procedures unless more specifi c procedures are followed: 
5.207 § 211.84(c)(2) Containers are cleaned before samples are 
removed. 
5.208 § 211.84(c)(4) Stratifi ed samples are not composited for 
analysis. 
5.209 § 211.84(c)(5) Containers from which samples have been 
taken are so marked indicating date and approximate 
amount taken. 
TABLE 6 Continued

Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
5.210 Each sample container is clearly identifi ed by material or 
component name, lot number, date sample taken, name 
of person taking sample, and original container 
identifi cation. 
5.211 § 211.84(d)(1)(2) At least one test is conducted to confi rm 
the identity of a raw material (bulk chemical or 
pharmaceutical) when a Certifi cate of Analysis is 
provided by supplier and accepted by QA. 
5.212 If a Certifi cate of Analysis is not accepted for a lot of 
material, then additional testing is conducted by a written 
protocol to determine suitability for purpose. 
5.213 § 211.84(d)(6) Microbiological testing is conducted where 
appropriate. 
5.3 Material Component Storage and Handling 
Verify that materials and components are stored and 
handled in a way that prevents contamination, mixups, 
and errors. 
5.301 § 211.42(b) Are incoming material and components 
quarantined until approved for use? 
5.302 Are all materials handled in such a way to prevent 
contamination? 
5.303 Are all materials stored off the fl oor? 
5.304 Are materials spaced to allow for cleaning and inspection? 
5.305 § 211.122(d) Are labels for different products, strengths, 
dosage forms, etc., stored separately with suitable 
identifi cation? 
5.306 Is label storage area limited to authorized personnel? 
5.307 § 211.89 Are rejected components, material, and containers 
quarantined and clearly marked to prevent their use? 
5.4 Inventory Control Program 
5.401 § 211.142 Are inventory control procedures written? 
5.402 Does the program identify destruction dates for obsolete 
or out - dated materials, components, and packaging 
materials? 
5.403 § 211.150(a) Is stock rotated to ensure that the oldest 
approved product or material is used fi rst? 
5.404 § 211.184(e) Is destruction of materials documented in a 
way that clearly identifi es the material destroyed and the 
date on which destruction took place? 
5.5 Vendor (Supplier) Control Program 
5.501 Are vendors periodically inspected according to a written 
procedure? 
5.502 Is the procedure for confi rming vendor test results written 
and followed? 
6.0 Operational control 
TABLE 6 Continued 
AUDIT CHECKLIST FOR DRUG INDUSTRY 233

234 ROLE OF QUALITY SYSTEMS AND AUDITS 
Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
6.1 Material/Component/Label Verifi cation, Storage, and 
Handling 
6.101 § 211.87 Do written procedures identify storage time beyond 
which components, containers, and closures must be 
reexamined before use? 
6.102 § 211.87 Is release of retested material clearly identifi ed for 
use? 
6.103 Are retesting information supplements originally obtained? 
6.104 Do written procedures identify steps in the dispensing of 
material for production? 
6.105 Do these procedures include (1) release by QC, (2) 
documentation of correct weight or measure, and (3) 
proper identifi cation of containers? 
6.106 Does a second person observe weighing/measuring/ 
dispensing and verify accuracy with a second signature? 
6.107 § 211.101(c) Is the addition of each component documented 
by the person adding the material during manufacturing? 
6.108 § 211.101(d) Does a second person observe each addition of 
material and document verifi cation with a second 
signature? 
6.109 § 211.125(a) Does a written procedure specify who is 
authorized to issue labels? 
6.110 § 211.125(a) Does a written procedure specify how labels 
are issued, used, reconciled with production, returned 
when unused, and the specifi c steps for evaluation of any 
discrepancies? 
6.111 § 211.125(d) Do written procedures call for destruction of 
excess labeling on which lot or control numbers have 
been stamped or imprinted? 
6.2 Equipment/Line/Area Cleaning, Preparation, and Clearance 
6.201 § 211.67(b)(5) Do written procedures detail how equipment 
is to be checked immediately prior to use for cleanliness, 
removal of any labels, and labeling from prior print 
operations? 
6.202 § 211.67(b)(3) Do written procedures detail any 
disconnection and reassembly required to verify readiness 
for use? 
6.3 Operational Process Validation and Production Change 
Order Control 
6.301 Have production procedures been validated? (Review 
selected procedures for validation documentation. 
Adequate?) 
6.302 § 211.100(a) Does the process control address all issues to 
ensure identity, strength, quality, and purity of product? 
6.303 § § 211.101(a) Does the procedure include formulation that 
is written to yield not less than 100% of established 
amount of active ingredients? 
TABLE 6 Continued

Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
6.304 § 211.101(c) Are all weighing and measuring preformed by 
one qualifi ed person and observed by a second person? 
6.305 § 211.101(d) Have records indicated preceding policy been 
followed by presence of two signatures? 
6.306 § 211.103 Are actual yields calculated at the conclusion of 
appropriate phases of the operation and at the end of the 
process? 
6.307 § 211.103 Are calculations performed by one person? Is 
there independent verifi cation by a second person? 
6.4 In - Process Inspection, Sampling, and Laboratory Control 
6.401 § 211.110(a) Are written procedures established to monitor 
output and validate the performance of manufacturing 
procedures that may cause variability in characteristics of 
in - process materials and fi nished drug products? 
6.402 § 211.110(c) Are in - process materials tested at appropriate 
phases for identity, strength, quality, purity, and are they 
approved or rejected by Quality Control? 
6.403 § 211.160(b) Are there laboratory controls including 
sampling and testing procedures to assure conformance 
of components, containers, closures, in - process materials, 
and fi nished product specifi cations? 
6.5 Reprocessing/Disposition of Materials 
6.501 § 211.115(a) Do written procedures identify steps for 
reprocessing batches? 
6.502 § 211.115(b) Are quality control review and approval 
required for any and all reprocessing of material? 
6.503 Does testing confi rm that reprocessed batches conform to 
established specifi cation? 
6.504 Does a written procedure outline steps required to 
reprocess returned drug products (if it can be determined 
that such products have not been subjected to improper 
storage conditions)? 
6.505 Does Quality Control review such reprocessed returned 
goods and test such material for conformance to 
specifi cations before releasing such material for resale? 
7.0 Finished product control 
7.1 Finished Product Verifi cation, Storage, and Handling 
7.101 § 211.30 Do written procedures indicate how and who 
verifi es that correct containers and packages are used for 
fi nished product during the fi nishing operation? 
7.102 § 211.134(a) In addition, do written procedures require that 
representative sample of units be visually examined upon 
completion of packaging to verify correct labeling? 
7.103 § 211.137(a) Are expiration dates stamped or imprinted on 
labels? 
7.104 § 211.137(b) Are expiration dates related to any storage 
conditions stated on the label? 
TABLE 6 Continued 
AUDIT CHECKLIST FOR DRUG INDUSTRY 235

Question 
Instructions/Questions 
(note any exceptions and comments in notebook) Yes, No, or NA 
7.105 § 211.142(a) Are all fi nished products held in quarantine 
until QC has completed its testing and releases product 
on a batch - to - batch basis for sale? 
7.106 § 211.142(o) Is fi nished product stored under appropriate 
conditions of temperature, humidity, light, etc. 
7.2 Finished Product Inspection, Sampling, Testing, and 
Release for Distribution 
7.201 § 211.166 Has the formulation for each product been tested 
for stability based on a written protocol? (Containers 
must duplicate those used in fi nal product packaging.) 
7.202 § 211.166 Are written sampling and testing procedures and 
acceptance criteria available for each product to ensure 
conformance to fi nished product specifi cations? 
7.203 § 211.170(a) Is a quantity of samples equal to at least twice 
the quantity needed for fi nished product release testing 
maintained as a reserve sample? 
7.204 § 211.167(a) Are sterility and pyrogen testing performed as 
required? 
7.205 § 211.167(b) Are specifi c tests for foreign particles or 
abrasives included for any ophthalmic ointments? 
7.206 § 211.167(c) Do controlled release or sustained release 
products include tests to determine conformance to 
release time specifi cation? 
7.3 Distribution Controls 
7.301 § 211.150(a) Does a written procedure manage stocks to 
ensure that oldest approved product is sold fi rst? 
7.302 § 211.150(a) Are deviations to the policy above 
documented? 
7.303 § 211.150(a) Does a written procedure identify the steps 
required if a product recall is necessary? 
7.304 Is the recall policy current and adequate? 
7.4 Marketing Controls 
7.401 The current regulation does not address marketing controls 
per se except that all fi nished products must meet their 
specifi cations. 
7.5 Complaint Handling and Customer Satisfaction Program 
7.501 § 211.198(a) Are complaints, whether received in oral or 
written form, documented in writing, and retained in a 
designated fi le? 
7.502 § 211.198(a) Are complaints reviewed on a timely basis by 
the Quality Control unit? 
7.503 § 211.198(b)(1) Is the action taken in response to each 
complaint documented? 
7.504 § 211.198(b)(3) Are decisions not to investigate a complaint 
also documented and the name of the responsible person 
documented? 
7.505 § 211.198(b)(2) Are complaint investigations documented 
and do they include investigation steps, fi ndings, and 
follow - up steps, if required? Are dates included for each 
entry? 
TABLE 6 Continued 
236

7. The references to sections in the GMP regulation are for your convenience 
should a question arise. In some instances, two or more sections within the 
GMP regulation may have bearing on a specifi c subject. The headings in the 
GMP regulation will usually offer some guidance on the areas covered in each 
section. 
8. A general suggestion for a successful audit is to spend most of your time on 
major issues and a smaller portion of your time on small issues. There may be 
observations that you may wish to point out to supervisory personnel that 
deserve attention but do not belong in an audit report because they are relatively 
insignifi cant. By the same token, too many small items suggests a trend 
of noncompliance and deserve attention as such. When citing these, be 
specifi c. 
REFERENCES 
1. U.S. Code of Federal Regulations (CFR) , Title 21, Part 211, Current good manufacturing 
practice for fi nished pharmaceuticals, available: http://www.accessdata.fda.gov/scripts/ 
cdrh/cfdocs/cfcfr/CRFSearch.cfm?CFRPart=211 , accessed Dec. 5, 2006 . 
2. American National Standards Institute (ANSI) ( 2000 ), Quality management system — 
Requirements, ANSI/ISO/ASQ Q9001 - 2000, ANSI, New York. 
3. American National Standards Institute (ANSI) ( 2000 ), Quality management system — 
Fundamentals and vocabulary, ANSI/ISO/ASQ Q9000 - 2000, ANSI, New York. 
4. International Organization for Standardization (ISO) , Application of risk management 
of medical devices, ISO 14971:2000, ISO, Geneva. 
5. U.S. Department of Health and Human services, U.S. Food and Drug Administration , 
Pharmaceutical cGMPs for the 21st century — A risk - based approach, Final Report — 
Fall 2004, September 2004 , available: http://www.fda.gov/cder/gmp/gmp2004/GMP_ 
fi nalreport2004.htm , accessed Dec. 5, 2006. 
6. U.S. Department of Health and Human Services (DHHS) , Food and Drug Administration 
( 2006 , June), Guidance for industry: Q9 Quality risk management, DHHS, Rockville, 
MD. 
7. U.S. Department of Health and Human Services (DHHS) , Food and Drug Administration 
( 2004 , Sept.), Guidance for industry: PAT — A framework for innovative pharmaceutical 
development, manufacturing, and quality assurance, DHHS, Rockville, MD. 
8. U.S. Department of Health and Human Services (DHHS) , Food and Drug Administration 
( 2006 , Sept.), Guidance for industry: Quality systems approach to pharmaceutical 
CGMP regulations, DHHS, Rockville, MD. 
9. U.S. Code of Federal Regulations (CFR) , Title 21, Part 820, Quality system regulation 
for medical devices, available: http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/ 
CFRSearch.cfm?CFRPart=820 , accessed Dec. 5, 2006 . 
10. International Organization for Standardization (ISO) , ( 2003 ), Medical devices — Quality 
management systems — Requirements for regulatory purpose, ISO 13485:2003, ISO, 
Geneva. 
11. Juran J. M. , and Godfrey , A. B. , Eds. ( 1999 ), Juran ’ s Quality Handbook , 5th ed. , McGraw - 
Hill , New York . 
12. FDA compliance policy guide section 130.000, FDA access to results of quality assurance 
program audits and inspections (CPG 7151.02), available: http://www.fda.gov/ora/ 
compliance_ref/cpg/cpggenl/cpg130 - 300.html , accessed Dec. 5, 2006 . 
REFERENCES 237

238 ROLE OF QUALITY SYSTEMS AND AUDITS 
13. International Organization for Standardization (ISO) , ( 2002 ), Guidelines for quality 
and/or environmental management systems auditing, ISO 19011:2002, ISO, Geneva. 
14. The Global Harmonization Task Force, SG4, Training requirements for auditors (guidelines 
for regulatory auditing of quality systems of medical device manufacturers — Part 1: 
General requirements — Supplement 3), available: http://www.ghtf.org/sg4/inventorysg4/ 
trainingfi nal.pdf , accessed Dec. 5, 2006 . 
15. FDA compliance program guidance manual for FDA staff: Drug manufacturing inspections 
program (7356.002), 2/1/ 2002 , available: http://www.fda.gov/cder/dmpq/compliance_ 
guide.htm , accessed Dec. 5, 2006. 
16. U.S. Department of Health and Human Services (DHHS) , Food and Drug Administration 
( 2004 , Sept.), Guidance for industry: Sterile drug products produced by aseptic processing 
— Current good manufacturing practice, DHHS, Rockville, MD. 
17. U.S. Department of Health and Human Services (DHHS) , Food and Drug Administration 
( 2001 , Aug.), Guidance for industry: Q7A good manufacturing practice guidance for 
active pharmaceutical Ingredients, DHHS, Rockville, MD. 

239 
3.3 
CREATING AND MANAGING A 
QUALITY MANAGEMENT SYSTEM 
Edward R. Arling , Michelle E. Dowling , and Paul A. Frankel 
Amgen, Inc., Thousand Oaks, California 
Contents 
3.3.1 Introduction 
3.3.2 Understanding a Quality Management System 
3.3.2.1 Defi ning Quality Management Systems 
3.3.2.2 Synthesis versus Analysis 
3.3.2.3 System versus Process 
3.3.2.4 Business Benefi ts of Establishing a Robust Quality Management System 
3.3.2.5 Industry and Regulatory Expectations 
3.3.3 Management and Staff: Leadership and Support 
3.3.3.1 Outlining Benefi ts to the Enterprise 
3.3.3.2 Speaking Management Language 
3.3.3.3 Translating Benefi ts to Staff 
3.3.3.4 Ensuring Staff Support and Management Leadership 
3.3.3.5 Traps to Avoid 
3.3.4 Establishing Quality Management System Scope 
3.3.4.1 Defi ning Business Requirements 
3.3.4.2 Integrating Quality Management System into Quality Plans 
3.3.4.3 Determining Process Resolution Requirements 
3.3.4.4 Scalability to Enterprise 
3.3.5 System and Process Ownership: Roles and Responsibilities 
3.3.5.1 Quality Management System Ownership and Management 
3.3.5.2 Process Ownership 
3.3.5.3 Process Owner Selection 
3.3.5.4 Stakeholder/Process Owner Integration 
3.3.5.5 Decision Authority 
3.3.5.6 Industry Knowledge 
3.3.5.7 Regulatory Inspection and Audit Lead 
3.3.5.8 Subject Matter Expert 
3.3.5.9 Metric Ownership 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

240 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
3.3.5.10 Documentation Ownership 
3.3.5.11 Training 
3.3.5.12 Risk Management 
3.3.5.13 Continuous Improvement and Project Management 
3.3.5.14 Nonconformance / CAPA / Planned Deviation Ownership 
3.3.6 Change Management/Communication 
3.3.6.1 Managing Organizational Change 
3.3.6.2 Communication 
3.3.6.3 Feedback and Alignment 
3.3.6.4 Training 
3.3.7 Measuring Success through Meaningful Metrics 
3.3.7.1 Performance Metric Development 
3.3.7.2 Metric Review 
3.3.7.3 Maturity Model 
3.3.7.4 Meeting Process Maturity Requirements 
3.3.8 Driving Continuous Improvement: Projects 
3.3.8.1 Process Improvements 
3.3.8.2 Process Improvement Proposal 
3.3.8.3 Task versus Project 
3.3.8.4 Project Metrics 
3.3.9 Ensuring Ongoing Success 
3.3.9.1 Establishing Mutual Goals 
3.3.9.2 Rewards and Recognition 
3.3.9.3 Ensuring Ongoing Program Continuity 
3.3.9.4 Program Institutionalization 
References 
3.3.1 INTRODUCTION 
The world ’ s population continues to grow and the average life expectancy continues 
to increase. Pharmaceutical and biopharmaceutical products are more in demand 
as the population expands, requiring novel and specialized medications to treat 
common and debilitating diseases. The industry is challenged to rapidly discover 
and commercialize products to treat existing unmet medical needs and emerging 
threats as viruses mutate into new diseases that threaten the stability of the world 
as we know it. 
At the same time, the global marketplace continues to increase its demand on 
the industry. Government, consumer, and wholesale buying pressures demand lower 
prices. Higher quality standards are expected by regulators and consumers. Competition 
continues to increase from generic, biosimilar, and counterfeit producers. 
Developing nations, with lower cost overheads, are developing economical production 
capabilities. Meanwhile, research and development costs are increasing. 
This chapter will outline the concepts, benefi ts, and practical implementation 
steps for developing a comprehensive quality management system (QMS) that supports 
pharmaceutical and biopharmaceutical manufacturing operations. The material 
presented is universal in its utility, applicable to small and large companies, 
development, and commercial enterprises. A QMS is a proactive, structured approach 

UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 241 
to supporting development and manufacturing operations. It includes all processes, 
metrics, management review, and continuous improvement activities. The QMS, as 
described in this chapter, is further supported through an active change management 
program and application of annual quality plans to ensure ongoing system 
sustainability. 
A well - designed QMS, with mature, developed processes, provides the required 
infrastructure and support necessary for successful manufacturing operations. Integrated 
processes, proactively managed, that can be quickly modifi ed to meet changing 
business and regulatory demands will support ongoing manufacturing operations 
and provide competitive advantage. This chapter provides guidance on creating and 
managing a robust QMS that supports manufacturing operations in the pharmaceutical 
and biopharmaceutical industry. 
3.3.2 UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 
Every development, testing, manufacturing, packaging, warehouse, or distribution 
facility has its own unique role in producing an output or product for consumption 
by a customer somewhere in the pharmaceutical or biopharmaceutical supply chain. 
Each facility and organization is critically dependent upon several different processes 
that function interdependently producing the desired output. Organizations ’ 
survival and profi tability are directly linked to the effi ciency of design, execution, 
performance, and interrelational attributes of these processes. Throughout a product 
life cycle, from early discovery through development, scale up, clinical testing, 
product technology transfer, registration, approval, commercialization, and eventually 
product discontinuance, robust processes are the foundation supporting the 
successful enterprise. 
Manufacturing support processes are discrete in their output, but interrelated in 
their overall effect. Weak or ill - defi ned processes have a diminishing overall effect 
on the organization and its product. It manifests itself as increased rework, rejected 
material, extended cycle times, delayed disposition, high nonconforming performance 
metrics, complaints, recalls, or other inabilities to meet customer or market 
demands. A comprehensive QMS may encompass all the processes supporting 
development and manufacturing. It includes the standards, policies, and procedures 
required to measure those processes for performance and maturity. It provides 
metrics necessary for leadership to perform risk - based prioritization and focus 
resources for business improvement and regulatory compliance. 
Robust processes will have owners that have defi ned roles, responsibilities, and 
accountabilities. These process owners must be fully dedicated to their process. They 
must know their process capabilities and expectations, the interrelationship between 
their process and other processes and manage them like a business unto themselves. 
Functional management must support process owners, and leadership must understand 
and lead the QMS effort as an ongoing program, treating it as the integral 
part of the business that it is. 
A QMS is an organizational approach consisting of people, interrelated processes, 
process inputs and outputs, and structured review programs that lead to 
ongoing continuous improvement. This complexity of processes requires a programmatic 
organization and management to effectively interrelate its components. A 

242 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
QMS program offi ce is required to provide the organizational benefi ts expected 
from well - managed processes and should be one of the fi rst elements established 
when instituting the program. 
A QMS and the processes comprising it are not the sole responsibility of the 
quality function or a single functional group. Inherently, these processes have no 
bounds in the organization. The concept must be owned, managed, or executed by 
all staff from leadership to the most entry - level manufacturing associate. A quality 
mindset must be part of every employee that contributes to the discovery, manufacture, 
packaging, testing, warehousing, and shipping of a product or output. A 
culture of quality and understanding of the processes in which personnel work are 
essential to advance the QMS to maximize benefi ts to the enterprise and remain 
competitive. 
Instituting a QMS through a holistic approach that supports manufacturing operations 
has the potential to meet and exceed customer, patient, shareholder, and 
employee expectations. It requires a cross - functional team approach, with proactive 
management of all the processes responsible for manufacture, including functional 
support from development, manufacturing, analytical, engineering, and quality 
assurance. The development and maintenance of a tested, robust QMS requires time 
and resources. Full maturation of processes and organizational culture change may 
take, in some cases, years to fully implement and realize benefi ts, but worth the effort 
and time. Signifi cant QMS issues should not be addressed as one - off fi xes. Rather, 
action taken to remediate defi cient processes should be approached as long - term 
corrections, addressing the root cause of the failed process, so they do not repeatedly 
plague the organization. 
The ultimate responsibility for a robust, functional QMS lies with top management. 
The organization follows the leadership, and therefore, leadership must 
support a QMS that is specifi cally designed for the organization, be aware of and 
monitor its progress and contribution to the organization, and frequently support, 
guide, and maintain it. Doing so ensures viability of the QMS, and in turn the QMS 
will provide leadership the data and guidance necessary to effectively manage the 
organization. 
3.3.2.1 Defi ning Quality Management Systems 
The term system or quality system is used with surprising inconsistency throughout 
the pharmaceutical and biopharmaceutical industry and by government regulators. 
Even within a single company or within a department, the terms can be nebulous 
in their use and interpretation. System is often used to describe an individual process 
or unit operation. Often, the term system is used so narrowly as to describe an 
individual policy, standard, or even a single procedure. 
Recent initiatives by global organizations such as ISO (International Organization 
for Standardization, www.iso.org ) and ICH (International Conference on 
Harmonization, www.ich.org ) are attempting to bring consistency in concept and 
standardization in defi nition to the QMS. In 2004, the Pharmaceutical Inspection 
Co - Operation Scheme (PIC/S, www.picscheme.org ) issued its recommendation on 
Quality System Requirements for Pharmaceutical Inspectorates. The U.S. Food and 
Drug Administration (FDA) initiated inspection surveillance approaches based 
upon QMS organization and is another source of defi nition and interpretation. 

UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 243 
Inconsistency in language and expectations continues to exist; however, efforts are 
progressing to minimize distinctions and globally harmonize efforts, structure, and 
language concerning quality systems. 
According to Webster ’ s dictionary, system is defi ned as a regularly interacting or 
interdependent group of items forming a unifi ed whole; a group of interacting 
bodies under the infl uence of related forces . . . an assemblage of substances that is 
in or tends to equilibrium . . . a group of organs that, when together, perform one or 
more vital functions . . . an organization forming a network especially for distributing 
something or serving a common purpose . . . an organized set of doctrines, ideas, 
or principles usually intended to explain the arrangement or working of a systematic 
whole [1] . 
The vocabulary and defi nitions used in this chapter defi nes a quality management 
system as the compilation of all the processes required to support the manufacture, 
packaging, testing, release, and distribution of an active pharmaceutical ingredient 
(API) or drug product. It is aligned with that of the FDA Center for Drug Evaluation 
and Research (CDER) compliance program 7356.002, issued to investigators 
for the inspection of pharmaceutical and biopharmaceutical manufacturing plants 
( www.fda.gov/IOM 7356.002). The CDER inspection program subdivides the processes 
comprising the QMS into six subsystems: quality, facilities/equipment, production, 
materials control, laboratory controls, and packaging and labeling. 
There are no specifi c CDER requirements as to which processes belong under 
each subsystem; however, one can easily follow the outline provided in 21 CFR Part 
211, the regulations applicable to human drug product manufacture, to aid in the 
determination of processes likely to be inspected during a regulatory inspection 
( www.fda.gov ). The FDA subdivides all the processes comprising a company ’ s QMS 
into six subsystems to ensure adequate and varied coverage during inspections. See 
Figure 1 . Using the same process organization structure and vocabulary as regulators 
provides an enterprise the advantage of more effi cient inspection preparation 
and avoidance of miscommunication during and after regulatory inspections. 
The CDER subsystem organization provides regulators and management the 
ability to focus attention to specifi c functional areas. Table 1 is an example of the 
processes, organized under appropriate subsystems, supporting a typical API or drug 
fi ll - and - fi nish operation. These subsystems are organized according requirements 
found in regulations used by investigators during inspections, 21 CFR Part 210, 211, 
and the unit operations and support processes necessary for production. 
One size does not fi t all situations. Each enterprise has the responsibility and 
latitude to design a QMS to meet its specifi c needs. Even facilities with very similar 
manufacturing operations may require different processes to support the business. 
Each manufacturing organization requires a customized set of processes which will 
comprise its QMS. The management group responsible for the QMS should be able 
to identify and justify the processes comprising the system. There is not a single set 
of processes that can be universally applied to all operations, as each organization 
is unique in its business, product output, organization, culture, as well as local and 
global regulatory and customer requirements. 
Processes identifi ed as part of the QMS can be organized into the appropriate 
CDER subsystems for the purpose of aligning with the methodology used during 
inspections. It also provides management the ability to determine areas of strength 
or opportunities for improvement within the QMS. Regulators will always include 

244 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
FIGURE 1 Subsystems and management relationship. 
TABLE 1 Quality Management System Subsystems and Processes 
Quality Facilities/equipment 
Audits and inspections Facility and equipment design 
Management review Equipment maintenance 
Risk management Equipment cleaning 
Organization and personnel Calibration 
Training Materials control 
Document management Supplier quality management 
Change control Sampling and inspection 
Nonconformances Receiving, warehouse, and storage 
Corrective and preventative actions Inventory management 
Biological product deviation Transport 
Product disposition Return and salvage 
Validation Laboratory controls 
Production Laboratory testing 
Manufacturing Sample management and sample plans 
Process monitoring Stability program 
Environmental and gowning monitoring Packaging and labeling 
In - process controls Labeling controls and approvals 
Gowning Package development 

UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 245 
a focus on the processes within the quality subsystem. Other subsystems will be 
reviewed during inspections based upon the type of inspection and compliance 
history of the enterprise. More information on how the FDA focuses inspections 
based on quality system and subsystem organization is available at the FDA website 
( www.fda.gov ) or articles written on this subject [2] . 
To maximize the effect of a QMS, it should be designed to be scalable and transferable 
throughout the enterprise and easy to understand and execute. An adequately 
designed QMS results in increased effi ciency, a compliant operation, and 
staff satisfaction. 
3.3.2.2 Synthesis versus Analysis 
With systems thinking, the whole is greater than the sum of its parts. Systems rely 
upon the interaction of several processes. An individual process has limited value 
on its own, regardless of the level of development it has achieved. Processes provide 
value to the system through synthesis with other processes. 
In the early twentieth century, researchers began to recognize the existence of 
interdependent relationships and organizational patterns among seemingly discrete 
parts. It is the relationships that allow parts to function as a whole. The “ perceived 
whole ” is a system. Systems thinking involves considering the parts in the context 
of that whole. In systems thinking: 
• Everything in a system is related to everything else in the system. 
• The parts of a system work together to achieve the overall objective of the 
whole system. 
• In addition to the immediate effects of an action, there will be other consequences 
that ripple through the system. 
• Every change brings benefi ts and consequences. 
• Changing or reinforcing patterns and relationships within a system is as necessary 
to achieving the goals of the system as changing or retaining the parts of 
the system. 
• Systems are “ living ” entities that sustain themselves through self - regulating 
dynamic equilibrium and organize to respond to externally imposed change. 
Viewing a QMS in this context is benefi cial to organizational leadership and management 
responsible for the system. It puts into perspective the overall effect on an 
organization that is achievable by individual processes alone and what can be 
achieved and sustained through active management and the interaction between 
those processes. 
3.3.2.3 System versus Process 
Traditional industry paradigm has the Quality Department responsible for quality 
and the Manufacturing Department responsible for producing product. Inherent 
confl ict exists in this model due to competing functional priorities. By building 
quality concepts and accountabilities into production processes responsible for 
production, quality becomes infused into the organization. Both Quality and 

246 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
Manufacturing therefore share the common goal of supplying high - quality product 
through the effi cient execution of their processes. 
Historically, very few processes were regarded as “ quality systems, ” and they were 
viewed as something owned by the Quality Department. These “ systems ” were in 
fact ill defi ned and nonrelated processes used to monitor or detect individual actions 
and activities occurring in the manufacturing environment. These systems were 
based on quality control (QC) type of responsibilities for testing quality into the 
product. Examples include raw material testing, in - process and fi nished - product 
testing, nonconforming material review, environmental monitoring, and release and 
distribution. Few were interrelated with other processes, actively supported by 
management, or reviewed by leadership for performance or compliance. 
The QC monitoring processes described above, if supported, were limited in their 
ability to support improvements and could only lead to action that was reactive in 
nature. Process integration is weak or nonexistent. Neither process maturity and 
development nor proactive system management is achievable. In the past, QMS 
enhancement was viewed as an expense and not seen as a relational contributor to 
the value chain. Aware management now realizes, through regulatory action, penalty 
and fi nes, delayed product approvals, recalls, and the like that establishment of a 
comprehensive QMS is essential to survive in the current regulatory environment 
and remain competitive in the business environment. 
With the advancement of quality assurance (QA) principles and concepts at the 
end of the last century, QMSs have evolved to be more proactive to include change 
control, supplier and internal auditing, risk management, lagging and leading metric 
collection, and review. Review of predictive metrics has become the basis for preventive 
action and continuous improvement programs. Today ’ s competitive environment 
obligates leading manufacturers and world - class organizations to apply 
proactive system thinking to expand their focus to include all processes that support 
product quality, irregardless of the stage of development or manufacture. Early 
implementation of appropriate processes supports quality - by - design concepts and 
practice, within the framework of a QMS and ensures quality in all processes and 
provides the foundation for good investigations and continuous improvement. 
A QMS should be comprised of all the processes supporting that business and 
include an effective management review of those process metrics. Management 
needs to be aware of and understand process performance through structured 
metrics review programs in order to take appropriate action, providing resources 
and capital to improve the QMS. This hierarchy is illustrated in Figure 2 . 
Processes supporting and applicable to pharmaceutical and biopharmaceutical 
manufacturing are easily determined by examining the business needs of the organization 
and the regulations governing them. A carefully designed QMS will consider 
the needs of the enterprise as a whole, as well as that of the individual unit 
operations comprising the enterprise. If the QMS design is comprehensive, it will 
provide signifi cant value to global and local management. It will support staff by 
standardizing processes, requirements, and expectations and provide leadership 
meaningful and comparable metrics on system and process performance. Changes 
can be quickly facilitated and implemented when process modifi cations are required. 
A consistent representation of processes to regulators builds confi dence and trust 
that the enterprise is capable to produce the product for which approval has been 
granted. 

UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 247 
3.3.2.4 Business Benefi ts of Establishing a Robust Quality Management System 
The competitive nature of the pharmaceutical business demands capable and effi - 
cient processes supporting discovery, development, technology transfer and scale - 
up, and commercial manufacturing and distribution. Execution of effi cient processes 
is the foundation for new and ongoing enterprises to be successful. It is the basis 
for successful manufacture and the bedrock upon which management and regulators 
can gauge the capability level of the enterprise. Providing patients with needed 
medicines in a timely, cost - effi cient manner, without delay due to manufacturing or 
compliance issues, should be a primary driving force behind the pharmaceutical and 
biopharmaceutical industry. 
Leadership may ask the question: Why implement a quality management system? 
The answer is that a well - designed system is necessary to establish a state of control 
to ensure that a high quality, safe, and effi cacious product is produced and available 
for patients. Quality systems as described in the forthcoming ICH Q10 guidance is 
the logical complement to its predecessors, ICH Q8 (Product Development) and 
ICH Q9 (Risk Management) ( www.ich.org ). These three guidance documents build 
upon each other from quality - by - design activities in development through the entire 
product life cycle. When used together, the guidance documents maximize their 
benefi ts to the enterprise through better process understanding, less regulatory 
scrutiny, and increased freedom to operate. Together, these guidance ’ s support more 
effi cient product life - cycle management from discovery through development and 
commercialization. 
Ineffi cient operations cost businesses untold amounts in fi nancial and human 
capital. A poorly designed system coupled with ineffi cient processes may result in 
rework of development and commercialization activities, data integrity issues, inef- 
fi cient use of resources, and delay in approval. Poorly designed processes may also 
FIGURE 2 Quality management system hierarchy. 

248 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
lead to loss of future revenue with business partners and have a negative regulatory 
consequence. 
A recent study conducted by the Pharmaceutical Manufacturing Research 
Project, a joint venture by Georgetown University and Washington University in St. 
Louis business schools, collected data from 42 manufacturing facilities owned by 19 
companies to determine factors that affected industry performance. The Final 
Benchmarking Report assessed performance in terms of manufacturing times, frequency 
of deviations from manufacturing standards, reasons for deviations, manufacturing 
yield, and rates of improvement for those metrics. 
The study determined that improvements in manufacturing process could save 
industry more than $ 50 billion in manufacturing costs, which the researchers believe 
could result in lower drug prices and more money for R & D. The report received no 
industry or government funding [3] . 
Leadership is both challenged and rewarded for supporting the development of 
a robust QMS. On one hand, it takes time and resources to design and develop a 
comprehensive program. Immediate return on this investment is not usually forthcoming. 
Management is typically under pressure to deliver aggressive results in a 
short time period, which is counterintuitive to careful planning and long - range 
development. Conversely, proactively formalizing and supporting a robust QMS 
will, in the long run, ensure the operations freedom to operate (regulatory compliance) 
and deliver business effi ciencies. 
In the pharmaceutical and biopharmaceutical manufacturing industry, the perception 
of quality has dramatically changed over the past several years, and loss of 
market capitalization can be a direct correlation to this perception. Large pharmaceutical 
companies have gone from some of the world ’ s most admired companies to 
losing signifi cant percentage of their value, based on consumer, media, and investor 
perceptions of quality and ethics. Speaking at a recent Parenteral Drug Association 
(PDA)/FDA joint regulatory conference, Daniel Diermeier, IBM Distinguished Professor 
of Regulation and Competitive Practice, Northwestern University stated: “ The 
perception of quality on the pharmaceutical value chain is greater than in other 
industries (auto, furniture, etc.). Patients cannot assess the quality of drugs as they 
can a car or hotel room. In healthcare, the ‘ value proposition ’ is higher than other 
industries and the Quality [Management] System is a critical subset of that perception 
” [4] . Dr. Diermeier goes on to suggest a QMS include processes for decision 
and detection to further protect the “ value proposition ” of the enterprise. 
Enterprises lacking individual capable processes experience degrees of negative 
effects throughout the organization. This is true for processes that support discovery, 
development, manufacturing, or marketing. Recent examples of fi nes imposed by 
regulators for poor processes supporting the QMS are increasing (see Table 2 ). 
These costs are only indicative of the fi ne itself and do not include lost revenue, cost 
of consultancy for remediation, decreased shareholder value, and diminished staff 
morale and support. These costs are typically an order of magnitude or more greater 
than the fi ne itself. 
A common misconception of pharmaceutical and especially smaller biopharmaceutical 
companies is that the implementation of a robust QMS is not required in 
areas other than commercial manufacturing. Small, biotech start - up companies also 
tend to delay the implementation of well - designed processes until they near the 
approval stage, focusing the organization instead for product approval or sale. This 

UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 249 
can become a costly miscalculation, as speed to market and limited capital demand 
processes supporting effi cient development, clinical and regulatory submission processes 
be executed with minimal waste or rework. 
Although QMSs are routinely identifi ed with commercial manufacturing, it is 
critical to establish process parameters for discovery, development, and technology 
transfer, including scale - up, characterization of process, analytical methodology, and 
validation. Development activities are executed more effi ciently through the application 
of robust processes and ultimately become the foundation for robust manufacturing. 
Failed development studies, inadequate comparability reports, clinical 
studies requiring repeated, or poorly supported analytical and process characterization 
contribute to delayed submissions and weak regulatory submission and inspection 
presentation. The identifi cation of processes supporting these activities, owner 
identifi cation and accountability and support will ensure success of the enterprise 
and reduce the anxiety and uncertainty that is inherent in development and approval 
activities. 
Several opportunities exist for pharmaceutical and biopharmaceutical manufacturing 
plants to improve effi ciency and cost savings, which ultimately validate the 
program ’ s benefi ts and supports leadership in achieving their fi nancial goals. Traditionally, 
the industry environment is heavily regulated and has been very risk 
adverse. These two elements combine to offer countless opportunities to improve 
ineffi cient and ill - defi ned processes, clarify process scope, defi ne process owner 
accountabilities and responsibilities, and remediate process duplication or gaps. 
Performing ineffi cient processes for the sake of avoiding regulatory scrutiny or 
attempting to defend poorly characterized processes without adequate data and 
interpretation becomes self - defeating to the industry. Poor prioritization of work, 
ill - defi ned process relationships, and functional management interference or neglect 
may also contribute to ineffi ciency. Staff requires processes that are easy to execute, 
well integrated, and result in value - added activities. This can only be accomplished 
through the design and execution of effi cient processes that are interrelated, bringing 
value to the enterprise, process owners, and stakeholders. 
An example of a robust process is the design, development, and operation of a 
nonconformance process. Regulations require an operational process to identify, 
document, and correct nonconformances occurring in licensed pharmaceutical manufacturing 
facilities for approved products. Companies spend signifi cant human 
TABLE 2 Potential Financial Impacts 
Company Compliance Issue Type of Impact 
Cost to Business 
( $ Mil) 
A Failure to follow procedure 
Inadequate training 
Multiple 483 
observations 
< 1 
B Inadequate process defi nition, 
controls, and oversight 
Warning letter > 1 
C Repeat observations — direct product 
impact Failure to meet warning 
letter commitments 
Consent decree > 100 
D Plant shutdown Direct fi nes product 
stock - out 
> 500 

250 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
capital identifying, documenting, and tracking nonconformances. But how much is 
actually being done to remediate these nonconformances? Can the nonconformances 
be related to previously completed development or commercialization 
studies? Is the nonconformance process suffi ciently related to an effective corrective 
or preventive action (CAPA) process? Does the preventive action interrelate 
effi ciently with an effi cient change control process to ensure proposed changes 
remain in compliance with registrations? Are the documentation and training processes 
suffi cient to support approved changes? An adequately designed QMS will 
ensure the supporting processes are present and that functional and interrelationships 
established. A systems implementation provides a holistic approach, which 
results in both building effective individual processes and interrelating those processes 
to maximize their effect on the business, driving effi cient and science - based 
activities. 
Maintaining good manufacturing practice (GMP) compliance is essential for 
pharmaceutical and biopharmaceutical companies. Results of noncompliance are 
costly fi nes, loss of revenue, higher overhead costs, delayed approvals, and poor 
customer and regulatory perceptions. Poor compliance results from an inadequately 
designed QMS that lacked the processes and management review required to 
support the enterprise. Processes supporting compliance include self - audits, change 
control, document revision and approval, and staff training programs. Regular management 
review of these processes will ensure resources are allocated to appropriate 
initiatives and there should be no surprises during inspections. A well - designed 
QMS should prevent negative regulatory consequences. Effi cient and compliant 
processes support lean manufacturing efforts through the documentation and 
understanding of processes. Management review of these processes ensures that 
leadership awareness, support, and action is taken by the organization when 
appropriate. 
Figures 3 and 4 illustrate how a biennial document review process and document 
processing cycle time metrics faltered in their early stages due to lack of process 
ownership, defi nition, and management review. This situation presented a compliance 
risk to the organization and resulted in poor business effi ciencies. Improve- 
FIGURE 3 Biennial document review process. 
11% 
19% 
15% 
55% 
87% 
92% 
84% 
88% 
0% 
10% 
20% 
30% 
40% 
50% 
60% 
70% 
80% 
90% 
100% 
Q1/05 Q2/05 Q3/05 Q4/05 Q1/06 Q2/06 Q3/06 Q4/06 
Reviews complete 
Actual 
Target

UNDERSTANDING A QUALITY MANAGEMENT SYSTEM 251 
ment was attained by assigning a process owner who defi ned and improved each 
process, developed meaningful metrics, and presented those metrics to management. 
Management became aware of process performance, understood the compliance 
risk and business impact and took appropriate action to focus staff efforts to meet 
process requirements. Results were improved document review cycles, proactive 
compliance with internal procedures and regulatory requirements, and the satisfaction 
of knowing that no additional effort was required to achieve better business 
results and regulatory compliance. 
3.3.2.5 Industry and Regulatory Expectations 
While there are no requirements for a “ quality system ” in current FDA regulations 
applicable to pharmaceutical and biopharmaceutical manufacturing, regulatory 
agencies and industry trade organizations are increasingly recognizing the importance 
of robust, functioning quality systems in support of manufacturing the world ’ s 
medicinal products. The FDA realizes not all quality principles are represented in 
current GMP regulations for drug products (21 CFR Part 211), which were last 
updated in 1978. 
Quality management system issues and their association with risk management 
are common topics discussed in trade and regulatory seminars and conferences. 
Recent guidelines such as FDA “ Quality Systems Approach to Current Good Manufacturing 
Practice Regulations ” found on the FDA website and part of FDA ’ s initiative 
titled “ GMP ’ s for the 21st Century ” was written to complement existing 
regulations. While the FDA guidance may change or even become redundant with 
the issuance of ICH Q10, there is common intent among industry and government 
to advance quality management systems. According to Joe Famulare, Director 
DMPQ, FDA, the “ FDA wanted to write a comprehensive Quality System model 
that would support and correlate with CGMP regulations. The guidance is consistent 
with defi ning a state of control; facilitate quality efforts, change control, Quality by 
Design, and risk management ” [4] . 
In discussing quality systems at a recent industry conference on GMPs, Chris 
Joneckis of the FDA CBER (Center for Biological Evaluation and Research) had 
FIGURE 4 Document review cycle time. 
0 
10 
20 
30 
40 
50 
60 
Q1/05 Q2/05 Q3/05 Q4/05 Q1/06 Q2/06 Q3/06 Q4/06 
Month 
Number of days 
Actual 
Target

252 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
this to say: “ A robust Quality Management System makes a strong case for quality 
product. It is a win, win, win — for patient, industry and regulators. It benefi ts technology 
transfer, process control, monitoring, capability, improves manufacturing, 
fewer nonconformances and better quality of investigations. Regulatory benefi ts 
include enhanced Chemistry, Manufacturing, Controls (CMC) review, change 
control, and submission of postapproval changes ” [5] . 
Regulatory and industry guidance documents have been generated in support of 
developing and organizing quality systems. In the late 1990s, the system - based 
inspection approach was formalized by the Center for Devices & Radiological 
Health (CDRH) of the FDA [6] . These regulations were codifi ed as QSR, Quality 
Systems Regulations, and are included in Part 820 of the Code of Federal Regulations 
(CFR). 
The CDER and CBER soon followed the CDRH approach and issued their own 
Compliance Program Guidance Manuals, 7356.002 [7] and 7345.848 [8] , respectively, 
which were modeled on the CDRH QSR approach. The CDER and CBER are 
responsible for ensuring the biennial inspection of pharmaceutical and biopharmaceutical 
manufacturing facilities. The guidelines listed here are used by investigators 
during manufacturing inspections. Process owners and stakeholders as well as management 
and leadership should be familiar with these compliance manuals and how 
investigators plan to use them during inspections. 
Current FDA inspectional surveillance, based on the models described above, 
requires investigators evaluate the processes within the subsystems defi ned by the 
QMS to determine compliance and risk to patient safety. This is different than the 
traditional approach of reviewing individual products during inspections. There is 
subtle, yet signifi cant advantage to both the regulating agencies and compliant 
companies by using a system approach, as the inspections are designed to be faster 
and cover many product types during one inspection. Companies with compliant 
histories can benefi t with nominal inspections, whereas companies with noncompliant 
histories will receive more regulatory scrutiny and possible regulatory action. 
The movement by industry groups such as the ISO, which attempts to provide 
recognized standards for many industries, was also grounded in a systems approach 
with the publication and certifi cation of ISO 9000 series and later with ISO 2000:9004 
( www.iso.org ), which is based on QMS establishment and eventually continuous 
improvements once processes become stable. 
The ICH, a joint regulatory – industry initivative on international harmonization 
for drug development and approval, also recognizes the value and contribution of 
a quality systems approach through its guidance development on this topic (ICH 
Q10). The pharmaceutical and biopharmaceutical industry and regulatory agencies 
are collaborating to fi nalize the guidance sometime in 2008. ICH Q10 is focused on 
pharmaceuticals and is intended to align GMP requirements with a quality system 
approach. It will be applicable to drug substance and drug product, large and small 
molecule products, and harmonize one approach to quality systems. It also will 
complement ICH Q8 and ICH Q9. ICH Q10 contains a pharmaceutical context 
emphasizing a comprehensive approach; key elements included are management 
response and continuous improvement. Several ICH guidance documents are 
already adopted by regulatory agencies, such as ICH Q7A, for the manufacture of 
APIs. As these guidance documents are adopted, they often become the basis for 
regulatory expectations and inspections. 

3.3.3 MANAGEMENT AND STAFF: LEADERSHIP AND SUPPORT 
All manufacturing operations operate, to some extent, with elements and components 
of a quality management system. Those elements and processes may not be 
recognized or managed as though they are an integral part of a larger system and 
may be primarily reactive in nature. Signifi cant time and resources are required to 
change an organization ’ s culture and practices to move existing elements from a 
fragmented, reactive program to a defi ned structure that is proactively managed. 
The degree to which a program is proactively managed and supported by its leadership 
is directly related to the benefi ts experienced by the organization. 
Three distinct levels of support are required for successful implementation of a 
QMS program: executive leadership, functional management, and operational staff. 
All three levels of the organization must support the effort to attain success. Delivering 
program understanding and benefi ts to each should be a priority to ensure 
acceptance and continuity. Motivating staff and leadership, through benefi ts and 
business results, is important to ongoing program sustainability. 
Leadership requires capable and dedicated staff to design and maintain a dynamic 
QMS program. Leadership must embrace the program and support it throughout 
the organization. Functional management must understand the program in order to 
support it and direct its staff in execution of the program. Staff must understand 
what the program means to them and experience and realize the benefi ts in order 
to support it. 
The quality organization must be seen as a partner in assuring product quality, 
not the department that disseminates quality. Within a QMS, certain processes are 
owned by the quality function, just as manufacturing, engineering, development, 
technical support, and facilities own processes within the system. All functional 
groups should have defi ned roles and responsibilities to ensure quality product is 
produced. Cross - functional support and delineation of responsibilities ensure quality 
is built into every process, and each process owner is ultimately responsible for his 
or her process output. Leadership that understands and embraces this concept will 
support and infuse a culture of quality throughout the organization, maximizing the 
probability of success and competitive advantage. 
The organizations leadership, management, staff, and QMS program group must 
work together to develop and progress the QMS. A successful program should detail 
expected benefi ts for all stakeholders in the organization and provide ongoing 
results demonstrating functionality and utility. 
3.3.3.1 Outlining Benefi ts to the Enterprise 
Establishing a formal, structured QMS for an organization requires leadership 
approval, resources, and capital. Leadership support and approval is the place to 
initiate the program to ensure all program efforts are supported and the proposed 
system meets the business needs. This includes having dedicated resources that can 
focus their efforts to design and manage the program and operate and manage the 
processes. 
Leadership has visibility to present business needs and budget and the vision and 
insight for the organizations ’ future. Quality management system design needs to 
fulfi ll present and future needs to be robust and value added. A gap analysis on 
MANAGEMENT AND STAFF: LEADERSHIP AND SUPPORT 253

254 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
current business processes can help leadership understand where opportunities exist 
for improving processes. These gaps can be determined by analyzing the purpose of 
the organization and its ability to deliver quality results on time and on budget. 
Manufacturing areas to examine for operational improvement are regulatory 
compliance, audit fi ndings, rework, nonconformances, document revisions, disposition 
timeliness, complaints received, inventory on hand, equipment failures, manufacturing 
cycle times, employee turnover, and training opportunities. Additional 
areas targeted for improvement may come from benchmarking key manufacturing 
parameters against industry peers. Results of a gap analysis begin the dialogue 
regarding process performance and the need for process improvements. Leadership 
must be convinced there is opportunity for fi nancial and competitive gain, and the 
resource investment to operate the QMS will be outweighed by the program bene- 
fi ts received. 
Management at the highest level in the organization must understand, support, 
and lead the strategy to implement systems across the enterprise. More often than 
not, this requires some level of business transformation, a cultural and behavioral 
shift, and a certain level of risk. The risk associated in implementing change is 
minimal compared to that of not having a robust system, as outlined in the benefi ts 
section. 
3.3.3.2 Speaking Management Language 
Without upper management championing the establishment of systems, midlevel 
management will not support the effort, dedicate the time required, nor practice the 
behaviors essential to establish and maintain the processes. Leadership needs to be 
cognizant of the benefi ts and consequences of nonimplementation and be clear and 
unwavering in its support, delivering frequent consistent messaging to management 
and staff. Leadership requires tangible and intangible benefi ts to be convinced that 
the efforts are worthwhile and working and to regularly convey results to staff. 
Tangible benefi ts should include metrics and improvements demonstrating 
process and system cost savings, compliant inspections and customer audits, faster 
product approvals and manufacturing throughput, less rejected material, reduced 
nonconformance issues, and more effi cient continuous improvement and project 
implementation. Intangible benefi ts include improved staff morale, faster, more 
accurate transparent decision making, less employee turnover, increased staff 
accountability, and an enhanced culture of quality throughout the organization. The 
“ feeling ” conveyed by an organization that is reactive, stressed, and without well - 
structured processes is much different than that of a proactive organization with 
simple processes that are easily and successfully executed by trained staff. 
Systems thinking allows decision making and process management to occur at 
the process owner level, not the functional management level. This is a cultural shift 
for many organizations but brings with it many benefi ts. Faster decision making, by 
subject matter experts is valuable to organizations. It can benefi t both on a day - to - 
day, lot - to - lot basis as well as provide long - term strategic direction to leadership. 
Taking the burden off functional management and defi ning process owner responsibilities 
allows functional management to manage resource and personnel issues 
and not split time and attention between resources, personnel, technical, and process 
issues. 

3.3.3.3 Translating Benefi ts to Staff 
Similar to leadership and management requirements regarding system understanding 
and benefi ts, staff requires understanding prior to accepting the cultural changes 
that a system - based approach will bring to the organization. Once the program is 
initiated, tangible and intangible benefi ts must be realized and appreciated in order 
for staff to continually support the program. Staff support, through benefi t realization 
and management direction, will ensure program execution, ultimately delivering 
the expected business results. 
Transforming disparate processes into processes that are simple to understand, 
easy to execute, and provide a sense of accomplishment meet one of management ’ s 
obligations to staff. Staff interest lies in the ability to perform their work, contribute 
to continuous improvement, and have a reasonable work – life balance. Finally, they 
want to be able to contribute to their careers, have defi ned career paths, and have 
attainable development goals for advancement. A well - designed quality management 
system can contribute to provide all these employee benefi ts. 
Staff benefi ts should be designed into the QMS. An outline of expected benefi ts 
should be presented to staff to gain their support of the system initiative. Accomplishments 
should be advertised and rewarded. Establishing well - defi ned processes 
empowers employee involvement, participation, and contribution to the organization. 
It reinforces a culture of quality throughout the organization, and provides a 
conduit for their contribution. 
3.3.3.4 Ensuring Staff Support and Management Leadership 
Management ’ s responsibility includes providing staff robust tools and processes 
necessary to accomplish their jobs effi ciently. Complex, missing, or fragmented 
processes do not allow for easy operational execution, the ability to leave work at 
reasonable times, and may result in poor - quality output or rework. This type of 
environment quickly becomes dissatisfying to employees and results in poor morale, 
low effi ciency, and ultimately lack of interest and loss of staff. 
Staff empowerment allows pride in workmanship. Well - designed quality systems 
make clear to staff where decision authority and process accountability lies, provide 
clear expectations of the process and process owners, and provide personnel a clear 
development path to process ownership. 
Clearly identifi ed process attributes provide organizations more than tribal 
knowledge to pass onto the next process owner. They provide clear structure, process, 
and other attributes critical to the ongoing success of the enterprise. The organization 
becomes reliant on their system and processes not people ’ s personal knowledge, 
which can be lost with staff turnover. 
Ensuring leadership and staff support requires that a well - defi ned plan be 
designed and shared throughout the organization. A long - range plan, spanning 
several years may benefi t the organization to maintain perspective and govern 
expectations. An annual quality plan should encompass all aspects of the QMS and 
contain detailed periodic goals and objectives. Progress against the quality plan 
needs to be advertised and celebrated. Quality plan leadership should be recognized 
for its efforts and accomplishments. Advertising wins and accomplishments in both 
small group and large settings should be designed into the communication and 
change management program. 
MANAGEMENT AND STAFF: LEADERSHIP AND SUPPORT 255

256 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
Table 3 provides an outline of a long - term vision and goals for a quality management 
system. A long - term strategy provides leadership, management, and staff with 
an understanding of the program and anticipated timelines for implementation and 
benefi t expectations. Annual quality plans become the short - term strategic milestone 
vehicle to achieve the long - term strategic vision. 
3.3.3.5 Traps to Avoid 
Several challenges and requirements present themselves when establishing a formal 
QMS. A primary requirement is a skilled team that understands the needs of the 
organization, regulatory, and customer requirements. It should have the skill, experience, 
and expertise to design a robust system and identify processes that support 
the enterprise. A mismatch of team skills with enterprise needs may result in a 
nonviable system that is not supported by leadership and staff, leading to failure 
and disuse over time. 
Quality management system design must be well thought out and tested. Pilot 
programs are crucial to test system robustness and reliability, staff and management 
acceptance, and the ability to produce the desired results. Time spent in system 
design will pay dividends for years to come and increase staff support and critical 
mass throughout the enterprise supporting the program efforts. Avoid implementing 
any system or process design that has not been well thought out, does not have input 
from the stakeholders using the system, or has not been piloted prior to a full - scale 
implementation. Typically, a single opportunity exists to introduce a new program 
before staff and management either accept it or reject the ideas and concepts. 
Rebuilding interest and trust of a failed system is diffi cult. The probability for successful 
reintroduction is minimized. Taking suffi cient precaution for correct implementation 
the fi rst time is important. 
TABLE 3 Long - Term Strategic Vision 
Year 1 Year 2 Year 3 Year 4 Year 5 
Gain 
management 
support 
Create QMS 
offi ce 
Identify site 
processes and 
resources 
Develop 
communication 
and change 
management 
plan 
Implement 
program 
Train 
management, 
process 
owners, QA, 
and support 
staff 
Focus on 
maturing 
high - risk/ 
impact 
processes 
Reward and 
recognize 
QMS efforts 
Indoctrinate 
remaining 
processes into 
program 
Document and 
communicate 
cost/resource 
savings 
Begin 
integrating 
processes 
across the 
organization 
Focus on key 
projects based 
on QMS 
portfolio and 
management 
review 
Provide ongoing 
training, 
communications, 
and change 
management 
Adapt to changing 
business and 
regulatory 
environment 
Provide leadership to 
industry on QMS 
paradigm

Change management is another very important consideration when implementing 
a QMS because of the culture change required from the organization. Several 
resources can assist in managing change, and these should be incorporated into the 
system design. It is important to be cognizant that successful implementation 
requires change at all three levels of the organization; leadership, functional management, 
and operational staff. Each will need different messages, encouragement, 
rewards, and benefi ts. Consideration to deliver both tangible and intangible benefi ts 
to stakeholders is necessary. 
Leadership support from the highest level is required. Middle management will 
not support an effort that is not supported by its leadership. Leadership must 
provide unwavering support, not provide mixed messages, continue to advertise and 
celebrate success, and support the program through rough times. Consistency in 
language and deeds from management supports understanding and appropriate risk 
taking by management and staff. 
Functional management must also support system efforts and long - term strategies, 
to ensure that staff, who are critical to execution of the processes, know that 
their support and efforts are expected. Functional management send powerful messages 
to staff, and their support of the long - term plan and annual quality plan are 
essential. Specifi c system objectives, included in leadership, management, and staff 
goals reinforce the commitment and help ensure success of the program. 
The system needs to remain fl exible. Having a long - term plan and vision is necessary 
to provide a roadmap to the future. That roadmap may need to change as the 
business environment and enterprise needs change. The long - term plan and vision 
should be written at the level that it changes very little, but fl exibility is maintained 
through the preparation of an annual quality plan that is capable of addressing 
temporal issues and business needs. 
Prior to implementing any QMS initiative, one must understand what leadership, 
management, staff, and customers require. Knowing which processes are required 
to support customer needs and the impact of those processes upon each other is 
essential to system design. Developing process owners that understand their roles 
and deliverables in the organization, eliminating constraints so they may meet their 
goals is essential for success. Process owners must understand product and process 
priorities so signifi cant benefi ts may be realized. These are important considerations 
in designing system and processes that support the organization, produce meaningful 
metrics, and demonstrate progress. Consideration of these important points 
prevents system initiatives from failing and interpreted as another burden to the 
already overburdened work and demands placed upon the organization. 
3.3.4 ESTABLISHING QUALITY MANAGEMENT SYSTEM SCOPE 
In many pharmaceutical and biopharmaceutical manufacturing operations, duplicity 
exists in some processes and gaps are present between others. Often it is unclear 
exactly what boundaries or scope constitute a process, the expected outputs, who 
are the customers, who is the owner, and who is responsible for continuous improvement. 
Duplicity is ineffi cient and costly. Examples include multiple layers of an 
organization performing data reviews as documentation or information moves 
through the value chain. Regulatory submissions for analytical validation are an 
ESTABLISHING QUALITY MANAGEMENT SYSTEM SCOPE 257

258 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
example, where raw data may be checked at the laboratory, supervisor, quality assurance, 
compliance, and regulatory group levels. On the other hand, gaps may exist 
where each functional group listed above assumes data verifi cation is occurring with 
another group, and in fact there are gaps in data integrity. In this case, the result can 
be tremendously expensive if, upon regulatory inspection, errors are found and it 
appears data integrity issues are ubiquitous in a submission. 
This section will discuss the importance of defi ning business requirements to 
ensure processes comprising the QMS are designed to support the enterprise, integrated 
into a quality plan, suffi ciently defi ned to provide adequate resolution and 
are transferable and scalable throughout the enterprise. 
3.3.4.1 Defi ning Business Requirements 
The QMS and the processes that comprise it must be custom designed for the needs 
of the business. One size does not fi t all situations. The requirements of an enterprise 
vary across sites and the phases of a product life cycle. A comprehensive system will 
ensure a holistic programmatic approach in its support to the enterprise. This does 
not mean that every phase of the product life cycle (discovery, development, commercial 
manufacturing) will utilize all the processes that comprise the system. Nor 
does it require that all commercial manufacturing sites will necessarily implement 
all processes. It does, however, provide a common platform and expectation for all 
processes, owners, metrics review programs, continuous improvement efforts, and 
the like when they are implemented. 
The fi rst step in designing a QMS is determining business needs and the processes 
required to support the enterprise. Important consideration must be given to ensure 
that all processes are included in the assessment. The assessment must include all 
activities that affect product quality at corporate, business, manufacturing, distribution, 
contractors, or joint venture sites. Processes controlling incoming materials 
from vendors, laboratory services, contractual support, and other inputs should also 
be included in the initial assessment. 
Upon identifi cation of the processes required to support the enterprise, the next 
step is to defi ne exactly what is in and out of scope for each process. Mapping all 
the processes and their inter relationship with other processes will determine if any 
gaps or duplication exists in the system. Duplication may be warranted or eliminated. 
Gaps between processes require remediation. For example, a nonconformance 
process should have direct linkage into a corrective action process. A 
well - operating nonconformance process without an active, integrated corrective/ 
preventive action process will yield little benefi t to the organization and efforts 
expended on the nonconformance process will be nominal in their overall positive 
business impact. 
This comprehensive approach allows for effi cient integration between processes, 
different phases of product life cycle, and integration between different sites in the 
supply chain. This integration provides opportunity for effi ciency in that process 
owners are integrated with each other ’ s needs and expectations. Duplication of 
effort is avoided and effi ciencies gained. Quality outputs from one process become 
reliable inputs into the next process. Management and leadership will have access 
and insight into compliance, infrastructure, and performance metrics of all processes 
on a comparable basis. This provides leadership the opportunity for risk - based 
resource allocation to appropriate areas of the enterprise. 

Process mapping of the enterprise ’ s requirements to supply product enables 
design of the processes required for the system. Staff, management, and leadership 
input into the business needs provide additional guidance into processes 
attributes. 
3.3.4.2 Integrating Quality Management System into Quality Plans 
A quality plan is required by the regulations governing medical devices (QSR) but 
can readily be adopted as a useful tool for pharmaceutical and biopharmaceutical 
manufacturing operations. A quality plan is the documented plan and goals for 
enhancing and advancing the QMS. It can provide the outline and requirements of 
the organization ’ s purpose, mission, product, and business practices used to produce 
a quality product. A quality plan can detail the processes that comprise the QMS, 
the maturity level required for each process, organizational structure, and other 
requirements needed to meet the organization ’ s purpose. Included in the quality 
plan are the elements of the business including location, size, products, and expectations. 
It also includes its structure and support functions, values, and other attributes 
of the organization. 
An annual quality plan can be the detailed execution plan of the organization ’ s 
long - term quality vision for the QMS. It provides staff and management the outline 
and goals for improving the QMS. It enables employees to see the big picture, how 
they fi t into the organization, and the organization ’ s expectations. Within the quality 
plan attributes of the QMS should be described, including functional management 
responsibility. This then becomes the foundation for further defi nition of processes, 
description of management review and responsibility, and continuous improvement 
programs. The preparation of a quality plan begins defi ning what is assumed to be 
known by all levels of the organization. It is the mechanism for ensuring requirements 
are addressed and gaps in the organization do not exist. 
A quality plan may outline the organization ’ s long - term (several years) and 
short - term (annual) goals through a risk - based approach to improving product 
quality. It is the foundation for the manufacturing structure and support processes. 
A quality plan ensures integration of personnel, their qualifi cations, product requirements, 
quality management system, and regulatory and compliance infrastructure. 
An example of an outline of a quality plan is in Table 4 . Leadership review and 
approval of the quality plan is required to ensure that mission, scope, expectations, 
and division of labor in the organization is consistent and supported. 
In larger organizations, site or suborganization - based quality plans can be 
designed to support the scaling of the QMS across all components of the enterprise. 
The individual site plans provide focus on process challenges that are more critical 
than at other sites due to variations in business and compliance environments. While 
the specifi c plans emphasize goals based on site priorities, they also connect 
the members of an organization to the mission of the greater QMS, as shown in 
Figure 5 . 
3.3.4.3 Determining Process Resolution Requirements 
Leadership expects cost - effi cient reliable results from their manufacturing operations. 
Management requires a capable workforce, equipment, facilities, and materials 
to manufacture the product. Employees require robust processes that are easy to 
ESTABLISHING QUALITY MANAGEMENT SYSTEM SCOPE 259

260 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
execute to perform their jobs. All this needs to be considered in the design of the 
processes that comprise the quality management system. 
Complex processes may need to be managed as distinct subprocesses in order to 
provide process owners the ability to accomplish their work with specifi c focus and 
expertise. Management and leadership may require data and metrics on specifi c 
areas of the process that are not available if the process is too complex and large. 
Dividing a complex process into simpler, more manageable processes also allows 
for scalability and transferability throughout the organization. 
Once processes have been defi ned for the enterprise, suffi cient system resolution 
should be determined. This is accomplished by evaluating the ability of the process 
owner to manage and execute the process requirements. Another factor in this 
determination is the data and metrics needed from the process by management and 
leadership. An example of a complex process that benefi ts the organization by being 
managed through distinct subprocesses is validation. 
Validation is a regulatory requirement and has become an industry standard for 
ensuring product consistently meets quality attributes and regulatory requirements. 
Validation requirements are woven throughout the manufacturing supply chain 
encompassing many different subprocesses. The validation process may best be 
TABLE 4 Elements of QMS Annual Plan 
Element Defi nition 
Introduction Purpose of plan and defi nitions for clarity 
Plan Planned activities for the calendar year 
Goals Specifi c/cascading goals of the site 
Projects Major projects in support of the goals 
Metrics Key metrics with defi ned targets 
Approvals Site/plant management 
FIGURE 5 Scaling the QMS through site quality plans. 
Site 2 
Quality 
plan 
Site 6 
Quality 
plan 
Site 5 
Quality 
plan 
Site 7 
Quality 
plan 
Site 4 
Quality 
plan 
Site 3 
Quality 
plan 
Site 8 
Quality 
plan 
Site 1 
Quality 
plan 
Corp. 
Quality 
Plan

managed by dividing it into manageable subprocesses. This allows for effi cient management 
and execution of the subprocesses, and the metrics reported for those 
subprocesses are meaningful and specifi c. See Figure 6 , which illustrates one potential 
organization of the validation subprocesses. Subprocesses contained within the 
validation system could be cleaning, computers, automation, analytical, packaging, 
process, transport validation, etc. Manufacturing is another example of a large, 
complex process that may best be subdivided to support better management and 
more meaningful metrics to management. 
By dividing a larger process into manageable and specifi c subprocesses, management 
can assign appropriate subject matter expertise to lead and manage each 
subprocess. The metrics measuring subprocess performance can be uniquely 
reviewed, evaluated, and compared to similar subprocess metrics at other sites or 
companies. Valuable, meaningful comparisons can be obtained for process and subprocess 
performance that would otherwise be blinded or diluted, if they were summarized 
within the higher level process metrics. 
An additional advantage of establishing subprocesses is that it affords the opportunity 
for rapid assimilation and transfer of the subprocess at various sites within 
the enterprise. An example of this is the comparison of a bulk manufacturing facility 
with that of a distribution center. Both will need to implement aspects of the subprocess 
“ transport validation, ” however, the distribution center will not need to 
implement other subprocesses such as process or packaging validation. As the subprocess 
transport validation is designed and implemented at one site, that infrastructure 
and knowledge transfer to the other site is rapid, avoiding duplication of efforts. 
Sharing of information and expectations of the two sites becomes a common goal 
and format. Management can, therefore, compare transport validation needs and 
maturity levels between sites equally. 
Once all the processes and subprocesses supporting an operation have been 
defi ned, another gap analysis may be conducted to ensure that there are no assumptions, 
and all required processes and subprocesses required to support the business 
are included in the scope of the QMS. This can easily be accomplished by listing all 
the business drivers for an operation and comparing that against the processes 
FIGURE 6 Validation subprocesses. 
Computer 
validation 
Transport 
validation 
Packaging 
validation 
Analytical 
validation 
Automation 
validation 
Process 
validation 
Cleaning 
validation 
ESTABLISHING QUALITY MANAGEMENT SYSTEM SCOPE 261

262 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
established. Written defi nition of the process and subprocess scope is required. 
Stakeholders, owners, and users of the processes should be involved to ensure clear 
defi nition and understanding of process scope. Questions to be asked are: Are all 
the business needs addressed? Have all our activities and operations been included 
in the assessment? Is there any duplication in process expectations? Are there any 
gaps between process outputs and inputs? Once this evaluation has been concluded, 
it can be easily determined if any existing work process have been overlooked and 
the system requires further modifi cation. 
3.3.4.4 Scalability to Enterprise 
Well - designed processes and subprocesses are scalable to the enterprise. A comprehensive 
design will allow for replication and comparison of processes and subprocesses 
between multiple sites. This allows for rapid implementation of new technology, 
sharing of best practices, and comparison of similar metrics to determine compliance, 
infrastructure, and performance. A comprehensive system allows for each unit 
operation or site within the enterprise to have the fl exibility to apply applicable 
processes and subprocesses, yet continue operating within the defi ned structure of 
the QMS. For example, a manufacturing site may utilize almost all of the processes 
discussed in the validation process, whereas a distribution site may only utilize the 
transport process. Both sites, however, implement the same structure for the transport 
process, allowing for meaningful comparison of data and metrics and rapid 
implementation of any required changes to that process. 
Well - designed quality management systems support structured organic growth 
and are valuable in evaluating and integrating manufacturing acquisition opportunities. 
Business and manufacturing management should utilize the QMS and its standards 
whenever evaluating external facilities for appraisal, approval, integration, or 
expansion. Meaningful metrics obtained from a QMS provides the standard to 
make critical decisions affecting multiple internal or external manufacturing 
capabilities. 
Documented process structure provides rapid employee assimilations when 
transferring employees between sites. New employees, replacing existing process 
owners, are enabled to rapidly execute process responsibilities due to the abbreviated 
learning curve when processes have been well defi ned and documented. Systems 
designed as described here provide meaningful and comparable metrics for leadership 
to evaluate progress, compliance, and performance. 
3.3.5 SYSTEM AND PROCESS OWNERSHIP: 
ROLES AND RESPONSIBILITIES 
A well - designed QMS and the processes that comprise it require competent ownership 
with defi ned roles and responsibilities for program success. This combination 
ensures that the system and processes are established, maintained, improved, and 
remain current with industry practices and business expectations. Operational 
execution of the QMS and the processes comprising it will engage stakeholders, 
management, and leadership, provide business results, and support and ensure 
compliance. 

3.3.5.1 Quality Management System Ownership and Management 
The QMS is best owned at the highest level in the organization. At a minimum it 
should be owned at a level in the organization above manufacturing and quality. 
The owners ’ main responsibility is to champion the program and ensure organizational 
alignment. Regulatory investigators expect processes supporting manufacture 
are fully incorporated into the QMS. They also expect leadership to have signifi cant 
knowledge of the operations and interact with investigators during inspections with 
some degree of familiarity with the processes supporting manufacture. At the conclusion 
of an inspection, regulators issue inspectional fi ndings and, if appropriate, 
take regulatory action against the most senior member of the leadership group. 
Through high - level leadership ’ s active involvement and ownership, the QMS 
program and enterprise will be successful. 
As mentioned previously, the QMS is best managed by a group dedicated to the 
program. The QMS program offi ce should have defi ned roles and responsibilities. 
In the FDA regulation 21 CFR Part 211.22, the responsibility of the quality unit is 
described. It is the only functional group in a manufacturing organization that has 
its job description codifi ed in federal regulations. These responsibilities should not 
enable or dilute the responsibility for ensuring quality of other functional groups in 
the organization. All functional groups supporting manufacture should be applying 
their trade to the GMP world. Regulators expect the Quality Department to have 
oversight and approval of all processes affecting product quality. Program management 
is important because of the need for coordination and accountability to bring 
individual processes, long - term system strategy, yearly quality plans, and goals 
together to accomplish the program ’ s objectives. These activities and benefi ts cannot 
be realized from individual process owners. 
The program management of the QMS can be managed just as a process, with 
predefi ned expectations, metric collection, and management review, culminating in 
risk management application to continuous improvement programs. These metrics 
and improvement initiatives need to be vetted through leadership review and input 
to ensure alignment throughout the organization. 
An outline for the roles and responsibilities for the QMS program offi ce is illustrated 
in Table 5 . By establishing the roles and responsibilities of the program offi ce 
a defi ned point of contact and accountability is established for program execution. 
It establishes strong linkage and focus on the program objectives for process owners, 
training, functional management, and leadership. Similar program management 
structures are required at manufacturing sites and corporate functions to maximize 
benefi ts of the program through establishing common and specifi c goals and 
TABLE 5 QMS Program Offi ce Roles and Responsibilities 
Subject matter expert for QMS program 
Develop and execute communication plan 
Initial and ongoing training 
Facilitate management review process 
Identify process maturity goals and metrics 
Develop long - term strategic vision 
Create and execute annual action plan 
SYSTEM AND PROCESS OWNERSHIP: ROLES AND RESPONSIBILITIES 263

264 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
providing a platform for sharing best practices and knowledge. As organizations 
grow in complexity, additional management may be required to ensure that the 
elements of the QMS are integrated, functioning, and delivering the results 
expected. 
3.3.5.2 Process Ownership 
Designing a QMS that mandates and assigns process ownership to designated individuals 
is a signifi cant strategic decision in the establishment of a successful quality 
management system. It provides effi ciency, expertise, dedication to the process, and 
focused ownership for documentation, improvements, benchmarking, and compliance. 
Without defi ned and assigned process ownership functional management 
becomes the de facto process owner. This is problematic in that functional management 
is already overburdened with personnel and business management issues, 
unable to adequately focus and deliver the demanding process owner requirements 
required in today ’ s manufacturing environment. Several processes typically are 
organized under an individual functional manager, further diluting focus, attention, 
and expertise if functional management is relied upon as a process owner. 
3.3.5.3 Process Owner Selection 
Process owner selection requires program management to establish defi ned criteria 
for the selection process. Criteria include the capability to perform process owner 
roles and responsibilities, including self - development and decision making. Empowered 
process owners are accountable for maintaining and executing the processes 
that management relies upon to deliver business results. This accountability ensures 
that staff, management, and leadership know who to solicit for answers to process - 
related questions and issues. It also provides the best representation to regulators, 
clients, and customers. Effi ciencies are gained and current trends maintained with 
an active owner, with defi ned responsibilities. 
Selection criteria may include attributes of technical, interpersonal, and management 
skills. The capabilities needed for different process will vary and should be 
considered in the selection process. At the end of the selection process, functional 
management and the process owner may consider inclusion of the process owner ’ s 
roles and responsibilities into the process owner ’ s job description. Personal goals 
and development activities should be based on improving the process owner ’ s capabilities 
to manage the process and develop future process owners through active 
mentoring and talent development programs. 
Process owners need to be dedicated to their process. They must be empowered 
and held accountable for all the attributes listed in their roles and responsibilities. 
Process owners may have ownership of more than one process and may have other 
job responsibilities, but it must be clear throughout the organization as to who has 
full authority for the process. 
Process owners require a defi ned set of responsibilities to maintain a vibrant and 
effective process that continues to support product quality deliverables. Having 
roles and responsibilities defi ned provides owners with the structure and parameters 

needed to be effective. Examples of owner responsibilities include identifi cation of 
stakeholders, defi ned decision authority, document ownership, nonconformance 
ownership, knowledge of regulations and industry trends, subject matter expertise, 
training content, metric ownership, and representation to internal auditors and 
external regulators. Identifying, training, and development of the process owner on 
his or her roles and responsibilities is similar to assembling the piece of a puzzle. If 
one piece is missing, the effectiveness of the process owner will be minimized (see 
Figure 7 ). Developing a sound methodology for process owner selection ensures 
objectivity and is critical to the success of the program. A brief discussion on each 
process owner roles and responsibilities follows. 
3.3.5.4 Stakeholder/Process Owner Integration 
Process owners must identify the stakeholders of their process and ensure design 
and output of the process meets stakeholder needs. Regular communication, interaction, 
and support are maintained with stakeholders through scheduled meetings 
to discuss process status and improvements. Any changes to the process are vetted 
through the stakeholder group. Typical stakeholders include the QA unit that is 
responsible for review and approval of the process components, suppliers to and 
receivers of the process (i.e., owners of other processes that interact with the 
process), customers, management, and leadership. Each stakeholder roles and 
responsibilities also require defi nition. 
Including the key stakeholders in decisions affecting process design or changes 
ensures efforts by the process owner are applied correctly. Process robustness is 
dependent upon meeting business and customer needs, and the process owners 
require input and support from the stakeholder group. 
FIGURE 7 Process owner roles and responsibilities. 
Accountable 
ownership 
Risk 
management 
Nonconformance 
& corrective 
action 
ownership 
Inter/Intra 
process 
expertise 
Metrics 
& process 
improvements 
Inspection 
& audit 
point of contact 
Stakeholder 
management 
Document 
& training 
content 
Decision 
authority 
SYSTEM AND PROCESS OWNERSHIP: ROLES AND RESPONSIBILITIES 265

266 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
3.3.5.5 Decision Authority 
Each process owner requires a defi ned level of decision authority. This authority 
level delineates the bounds of decision making granted by the organization to the 
process owner. Business needs and risk assessment must be incorporated into the 
design of the decision authority granted to a process owner. Table 6 is an example 
of a decision authority matrix design for a process owner. It requires cross - functional 
management support to be effective. 
Preparing a decision matrix that is shared and agreed upon by the stakeholder 
group and functional management ensures decisions are made and communicated 
quickly by appropriate persons. It removes the burden of making every technical 
process decision from functional management. It is important to outline the process 
owner ’ s role in the decision - making process, as well as conditions for escalation. 
Effective process management is realized when the culture of an organization can 
support the outline of the decision matrix and not continually rely upon functional 
TABLE 6 Decision Authority Matrix 
Decision 
Category Defi nition 
Decision 
Maker 
Decision 
Support 
Required 
Informed of 
Decision 
Company 
standards 
Process - related 
global 
standards 
relevant to all 
manufacturing 
sites 
Corporate 
process 
owner 
Site process 
owners 
Site quality 
assurance 
counterpart 
Process 
stakeholders 
Impacted staff 
Standard 
operating 
procedures 
SOPs related to 
the specifi c 
process 
Site process 
owner 
Process 
stakeholders 
Site quality 
assurance 
counterpart 
Management 
review 
Corporate 
quality 
assurance 
counterpart 
Impacted staff 
Training Training on 
processes or 
procedures 
Corporate and 
site process 
owners 
Training 
Technical 
system 
matter expert 
Process 
stakeholders 
Corporate 
quality 
assurance 
counterpart 
Impacted staff 
Site projects All projects 
related to the 
existing 
process or the 
projected 
improved 
state of the 
process at a 
specifi c site 
Site head Process owner 
Leadership 
team 
Site project 
portfolio 
manager 
Process 
stakeholders 
Corporate 
quality 
assurance 
counterpart 
QMS offi ce 

management. If functional management continues to be relied upon and seen as the 
process decision makers, efforts and progress by the process owner will be nominal. 
The organization ’ s culture must support the process owner at all levels of the enterprise 
for the owners to be successful. 
3.3.5.6 Industry Knowledge 
The c in c GMP represents the notion of current industry practices. Process owners 
must work to remain current with industry and regulatory trends affecting their 
process and its overall effect on the QMS and the business. Awareness of process 
capability and comparability with other like processes within and outside the pharmaceutical 
and biopharmaceutical industry is essential. Regulators will compare an 
owner ’ s process against other similar processes in which they have experience when 
formulating value judgments. Benchmarking against similar processes provides 
process owners the data needed to determine adequacy of their process with industry 
peer groups. 
Where technology, effi ciency, performance, or compliance can be enhanced, it 
should be considered by an aware and informed process owner. Functional management 
cannot keep pace with the changes occurring with all the processes supporting 
manufacturing. Ensuring process owners dedicate suffi cient time to keeping current 
with process - related external events will ensure process success. This may include 
review of industry periodicals, attendance at seminars and regulatory presentations, 
and routine self - evaluation and benchmarking against relative processes. 
Often, the best examples of process effi ciency can be found outside the pharmaceutical 
and biopharmaceutical industries. Other fi elds such as electronics, space, 
and software industries have evolved their documentation, training, quality, and 
change control systems to the point of best in class. These industries are more time 
sensitive to get product to market and have often evolved their processes to be 
effi cient and decision processes to be very quick. Process owners may expand their 
knowledge by investigating other industries to fi nd best practices and apply them 
internally. 
3.3.5.7 Regulatory Inspection and Audit Lead 
Process owners play a critical role during regulatory inspections and customer 
audits. Process owners are the best choice to represent process attributes and performance 
to interested parties. Process owners provide regulators and auditors with 
a capable, knowledgeable resource to represent the process and answer detailed 
questions. The process owner should be aware of process history, requirements, 
operations, exceptions, changes, and nonconformances. The process owner will have 
detailed knowledge of process operations, compliance, and be able to defend the 
process. Providing an accurate answer the fi rst time to regulators and auditors is 
essential in building trust and representing competence. 
Each process owner is required to work closely with his or her QA counterpart. 
This ensures design and operational issues are clearly reviewed and approved by a 
representative from the quality assurance function, a regulatory expectation. The 
quality assurance counterpart must be familiar with the process, understand documentation 
supporting the process, and able to convey what approval the Quality 
SYSTEM AND PROCESS OWNERSHIP: ROLES AND RESPONSIBILITIES 267

268 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
Department has conveyed on the process and meaning of that approval. The QA 
counterpart to a process should also have defi ned and documented roles and 
responsibilities. 
When teamed together, the process owner and the quality assurance counterpart 
for a process will make a favorable impression upon regulators, be able to 
explain all operations involving the process, supporting documentation, and any 
ongoing projects or process improvements. This pair is the best to evaluate and 
consider any deviations to the process or recommendations for continuous 
improvement. 
In most all cases, the process owner and QA contact will posses more information 
about the process than regulators and be capable to defend the process design and 
operation. Should regulators have suggestions on process design or functionality, 
the process owner may consider them. If appropriate, any recommendations or 
observations made by regulators or auditors can be incorporated into the process 
design. However, it is critical for the process owner and stakeholders to evaluate 
proposed changes to avoid reactive management commitments, which could be 
deleterious to the effi cient operation and output of the process. 
3.3.5.8 Subject Matter Experts 
Process owners, through their selection and development, become the subject matter 
experts for the process. It is more effi cient for an enterprise to focus its expertise 
on individuals that have the authority and accountability described in this section, 
rather than dilute those attributes and accountabilities, thereby risking poor process 
execution and management. 
As a subject matter expert on the process, the owner has the capacity to deliver 
results outlined in the list of owner responsibilities, mentor future process owners, 
assist in staff development, and accurately guide management in its strategy related 
to the process. The process owner ’ s personal development of his or her process 
expertise is essential in delivering operational results and providing direction for 
future strategic changes to the process. 
3.3.5.9 Metric Ownership 
Process owners ’ responsibilities include determining appropriate metrics for their 
process. These metrics should include lagging and leading metrics that are meaningful 
to the process owner and management in determining performance, compliance, 
and infrastructure of the process. 
The process owner should represent and interpret these metrics to the organizations 
leadership. The metric output from a process is the basis for management 
and leadership ’ s action in resource deployment and approval of continuous improvement 
projects. Key operating parameters such as number of nonconformances and 
regulatory observations against the process should be tracked and factored into the 
maturity of the process. 
Every process owner needs to base their continuous improvement plan for the 
process based upon metrics collected from the process output. The metrics must be 
designed to assist in these decisions and be readily available for review, presentation, 
and interpretation. 

3.3.5.10 Documentation Ownership 
Process owners are the most appropriate owners for all documentation supporting 
their processes. This includes having either direct ownership or controlling infl uence 
over guidance and execution documentation such as corporate policies and standards, 
local requirements and standard operating procedures (SOPs), logs, and 
records. 
For the manufacturing process owner, this means owning the master manufacturing 
records and executed batch records, SOPs, use logs, and related training documents 
for their process. Combining responsibilities for process management and 
process ownership results in true accountability for the process owner. It also allows 
for progress and continuous improvement of the QMS. Removing questions 
of responsibility and accountability ensures integration between requirements 
(standards, policies, procedures) and execution (training, performance, and 
documentation). 
3.3.5.11 Training 
Assurance of adequate training for process users is an important responsibility of 
a process owner. Process owners must have a clear understanding of the requirements 
of their process and its operation. This understanding requires translation 
into executable training. Users must be able to understand and apply the training. 
Complicated processes coupled with ambiguous training will lead to confusion and 
an inability to properly execute a process, which eventually constitutes failure for 
the organization. A simple process, with easy to understand process steps, that are 
consistent with instructions and documentation requirements will support success, 
reduce production costs, minimize nonconforming events, and allow for employee 
satisfaction. 
Process owners are subject matter experts and should infl uence and provide 
consulting for training on the process. They may also participate in training delivery. 
Ensuring adequate training on a process is a key goal for system effi ciency and 
regulatory compliance. Process owners, capable of explaining the reasoning behind 
the process requirements, enhance the training experience for process users. Process 
owners should include effective presentation and training skill development into 
their personal development programs. 
3.3.5.12 Risk Management 
Process owners require basic understanding of risk management and its application 
to process design and continuous improvement prioritization. Several industry and 
regulatory resources exist, such as ICH Q9, that provide understanding on risk 
assessment, identifi cation, control, methodology, and the overall risk management 
process. Process owners should be familiar with risk management techniques and 
tools and apply them to their process management when designing, executing, or 
managing improvement efforts for their process. 
Risk management is especially important for the presentation of process improvement 
proposals to management where resources are required. The ability to quantify 
risk and demonstrate continuous improvement benefi ts is essential to project 
SYSTEM AND PROCESS OWNERSHIP: ROLES AND RESPONSIBILITIES 269

270 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
and resource approval. Risk analysis, management, and presentation constitute 
guiding leadership to work on the right things at the right time and then improve 
it. 
3.3.5.13 Continuous Improvement and Project Management 
Instituting quality - by - design efforts early in the design of a process should negate 
the need for major process improvements. However, over time, due to business 
needs, regulatory changes, or technology improvements, processes will require some 
form of change to ensure compliance or performance enhancement. As part of 
executing and maintaining their processes, owners need to collect and report performance 
metrics to management and staff. These metrics will inevitably direct 
attention to opportunities for improvement that require capital and human resources. 
Process owners are the best leaders of continuous improvement projects due to their 
intimate knowledge of the process and accountability for process output. 
Managing or leading a continuous improvement project requires process owners 
be knowledgeable in project management and team - leading skills. Improvement 
projects typically require cross - functional support and expertise from areas such as 
information systems, project management, manufacturing, engineering, and development. 
It is essential that continuous improvement efforts name the process owner 
as the project lead to ensure the required output of the project meets the process 
owner, stakeholders, and enterprise needs. Often, projects are completed and 
declared a success, delivering a substandard result that the process owner, users, and 
stakeholders fi nd inadequate to meet process requirements. 
Upon completion of continuous improvement projects, process owners ’ responsibilities 
include monitoring the changes made to the process to determine the 
impact of improvements. Metrics monitoring changes to the process, pre - and postimplementation, 
should be incorporated into existing performance metrics and 
reported during regular management reviews. 
3.3.5.14 Non Conformance/ CAPA /Planned Deviation Ownership 
An important barometer of process performance is the number of nonconformances, 
corrective and preventive actions taken, and planned deviations initiated 
against a process. These types of process artifacts must be known and owned by the 
process owner and stakeholders. The process owner must consider these process 
metrics for evaluation of and changes to process design, training, documentation, 
and performance. 
Nonconformances may fall into the category of manpower, machinery methods, 
materials, etc. Employees not following procedures or unable to execute required 
steps of the process indicate a poorly designed process requiring modifi cation and/or 
improved training. Machinery failures often indicate poor qualifi cation, validation, 
calibration, or maintenance programs. Unexpected results or outcomes are 
indicative of poor process design, characterization, or a break down between 
processes. 
Although planned deviations are frowned upon by many in the industry and 
regulators, there are times when temporary changes to a process must be employed 
to support the business. Permanent changes must be made through a formal change 

control process. When the use of a planned deviation is required, the affected 
process owner should be aware of and own the change. This provides owner control 
over the duration and extent of the change to the process and provides data for 
possible consideration in making a permanent change to the process. A planned 
deviation should be rare and monitored closely as it affects previously established 
standards, expectations, and training. 
Process owners must be capable to evaluate and interpret the effect of nonconformances 
and planned deviations on their systems. Process owners can evaluate 
the need and lead efforts for corrective or preventive action, ensuring adequate 
corrections and improvements are implemented. An effective QMS ensures deviations 
from approved processes are owned and adequately investigated by the process 
owner ’ s and ultimately approved by their quality assurance counterpart. The knowledge 
of these events is the basis and foundation for the process owners to make a 
risk - based evaluation on whether or not process changes are required, documentation 
or training require modifi cation, or continuous improvement efforts are 
warranted. 
A well - designed QMS will include identifi ed process owners with defi ned roles 
and responsibilities. Process owners require support from management, their customers, 
stakeholders, and quality assurance. Accountability and decision - making 
parameters will empower process owners to drive execution and improvements to 
their process, delivering the business results expected. Without these process owner 
attributes and support, minimal results will be achieved, and functional management 
will be burdened with and assume the responsibility for making decisions that 
should be in the hands of capable process owners. 
3.3.6 CHANGE MANAGEMENT/COMMUNICATION 
Establishing and maintaining an effective QMS, as this chapter describes, 
requires a signifi cant cultural shift. Many employees and functional management 
will fi nd the business transformation of defi ning processes, assigning ownership, 
delegating authority, and responsibility for process performance within the QMS 
is a signifi cant change in business conduct. The most signifi cant change results 
from the shift of control in process expertise and decision - making authority 
from functional management to process owner. Signifi cant business transformation 
may result by assigning responsibility and accountability to the process owner, 
and management ’ s support of process owners who drive continuous process 
improvement. 
In The Second American Revolution , Rockefeller describes the conservatism of 
organizations: “ An organization is a system, with a logic of its own, and all the weight 
of tradition and inertia. The deck is stacked in favor of the tried and proven way 
of doing things and against the taking of risks and striking out in new directions ” 
[9 ]. If an organization is not already practicing principles of delegation, process 
ownership, established metric collection, management review, and continuous 
improvement, barriers within the organization will need to be addressed and broken 
down in order to establish new behaviors. These barriers to change will exist within 
and between functions, functional management and staff, and possibly between 
companies and regulators. 
CHANGE MANAGEMENT/COMMUNICATION 271

272 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
Although expected benefi ts are signifi cant when implementing a QMS and the 
end result desirable for employees and management, describing the desired 
state and motivating personnel to change and implement new behaviors contains 
signifi cant challenges. A successful business transformation requires a robust 
change management and communication plan that includes support for all staff 
affected. 
3.3.6.1 Managing Organizational Change 
Integration of the skill sets of human resources, training, and change management 
groups will signifi cantly augment efforts toward cultural change and acceptance. 
Often personality profi ling tools are effective to gauge the organization ’ s preferences, 
learning styles, and adoption tendencies. These types of tools should be considered 
in the overall change management program, used where applicable, and 
program modifi cations made based upon their results. 
The fi rst and critical step in developing a successful change management plan is 
to obtain initial support from the corporation ’ s leadership and functional management. 
Without this support the QMS will not gain critical mass and may not deliver 
the desired effects or changes. To acquire this support, the implementation team 
must put together a strong business case that speaks to the leadership ’ s needs and 
wants. The business case must include a risk assessment against compliance and the 
benefi ts of the fi nancial gains. It is important to be honest, consider current system 
status and future requirements, and include a long - term strategy that addresses costs 
and benefi ts. The change management plan must include frequent and repetitive 
communications, to all levels of the organization, of the cost/benefi ts and successes 
expected and realized by the program. 
Functional management support is also critical to the success of a new program 
such as a quality management system. Any time a staff member is asked to take on 
a new role or responsibility, he or she needs to be supported by the functional 
manager as well as leadership within the organization. Corporations are resource 
limited and necessarily need to continually prioritize where to allocate resources. 
Staff will only take on roles or responsibilities that they believe are supported by 
their functional manager in an effort to successfully meet their perceived immediate 
goals. Quantifi able support from leadership and functional management can be 
directly correlated to the success or failure of the QMS program. 
Signifi cant work is involved in training new process owners, functional managers, 
leadership, support organizations, and actualizing their new behaviors. A support 
system must be in place for the process owners, stakeholders, and management to 
guide and reinforce the new behaviors and maintain the process effectiveness. It is 
preferable that this support system be established through a dedicated team 
that can be fully attentive to all their needs. Without a single source to lead the 
efforts, diversity in interpretation and implementation will dilute the program, 
within different functions and sites, and its effectiveness and outcomes will be 
diminished. 
Establishing an organization to lead the systems initiative is important. That 
organization requires management, standards, and parameters similar to managing 
an individual quality process. It requires roles and responsibilities be established, 

metrics be determined, collected, reviewed and acted upon, and receive management 
and leadership visibility and support. The QMS program is best organized as 
a function within the Quality Department and be regarded as an ongoing program, 
not a short - term project or effort with limited shelf life. The group must be led by 
competent persons who are familiar with quality concepts and applications, regulatory 
expectations and requirements, needs of the enterprise, good communication 
and infl uencing skills, and are fl exible and enduring. 
3.3.6.2 Communication 
Trying to get people to comprehend a vision of an alternative future is also a communications 
challenge of a completely different magnitude from organizing them 
to fulfi ll a short - term plan. It is much like the difference between a football quarterback 
attempting to describe to his team the next two or three plays versus his 
trying to explain to them a totally new approach to the game to be used in the 
second half of the season. Aligning the organization to accept and implement a 
system - based approach requires careful messaging coupled with management 
support and results. 
Messages are not necessarily accepted just because they are understood. Another 
big challenge for leadership is credibility and getting people to believe the message. 
Aligning words and deeds supports the worthiness and credibility of the messaging. 
People have learned from experience that even if they correctly perceive important 
external changes and then initiate appropriate actions, they are vulnerable to 
someone higher up who does not like what they have done. Reprimands can take 
many forms: “ That ’ s against policy, ” or “ We can ’ t afford it, ” or “ Shut up and do as 
you ’ re told ” [10] . 
Having established a dedicated team that provides overall program management, 
it is imperative that the team outline a strategic plan for presentation to leadership. 
Without a vision and long - term plan, which is supported by the enterprise leadership, 
quality system initiatives will become diffi cult. The plan needs to be comprehensive 
in nature, yet broad enough to convey purpose, mission, and benefi ts at a 
high enough level to be understood and supported. An outline such as this provides 
framework and direction for the program management team and leadership. It 
also guides the program management team to developing annual goals and quality 
plans that fi t into the overall strategy and provide momentum and results to the 
organization. 
Annual quality plans should be prepared by the QMS program offi ce that 
address the long - term strategy and intermediate goals that come to surface during 
program implementation. Training, changes in regulatory requirements, metric - 
driven projects, and special circumstances warranting process changes such as implementation 
of new technology or programs should be included into the annual 
quality plan. 
Long - term strategy documents and annual quality plans require leadership and 
functional management support and approval. These documents must be reviewed 
and discussed with the leadership of the organization, modifi ed to meet the business 
and regulatory requirements, and then have full support through upper management 
approval. In this way, the goals are being led by top leadership and management 
and not any individual group in the organization. Once top leadership signs 
CHANGE MANAGEMENT/COMMUNICATION 273

274 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
onto the program, it can be shared throughout the organization in a number of 
ways. 
If leadership can support the long - term strategy and annual quality plans to 
accomplish the vision, then the foundation for change management and cultural 
shift is in place. Leadership will need to continually discuss the need for systems 
implementation, in front of a variety of audiences. This includes leadership staff 
meetings, management, and employee meetings. The importance of leadership 
support cannot be overlooked. Without consistent visible leadership and management 
support process owners and staff will revert to old behaviors, become reactive, 
and perhaps unrelated in their process integration efforts. Leadership needs to 
require aspects of the program be included into functional management annual 
objectives with defi ned deliverables outlined and evaluated. Likewise, functional 
management should require staff to include appropriate aspects and objectives of 
the QMS program into their individual goals and work to accomplish them. 
3.3.6.3 Feedback and Alignment 
Managing the changes required to fully implement a QMS can include several forms 
of communication and feedback. A detailed annual communications plan can aid 
the QMS ’ s group in identifying specifi c target groups, methods, and frequencies of 
communications, messaging types, and feedback mechanisms to monitor progress 
for program modifi cations. Table 7 is an example of an annual communications plan 
that supports efforts to keep internal audiences informed, aligned, and engaged. 
Each target audience requires specifi c messaging that connects with its needs. 
Failure to get the appropriate message, that is, what is the program bringing to them, 
will minimize support for the program. This plan should include face - to face and 
written communications addressing multiple audiences and media types. Face - to - 
face meetings can include presentations to steering committees, process owners, 
functional departments, and all staff meetings. Written communications can include 
sitewide communications, poster sessions, and newsletters. The communications 
should speak to all audiences — “ what ’ s in it for me? ” Topics can include leadership ’ s 
commitment (direct quotes or actions taken); spotlight on successes (real - life stories 
from process owners); impact to the site (process improvements or risk mitigation); 
and progress to the program (metrics and successes). The progress and success of 
the QMS cannot be overcommunicated. 
Another useful tool to help the message and modify the program is the use 
of a feedback survey. If properly designed and distributed to a defi ned set of stake- 
TABLE 7 Communication Plan 
Vehicle Communication Type Frequency Date 
Functional metrics meeting Face to face Monthly First week of month 
Management interviews Face to face Annually January 1 – 31 
QMS newsletter Written Quarterly First week of 
quarter 
All staff meeting Presentation Semiannually March & September 
Poster session Written/face to face Annually July 

holders and employees, the survey can provide valuable insight into how staff and 
management view the program, its progress, and suggestions for modifi cations. 
If surveys are distributed electronically and offer only one - way communication, 
the benefi ts may be limited as the respondents are limited in their ability to fully 
convey their impressions or offer effective feedback. An electronic feedback survey 
may be a fi rst good step in understanding the thoughts and concerns of the 
stakeholders. 
Another suggestion or follow - up to the electronic survey is to utilize focus groups 
that have the ability to interact with the program questioners. This two - way conversation, 
verbal dialogue, allows further understanding of the program by the participants 
that follows with more meaningful feedback to the program administrators. 
Focus groups should be selected at different levels within the organization, including 
process owners, stakeholders and users of the system, leadership, functional management, 
and the general populace of employees. Focus groups provide valuable 
input into programs that the program administrators may be unaware of and can 
provide program redirection. 
Once suggestions are received on the program, it is essential to consider and 
incorporate those ideas and modifi cations that make sense to implement. Those 
changes need to be communicated and seen by the focus group members to ensure 
that their time and effort has not been wasted and their suggestions have been 
heard. This is one of the best ways to spread the word about the QMS program and 
garner grassroots support. 
3.3.6.4 Training 
A training plan should be developed to identify the needs of the staff and affected 
functional areas required to support the successful implementation of a QMS. It is 
the responsibility of the corporation to adequately support staff with training and 
tools when staff is expected to take on new roles, responsibilities, or behaviors. The 
training plan should consist of targeted training for general staff, process owners, 
and functional management of the process owners. 
At a minimum, all staff should be introduced to the purpose, goals, and requirements 
of the QMS. This training should be a high level explanation of the program 
looking to gain understanding and support for the program by communicating why 
it is important and what are the risks of not adopting the program. This can be 
accomplished by instructor - led training or an electronic, Web - based learning module 
depending on the size of the corporation. 
Process owners require more comprehensive levels of training to fully understand 
their role and responsibilities within the program. Process owner training 
should teach key concepts and tools that owners will need to evaluate and support 
their processes. This training can be done in a phased approach to support the elevation 
and advancement of a process within the organization ’ s chosen maturity 
model. 
Training should be provided to offer functional managers supervising process 
owners a thorough understanding of the QMS. Training should address new roles 
and responsibilities of staff, time demands on process owners, overall program 
timelines, and impact to functional areas. The acceptance and support of functional 
management is critical to successful implementation of a QMS. 
CHANGE MANAGEMENT/COMMUNICATION 275

276 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
Managing organizational change demands a well - written strategy, skill set, and 
resources to ensure changes that come with system implementation are understood, 
supported, and maintained. Starting with high - level overviews of system design, 
benefi ts and timelines for implementation are the foundation for management 
understanding and support. Detailed annual quality plans can be the tactical vehicle 
for program implementation. Leadership support through understanding and 
approval of the annual quality plan, inclusion of program objectives into management 
goals, and frequent verbal and visual support of the program are essential to 
success. Building the program infrastructure is a signifi cant undertaking. Inclusion 
of a comprehensive training, communication, and change management plan should 
be built into the overall goals of the program and routinely evaluated and 
delivered. 
3.3.7 MEASURING SUCCESS THROUGH MEANINGFUL METRICS 
Successful implementation of a comprehensive QMS can be determined by the 
establishment of a meaningful metrics program. The purpose of a metrics program 
is twofold: fi rst, to allow an organization to evaluate its progress toward meeting its 
goals in an objective, data - driven manner and, second, to monitor the performance 
of each process to ensure continuous improvement. By evaluating metrics for the 
QMS and its processes, the enterprise has the knowledge and understanding of the 
overall health of its system and processes and can develop strategies based on risk 
for continuous improvement of the system and processes. 
Once the metrics program is in place, the system and process metrics require 
visibility to process owners, upper management, and stakeholders. Process owners 
require understanding of the metrics ’ trends, issues, and associated risks. Stakeholders 
must work with the process owner to identify and propose process improvement 
opportunities. Leadership is accountable to understand the issues and associated 
risks and responsibly apply resources for remediation efforts. 
3.3.7.1 Performance Metric Development 
Quality and business indicating metrics should also be reviewed on a routine basis. 
These may include the following: 
• Quality indicating: ability to meet quality standards and procedures 
• Supply: ability to meet demand 
• Cost: savings as well as avoidance 
• Safety: near misses and incidents against process 
The guiding principle of metric development is to have a stable system or process 
to collect, review, and draw conclusions. All metrics should be developed with stakeholders 
input taking into account the requirements and needs of the customers. This 
includes the touch points of the downstream quality processes. Without this input 
and understanding metrics may be developed within a silo and hold little value, 
causing both frustration at the leadership as well as the staff level. Without proper 

design, metrics may become a check box activity that results in minimal or no action 
by management to support efforts by a process owner. 
Metrics can either fall into one of two categories: lagging or leading indicators. 
Both types are important to the process owner and management. Lagging indicators 
are metrics that represent the process ’ s ability to deliver results or outputs. They 
indicate the performance of the system in the past. They can assist process owners, 
management, and leadership in determining if goals have been met, objectives 
attained, or existing standards or expectations have been met. Leading metrics focus 
on the inputs and suppliers of a process. These metrics are important indicators to 
proactively allow owners and management to take action on a process prior to violating 
a standard, objective, or goal. A successful process owner will understand the 
relationship of leading metrics and their affect on the lagging metrics and process. 
Metrics need to be designed to meet the needs of the organization, be simple to 
track and present, and be regularly reviewed. 
3.3.7.2 Metric Review 
Ignorance of system and process performance leads to ineffi ciency, poor compliance, 
and low employee morale. It is good business practice to have regular review of 
process metrics to gauge the health and output of the system and processes that 
drive the organization. 
Process owners should be aware of all the metrics affecting their process and 
have a conduit to present the critical metrics to upper management. There are 
examples in the industry where process owners responsible for execution of a 
process are not aware of the metrics being collected, if any are, and have no basis 
for judging the adequacy of their process or its performance. 
Regulatory agencies hold management accountable for the operations of an 
organization. It is the fi duciary responsibility for process owners to share the output 
and performance of the operation with management and be able to explain and 
interpret those metrics. Management has the responsibility to know the operations, 
its performance, and take appropriate action to ensure compliance with government, 
industry, and company policies and regulations. 
Regulations require an annual product review be conducted of pharmaceutical 
products to determine and assess changes made to processes that may affect product 
quality. However, good industry practices would mandate quarterly or monthly 
review for faster detection, decision, and action. Reviews need to include metrics on 
key operating parameters and critical quality attributes to ensure product safety and 
effi cacy. Several other key business metrics also benefi t the organization and should 
be included in the metrics review program. The metrics collected should easily 
provide the process owner and management with an indication if the process is in 
control and delivering the desired results. If not, the process owner needs to present 
management a proposal to pursue continuous improvement opportunities and be 
able to describe required changes necessary to realize process enhancement. 
3.3.7.3 Maturity Model 
A maturity model is a useful management tool to determine process status and 
provides a standard in which to value processes. It provides a standard in determin- 
MEASURING SUCCESS THROUGH MEANINGFUL METRICS 277

278 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
ing the overall robustness and progression of a process and assists in the determination 
of resource prioritization. It provides the basic framework to apply risk 
management in determination of process development. For example, development 
of high - risk systems, such as aseptic fi lling where high patient and business risk exist, 
should be developed to a higher maturity than other processes with less patient or 
regulatory risk. Business demands placed on the pharmaceutical and biopharmaceutical 
industry limit resources in development, quality, and manufacturing requiring 
wise deployment of these resources to the areas that can best benefi t the 
organization. 
An example of a maturity model can be seen in Figure 8 . This example provides 
the QMS program group and leadership the ability to evaluate processes based on 
an objective standard. It is divided into fi ve general levels, moving from informal, 
unstructured to best in class. It includes specifi c deliverables for each level of the 
model to be completed before a process can be considered to have achieved that 
level. This model can also be divided into distinct subcategories, for instance, infrastructure, 
performance, and compliance, which are depicted in Table 8 . Each subcategory 
can be designed to provide meaningful information to the process owners 
FIGURE 8 Maturity model overview. ( Source : Adapted from Capability Maturity Model 
Integration, www.sei.cmu.edu . ) 
TABLE 8 Example Maturity Model 
Theme 
Level 1 (No 
Formal 
Approach) 
Level 2 
(Process 
Defi ned) 
Level 3 
(Proactively 
Managed) 
Level 4 
(Continuous 
Improvement) 
Level 5 (Best 
in Class) 
Compliance 
Infrastructure 
Performance 
Source : Adapted from Capability Maturity Model Integration, www.sei.cmu.org . 

and management. The maturity model is an excellent metric to measure development 
of the QMS and focus leadership in deployment of resources. 
The subcategories of the model demonstrate, through defi ned attributes that 
must be in place, specifi c areas required of a robust process. The infrastructure category 
includes a capable owner of the process is in place and a quality assurance 
counterpart is identifi ed, the process owner has a strong understanding of the 
process fl ow, scope, process boundaries, suppliers, customers, and roles and responsibilities. 
The goal is to develop a highly integrated process that is fully transferable 
and scalable. 
Compliance is a key process attribute for a process in the pharmaceutical and 
biopharmaceutical industry. Process maturity determination related to compliance 
can include documentation such as standards and SOPs, number of observations 
written against the process from internal audits, supplier audits, regulatory inspections, 
nonconformances, and a risk assessment on the process against patient safety 
and effi cacy. Training programs are also required as part of the process compliance. 
Audit and inspection observations written against a process are key metrics indicating 
maturity. Processes that can meet high maturity level for compliance represents 
a well - managed process that is consistently delivering a compliant and quality 
output. 
Performance metrics determine process performance, preferably against 
predetermined standards or expectations. Performance metrics should be indicators 
as to the health and robustness of the system. Performance metrics may include 
cycle turn around time, time to disposition from end of manufacture, and a risk 
assessment against business drivers. The purpose is to raise target performance 
objectives, developing a strategic approach, reducing variability, and improving 
effi ciencies. Effi ciencies gained in process performance contribute to the business 
needs. 
Advantages to utilizing a maturity model are that it provides a useful methodology 
for the QMS program group and leadership to evaluate, grade and provide 
process owners a goal for process development. Again, using a risk - based mindset, 
the entire inventory of process can be evaluated by leadership to determine where 
to place resources and to what maturity level each process is best positioned to 
support the enterprise. 
Maturity - level goals are best made by process owners, the QMS program group, 
and leadership. It is recommended that all processes are assessed against risk to the 
customer and the business. This allows the QMS to prioritize the processes and 
identify which of the processes need to be elevated to a higher maturity level in the 
maturity model. Upon completion of the risk assessment, results should be reviewed 
by leadership to determine if the processes have been prioritized appropriately and 
meet the corporation ’ s goals. This feedback forum will ensure that leadership supports 
process owners as they endeavor to achieve a higher level of process maturity. 
A well - designed QMS will allow for two - way conversation between the leadership 
and the process owners. It is as important for the leadership to communicate priorities 
to the process owners as well as having the process owners communicate issues 
and concerns that need to be addressed to the leadership. This will improve the 
alignment of priorities between the leadership and process owners. This integration 
ensures that the corporation is working on the right things at the right time with 
the right people. 
MEASURING SUCCESS THROUGH MEANINGFUL METRICS 279

280 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
3.3.7.4 Meeting Process Maturity Requirements 
A dedicated team or review board should be developed to review and approve all 
maturity - level deliverables upon completion of the attributes for the current level. 
This review board ’ s purpose is to ensure that all deliverables meet a consistent level 
of quality and documentation. This board can provide feedback to process owners 
or QMS program group to communicate best practices and lessons learned. 
A well - designed metric review program is essential to the success of the QMS. 
The program should include metrics for the QMS, process maturity - level assessment 
and process performance, infrastructure, and compliance metrics. These metrics are 
the basis for evaluating system progress against long - term vision and annual quality 
plans. The metrics provide leadership and process owners specifi c and objective data 
to determine program goal achievement. Leadership will have visibility and comparability 
of process performance within and between sites and have risk - based data 
to support their deployment of resources in addressing business issues. 
3.3.8 DRIVING CONTINUOUS IMPROVEMENT: PROJECTS 
Pharmaceutical and biopharmaceutical companies are under signifi cant pressure to 
deliver consistent quality product as well as drive the overall product cost down. 
The goal of implementing ICH Q8, Q9, and ultimately Q10 is to characterize processes 
based on risk assessments and improve them through a well - designed QMS. 
There are regulatory and business drivers to continually improve the QMS processes 
by building in quality and improve process effi ciency. The regulatory agencies 
are now focused on ensuring systems are in place that protect the public health by 
assuring both the safety and effi cacy of products. Understanding manufacturing 
processes, through well - designed characterization studies, is one of the most effi cient 
and effective methods to ensure process effi ciency. To meet business and consumer 
demands as well as regulatory guidance and expectations, the implementation of 
continuous improvement through risk - managed evaluations of manufacturing processes 
is expected. 
3.3.8.1 Process Improvements 
A quality management system ’ s process should follow a standard Six Sigma process 
improvement life cycle that includes the following steps: defi ne (process and metrics), 
measure and control (identify problems and issues), analyze (analyze problems and 
issues), and improve (implement) circling back to measure and control [11] . An 
example of a process improvement life cycle can be seen in Figure 9 . 
The basic foundation of continuous improvement begins with a process owner 
who fully understands the process and recognizes how the process impacts other 
processes within the QMS. Understanding this cause - and - effect relationship between 
processes requires close integration between process owners and stakeholders. This 
integration is critical throughout the entire life cycle of a process, from design 
through development and management. 
Prior to process improvements the process must be well - defi ned and predictable. 
This does not mean that the process or output is desirable but instead well under

stood and predictable. It is through metrics, trends, and risk assessments that issues 
and concerns should be evident. Process owners can use the management review 
forum to present a proposal for process improvements. 
3.3.8.2 Process Improvement Proposal 
The process owner with stakeholders will need to provide a process improvement 
proposal if the issue or change requires prioritization due to funds or additional 
resources from the enterprise. The proposal should include, at minimum, the problem 
and or opportunity statement, impact to the site based on risk, and proposal of an 
action and/or project, including both cost and resource requirements. 
During the development of the proposal, the process owner should consider 
requesting subject matter experts to assist with the development of the problem 
statement, risk assessment, and cost avoidance or savings. Many times a process 
owner ’ s core competencies align closely with the process but may lack business or 
project management skills. The process owner may need assistance to clearly articulate 
to the leadership what the benefi ts are to accept the proposed change versus 
the risks for not adopting the proposal. 
Risk assessment tools such as a nine - block risk assessment (Table 9 ) or a failure 
mode and effect analysis (FMEA) are available to assist the process owner with 
the evaluation of the process or issue to better understand and communicate the 
FIGURE 9 Continuous improvement process. 
Monitor / 
control Analyze 
problem 
Improve / 
implement 
Identify 
problems/ 
opportunities 
Define process 
and metrics 
TABLE 9 Nine - Block Risk Assessment Matrix 
Severity 
Minor Major Severe 
Frequency 
Probable Medium High High 
Occasional Low Medium High 
Remote Low Low Medium 
DRIVING CONTINUOUS IMPROVEMENT: PROJECTS 281

282 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
probability of failure and the severity of a process issue. These analyses assist with 
the prioritization of issues and identifi cation of specifi c actions required to mitigate 
risks or the identifi cation of contingency plans for issues that are not mitigated. In 
a pharmaceutical and biopharmaceutical environment it is important that all risk 
tools are completed assessing impact to product safety and effi cacy as well as the 
business drivers. Any steps or issues in the process that can negatively impact the 
safety and effi cacy of the product require immediate elevation to leadership and 
must be addressed immediately. If the proposal requires prioritization, the process 
owners should clearly identify potential cost savings or avoidance through a cost of 
quality model to further engage senior leadership. The combination of risk and costs 
is an effective way to gain leadership support and attention. 
Leadership ’ s role in the process improvement proposal is to understand the issue 
or opportunity, understand associated risk(s), and approve or redirect a proposed 
action or project and provide appropriate funding and or resources. As action items 
and proposals are approved and initiated, the progress should be monitored on a 
routine basis to ensure appropriate progress is made. 
3.3.8.3 Task versus Project 
Process improvements may be conducted by the completion of a task or a project. 
A task is an activity that can be completed by the process owner with minimal cost 
and/or resources over a short period of time. A project is defi ned as temporary work 
to provide a product or service that is beyond the process owner ’ s support. In 
general, a project requires more than one full - time equivalent (FTE), crosses over 
multiple functional organizations, and the duration of the effort spans over a longer 
period of time. Improvement status, updates, and issues should be discussed on a 
regular basis by a management forum or steering committee. Tasks and projects 
should be prioritized based on the risk against patient safety and effi cacy and 
compliance. 
If the process improvement meets the requirements of a project, a project 
manager should be identifi ed. Formal project management allows for a holistic and 
integrated approach to the change. The project manager should not replace the 
process owner but ensure that the issues are identifi ed, prioritized, and resources 
are applied, milestones are met, issues escalated and resolved, and progress reported. 
The process owner needs to be the project lead with the stakeholders or steering 
committee, providing support and guidance. This allows the process owner to focus 
on the issues and improvements (their core competencies) and allows the project 
manager to move the project forward in a methodical manner. During the project 
it is critical that success is defi ned and measured. 
3.3.8.4 Project Metrics 
Project metrics should be identifi ed to measure the actual benefi t of the change 
versus the expected result following the implementation. Many times, corporations 
implement a change and move on to the next project without fully understanding 
whether or not the changes achieved the desired result. A project that does not 
achieve the expected benefi ts can lead to an ineffective process, confl icts with associated 
touch points with other processes, or frustration from staff and customers. 

Applying a systems - based approach to continuous improvement of the QMS, 
utilizing formal risk management tools benefi ts the overall effi ciency of the organization. 
Process owners are accountable and empowered to drive continuous improvements. 
Metrics are utilized to identify trends, issues, and opportunities. Stakeholders 
are engaged throughout the process, and management is involved in the prioritization 
and staffi ng of the task or project. The processes are continually managed and 
evaluated. Continuous improvements based on risk allow the organization to apply 
resources and money to the most critical projects that will make the most impact. 
As process improvements are implemented, staff will benefi t from a predictable, 
lean process allowing them to focus on the proactive nature of their work as 
opposed to the high stress of reacting to the issue of the day. The process owner will 
gain credibility as he or she demonstrate the ability to ensure that the right people 
are making the right decisions in a timely manner and that process improvements 
are addressing systemic process problems and not superfi cially addressing issues 
that will resurface again. 
3.3.9 ENSURING ONGOING SUCCESS 
Building infrastructure to establish and maintain a quality management system 
requires resources and resolve from leadership and staff. The current pharmaceutical 
and biopharmaceutical global and regulatory environment requires an organization 
invested in developing and maintaining a robust system and processes meeting 
the organization ’ s requirements for producing quality product. Future competition, 
shorter time to market, effi cient development, and fi rst - pass approval expectations 
exacerbate the need for robust processes. The global marketplace continues its pressure 
on industry to deliver lifesaving and life - style changing medicines faster and 
cheaper. 
3.3.9.1 Establishing Mutual Goals 
Companies that have designed, developed, and established QMSs and processes 
that are simple to execute, easy to understand, and deliver the business and regulatory 
results will have competitive advantage over their industry peers. They will be 
faster and more effi cient at adapting new technologies, assimilating new organizations 
through merger and acquisitions, able to apply adequate resources to appropriate 
business needs, and most importantly quickly modify and adapt to changing 
marketplace demands. Dependence on people, fragmented procedures, or tribal 
knowledge, rather than integrated, functional processes, will bring undesirable 
results to all levels of the organization. 
Ensuring ongoing success requires establishing mutual goals for the organization 
from the beginning. These goals must satisfy the needs of the business, the employees, 
and the shareholders. Well - designed processes with accountable ownership that 
have been established through discussion, design, and support of leadership, 
functional management, operational stakeholders, and general staff provide the 
foundation for common shared needs (Figure 10 ). If anyone of these groups is not 
considered, nominal support and eventually failure of the program can be 
expected. 
ENSURING ONGOING SUCCESS 283

284 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
These shared goals need to be memorialized through documentation of the 
program. This includes outlining the long - term objectives of the program, benefi ts 
required to be achieved for stakeholders, and annual quality plans for achieving 
milestone goals and success. Program success comes from leadership support, robust 
system design, adequate training for employees, and meaningful metrics to measure 
performance and continuous improvement efforts. 
Mechanisms to determine stakeholder feedback on program acceptance, clarity, 
and improvement opportunities need incorporation into the ongoing maintenance 
of the QMS. Focus groups are one method of obtaining this type of information. 
Another opportunity exists with regulatory inspections and customer audits. Taking 
appropriate action to implement program changes and enhancements while recognizing 
contributors will ensure stakeholder support and participation. 
Mutual goals will drive success of the program and provide the reference for why 
the system approach is needed and the benefi ts it can bring. There is no better situation 
than having an entire organization aligned around the business design, and 
executing against it, while supporting each other. 
3.3.9.2 Rewards and Recognition 
Process owners ’ responsibilities are signifi cant. Owners need to be selected from 
predetermined criteria that are discriminatory in nature. Process owners are the 
drivers of the operations and therefore need to be recognized for their special 
efforts and responsibilities. This recognition can take many forms. A signifi cant distinction 
in base qualifi cations and rewards is a valuable incentive to becoming a 
process owner. Ongoing development for process owners is another incentive and 
reward for the process owner. In addition to the fi nancial and tangible rewards, 
being recognized by the organization to have the confi dence of management is also 
another form of reward and recognition. Inevitably, processes contain waste and 
FIGURE 10 Process owner support model. 

ineffi ciencies, thereby providing another opportunity for owners to improve their 
process and be recognized for that improvement. 
Public recognition of system program and process owner accomplishments is 
essential. This can easily be accomplished through regular review sessions, at metric 
review meetings, through staff meetings and updates, poster sessions, newsletters, 
and departmental meetings. Simple recognition and small gifts are appreciated and 
reinforce management ’ s support and commitment to the program. Process owners 
and stakeholders are the most infl uential group to spread the word on the usefulness 
of the program and must be cultured to ensure ongoing success. 
Studies indicate fi nancial rewards alone cannot provide employee satisfaction 
and retention. High employee turnover costs companies tremendous fi nancial and 
competitive resources. Many employees faced with equal or higher pay but unsatisfying 
work will move onto another company or position. A poorly integrated QMS 
with complicated processes is often the foundation for that dissatisfaction. To repeat 
work, lose valuable time, or deliver substandard product does not satisfy today ’ s 
highly educated and competitive worker in the pharmaceutical and biopharmaceutical 
industry. The cost to recruit, replace, relocate, and retrain employees is signifi - 
cant. Avoidance of these costs can be used as a partial basis for support of the 
program. 
3.3.9.3 Ensuring Ongoing Program Continuity 
Accomplishments of a comprehensive QMS program should be shared between 
locations and be consistent. Common, competent leadership for the enterprise will 
ensure consistency. A consistent QMS program also allows for transfer of staff 
between sites with little or no training and assimilation requirements. Divergent 
evolution will dilute the QMS effort and support. Flexibility to execute is important, 
however, caution must be exercised to restrict diverging language, interpretation, 
and philosophy. Within a short time of a global execution, effi ciencies will be quickly 
realized. Ensuring consistency also increases the number of process users with 
similar experiences and leverages focus for process improvements and therefore 
support. 
Regulators and customers require assurance in consistency of pharmaceutical 
and biopharmaceutical manufacturing operations. Today ’ s manufacturing supply 
chains require multiple sites in varying locations to produce a product. Quality 
systems must be perceived as an integral part of the value chain. This requires that 
all sites be compliant in their operations and systems. Strong areas in one location 
do not make up for weak or absent systems in another location. Fines are levied 
and business is made or lost based on the individual site or weakest link in the 
supply chain. Management must have a mechanism to measure its processes, and a 
comprehensive QMS is the mechanism to demonstrate capability. 
3.3.9.4 Program Institutionalization 
Program institutionalization is realized with time. All levels of the organization 
need to recognize and verbalize that the quality management system approach is 
the way business is conducted. This way of doing business will become part of 
the culture to the point at which it is second nature to leadership, management, 
ENSURING ONGOING SUCCESS 285

286 CREATING AND MANAGING A QUALITY MANAGEMENT SYSTEM 
and staff. Regulators and customers will recognize the benefi ts, as do the shareholders 
and patients. 
REFERENCES 
1. Webster ’ s New Collegiate Dictionary , ninth edition, 1986 , p. 1199 . 
2. Arling , E. R. ( 2004 ), Integrating QSIT into quality plans , Biopharm. Int. , June, 44 – 46, 48 , 
50 – 52 . 
3. Drug Industry Daily , Oct. 12, 2006, Vol. 5, No. 200, Washington Business Information. 
4. 2006 PDA/FDA joint regulatory conference , Sept. 11, 2006, Washington, DC. 
5. Joneckis , C. ( 2006 ), Ph.D. presentation at 11th Annual GMP by the Sea, Aug. 28 – 30, 
Cambridge, MD. 
6. Quality systems regulations CDRH , available: www.fda.gov . 
7. CPGM 7356.002, CDER, available: www.fda.gov . 
8. IOM biological inspections 7345.848 CBER, available: www.fda.gov . 
9. Rockefeller , J. D. , III ( 1973 ), The Second Revolution , Harper - Row , New York . 
10. Kotter , J. P. ( 2001 ), Leadership insights , Harvard Bus. Re. , p. 29 . 
11. George , M. (2004), Lean Six Sigma Pocket Tool Book , McGraw - Hill Professional , New 
York , p. 4 . 

287 
3.4 
QUALITY PROCESS IMPROVEMENT 
Jyh-hone Wang 
University of Rhode Island, Kingston, Rhode Island 
Contents 
3.4.1 Diagnosing a Process 
3.4.1.1 Introduction 
3.4.1.2 Basic Tools for Diagnosing a Process 
3.4.2 Stabilizing and Improving a Process 
3.4.2.1 Introduction 
3.4.2.2 Control Charts for Attributes 
3.4.2.3 Control Charts for Variables 
3.4.2.4 Special Control Charts 
3.4.3 Improving Performance of a Process 
3.4.3.1 Introduction 
3.4.3.2 Process Capability and Improvement Studies 
Bibliography 
3.4.1 DIAGNOSING A PROCESS 
3.4.1.1 Introduction 
Quality process improvement starts with a diagnostic journey where problems are 
identifi ed. Remedial activity will be taken and the process will be continuously 
monitored afterward. The common activities taken in the diagnostic journey are 
analyzing symptoms, formulating hypotheses, testing hypotheses, and identifying 
causes. Table 1 describes basic tools for the diagnostic journey. A description of them 
is given in Section 3.4.1.2 . 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

288 QUALITY PROCESS IMPROVEMENT 
3.4.1.2 Basic Tools for Diagnosing a Process 
Cause - and - Effect Diagram A cause - and - effect diagram relates potential causes 
of a problem to their effects. This is a tool that could be very useful in diagnosing 
a process. It focuses on the possible causes of a specifi c problem in a structured and 
systematic way. The following steps are suggested for constructing a cause - and - 
effect diagram: 
1. Defi ne the problem (effect). 
2. Write problem on the right side and draw an arrow from the left to the right 
side. 
3. Brainstorm the main categories of causes of problems and draw major branch 
arrows to the main arrow. 
4. For each major branch, detailed causal factors (subcauses) are drawn as 
subbranches. 
5. Write sub - subcauses branching off the subcauses. 
6. Ensure all the items that may be causing the problem are indicated in the 
diagram. 
Figure 1 shows a cause - and - effect diagram which is used to identify causes to yield 
a problem in a biopharmaceutical manufacturing process. Possible main causes and 
subcauses are identifi ed. Once the causes are identifi ed, other tools are employed 
to determine the contribution of various causes to the effect. Actions are taken to 
eliminate or minimize the impact of these causes. 
Pareto Chart The Pareto principle suggests a problem (effect) can be attributed 
to relatively few causes. In quantitative terms, 80% of the problems come from 20% 
of the causes (machines, raw materials, operators, etc.); therefore effort aimed at the 
right 20% can solve 80% of the problems. A Pareto chart includes three basic elements: 
(1) the causes to the total effect, ranked by the magnitude of the contribution; 
(2) the frequency of each cause; and (3) the cumulative - percent - of - total effect of 
TABLE 1 Basic Quality Process Improvement Tools during Process Diagnosis 
Common Activities to 
Diagnose Cause 
Basic Tools for Quality Process Improvement 
Cause – Effect 
Diagram 
Pareto Chart 
Histogram 
Scatter Diagram 
Normal 
Probability Plot 
Flow Diagrams 
Data Collection 
Box Plot 
Stratifi cation 
Analyzing symptoms • • • • • • • 
Formulating hypotheses • • • • 
Testing hypotheses • • • • • • 
Identifying cause(s) • • • • • 
Note : ( • ) major; ( ) minor. 
0 0 
0 0 0 0 0 
0 0 0 
0 0 0 0 
0

DIAGNOSING A PROCESS 289 
the ranked causes. Figure 2 gives an example of a Pareto chart which exhibits errors 
found in a pharmacy store chain in one month. 
Histogram A histogram is a graphic summary of variation in a set of data. Data 
are clustered into categories and the values of individual clusters are plotted to give 
a series of bars. For illustration, Table 2 presents 40 observations on the shelf life of 
a certain drug and their frequency distribution. Figure 3 gives a histogram for the 
drug shelf life data. 
Scatter Diagram A scatter diagram is a basic tool to identify the potential relationship 
between two variables. Scatter diagrams are similar to line graphs in that 
they use horizontal and vertical axes to plot data points. However, they have a very 
specifi c purpose. Scatter diagrams show how much one variable is affected by 
another. The relationship between two variables is called their correlation. The 
FIGURE 1 Cause - and - effect diagram. 
Agitation pH Coolant 
flow 
Concentration 
Substrate 
Flow rate 
Feed 
Temperature 
Aeration
Oxygen Water 
temperature 
Yield 
FIGURE 2 A Pareto chart showing pharmacy errors. 
Count 
Percent 
Pharmacy error
Count 
29.8 12.4 6.3 2.1 1.2 
Cum % 48.1 78.0 90.4 96.7 
2452 
98.8 100.0 
1520 632 320 108 62 
Percent 48.1 
Miss 
ed 
drug 
allerg 
ies 
W 
rong 
patient 
Mixing 
up 
prescriptions 
Incorrect 
label 
Incorrect 
dosing 
Misread 
pres 
cription 
5000 
4000 
3000 
2000 
1000
0 
100 
80 
60 
40 
20 
0

290 QUALITY PROCESS IMPROVEMENT 
closer the data points come when plotted to making a straight line, the higher the 
correlation between the two variables. If the data points make a straight line going 
from the origin out to high x and y values, then the variables are said to have a 
positive correlation. If the line goes from a high value on the y axis to a high value 
on the x axis, the variables have a negative correlation. Figure 4 gives a few examples 
of scatter diagrams. 
Normal Probability Plot The normal probability plot is a graphical technique for 
assessing whether or not a data set is approximately normally distributed. The data 
are plotted against a theoretical normal distribution in such a way that the points 
form an approximate straight line. Departures from this straight line indicate departures 
from normality. The normal probability plot is important for quality process 
improvement since many other tools require the normality assumption. A normal 
TABLE 2 Drug Shelf Life (days) 
102.2 104.1 103.5 104.5 103.2 103.7 103.0 102.6 
103.4 101.6 103.1 103.3 103.8 103.1 104.7 103.7 
102.5 104.3 103.4 103.6 102.9 103.3 103.9 103.1 
103.3 103.1 103.7 104.4 103.2 104.1 101.9 103.4 
104.7 103.8 103.2 102.6 103.9 103.0 104.2 103.5 
Range Midpoint Frequency Cumutative % 
101.5 . x < 102.0 101.75 2 5.00 
102.0 . x < 102.5 102.25 2 10.00 
102.5 . x < 103.0 102.75 5 22.50 
103.0 . x < 103.5 103.25 15 60.00 
103.5 . x < 104.0 103.75 8 80.00 
104.0 . x < 104.5 104.25 6 95.00 
104.5 . x < 105.0 104.75 2 100.00 
FIGURE 3 Histogram of drug shelf life. 
0
2
4
6
8 
10 
12 
14 
16 
101.75 102.25 102.75 103.25 103.75 104.25 104.75 
Shelf life 
Frequenc 
. 00% 
20.00% 
40.00% 
60.00% 
80.00% 
100.00% 
120.00% 
Frequency 
Cumulative %

DIAGNOSING A PROCESS 291 
probability plot of the drug shelf life in Table 2 is given in Figure 5 . As seen from 
the plot, the observations follow a straight line and are contained in the 95% confi - 
dence interval. It can thus be said that the shelf life of this drug follows a normal 
distribution. 
Other Tools 
Box Plot This plot is useful when analyzing the pattern of the data. It displays 
several important features of data such as central tendency, variability, departure 
from symmetry, and presence of outliers. 
Flow Diagrams A process fl ow diagram can be used to study and understand 
the process. 
Data Collection Data are essential for making a proper evaluation of the current 
process. Tools for data collection include checklists and data sheets. 
FIGURE 4 Scatter diagrams. 
X 
Y 
No correlation 
X 
Y 
Negative correlation 
X 
Positive correlation 
Y

292 QUALITY PROCESS IMPROVEMENT 
Stratifi cation This technique is used to separate data into groups based on categories 
or characteristics. It is the basis for the application of other tools or it 
can be used with other data analysis tools such as scatter diagrams. 
3.4.2 STABILIZING AND IMPROVING A PROCESS 
3.4.2.1 Introduction 
Basic Concepts of a Control Chart The control chart is one of the main tools for 
quality process improvement. It is used to assess the nature of variation in a process 
and to facilitate the forecasting and management of a process. Values of the quality 
characteristic are plotted against the sample number or time, as shown in Figure 6 . 
The centerline represents the process average. The upper and lower control limits 
(UCL and LCL) are usually chosen as three standard deviations (SDs) above and 
below the centerline so they can be used to detect “ out - of - control ” situations without 
causing mamy false alarms. An out - of - control situation is usually signaled by a 
plotted point falling outside the control limits or a cluster of plotted points forming 
an abnormal pattern. 
FIGURE 5 Normal probability plot of drug shelf life with 95% confi dence interval. 
Shelf life 
Percent 
106.0 105.5 105.0 104.5 104.0 103.5 103.0 102.5 102.0 101.5 101.0 
99 
95 
90 
80 
70 
60 
50 
40 
30 
20 
10
5
1 
95% confidence interval 
FIGURE 6 Normal curve - based control chart. 
Upper control limit (UCL) 
Center line 
Lower control limit (LCL) 
Sample number 
X 
68.26% 
95.46% 
99.73% 
–3. –2. –1. X +1. +2. +3.

Plotted points on a control chart are usually based on data collected from samples 
in a process. After a suffi cient number of samples are collected and the data are 
plotted on a control chart, the stability of the process can be evaluated. A stable 
process is “ in control ” while an out - of - control process is unstable. Depending on 
the type of quality characteristic, control charts can be divided into two groups: 
variable control charts and attribute control charts. Variable control charts are used 
to monitor quality characteristic that are continuously varying in nature; attribute 
control charts are used to monitor those quality characteristics that are not numerically 
measurable. The determination of the centerline and control limits are described 
in Sections 3.4.2.2 and 3.4.2.3 with respect to different types of control charts. 
Applications of Control Charts Control charts serve to direct management attention 
toward special causes of variation in a process when they appear. In evaluating 
control charts, the following symptoms could indicate a process that is out - 
of - control: 
• Outlier One or more point(s) that fell outside the control limits. 
• Run A series of plotted points above or below the centerline. 
• Trend A continual rise or fall of plotted points. 
• Cyclicity A pattern that repeats itself over time. 
The following steps are usually followed in a control chart ’ s development and 
application: 
• Determine a “ base period ” for initial control chart development. 
• Collect sample data from the base period. 
• Calculate the parameters for the control chart, that is, centerline and control 
limits. 
• Plot collected sample points on the chart with the centerline and control 
limits. 
• Determine whether the chart parameters can be used to monitor the process; 
revise the parameters if necessary. 
• Collect ongoing samples and continue monitoring the process using the 
developed control chart. 
• Conduct periodic audits on the parameters of the control chart. 
Variable control charts are widely applied in many manufacturing and nonmanufacturing 
settings. They can be used to monitor, for example, the inside diameter of 
an aircraft bearing, the moisture content of a drug tablet, the net weight of a pharmaceutical 
product, the processing time of phone inquiries, and the satisfaction level 
of customers. The latter two are examples of nonmanufacturing applications. 
Attribute control charts are less used compared to variable control charts. When 
it is not possible or practical to measure the quality characteristic of a product, 
attribute control charts are often applied. Examples of their application include 
monitoring the fraction of nonconforming of a certain sensor production, the number 
of defective diodes in an electronic assembly, the number of imperfections in textile 
STABILIZING AND IMPROVING A PROCESS 293

294 QUALITY PROCESS IMPROVEMENT 
production, the fraction of defective batches in a biomedical manufacturing production, 
and the number of errors found in a pharmacy store. 
In most applications, the choice between a variable control chart and an attribute 
control chart is clear - cut. In some cases, the choice will not be obvious. For instance, 
if the quality characteristic is the softness of an item, such as the case of pillow 
production, then either an actual measurement or a classifi cation of softness can be 
used. Quality managers and engineers will have to consider several factors in the 
choice of a control chart, including cost, effort, sensitivity, and sample size. Variable 
control charts usually provide more information to analysts but cost more to implement 
and use. Attribute control charts are less sensitive and expensive but usually 
requires large samples to reach certain statistical signifi cance. 
3.4.2.2 Control Charts for Attributes 
Control charts based on attribute data include the p chart, np chart, c chart, and u 
chart. The former two are applied when fraction nonconforming or number of nonconforming 
is a concern, and the latter two are used to deal with the nonconformities. 
Most pharmaceutical manufacturing industries employ one or more of these 
charts. 
p Control Chart A p chart can be used to monitor the fraction nonconforming of 
a process. Fraction nonconforming is defi ned as the ratio of the number of nonconforming 
items in a population to the total number of items in that population. In 
pharmaceutical manufacturing, an item will be classifi ed as nonconforming if it fails 
to conform to standards on one or more attributes, for example, fi ll volumes of vials, 
moisture content, hardness, and solubility. 
Let us suppose a random sample of n items is selected and examined from a 
process running with a stable nonconforming rate p and D units of nonconforming 
items are found; then D is a random variable following a binomial distribution 
with parameters n and p . If the true fraction nonconforming, p , is known, then the 
parameters of the p chart are 
UCL 
Centerline 
LCL
= + 
. 
=
= . 
. 
p 
p p 
n 
p
p 
p p 
n 
3 
1 
3 
1
( ) 
( ) 
(1) 
In practice, the fraction nonconforming, p , is unknown most of the time and is thus 
needed to be estimated from the sample data. An estimated p i can be calculated for 
the i th sample collected and an average p. value can be obtained as an arithmetic 
average of those individual p i found from the m samples: 
p 
D 
mn 
p 
m 
i
m 
i i
m 
i = = = = . . 1 1 
(2) 

The p. can then be used in place of p in Equation (1) in the application. It should be 
noted that the p. value needs to be assessed periodically to assure its representativeness 
of the average process fraction nonconforming. 
np Control Chart An np chart is used to monitor the number of nonconforming 
items produced in a process. Very similar to the p chart, the parameters of an np 
chart are 
UCL 
Centerline 
LCL
= + . 
=
= . . 
np np p 
np 
np np p 
3 1 
3 1
( ) 
( ) 
(3) 
As in the p chart, if the actual p value is not available, p. can be used in the 
calculation. 
c Control Chart A c chart can be used to monitor the number of nonconformities 
(defects) per inspection unit. An inspection unit can be a single unit of product, a 
batch of multiple products, or a certain measured volume (weight) of product. Many 
pharmaceutical manufacturing processes are lot based where raw material or semiproduct 
passes from one process to the next. For example, an inappropriately coated 
tablet in a coating process can be considered as a nonconformity (defect) where an 
inspection unit might be defi ned as 1 kg of the tablet. 
Suppose an inspection unit of a certain product is selected and examined from a 
process running with a stable nonconformity rate c per inspection unit and X nonconformities 
are found. Then X is a random variable following a Poisson distribution 
with parameter c . If the true nonconformity level c is known, then the parameters 
of the c chart are 
UCL 
Centerline 
LCL
= + 
=
= .
c c 
c
c c 
3
3 
(4) 
If the actual nonconformity level c is unknown, it can be estimated by using average 
c values obtained from m inspection units collected in a base period: 
c 
C 
m
i
m 
i = = . 1 
(5) 
The c. can then be used in place of c in Equation (4) in the application. Since it is 
possible to obtain a negative LCL using Equation (4) , a value of zero should be 
used in that case. 
u Control Chart A u chart is used to monitor the rate of nonconformities. The 
rate of nonconformities ( u ) is the number of nonconformities ( x ) in an inspection 
unit divided by the number of physical units ( n ) inspected (e.g., 100 ft of pipe, 100 
items in a batch). Similar to the c chart, the parameters of a u chart are 
STABILIZING AND IMPROVING A PROCESS 295

296 QUALITY PROCESS IMPROVEMENT 
UCL 
Centerline 
LCL
= + 
=
= . 
u 
u
n 
u
u 
u
n 
3
3 
(6) 
If the actual u value is not available, can be used in Equation (6) . 
Example 1 A medical device manufacturer is concerned about the nonconforming 
(defective) and the nonconformity (defect) produced in its recently set - up production 
line. Twenty batches of this medical device were randomly selected from the production 
line. Each batch contained 100 units. Each unit is inspected and is classifi ed as 
either “conforming” or “nonconforming. ” During the inspection, the number of nonconformities 
(defects) was also counted. The data collected are shown in Table 3 . 
3.4.2.3 Control Charts for Variables 
Control charts based on variable sample data include the x. chart and the s chart. 
When dealing with a numerically measurable quality characteristic, the x. chart is 
usually employed to monitor the process average and the s chart is used to monitor 
the process variability. When there is only one observation in each sample, the individual 
measurement chart ( I chart) and moving range chart (MR chart) are used to 
monitor the process average and variability. It should be noted that due to the poor 
TABLE 3 Nonconforming and Nonconformity Counts of 20 Batches of Medical Device 
Batch number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 
Nonconformings 3 2 4 2 5 2 1 2 0 5 2 4 1 3 6 0 1 2 3 2 
Nonconformities 9 7 13 8 6 8 10 10 7 10 12 9 11 15 8 12 11 8 7 15 
FIGURE 7 p Chart for medical device manufacturing example. 
Sample no. 
Fraction nonconforming 
19 17 15 13 11 9 7 5 3 1 
0.08 
0.06 
0.04 
0.02 
0.00 
_
P=0.025 
UCL=0.07184 
LCL=0 
Based on the data in Table 3 , the average fraction nonconforming, p. , can be 
obtained as 2.5%; the average nonconformity per batch, c. , is 9.8; and the average 
nonconformity per unit, , is 0.098. The resulting control charts are shown in Figures 
7 – 10 . These charts indicate that the process is in control and thus the parameters 
established here can be used to monitor future productions. 
u 
u

FIGURE 8 np Chart for medical device manufacturing example. 
Sample no. 
No. of nonconformities 
19 17 15 13 11 9 7 5 3 1 
8
6
4
2
0 
__ 
np =2.5 
UCL=7.184 
LCL=0 
FIGURE 9 c Chart for medical device manufacturing example. 
Sample no. 
No. of nonconformities 
19 17 15 13 11 9 7 5 3 1 
20 
15 
10
5
0 
_
C=9.8 
UCL=19.19 
LCL=0.41 
FIGURE 10 u Chart for medical device manufacturing example. 
Sample no. 
Nonconformity rate 
19 17 15 13 11 9 7 5 3 1 
0.20 
0.15 
0.10 
0.05 
0.00 
_
U=0.098 
UCL=0.1919 
LCL=0.0041 
sample size, the I and MR charts are less sensitive in detecting if the process is out 
of control than the x. and s charts. 
x. Control Chart An x. chart is used to monitor the process average. It is usually 
used in pharmaceutical manufacturing where multiple units are collected in each 
sample (e.g., a sample of multiple tablets formed by dry powders or wet granules.) 
Due to contamination risk and cost of sampling (including product loss due to 
STABILIZING AND IMPROVING A PROCESS 297

298 QUALITY PROCESS IMPROVEMENT 
sample volumes and incurred labor cost of laboratory analysis), the sample size is 
usually kept small. 
Sample means x. are plotted on the x. chart. Assume that random samples of n 
items are collected and examined from a stable process with a process mean . and 
standard deviation . . Then x. can be considered as a random variable following a 
normal distribution with mean . x. , and standard deviation . x. , where 
. . . 
. 
x x n 
= = and 
(7) 
If the true process mean . and standard deviation . are known, then the parameters 
of the x. chart are 
UCL 
Centerline 
LCL
= + = + 
= = 
= . = . 
. . . 
. 
. . 
. . . 
. 
x x 
x
x x 
n
n 
3 3 
3 3 
(8) 
Since . and . are not usually known, estimators of them can be obtained from the 
sample means ( x. ) and sample standard deviations ( s ) of the m samples collected in 
the base period: 
Estimator of 
Estimator of 
.
.
= = 
= 
. . 
= 
= 
= 
. 
. 
x 
x 
m
s m 
n n 
s 
c 
i
m 
i 
i
n 
i 
1 
1 
4 4 1 4 3 ( ) 
(9) 
Using the estimators, the parameters of the x. chart are now 
UCL 
Centerline 
LCL
= + = + 
=
= . = . 
x 
s c
n 
x A s 
x
x s c
n 
x A s 
3
3 
4 
3 
4 
3 
(10) 
The common values of constants c 4 and A 3 are tabulated in Table 4 for sample sizes 
from 2 to 10. Like other control charts, the values of x and s. should be periodically 
verifi ed to assure that they can be used to derive good estimators for the process 
average and process standard deviation. 
s Control Chart An s chart is used to monitor the process variability. Since it is 
equally important to ascertain that the variability and the mean of a process are in 
control, an s chart is usually used in conjunction with the x. chart. Sample standard 
deviations are plotted on the s chart. Consider s as a random variable with mean . s 
and standard deviation . s . Then the parameters of the s chart can be stated as

UCL 
Centerline 
LCL
= + 
=
= . 
. . 
.
. . 
s s 
s
s s 
3
3 
(11) 
In practice, the parameters of the s chart can be estimated using s. as 
UCL 
Centerline 
LCL
= + . = 
=
= . . = 
s 
s 
c 
c Bs 
s
s 
s 
c 
c Bs 
3 1 
3 1 
4 
4
2 
4 
4 
4
2 
3 
(12) 
If the LCL calculation results in a negative value, use zero as the LCL . 
TABLE 4 Values of Constants in Variable Control Chart 
Parameters 
n A 3 c 4 B 3 B 4 d 2 d 3 
2 2.659 0.798 0 3.267 1.128 0.853 
3 1.954 0.886 0 2.568 1.693 0.888 
4 1.628 0.921 0 2.266 2.059 0.880 
5 1.427 0.940 0 2.089 2.326 0.864 
6 1.287 0.952 0.030 1.970 2.534 0.848 
7 1.182 0.959 0.118 1.882 2.704 0.833 
8 1.099 0.965 0.185 1.815 2.847 0.820 
9 1.032 0.969 0.239 1.761 2.970 0.808 
10 0.975 0.973 0.284 1.716 3.078 0.797 
Example 2 In a Pet Tabs (pet vitamin tablets) production, the pharmaceutical 
manufacturer is using milling and micronizing machines to pulverize raw materials 
into fi ne particles. These fi nished particles are combined and processed further in 
mixing machines. The mixed ingredients are then pressed into tablets, dried, and 
sealed in packages. A normally distributed quality characteristic, moisture content, 
is monitored. Samples of n = 4 tablets are taken from the manufacturing process 
every hour. The data after 25 samples have been collected are shown in Table 5 . 
From these data, it is found that x = 10 254 . and s. = 0.926. Using Equations (10) 
and (12) , the parameters of the x. and s charts are found as: 
x. Chart s Chart 
UCL 11.761 2.098 
Centerline 10.254 0.926 
LCL 8.747 0 
The control charts are shown in Figure 11 . The x. and s charts show that the process 
is in control and thus the parameters established here can be used to monitor future 
productions. 
STABILIZING AND IMPROVING A PROCESS 299

300 QUALITY PROCESS IMPROVEMENT 
TABLE 5 Moisture Content (%) of 25 Samples of Pet Tabs 
Sample Number 
Observations 
x. s 1 2 3 4 
1 7.84 11.01 10.14 9.41 9.600 1.343 
2 10.51 9.1 9.52 10.83 9.990 0.814 
3 9.74 10.39 9.62 11.16 10.228 0.708 
4 10.71 11.41 10.71 8.63 10.365 1.203 
5 9.93 10.95 8.99 10.73 10.150 0.889 
6 9.94 10.27 9.35 9.42 9.745 0.438 
7 12.11 9.72 8.89 9.75 10.118 1.387 
8 9.61 8.93 11.12 8.75 9.603 1.077 
9 9.17 10.87 9.97 10.79 10.200 0.798 
10 11.41 10.39 8.83 12.19 10.705 1.451 
11 8.43 9.48 10.56 10.2 9.668 0.939 
12 9.92 10.13 9.66 8.21 9.480 0.868 
13 8.39 9.94 10.4 8.69 9.355 0.967 
14 10.42 10.27 10.94 10.91 10.635 0.341 
15 10.98 12.57 11.14 8.97 10.915 1.481 
16 9.73 10.05 12.82 12.43 11.258 1.592 
17 11.36 8.91 10.08 10.55 10.225 1.024 
18 9.42 11.12 9.01 10.52 10.018 0.973 
19 10.15 10.08 10.12 9.88 10.058 0.122 
20 11.73 11.1 10.75 9.94 10.880 0.746 
21 11.52 9.11 9.88 11 10.378 1.087 
22 11.29 10.43 11.6 11.74 11.265 0.588 
23 9.39 12.96 11.42 10.28 11.013 1.541 
24 10.26 9.59 9.33 9.26 9.610 0.456 
25 11.25 10.65 11.06 10.63 10.898 0.307 
FIGURE 11 x. and s charts for Pet Tab manufacturing example. 
Sample no. 
Sample mean 
25 23 21 19 17 15 13 11 9 7 5 3 1 
12 
11 
10
9 
__
X=10.254 
UCL=11.761 
LCL=8.747 
Sample no. 
Sample SD 25 23 21 19 17 15 13 11 9 7 5 3 1 
2.0 
1.5 
1.0 
0.5 
0.0 
_S
=0.926 
UCL=2.098 
LCL=0

Individuals Control Charts In some chemical and biopharmaceutical manufacturing 
processes involving lengthy and expensive procedures, it is not feasible to form 
a sample of size greater than one because only one product or one batch is available 
each time. When the sample size used for statistical process monitoring is limited 
to one, individual control charts, I and MR charts, are needed. 
The I chart is serving the same function as the x. chart except that now x is the value 
of the individual measurement. Assuming that x follows a normal distribution with 
mean . and standard deviation . , the theoretical parameters of the I chart are 
UCL 
Centerline 
LCL
= + 
=
= . 
. . 
.
. .
3
3 
(13) 
The process average . can be estimated by x. , which is 
.. = . = . x 
x 
m
i
m 
i 1 
(14) 
Since only individual measurements are available, moving ranges need to be calculated 
for the estimation of process standard deviation . . A k - point moving range, 
MR k , can be calculated as 
MR max( , . . . , min , . . . , k i i k i i k x x x x = . + + ) ( ) (15) 
For m individual measurements, there are m . k MR k available, and the process 
standard deviation . can be estimated as 
.. = = 
. =
. . MR MR k i
m k 
ki 
d 
m k 
d 2 
1 
2 
(16) 
The estimated process mean and standard deviation can be used to calculate the 
practical parameters for the I chart in Equation (13) . The constant d 2 value is determined 
by k and can be found by using k as n in Table 3 . Common k values can range 
from 2 to 5. 
The MR chart is used to monitor process variability. Considering MR k as a 
random variable with a mean of .MRk and a standard deviation of .MRk, the theoretical 
parameters of the MR chart can then be stated as 
UCL 
Centerline 
LCL 
MR MR 
MR 
MR MR 
= + 
=
= . 
. . 
.
. . 
k k 
k
k k 
3
3 
(17) 
Since .MRk and .MRk are not usually available, they can be estimated as 
. . . . MR MR MR and MR k k k k 
d
d 
= =3
2 
(18) 
STABILIZING AND IMPROVING A PROCESS 301

302 QUALITY PROCESS IMPROVEMENT 
The constant value of d 3 can also be found in Table 3 . If a negative LCL was 
obtained, use zero. 
3.4.2.4 Special Control Charts 
The control charts discussed earlier are very useful in the diagnostic aspects of 
quality process improvement. They can be used to stabilize a process by identifying 
out - of - control situations. After the process is stabilized and brought in control, 
further improvement of the process can be achieved by using some special control 
charts such as the cumulative sum (CUSUM) control chart and the exponentially 
weighted moving average (EWMA) control chart. These control charts can be used 
when “ small shifts ” in a process are of interest. 
CUSUM Control Chart A CUSUM chart provides an effi cient way of detecting 
small shifts in the mean of a process ( < 1/2 . ). For larger shifts ( > 1/2 . ), the x. chart 
is usually used. The CUSUM chart incorporates information contained in a sequence 
of sample points. It keeps track of the cumulative sum of the deviations between 
each sample point (a sample mean) and a target value. Unlike the x. chart, 
which often bases its out - of - control decision on just the most recently collected 
sample, the CUSUM calculated for a sample point carries the “ history ” prior to 
that sample. For example, a sequence of sample points above the centerline can 
trigger an out - of - control signal although all of them stayed well below the UCLs of 
the x. chart. 
There are two forms of the CUSUM chart, the tabular form and the V - mask form. 
Due to its practicality, the tabular form is more preferred in industrial settings. The 
tabular CUSUM accumulates deviations from a target value (or a known process 
mean . 0 ). Deviations above that target value are cumulated as a one - sided upper 
CUSUM ( C + ) and deviations below the target value are cumulated as one - sided 
lower CUSUM ( C . ): 
C x k C 
C k x C 
i i x i 
i x i i 
+ 
. 
+ 
. 
. 
. 
= . + + 
= . . + 
max[0, 
max , 
( ) ] 
[ ( ) ] 
. . 
. .
0 1 
0 1 0 
(19) 
where C C 0 0 0 + . = = . 
The parameter k is called the allowance and is usually determined as the magnitude 
of the shift to be detected in terms of . x. . If either Ci
+ or Ci
. exceeds a decision 
interval h , the process is considered out of control. In other words, the value of h is 
considered a UCL and . h is considered an LCL. Its centerline is always at zero. A 
reasonable value for h is fi ve times the process standard deviation . . 
EWMA Control Chart An EWMA control chart plots weighted moving average 
values for variables data. A weighting factor is chosen by the user to determine how 
older data points affect the mean value compared to more recent ones. Because the 
EWMA chart uses information from all samples, it is a good alternative to the 
CUSUM chart in detecting smaller process shifts. 
The EWMA for sample i ( z i ) is plotted on the chart and is defi ned as z i = . x. i + 
(1 . . ) z i . 1 , where z 0 = . 0 . The constant . defi nes the weight assigned to the current 

sample (0 < . . 1) and 1 . . is the weight assigned to earlier samples. Parameters 
of the EWMA are 
UCL 
Centerline 
LCL
= + 
. 
. . 
=
= . 
. 
. . 
. . 
. 
. 
. 
.
. . 
. 
. 
0 
2 
0
0 
2 
1 1 
2 
1 1 
L
L 
x 
i 
x 
[ ( ) ] 
[ ( .) ] 2i 
(20) 
where L is a design parameter that defi nes the width of the control limits. The choice 
of L = 3 and 0.05 < . . 0.25 is reasonable. The control limits will become wider when 
the sample number i is getting larger and fi nally reach constant values as 
UCL 
Centerline 
LCL
= + 
. 
=
= . 
. 
. . 
. 
. 
.
. . 
. 
. 
0
0
0 
2
2 
L
L 
x
x 
(21) 
Example 3 The data in Example 2 are now analyzed by CUSUM and EWMA 
charts. Table 6 shows calculated CUSUM and EWMA values. The value of h in 
CUSUM is chosen as 5 times the standard deviation of x. ( . . .x = 0 5027) and the value 
of k is chosen as 0.5. The Ci
+ and Ci
. are calculated using a target value . 0 = 10. 
The CUSUM chart is shown in Figure 12 . The value of . in EWMA is chosen as 0.2 
and L is chosen as 3. The UCL and LCL for individual samples are shown in Table 
6 and the EWMA chart is shown in Figure 13 . Although the x. and s charts in Figure 
6 indicate that the process is in control, both CUSUM and EWMA gave out - of - 
control signals at sample point 22. A small process shift has occurred after sample 
21. 
IMPROVING PERFORMANCE OF A PROCESS 303 
3.4.3 IMPROVING PERFORMANCE OF A PROCESS 
3.4.3.1 Introduction 
Basic Concepts After a process is diagnosed, corrected, and brought into statistical 
control, the next question is “ How can the performance of a process be improved? ” 
To answer this question, quality managers and engineers need fi rst measure the 
present process performance. This measurement can be achieved through a process 
capability study which gauges the ability of a process to produce products according 
to the specifi cations. A process can achieve a state of statistical control but still exhibit 
a poor capability due to the variability in the process. It will be necessary to reduce 
variability to improve the process capability. Designed experiments based on statistical 
principles can offer helps toward reduction of variability and optimization of the 
process. Employing designed experiments, intentional changes can be made in various 
places in the process; results gathered from these experiments can lead to further 
process improvement and bring it to the next level. This section presents commonly 

304 QUALITY PROCESS IMPROVEMENT 
TABLE 6 CUSUM and EWMA Values for Pet Tabs Example 
Sample Number x. 
CUSUM EWMA 
Ci
+ Ci
. z i UCL LCL 
1 9.600 0.000 0.149 9.920 10.302 9.698 
2 9.990 0.000 0.000 9.934 10.386 9.614 
3 10.228 0.000 0.000 9.993 10.432 9.568 
4 10.365 0.114 0.000 10.067 10.458 9.542 
5 10.150 0.012 0.000 10.084 10.475 9.525 
6 9.745 0.000 0.004 10.016 10.485 9.515 
7 10.118 0.000 0.000 10.036 10.491 9.509 
8 9.603 0.000 0.146 9.950 10.495 9.505 
9 10.200 0.000 0.000 10.000 10.498 9.502 
10 10.705 0.454 0.000 10.141 10.500 9.500 
11 9.668 0.000 0.081 10.046 10.501 9.499 
12 9.480 0.000 0.350 9.933 10.501 9.499 
13 9.355 0.000 0.744 9.817 10.502 9.498 
14 10.635 0.384 0.000 9.981 10.502 9.498 
15 10.915 1.047 0.000 10.168 10.502 9.498 
16 11.258 2.054 0.000 10.386 10.502 9.498 
17 10.225 2.027 0.000 10.354 10.502 9.498 
18 10.018 1.794 0.000 10.286 10.502 9.498 
19 10.058 1.600 0.000 10.241 10.502 9.498 
20 10.880 2.229 0.000 10.368 10.502 9.498 
21 10.378 2.355 0.000 10.370 10.503 9.497 
22 11.265 3.369 0.000 10.549 10.503 9.497 
23 11.013 4.130 0.000 10.642 10.503 9.497 
24 9.610 3.489 0.139 10.435 10.503 9.497 
25 10.898 4.135 0.000 10.528 10.503 9.497 
h = 2.513 . = 0.2 
k = 0.5 L = 3 
FIGURE 12 CUSUM chart for Pet Tabs manufacturing example. 
Sample no. 
Cumulative sum 
25 23 21 19 17 15 13 11 9 7 5 3 1 
5
4
3
2
1
0 
-1 
-2 
0
UCL=2.513 
LCL=-2.513

FIGURE 13 EWMA chart for Pet Tabs manufacturing example. 
Sample no. 
EWMA 
25 23 21 19 17 15 13 11 9 7 5 3 1 
10.75 
10.50 
10.25 
10.00 
9.75 
9.50 
UCL=10.503 
LCL=9.497 
Target = 10 
used methods in process capability studies. Design - of - experiment techniques can be 
found elsewhere in this handbook and in many other textbooks. 
Specifi cation Limits, Control Limits, and Natural Tolerance Limits To conduct 
a process capability study, it is important to distinguish the specifi cation limits of a 
product, the control limits of the process producing the product, and the natural 
tolerance limits (NTLs) of the product. In general, specifi cation limits are given by 
customers or prescribed by in - house design engineers before production. A product 
that failed to meet the specifi cations is a nonconforming product. Control limits are 
usually determined by samples collected from a process during a base period. A 
sample point that fell outside the control limits will trigger an out - of - control state; 
however, a product produced in the out - of - control state is not necessarily a nonconforming 
product. It should also be noted that a sample point plotted in a control 
chart usually represents a statistic of the sample such as the sample mean. In other 
words, a single product that fell outside of the control limits will neither cause the 
process to be out of control nor become nonconforming. The variability of products 
produced can usually be described by its natural tolerance limits. It is commonly 
acceptable that the ± 3 standard deviations from the process mean be used as the 
natural tolerance limits. 
Example 4 Following Example 2 , the specifi cation limits are specifi ed as 10.00 ± 
2.00, where: 
Nominal or target value ( . 0 ) = 10.00 
Upper specifi cation limit (USL) = 10.00 + 2.00 = 12.00 
Lower specifi cation limit (LSL) = 10.00 . 2.00 = 8.00 
The control limits for the x. chart are: 
Center line ( x ) = 10.254 
Upper control limit (UCL) = 11.761 
Lower control limit (LCL) = 8.747 
IMPROVING PERFORMANCE OF A PROCESS 305

306 QUALITY PROCESS IMPROVEMENT 
3.4.3.2 Process Capability and Improvement Studies 
Process Capability Indices Process capability indices provide a quantitative measure 
to assess the ability of a process to produce products that meet the specifi cations. A 
commonly used process capability index, denoted as C p , can be calculated as 
Cp = 
. USL LSL 
6. 
(22) 
where USL is the upper specifi cation limit, LSL is the lower specifi cation limit, and 
. is the process standard deviation. Since . is not usually known, it can be estimated 
by .. = s c / 4 . A C p = 1 means that the process is just capable. If the process is centered 
at its nominal value, it will produce 2,700 nonconforming products out of one million 
(PPM). The target value C p is usually set at 1.33 for an existing process and 1.50 for 
a new process. 
It should be noted that the C p value could not indicate the proper process capability 
if the process is not centered since C p does not account for where the process 
mean is with respect to the specifi cations. To alleviate this issue, another process 
capability index, C pk , is used: 
C C C pk pu pl = min( , ) 
Using the x ± 3. 
natural tolerance limits, they can be obtained as: 
Process mean ( x ) = 10.254 
Upper natural tolerance limit (UNTL) = 13.270 
Lower natural tolerance limit (LNTL) = 7.238 
The relationships among the three sets of limits are illustrated in Figure 14 . As can 
be seen from this fi gure, the current process is not centered at its nominal value and 
its specifi cation limits are tighter than its natural tolerance limits. Due to this, a 
portion of manufactured products ( . 5.4%) will not be able to conform to the 
specifi cations. 
FIGURE 14 Specifi cation limits, control limits, and natural tolerance limits for Pet Tabs 
manufacturing example. 
LSL m0 USL 
LNTL LCL x UCL UNTL

where 
C C pu pl = 
. 
= 
. USL 
and 
LSL 
3 
. 
. 
. 
. 3 
(23) 
The . value can be estimated by x and the . value can be estimated as discussed 
earlier. In general, a process is considered “ centered ” at the nominal value of the 
specifi cations when C p = C pk and “ off centered ” when C p < C pk . The relationships 
between C p and C pk are further illustrated in Figure 15 where the process mean has 
shifted from . 0 to . 0 + 2 . to . 0 + 4 . . As noted from the fi gure, C p remains the same 
regardless of the shift but C pk is signifi cantly reduced. 
FIGURE 15 Relationships between C p and C pk . 
0 2 4 6 8 10 12 14 16 18 20 
0 2 4 6 8 10 12 14 16 18 20 22 
0 2 4 6 8 10 12 14 16 18 20 22 24 
LSL m0 
s = 2 
USL
Cp = 1.667 
Cpk = 1.667 
Cp = 1.667 
Cpk = 1.0 
Cp = 1.667 
Cpk = 0.333 
IMPROVING PERFORMANCE OF A PROCESS 307 
If one - sided specifi cations are used, one - sided process capability can also be 
defi ned by Equation (23) where C pu is for upper specifi cation and C pl for lower 
specifi cation. 
Interpretation and Improvement of Process Capability Evaluation and interpretation 
of process capability represent an important step in process quality 

308 QUALITY PROCESS IMPROVEMENT 
improvement. A process must have its source of instability eliminated before it can 
be improved. Results obtained from process capability studies can help determine 
whether the process is stable and meeting its specifi cations. It should be noted that 
a valid process capability study is based on the normality assumption of the process. 
The normality assumption will need to be checked before proceeding to the next 
step. 
Conclusions regarding whether the process is centered at the target and is meeting 
the specifi cations can be drawn from the process capability study. When C p = C pk , 
the process is centered. When C p has a value of 1.0 or greater, the process is capable 
of producing products meeting specifi cations; otherwise, it is not capable. 
Example 5 Following Example 2 , C p and C pk are calculated as 
Cp = 
. 
= 
. 
. 
= USL LSL 
6. 
12 8 
6 1 0054 
0 6633 
. 
. 
C C C pk pu pl = = = min( , min , ) (. . ) . 0 5790 0 7476 0 5790 
where 
Cpu = 
. 
= 
. 
. 
= USL . 
. 3 
12 10 254 
3 1 0054 
0 5790 
. 
. 
. 
Cpl = 
. 
= 
. 
. 
= 
. 
. 
LSL 
3 
10 254 8 
3 1 0054 
0 7476 
. 
. 
. 
Figure 16 shows the histogram of the data in relation to the specifi cations. The x. and 
s charts in Figure 11 show that the process is in statistical control. However, since 
C p < C pk , the process is not centered. With a C pk value of 0.579, it is expected to have 
53,711 nonconforming Pet Tabs manufactured out of one million parts in this production 
line. 
FIGURE 16 Process capability plot for Pet Tabs manufacturing example. 

BIBLIOGRAPHY 
Aft , L. S. ( 1997 ), Fundamentals of Industrial Quality Control , 3rd ed., CRC Press , Boca Raton, 
FL . 
DeVor , R. E. , Chang , T. H. , and Sutherland , J. W. ( 2006 ), Statistical Quality Design and Control , 
2nd ed., Prentice - Hall , Upper Saddle Brook, NJ . 
Gitlow , H. , Gitlow , S. , Oppenheim , A. , and Oppenheim , R. ( 1989 ), Tools and Methods for the 
Improvement of Quality , Irwin , Boston . 
Grant, E. L. , and Leavenworth, R. S. (1996), Statistical Quality Control , 7th ed., McGraw-Hill, 
New York . 
Ishikawa , K. ( 1982 ), Guide to Quality Control , 2nd ed., Asian Productivity Organization , 
Tokyo, Japan . 
Juran , J. M. , and Godfrey , A. B. ( 1998 ), Juran ’ s Quality Handbook , 5th ed., McGraw - Hill , New 
York . 
Ledolter , J. , and Burrill , C. W. ( 1998 ), Statistical Quality Control: Strategies and Tools for 
Continual Improvement , Wiley , New York . 
Montgomery , D. C. ( 2004 ), Design and Analysis of Experiments , 6th ed., Wiley , New York. 
Montgomery , D. C. ( 2001 ), Introduction to Statistical Quality Control , 4th ed., Wiley , New 
York . 
Ryan , T. P. ( 2000 ), Statistical Methods for Quality Improvement , 2nd ed., Wiley , New York. 
Smith , G. M. ( 2003 ), Statistical Process Control and Quality Improvement , 5th ed., Prentice - 
Hall , Upper Saddle Brook, NJ . 
Summers , D. C. S. ( 2006 ), Quality , 4th ed., Prentice - Hall , Upper Saddle Brook, NJ . 
Tague , N. R. ( 2005 ), Quality Toolbox , ASQ Quality Press , Milwaukee . 
Thompson , J. R. , and Koronacki , J. ( 2001 ), Statistical Process Control: The Deming Paradigm 
and Beyond , 2nd ed., Chapman & Hall , New York . 
To improve the process capability, the process needs to be centered fi rst. This 
usually involves adjusting the process settings. Cause - and - effect diagrams, Pareto 
charts, and other tools discussed earlier in the chapter can be employed to fi nd 
causes to the “ off - centering ” problems. After the process is brought back to its 
nominal (10), the total nonconforming Pet Tabs produced will be dropped to 
46,673 PPM. This is still far from the 2700 PPM for a “ just capable ” process ( C p = 1), 
not to mention reaching the goal of 63 PPM at C p = 1.33. 
To further improve the process capability, the variability needs to be reduced. 
This can be achieved by designed experiments. Design of experiment (DOE) is a 
systematic approach that allows engineers and managers to make intentional 
changes in some process settings and assess the effects of those changes. An experiment 
can be designed in this example by varying a few key process settings such as 
drying time, mixing time, and temperature. Through a series of experimentations, 
optimum settings are found for these process variables and the variability of the 
process is reduced by 50%. With this reduction in process variability, the process is 
now exhibiting a C p of 1.265 with 694 PPM. This example highlights the benefi ts of 
process improvement. The move from an off - centered state to a centered state 
resulted in a reduction of process fall - out by 13.1%. With designed experiments, the 
process variability was cut in half and the process fall - out was signifi cantly reduced 
by 98.5%. 
BIBLIOGRAPHY 309


PROCESS ANALYTICAL 
TECHNOLOGY ( PAT ) 
SECTION 4


CASE FOR PROCESS ANALYTICAL 
TECHNOLOGY: REGULATORY AND 
INDUSTRIAL PERSPECTIVES 
Robert P. Cogdill 
Duquesne University, Center for Pharmaceutical Technology, Pittsburgh, Pennsylvania 
Contents 
4.1.1 Introduction 
4.1.2 Basis for Process Analytical Technology 
4.1.2.1 Process Analytical Chemistry 
4.1.2.2 Quality Management 
4.1.2.3 Lean Manufacturing 
4.1.3 Historical Factors Limiting Implementation of PAT 
4.1.3.1 Real and Perceived Technological Barriers 
4.1.3.2 Lack of Economic Incentive 
4.1.3.3 Regulatory Disincentives 
4.1.4 FDA Twenty - First - Century cGMPs Initiative 
4.1.4.1 Conception of the Initiative 
4.1.4.2 Risk - Based Orientation 
4.1.4.3 Quality Systems Approach 
4.1.4.4 Science - Based Policies 
4.1.4.5 International Collaboration 
4.1.5 PAT Evolution in Pharmaceutical Manufacturing 
4.1.5.1 Process Understanding 
4.1.5.2 PAT Principles and Tools 
4.1.5.3 Strategy for Implementation 
4.1.6 PAT Implementation Process 
4.1.6.1 Preparation 
4.1.6.2 Assessment 
4.1.6.3 Analyze 
4.1.6.4 Control 
4.1.6.5 Release Philosophy 
4.1.6.6 Optimization 
4.1 
313 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

314 REGULATORY AND INDUSTRIAL PERSPECTIVES 
4.1.7 Perspectives on the Impact of PAT 
References 
4.1.1 INTRODUCTION 
The implementation of process analytical technology (PAT) is occurring in what is 
perhaps the most exciting period of change in pharmaceutical manufacturing of the 
past three decades. A host of technological, regulatory, and market forces have 
converged during the last fi ve years, yielding new opportunities for innovation in 
the development and operation of pharmaceutical production processes. A major 
driving force for change is the Food and Drug Administration (FDA) initiative to 
implement a modern, risk - based framework for regulation and oversight of pharmaceutical 
manufacturing [1] . The objectives of this section are to outline the historical 
background of process analytics, to provide an overview of PAT in the pharmaceutical 
industry and the business drivers for change, to summarize the FDA ’ s new initiative 
and the PAT guidance [2] , and to present a basic plan for PAT implementation. 
While the focus of this chapter is PAT, it should be kept in mind that PAT is an 
important part of the much broader and risk - based paradigm introduced by the 
twenty - fi rst - century current good manufacturing practices (cGMPs) initiative. 
4.1.2 BASIS FOR PROCESS ANALYTICAL TECHNOLOGY 
Despite the fact that the FDA ’ s PAT framework (and guidance) began to take form 
just ahead of the creation of the twenty - fi rst - century cGMPs initiative in 2001, it is 
well known that several of the core concepts were pioneered decades ago by other 
manufacturing industries such as fi ne chemicals, semiconductors, petroleum, and 
consumer products. The main concepts that differentiate PAT from the traditional 
industrial pharmacy skill set (including pharmaceutical and materials science, chemistry, 
and engineering) are process analytical chemistry (PAC) and advanced manufacturing 
science (Figure 1 ). 
For the purpose of this discussion, the term manufacturing science is meant to 
describe the science and technology related to modern innovations in the design 
and management of manufacturing processes. Since it is neither practical nor necessary 
to cover all aspects of modern pharmaceutical manufacturing science in detail, 
the following sections are intended to introduce two specifi c topics which are popular 
in the current industrial vernacular but are not covered in detail in the pharmaceutical 
literature: quality management systems and “ lean ” manufacturing. 
4.1.2.1 Process Analytical Chemistry 
Process analytical chemistry generally describes the science and technology associated 
with displacement of laboratory - based measurements with sensors and instrumentation 
positioned closer to the site of operation. Although industrial process 
analyzers have been in use for more than 60 years [3] , the modern period of PAC 
essentially began with the formation of the Center for Process Analytical Chemistry 
(CPAC) in 1984 [4] . As described by Callis, Illman, and Kowalski [5] , the goal of 

BASIS FOR PROCESS ANALYTICAL TECHNOLOGY 315 
PAC is to “ supply quantitative and qualitative information about a chemical process ” 
for monitoring, control, and optimization: they went on to defi ne fi ve “ eras ” of PAC: 
(1) off line, (2) at line, (3) online (4) inline, and (5) noninvasive, which describe the 
evolution of sensor technologies. In addition, they discussed the importance of issues 
beyond chemical sensing, such as sampling, extraction of information from data 
(chemometrics), integration with process controls, as well as the sociological aspects 
of PAC deployment (e.g., gaining the trust of plant operators). 
The industrial PAC movement has been bolstered by two decades of advances 
in materials science, electronics, and chemometrics. Since the inception of CPAC, 
the pace of innovation in sensors, instrumentation, and analytics has quickened 
dramatically. The development of more robust, sensitive photodetector materials, 
microelectromechanical systems (MEMSs), and fi ber optics and the perpetual 
advancement of computing power (as predicted by Moore ’ s law) have both increased 
the performance and reduced the cost of PAC. As a result, PAC is now a critical 
part of routine operations within the realm of industrial chemistry. Many general 
reviews on the subject of PAC (and PAT) have been published [6 – 10] . A series of 
literature reviews on the subject of PAC have been published regularly in Analytical 
Chemistry . 
The fi rst review [11] listed manuscripts published between 1987 and 1992, covering 
seven specifi c topics (general PAC, chromatography, optical spectroscopy, fi ber 
optics, mass spectrometry, chemometrics, and fl ow injection analysis), along with a 
section on needs for the future of PAC; in all, the fi rst review included 507 references. 
Subsequent reviews were published in 1995 [12] , 1999 [13] , 2001 [14] , 2003 
[15] , and 2005 [16] . The review series is an essential resource for scientists seeking 
information on specifi c PAC methods; in total, 2650 references covering more than 
16 topics were catalogued by the authors. 
Currently there are three major consortia involving university, government, 
and industrial partners — CPAC, the Measurement & Control Engineering Center 
(MCEC), and the Control Theory and Applications Centre (CTAC) — along with an 
annual conference, the International Forum on Process Analytical Chemistry 
(IFPAC), and numerous online resources that are devoted to issues related to 
process analytics [16] . In parallel with the FDA ’ s initiative, the term process 
FIGURE 1 Multidisciplinary components of PAT in pharmaceutical manufacturing. 

316 REGULATORY AND INDUSTRIAL PERSPECTIVES 
analytical chemistry is gradually being replaced in the industrial vernacular by 
process analytical technology . This refl ects the expansion of the fi eld as the importance 
of physical characterization, risk analysis, and manufacturing science is 
recognized. 
4.1.2.2 Quality Management 
Many of the quality improvement goals for implementation of PAT in the pharmaceutical 
industry have been achieved by companies in other industries, such as 
automobile production and consumer electronics, as a direct result of adopting principles 
of quality management. The lineage of modern quality management can be 
traced to the work of Walter Shewhart, a statistician for Bell Laboratories in the 
mid - 1920s [17] . His observation that statistical analysis of the dimensions of industrial 
products over time could be used to control the quality of production laid the 
foundation for modern control charts. Shewhart is considered to be the father of 
statistical process control (SPC); his work provides the fi rst evidence of the transition 
from product quality (by inspection) to the concept of quality processes [18, 19] . 
Shewhart ’ s methodologies were adopted and expanded by W. Edwards Deming 
[20] and Joseph M. Juran [21] , who are credited with the birth of the “ total quality ” 
(TQC) approach in Japan following World War II. Successors to the total quality 
movement include management by objectives (MBO) (1960 ’ s), Crosby ’ s zero - 
defects (ZD) movement (1970s), the American incarnation of total quality management 
(TQM) (1970s – 1980s), quality circles (1970s), quality function deployment 
(QFD) (1980s), the International Organization for Standardization (ISO) 9000 
series (1987), and the Malcolm Baldridge National Quality Award (1987 – present). 
The most recent major quality management methodology, Six Sigma (6 . ) [22, 23] , 
pioneered by Motorola, has become immensely popular because of the litany of 
corporate CEOs (e.g., Thomas Galvin, Jack Welch) who have openly credited their 
internal 6 . initiatives for dramatic improvements in bottom - line performance. All 
of these quality movements [24] , however, as well as PAT, are related to the principles 
of Shewhart, Deming, Juran, Crosby, Taguchi [25, 26] , and others, in that they 
are based on systematic methods for understanding the sources of variability in 
processes and minimizing their impact on product quality. 
The so - called DMAIC (defi ne, measure, analyze, improve, and control) methodology 
is a common framework used by improvement teams in many industries to 
apply the concepts of quality management to systematically identify, prioritize, 
and eliminate the root cause of quality problems. A variant of DMAIC, known as 
DMADV (defi ne, measure, analyze, design, and verify), is sometimes used when a 
process or operation requires complete redesign to bring about the desired quality 
improvement and is a central concept of the DFSS (design for six sigma) movement. 
The origins of DMAIC, DMADV, DFSS, and other various quality management 
cycles can be traced to the “ Shewhart cycle ” of (1) plan, (2) do, (3) study, and (4) 
act [24] . 
Arguably, the most important aspects of quality management for PAT are the 
concepts of quantitative process performance characterization using process capability 
indices as universal descriptors, which form the basis of the “ measure ” and 
“ analyze ” portions of the DMAIC model. Process capability indices consider simultaneously 
both process variability and process specifi cations to determine whether 

BASIS FOR PROCESS ANALYTICAL TECHNOLOGY 317 
the process is “ capable ” [27] . A process is said to be capable if the quality measurements 
for nearly all samples are within the specifi cation limits. A common version 
of the process capability index, Cpk , is calculated according to 
Cpk = 
. . ... 
... 
min 
USL 
, 
LSL 
3 
. 
. 
. 
. 3 
where . and . are the mean and standard deviation and USL and LSL are the upper 
and lower specifi cation limits, respectively, for a product quality measurement. 
Process capability indices are useful for process improvement studies because they 
transform diverse measures of quality (e.g., weight, concentration, rate) into dimensionless 
units, thereby allowing investigators to pinpoint major sources of variation 
in a process (operations which have the lowest Cpk scores) when many measurement 
systems and quality attributes are involved. 
The process capability index, Cpk , is related to the so - called “ process sigma ” such 
that a 6 . process corresponds to a Cpk of exactly 2.00, or 2.0 defective parts per 
billion (PPB), assuming .N (0, . ) quality variance distribution (can alternative calculation 
for process sigma estimates 3.4 defective parts per million for a 6 . process). 
Examples of the correspondence between Cpk , process sigma, and defect rate for 
.N (0, . ) distributions are shown in Figure 2 . The process capability (based on 
observed yield) of pharmaceutical manufacturers has been cited by some benchmark 
studies to be roughly 0.7 (2.1 . ) [28] . 
While industrial benchmarks clearly indicate that pharmaceutical manufacturers 
have many opportunities to improve quality control, direct comparison with other 
industries may be somewhat misleading. As opposed to such industries as semiconductor 
manufacturing, where defective parts are often readily apparent at some 
FIGURE 2 Graphical illustration of the correspondence between the defects per million 
opportunity (DPMO) process capability (C pk ) and process sigma (assuming normally distributed 
quality variation). 

318 REGULATORY AND INDUSTRIAL PERSPECTIVES 
point in the value chain (i.e., the device built from the part will fail), drug products 
suffer from a high degree of ambiguity in their quality specifi cations. 
For example, fi nished - product release specifi cations such as content uniformity 
are rarely correlated to clinical evidence; rather, they are set according to compendial 
test standards. Furthermore, the functional relationship between in - process 
material characteristics and fi nished - product quality is seldom known at a high level; 
hence, the assigned in - process specifi cations for some operations may over - or 
underestimate the true level of process capability. As the level of process understanding 
in the pharmaceutical industry increases, development of science - and 
evidence - based in - process and release specifi cations will improve the reliability of 
C pk as a tool for process characterization. 
For further information, the NIST/SEMATECH Handbook of Engineering Statistics 
, which is freely available online [23] , and the American Society for Quality 
( www.ASQ.org ), are excellent sources for background information and technical 
details related to quality management. 
4.1.2.3 Lean Manufacturing 
In contrast to quality management systems, which have clear parallels with PAT (i.e., 
reduction of quality variation), the links between PAT and lean manufacturing are 
less direct. In fact, while quality management systems are concerned with process 
analysis of quality variation, lean fl ow path management is concerned with process 
analysis of production time variation. Furthermore, the core concepts of lean manufacturing, 
however, provide the technology platform which the pharmaceutical 
industry will use to derive gains in production effi ciency from the adoption of PAT. 
Without considering the impact of PAT on production effi ciency [i.e., the return on 
investment (ROI) from implementing PAT], industry would have very little impetus 
to voluntarily embrace PAT. The following paragraphs are intended to provide a 
brief introduction to lean manufacturing; later portions of this chapter will discuss 
the business drivers for implementation of PAT. 
Lean manufacturing, or “ lean, ” is often misunderstood (not unlike TQM or 6 . ); 
for some people, lean business initiatives conjure “ slash - and - burn ” management 
tactics to reduce workforce levels or shut down low - productivity operations. In fact, 
lean manufacturing has been characterized as “ an amalgam of methodologies 
including industrial engineering, just - in - time (JIT) (Osadas ’ s) 5 - S ’ s, TQC, continuous 
quality improvement (CQI), Visual Control, Total Productive Maintenance 
(TPM), Quality Circles, and Kaizen ” [24] . 
The origins of lean manufacturing are often ascribed to the creation of the Toyota 
Production System (TPS) by the Toyota Motor Corporation. However, the history 
of lean manufacturing can be traced back to industrial developments which occurred 
more than 150 years before TPS. The foundation for modern manufacturing was 
laid by Eli Whitney in 1798; while Whitney is best known for his invention of the 
cotton gin, it is his invention of interchangeable parts and uniform production which 
revolutionized mass production ( www.EliWhitney.org ). 
Nearly a century later, Frederick W. Taylor introduced the concepts of time study 
and standardized work, coining the term scientifi c management . It was not until 1908, 
with Henry Ford ’ s introduction of the Model T, that the value of lean manufacturing 
was recognized worldwide. Henry Ford is considered by some to be the fi rst practi

BASIS FOR PROCESS ANALYTICAL TECHNOLOGY 319 
tioner of JIT manufacturing; furthermore, his manufacturing system has been 
described as the inspiration for TPS [29] . More recently Ford Motor Company has 
developed a modernized version of Henry Ford ’ s original system, the Ford Production 
System [24] , which borrows heavily from TPS. 
As a discipline of manufacturing science, lean manufacturing is a technical 
philosophy focused on the reduction of seven types of waste, or “ muda, ” in manufacturing: 
overproduction, waiting, transport, inappropriate processing, unnecessary 
inventory, excess motion, and defects. The transformation of a process to lean operation 
is accomplished using many tools and strategies. Arguably, the most important 
mechanism for change is to replace traditional “ make to forecast ” or “ push ” production 
scheduling with “ pull ” strategies, such as “ kanban ” cards. The principles of lean 
have been applied with success in manufacturing and service industries, as well as 
governmental entities. Not unlike quality management, there are literally hundreds 
of books describing the various tools and techniques used to apply lean methodologies. 
The Society of Manufacturing Engineers ( www.SME.org ) maintains publications, 
conferences, and a technical community devoted to production management 
and is a good fi rst source for more information on lean manufacturing. 
Compared with other industries, pharmaceutical manufacturers have been relatively 
late to adopt lean manufacturing; consequently, pharmaceutical cycle times 
are extremely long when compared with other industries [30, 31] . By comparing the 
ratio of total cost of goods sold (COGS) to inventory value for the top 22 publicly 
traded branded, generic, and biotech pharmaceutical companies to the reported 
fi gures for other process industries, a rough indication can be gained of how much 
less effectively pharmaceutical manufacturers manage their supply chains (Figure 
3 ). Furthermore, it would not be diffi cult for most industrial pharmaceutical scientists 
to fi nd common examples of each of the “ seven wastes ” in a typical pharmaceutical 
manufacturing facility. 
Admittedly, there are some constraints intrinsic to the industry which may ultimately 
prevent pharmaceutical manufacturers from achieving “ world - class ” supply 
chain and manufacturing performance. Furthermore, application of lean and quality 
management tools to pharmaceutical manufacturing is proving to be a unique challenge. 
A recent survey of 1500 pharmaceutical manufacturing professionals indicated 
that, while more than half of the companies surveyed have implemented lean, 
6 . , or operational excellence, less than half of those programs have yielded satisfactory 
results [32] . 
While these data seem to suggest that lean manufacturing is not suited to pharmaceutical 
manufacturing, it is important to consider that most lean methodologies 
(e.g., TPS) were developed for high - volume production of uniform products. 
Although many “ blockbuster ” drugs are produced in dedicated facilities or in plants 
specializing in only a few products, it is quite common for pharmaceutical manufacturers 
to produce many products in a single plant, having a high proportion of 
shared equipment. Traditional lean methods, such as kanban cards, are diffi cult to 
manage in a complex, “ high - mix ” production environment. In order to solve these 
limitations, innovative software algorithms for “ fl ow path management ” [33] have 
been developed to simulate, design, and optimize pharmaceutical production processes 
according to lean manufacturing principles. 
Furthermore, the effectiveness of lean manufacturing is limited by variability in 
the cycle time (C/T) for individual unit operations as well as by the fi nite risk of 

320 REGULATORY AND INDUSTRIAL PERSPECTIVES 
batch failure during production; this is true regardless of the complexity of the fl ow 
path or of the degree to which equipment is shared. Pharmaceutical manufacturers 
cope with such risks by building up long production queues to accumulate work in 
process (WIP) ahead of unit operations. While this helps to improve capacity utilization 
and overall equipment effectiveness (OEE), it decreases effi ciency by consuming 
working capital and increasing the intensity of overhead operations (required 
to fi nance, transport, and warehouse WIP). In order to gainfully implement lean, 
pharmaceutical manufacturers must fi rst minimize C/T variation and risks to product 
quality. 
Finally, it is well known that a signifi cant portion of the typical production C/T 
reported by industry is consumed by the delay between completion of a unit operation, 
sampling, analysis, reporting, and in - process or fi nished - product release. In 
some scenarios, PAT will enable manufacturers to release fi nished products to the 
market immediately, with no delay for manual, offl ine testing; this is the so - called 
real - time release (RTR) benefi t of PAT. Without PAT and RTR, the effectiveness 
of lean strategies in reducing C/T will be limited by the maximum rate of product 
inspection and release. Thus, it is critical for pharmaceutical manufacturers to deploy 
PAT and lean in parallel if real gains in process performance are to be realized. The 
lean – PAT concept is quite similar to lean – 6 . , or “ fusion management ” [24] . 
FIGURE 3 Ratio of total COGS to reported inventory value. The ratio of COGS to inventories 
is a rough indicator of supply chain velocity. A large ratio implies that inventories are 
small relative to COGS and are turned over frequently. Toyota Motors, for example, which 
is well known for effective supply chain management, has a much higher ratio of COGS to 
inventories than General Motors. 
Budweiser 
Pepsi 
Microsoft 
Tyson 
Toyota 
Kelloggs 
Kraft foods 
Groupe Danone 
DOW 
FMC 
Coca Cola 
Unilever 
BASF 
Proctor & Gamble 
HJ Heinz 
ConAgra 
Celgene 
Genzyme 
Amgen 
Biogen IDEC 
Genentech 
General Motors 
Gilead 
Watson Labs 
Barr Laboratories 
TEVA 
Alpharma 
Mylan 
Forest labs 
Johnson & Johnson 
Merck 
Bristol Myers Squibb 
Astra Zeneca 
Novartis 
Sanofi Aventis 
GSK 
Eli Lilly 
Pfizer 
Wyeth 
2.00 
0.00 
4.00 
6.00 
(COGS/Inventory) ratio 
8.00 
10.00 
12.00 
14.00

4.1.3 HISTORICAL FACTORS LIMITING IMPLEMENTATION OF PAT 
Despite the evidence of fi scal and competitive benefi ts enjoyed by the various 
industries which have embraced process analytics, pharmaceutical companies have 
been notoriously restrained in their efforts to deploy PAT. Indeed, the pharmaceutical 
industry has slipped so far behind peer industries that a well - known Wall Street 
Journal article from 2003 [34] characterized the manufacturing prowess of drug 
makers as lagging “ far behind potato - chip and laundry - soap makers. ” While the 
declaration was shocking to many, it was, nonetheless, an accurate assessment. 
Before indicting the industry for gross negligence, however, it is important to consider 
the various factors which have acted over time to create the current state of 
affairs. 
Over the years, dozens of excuses have been provided for the industry ’ s lack of 
manufacturing innovation; many of the reasons are well known and have been 
published elsewhere [35] . For the sake of simplicity, the factors limiting the adoption 
of PAT can be distilled into three categories: real and perceived technological barriers, 
lack of economic incentive, and regulatory disincentives. 
4.1.3.1 Real and Perceived Technological Barriers 
Despite the fact that near infrared spectroscopy (NIR) has been used industrially 
for decades [36] , there has been hesitance to accept and trust “ new ” process analytical 
measurement technologies as equivalent or superior to traditional methods. For 
example, when a discrepancy between online NIR and laboratory analyses is 
observed, it is rare that the destructive reference methods are ever targeted as the 
source of error, despite the fact that NIR is often the more precise method. The 
hesitance to trust more advanced, multivariate tools (which are perhaps less directly 
understood) has certainly been a detriment to progress in deploying PAT. 
Similar concerns persist with regard to chemometrics (multivariate data analysis), 
information technology (IT), and advanced controls. One reason for such behavior 
may be the practice of calibrating and validating PAT sensors by correlating their 
signals to traditional, laboratory - based reference methods and characterizing performance 
in terms of prediction error [37 – 39] . It is a truism of statistics that, no 
matter how sensitive or accurate the PAT sensor may be in detecting quality variation, 
the performance of the reference method will always limit the level of perceived 
accuracy. A much more accurate depiction of the performance of PAT sensors 
compared to reference techniques would be to compare analytical fi gures of merit, 
such as signal - to - noise ratio (S/N) or analytical sensitivity, which explicitly account 
for measurement precision [40, 41] . 
Even though the perceptions of PAT instrumentation have begun to improve, 
companies continue to worry that the intensity of product quality sampling afforded 
by PAT sensors will result in negative consequences, such as increased inspection 
and investigations. In other words, many companies continue to “ fear what they will 
fi nd ” if they begin to analyze their operations more closely. Prior to the introduction 
of rapid, nondestructive quality monitoring tools, there were few alternatives for 
effi cient quality assurance except to rely on batch release criteria, such as the well - 
known U.S. Pharmacopeia (USP) . 905 . procedure, which were based on extremely 
limited sampling (i.e., assay 10 individual dosage units from a 30 - unit sample of a 
production - scale batch). 
HISTORICAL FACTORS LIMITING IMPLEMENTATION OF PAT 321

322 REGULATORY AND INDUSTRIAL PERSPECTIVES 
Despite the fact that the operating characteristic (OC) curve of the USP . 905 . 
test guarantees a signifi cant portion of each batch will have poor quality before 
batch rejection is probable [42, 43] , companies have become comfortable with their 
odds. Process analytical monitoring tools such as NIR spectroscopy, which are 
capable of high - speed sampling in line, online, or at line, have been perceived as an 
additional burden on the rate of successful batch release. 
By forgoing real - time, pervasive quality monitoring, however, companies incur 
signifi cant opportunity costs in at least three ways. First, without continuous monitoring 
there are few feasible opportunities for implementing RTR; time delays 
related to offl ine release testing are one of the most signifi cant factors limiting 
supply chain velocity in pharmaceutical manufacturing. Second, while there is some 
potential for “ discovering ” a greater number of batches which do not meet release 
criteria, statistical simulations suggest that potentially fewer batches will be rejected 
when larger sample sizes are considered. In other words, when the impact of measurement 
imprecision and the true distribution of quality characteristics are considered, 
traditional release testing methods pose fi nite risks of failing passable batches 
(which otherwise should have passed) because the limited sample does not adequately 
represent the characteristics of the population (Figure 4 ). Finally, and 
perhaps most importantly, traditional sampling techniques are an effective barrier 
to continuous improvement; based on fundamentals of statistical theory, it can be 
shown that samples of at least hundreds of individuals are required to detect incremental 
changes in process capability (Figure 5 ). Hence, even if a company were to 
investigate potential process improvements, only process capability changes of 
improbable magnitude would be recognized with statistical confi dence. 
FIGURE 4 Comparison of operational characteristic (OC) curves for the USP . 905 . ( a ) 
and PAT - based ( b ) release strategies generated by Monte Carlo simulation. The USP OC 
curve ( a ) is based on the assumption of 2% RSD measurement precision; the PAT OC curve 
( b ) assumes NIR measurement of 800 tablets with 0.9% measurement precision; both curves 
were estimated using the same simulated populations of one million tablets having varying 
levels of quality uniformity. Each curve consists of four regions: the regions above and below 
the sigmoid curve correspond to proportions of batches accurately passed or rejected based 
on the release criteria. Along the sigmoid curve are regions related to the rates of false batch 
failure (lower side of curve) and false batch acceptance (upper side of curve). The jagged 
nature of the curves is related to the limitations imposed by fi nite iterations. The slope of the 
curves demonstrates the superior specifi city (or “ tunability ” ) of release tests optimized for 
PAT systems. 
100 
90 
80 
70 
60 
50 
40 
30 
20 
10
0 
Batches (%) 
98 96 94 92 90 88 
Within-batch true coverage (%) 
False 
fail lots
False pass 
lots 
(a) (b) 
98 96 94 92 90 88 
Within-batch true coverage (%)

4.1.3.2 Lack of Economic Incentive 
A common refrain within the industry has been that there simply is not suffi cient 
fi nancial return from investment in process analytics or manufacturing technology 
upgrades to justify spending. In some respects, this is a valid argument. Historically, 
many of the industries which have justifi ed signifi cant investment in process analytics 
utilized continuous manufacturing; it is far more diffi cult to effi ciently control continuous 
processes (relative batch production systems) without real - time process analytics 
[35] . Hence, while the pharmaceutical industry has been able to choose, many 
other manufacturers have been forced to integrate PAT into their operations. 
Since pharmaceutical investment in PAT continues to be an option rather than 
a priority for most companies, arguments justifying PAT spending are forced to 
compete with other spending initiatives for capital. During each planning cycle, 
company managers must decide whether to allocate additional capital toward 
diverse opportunities, such as greater research and development (R & D), improvements 
in manufacturing capabilities, or additional forces in sales and marketing [i.e., 
selling, general and administrative (SG & A)]. For any particular project to be funded, 
expected returns must not only exceed the company ’ s cost of capital [i.e., weighted 
average cost of capital (WACC)], winning projects may be required to exceed the 
company ’ s expected return on invested capital (ROIC) or at least provide expected 
returns in excess of other investment alternatives. A recent academic case study of 
the potential fi nancial returns on investment (ROI) in PAT and lean manufacturing 
in the pharmaceutical industry show, however, that many pharmaceutical manufacturers 
could ultimately benefi t tremendously by improving manufacturing performance 
[44] . 
FIGURE 5 Relationship between sampling rate and effective resolution of process capability 
assessment. The curve is based on the width of the confi dence intervals for estimation of 
mean and variance. The relationship shown does not consider the effect of reference measurement 
precision, which would further reduce the ability to discern changes in process 
capability. 
Detectable change in process capability (95% confidence) 
as a function of sampling rate 
Detectable change in Cpk (%) 
Samples assayed (N) 
10 100 200 300 400 500 600 700 800 900 1000 
1
2
3
4 5 
10 
20 
30 
50 
100 
USP <905>, 30 Samples 
HISTORICAL FACTORS LIMITING IMPLEMENTATION OF PAT 323

324 REGULATORY AND INDUSTRIAL PERSPECTIVES 
Unfortunately, proponents of PAT are only just beginning to develop the methods 
to quantify all of the potential opportunities for ROI. Furthermore, it is important 
to consider the relative level of risk posed by investment in PAT (as opposed to 
other alternatives). Unlike investments in sales or marketing, there remains 
considerable uncertainty in the industry regarding the likelihood of achieving ROI 
projections or the prospect of PAT investment creating new problems. For these 
reasons, management teams have typically found it easier to justify spending in 
R & D and marketing instead of PAT or manufacturing reforms. 
Besides concerns over the likelihood and magnitude of returns on PAT investments, 
it is often cited that manufacturing and optimizing the cost of production 
have simply not been a priority in the industry; manufacturing has often been 
viewed as a cost rather than a value - generating component. The distribution of 
corporate expenditures has been provided as evidence in support of this theory 
(Figure 6 ). Based on corporate annual income statements from 2005, the average 
expenditure on R & D and SG & A among the top - 10 branded pharmaceutical companies 
(by market capitalization, November 7, 2006) was nearly double their reported 
cost of goods sold. Another take on this theory is that institutional and individual 
investors (who own the pharmaceutical companies and supply the capital for their 
operation) and the boards of directors elected by them look favorably on the expansion 
of R & D and marketing investment while taking a more myopic view on the 
importance of manufacturing. It has sometimes been said that Wall Street rewards 
FIGURE 6 Distribution of the components of revenue (FY2005 annual data) for branded 
( a ), generic ( b ), and biotech ( c ) drug manufacturers. Companies are arranged according to 
market capitalization (as of November 2006). 
0 
10 
20 
30 
40 
50 
60 
70 
80 
90 
100 
Net profit 
Tax, interest, other 
Research & development 
Net profit 
Tax, interest, other 
Research & development 
Cost of goods sold 
Cost of goods sold 
Selling, general & administrative 
Selling, general & administrative 
JNJ 
AMGN DNA GILD CELG GENZ BIIB 
PFE GSK NVS SNY 
Company (ticker symbol) 
Company (ticker symbol) 
MRK AZN WYE LLY BMK 
(a) 
% of revenues 
0 
10 
20 
30 
40 
50 
60 
70 
80 
90 
100 
Net profit 
Tax, interest, other 
Net profit 
15% 
Other Exp. 
12%
R&D 
15% 
Branded 
Pharma 
COGS 
25% 
SG&A 
33% 
Net profit 
16% 
Other Exp. 
10%
R&D 
9% 
Generics 
COGS 
40%
SG&A 
25% 
Net profit 
19% 
Other Exp. 
14% 
R&D 
23% 
Biotech 
COGS 
16% 
SG&A 
28% 
R&d 
Cost of goods sold 
Selling, general & administrative 
TEVA FRX BRL 
Company (ticker symbol) 
MYL WPI ALO 
(b) 
% of revenues 
0 
10 
20 
30 
40 
50 
60 
70 
80 
90 
100 
(c) 
% of revenues

(pharmaceutical companies) for innovation in discovery and replication in manufacturing 
[45] . It is not completely coincidence, for example, that Merck ’ s appointment 
of its president of manufacturing, Richard T. Clark, to chief executive in May 
2005, which, according to fi nancial journalists, “ disappointed investors ” who apparently 
would have preferred someone with a “ research and development background 
” [46] , marked the beginning of a nearly 25% loss in market capitalization 
over the next six months. 
While the various reasons discussed for the pharmaceutical industry ’ s tepid 
approach to PAT and manufacturing reform are plausible, they are likely secondary 
to the real and perceived risks posed by the regulatory uncertainty surrounding 
innovation in manufacturing. For example, it is well known that many companies 
were beginning to use PAT tools long before the FDA ’ s initiative, which suggests 
that the economic benefi ts of process analytics have been recognized internally for 
some time. In response to the fear that their use of new technologies would spur 
additional investigations by the FDA, however, some of these companies operated 
in a “ Don ’ t use, or don ’ t tell ” manner with regard to PAT [45] . 
4.1.3.3 Regulatory Disincentives 
The real and perceived fear of regulatory noncompliance has arguably been one of 
the most important factors explaining the industry ’ s reluctance to pursue manufacturing 
innovation [1, 2] . While the fi rst 25 years of pharmaceutical GMP have been 
effective in ensuring the safety of prescription drug products for consumers, it has 
been achieved at the expense of innovation and fl exibility. Without the ability to 
adjust processes to account for changes in materials, operating conditions, or the 
level of process understanding, process analytics are of nearly no value since there 
is no capacity to act on new information (besides material/batch rejection). 
Furthermore, companies who dared to make changes or implement new technologies, 
whether conventional process improvements, new unit operations, or 
process analytics, were met with extensive supplemental documentation, FDA 
inspection, and the fi nite risk of production delays. Ultimately, the potential for 
regulatory action stifl ed the industry ’ s desire to pursue technologies which might 
have seemed extraordinary, such as real - time analytics or chemometrics. Finally, 
without the benefi ts conferred by the PAT guidance and risk - based cGMPs initiative, 
industry rarely had incentive to formally analyze the risk of established processes 
out of fear that what they might discover would be used against them in 
regulatory or legal actions. 
4.1.4 FDA TWENTY - FIRST - CENTURY c GMP s INITIATIVE 
The observation that the state of cGMP at the beginning of the twenty - fi rst - century 
was stifl ing innovation in pharmaceutical manufacturing did not go unnoticed by 
the FDA, which also saw opportunity in remodeling the regulatory framework. 
Since many changes, even minor operational modifi cations, required prior approval 
from the agency prior to implementation, regulators were swamped with thousands 
of supplements every year. Resources were stretched between processing of supplements, 
review and approval of new facilities, processes and documentation, and 
inspection; all the while, the FDA was being squeezed by external constraints on 
FDA TWENTY-FIRST-CENTURY cGMPs INITIATIVE 325

326 REGULATORY AND INDUSTRIAL PERSPECTIVES 
budget growth (Figure 7 ). As of 2001, FDA regulators were so burdened that they 
were unable to meet statutory biennial GMP inspections. Finally, the load of supplements, 
reviews, and inspections were acting as a signifi cant drag on the advancement 
to market of new pharmaceutical therapies. 
4.1.4.1 Conception of the Initiative 
The agency began a public dialogue on the state of pharmaceutical manufacturing 
and FDA regulation during discussions with the Advisory Committee for Pharmaceutical 
Science (ACPS) in July 2001, followed by further discussion within the FDA 
Science Board meetings in November 2001 and April 2002 [47] . A signifi cant focus 
of the discussions was the impact of the regulatory framework on innovation, quality, 
and effi ciency as well as opportunities for change. A new, risk - based paradigm which 
rewards innovative producers through opportunities for “ regulatory relief ” began to 
take shape, displacing the notion of regulatory compliance as a force for innovation. 
The new paradigm offered advantages to the FDA, as well, in that the level of inspection 
resources could be prioritized and allocated according to risk, thereby easing 
FIGURE 7 Trends in FDA workload and staffi ng resources. ( Adapted from L. X. Yu, Implementation 
of quality - by - design: Question - based review, Drug Information Association (DIA) 
42nd Annual Meeting, Philadelphia, PA, 2006 .) 
1000 
800 
600 
400 
200
0 
2001 2002 2003 2004 2005 
ANDAs 
Employees 
4000 
3500 
3000 
2500 
2000 
2001 2002 2003 2004 
Supplements

the strain on FDA resources. These changes signaled an evolution of what seemed 
to be an adversarial FDA – industry relationship toward greater cooperation. 
While the pharmaceutical incarnation of the term PAT was formally introduced 
during these meetings [48] , a signifi cant portion of the concepts which defi ne the 
core of PAT in pharmaceutical science were presented by industrial and academic 
scientists, many of whom had been building support for and working on these issues 
within their organizations for years. Industrial and academic presentations included 
topics such as total quality management [49] , new technologies for pharmaceutical 
manufacturing [50] , and QbD [51] , among others. 
In August 2002, the agency announced the Pharmaceutical cGMPs for the 21st 
Century initiative (or “ the initiative ” ), which began a two - year effort undertaken 
by a number of multidisciplinary working groups within the FDA, as well as the 
cGMP steering committee, to assess the current regulatory structure and defi ne the 
agency ’ s new vision for risk - based regulation of manufacturing and product quality. 
The new initiative, which was intended to modernize the FDA ’ s regulation of 
pharmaceutical quality for human, veterinary, and select human biological products, 
sought to reform the pharmaceutical as well as the chemistry, manufacturing, and 
controls (CMC) programs, with the following specifi c objectives: 
• Encourage the early adoption of new technological advances by the pharmaceutical 
industry. 
• Facilitate industry application of modern quality management techniques, 
including implementation of quality systems approaches, to all aspects of pharmaceutical 
production and quality assurance. 
• Encourage implementation of risk - based approaches that focus both industry 
and agency attention on critical areas. 
• Ensure that regulatory review, compliance, and inspection policies are based 
on state - of - the - art pharmaceutical science. 
• Enhance the consistency and coordination of the FDA ’ s drug quality regulatory 
programs, in part, by further integrating enhanced quality systems approaches 
into the agency ’ s business processes and regulatory policies concerning review 
and inspection activities. 
The result of the working groups ’ assessment enabled the development of the 
new framework embodied by the fi nalized twenty - fi rst - century cGMPs as well as 
the associated components, such as the PAT guidance. Throughout the assessment 
and development, and continuing during the “ implementation phase ” of the initiative, 
the following set of guiding principles has been maintained: 
• Risk - based orientation 
• Science - based policies and standards 
• Integrated quality systems orientation 
• International cooperation 
• Strong public health protection 
The fi nal report on the results and future plans for the initiative were released in 
September 2004. The report effi ciently describes the motives, origins, development 
FDA TWENTY-FIRST-CENTURY cGMPs INITIATIVE 327

328 REGULATORY AND INDUSTRIAL PERSPECTIVES 
process, and mechanisms for implementing and evaluating the initiative and can be 
found posted on the Center for Drug Evaluation and Research (CDER) Offi ce of 
Pharmaceutical Science (OPS) website ( http://www.fda.gov/cder/OPS/ ). 
Since it would be impractical to accurately describe all of the important aspects 
of the report within this space, the following sections are intended to detail some 
of the concepts and guiding principles of the initiative which are particularly important 
for understanding PAT. The organization of this summary is intended to effi - 
ciently describe selected concepts of the agency ’ s twenty - fi rst - century cGMPs and 
is not intended to mirror the structure or totality of the associated FDA documentation. 
All who are actively engaged in pharmaceutical manufacturing or are interested 
in PAT are encouraged to read the fi nal report [1] , which should be considered 
a primary source for direction. 
4.1.4.2 Risk - Based Orientation 
The FDA ’ s adoption of a risk - based orientation for regulation is the most important 
aspect of the twenty - fi rst - century cGMPs. It is a common misconception that the 
agency ’ s initiative describes a new set of practices for the industry. In fact, while the 
FDA is committed to encouraging innovation in the industry, the twenty - fi rst - 
century cGMPs initiative is entirely focused on changing the agency ’ s regulatory 
framework so that quality and innovation are rewarded with reduced oversight. 
Now that the agency has entered the implementation phase of the initiative, many 
of the previous regulatory disincentives have been eliminated. In other words, pharmaceutical 
companies are currently free to voluntarily choose whether or not to 
pursue innovative changes in their development, operation, and quality assurance 
of manufacturing processes such as PAT. 
Risk - Based Prioritization of c GMP Inspections The mechanism by which the 
FDA will encourage the industry to join in implementing the new methods is provided 
by the risk - based algorithm for prioritizing cGMP inspections. Incidentally, 
risk - based site selection is the same mechanism which will allow the agency to 
optimally allocate its limited oversight resources to achieve the greatest public 
health impact. Operational effi ciency is a major component of the FDA ’ s plans for 
the future. The key to the risk - based site selection program is the agency ’ s risk - 
ranking model, which has been deployed as a pilot program since the beginning of 
its 2005 fi scal year. 
The model is based on a hierarchical risk - ranking and risk - fi ltering method 
whereby a site risk potential (SRP) is estimated as a function of the weighted 
potentials for each of three top - level components of site risk — product, facility, and 
process (Figure 8 ). The risk potential for each of the three top - level components 
is calculated as a function of selected risk factors which are relevant to the component 
(specifi c to the site). A set of subcategories are defi ned for each top - level 
component; each subcategory is comprised of individual risk factors. The initial 
model weights (the actual risk scores at the lowest level) were optimized using 
a combination of empirical evidence and expert judgment. Examples of potential 
risk factors for each top - level component (and associated subcategories) were provided 
in a report which describes the fi rst iteration of pilot risk - ranking model in 
detail [52] . 

The results from the fi rst iteration of the risk - ranking model demonstrated the 
capability of the model to spread SRP scores for the purpose of fi ltering. Future 
iterations of the risk - ranking model will be generated by correlating predicted site 
risk potentials with data gathered by traditional oversight activities (e.g., cGMP 
compliance inspections) and adjusting the risk factor weights to maximize the effectiveness 
of SRP prediction (similar to multivariate linear regression). The selection 
of risk factors included in the fi rst iteration of the model was based on the availability 
of data. Some proposals for future iterations of the model include incorporating 
factors such as systems for continuous assessment of process capability as 
indicators of the site ’ s level of process understanding and control. Certainly, as the 
model is updated to capture the benefi ts of new best practices in manufacturing, 
such as PAT, the risk ranking will begin to provide effective incentive for producers 
to pursue innovation. 
4.1.4.3 Quality Systems Approach 
According to the FDA staff manual guide [53] , a quality system is a “ set of formal 
and informal business practices and processes that focus on customer needs, 
leadership vision, employee involvement, continual improvement, informed decision 
making based on real - time data and mutually benefi cial relationships with 
external business partners to achieve organizational outcomes. ” Based on this 
description, PAT should be considered to be an important tool for supporting a 
quality management system. As stated earlier, one of the FDA ’ s objectives in undertaking 
the initiative was to integrate quality systems and risk management approaches 
into its existing programs with the goal of encouraging industry to adopt modern 
and innovative manufacturing technologies, including industrial deployment of 
quality management systems such as those described earlier in this chapter (e.g., 
ISO 9000). 
In September 2006, the FDA released “ Guidance for Industry: Quality Systems 
Approach to Pharmaceutical cGMP Regulations ” [54] . The guidance is intended to 
“ help manufacturers implementing modern quality systems and risk management 
approaches to meet the requirements of the Agency ’ s cGMP regulations, ” in particular, 
Parts 210 and 211. In developing the guidance, the Quality System Guidance 
FIGURE 8 Schematic of FDA ’ s pilot risk - ranking model for calculation of site risk 
potential. 
CD1 CD2 CP1 CP2 CF1 CF2 
Top-level 
components 
Categories of 
risk factors 
Risk factors 
(quantitative or 
qualitative 
variables) 
Site risk potential 
Product Process Facility 
FDA TWENTY-FIRST-CENTURY cGMPs INITIATIVE 329

330 REGULATORY AND INDUSTRIAL PERSPECTIVES 
Development (QS) working group “ mapped ” the relationship between cGMP 
regulations and various quality system models both internal and external to the 
FDA. Their result is a comprehensive model which allows producers seeking to 
implement their own quality management systems to quickly identify those aspects 
of quality systems which are, and are not, correlated with cGMP. 
The QS guidance begins by defi ning critical concepts of modern quality systems, 
including quality, QbD and product development, quality risk management, corrective 
and preventative action (CAPA), change control, the “ quality unit, ” and the 
six - system inspection model. The discussion of the quality unit describes its relationship 
with the concepts of quality control (QC) and quality assurance (QA) and the 
relationship between the quality unit and the other units within the pharmaceutical 
manufacturing organization. The six - system inspection model is described as a 
blueprint for how compliance inspections will be organized under the new quality 
systems approach and should be considered a template for internal verifi cation of 
compliance within pharmaceutical organizations adopting quality management 
systems (Figure 9 ). 
The majority of the QS guidance is devoted to describing the essential components 
of modern quality systems, including four major factors which must be 
addressed: management responsibilities, resources, manufacturing operations, and 
evaluation activities. Each factor is described in detail, including aspects which 
overlap with cGMP regulations (for each factor there is a table listing the related 
regulatory citations). In particular, the manufacturing section describes aspects of 
quality systems (and related cGMPs) which are closely related to PAT, including 
raw materials analysis, operations monitoring, and procedures for addressing nonconformities. 
Finally, the guidance includes many important references and related 
guidance documents which should be considered by companies seeking to implement 
a quality management system. 
FIGURE 9 FDA ’ s six - system inspection model. 

4.1.4.4 Science - Based Policies 
Continuous improvement, which the agency describes as an “ essential element in a 
modern quality system, ” is aimed toward improving effi ciency by “ optimizing a 
process and eliminating wasted efforts in production ” [1] . One of the unintended 
consequences of the regulatory system (prior to the new initiative) had been the 
suppression of nearly all opportunities for continuous improvement in manufacturing 
once a pharmaceutical product has been approved for market. Changes to formulations 
and processes needed to be justifi ed regarding their impact on product 
quality, often requiring time - consuming postapproval supplements. Producers in 
most other modern industries (many of whom deal with public safety risks on par 
with or exceeding those managed by the pharmaceutical industry) make it a practice 
to continuously fi ne tune and adjust their operations to maximize quality and effi - 
ciency. Pharmaceutical manufacturers, on the other hand, have largely been constrained 
to treat demonstrated processes as if they were set in stone. 
While there is some logic to limiting the scope and pace at which changes can be 
made to processes, there is obvious fallacy in the idea that the fi rst approved con- 
fi guration for a drug manufacturing operation will be optimal, especially considering 
the enormous fi nancial and ethical pressures on process development teams to 
quickly bring new drug therapies to market. This realization spurred the agency to 
begin the process of developing science - based policies and standards to facilitate 
innovation , which currently includes three new updated guidance documents: 
“ Sterile Drug Products Produced by Aseptic Processing — cGMP ” [55] , the PAT 
guidance, and the draft guidance on comparability protocols. Each guidance document 
encourages voluntary adoption of new technologies in pharmaceutical manufacturing 
by defi ning modern, science - based regulatory mechanisms which enable 
producers to implement strategic improvements with opportunities for more effi - 
cient regulatory compliance. 
Comparability Protocols In fact, pharmaceutical manufacturers have always had 
the option to explore changes to their production processes. The difference between 
the old regulatory paradigm and the twenty - fi rst - century cGMPs initiative is that 
producers who seek to improve the quality and effi ciency of their processes will be 
able to implement changes much more quickly while spending signifi cantly fewer 
resources to maintain compliance. The key to achieving these benefi ts is demonstrating 
that there is suffi cient understanding of the process and changes to be made and 
that implementation of the improvements poses very little risk to consumers. 
A new mechanism for implementing process changes, which refl ects the inclination 
for science - based policies, is detailed in the FDA ’ s draft guidance “ Comparability 
Protocols — Chemistry, Manufacturing, and Controls (CMC) Information ” . A comparability 
protocol (CP) is a “well-defi ned, detailed, written plan for assessing the 
effect of specifi c CMC changes in the identity, strength, quality, purity, and potency 
of a specifi c drug product as these factors relate to the safety and effectiveness of 
the product ” [56] . Submission of a CP by a producer is optional and may be used to 
facilitate changes in a manufacturing process, analytical procedures, manufacturing 
equipment or facilities, or container closure systems or for implementation of PAT. 
The benefi t for producers submitting a CP is that, upon approval of a CP, “ the 
FDA can designate, where appropriate, a reduced reporting category for future 
FDA TWENTY-FIRST-CENTURY cGMPs INITIATIVE 331

332 REGULATORY AND INDUSTRIAL PERSPECTIVES 
reporting of CMC changes covered by the approved CP ” . For example, changes that 
otherwise would require submission, review, and acceptance of a postapproval supplement 
(PAS) might be designated as annual report (AR) changes if they were provided 
for in an approved CP. The CP is one of the mechanisms by which the FDA intends 
to reduce the number of supplements requiring review. Additionally, the CP was 
designed to facilitate free fl ow of communication with the agency, thereby reducing 
the risk that process changes will lead to unexpected regulatory shutdown or delay. 
Process Validation In agreement with the pursuit of science - based policies, the 
FDA has begun to revise the 1987 “ Guideline of General Principles of Process Validation 
” and in March 2004 released a revision of the compliance policy guide (CPG) 
(Section 490.100) “ Process Validation Requirements for Drug Products and Active 
Pharmaceutical Ingredients Subject to Pre - Market Approval ” [52] . The current revisions 
are designed to support continuous improvement and replace the notion of 
“ three - batch ” validation. The CPG describes the concept that, after having identi- 
fi ed and established control of all critical sources of variability, conformance batches 
are prepared to demonstrate that under normal conditions and operating parameters 
the process results in the production of acceptable product. However, the CPG 
does not describe how many conformance batches are required; rather, the manufacturer 
is expected to provide “ sound rationale ” for the procedure they choose to 
follow in demonstrating validation. 
The ambiguity in the revised (CPG) regulations may seem to signify that manufacturers 
would need to undertake even more extensive validation exercises when 
in fact the CPG contains language providing a pathway for batch release to market 
distribution concurrent with the manufacture of initial conformance batches or with 
a single conformance batch [57] : 
Advanced pharmaceutical science and engineering principles and manufacturing 
control technologies can provide a high level of process understanding and control 
capability. Use of these advanced principles and control technologies can provide a 
high assurance of quality by continuously monitoring, evaluating, and adjusting every 
batch using validated in - process measurements, tests, controls, and process endpoints. 
For manufacturing processes developed and controlled in such a manner, it may not 
be necessary for a fi rm to manufacture multiple conformance batches prior to initial 
distribution. 
Interpretation of the CPG suggests that implementation of PAT can be an important 
consideration for streamlining process validation. Finally, a quotable interpretation 
of the new science - based paradigm suggests that (instead of validating the process) 
producers should “ control the process, and validate the controls. ” Beyond revision 
of the CPG, FDA is expected in the near future to release draft guidance on process 
validation, which will be closely aligned with concepts associated with PAT, QbD, 
and the rest of the 21 st century cGMPs. 
4.1.4.5 International Collaboration 
Recognizing the current realities of the global marketplace, the FDA has made 
coordination with international regulatory partners a priority of the twenty - fi rst - 
century cGMPs initiative. By increasing its collaboration with international health 
and regulatory partners, the FDA has been able to leverage its resources through 

increased sharing of information and harmonization of activities. The International 
Conference on Harmonization of the Technical Requirements for Registration of 
Pharmaceuticals for Human Use (ICH) ( www.ich.org ) has been the dominant mechanism 
for international cooperation among pharmaceutical regulatory authorities 
in Europe, Japan, and the United States. 
A consensus vision statement was drafted at the July 2003 ICH meeting with 
regard to the objective of the ICH in harmonizing the efforts of regulatory bodies 
to establish quality systems approaches in their operations: “ Develop a harmonized 
pharmaceutical quality system applicable across the life cycle of the product emphasizing 
an integrated approach to quality risk management and science. ” 
Three consensus guidelines defi ne the core of the ICH ’ s involvement in harmonization 
of pharmaceutical quality systems — Q8: Pharmaceutical Development, Q9: 
Quality Risk Management, and Q10: Pharmaceutical Quality Systems (in addition, 
each of the guidance documents cites critical areas of overlap with Q6A: Specifi cations: 
Test Procedures and Acceptance Criteria for New Drug Substances and New 
Drug Products: Chemical Substances). 
Q 8: Pharmaceutical Development According to the ICH Q8 guideline [58] , 
the aim of pharmaceutical development is to “ design a quality product and the 
manufacturing process to deliver the product in a reproducible manner. ” While 
QbD is not specifi cally mentioned in the guideline, the intent of the ICH Q8 expert 
working group (EWG) was to describe a system that would provide incentive for 
manufacturers to incorporate aspects of QbD and continuous improvement throughout 
the product life cycle. In achieving this goal, the guideline they produced 
describes the suggested contents for Section 3.2.P.2 of a regulatory submission in 
the ICH M4 common technical document (CTD) [59] and the FDA electronic 
common technical document (eCTD) [60] . 
The pharmaceutical development and quality overall summary (QOS) sections 
of the CTD (Figure 10 ) provide pharmaceutical scientists with dedicated channels 
to present regulators with the relevant knowledge and process understanding gathered 
during the development of a new product (which can be updated to support 
new knowledge gained over the life cycle of the product following approval). The 
knowledge communicated within these sections are important considerations for 
justifi cation of a lower site risk potential (i.e., SRP, with regard to risk - based inspection) 
and for facilitation of effi cient, question - based review (QbR) [61] . Question - 
based review is another mechanism by which the agency intends to streamline the 
regulatory process as well as reward producers for adopting best practices in quality 
management. 
In addition to facilitating risk - based oversight, the content of the pharmaceutical 
development and QOS sections of the CTD are critical to enabling continuous 
improvement and fl exible operation. The information and knowledge communicated 
within these sections provide scientifi c understanding to support the establishment 
of a manufacturing design space, in - process and release specifi cations, and 
manufacturing controls. 
As described within the Q8 guideline, a design space is the “ multidimensional 
combination and interaction of input variables and process parameters that have 
been demonstrated to provide assurance of quality. ” So long as process control is 
maintained within the bounds of the design space, operating parameters can be 
adjusted to improve product quality or manufacturing effi ciency. Based on the 
FDA TWENTY-FIRST-CENTURY cGMPs INITIATIVE 333

334 REGULATORY AND INDUSTRIAL PERSPECTIVES 
current defi nition, operation outside of the established design space would initiate 
a regulatory postapproval change process. Thus, complete and accurate communication 
of the knowledge supporting a company ’ s design space is vital for a company 
to maximize productivity while maintaining regulatory compliance. Furthermore, 
with the new communication pathways in place, companies have incentive to pursue 
manufacturing studies beyond marketing approval to expand their design space or 
to update specifi cations and controls. In addition to the product under review, if 
appropriate, experiences gained from the development (and manufacture) of similar 
drug products may be included. 
Q 9: Quality Risk Management The second working group (ICH Q9 EWG) is 
trying to better defi ne the principles by which risk management will be integrated 
into decisions by regulators and industry regarding quality, including cGMP compliance. 
In November 2005, the Q9 EWG released the “ Step 4 ” version of the Q9 
guideline which defi nes the two primary principles of quality risk management, 
provides a model for the quality risk management process (Figure 11 ), and describes 
the terminology and tools for risk assessment and management. In addition, the 
document includes a concise reference list for more detailed information on risk 
management methods, such as failure mode effect and criticality analysis (FMECA), 
which are important tools for prioritized implementation of PAT. While it is not 
intended to be a “ how to ” manual for risk management, the Q9 guideline is a valu- 
FIGURE 10 Schematic illustration of the ICH M4 common technical document (CTD); 
the contents of the Quality Overall Summary (2.3) and Quality (3) modules are most relative 
to PAT. 
Module 2 
Module 1 
Regional 
Administrative 
Information 
1 
1.1 Submission 
Module 3 Module 4 Module 5 
Not part of the CTD
CTD 
CTD Table of Contents 
2.1 
CTD Introduction 
2.2 
Quality 
Oveall 
Summary 
2.3 
Quality 
3 
Nonclinical 
Overview 
2.4 
Nonclinical 
Study Reports 
4 
Clinical 
Overview 
2.5 
Nonclinical Written 
and Tabulated 
Summaries 
2.6 
Clinical 
Summary 
2.7 
Clinical 
Study Reports 
5

able information source for companies seeking to incorporate quality risk management 
into their operations [62] . 
Q 10: Pharmaceutical Quality Systems While the Step 2 document for the third 
tripartite guideline, Q10: Pharmaceutical Quality Systems, has not yet been released, 
the fi nal concept paper has been available since 2005 [63] . Similar to the manner by 
which the FDA ’ s quality systems approach guidance mapped the relationship 
between cGMPs and other industrial quality management systems, the Q10 guideline 
is anticipated to serve as a bridge between the approaches to quality systems 
taken by the different regional regulations, thereby helping to achieve global harmonization 
of quality systems. The guideline is expected to strengthen and complement 
issues covered in Q6A, Q8, and Q9 and will provide a foundation for a 
pharmaceutical quality system based on elements from the ISO 9001 and 9004 
standards. The guideline is also expected to develop harmonized defi nitions for 
issues critical to PAT, including continuous improvement activities, data - gathering 
methods, and the approach to measurement system validation. 
4.1.5 PAT EVOLUTION IN PHARMACEUTICAL MANUFACTURING 
Though it may be tempting to characterize PAT as a revolutionary change in pharmaceutical 
manufacturing, history will likely show that the beginning of the twenty - 
fi rst - century cGMPs initiative and the development of the PAT guidance mark the 
FIGURE 11 Schematic of quality risk management process described within ICH Q9. 
Initiate 
Quality Risk Management Process 
Risk Assessment 
Risk Identification 
Risk Analysis 
Risk Acceptance 
Risk Evaluation 
Risk Control 
Risk Reduction 
Output / Result of the 
Quality Risk Management Process 
Risk Review
Review Events 
unacceptable 
Risk Management tools 
Risk Communication 
PAT EVOLUTION IN PHARMACEUTICAL MANUFACTURING 335

336 REGULATORY AND INDUSTRIAL PERSPECTIVES 
beginning of a period of rapid evolution in pharmaceutical manufacturing which 
will extend far into the future. Even though the twenty - fi rst - century cGMPs initiative 
is more extensive (with regard to changing the relationship between the FDA 
and the pharmaceutical industry), interest in the PAT guidance and the opportunities 
it presents for the industry were initially much greater. More recently, perhaps 
in parallel with some changes in leadership in the agency, there has been a palpable 
shift of emphasis toward QbD, which was barely mentioned in many of the twenty - 
fi rst - century cGMPs documents. It is important to keep in mind that, just as most 
industries have seen a parade of “ new ” quality systems initiatives over the years 
since Shewhart ’ s fi rst methods were published, the principles upon which PAT and 
QbD are built, such as robust process design, quality monitoring, and effective controls, 
will persist regardless of the name of the initiative. Furthermore, as with PAT, 
QbD is not a new concept. Indeed, Dr. Genichi Taguchi, who has been credited by 
some as the father of QbD, began applying QbD in pharmaceutical manufacturing 
while working as a statistical consultant for Morinaga Pharmaceuticals Company of 
Japan from 1947 – 1949 [25] . 
The PAT guidance is unique when compared with typical FDA guidance documents 
in that it is not instructive or limiting per se; rather, the guidance describes 
the principles and tools upon which the PAT framework is built, with the goal of 
“ highlighting opportunities and developing regulatory processes that encourage 
innovation. ” The FDA ’ s goal in developing the PAT guidance was to eliminate the 
specter of regulatory uncertainty which has been identifi ed as a major factor limiting 
innovation in pharmaceutical manufacturing. The guidance works with existing 
regulations and was designed to be consistent with the agency ’ s twenty - fi rst - century 
cGMPs initiative. Furthermore, the guidance emphasizes that the decision on the 
part of manufacturers to work with the agency to implement PAT is voluntary. Since 
the guidance is not prescriptive in nature, it neither describes “ how to do PAT ” nor 
identifi es any particular practice or technology as “ approved for PAT. ” 
4.1.5.1 Process Understanding 
The agency considers PAT to be a “ system for designing, analyzing, and controlling 
manufacturing through timely measurements of critical quality and performance 
attributes of raw and in - process materials and processes, with the goal of ensuring 
fi nal product quality. ” Based on this defi nition, it would be practical to consider PAT 
to be an expansion of PAC; PAT builds on the measurement and control aspects of 
PAC by incorporating additional emphasis on QbD and process understanding. 
According to the PAT guidance, a process is generally considered well understood 
when: 
1. All critical sources of variability are identifi ed and explained. 
2. Variability is managed by the process. 
3. Product quality attributes can be accurately and reliably predicted over the 
design space established for materials used, process parameters, manufacturing, 
environmental, and other conditions. 
Furthermore, according to the guidance, the ability to predict “ refl ects a high degree 
of process understanding. ” 

Possession of a predictive model (for product quality attributes) alone does 
not necessarily constitute process understanding, however. A relatively common 
example would be prediction of material or product performance characteristics 
using multivariate measurements, such as prediction of tablet dissolution rate using 
NIR spectroscopy. Multiple researchers have demonstrated that (in some cases) it 
is possible to predict drug release from tablets in vitro using nondestructive NIR 
spectra by generating a calibration model for dissolution rate. Without demonstrating 
at least mechanistic understanding of the physicochemical feature (correlated 
to dissolution rate) being detected by NIR, the calibration model would constitute 
nothing more than pattern recognition (Figure 12 ) [64] . While such a calibration 
may be useful, without greater insight as to the basis for correlation, it would not 
likely be a useful demonstration of process understanding. 
Design Space and Quality by Design The concept of a multidimensional space of 
acceptable operating conditions, or design space, is perhaps one of the most important 
aspects of the twenty - fi rst - century cGMPs which facilitates continuous improvement. 
In a PAT - enabled environment, the process design space must provide 
evidence of QbD [65] and should be the mathematical medium by which process 
understanding and real - time control decisions are communicated (Figure 13 ). 
The current ICH Q8 defi nition of design space, unfortunately, offers little guidance 
with regard to the aspects of a process design space which are required for 
implementation. As a result, a variety of interpretations of what constitutes a suitable 
process design space have recently surfaced among industry participants. One 
of the most popular misconceptions is that an effective design space for a process 
or unit operation can be determined by the common trajectory of PAT measurements 
(i.e., “ process signature ” ) related to product batches known to have acceptable 
quality (i.e., “ golden path ” ). While such data are useful for monitoring, they 
are nothing more than a modern version of “ 3 - batch ” process validation. Golden 
paths or process trajectories are not suffi cient for control since 1) the path itself is 
not necessarily predictive and 2) such controls would imply that a process is limited 
by its historical path in the space of process parameters. Originally, the term process 
signature was defi ned as a multivariate process measurement, that is, NIR spectrum, 
which contained features useful for describing the impact of the process on the 
chemical and physical aspects of the processed material [38] . 
FIGURE 12 Illustration of aspects of method understanding which must be in place to 
justify product performance measurements using indirect and/or nondestructive analyses. 
PAT EVOLUTION IN PHARMACEUTICAL MANUFACTURING 337

338 REGULATORY AND INDUSTRIAL PERSPECTIVES 
While it is perhaps too early to posit a conclusive standard for pharmaceutical 
process design space development, the following minimum criteria should be 
achieved for a process design space to be suitable for process control: 
• The process design space should be expressed in the form of a mathematical 
model which quantitatively links process capability , quality of input materials, 
and process operating parameters. 
• Relevant critical - to - quality product attributes should be considered by the 
design space model (e.g., content uniformity, bioavailability, stability). 
• Borrowing from a famous quote by Albert Einstein, the (design space) model 
should be as complex as necessary (for accurate prediction), but no less. 
• Product attributes that are superfl uous or are not known to be critical to quality 
should not be considered by the design space model (there should not be a 
penalty for monitoring such parameters, however). 
• In the same way that in vitro – in vivo correlation (IVIVC) is required to be 
granted a biowaiver for implementation of postapproval changes, the ability of 
the design space model to predict the quality of fi nished goods must be validated 
prior to implementation. 
• If the accuracy of the design space model cannot be established a priori with 
statistical signifi cance within portions of the parameter hyperspace, operation 
in such regimes should initiate supplementary quality assurance (inspection) 
activities until the design space model can be updated and revalidated. 
• If unacceptable product quality is observed during operation within a region 
of the design space expected to yield acceptable quality, the design space should 
be considered unsuitable for process control (due to drift or the appearance of 
new factors in the parameter space) until the missing factor(s) can be identifi ed 
and incorporated into the model and the model is revalidated. 
If such a model - based process design space includes a suffi cient portion of the 
factors affecting product quality variance, the process control space can be projected 
to defi ne the bounds of normal operation. Based on this defi nition, the control 
FIGURE 13 Interrelation between design space, PAT, and process control in a manufacturing 
system based on quality - by - design. ( Source : R. C. Lyon, Process monitoring of pilot - scale 
pharmaceutical blends by near - infrared chemical imaging and spectroscopy, Eastern Analytical 
Symposium (EAS), Somerset, NJ, 2006 .)

model algorithm for each operation in the manufacturing process would be generated 
from a subset of the control space spanned by the material qualities and processing 
parameters which impact that operation. Each unit operation control model 
seeks to adjust process parameters in a timely manner in response to changes in 
raw material (feedforward) or fi nished - product (feedback) quality. In other words, 
control the process and validate the controls. 
The mathematical linkage of the design space, process, and control models 
enables continuous optimization of product quality by seeking the optimal point 
within the control space. As the level of process understanding increases or as 
processing conditions evolve, factors might be added or removed from the design 
space and the process and control models updated. Furthermore, by considering 
other factors such as yield, effi ciency, or C/T as a function of the variables spanned 
by the process design space, the process might be co - optimized for quality and 
profi tability. 
It is likely that many pharmaceutical manufacturing operations are not understood 
in a way that product quality variance can be fully described in functional 
form (e.g., transfer functions); attaining such a level of manufacturing knowledge 
should be a goal for the industry. Using functional representations of process understanding 
as the basis set for a process design space, rather than historical performance, 
offers many operational advantages: 
• Effi cient Process Development While the current defi nition of design space 
does not preclude the incorporation of knowledge from other products and 
processes, model - based knowledge representation offers a more robust framework 
for incorporation of external or a priori information. Even though the 
level of quality expected by a particular combination of input and process 
parameters from another product is not likely to transfer to a new product or 
process (in absolute terms), the functional relationships which predict quality 
may be quite similar. Furthermore, model - based design space development 
enables direct incorporation of fi rst principles and mechanistic knowledge, 
which might signifi cantly reduce the complexity of experimental designs 
required for process development since signifi cant terms may be identifi ed in 
silico. 
• Quality by Design The incorporation of functional relationships between 
inputs, parameters, and product quality (or effi ciency), which inherently imply 
magnitude and directionality, enables the use of a process design space as a 
tool for multiobjective process optimization. Furthermore, the model - based 
representation of knowledge is compatible with concepts of risk management, 
enabling more fl exible operation since the risk associated with extrapolation 
could be predicted. 
• Control System Development Model - based design space development offers 
an ideal segue between process and control development. Quite literally, 
a model - based design space would provide the template for development 
of feedforward process control models. Moreover, development of a process 
design space using a model - based framework would facilitate control 
system validation and identifi cation of science - based, in - process, and release 
specifi cations. 
PAT EVOLUTION IN PHARMACEUTICAL MANUFACTURING 339

340 REGULATORY AND INDUSTRIAL PERSPECTIVES 
• Scaling and Technology Transfer Within the current system for process development, 
it is common to use designed experiments (i.e., DOE) where some 
input variables are product specifi c (e.g., excipient “ grade ” ) or process parameters 
are device dependent (e.g., chopper speed, damper angle). In a model - 
based paradigm, however, a process design space would ideally be generated 
using product - and device - independent units which have more basic physical 
meaning (e.g., modulus, viscosity, energy, or work). Designing and describing 
production processes in fundamental terms or, perhaps, standardized dimensionless 
units would facilitate scaling and transfer of design space and process 
control models to similar manufacturing processes that are based on the same 
physical operating principles. 
Academic research is currently underway to further develop the model - based 
design space concept. Working within the limits of the current system, though, producers 
who are able to demonstrate process understanding or are willing to invest 
in a PAT system to facilitate their development of process understanding can use 
the tools and provisions of the framework to pursue innovation and continuous 
improvement with more effi cient regulatory oversight (i.e., the ability to make 
changes without supplemental review). The PAT framework is described as consisting 
of two components: (1) a set of scientifi c principles and tools supporting innovation 
and (2) a strategy for regulatory implementation that will accommodate 
innovation. The following paragraphs will describe selected aspects of both components 
in detail. 
4.1.5.2 PAT Principles and Tools 
Central to the PAT framework is the acceptance that certain physical and mechanical 
attributes of pharmaceutical ingredients are not necessarily well understood and 
that even processes which have achieved signifi cant process understanding are 
subject to a fi nite level of stochastic variation. Thus, the core of the PAT guidance 
is allocated to describing the principles and tools, such as process analyzers and risk 
analysis, which producers can employ to augment process understanding and mitigate 
latent risks to product quality. 
PAT Tools The guidance describes four categories of PAT tools: 
• Multivariate tools for design, data acquisition, and analysis 
• Process analyzers 
• Process control tools 
• Continuous improvement and knowledge management tools 
Since each of the four categories draws upon methods and technology which are 
already established in other fi elds such as PAC, the discussion of each category 
within the guidance is focused on aspects which are unique or signifi cant to pharmaceutical 
manufacturing, such as process signature [2] . Furthermore, in keeping 
with the spirit of the framework as a catalyst for innovation, the agency made an 
effort to avoid mention of any particular tool or technology in the fi nal version of 

the PAT guidance. The PAT tools section of the guidance does, however, include 
cross - references to relevant portions of current regulations which should be considered 
by a manufacturer developing a PAT strategy or system. 
Standards for Pharmaceutical Applications of PAT During the early stages of 
developing the PAT framework, the agency was aware that the lack of international 
standards was a signifi cant impediment to regulatory coordination and implementation 
of PAT in the global pharmaceutical industry. In 2003, the FDA ’ s PAT team 
worked with ASTM International to form Technical Committee E55 on Pharmaceutical 
Application of Process Analytical Technology. The E55 committee addresses 
issues related to process control, design, and performance as well as quality acceptance/
assurance tests for the pharmaceutical manufacturing industry. Stakeholders 
in the committee include manufacturers of pharmaceuticals and pharmaceutical 
equipment, federal agencies, design professionals, professional societies, trade associations, 
fi nancial organizations, and academia ( www.ASTM.org ). 
As of mid - 2006, there were three subcommittees of E55: PAT system management, 
PAT system implementation and practice, and PAT terminology. The PAT 
team has been represented on E55 committees with a goal to ensure that standards 
developed are aligned with the PAT guidance and acceptable to the FDA. To date, 
one active standard has been published, while 16 additional standards have been 
proposed. The ASTM International provides another venue for international cooperation 
(consistent with the twenty - fi rst - century cGMPs initiative); the defi nitions 
of PAT (in the FDA guidance and ASTM E55) as well as other concepts are being 
incorporated into the ICH Q8 guidance. 
Real - Time Release ( RTR ) The PAT guidance defi nes RTR as “ the ability to 
evaluate and ensure the acceptable quality of in - process and/or fi nal product based 
on process data. ” Whereas fi nished products are typically released for marketing 
only after sampling, inspection (i.e., laboratory - based QC testing), and review, 
implementation of an RTR system enables release of fi nished products concurrent 
with the completion of manufacturing operations. Practically speaking, RTR is one 
of the most signifi cant, tangible benefi ts for producers who implement PAT, because 
it can facilitate dramatic reductions in process C/T. 
Real - time release is considered by the guidance to be comparable to alternative 
analytical procedures for fi nal product release and is defi ned within the guidance as 
an extension of parametric release. The defi ning characteristic of RTR is that it 
considers simultaneously the degree to which material attributes and process parameters 
are measured and controlled during manufacturing. It was not intended that 
RTR be implemented by simply installing a rapid measurement system at the end 
of a manufacturing process; such uses for PAT tools would be tantamount to inspection 
and would do nothing to improve quality management. 
The guidance does suggest, however, that it may be feasible to implement RTR 
without fi nished - product quality monitoring by using “ the combined process measurements 
and other test data gathered during the manufacturing process. ” Similar 
language is found in the USP general notices, where it is suggested that data derived 
from “ [validation studies and] in - process controls may provide greater assurance 
that a batch meets a particular monograph requirement than analytical data derived 
from an examination of fi nished units drawn from that batch. ” It would not be 
PAT EVOLUTION IN PHARMACEUTICAL MANUFACTURING 341

342 REGULATORY AND INDUSTRIAL PERSPECTIVES 
diffi cult to create a system more capable of detecting quality variation than current 
methods based on inspection. Recent statistical analyses [42] have demonstrated 
that, for determining batch quality, the traditional USP . 905 . method of content 
uniformity testing may indeed have little more statistical power than a coin toss 
until more than 5% of the product exceeds specifi cation limits (corresponds to 
within - batch C pk of approximately 0.65, only slightly worse than has been observed 
in a recent industry benchmarking study [28] ). 
On the other hand, deployment of an RTR system without fi nished - product 
monitoring would require the manufacturer to demonstrate a very high level of 
process understanding based on, for example, their development of a comprehensive 
design space and/or a well - validated process model. Even though it may be 
feasible to implement RTR without end - of - process monitoring, a well - designed PAT 
system will typically include some form of fi nal product quality monitoring as a 
means for mitigating latent risk and creating strategic redundancy in process controls 
and as an additional tool to bolster process understanding. 
4.1.5.3 Strategy for Implementation 
One of the FDA ’ s goals for the PAT guidance is to “ tailor the Agency ’ s usual regulatory 
scrutiny to meet the needs of PAT - based innovations that (1) improve the 
scientifi c basis for establishing regulatory specifi cations, (2) promote continuous 
improvement, and (3) improve manufacturing while maintaining or improving the 
current level of product quality. ” Recognizing that the achievement of this goal 
requires a unique interface between regulators and manufacturers seeking to implement 
PAT, a strategy for implementation based on the integrated systems approach 
was developed. An objective of the strategy for implementation is to facilitate clear, 
effective, and meaningful communication between the agency and industry, for 
example, in the form of meetings or informal communication. 
In practice, the strategy breaks with traditional industry – FDA modes of communication; 
whenever PAT is concerned, it is anticipated that regulators will communicate 
directly with the pharmaceutical scientists and engineers involved with 
development and operation of the PAT system rather than indirectly via a department 
of regulatory affairs. The components of the agency ’ s regulatory strategy 
include: 
• A PAT team approach for CMC review and cGMP inspections 
• Joint training and certifi cation of PAT review, inspection, and compliance 
staff 
• Scientifi c and technical support for the PAT review, inspection, and compliance 
staff 
• Recommendations provided within the PAT guidance 
PAT Team Approach FDA ’ s assembly of the PAT team was one of the most signifi 
cant incentives for the industry to pursue manufacturing innovation as described 
in the twenty - fi rst - century cGMPs initiative and the PAT guidance. The PAT team 
was put in place to ensure that industrial PAT applications were handled with expediency 
and accuracy by scientists familiar with the most up - to - date PAT methods. 

At one point the PAT team included more than 20 scientists, including investigators, 
compliance offi cers, reviewers, training coordinators, and a policy development 
team. More recently the agency has begun steps to “ sunset ” the PAT team, the duties 
of which will ultimately be handled by FDA staff trained in PAT systems. A comprehensive 
scientifi c training program was developed for the PAT team with guidance 
from the ACPS PAT subcommittee. Initial training began in January 2006, with 
plans for further training to be provided by faculty at Duquesne and Delaware 
Universities [47] . 
Research Data Provision In developing the PAT guidance, the FDA recognized 
that, even with the guidance in place, manufacturers seeking to evaluate the suitability 
or potential value of new technologies for process control may be hesitant, 
fi guring that such data will be subject to cGMP inspection, thereby increasing their 
liability with respect to regulatory actions. To allay these fears, the agency included 
a statement which applies to investigational deployment of new technologies [2] : 
Data collected using an experimental tool should be considered research data. If 
research is conducted in a production facility, it should be under the facility ’ s own 
quality system. . . . FDA does not intend to inspect research data collected on an existing 
product for the purpose of evaluating the suitability of an experimental process 
analyzer or other PAT tool. FDA ’ s routine inspection of a fi rm ’ s manufacturing process 
that incorporates a PAT tool for research purposes will be based on current regulatory 
standards (e.g., test results from currently approved or acceptable regulatory methods). 
Any FDA decision to inspect research data would be based on exceptional situations 
similar to those outlined in Compliance Policy Guide Sec. 130.300. Those data used to 
support validation or regulatory submissions will be subject to inspection in the usual 
manner. 
4.1.6 PAT IMPLEMENTATION PROCESS 
The PAT guidance identifi es three possible plans for companies seeking to implement 
PAT: 
• PAT can be implemented under the facility ’ s own quality system; cGMP inspections 
by the PAT team or PAT - certifi ed investigator can precede or follow PAT 
implementation. 
• A changes being effected (CBE), CBE in 30 days (CBE - 30), or prior approval 
(PAS) supplement can be submitted to the agency prior to implementation, 
and, if necessary, an inspection can be performed by a PAT team or PAT - 
certifi ed investigator before implementation. 
• A comparability protocol (CP) can be submitted to the agency outlining PAT 
research, validation, and implementation strategies and time lines. Following 
approval of this comparability protocol by the agency, one or a combination of 
the above regulatory pathways can be adopted for implementation. 
Refl ecting its nonprescriptive nature, the three implementation plans are essentially 
the only “ how to ” portions of the PAT guidance. This leaves industrial (and academic) 
scientists and engineers with the burden of determining how best to proceed 
PAT IMPLEMENTATION PROCESS 343

344 REGULATORY AND INDUSTRIAL PERSPECTIVES 
in the deployment of a PAT system. Despite the fact that some pioneering companies 
have been incorporating aspects of PAT in their operations since long before 
the start of the FDA ’ s twenty - fi rst - century cGMPs initiative, there continues to be 
signifi cant diversity in their approaches to implementation. While perhaps the ambiguity 
(in how best to proceed) has slowed the uptake of PAT to some degree, in the 
long run, the latitude is preferable since the optimal path of implementation will 
likely be unique for most facilities. 
With regard to drug manufacturers ’ implementation of PAT, a list of 10 questions 
has been presented which provides an initial checklist for companies seeking 
approval of their plans [10, 66] : 
1. Is this a PAT system? 
2. Does it have aspects of design, measurement, and manufacturing control? 
3. Are PAT principles and tools used? 
4. Which tools specifi cally are used for manufacturing control? 
5. How are continuous improvement and knowledge management 
performed? 
6. What risk - based approach has the company taken — assessment, prevention, 
and management? 
7. How are the PAT systems integrated? 
8. What kind of RTR is being proposed or used? 
9. What regulatory process is being considered? 
a. Can the companies ’ quality systems manage the PAT change? 
b. Are the submission proposals appropriate and justifi ed? 
10. What are the critical aspects that will be evaluated during site visits/ 
inspections? 
Drawing from aspects of the DMAIC model, as well as the risk - based orientation 
and quality systems approach espoused by the FDA ’ s twenty - fi rst - century cGMPs 
initiative, the Duquesne University Center for Pharmaceutical Technology (DCPT) 
has proposed a six - phase, iterative cycle for process improvement based on PAT 
(Figure 14 ). While there are certainly many acceptable variants of this strategy, some 
of which have begun to appear in conferences and the industrial literature, any successful 
PAT deployment, large or small, will most likely include some combination 
of these elements. In addition, each project phase will necessarily include one or 
more modules of training. Finally, while the project phases are presented as being 
discrete, most of the phases will overlap to some degree. In particular, consideration 
of the objectives for control, release strategies, and plans for continuous improvement 
should begin, along with management buy - in, early in the cycle. 
4.1.6.1 Preparation 
The preparation phase is arguably the most critical step in the path toward PAT 
implementation. Process analytical technology projects are inherently multidisciplinary, 
requiring acceptance and buy - in from corporate divisions which sometimes 

FIGURE 14 PAT implementation cycle with examples of associated activities for each 
phase. 
operate with rather divergent goals and procedures. Most importantly, those who 
are seeking to initiate a PAT project will need to obtain management buy - in at a 
level high enough in the corporate structure to ensure suffi cient resources will be 
available and that the company will be committed to positive change. During the 
preparation phase, a PAT team having a diverse background and critical skills 
should be assembled, and formal planning of the project should begin, including 
selection of the product and process to address. Ideally, dialogue with the FDA PAT 
team should begin early in the preparation phase. 
4.1.6.2 Assessment 
The PAT guidance clearly states that industrial implementations should be risk 
based. Soon after the PAT team and objective have been identifi ed, the project 
should commence with a formal risk assessment. The risk assessment should be 
focused on identifying and characterizing the failure modes which present risks to 
product quality; the outcome of the risk assessment will provide a means prioritizing 
the allocation of PAT resources and a baseline for review of the effect of PAT in 
mitigating risks to quality. 
4.1.6.3 Analyze 
The “ analyze ” phase of the project consists of the activities which are typically 
associated with PAC, including identifi cation and assessment of potential sensor 
technologies, method development, qualifi cation, and validation. In addition, 
designed experiments (DOE) or data - mining exercises may be performed to 
PAT IMPLEMENTATION PROCESS 345

346 REGULATORY AND INDUSTRIAL PERSPECTIVES 
generate process understanding or to support PAT goals. Plans for the IT infrastructure, 
sampling protocols, and development of controls should also be considered. 
4.1.6.4 Control 
The implementation of controls begins as each new analytical method or technology 
is deployed. Controls may be as simple as automated termination of a unit operation 
upon reaching an endpoint. With greater process understanding, more complex 
controls can be deployed, including feedback (e.g., control of punch force during 
tablet compaction, control of temperature or airfl ow during fl uid bed processing) 
or feedforward controls (e.g., adjustment of process parameters based on incoming 
raw - material quality). The development and implementation of controls should also 
consider operating procedures for adverse situation management and should initiate 
a reassessment of risk to determine the suitability of controls. 
4.1.6.5 Release Philosophy 
For PAT projects including implementation of RTR or some modifi cation of a preexisting 
release mechanism for an approved process, additional method development 
and validation procedures will be required. The real - time release decision will 
typically be determined by a process model, which can be a mathematical equation 
or algorithm within the control system; furthermore, the IT system must accurately 
convey the release decision and supporting data to downstream operations (i.e., 
warehouse, logistics), upstream operations (i.e., production scheduling, accounting), 
or the facility information repository. The interwoven IT and scientifi c components 
require an integrated systems approach to development, validation, deployment, 
and operation. Finally, implementation of PAT systems enables redefi nition of 
product quality acceptance criteria for release; the task of identifying robust release 
criteria suitable for large sample sizes, for example, continues to merit examination 
[43] . 
4.1.6.6 Optimization 
The optimization phase of the project provides an opportunity to assess the performance 
of the PAT system relative to the goals of the project as well as the level of 
latent risk in the system. Ideally, with the PAT s2ystem in place, the level of process 
understanding will be improving as more data are collected for every batch. The 
added insight into the operation may yield new opportunities for improving quality 
or effi ciency or for solving similar problems with another product. The key to success 
in the optimization stage is realizing that it is only the beginning of continuous 
improvement. 
4.1.7 PERSPECTIVES ON THE IMPACT OF PAT 
PAT and the twenty - fi rst - century cGMPs initiative have clearly made an impact 
within the pharmaceutical and associated industries. Signifi cant sums of capital 
are now fl owing in new directions to meet the challenges and opportunities pre

sented by the changes. Some people within the industry, however, question 
whether there will be much of a long - term impact, citing the litany of new eras 
in the industry (and their careers) that turned out to be more of the same. With 
just a bit of observation, though, it is not hard to see that it really is different 
this time. 
The modern pharmaceutical manufacturing industry fi nds itself in a diffi cult situation 
that perhaps few anticipated just 10 or 15 years ago. The rate of new blockbuster 
drug approvals has continued to wane, while new drug therapies become 
inexorably more expensive to discover and develop. Despite the fact that the market 
for drug sales has never been larger, drug company profi t margins are shrinking 
while consumers, feeling that pharmaceutical company profi ts are unjust, have 
reached new lows in their opinion of the industry. A recent survey by the Kaiser 
Family Foundation placed pharmaceutical companies just above oil and tobacco 
companies, and right below health management organizations (HMOs), in terms of 
public opinion [67] . Entities of signifi cant magnitude in both the public and private 
sector are increasingly applying pressure to capture an even greater portion of the 
industry ’ s compensation. Indeed, there is no shortage of industrial and fi nancial 
publications which have chronicled the pharmaceutical industry ’ s troubles [34, 45, 
68, 69] . 
The pharmaceutical industry is fortunate, perhaps, to follow (rather than lead) 
most other major industries in adopting truly automated controls, process analytics, 
quality management, and lean manufacturing. The performance of pharmaceutical 
companies relative to the benchmarks for world - class manufacturers provides a 
roadmap for improvement. If the pharmaceutical industry, as a whole, were able to 
at least approach the benchmarks for world - class manufacturing performance (by 
implementing PAT), the savings returned to consumers and shareholders would be 
immense (Figure 15 ). Finally, the returns on investment in PAT are not limited to 
major producers. Estimates based on recent benchmarks suggest that, by successfully 
transforming operations through the deployment of PAT and lean, a typical 
small or mid - sized pharmaceutical manufacturer could improve operating margins 
by up to 600 basis points [44] . 
Forces which are out of the industry ’ s control are providing more reasons 
than ever before to seek effi ciency in pharmaceutical manufacturing, and the 
FDA is doing its part to clear the way. While the pharmaceutical industry has 
likely been unjustly cast as a culprit behind America ’ s fi scal crisis in health care, the 
industry has ample opportunity to change for the benefi t of patients as well as 
investors. 
ACKNOWLEDGMENTS 
The author would like to thank the following reviewers for their input, which was 
essential to the quality of this manuscript: James K. Drennen, III, Ph.D., Director, 
Duquesne University Center for Pharmaceutical Technology, Senior Consultant, 
Strategic Process Control Technologies; Robbe C. Lyon, Ph.D., Deputy Director, 
Division of Product Quality Research FDA/CDER; D. Christopher Watts, Ph.D., 
Team Leader, Standards & Technology, FDA/CDER/OPS; and Tom Knight, Founder 
& Chief Strategy Offi cer, Invistics Corp. 
ACKNOWLEDGMENTS 347

348 REGULATORY AND INDUSTRIAL PERSPECTIVES 
FIGURE 15 Potential fi nancial returns from deployment of PAT and lean. The curves are 
calculated based on the aggregate COGS and inventories reported in the 2005 annual reports 
of the top 16 branded and generic pharmaceutical manufacturers (according to market capitalization). 
It is important to keep in mind that working capital savings are a one - time - only 
benefi t, while cost of quality and inventory fi nancing and overhead savings represent on - going 
returns on investment. Furthermore, while the curves may overestimate savings because of 
innacuracies in benchmark data or the limits on the opportunities for PAT implementation, 
they do not account for numerous other potential pathways for returns from PAT such as 
capacity increase, labor productivity enhancement, reduction of QC expense, or decreased 
time to market. 
Cost of Quality Saving ($ Billions) 
10
9
8
7
6
5
4
3
2
1
0
0.5 0.6 0.7 0.8 0.9 1 
Process Capability (Cpk) 
2 4 6 8 10 12 14 
Total Inventory Turn Rate (Turns/Year) 
35 
30 
25 
20 
15 
10
5
0 
Supply Chain Optimization Savings ($ Billions) 
Working Capital Savings 
Financing & Overhead Savings

REFERENCES 
1. U.S. Department of Health and Human Services ( 2004 ), Pharmaceutical CGMPs for 
the 21st century — A risk - based approach, Final report, Food and Drug Administration, 
Rockville, MD. 
2. U.S. Department of Health and Human Services ( 2004 ), Guidance for industry: PAT — A 
framework for innovative pharmaceutical development, manufacturing, and quality 
assurance, Food and Drug Administration, Rockville, MD. 
3. Clevett , K. J. ( 1986 ), Process Analyzer Technology , Wiley , New York . 
4. Illman , D. L. ( 1986 ), CPAC: An industry — university cooperative research center for 
process analytical chemistry , TrAC Trends Anal. Chem. , 5 , 164 . 
5. Callis , J. B. , Illman , D. L. , and Kowalski , B. R. ( 1987 ), Process analytical chemistry , Anal. 
Chem. , 59 , 624A . 
6. Balboni , M. L. ( 2003 ), Process analytical technology: concepts and principles, Pharm. 
Technol . 
7. Koch , M. V. ( 2006 ), Optimizing the impact of developments in micro - instrumentation on 
process analytical technology: a consortium approach , Anal. Bioanal. Chem. , 384 , 1049 . 
8. K u ppers , S. , and Haider , M. ( 2003 ), Process analytical chemistry — future trends in industry 
, Anal. Bioanal. Chem. , 376 , 313 . 
9. K u ppers , S. , and Haider , M. ( 2006 ), Process analytical chemistry , Anal. Bioanal. Chem. , 
384 , 1034 . 
10. Hinz , D. C. ( 2006 ), Process analytical technologies in the pharmaceutical industry: the 
FDA ’ s PAT initiative , Anal. Bioanal. Chem. , 384 , 1036 . 
11. Beebe , K. R. , et al ., ( 1993 ), Process analytical chemistry , Anal. Chem. , 65 , 199R . 
12. Blaser , W. W. , et al., (1995), Process analytical chemistry , Anal. Chem. , 67 , 47R . 
13. Workman , J. J. , et al ., ( 1999 ), Process analytical chemistry , Anal. Chem. , 71 , 121R . 
14. Workman , J. J. , et al ., ( 2001 ), Process analytical chemistry , Anal. Chem. , 73 , 2705 . 
15. Workman , J. J. , Koch , M. V. , and Veltkamp , D. J. ( 2003 ), Process analytical chemistry , Anal. 
Chem. , 75 , 2859 . 
16. Workman , J. J. , Koch , M. V. , and Veltkamp , D. J. ( 2005 ), Process analytical chemistry , Anal. 
Chem. , 77 , 3789 . 
17. Gluckman , P. , Roome , D. R. , Deming , W. E. , and Delavigne , K. ( 1993 ), Everyday Heroes 
of the Quality Movement: From Taylor to Deming, the Journey to Higher Productivity , 
2nd ed., Dorset house , New York . 
18. Shewhart , W. A. ( 1980 ), Economic Control of Quality of Manufactured Product , ASQC 
Quality Press , Milwaukee, WI . 
19. Shewhart , W. A. ( 1986 ), in Deming W. E. ed., Statistical Method from the Viewpoint of 
Quality Control , Dover , New York . 
20. Scherkenbach , W. W. ( 1991 ), The Deming Route to Quality and Productivity: Road Maps 
and Roadblocks , ASQC Quality Press , Milwaukee, WI . 
21. Juran , J. M. ( 2003 ), Architect of Quality: The Autobiography of Dr. Joseph M. Juran , 
McGraw - Hill , New York . 
22. Bhote , K. R. ( 2002 ), What is Six Sigma? The Ultimate Six Sigma . Vol 1. New York : 
AMACOM ; 9 – 14 . 
23. Galvin , R.W. ( 2002 ), Forward . The Ultimate Six Sigma . Vol 1. New York : AMACOM ; 
xxi – xxii . 
24. Marash , S. A. , Berman , P. , and Flynn , M. ( 2004 ), Fusion Management: Harnessing the 
Power of Six Sigma, Lean, ISO 9001:2000, Malcom Baldridge, TQM and Other Quality 
Breakthroughs of the Past Century , QSU Pub., Fairfax, VA . 
REFERENCES 349

350 REGULATORY AND INDUSTRIAL PERSPECTIVES 
25. Ealey , L.A. ( 1988 ), Quality by design. Taguchi Methods and U.S. industry . Dearborn, MI : 
ASI Press . 
26. Taguchi , G. , Chowdhury , S. , and Wu , Y. ( 2005 ), Taguchi ’ s quality engineering handbook . 
Hoboken, NJ : John Wiley & Sons . 
27. e - Handbook of Statistical Methods. NIST/SEMATECH . Available at: http://www.itl.nist. 
gov/div898/handbook/ . 
28. Macher , J. , and Nickerson , J. ( 2006 ), Pharmaceutical manufacturing research project: Final 
benchmarking report, Georgetown University, McDonough School of Business and 
Washington University Olin School of Business, Washington DC. 
29. Levinson, W. A. (2002), Henry Ford ’ s Lean Vision , Productivity Press , New York . 
30. Leiper , K. ( 2006 ), paper presented at the The Heidelberg PAT Conference, Heidelberg, 
Germany. 
31. Lewis , N. A. ( 2006 ), A tracking tool for lean solid - dose manufacturing , Pharm. Technol. 
30 , 94 – 108 . 
32. Roumeliotis , G. ( 2006 ), Lean proves mean in drug manufacturing , available: in - 
PharmaTechnologist.com . 
33. Gerecke , G. , and Knight , T. , Improving performance and reducing cycle time using fl ow 
path management: A case study , Pharm. Eng. 21 . 
34. Abboud , L. , and Hensley , S. ( 2003 ), Factory shift: New prescription for drug makers: 
Update the plants, Wall Street J. , Sept. 3, p. 1. 
35. Cooley , R. E. , and Egan , J. C. ( 2004 ), The impact of process analytical technology (PAT) 
on pharmaceutical manufacturing , Am. Pharm. Rev. , 7 , 62 – 68 . 
36. Cogdill, R. P. (2006), in Brittain, H. G. , Ed., Spectroscopy of Pharmaceutical Solids , Taylor 
& Francis Group , New York , pp. 313 – 412 . 
37. Cogdill , R. P. , et al ., ( 2005 ), Process analytical technology case study, Part II: Development 
and validation of quantitative for tablet API content and hardness , AAPS Pharm. Sci. 
Tech. , 6 , Article 38. 
38. Cogdill , R. P. , et al ., ( 2005 ), Process analytical technology case study, Part I: Feasibility 
studies for quantitative NIR method development , AAPS Pharm. Sci. Tech. , 6 , Article 
37. 
39. Cogdill , R. P. , Anderson , C. A. , and Drennen , J. K. ( 2005 ), Process analytical technology 
case study, Part III: Calibration monitoring and transfer , AAPS Pharm. Sci. Tech. , 6 , 
Article 39. 
40. Lorber , A. ( 1986 ), Error propagation and fi gures of merit for quantifi cation by solving 
matrix equations , Anal. Chem. , 58 , 1167 . 
41. Braga , J. W. B. , and Poppi , R. J. ( 2004 ), Figures of merit for the determination of the 
polymorphic purity of carbamazepine by infrared spectroscopy and multivariate calibration 
, J. Pharm. Sci. , 93 , 2124 . 
42. Lunney , P. , and Drennen , J. K. I. ( 2005 ), A prevention based strategy for quality control 
using PAT , NIR News , 16 , 7 . 
43. Sandell , D. , Diener , M. , Vukovinsky , K. , Hofer , J. , and Pazdan , J. ( 2006 ), Development of 
a content uniformity test suitable for large sample sizes , Drug. Info. J. , 40 , 337 . 
44. Cogdill , R. P. , Knight , T. P. , Anderson , C. A. , and Drennen , J. K. ( 2007 ), The fi nancial 
returns on investments in process analytical technologies and lean manufacturing: Benchmarks 
and case study , J. Pharm. Innov. , 2(1–2) , 38 – 50 . 
45. du Pre Gauntt , J. ( 2005 ), Quality manufacturing: A blockbuster opportunity for pharmaceuticals, 
Economist Intelligence Unit. 
46. Steyer , R. ( 2005 ), Merck names clark CEO, available: TheStreet.com , accessed May 5, 
2005. 

47. Watts , D. C. ( 2006 ), PAT — An FDA paper presented at the The Heidelberg PAT Conference, 
Heidelberg, Germany. 
48. Hussain , A. S. ( 2001 ), Emerging science issues in pharmaceutical manufacturing: Process 
analytical technologies , paper presented at the Science Board Presentations to FDA, 
Rockville, MD. 
49. Chisholm , R. S. ( 2001 ), TQMS, statistically based in-process control with real time quality 
assurance, the AstraZeneca total quality management strategy , paper presented at the 
Science Board Presentations to FDA, Rockville, MD. 
50. Raju , G. K. ( 2001 ), Pharmaceutical manufacturing: New technology opportunities , paper 
presented at the Science Board Presentations to FDA, Rockville, MD. 
51. Scherzer , R. H. ( 2002 ), Quality by design: A challange to the pharma industry , paper 
presented at the Science Board Presentations to FDA, Rockville, MD. 
52. U.S. Department of Health and Human Services (2004), Risk-based method for prioritizing 
cGMP inspections of pharmaceutical manufacturing sites — A pilot risk ranking 
model, Food and Drug Administration, Rockville, MD. 
53. U.S. Department of Health and Human Services ( 2006 ), SMG 2020 — FDA quality 
system framework for internal activities, Food and Drug Administration, Rockville, 
MD. 
54. U.S. Department of Health and Human Services ( 2006 ), Guidance for industry: Quality 
systems approach to pharmaceutical cGMP regulations, Food and Drug Administration, 
Rockville, MD. 
55. U.S. Department of Health and Human Services ( 2004 ), Guidance for industry: Sterile 
drug products produced by aseptic processing — Current good manufacturing practice, 
Food and Drug Administration, Rockville, MD. 
56. U.S. Department of Health and Human Services ( 2003 ), Guidance for industry: Comparability 
protocols — Chemistry, manufacturing, and controls information, draft guidance, 
Food and Drug Administration, Rockville, MD. 
57. U.S. Food and Drug Administration (FDA) ( 2004 , Mar.), Sec. 490.100 Process validation 
requirements for drug products and active Pharmaceutical ingredients subject to pre - 
market approval (CPG 7132c.08), FDA, Rockville, MD. 
58. International Conference on Harmonization (ICH) ( 2004 ), Q8: Pharmaceutical development, 
International Conference on Harmonization of Technical Requirements for Registration 
of Pharmaceuticals for Human Use, ICH, Geneva. 
59. International Conference on Harmonization (ICH) ( 2004 ), M4: Organisation of the 
common technical document for the registration of pharmaceuticals for human use, 
International Conference on Harmonization of Technical Requirements for Registration 
of Pharmaceuticals for Human Use, ICH, Geneva. 
60. U.S. Department of Health and Human Services ( 2006 ), Guidance for industry: Providing 
regulatory submissions in electronic format — Human pharmaceutical product applications 
and related submissions using the eCTD specifi cations, Food and Drug Administration, 
Rockville, MD. 
61. Yu , L. X. ( 2006 ), Implementation of quality - by - design: Question - based review , paper 
presented at the Drug Information Association 42nd Annual Meeting, Philadelphia, 
PA. 
62. International Conference on Harmonization (ICH) ( 2005 ), Q9: Quality risk management, 
International Conference on Harmonization of Technical Requirements for Registration 
of Pharmaceuticals for Human Use, ICH, Geneva. 
63. International Conference on Harmonization (ICH) ( 2005 ), Q10: Pharmaceutical quality 
systems, fi nal concept paper, International Conference on Harmonization of Technical 
Requirements for Registration of Pharmaceuticals for Human Use, ICH, Geneva. 
REFERENCES 351

352 REGULATORY AND INDUSTRIAL PERSPECTIVES 
64. Hussain , A. S. ( 2006 ), Quality by design and bioequivalence/bioavailability assessment , 
paper presented at the The Heidelberg PAT Conference 2006, Heidelberg, Germany. 
65. Lyon , R. C. , and Hammond , S. ( 2006 ), Process monitoring of pilot - scale pharmaceutical 
blends by near - infrared chemical imaging and spectroscopy , paper presented at the 
Eastern Analytical Symposium, Somerset, NJ. 
66. D ’ Sa , A. ( 2005 ), Process analytical technology (PAT): regulatory process, review and 
inspection , paper presented at the 19th International Forum on Process Analytical Technology
—IFPAC 2005, Arlington, VA. 
67. Views on prescription drugs and the pharmaceutical industry, The Kaiser Family Foundation, 
2005 . 
68. Arlington , S. , Barnett , S. , Hughes , S. , Palo , J. , and Shu , E. ( 2002 ), Pharma 2010: The threshold 
of innovation, IBM Business Consulting Services, Somers, NY. 
69. Arlington , S. , et al ., ( 2005 ), The metamorphosis of manufacturing, IBM Business Consulting 
Services, Somers, NY. 

353 
4.2 
PROCESS ANALYTICAL TECHNOLOGY 
Michel Ulmschneider and Yves Roggo 
F. Hoffmann-La Roche Ltd, Basel, Switzerland 
Contents 
4.2.1 Basic Concepts and Impact 
4.2.1.1 Defi nition 
4.2.1.2 What Motivated PAT? 
4.2.1.3 Root - Cause Analysis and Process Control 
4.2.1.4 When to Introduce PAT 
4.2.1.5 PAT Enhances Process Understanding 
4.2.1.6 Changing Current Practice Using PAT 
4.2.1.7 Promoting Physical Pharmacy and Pharmaceutical Sciences 
4.2.1.8 Data Mining 
4.2.1.9 Data Warehousing 
4.2.1.10 Data - Mining Methods for Pharmaceutical Processes 
4.2.1.11 Data - Mining Practice 
4.2.1.12 Comments about Data Mining 
4.2.1.13 PAT Methods 
4.2.1.14 Conclusion 
4.2.2 Vibrational Spectroscopy 
4.2.2.1 Introduction 
4.2.2.2 IR Spectroscopy Theory 
4.2.2.3 Mechanical Model of IR Vibration 
4.2.2.4 Quantum Mechanical Model 
4.2.2.5 Anharmonicity 
4.2.2.6 Structure Elucidation Using MIRS 
4.2.2.7 Extending Use of MIRS 
4.2.2.8 Raman Spectroscopy 
4.2.2.9 Introducing NIRS 
4.2.2.10 Benefi ts of NIRS 
4.2.2.11 Introducing MIR/NIR Chemical Imaging 
4.2.2.12 Design of MIR Instruments 
4.2.2.13 Conclusion 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

354 PROCESS ANALYTICAL TECHNOLOGY 
4.2.3 Chemometrics 
4.2.3.1 Introduction 
4.2.3.2 From Univariate to Multivariate Regression 
4.2.3.3 Sample Quality and Data Error 
4.2.3.4 Mathematical Preprocessing of Spectroscopic Data 
4.2.3.5 Preprocessing NIR Data 
4.2.3.6 Mathematical Pretreatment and Transformation 
4.2.3.7 Principal - Component Analysis 
4.2.3.8 PCA Practice for NIRS 
4.2.3.9 Pattern Recognition 
4.2.3.10 SIMCA Classifi cation 
4.2.3.11 Regression 
4.2.3.12 Multiple Linear Regression 
4.2.3.13 PCR and PLS Regression 
4.2.3.14 Regression Practice in NIRS 
4.2.3.15 Some Pitfalls 
4.2.3.16 Example Analytical Applications of NIRS 
4.2.3.17 Conclusion 
Bibliography 
4.2.1 BASIC CONCEPTS AND IMPACT 
4.2.1.1 Defi nition 
Process analytical technology (PAT) is one of the objectives contained in the Initiative 
for Pharmaceutical cGMPs for the 21st Century published by the Food and Drug 
Administration (FDA). In a few words and according to the FDA ’ s guideline, PAT 
can be defi ned as a system for designing, analyzing, and controlling pharmaceutical 
manufacturing through the measurement of critical quality and performance parameters. 
The measurements performed on raw and in - process materials or process 
parameters are intended to enhance fi nal product quality. 
Process analytical technology encourages technological innovation, specifi cally 
the adoption of new analytical techniques by the pharmaceutical industry designed 
to improve the understanding and control of manufacturing processes. Both the 
FDA and industry experts expect benefi ts over conventional manufacturing practices: 
higher fi nal product quality, increased production effi ciency, decreased operating 
costs, better process capacity, and fewer rejects. Correspondingly, fundamental 
changes are also expected within the regulatory framework. The future of pharmaceutical 
production will require innovative technological approaches and more 
science - based processes. PAT will boost collaboration between research and development 
(R & D) and manufacturing departments inside companies and increase 
overall effi ciency. Approvals and inspections will increasingly focus on scientifi c and 
engineering principles. As a result, regulators will set higher expectations for new 
products from the outset. 
4.2.1.2 What Motivated PAT ? 
Preliminary discussions of PAT concepts between the FDA and certain pharmaceutical 
companies already active in this fi eld date back to the late 1990s. In September 

BASIC CONCEPTS AND IMPACT 355 
2004 the FDA released a document for the industry entitled “ PAT Guidance for 
Industry: A Framework for Innovative Pharmaceutical Development , Manufacturing, 
and Quality Assurance. ” PAT is clearly anchored in FDA corporate culture. 
Pharmaceutical companies are facing growing demands for increased productivity 
and reduced manufacturing costs. They also have to meet the evolving need for 
higher quality standards and higher drug expectations. At the same time the quest 
for new active substances remains a signifi cant issue. Reducing the attrition rate 
among selected candidates will bring more new medicines onto the market. In terms 
of drug marketing, the goal is to improve formulations so as to offer patients innovative 
and more effi cient solutions, and thus achieve commercial success or breakthrough. 
By prioritizing science - based design and introducing novel or improved 
process techniques, backed by the generation of increased critical data throughout 
a drug ’ s life cycle, the aim of the emerging PAT strategy is to direct the drug industry 
toward these essential goals. 
Because they have been used for many years, a variety of existing experimental 
methods and manufacturing processes are considered well established. They are 
trusted to generate few errors and make only modest contributions to process variation. 
Due to their longevity, they continue to be widely used in recent drug developments. 
Improvements in existing technologies are always possible and are constantly 
being made. However, this makes it diffi cult to consider or identify potential technological 
alternatives without critical review or a voluntary management decision 
to replace well - established techniques. The FDA noticed that nearly all recent drug 
developments lacked the possibility of enhancing and extending process capabilities 
toward newer or alternative technologies. More specifi cally, the FDA wanted to 
encourage drug manufacturers to achieve more innovation and improve risk management 
when releasing new medicines on the market. 
4.2.1.3 Root - Cause Analysis and Process Control 
When a quality problem arises in present - day production, it is increasingly diffi cult 
to identify the root cause. Thorough understanding of process and product performance 
often comes up against knowledge barriers, whether due to the escalating 
documentation burden, lack of time, or loss of expertise. The goal of PAT is to 
enhance process control and understanding so that procedures can be performed 
differently and more effi ciently. The PAT initiative facilitates and encourages the 
introduction of innovative approaches. It makes it possible to consider shifting from 
validation to continuous verifi cation. The next step is effective real - time release with 
continuous processing as an alternative to the conventional batch - after - batch production 
scheme. 
4.2.1.4 When to Introduce PAT 
Building quality into a pharmaceutical product has to be considered from the very 
beginning of the product ’ s life. Essential preconditions are the equal involvement 
of — and seamless communication between — R & D and manufacturing. One purpose 
of PAT is to provide a motivating framework to bring quality into a product from 
the outset. It is thus essential for it to be involved in the R & D phase. If product 
quality requirements are understood and implemented from the beginning, 

356 PROCESS ANALYTICAL TECHNOLOGY 
root - cause analysis of quality or process failure after scale - up to commercial manufacturing 
will be much easier. This is why PAT could play an even more important 
role in the design and analysis of manufacturing processes, enabling performance 
control to be based on timely measurement of well - described critical processing 
data. 
Data processing needs should also be considered in the context of overall process 
analysis strategy to meet emerging requirements for the speed and volume of data 
collection. Real - time analysis supported by knowledge management requires collecting 
and gathering all production batch information, for example, by data warehousing. 
Thus, a PAT data management strategy based on online process analysis 
or data mining can be set up long before generating large sets of measurement data. 
Historical data analysis should aim to cover method development, method validation, 
and ongoing performance monitoring, as well as routine results for a given 
manufacturing process. 
4.2.1.5 PAT Enhances Process Understanding 
Process analytical technology can greatly enhance process understanding. In fact, 
introducing PAT can act as a key driver to better process knowledge. The expected 
steps in implementing the PAT approach are the collection of online, in - line, and 
at - line data (Figure 1 ) on critical attributes, extraction of information, and analysis 
of process status data, ending with closure of the loop by dynamic process control. 
Innovating during development, applying cutting - edge techniques, and process 
modeling whenever possible, all contribute to a more fundamental exploration of 
the science behind the process. It is important to realize that PAT is not only the 
straightforward introduction of additional analytical techniques into a process but 
also the development of methods to predict future behavior according to given settings 
of the critical parameters. That means being able to predict fi nal product 
quality. For example, while implementing the process, it is important to explore all 
sources of component variation as well as their effect on the fi nished product in 
order to select which quality parameters (i.e., attributes) have to be measured for 
optimal and realistic process control. 
Science, engineering, and control technologies can provide a very high level of 
process understanding and control capability. A process is well understood when all 
FIGURE 1 In - line, online, and at - line process measurements. 
Spectrometer Spectrometer Spectrometer 
Reactor 
Inline Online Atline
Sampling 
Process flow

BASIC CONCEPTS AND IMPACT 357 
critical sources of variability are identifi ed and explained. The process should be 
robust enough to manage this variability. It is also expected that critical quality 
attributes can be accurately and reliably predicted in an adequate design space 
when other unexpected variables are encountered (e.g., change of raw material 
supplier). 
4.2.1.6 Changing Current Practice Using PAT 
An approach integrating R & D and manufacturing will enhance process understanding 
and make acceptable risk management possible. By establishing transferable 
process models, it will be possible to develop and implement adequate measurement 
technologies that match process needs rather than vice versa. More effi cient and 
cost - effective technology transfers will facilitate process knowledge, continuous 
process verifi cation, and compliance, thereby enhancing fi nal product quality. Better 
process understanding makes it possible to operate by continuous process verifi cation 
instead of three - batch validation. Measurement technique selection and integration 
occur very early. Accumulated pertinent knowledge is readily available 
through data - mining techniques to confi rm or control processing. A series of dynamic 
closed control/compliance loops at the process steps identifi ed as critical will increase 
confi dence in fi nal product quality. In addition knowledge accumulated over time 
will provide a basis for immediate and rapid intervention in the event of deviation 
or failure. 
A typical illustration of a PAT approach to quality improvement is the use of 
near - infrared spectroscopy (NIRS) to qualify excipients and active principles just 
before they enter the production process, for example, in dispensing. As discussed 
in the next part, near - infrared (NIR) spectra are informative about product structure 
and overall quality. Because with substances such as excipients the quality 
range was investigated at some time in the past and fi xed into a calibration, NIR 
measurement can provide simultaneous nondestructive confi rmation of the predominant 
physical and chemical parameters. This is an effective method of reducing 
uncertainties about possible causes of failure or poor quality during production. 
Each time a given excipient fails its quality requirements at the moment of use, 
immediate action can be taken. Control is possible before the risk of failure is 
increased. Such an approach is complementary to container - wise identifi cation of 
materials on delivery to a warehouse. 
4.2.1.7 Promoting Physical Pharmacy and Pharmaceutical Sciences 
Process analytical technology supposes a more science - based approach to pharmaceutical 
processes. As a matter of fact, it underlines the observed weakness in formal 
knowledge of the physical phenomena behind pharmaceutical processes. The physics 
is less well understood than the chemistry. Conventional physics has moved increasingly 
into the fi eld of activity of engineers and technologists. Formal approaches are 
lacking. As a consequence, much highly valuable knowledge of physical phenomena 
is dispersed across various disciplines. Expertise in physics is often purely technological 
rather than being formalized and integrated into a specifi c discipline. 
Just as the boundaries of physics and chemistry once merged to create 
physical chemistry, there is an opportunity now for assembling complementary 

358 PROCESS ANALYTICAL TECHNOLOGY 
scientifi c knowledge from various disciplines. It is a major challenge to improve 
understanding through in - depth investigation of the physical phenomena behind 
pharmaceutical processes. This objective motivates the enforcement of physical 
pharmacy to improve process understanding through a grounding in theoretical 
physics. 
One major issue is the science and technology of solid particles and powders: 
characterization, size and shape analysis, processing understanding, and so forth. 
Others include particle formation and fl uid – particle separation, mixture stability, 
and understanding and simulating the dynamics of powder mixtures. For example, 
the compaction state of powders and mixtures may change rapidly depending on 
storage time and conditions. Time to use is not always under control and unexpected 
changes may occur. Stirring a mixture of two free - fl owing powders of different size 
may result in segregation rather than improved mixture quality. The fl ow properties 
of powders depend not only on intrinsic characteristics of the different materials, 
such as particle size distribution, particle shape, and surface properties, but also on 
external conditions, such as humidity or compaction status. Further areas of interest 
include liquid drops, emulsions and colloids, bubbles, and polymers, as well as 
surface properties, surface analysis, interfacial and electrostatic phenomena, surface 
reactivity, wet chemistry properties, and solubility. 
4.2.1.8 Data Mining 
Complex processes generate large volumes of data over time. As ever - increasing 
volumes are collected and stored, the gap between buried information and usable 
accessible knowledge can quickly expand if care is not taken. Data mining extracts 
new knowledge out of accumulated observations and thus provides a basis for decision 
making and action. How to turn understanding of buried knowledge to best 
use? How to extract operational feedback from preexisting, but latent, dormant 
empirical knowledge? Such questions precede any data - mining project. 
As a multidisciplinary technique, data mining sits at the interface between statistics, 
mathematics, and computer science. It is a collection of methods for detecting 
regularities and patterns and for extracting knowledge from massive databases 
using conventional and advanced analytical tools. Another approach to data mining 
is to view it as the multivariate modeling of a real environment on the basis of 
multidimensional and accumulated historical data. Thus, data mining is similar to 
explorative data analysis. It is driven by the data itself. However, it must be considered 
as different from conventional statistics due to the huge volume of processed 
data, far above the megabyte scale. Beyond this critical database dimension, most 
conventional statistical packages exceed their operational limit. Data mining can 
also be performed without the help of professional statisticians. It runs according 
to semiautomatic procedures, which makes it widely attractive and more likely to 
be used in an industrial environment. 
Such situations are characteristic of pharmaceutical processes which accumulate 
a variety of historical data without consideration of pertinence. Accumulation is 
systematic and exhaustive. However, cross - links between data sources or types may 
not be established, leading to irrelevant and undetected redundancies. Reliability 
of the collected data is not clearly established over time and variations may not be 
detected. 

BASIC CONCEPTS AND IMPACT 359 
4.2.1.9 Data Warehousing 
The 1990s saw the development of data warehouses. An ideal data warehouse is a 
collection of historical data varying with time, organized by topic, aggregated in a 
unique database, and stored in a way that facilitates decision making (Figure 2 ). 
Three main functions are required to manage data warehouses. First, the data must 
be collected or else accessed by an alternative method, for example, as preexisting 
databases or fi les. Second, the data warehouse requires management and control 
tools. Only then can the third function operate, namely data analysis for the purpose 
of decision making and new knowledge. Dedicated information management tools 
mediate all external, operational, and historical data to the warehouse. Decisional 
information management components are used to extract and visualize the data 
warehouse information. Online analysis processing (OLAP) consists of the real - 
time analysis and visualization of the historical data. Data mining involves the 
extraction of rules and models constructed from the collected data. 
Online analysis processing mainly comprises the interactive exploration of multidimensional 
data sets, or data cubes, which are manipulated by operations from 
matrix algebra, for example, slice - and - dice, roll - up, and drill - down. Computing performance 
is related to data warehouse size and also data quality, for example, 
missing data, unsharpness, and redundancy. The multidimensionality issue is critical 
for extracting pertinent information and selecting the results to be stored and 
visualized. 
The data - mining tools now incorporated in much commercial software are a set 
of techniques and algorithms for exploring large databases in order to extract 
semantic links pertinent to event explanation and new knowledge acquisition. The 
more general goal of data mining is to extract rules and models for understanding 
connections and assisting decision making. There are numerous fi elds of application: 
risk analysis, manufacturing trends, raw material management, maintenance, process 
validation, development, quality control, and so forth. The idea behind data mining 
consists in introducing or proposing rules associated with likelihood coeffi cients 
established from a large set of existing (i.e., historical) data. The techniques used 
FIGURE 2 Schematic structure of a data warehouse. 
Data 
warehouse 
External data sources Internal data sources Operational data sources 
Data-mining 
modeling of rules 
prediction 
OLAP 
Data analysis 
visualization 
Transformation or 
pretreatments

360 PROCESS ANALYTICAL TECHNOLOGY 
are drawn from the fi elds of artifi cial intelligence and numerical and statistical data 
analysis, for example, functional modeling, learning machines, neuronal networks, 
Bayesian networks, support vector machines, modeling of associations, and explanatory 
rules, classifi cations, and segmentations. Their computing complexity derives 
from the dramatic up scaling from database to data warehouse level (from megabase 
to petabase, i.e., 10 6 . 10 15 ). 
4.2.1.10 Data - Mining Methods for Pharmaceutical Processes 
The data warehouse is a central repository of data accumulated over time from 
various origins: quality control, quality assurance, production, development, and the 
like. The accumulated data represent a potential gold mine, conferring competitive 
advantage by facilitating understanding of pharmaceutical process and optimizing 
it in the light of buried empirical knowledge. 
Data mining is used to extract previously unexploited data and knowledge. Its 
potential for acquiring knowledge and generating explanatory rules can overcome 
the loss of data or underused accumulated data. There are two ways of proceeding. 
The fi rst is proactive or directed, for example, hypothesis testing. Particular groupings 
or features are suspected, and verifi cation or confi rmation of identity is sought. 
The second is reactive or undirected, consisting of simple data exploration. Groupings 
are unknown, properties undetected or latent, and patterns unidentifi ed. Alternative 
terms for these approaches are supervised and unsupervised learning, 
respectively. Top - down and bottom - up approaches complement one another. For 
example, the confi rmatory tools of supervised learning can be used to verify and 
certify the quality of the discoveries obtained using the exploratory approach. 
What can be obtained using data - mining tools? Here is a short list of achievable 
goals: 
• Data characterization to extract or determine descriptors or indicators, for 
example, by generalizing, summarizing, or grouping 
• Establishment of associative and explanatory rules 
• Classifi cation (supervised learning) of items or objects in classes according to 
a given probability 
• Clustering of data items (unsupervised learning) in classes, after establishing 
class limits inductively from existing data sets 
• Detection of similarities in time series 
• Pattern recognition 
Data from external and internal sources is integrated, aggregated, or associated in 
time series. Data items may contain errors or the data may be missing, unsharp, 
redundant, or contradictory. A language with operators and variables is required to 
establish models. Validity levels also have to be defi ned using suitable optimization 
and validation criteria. In addition, a search method is required to extract the data 
from the data warehouse and prepare it for analysis. 
Data mining can, therefore, be considered as a three - step operation. Prior to any 
analysis, the collected data is preprocessed to integrate the warehouse, and some 
verifi cation is performed to maintain the data level: for example, integration, 

BASIC CONCEPTS AND IMPACT 361 
aggregation, or grouping of data from different internal and external sources. The 
data is then selected and data mining performed applying the appropriate algorithms 
or models. Results are visualized and interpreted for experts in the fi eld. 
4.2.1.11 Data - Mining Practice 
Data mining is part of an action process known as a business intelligence chain 
(Figure 3 ). Data mining is a fl exible solution to the recurrent problem of how to 
derive knowledge from data. The source for data mining is the existence of a large 
but buried data set. The corresponding data analysis is an intellectual method that 
applies only if integrated into the current operational process. Hypothesis testing, 
knowledge acquisition, and the generation of explanatory rules are directed by 
active collaboration between different process actors. Data mining is teamwork that 
requires expertise in various areas, such as information technology (IT), database 
management, and data analysis. However, the methods are available in commercial 
packages and may not require the expertise of traditional statisticians. It is the 
computer which is responsible for discovering patterns or identifying rules or features. 
In summary, data mining is a logical loop involving the following steps: 
• Business understanding 
• Precise setting of the data - mining project, for example: 
Defi nition of realistic objectives 
Field of treated data 
Inventory of available or usable data 
• Data preparation 
Extraction from internal or external sources 
Verifi cation and correction 
Pretreatment 
• Warehouse construction 
• Modeling, for example: 
Description and visualization 
Affi nity grouping 
Rules of association, explanatory rules 
Clustering 
FIGURE 3 Place of data mining in the decision chain. 
Data Information Knowledge Decision Action 
Business intelligence chain 
Data mining 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 

362 PROCESS ANALYTICAL TECHNOLOGY 
Classifi cation 
Estimation 
Prediction 
• Evaluation and comparison of models 
• Documentation and presentation of results 
• Deployment for action 
• Back to business understanding. 
4.2.1.12 Comments about Data Mining 
Data mining provides an explanatory analysis from a confi rmatory analysis. It is 
tempting to extract maximal value from available resources such as any kind of 
accumulated data. But maximal effi ciency requires critical insight into the expertise 
actually buried in data collections or warehouses. The goal of data exploration is to 
access the buried data to acquire the knowledge that will make explanation, prediction, 
or estimation possible. That is why data mining requires team effort from data 
specialists, users, information technologists, and specialists in the relevant fi eld (in 
this case, pharmaceutical process). It also requires senior management support 
throughout the organization. Mining is a matter of good practice according to established 
rules but also a challenge for innovative mathematical techniques. Not all 
patterns or rules found by data mining are interesting, although the results should 
remain logical and actionable by experts in the relevant fi eld. Because the 
algorithms involved tend to be complex and the data volume is huge, software 
implementation together with the level of information technology are major 
considerations. 
Data mining is driven by the accumulated data but always directed at solving a 
process, business, or research problem. The results are designed to make it easier to 
reach a diagnosis or make a decision. They are only likely to be useful in context: 
that is, they are not simply numbers and graphics but an aid to insight for experts 
in the relevant fi eld. Also, no single mining technique is equally applicable. A range 
of different methods or algorithms should be considered, as no one particular technique 
will work equally well or outperform all other techniques on all problems. 
Nor will the value of an analytical technique exceed that of the data upon which it 
is based. 
4.2.1.13 PAT Methods 
Almost any existing analytical method can serve the objectives of PAT. Many online 
applications already exist. With newer techniques, like NIR imaging or matrix - 
assisted laser desorption/ionization time - of - fl ight (MALDI - TOF) mass spectrometry, 
there are technological problems about performing online or inline analytics. 
Implementing a given analytical technique close to or during process does not 
always provide better process understanding. Attributes which are not informative 
should not be measured at all and are not worth the burden of complex process 
implementation. 
Use of the various techniques listed in Table 1 depends on process requirements. 
The validity of a given technique or analytical application is challenged by every 
0 
0 
0 

TABLE 1 Analytical Methods for PAT 
Method Description Online Application Chemical Identifi cation 
Pharmaceutical Application 
Examples 
Infrared, 
near - infrared, 
and 
Raman spectroscopy 
Vibrational spectroscopy 
(discussed in this chapter) 
. 
. 
Reaction monitoring 
Polymorphism 
Content determination 
Process monitoring (drying, 
granulation, blending) 
Hyperspectral imaging 
Vibrational spectroscopy 
coupled with a spatial 
analysis (cf. chemical 
imaging chapter) 
. 
Chemical compound distributions 
Counterfeit detection 
UV – Vis spectroscopy Photoelectron spectroscopy 
. 
. 
Color measurement 
Dissolution testing 
Cleaning validation (ppm - level 
detection) 
Terahertz spectroscopy Far - infrared spectroscopy; 
3D imaging 
. 
Polymorphism 
Coating integrity and thickness 
API distribution possible 
Laser - induced breakdown 
spectroscopy 
Plasma generated by a laser 
pulse and detection of the 
emitted light (destruction 
of sample) 
. 
Drug development 
Process troubleshooting 
Laser diffraction Interaction of a laser beam 
with particles and 
detection of the scattered 
light 
. 
Particle size determination 
Effusivity Combines thermal 
conductivity, density, and 
heat capacity 
. 
Mixing, 
blending, 
granulation 
monitoring 
Acoustic methods 
Active or passive 
. 
Solid, 
semisolid, 
and high viscose 
sample 
High shear granulation 
monitoring Crystallization 
monitoring 
363

364 PROCESS ANALYTICAL TECHNOLOGY 
technological advance or new analytical technique. Innovation continuously drives 
optimization of overall process performance. 
4.2.1.14 Conclusion 
Process Analytical Technology can be viewed as a constellation placing greater or 
less emphasis on a given activity depending on the current problem or situation 
(Figure 4 ). There is no written rule or straightforward path to progress through PAT. 
Experience and expertise are necessary, together with a good knowledge of the 
pharmaceutical environment. Once a pharmaceutical company has decided 
to implement PAT, continuous management support for the development and 
maintenance of PAT - related activities is critical. It is a strategic and necessary step 
for the future success of PAT to encourage, stimulate, and initiate scientifi c collaboration 
and interaction as well as the relevant education and training. Better understanding 
and control of chemical and pharmaceutical processes are greatly needed, 
as well as the development of advanced measurement tools and data analysis 
methods. 
A summary of PAT benefi ts follows: 
• Immediate action if quality is not met 
• Better process control and understanding 
• Less uncontrolled variation and less production waste 
• Better and more stable products 
• Data collection and improved historical knowledge 
Process analytical technology continuously improves product quality, extends the 
acquired knowledge base for new projects, and shortens time to market. 
FIGURE 4 PAT constellation (DoE, design of experiments). 
DoE 
Risk 
analysis 
Analytical 
methods 
PAT 
Chemometrics 
Data 
mining 
Physical 
pharmacy 
Pharmaceutical 
sciences 
Sensor 
technology 
Pharmaceutical 
technology

4.2.2 VIBRATIONAL SPECTROSCOPY 
4.2.2.1 Introduction 
Modern infrared (IR) spectroscopy is a versatile tool applied to the qualitative and 
quantitative determination of molecular species of all types. Its applications fall into 
three categories based on the spectral regions considered. Mid - IR (MIR) is by 
far the most widely used, with absorption, refl ection, and emission spectra being 
employed for both qualitative and quantitative analysis. The NIR region is particularly 
used for routine quantitative determinations in complex samples, which is of 
interest in agriculture, food and feed, and, more recently, pharmaceutical industries. 
Determinations are usually based on diffuse refl ectance measurements of untreated 
solid or liquid samples or, in some cases, on transmittance studies. Far - IR (FIR) 
is used primarily for absorption measurements of inorganic and metal - organic 
samples. 
Within the electromagnetic spectrum (Figure 5 ), the IR region ranges from 12,800 
to 10 cm . 1 or from 0.78 to 1000 . m. The IR domain is conveniently subdivided into 
NIR, MIR, and FIR, respectively, with the following limits: 
Near 0.78 – 2.5 4000 – 12,800 
Mid 2.5 – 50 200 – 4000 
Far 50 – 500 20 – 200 
Methods and applications differ with the IR subregion considered. Academia and 
analytical chemists commonly consider MIR as the default region of interest. 
Current MIR instruments are completely different from traditional grating spectrophotometer 
technology. The generalization of Fourier transform (FT) – based spectrometers 
in the early 1980s lowered instrument prices and increased the number 
and types of MIR applications, in particular thanks to the use of interferometers in 
improving signal - to - noise ratios and detection limits. IR applications were originally 
limited to qualitative organic analysis. Almost from the outset, absorption MIR 
became a well - established application for structure elucidation. Organic chemists 
were trained in the visual and direct interpretation of MIR spectra. Nowadays mid - 
IR spectroscopy (MIRS) tends to be more viewed as a useful tool for the quantitative 
analysis of complex samples by absorption and emission spectrometry, which 
may require calibration and data pretreatment. 
Near - IR measurements can be performed similarly to those using dedicated 
ultraviolet (UV) or visible spectrophotometers. Historically, the most important 
FIGURE 5 Limits and designation of the spectroscopic domains. 
cm-1 3.3 20 200 4 000 12 500 25 000 10 5 
Far Middle Near 
Microwaves IR VIS UV 
mm 3 000 500 50 2.5 0.8 0.4 0.1 
VIBRATIONAL SPECTROSCOPY 365

366 PROCESS ANALYTICAL TECHNOLOGY 
application was quantitative analysis in the food and feed industries. Only more 
recently have the chemical and pharmaceutical industries shown increasing interest 
in the NIR range. The major reason for the delay is in the type of information 
delivered. All observed bands result from overtones or combinations of overtones 
originating in the fundamental MIR region of the spectrum. Because the measurement 
method is nondestructive, samples are measured with little or no specifi c 
preparation. NIR spectra contain chemical and physical information on the sample. 
Direct interpretation is limited, if not impossible, meaning that multivariate data 
processing is routinely required to extract the relevant information. This led most 
analytical chemists to ignore the potential of NIR. Until the early 1990s, NIR spectrophotometers 
tended to be the dispersive type based on diffraction gratings. Subsequent 
technological advance has brought FT and diode array instruments. Filter 
instruments remain used for ultrarapid measurement of material composition in the 
food and feed industries. 
Being at the edge of the IR region, FIR is believed to have less industrial potential. 
This is partly due to unresolved experimental and technological diffi culties. FIR 
may provide relevant information, but at the cost of disproportionate effort. Routine 
use in the pharmaceutical environment is not anticipated in the near future, and for 
this reason we shall not discuss FIR further. 
The most recent developments in IR/NIR technology include imaging large sample 
surfaces, nondestructive analysis of solids by attenuated total refl ectance (ATR), and 
photoacoustic measurement. Instrument performance continues to increase, with particular 
respect to reliability and modularity. Spectrometer downsizing, speed of measurement, 
and mobility no longer represent critical challenges. However, what has 
really expanded the scope of MIR applications, and use of the full NIR region, has 
been the constant increase in computing power. The fi eld of application of IR spectroscopy 
is moving toward the quantitative analysis of complex samples in various 
measurement modes. These types of samples are characteristic of the pharmaceutical 
industry. Noninvasive spectral sampling using light probes is at last making in situ 
analytics attractive, for example, for performing online real - time measurements. 
Infrared microscopy was introduced in the early 1980s. Two microscopes, an 
ordinary optical microscope and an FT IR instrument with refl ection optics, were 
combined. The optical microscope is used to visually locate the spot of interest. The 
spot is then irradiated with the IR or NIR beam. There are numerous applications 
for noninvasive measurement, including of contaminants, particles, imperfections, 
and for fi ber identifi cation. Chemical imaging systems (CIS) are a refi nement of the 
technique. Spectra are collected from adjacent areas (pixels) on a larger surface. In 
practice, an imaging breakthrough became possible after moving away from pixel - 
after - pixel scanning. CIS fl exibility and speed of acquisition improved with the 
introduction of new detectors, for example, focal plane array (FPA) detectors. Multiple 
IR/NIR spectra (up to many thousand) are scanned in a single step on the 
sample surface. With image analysis algorithms and fast computers, current NIR/IR 
imaging techniques hold fresh promise for resolving quality problems. 
4.2.2.2 IR Spectroscopy Theory 
In a typical IR absorption spectrum of an organic substance (Figure 6 ), the ordinate 
is transmittance and the abscissa is the wavenumber. A linear wavenumber scale is 

preferred because of the linear relationship between wavenumber and energy and 
frequency. The frequency of an absorbed radiation is the molecular vibrational frequency 
actually responsible for the observed absorption. 
Infrared absorption, emission, or refl ection for molecular species can be explained 
by assuming transitions from one rotational or vibrational energy state to another. 
IR radiation is not energetic enough to produce electronic transitions similar to 
those resulting from UV, visible (Vis), or X - ray radiation. Absorption of IR radiation 
is limited to molecular species with small energy differences between various 
vibrational and rotational states. In order to absorb IR radiation, a molecule must 
undergo a net change in dipole moment as a consequence of its vibrational or rotational 
motion. Under these circumstances an alternating electrical fi eld interacts 
with the molecule and causes changes in the amplitude of one of its motions. The 
dipole moment is determined by the magnitude of the charge difference and the 
distance between the two charge centers. In addition, regular fl uctuation in dipole 
moment occurs, and a fi eld is established which interacts with the electrical fi eld 
associated with the incident radiation. If the radiation frequency exactly matches a 
natural vibrational frequency of the molecule, a transfer of energy takes place that 
changes the amplitude of molecular vibrations and absorption of radiation results. 
Similarly, the rotation of asymmetric molecules around their centers of mass results 
in periodic dipole fl uctuations which interact with radiation. Homonuclear species 
are not concerned and such compounds cannot absorb in the IR. 
The amount of energy required to cause a change in energy level is approximately 
equivalent to radiation of 100 cm . 1 or less. The relative positions of atoms in a molecule 
fl uctuate continuously, and multiple types of vibrations and rotations about 
the bonds in the molecule are possible. Exact analysis of all movements becomes 
FIGURE 6 Typical example of an infrared absorption spectrum. 
55 
60 
65 
70 
75 
80 
85 
90 
95 
%T 
500 1000 1500 2000 2500 3000 3500 
Wavelength ( cm–1) 
VIBRATIONAL SPECTROSCOPY 367

368 PROCESS ANALYTICAL TECHNOLOGY 
impossible for molecules comprising several atoms. Not only do larger molecules 
have more vibrating possibilities, but intercenter interactions occur that must be 
taken into account. Vibrations may be of the stretching and bending variety. Stretching 
vibration involves a continuous change in interatomic distance along the axis of 
the bond between the atoms. Bending vibration is characterized by a change in the 
angle between two bonds and comes in four types: scissoring, rocking, wagging, and 
twisting. All vibration types may be possible in a molecule containing more than 
two atoms. In addition, vibration interaction or coupling may occur if the vibrations 
involve bonds to a single central atom with a change in the characteristic of the 
vibrations concerned. 
4.2.2.3 Mechanical Model of IR Vibration 
Infrared spectra result from light absorption by organic molecules. The easiest way 
to describe vibrational spectroscopy from a theoretical perspective is to consider 
the isolated vibrations of a mechanical model called the harmonic oscillator. Atomic 
stretching vibration behavior can be approximated by a mechanical model consisting 
of two masses, m 1 and m 2 , connected by an ideal spring. Displacement of one 
such mass along the spring axis results in harmonic motion. Many fundamental frequencies 
may be calculated by assuming that band energies arise from the vibration 
of the ideal diatomic harmonic oscillator (Figure 7 ), obeying Hooke ’ s law, that is, 
. . = 1
2 
k
u 
where . is the vibrational frequency, k the classical force constant, and 
u mm m m = + ( ) 1 2 1 2 , the reduced mass of the two atoms. 
The model provides a good description of true diatomic molecules and is not far 
from the average value of two atoms stretching within a polyatomic molecule. The 
corresponding potential - energy curve is the typical parabola illustrated in Figure 8 . 
This approximation gives the average vibration frequency of the bond. For example, 
the reduced masses for C — H, O — H, and N — H are 0.85, 0.89, and 0.87. These 
fi gures are similar, so the frequencies would be quite similar too. However, the 
electron - withdrawing and - donating properties of neighbors within molecules act 
FIGURE 7 Ideal diatomic harmonic oscillator. 
x 
x2 
0
requilibrium 
r 
x 
m1 m2 
x 1 
2 1

on the observed band strength, length, and frequency. An average value is of little 
use in structural determinations and these differences cause a real spectrum to 
develop. The force constant k is a measure of the stiffness of the chemical bond and 
is the equivalent of the force constant of the spring in the harmonic model. The k 
values vary widely and cause energy differences which can both be calculated and 
utilized in spectral interpretation. It has been possible to evaluate some force constants 
for various types of chemical bonds by IR spectroscopy. Generally, k has been 
found to range between 3 . 10 2 N/m and 8 . 10 2 N/m for most single bonds (average: 
5 . 10 2 N/m). Double and triple bonds are found to have k values two and three 
times this average, respectively. In practice, these average experimental values can 
be used to estimate the wavenumbers of fundamental absorption peaks, that is, 
peaks of the transition from the ground state to the fi rst excited state, for a variety 
of bond types. 
Classical mechanics does not apply to the atomic scale and does not take the 
quantized nature of molecular vibration energies into account. Thus, in contrast to 
ordinary mechanics where vibrators can assume any potential energy, quantum 
mechanical vibrators can only take on certain discrete energies. Transitions in vibrational 
energy levels can be brought about by radiation absorption, provided the 
energy of the radiation exactly matches the difference in energy levels between the 
vibrational quantum states and provided also that the vibration causes a fl uctuation 
in dipole. 
4.2.2.4 Quantum Mechanical Model 
Unlike the classical spring model for molecular vibrations, there are not an infi nite 
number of energy levels. Instead of a continuum of energies, there are discrete 
energy levels described by quantum theory. The time - independent Schr o dinger 
equation is solved using the vibrational Hamiltonian for a diatomic molecule. Values 
for the ground state ( . = 0) and succeeding excited states can be calculated by 
solving the equation (Figure 8 ). Absorption of a photon of the correct energy can 
cause the molecule to change between vibrational energy levels. At room temperature 
only the ground state has a signifi cant population, and so transitions due to 
absorption at these temperatures occur from the ground state. Transitions between 
ground state to energy level 1 give the fundamental absorption if this leads to a 
FIGURE 8 Energy diagram of the ideal diatomic oscillator. 
Potential energy 
V=0 
V=1 
V=2 
V=3 
Interatomic distance
Energy 
level 
VIBRATIONAL SPECTROSCOPY 369

370 PROCESS ANALYTICAL TECHNOLOGY 
change in molecular dipole moment. Transitions between ground state and energy 
level 2 or above give overtones. Transitions between multiple states can occur and 
give rise to combination bands. 
A simplifi ed version of the energy levels may be written for the energy levels of 
a diatomic molecule: 
E 
h k
u . . 
. 
. = + ( ) = 1
2 2 
0, 1, 2, . . . 
in which Hooke ’ s law terms can be seen. Rewritten using the quantum term 
hV h k u =( ) 2. , the equation reduces to 
E hV . . . = + ( ) = 1
2 
0, 1, 2, . . . 
In the case of polyatomic molecules, the energy levels become quite numerous. 
Ideally, one can treat such a molecule as a series of diatomic, independent, harmonic 
oscillators and the above equation can be generalized: 
E hV 
i
N 
i i . . . . . . . 1 2 
1 
3 6 
1 
1
2 
0 , , , . . . , , , . . . , 1, 2, 3, . . . 3 23 ( )= + ( ) = 
=
. 
. 
Any transition of an energy state from 0 to 1 in any one of the vibrational states 
( . 1 , . 2 , . 3 , . ) is fundamental and allowed by selection rules. Where the transition 
is from the ground state to . i = 2, 3, and so on and all others are zero, it is known 
as the fi rst overtone, the second overtone, and so on. Transitions from the ground 
state to a state for which . i = 1 and . j = 1 simultaneously are combinations. Other 
combinations, such as . i = 1, . j = 1, . k = 1, or . i = 2, . j = 1, and so forth are also 
possible. Typically, NIR spectra will contain these overtones and combinations 
derived from the fundamental vibrations which appear in the MIR. Overtones and 
combinations are not allowed, but appear as weak bands due to anharmonicity or 
Fermi resonance. As a rule, overtones occur at one - half and one - third of the fundamental 
absorption wavelength or 2 and 3 times the frequency. The majority of 
overtone peaks arise from the R — H stretching and bending modes because the 
dipole moment is high: O — H, C — H, S — H, and N — H are strong NIR absorbers 
and form most NIR bands. Since most absorption is repeated in the NIR range, this 
region is likely to be used to identify a molecule, as with MIR. As a consequence, 
IR bands are traditionally used to identify functional groups which have characteristic 
frequencies. NIR spectra are more overlapping, and, although bands can be 
identifi ed, they cannot be placed in relation to the rest of the molecule. NIR spectra 
are, therefore, mainly used to confi rm the identity of a material, as for true 
identifi cation. 
As given from the quantum mechanics equations, the energy for transition from 
energy levels 1 to 2 or 2 to 3 should be identical to that for transition from 0 to 1. 
Furthermore, quantum theory states that the only transitions that can take place 
are those for which, according to vibrational quantum theory, the vibrational 
quantum number changes by unity. This is the so - called selection rule. 

So far we have illustrated the classic and quantum mechanical treatment of the 
harmonic oscillator. The potential energy of a vibrator changes periodically as 
the distance between the masses fl uctuates. In terms of qualitative considerations, 
however, this description of molecular vibration appears imperfect. For example, as 
two atoms approach one another, Coulombic repulsion between the two nuclei adds 
to the bond force; thus, potential energy can be expected to increase more rapidly 
than predicted by harmonic approximation. At the other extreme of oscillation, a 
decrease in restoring force, and thus potential energy, occurs as interatomic distance 
approaches that at which the bonds dissociate. 
In theory, the wave equations of quantum mechanics can be used to derive near - 
correct potential - energy curves for molecular vibrations. Unfortunately, the mathematical 
complexity of these equations precludes quantitative application to all but 
the very simplest of systems. Qualitatively, the curves must take the anharmonic 
form. Such curves depart from harmonic behavior by varying degrees, depending 
on the nature of the bond and the atom involved. However, the harmonic and 
anharmonic curves are almost identical at low potential energies, which accounts 
for the success of the approximate methods described. 
Anharmonicity leads to deviations of two kinds. At higher quantum numbers, . E 
becomes smaller, and the selection rule is not rigorously followed; as a result, transitions 
of . ± 2 or ± 3 are observed. Such transformations are responsible for the 
appearance of overtone lines at frequencies approximately two or three times that 
of the fundamental line; the intensity of overtone absorption is frequently low, and 
the peaks may not be observed. Vibrational spectra are further complicated by the 
fact that two different vibrations in a molecule can interact to give absorption peaks 
with frequencies that are approximately the sums or differences of their fundamental 
frequencies. Again, the intensities of combination and difference peaks are generally 
low. 
It is ordinarily possible to deduce the number and kinds of vibrations in simple 
diatomic and triatomic molecules and determine whether these vibrations contain 
several types of atoms as well as bonds; for these molecules, the multitude of possible 
vibrations gives rise to IR spectra that are diffi cult, if not impossible, to analyze. 
The number of possible vibrations in a polyatomic molecule can be calculated as 
follows. Three coordinates are needed to locate a point in space; fi xing N points 
requires 3 N coordinates. Each coordinate corresponds to one degree of freedom for 
one of the atoms in a polyatomic molecule; for this reason, a molecule containing 
N atoms is said to have 3 N degrees of freedom. A molecule features three types of 
motion. First, the motion of the entire molecule through space; second, the rotational 
motion of the entire molecule around its center of gravity; and, third, the 
vibrations of each of its atoms relative to the other atoms. Since all atoms in the 
molecule move in concert through space, defi nition of translational motion requires 
three of the 3 N degrees of freedom. Another 3 degrees of freedom are needed to 
describe the rotation of the molecule as a whole. The remaining 3 N . 6 degrees of 
freedom involve interatomic motion and hence represent the number of possible 
vibrations within the molecule. In a linear molecule 2 degrees of freedom suffi ce to 
describe rotational motion. Thus, the number of vibrations for a linear molecule is 
3 N . 5. Each of the 3 N . 6 or 3 N . 5 vibrations is a normal mode. For each normal 
mode of vibration there is a potential energy relationship. In addition, to the extent 
that a vibration approximates harmonic behavior, the differences between the 
VIBRATIONAL SPECTROSCOPY 371

372 PROCESS ANALYTICAL TECHNOLOGY 
energy levels of given vibrations are the same; that is, a single absorption should 
appear for each vibration in which there is a change in dipole. 
However, fewer experimental peaks may be observed than would be expected 
from the theoretical number of normal modes. Fewer peaks can be found when the 
symmetry of the molecules is such that no change in dipole results from a particular 
vibration. The energies of two or more vibrations can be identical or nearly identical. 
In some cases absorption intensity is too low to be detected by ordinary means. It 
may also happen that the vibrational energy is in a wavelength region which is 
beyond the range of the instrument. 
Conversely, more peaks may be found than expected from the number of normal 
modes. This is the typical situation that concerns the NIR domain. Overtone peaks 
at two or three times the frequency of a fundamental peak, or addition combination 
bands at approximately the sum or difference of two fundamental frequencies, are 
sometimes encountered. The energy of a vibration and thus the wavelength of its 
absorption peak may be infl uenced by, or coupled with, other vibrators in the molecule. 
A number of factors infl uence the extent of such coupling. Vibration coupling 
is a common phenomenon. As a result, the position of an absorption peak corresponding 
to a given organic functional group cannot always be specifi ed exactly. 
While interaction effects may lead to uncertainties in the identifi cation of functional 
groups contained in a compound, it is this very effect that provides the unique features 
of an IR absorption spectrum that are so important for the positive identifi cation 
of a specifi c compound. 
4.2.2.5 Anharmonicity 
The ideal harmonic oscillator is a somewhat limited model. As the oscillating masses 
get very close, real compression forces — which are neglected in calculations — fi ght 
against the bulk of the spring. As the spring stretches, it eventually reaches a point 
where it loses its shape and fails to return to its original coil. This ideal case is shown 
in Figure 9 . The barriers at either end of the cycle are approached in a smooth and 
orderly fashion. Likewise, in molecules, the respective electron clouds of the two 
bound atoms limit approach by the nuclei during the compression step, creating an 
energy barrier. At extension of the stretch, the bond eventually breaks when the 
vibrational energy level reaches the dissociation energy. The barrier at smaller dis- 
FIGURE 9 Energy diagram of the anharmonic diatomic oscillator. 
Interatomic distance 
Potential energy 
V=0
V=1
V=2
V=3 
Energy 
level

tances increases at a rapid rate, while the barrier at the far end of the stretch slowly 
approaches zero (Figure 9 ). The shape of the potential energy curve is typical of an 
anharmonic oscillator. 
Energy levels in the anharmonic oscillator are not equal, although they become 
slightly closer as energy increases. This phenomenon can be seen in the following 
equation: 
E hW W X e e e . . . = + ( ) . + ( ) + 1
2 
1
2 
2 
higher terms 
where W Ku e e = ( ) 
1
2 
1 2 . is the vibrational frequency, W e X e the anharmonicity 
constant, K e the anharmonicity force constant, and u the reduced mass of the two 
atoms. In practice, anharmonicity is between 1 and 5%. Thus, the fi rst overtone of 
a fundamental vibration set, for example, at 3500 nm would be 
. = + .[ ] ( ) 
3500 
2 
3500 0 01 . , 0.02, . . . 
Depending on structural or steric conditions, the number may range from 1785 to 
1925 nm for this example. However, it would generally appear at 3500/2, plus a relatively 
small shift to a longer wavelength. As forbidden transitions, the overtones are 
between 10 and 1000 times weaker than the fundamental bands. Thus, a band arising 
from bending or rotating atoms would have to be in its third or fourth overtone to 
be seen in the NIR region of the spectrum. For example, a fundamental carbonyl 
stretching vibration at 1750 cm . 1 or 5714 nm would have a fi rst overtone at approximately 
3000 nm, a weaker second overtone at 2100 nm, and a third very weak overtone 
at 1650 nm. The fourth overtone, at about 1370 nm, would be so weak as to be 
useless. These fi gures are based on an illustrative 5% anharmonicity constant. 
The detailed examination of the spectra of simple molecules is a direct source to 
determine the characteristic NIR frequencies for selected vibration modes. For 
qualitative and quantitative analyses there is the requirement to interpret as much 
as possible the NIR spectrum. Although interpretation of spectra in a manner 
analogous to MIR is not conceivable, attempts exist to defi ne and categorize 
observed NIR frequencies. Examples of reported frequencies for aliphatic hydrocarbons 
are given in the following list: 
8547 cm . 1 C — H second overtone in CH —— CH 
8474 cm . 1 C — H group in cis olefi ns 
7700 – 9000 cm . 1 C — H second overtone 
8696 cm . 1 Second overtone of CH 2 antisymmetric stretching 
8285 cm . 1 Second overtone of CH 2 symmetric stretching 
1080 – 1140 cm . 1 Second overtone olefi n 
7692, 8237, 8576 cm . 1 C — H stretching second overtone in CH 2 
4.2.2.6 Structure Elucidation Using MIRS 
Mid - IR absorption and refl ectance spectroscopy is typically used for determining 
the structure of organic and biochemical species. When used in conjunction with 
VIBRATIONAL SPECTROSCOPY 373

374 PROCESS ANALYTICAL TECHNOLOGY 
other analytical methods, such as mass spectroscopy, nuclear magnetic resonance, 
and elemental analysis, IR spectroscopy usually achieves positive species identifi cation. 
Spectra are obtained after sample preparation, usually involving dilution of 
the analyte. Sample handling is the diffi cult and time - consuming part of the analysis. 
Organic samples exhibit numerous IR absorption peaks used for qualitative structure 
confi rmation. First, presumptive functional groups are identifi ed by examining 
their frequency region from about 3600 to 1200 cm . 1 . As mentioned earlier, the frequency 
at which an organic functional group absorbs radiation can be approximated 
from the atomic masses and bond forces between them. These group frequencies 
are not totally invariant because of interactions with other vibrations. However, such 
interaction effects are small, and a range of frequencies can be assigned within which 
it is highly probable that the absorption peak for a given functional group will be 
found. Group frequencies are listed in correlation charts, which serve as a starting 
point in the identifi cation process. 
Second, the spectrum of the unknown is compared with the spectra of reference 
compounds featuring all the functional groups found in the fi rst step. The fi ngerprint 
region from 1200 . 1 to 600 cm . 1 is extremely useful because small differences in 
structure and constitution produce signifi cant changes in the appearance and distribution 
of absorption peaks in this region. Most single bonds give rise to absorption 
bands at these frequencies. Because their energies are about the same, strong interaction 
occurs between neighboring bonds. The absorption bands are thus composites 
of these various interactions and depend upon the overall skeletal structure of 
the molecule. Exact interpretation in this region is seldom possible because of spectral 
complexity. On the other hand, it is this complexity that leads to uniqueness and 
the consequent usefulness of the region in fi nal identifi cation. A close match between 
two spectra in the fi ngerprint region constitutes almost conclusive compound 
identifi cation. 
In employing group frequencies it is essential that the entire spectrum rather 
than a small isolated portion be considered and interrelated. Correlation charts 
serve only as a guide for further and more careful study. Catalogs of IR spectra that 
assist in qualitative identifi cation by providing comparison and reference spectra 
for a large number of pure compounds are commercially available on electronic 
media. Optimized search systems for identifying compounds from IR spectral databases 
and algorithms for the matching step produce rapid and reliable potential 
hits. 
4.2.2.7 Extending Use of MIRS 
Organic and inorganic molecular species (except homonuclear molecules) absorb 
in the IR region. IR spectroscopy has the potential to determine the identity of an 
unusually large number of substances. Moreover, the uniqueness of a MIR spectrum 
confers a degree of specifi city which is matched or exceeded by relatively few other 
analytical methods. This specifi city has found particular applications for the development 
of quantitative IR absorption methods. However, these differ from quantitative 
UV/Vis techniques in their greater spectral complexity, narrower absorption 
bands, and the technical limitations of IR instruments. Quantitative determinations 
obtained from IR spectra are usually inferior in quality and robustness to those 
obtained with UV/Vis and NIR spectroscopy. In addition, univariate or linear cali

bration curves require meticulous attention to numerous details. One cause of 
failure is the frequent nonadherence to Beer ’ s law due to the inherent complexity 
of IR spectra, featuring overlapping absorption peaks or disturbance by stray radiation. 
Analytical uncertainties cannot be reduced to a level which is comparable to 
other methods, despite considerable effort or care. 
Diffuse - refl ectance MIRS has found a number of applications for dealing with 
hard - to - handle solid samples, such as polymer fi lms, fi bers, or solid dosage forms. 
Refl ectance MIR spectra are not identical to the corresponding absorption spectra, 
but suffi ciently close in general appearance to provide the same level of information. 
Refl ectance spectra can be used for both qualitative and quantitative analysis. Basically, 
refl ection of radiation may be of four types: specular, diffuse, internal, and 
attenuated total. 
Specular refl ection is encountered when the refl ecting medium is a smooth polished 
surface. The angle of refl ection is identical to the incident angle of the radiation 
beam. If the surface is IR absorbent, the relative intensity of refl ection is less 
for wavelengths that are absorbed than for wavelengths that are not. Thus, the plot 
of refl ectance R , defi ned as the fraction of refl ected incident radiant energy versus 
the wavelength (or wavenumber) appears similar to a transmission spectrum for the 
sample. 
Diffuse - refl ectance spectra are obtained directly from powder samples after a 
minimum of preparation. In addition to the time saved, measurement is nondestructive, 
leaving the sample intact for further analysis. The widespread use of diffuse 
refl ectance was only possible with the introduction of the FT technique. Refl ected 
radiation from powders is too low to be measured at medium resolutions or inadequate 
signal - to - noise ratios. Diffuse refl ectance (Figure 10 ) occurs when a beam of 
radiation strikes the surface of a fi nely divided powder. With this type of sample, 
specular refl ection occurs at each plane surface. However, since there are many of 
these surfaces and they are randomly oriented, radiation is refl ected in all directions. 
The intensity of the refl ected radiation is independent of the viewing angle. If peak 
locations are identical in refl ectance and transmittance spectra, relative peak heights 
differ considerably. For example, minor transmittance peaks generally appear larger 
in refl ectance spectra. 
Internal - refl ection spectroscopy is used to obtain IR spectra of hard - to - handle 
or hard - to - prepare samples such as solids with limited solubility, fi lms, pastes, adhesives, 
and powders. Refl ection occurs when a beam of radiation passes from a denser 
to a less dense medium. The fraction of incident beam which is refl ected increases 
as the angle of incidence becomes larger. Beyond a certain critical angle, refl ection 
is complete. During the refl ection process the beam penetrates a small distance into 
FIGURE 10 Diffuse refl ectance, transfl ectance, and transmittance measurements. 
Diffuse 
reflectance 
Transflectance 
Diffuse 
transmittance 
VIBRATIONAL SPECTROSCOPY 375

376 PROCESS ANALYTICAL TECHNOLOGY 
the less dense medium before refl ection occurs. The depth of penetration varies 
from a fraction of a wavelength up to several wavelengths and depends on the 
wavelength of incident radiation, the refraction indices of the two materials, and the 
angle of incident beam with respect to the interface. 
Attenuated total refl ection (ATR) is the most common refl ectance measurement 
modality. ATR spectra cannot be compared to absorption spectra. While the same 
peaks are observed, their relative intensities differ considerably. The absorbances 
depend on the angle of incidence, not on sample thickness, since the radiation penetrates 
only a few micrometers into the sample. The major advantage of ATR spectroscopy 
is ease of use with a wide variety of solid samples. The spectra are readily 
obtainable with a minimum of preparation: Samples are simply pressed against the 
dense ATR crystal. Plastics, rubbers, packaging materials, pastes, powders, solids, and 
dosage forms such as tablets can all be handled directly in a similar way. 
4.2.2.8 Raman Spectroscopy 
When radiation passes through a transparent medium, a fraction of the beam scatters 
in all directions. A small fraction of the scattered radiation differs from the 
incident beam, showing shifts in wavelength determined by the chemical structure 
of the molecules in the medium. The same types of quantized vibrational changes 
associated with IR absorption occur, and the difference in wavelengths between 
incident and scattered radiations corresponds to wavelengths in the MIR. The 
Raman scattering spectrum and IR absorption spectrum for a given species are very 
similar. Figure 11 illustrates a typical Raman spectrum. IR is generally the method 
of choice, but in some cases Raman spectroscopy offers more information about 
certain types of organic compounds. For example, it is sensitive to conformational 
FIGURE 11 Example plot of two Raman spectra (two polymorphic forms of an 
excipient). 
12,000 
14,000 
16,000 
18,000 
20,000 
22,000 
24,000 
26,000 
28,000 
30,000 
32,000 
34,000 
36,000 
38,000 
400 600 800 1000 1200 1400 1600 

and environmental information. Peak overlap in compound mixtures is less likely, 
and quantitative determinations are easier. In particular, accurate quantitative 
determination can be performed on very small samples. Despite these advantages, 
Raman spectroscopy has not yet been exploited due to the rather high cost of the 
instruments. 
There are differences between the kinds of groups that absorb in the IR and 
those that are Raman active. Parts of Raman and IR spectra are complementary, 
each being associated with a different set of vibrational modes within a molecule. 
Other vibrational modes may be both Raman and IR active. The intensity or power 
of a Raman peak depends in a complex way on the polarizability of the molecule, 
the intensity of the source, and the concentration of the active group, as well as 
other factors. Raman intensities are usually directly proportional to the concentration 
of the active species. 
In Raman spectroscopy, the excitation radiation occurs at a wavelength distant 
from any absorption peaks of the analyte. The mechanism which leads to Raman 
spectra is different from that of MIR spectra, although dependent upon the same 
vibrational modes. IR absorption requires a change in dipole moment or its associated 
charge distribution. Only then can radiation of the same frequency interact 
with the molecule to promote an excited vibrational state. In contrast, scattering 
involves momentary distortion of the electron cloud distributed around a bond in 
a molecule, followed by reemission of the radiation as the bond returns to its normal 
state. In the distorted form the molecule is temporarily polarized. A dipole is 
momentarily induced which disappears upon relaxation and reemission. Thus, the 
Raman activity of a given vibrational mode may differ markedly from its IR activity. 
For example, a homonuclear molecule has no dipole moment either in equilibrium 
or when stretched, and IR absorption of radiation at the exciting frequency cannot 
occur. On the other hand, the polarizability of the bond between the two atoms of 
such a molecule varies periodically in phase with the stretching vibrations, reaching 
a maximum at the greatest separation and a minimum at the closest position. A 
Raman shift corresponding in frequency to that of the vibrational mode results. 
Raman shift magnitude is independent of excitation wavelength. Thus shift patterns 
are identical regardless of the laser used for excitation. 
In MIR spectroscopy water always causes interference, which is not the case with 
Raman scattering. Thus, Raman spectra can be obtained directly from aqueous solutions. 
In addition, glass or quartz cells can be employed. The development of Raman 
spectroscopy was closely associated with the availability of readily usable laser 
beams. Raman spectra are obtained by irradiating the sample with a laser source of 
visible or NIR monochromatic radiation. During irradiation, the scattered radiation 
is acquired at some angle (e.g., 90 ° ) with a suitable device. Raman lines are 0.001% 
or less intense than the source and more diffi cult to detect than IR spectra. Raman 
measurement can be restricted by fl uorescence or impurities in the sample. This 
problem has been partly solved by the use of NIR laser sources, which operate at 
longer wavelengths. Much higher power can irradiate the sample without causing 
photodecomposition or simply heating. NIR lasers are not energetic enough to 
populate a signifi cant number of fl uorescence - producing excited electronic energy 
states in most molecules. Fluorescence is less intense or nearly nonexistent. 
Scattered radiation is of three types: Stokes, anti - Stokes, and Rayleigh. The 
wavelength of Rayleigh scattering is identical to that of the excitation source and 
VIBRATIONAL SPECTROSCOPY 377

378 PROCESS ANALYTICAL TECHNOLOGY 
signifi cantly more intense than either of the other two types. By convention Raman 
spectra are plotted with the abscissa defi ned as the difference in wavenumbers 
between the observed radiation and that of the source. As anti - Stokes lines are less 
intense than the corresponding Stokes lines, only this part of the spectrum is used. 
However, when fl uorescence occurs and interferes with the observations of the 
Stokes lines, the anti - Stokes part of the spectrum proves helpful, despite the lower 
intensities involved. 
4.2.2.9 Introducing NIRS 
Until recently the NIR region was not considered particularly useful for the spectroscopy 
of organic compounds. The combination and overtone bands of molecular 
vibrations occur in the relatively narrow region of 750 – 3000 nm compared to the 
fundamental bands occurring at 2800 – 50,000 nm. This was considered as a drawback. 
In addition it was observed that NIR bands severely overlapped, were diffi cult to 
resolve, and, once resolved, diffi cult to interpret. If samples were not dried prior to 
NIR analysis, the changes in hydrogen bonding due to the effects of sample temperature, 
ionic strength, and analyte concentration could also complicate interpretation 
of frequently overlapping NIR spectral bands. Changes in hydrogen bonding 
produce band shifts as well as fl attening or broadening of band shapes. The overtone 
and combination molecular absorptions found within the NIR region are inherently 
much less intense than the fundamental IR absorptions. Thus, the changes in absorbance 
in the NIR region are quite small with respect to changes in concentration. 
The relatively small extinction coeffi cients due to combination and overtone NIR 
bands severely restrict the allowable noise levels and stability of any NIR instrument 
that is used for quantitative work. Sample presentation and the relatively 
straightforward refl ectance measurement involved aspects which were not expected 
with traditional IR spectroscopy. 
If analytical information is obtained with better resolution in the IR region, why 
should a chemist be interested in NIRS? Numerous diffi culties inherent in the use 
of qualitative NIRS led to its rejection by analytical chemists. Karl Norris, an engineer 
at the U.S. Department of Agriculture, demonstrated the potential value of the 
NIR region for quantitative work by making measurements of agricultural products 
in the 1960s. The basic idea was to provide various research and production facilities 
with online NIR measurements of agricultural products. NIRS has now become 
widespread in the chemical and pharmaceutical industry, with the publication of 
multiple practical applications and a massively increased presence in specialized 
journals. 
4.2.2.10 Benefi ts of NIRS 
The NIR region is of great interest for pharmaceutical applications. NIRS is fast, 
nondestructive, and cost effective. Samples require no preparation and can be measured 
as such, intact and available for further analysis. NIRS can be performed in - , 
on - , and offl ine. Also, glass fi ber optics can be used to perform remote analysis, thus 
bringing radiation directly to the sample. Many more advantages can be cited when 
considering the practical use of NIR in a pharmaceutical process, depending on the 
particular objective. 

As already mentioned, absorption bands in this region are overtones or combination 
bands of fundamental stretching bands that occur in the 3000 – 1700 cm . 1 region 
(Figure 12 ). The bonds involved are usually C — H, N — H, and O — H. Because the 
bands are overtones or combinations, their molar absorptivities are low and detection 
limits are around 0.1%. 
Sample handling is simplifi ed as glass can be used for windows, lenses, and any 
other optical components. In addition, the laser source is easily focused on small 
sample area. Very small samples can be investigated without time - consuming preparation. 
It is also possible for the source radiation to be transmitted through optical 
fi bers. The fi ber - optic probe can be in contact with the sample or immersed in it. 
The probe consists of input fi bers surrounded by several collection fi bers that transport 
the scattered radiation to the monochromator. This makes it possible to collect 
spectra directly under relatively adverse conditions. 
In contrast to its MIR counterpart, an important application of NIRS is the 
routine quantitative determination of species, such as water, proteins, hydrocarbons, 
and fats, for example, in food or feed products, but also in the petroleum and chemical 
industries. Figure 13 illustrates a collection of spectra from a pharmaceutical 
product with varying water content. It shows that a quantitative application can be 
FIGURE 12 Typical NIR spectrum. 
0 
0,5
1 
1,5
2 
2,5
3
700 900 1100 1300 1500 1700 1900 2100 2300 2500 
Wavelength (nm) 
log(1/R) 
First overtone Third overtone 
Second overtone Combinations 
FIGURE 13 Collection of NIR spectra. The varying water band near 1950 nm can be clearly 
identifi ed. 
Absorbance 
1400 1500 1600 1700 1800 1900 2000 
-0.5
0 
0.5
1 
1.5
2 
Wavelength (nm) 
1950 nm 
VIBRATIONAL SPECTROSCOPY 379

380 PROCESS ANALYTICAL TECHNOLOGY 
developed once a given physical parameter varies, which can then be specifi cally 
modeled. The spread of the NIR technique in the pharmaceutical industry was 
encouraged by advances in computer technology and data handling. Both diffuse 
refl ectance and transmittance measurements (Figure 10 ) are used, although diffuse 
refl ectance much more widely because of its ease of use. Many spectrometers were 
specifi cally designed for the NIR range. Instrument variety is wide and much more 
varied than in the MIR range. The most sophisticated are dual - beam grating, diode - 
array, and FT instruments. Simpler fi lter instruments are still available and remain 
valuable tools. These instruments can basically be viewed as similar to those used 
for UV/Vis analysis. Tungsten – halogen lamps with quartz windows serve as sources. 
Detectors are usually lead sulfi de (PbS) or arsine – gallium (AsGa) detectors and 
many instruments are designed to operate from 180 to 2500 nm. 
The NIR spectra are less useful for identifi cation and more useful for quantitative 
analysis of compounds containing functional groups made up of hydrogen bonded 
to carbon, nitrogen, or oxygen. Such compounds can often be determined with an 
accuracy and precision equivalent to UV/Vis spectroscopy, rather than with MIR 
spectroscopy. Many currently implemented applications concern the determination 
of water in variety of samples. Refl ectance NIRS has become a reliable tool for the 
quantitative determination of constituents in solids. The fi rst use of this fast technique 
was for the determination of protein, moisture, starch, oil, lipids, and cellulose 
in feed and food products. Sample handling, measurement procedure, and data 
treatment were established with these fi rst applications. Typically in NIR the solid 
sample is irradiated with one or more narrow bands or the full range of radiation 
from 1 to 2.5 . m. Diffuse refl ectance occurs in which the radiation penetrates the 
surface layer of the particles, excites the vibrational modes of the analyte molecule, 
and is then scattered in all directions. A refl ectance spectrum is thus produced that 
is dependent upon the sample composition. In this case the ordinate is the logarithm 
of the reciprocal of refl ectance R , log(1/ R ), where R is the ratio of the intensity of 
radiation refl ected from the sample to refl ectance from a standard refl ector, such as 
fi nely ground barium sulfate or magnesium oxide. The typical refl ectance band at 
1940 nm is a water peak used for moisture determination, as can be seen in Figure 
13 . 
Many diffuse - refl ectance instruments are available. Some employ several interference 
fi lters to provide narrow bands of radiation. Others are equipped with 
grating monochromators. Ordinarily, calibration is often a stringent requirement as 
samples must be acquired of the material for analysis that contain the range of 
analyte concentrations likely to be encountered. It may be useful to grind solid 
samples to a reproducible particle size. Equations are developed and used for the 
analysis. Once method development has been completed and validated, solid samples 
can be analyzed in a few minutes. Accuracy and precision are reported to be of 1 
to 2% relative. 
In addition to the somewhat empirical and diffi cult development of NIR applications, 
thorough documentation must be produced. NIR methods have to comply 
with the current good manufacturing practice (cGMP) requirements used in the 
pharmaceutical industry. Various regulatory aspects have to be carefully considered. 
For example, NIR applications in classifi cation, identifi cation, or quantifi cation 
require extensive model development and validation, a study of the risk impact of 
possible errors, a defi nition of model variables and measurement parameters, and 

comprehensive data analysis. Further, clearly defi ned operating procedures and user 
training are required for routine analysis. General regulatory requirements also 
require valid documentation to have been maintained on the life cycle of the NIR 
model, spectrometer, computer, and the like. 
4.2.2.11 Introducing MIR / NIR Chemical Imaging 
Chemical compound homogeneity is an important issue for pharmaceutical solid 
forms. A classical spectrometer integrates the spatial information. In solid form 
analysis, use of a mean spectrum on a surface can be a drawback. For example, in 
the pharmaceutical industry it is important to map the distribution of active ingredients 
and excipients in a tablet so as to reveal physical interaction between the 
compounds and help to solve homogeneity issues. Spectroscopic imaging techniques 
that visualize chemical component distribution are thus of great interest to the 
pharmaceutical community. 
Vibrational hyperspectral imaging is the most recent development to combine 
chemical information from spectroscopy with spatial information in the sample. In 
principle, hyperspectral images can be collected using single - point detectors, that is, 
classical scanning or mapping with microscopes. The commercialization of FPA 
detectors promoted imaging using a more rewarding analytical method. Array 
detectors with multiple detector elements measure all pixels on the mapped surface 
simultaneously, thus drastically reducing recording time, and provide a uniform 
background to improve the signal - to - noise ratio. A complete spectrum is acquired 
for each pixel, meaning that a hyperspectral image is in fact a data cube, that is, a 
three - dimensional (3D) matrix. Hyperspectral imaging provides spatial and spectral 
information as well as qualitative and quantitative information. 
Imaging techniques are typically effective when applied to pharmaceutical tablets 
in order to explore qualitative aspects or visually address process issues: dissolution, 
polymorph distribution, moisture content determination, active pharmaceutical 
ingredient (API) localization and characterization, content uniformity, blending, 
and granulation. Conventional spectroscopy cannot provide this kind of information. 
Chemical compound identifi cation or particle size determination can also be 
estimated by imaging. Selection of an imaging technique will depend on several 
criteria, such as spatial and spectral resolutions, time of measurement, and wavelength 
range. 
The CISs are rapidly becoming more popular and reliable as their fi eld of application 
broadens. This is mainly due to the production of surface images by multipoint 
scanning and mapping. Hyperspectral imaging has proven its potential for 
qualitative analysis of pharmaceutical products and can be used when spatial information 
becomes relevant for an analytical application. Even if online applications 
and regulatory method validation require further development, the power of CIS 
in quality control and PAT needs no further demonstration, whatever the wavelength 
domain or method of spectra collection. 
4.2.2.12 Design of MIR Instruments 
Different types of IR instruments are available: dispersive grating spectrophotometers, 
FT instruments, and nondispersive photometers. Until the 1980s, and the 
VIBRATIONAL SPECTROSCOPY 381

382 PROCESS ANALYTICAL TECHNOLOGY 
introduction of more reliable interferometers, the dispersive type was the most 
widely used. FT spectrophotometers are preferred for many applications because 
of their speed, reliability, and convenience. They have largely displaced other equipments 
in the analytical laboratories. Source radiation is split into two beams whose 
path lengths can be varied periodically to produce interference patterns. The reversed 
FT algorithm is then used for data recovery. Early FT spectrometers required frequent 
adjustments of the optical parts, but their unique characteristics of speed, high 
resolution, sensitivity, and wavelength precision and accuracy made them indispensable. 
Recent instrument design is much more reliable and easier to adjust. Typically, 
FT spectrometers are based on the Michelson interferometer (Figure 14 ), although 
other types of optical systems are also encountered. Laboratory FT spectrometers 
are usually proposed as single - beam instruments. For determining transmittance or 
absorbance spectra with this type of instrument a reference interferogram, air, for 
example, must fi rst be measured and stored. The analyzed sample is then placed in 
the radiation path and the process repeated. Transmittance at various frequencies 
is obtained by computing the ratio of the sample and reference spectra. 
Over most of the MIR spectral range, FT instruments appear to have better 
signal - to - noise ratios than do good - quality dispersive instruments by more than one 
order of magnitude. The enhanced signal - to - noise ratio can, of course, be traded for 
rapid scanning, producing spectra within a few seconds. Interferometric instruments 
also feature high resolutions (less than 0.1 cm . 1 ), high accuracy, and reproducible 
frequency determinations. FT instruments provide a much larger energy throughput 
(one or two orders of magnitude) than do dispersive instruments, which are limited 
in throughput by the necessity of narrow slit widths. However, this advantage is 
partially offset by the lower sensitivity of fast - response detectors required for 
interferometry. 
To perform chemical imaging of sample surfaces, FT spectrometers can be coupled 
with a microscope or macrochamber with an FPA detector. CIS are available for 
Raman, NIR, and MIR spectroscopy. Figure 15 illustrates an optical arrangement 
for chemical imaging. 
Instrument design depends on how measurements are performed (Figure 10 ). 
The ideal instrument has both transmittance and refl ectance capabilities. For 
FIGURE 14 Michelson interferometer. 
Lens 
Lens 
Moving mirror 
Fixed mirror 
Beamsplitter 
Source 
Sample

example, refl ectance measurements penetrate only 1 – 4 mm into the surface of solid 
samples. This shallow penetration of energy into a sample brings greater variation 
when measuring nonhomogeneous samples than transmittance techniques. In transmittance 
measurements, the entire path through the sample is integrated into the 
spectral measurement, thereby reducing error due to sample nonhomogeneity. 
Transmittance is suitable for measuring through compact samples, like intact tablets, 
but surface scattering induces a loss of transmitted energy with the net effect being 
a decrease in the signal - to - noise ratio. In some circumstances, particle size is so small 
that most of the energy striking the sample is scattered back. If the particle size is 
small enough, the instrument will not transmit enough energy through the sample 
for the detectors to record a signal. In transmittance, higher frequencies (800 – 
1400 nm) are used to increase the depth of penetration into the sample. But higher 
frequency energy is more susceptible to surface scattering than lower frequency 
energy. Transmittance measurements must therefore be optimized based on the 
relationships between the frequencies used for measurement, surface scatter, and 
sample path length. 
The fi rst NIR spectrometers utilized a tilting fi lter concept. This concept was 
refi ned into a spinning system with a range of fi lters mounted in an encoder wheel 
for greater positioning accuracy (wavelength reproducibility) and greater reliability. 
The use of interference fi lters represents another type of instrument design, for 
which prespecifi ed discrete interference fi lters are manufactured. The fi lters are 
mounted in a turret and rotated slowly to different positions during measurement 
scanning. These instruments are rugged, provided the calibrations are well established 
and specifi c to the analyte concerned. If the wrong interference fi lters are 
selected for the specifi c application, successful calibration is impossible. Systems 
currently exist confi gured from 6 to 44 discrete wavelength interference fi lters, in 
particular for routine application in the food and feed industries. 
Dispersive, grating, scanning NIR instruments have been available since the late 
1970s. These instruments varied in optical design, but all shared tungsten – halogen 
source lamps, a single monochromator with a holographic diffraction grating, and 
uncooled lead sulfi de detectors. This design dates back to the early 1980s. Figure 16 
FIGURE 15 IR microscope coupled with a FPA detector. 
Spectrometer 
Microscope FPAdetector 
Sample 
MCT array 
IR light 
VIBRATIONAL SPECTROSCOPY 383

384 PROCESS ANALYTICAL TECHNOLOGY 
shows a typical predispersive monochromator - based instrument where the light is 
dispersed before striking the sample, with a detector system for diffuse refl ectance. 
The monochromatic light beam illuminates the sample at 0 ° (normal incidence), and 
the detectors collect the refl ected light at 45 ° . Two to six detectors can be used, 
generally lead sulfi de detectors for measurement in the 1100 - to 2500 - nm region. A 
computer is required for data processing, calibration, and storage. The spectrum is 
the difference between the raw refl ectance measurement of the sample and the raw 
refl ectance measurement of the reference material. Refl ectance is converted to 
absorbance using the function absorbance = . log 10 (refl ectance), written as log 1/ R . 
Raw transmittance is converted to absorbance using the expression log 1/ T . 
Postdispersive design is an optical arrangement allowing the transmission of 
more energy in either a single fi ber - optic strand or fi ber - optic bundle. Radiation is 
piped through the fi ber - optic strand or bundle where it strikes the sample and 
returns to the dispersive element, a grating. After striking the grating, the light is 
separated into the various wavelengths before striking the detector(s). 
The integrating sphere is still a common sample presentation geometry for NIR 
measurements. The use of integrating spheres dates back to the fi rst commercial 
photometers. Their greatest advantage is that a detector placed at an exit port of 
the sphere is unlikely to lose energy. However, as detector technology improves, the 
advantages of integrating spheres for energy collection are no longer critical. Some 
special applications still require the use of spheres attached to fi ber - optic bundles. 
The use of a sphere allows internal photometric referencing — a sort of double - beam 
instrument. The application of FT techniques to NIR dates back to the late 1980s. 
Transmittance/diffuse - refl ectance FT - NIR instruments are widespread. As already 
mentioned, FT spectrometers differ from scanning spectrometers in that the recorded 
signal is an interferogram (details in the MIR section of this chapter). Other designs 
include diode array detectors (Figure 17 ) and NIR emitting diode sources. Acousto - 
optic tunable fi lters (AOTF) are devices based on diffraction. The NIR fi lter is in 
fact a transparent crystal in which an ultrasonic wave fi eld is created. Thus, the 
selected wavelength is a function of the fi eld intensity. AOTF scanning speed is 
FIGURE 16 Predispersive grating spectrometer for diffuse refl ectance. 
Sample 
Mirror 
Source 
Holographic 
network 
Diffracted light 
Polychromatic 
light 
Reference 
Detector 
Detector

measured in microseconds. This kind of spectrometer is considered very fast but 
requires extra care to perform reliably in a pharmaceutical process environment. 
Other techniques are available, such as ultrafast - spinning interference fi lter wheels, 
interferometers with no moving parts, and tunable laser sources. 
4.2.2.13 Conclusion 
To summarize, instruments for vibrational spectroscopy can be categorized by the 
optimization of their optical design for a specifi c type of sample presentation and 
for the solution of a multiplicity of measurement problems. For development purposes, 
the instrument of choice is generally one with broad capabilities. Once an 
application has been clearly defi ned, an instrument suited to the specifi c sample 
presentation geometry or sample type will give the most reliable results. However, 
at this stage it is worth mentioning that due to the optical design and the mode of 
measurement, it can be very tricky to transfer calibrations or analytical methods 
from one instrument to another. For example, calibration transfer is a recurrent 
issue in NIRS and concerns all instrument designs. This is important if planning to 
perform measurements of a given sample attribute at different locations with different 
instruments, for example, laboratory versus process. The issue is more stringent 
with complex pharmaceutical samples where the signal from the analyte or 
property of interest strongly interacts with other parameters. 
The choice of a given spectroscopy, that is, MIR, NIR, or Raman, has great 
importance for improving the monitoring and understanding of pharmaceutical 
processes. In Table 2 the major characteristics of use of MIR, NIR, and Raman 
spectroscopy are summarized. The most appropriate strategy is implemented according 
to the necessity and requirements of the manufacturing or analytical operations. 
In the selection of the spectroscopic technique, attention has to focus on criteria 
like chemical information content, location of the measurement, speed or real - time 
ability, robustness, ease of use, sample preparation, sample destruction, and the like. 
Multiconstituent analysis requires levels of accuracy which are comparable to those 
of the primary reference methods. Nowadays, a large number of technologies and 
numerous manufacturers exist. With this in mind, a given choice of equipment will 
probably match most signifi cant criteria for the analytical problem of interest and 
FIGURE 17 Diode - array spectrometer. 
Diode array 
Grating 
Slit of entry 
Diffracted light 
Sample 
Source 
VIBRATIONAL SPECTROSCOPY 385

386 PROCESS ANALYTICAL TECHNOLOGY 
at the same time require to accept some inherent measurement limits or restrictions 
in use. 
4.2.3 CHEMOMETRICS 
4.2.3.1 Introduction 
Chemometrics is a term which defi nes a discipline of chemistry involving mathematics 
and computer science in order to derive information from data of various type, 
origin, and complexity. Typical applications include relating the concentration of 
some analyte found in a sample to the sample ’ s spectral data or identifying some 
physical or chemical characteristics of the sample. Although performing chemometrics 
without a computer is effectively impossible, the basic multivariate mathematic 
calculation has been known since the early twentieth century. 
Near - IR spectroscopy is a typical application for chemometrics, for example, how 
to relate the NIR spectra, obtained at little expense, of the contents of a closed 
sterile vial with measurement of the concentration of a particular compound in the 
powdered mixture found by a reference method without opening the sample. To 
achieve this, the relationship between two data matrices has to be found and a 
quantitative calibration calculated. This task is a little different depending on 
whether the data has been generated using statistical experimental design (i.e., 
designed data) or has simply been collected, more or less at random, from a given 
population (i.e., nondesigned data). With designed data matrices, the variables are 
orthogonal by construction. Traditional statistical methods such as analysis of variance 
(ANOVA) and multiple linear regression (MLR) are well suited to fi nd the 
regression model in orthogonal data tables. In nondesigned data matrices the 
variables are seldom orthogonal but are more or less collinear. MLR is likely to 
fail in these circumstances, and the use of projection techniques such as principal - 
component regression (PCR) or partial least squares (PLS) is recommended. A 
regression model can then be used to predict new, that is, unknown, samples. Prediction 
is a useful technique because it can be used in place of costly and time - 
consuming measurements. 
Another example is the noninvasive identifi cation of species from NIR absorbance 
spectra. Classifi cation is simply a matter of fi nding out whether new samples 
TABLE 2 Comprison of MIR/NIR Roman Spectroscopic Techniques 
Spectroscopy NIR IR Raman 
Spectral range (cm . 1 ) 12,000–4,000 4,000–400 4,000–50 
Signal intensity ++ +++ + 
Microscopic analysis No Yes Yes 
Fiber - optic interfacing Yes Yes (limited length) Yes 
Sampling through glass Yes No Yes 
Qualitative application Yes Yes Yes 
Quantitative application Yes Diffi cult Yes 
Instrument robustness +++ + ++ 
Chemical Interpretation Chemometrics Direct Direct 

are similar to classes of samples that have been used to make the model. If a new 
sample fi ts a particular model well, it is said to be a member of that class. Many 
analytical tasks fall into this category. For example, raw materials like excipients 
may be sorted according to “ good and bad quality, ” fi nished products classifi ed into 
“ grades A, B, C, ” and so on. Principal - component analysis (PCA) is a typical mathematical 
procedure for resolving such sets of data into orthogonal components 
whose linear combinations approximate the original data with a desired degree of 
accuracy. As successive components are calculated, each accounts for the maximum 
possible amount of residual variance in the set of data. In NIRS, the data usually 
consist of large sets of recorded spectra, and the number of components will be 
smaller than or equal to the number of known variables or the number of spectra. 
4.2.3.2 From Univariate to Multivariate Regression 
In spectroscopy the simplest method of quantitative calibration is based on a single 
independent variable, for example, wavelength, since a sample attribute such as 
analyte concentration is a linear function of absorbance at a given wavelength. 
Modeling of the concentration requires conventional least - squares fi tting. A straight 
line is fi tted through a set of data points to minimize the sum of squares of deviations 
between estimated and known data points. In this approach, a wavelength is 
selected when it shows a high degree of correlation between concentration and 
absorbance. Correlation is an indicator of goodness of fi t between concentration 
and absorbance and of how well the calibration describes the data set. The linear 
relationship has a practical advantage in that it permits direct and visual estimation 
of goodness of fi t, thus enhancing the analyst ’ s trust in his data collection. Where 
pharmaceutical samples are concerned, the linear approach rapidly reaches its limits 
and another approach is required. 
Multiple linear regression extends linear regression to one wavelength by least 
squares, with more than one wavelength selected to perform a calibration. The 
complexity of the test samples is again an issue with this approach. MLR requires 
independent variables for its ability to explain the data set. As pharmaceutical 
samples comprise a complex matrix in which species interact to various degrees in 
the NIR range, it is impossible to fi nd appropriate wavelengths to select. Observed 
absorbance values are linked as they may describe related behaviors in the spectral 
data set. It is typical of NIR spectra of mixtures that collinearity is found among 
the wavelengths and MLR will never perform a usable linear calibration. 
Partial least squares multivariate regression uses many wavelengths without the 
limitation due to collinearity. In fact, collinearity makes PLS the tool of choice when 
considering segments or entire NIR spectra. PLS treats constituent data and spectral 
data on the same basis, without statistical prerequisites of standard distribution or 
noncollinearity. Errors are assumed equally and in both data sets. Spectral and 
constituent data are modeled simultaneously according to an iterative algorithm. 
This training step is required for the model to learn to predict the sample attribute 
of interest. For each iteration a part of the spectral data and a part of the corresponding 
constituent data are combined until the data set is optimally described. 
The nonexplained part of the data set is made up of residuals, which are a measure 
of the quality of the modeling. The original data is combined in factors or principal 
components. Coeffi cients are calculated — loadings for spectral data and scores for 
CHEMOMETRICS 387

388 PROCESS ANALYTICAL TECHNOLOGY 
constituent data — to indicate the extent of original data involvement in the computation 
of each factor. Finally, the amount of variance modeled, that is, the explained 
part of the data, is maximized for each factor and the residuals are minimized. 
The raw data is normally preprocessed to some extent before PLS calibration, as 
illustrated below. A critical step in PLS modeling is the selection of the number of 
factors. Selecting too few factors will provide an inadequate explanation of variability 
in the training data set, while too many factors will cause overfi tting and 
instability in the resulting calibration. Optimal factor number is estimated during 
validation of the calibration, which is part of the PLS algorithm. Either an additional 
and independent data set is used (external validation in one computation step) or 
the training data set is split into subsets for continuous internal validation at each 
iterative step (cross - validation). In addition, the resulting PLS calibration must 
prove able to predict unknown samples. It is also good practice to challenge the 
resulting calibration in a way that enhances trust in future results. In no way are 
PLS calibrations black boxes despite the fact that visualization is far from evident 
when compared to basic linear regression. 
4.2.3.3 Sample Quality and Data Error 
In an ideal case, a given resulting calibration equation provides an explicit value for 
the concentration of an unknown analyte in terms solely of measurable absorbances. 
Only the concentrations of the desired analyte need be known during calibration. 
The values for other sample constituents need not be determined, although in order 
to develop a good calibration equation, variations of these values should also be 
evenly distributed. Indeed, in some cases, even the number of other compounds in 
the calibration samples may not be known. Using stepwise multiple regression, 
features in the spectra which correlate most closely with the analyte concentration 
are selected for a particular sample set. Once optimal calibrations are computed, 
the NIR instrument can be used to predict unknown samples for the determination 
of the quantity of desired analyte. Thus, regression analysis is a method to develop 
the relationship (i.e., regression calibration equation) between several spectral features 
and the constituent being investigated. For diffuse refl ectance measurements, 
the effects of extraneous physical phenomena are superimposed on the absorbance 
readings and have to be taken into account. The equations can be formulated to 
account for these phenomena by implicitly including corrections for their effects in 
the calibration coeffi cients. 
One problem that chemometrics cannot deal with is the appropriate selection or 
preparation of samples. Careful sample selection increases the likelihood of extracting 
useful information from spectral data. Whenever it is possible to actively 
experiment with selected variables or parameters, the quality of the results is 
increased. The critical part is deciding which variables can be changed and what 
limits to fi x on their variation. Appreciation of which parameters may infl uence data 
depends mainly on the experience of NIR operators but is also a matter for specialists. 
Each problem will generate and implicate a specifi c set of more or less usable 
variables. The key to success lies in gaining experience in dealing with numerous 
and different problems. Experimental design could ideally help to generate the 
data indicating which parameters or design variables infl uence output or response 
variables. 

Understanding the interactions between design variables while identifying the 
optimal conditions for extracting the relevant information should in principle minimize 
the number of experiments required, thus reducing cost. An experimental 
design program offers appropriate design methods and encourages good experimental 
practice by performing a small number of useful experiments that span the 
important variations. Experimental design will not be discussed here, but only the 
way in which data should be collected and organized in practice for multivariate 
regression. Typically, with spectral data, the problem is to fi nd out which variables 
are really important for variation in the data matrix. The usual tasks when thinking 
about modeling a response variable are to fi nd out which variables are necessary 
for describing the samples adequately, which samples are similar to each other, and 
whether the data set contains groups of samples. A good way of fi nding this information 
is by decomposing the spectral data matrix into a structured part and a noise 
part, typically by using PCA. Another problem is how to establish a regression 
model between a spectral data matrix and a response variable matrix. 
Before performing multivariate data analysis, statistical analysis of sample 
response values or spectral data may help to check data quality. Descriptive statistics 
summarize the distribution of one or two variables at a time. They are not supposed 
to say much about the data structure, but they are useful for obtaining a quick look 
at each separate variable before starting an analysis. One - way statistics, that is, 
mean, standard deviation, variance, median, minimum, maximum, lower, and upper 
quartile, can be used to explore the data set and detect out - of - range values, abnormal 
spread, or asymmetry. These statistics reveal anything suspect in a data table 
and indicate whether a transformation might be useful. Two - way statistics, for 
example, correlation, show how variations in two variables may be linked in a data 
table. Checking these statistics is also useful to detect out - of - range values and 
outliers. 
Experience in NIRS practice shows that the critical step(s) for successful implementation 
vary with each specifi c problem to solve. However, the following procedure 
is recommended when analyzing nondesigned data. First, investigate the origin 
and availability of the data. In formulating the study problem, defi ne the precise 
objective of data collection and the expected analysis results. The data collection 
should span appropriate variation in the explored variable(s) or attribute(s). If the 
naturally available data do not span the expected variation, prepare or measure 
samples with corresponding experimental data. Raw spectra may have to be transformed 
and mathematical pretreatments performed. Calibrate and validate the 
model either using PCA or PLS . Explore and challenge the way the calibrations 
behave on real samples or data, for example, validate the method according to 
current pharmaceutical usage and requirements. 
There are three types of data error: random error in the reference laboratory 
values, random error in the optical data, and systematic error in the relationship 
between the two. The proper approach to data error depends on whether the 
affected variables are reference values or spectroscopic data. Calibrations are 
usually performed empirically and are problem specifi c. In this situation, the question 
of data error becomes an important issue. However, it is diffi cult to decide if 
the spectroscopic error is greater than the reference laboratory method error, or 
vice versa. The noise of current NIR instrumentation is usually lower than almost 
anything else in the calibration. The total error of spectroscopic data includes 
CHEMOMETRICS 389

390 PROCESS ANALYTICAL TECHNOLOGY 
sample - induced errors which can be much greater than the instrument ’ s noise level. 
Such errors include particle size effects, variations in the packing density of powders, 
effects of impurities, and effects due to changing physical characteristics of the 
sample (e.g., crystallinity). Practical experience shows that in many cases sample - 
induced error in the optical data remains small, prompting the empirical assumption 
that optical data error is always smaller than reference laboratory error. 
4.2.3.4 Mathematical Preprocessing of Spectroscopic Data 
As discussed above, the greatest source of error in NIR calibration is usually reference 
laboratory error, sample nonhomogeneity, and nonrepresentative sampling in 
the learning (training) set or calibration set population. Instrument quality and 
equation selection account for only a fraction of the variance or error attributable 
to NIR analytical technique in current routine application. 
When radiation is refl ected from solid matter surfaces, diffuse and specular 
refl ected energies are superimposed. The intensity of diffusely refl ected energy 
depends on the angles of incidence and observation but also on sample packing 
density, crystalline structure, refractive index, particle size distribution, and absorptive 
qualities. Thus, in practice, an ideal diffusely refl ecting surface can only be 
approximated, even with the fi nest possible grinding of the samples. There are 
always coherently refl ecting surface regions acting as elementary mirrors whose 
refl ection obeys Fresnel ’ s formulas. Radiation returning to the sample surface 
from the interior can be assumed to be largely isotropic, thus meeting the requirements 
of the Beer – Lambert law. It is assumed that radiant energy is continuously 
removed from the incident NIR beam and converted to thermal vibrational energy 
of atoms and molecules. Decrease in the intensity of the diffusely refl ected light 
depends on the absorption coeffi cient of the sample. The absorption coeffi cient ( K ), 
when taken as the ratio K/S , where S is the scattering coeffi cient, is proportional to 
the quantity of absorbing material in the sample. In the Kubelka – Munk theory, 
refl ectance ( R ) is related to absorption ( K ) and the scattering coeffi cient ( S ) by the 
equation 
K
S 
R
R 
F R = 
. ( ) 
= ( ) 
1 
2 
2 
Diffuse refl ectance R is a function of the ratio K/S and proportional to the addition 
of the absorbing species in the refl ecting sample medium. In NIR practice, absolute 
refl ectance R is replaced by the ratio of the intensity of radiation refl ected from the 
sample and the intensity of that refl ected from a reference material, that is, a ceramic 
disk. Thus, R depends on the analyte concentration. The assumption that the diffuse 
refl ectance of an incident beam of radiation is directly proportional to the quantity 
of absorbing species interacting with the incident beam is based on these relationships. 
Like Beer ’ s law, the Kubelka – Munk equation is limited to weak absorptions, 
such as those observed in the NIR range. However, in practice there is no need to 
assume a linear relationship between NIRS data and the constituent concentration, 
as data transformations or pretreatments are used to linearize the refl ectance data. 
The most used linear transforms include log 1/ R and Kubelka – Munk as mathemati

cal pretreatments. To some extent PCR, PLS, and multilinear regression compensate 
for nonlinearity. Calibration equations can be developed which compensate to some 
extent for the nonlinear relationship between analyte concentrations and log 1/ R 
or Kubelka – Munk transformed data. If a matrix absorbs at different wavelengths 
than the analyte, Kubelka - Munk can prove a useful linearization method for spectroscopic 
data. If the matrix absorbs at the same wavelength as the analyte, log 1/ R 
will prove a better choice for relating refl ectance to concentration. This transformation 
is basically well suited for diffuse refl ectance NIR spectra of most mixtures with 
absorbing matrices. Plots of F ( R ) versus concentration are less linear than plots of 
log 1/ R versus concentration. 
4.2.3.5 Preprocessing NIR Data 
When generating calibration equations using samples of known composition, the 
independent variable is represented by the spectroscopic readings (i.e., log 1/ R ) at 
specifi c wavelengths, while the concentration of the analyte of interest, for example, 
determined by traditional laboratory techniques, is the dependent variable. 
Spectral raw data may have a distribution or shape that is not optimal for analysis. 
Background effects, baseline shifts, measurements in different conditions, different 
variances in interfering variables and the like can complicate information extraction 
using multivariate methods. It is important to minimize the noise introduced by such 
effects. Preprocessing operations include centering, weighting, and numerous mathematical 
transformations. Mean centering consists of subtracting the average spectra 
from each individual spectrum. This ensures that all results will be interpretable in 
terms of variation around the mean considered as the model center. This is useful 
with models in which a linear relationship between spectral data and response data 
is supposed to go through the origin. Depending on the kind of information to be 
extracted from the spectral data, weights based on the standard deviation (i.e., 
square root of the variance, which expresses the variance in the same unit as the 
original variable) may be used for scaling. This may be a typical pretreatment for 
PCA, PLS, or PCR calibrations which are projection methods based on fi nding 
directions of maximum variation, thus depending on the relative variance of the 
variables. A possible weighting option is the 1/SD standardization, which gives all 
variables the same variance, that is, 1. In this case all variables are given the same 
chance to infl uence estimation of the components. This is advisable if the variables 
are measured with different units, have different ranges, or are of different types. It 
is also possible to fi x a constant weight for each variable manually. Weighting 
involves stretching and shrinking by measuring a position relative to the extremes 
in the actual spectral data table. By standardizing the spectra, variation in the data 
set is performed relative to the extremes in the data table. However, this procedure 
emphasizes the relative infl uence of unreliable or noisy attributes. 
4.2.3.6 Mathematical Pretreatment and Transformation 
A wide range of transformations can be applied to spectral data before they are 
analyzed. The main purpose of transformations is to make the latent variables better 
available for powerful analysis. One of the most widely used is logarithmic transformation, 
which is especially useful to make skewed variables more symmetrically 
CHEMOMETRICS 391

392 PROCESS ANALYTICAL TECHNOLOGY 
distributed. It is also indicated when the measurement error in a variable increases 
proportionally with the level of that variable. Taking the logarithm will achieve 
uniform precision over the whole range of variation. This particular application is 
also called variance stabilization. In case of limited asymmetries, a square root can 
be suffi cient. Smoothing is relevant for variables which are themselves a function 
of some underlying variable, for instance, of time. It is also one of the fi rst operations 
performed on recorded NIR spectra. It removes as much noise as possible without 
degrading important information content. Moving point average (MPA) is a classic 
smoothing method which replaces each observation with an average of the adjacent 
observations including it. The number of observations to average is a variable 
parameter. Polynomial smoothing, also called Savitzky – Golay smoothing, involves 
least - square fi tting of a polynomial equation to a window of n sequentially selected 
spectral data points. If the polynomial order is less than the number of data points, 
the polynomial cannot pass through all selected data points, and the least - square fi t 
gives a smoothed approximation to the original window. Normalization is a family 
of transformations which are computed sample wise. Here the purpose is to scale 
samples so as to improve specifi c properties. Mean normalization is the classic algorithm. 
It consists in dividing each row of a data matrix by its average, thus neutralizing 
the infl uence of possible hidden factors. It is equivalent to replacing the 
original variables by a profi le centered on 1. Only the relative values of the variables 
are used to describe the sample, and the information carried by their absolute level 
is dropped. It is indicated in the specifi c case where all variables are measured in 
the same unit, and their values are assumed to be proportional to a factor which 
cannot be directly taken into account in the analysis. Maximum normalization is an 
alternative procedure which divides each row by its maximum absolute value instead 
of the average. The maximum value becomes +1 and the minimum value becomes 
. 1. In range normalization, each row is divided by its range, that is, maximum value 
minus minimum value and the curve span becomes 1. 
More specifi c transformations for spectroscopic data are the refl ectance to absorbance, 
absorbance to refl ectance, and refl ectance to Kubelka – Munk transformations. 
Multiplicative scatter correction (MSC) is an additional transformation 
method used to compensate for additive and/or multiplicative effects in spectral 
data. MSC was originally designed to deal with multiplicative scattering alone. 
However, MSC successfully treats a number of similar effects such as path length 
problems, offset shifts, and interference. The idea behind MSC is to remove the two 
effects — amplifi cation, which is multiplicative, and offset, which is additive — from 
the spectral data table to prevent them from dominating the table ’ s information 
content. Derivation is typically relevant for spectral data that are a function of some 
underlying variable infl uencing absorbance at various wavelengths. Derivatives are 
also a simple but powerful technique for magnifying fi ne structure in raw spectra 
lacking structure, which is common in NIRS. By increasing the order of derivation, 
band structure resolution is increased. The inherent drawback is a decrease in the 
signal - to - noise ratio, which in the particular case of NIR spectra is not considered, 
since smoothing while performing a derivation may limit the effect of noise. On the 
other hand, it will tend to hide some weaker spectral features. The main advantage 
of the second derivative as usually performed is that band structure is maintained: 
Peak maxima still correspond while peak shape and resolution are improved. The 
Savitzky – Golay algorithm permits computation to higher order derivatives, includ

ing a smoothing factor which determines how many adjacent variables will be used 
to estimate the polynomial approximation for derivation. Norris derivation is an 
alternative algorithm for computing fi rst derivatives only. A baseline correction 
algorithm is the standard normal variate (SNV) method, which does not affect 
overall spectra layout. Averaging samples in case of replicates or variables for variable 
reduction, for example, to reduce the number of latent variables, is a method 
of obtaining more stable and more readily interpretable results. Figure 18 illustrates 
the effect of selected combined transformations to a collection of spectral data. 
4.2.3.7 Principal - Component Analysis 
Large data tables contain an amount of information which is partly hidden because 
the data complexity prevents ready interpretation. This is typical of NIR spectra 
collections. PCA is a projection method used to visualize all the information contained 
in the data table. It can be used to show in what respect one sample differs 
from another, which variables contribute most to this difference, and whether these 
variables contribute in the same way and are correlated or independent of each 
other. It also reveals sample patterns or groupings. In addition, it quantifi es the 
amount of useful information, as opposed to noise or meaningless variation, contained 
in the data table. Principal components are defi ned only for the data set from 
which they were computed. They may also hold for other data of identical type, but 
this is not guaranteed, and it is certainly not true for different types of data. 
FIGURE 18 Examples of combined spectral preprocessing. 
Standard normal variate (SNV) 
log(1/R) 
Wavelength 
(nm) 
SNV + detrending (SNVD) 
SNVD + 2nd derivative 
Raw NIR spectra 
Spectra 
# 
log(1/R) 
Wavelength 
(nm) 
Spectra 
# 
log(1/R) 
Wavelength 
(nm) 
Spectra 
# 
log(1/R) 
Wavelength 
(nm) 
Spectra 
# 
CHEMOMETRICS 393

394 PROCESS ANALYTICAL TECHNOLOGY 
The PCA modeling forms the basis for several classifi cation and regression 
methods. The underlying idea is to replace a complex multidimensional data set by 
a simpler version having fewer dimensions, but still fi tting the original data closely 
enough to be considered a good approximation. Extracting information from a data 
table consists of exploring variations between samples, that is, fi nding out what 
makes a sample different from, or similar to, another. Two samples can be described 
as similar if they have close values for most variables. From a geometric perspective, 
in the case of close coordinates in the multidimensional space of variables, the two 
points are located in the same area. Likewise, two samples can be described as different 
if their values differ greatly with respect to at least some variables. The two 
points have different coordinates and are located far away from each other in the 
multidimensional space. An illustration of this geometrical concept is proposed in 
Figure 19 . 
The principle of PCA consists of fi nding the directions in space — known as principal 
components (PCs) — along which the data points are furthest apart. It requires 
linear combinations of the initial variables that contribute most to making the 
samples different from each other. PCs are computed iteratively, with the fi rst PC 
carrying the most information, that is, the most explained variance, and the second 
PC carrying most of the residual information not taken into account by the previous 
PC, and so on. This process can go on until as many PCs have been computed as 
there are potential variables in the data table. At that point, all between - sample 
variation has been accounted for, and the PCs form a new set of axes having two 
FIGURE 19 Geometric illustration of PCA. Each spectrum is plotted as a point, whose 
coordinates are the intensities collected at three different wavelengths. For a given collection 
of spectra, the fi rst principal component is the direction in space which covers the greatest 
variation of the corresponding point set. 
B 
B 
A 
B' A' 
C' 
C 
rc 
PC1 
rb ra 
PC1 
A 
A2 
B3 
B2 
B1 
.2 
A1 
A3 
.3 
.1 
A1 
B1 
A2 
B2 
A
B 
.1 .2 .3 
.1 .2 .3 
A3 
A1 
A2 
A3 
B3 
B1
B2 
B3 
All spectra 
0.30 
0.25 
0.20 
0.15 
0.10 
0.05 
8000 9000 10000 11000

advantages over the original set. First, the PCs are orthogonal to each other. Second, 
they are ranked so that each carries more information than any of those following. 
It thus prioritizes their interpretation, starting with the fi rst PCs. This method of 
generation ensures that the new set of axes is more suitable for interpreting 
the data structure. Usually only the fi rst PCs contain pertinent information, with 
later PCs being more likely to describe noise. In practice, only the fi rst PCs are 
examined rather than the whole raw data table: Not only is it less complex, but it 
also ensures that noise is not mistaken for information. If all PCs were retained, 
there would be no approximation at all and no gain in simplicity either. Deciding 
on the number of components to retain in a PCA model is a compromise between 
simplicity, completeness, and effectiveness. The PCA model is only an approximation 
of reality. 
Each component of a PCA model is characterized by three complementary sets 
of attributes, that is, variances, loadings, and scores, respectively. The importance of 
a given PC is expressed by its variance. Loadings describe the relationships between 
variables, while scores describe sample properties. Variances are error measures 
which tell how much information is taken into account by successive PCs. The way 
they vary with the number of components can be studied to decide how complex 
the model should be. Residual variance designates the variation in the data that 
remains to be explained once the current PC has been taken into account, while 
explained variance, often measured as a percentage of total variance in the data, 
measures the proportion of variation in the data accounted for by the current PC. 
These variances can be considered either for a single variable or sample or for the 
whole data. They are computed as mean square variations, corrected for the remaining 
degrees of freedom. 
Loadings describe the data structure in terms of variable correlations. Each 
variable has a loading on each PC. This refl ects how much the variable contributed 
to that PC and how well that PC refl ects the variation of the variable considered. 
In geometric terms, loading is the cosine of the angle between the variable and 
the current PC, with a value that ranges by defi nition between . 1 and +1. The 
smaller the angle, the greater the link between variable and PC, and the greater the 
loading value. Variables with high loadings (i.e., close to +1 or . 1) for a given PC 
contribute greatly to the meaning of that particular PC. Thus, in studying correlations 
between variables, the loadings indicate their respective angles in the multidimensional 
space. For instance, if two variables have high loadings along the 
same PC, their angle is small, which in turn means that the two variables are 
highly correlated. If both loadings have the same sign, the correlation is positive; 
when one variable increases, so does the other. If negative, the variables are 
anticorrelated. 
Scores describe the data structure in terms of sample patterns and emphasize 
differences or similarities. Each sample has a score on each PC, which is the coordinate 
of the sample on the PC. Once the information carried by a PC has been 
interpreted with help of the loadings, the score of a sample along that PC can be 
used to characterize a given sample. It describes the major features of the sample, 
relative to the variables with high loadings on the same PC. Samples with close 
scores along the same PC are considered as similar because they have close values 
for the corresponding variables. Conversely, samples with greatly dissimilar scores 
differ greatly from each other with respect to those variables. 
CHEMOMETRICS 395

396 PROCESS ANALYTICAL TECHNOLOGY 
4.2.3.8 PCA Practice for NIRS 
As already mentioned, any multivariate analysis should include some validation, 
that is, formal testing, to extrapolate the model to new but similar data. This requires 
two separate steps in the computation of each model component: calibration, which 
consists of fi nding the new components, and validation, which checks how well the 
computed components describe the new data. Each of these two steps needs its own 
set of samples: calibration samples or training samples, and validation samples or 
test samples. Computation of spectroscopic data PCs is based solely on optic data. 
There is no explicit or formal relationship between PCs and the composition of the 
samples in the sets from which the spectra were measured. In addition, PCs are 
considered superior to the original spectral data produced directly by the NIR 
instrument. Since the fi rst few PCs are stripped of noise, they represent the real 
variation of the spectra, presumably caused by physical or chemical phenomena. 
For these reasons PCs are considered as latent variables as opposed to the direct 
variables actually measured. 
A convenient analogy for understanding latent variables is reconstructing the 
spectrum of a mixture from the spectra of the pure chemicals contained in 
the mixture. The spectra of these pure chemicals would be the latent variables of 
the measured spectrum because they are not directly accessible in the spectrum 
of the mixture. However, PCs are not necessarily the spectra of the pure chemicals 
in the mixtures representing the samples. PCs represent whatever independent 
phenomena affect the spectra of the samples composing the calibration set. If one 
sample constituent varies entirely independently of everything else, and this constituent 
has a spectrum of its own, then one of the PCs will indeed represent the 
spectrum of that constituent. It is most unusual for any one constituent to vary in 
a manner that is exactly independent of any other. There is inevitably some correlation 
between the various constituents in a set of specimens, and any PC will represent 
the sum of the effects of these correlated constituents. Even if full independence 
is accomplished, there is dependence in that the sum of all constituents must equal 
100%. Consequently, the PC representing that source of independent variability will 
look like the difference between the constituent of interest and all the other constituents 
in the samples. The spectrum of the constituent considered could be 
extracted mathematically, but the PCs will not look exactly like the spectrum of the 
pure constituent. 
Once samples have been collected and the corresponding NIR spectra stored in 
suitable fi les, building and using a PCA model involves three steps: selecting the 
appropriate preprocessing procedure(s), running the PCA algorithm and diagnosing 
the model, and interpreting the loading and score plots. Once the model is built, it 
is important to assess its quality before using it for interpretation. There are two 
steps in diagnosing a PCA model. The variances must be checked to determine how 
many components the model should include and to estimate how much information 
the selected components take into account. It is important to verify the variances 
calculated during validation. Then it is advisable to look for outliers, that is, samples 
that do not fi t into the general pattern. 
Total explained variance measures how much of the original variation in the data 
is described by the model. It expresses the proportion of structure found in the data 
by the model. Total residual and explained variances show how well the model fi ts 

the data. Models with small total residual variance (close to 0) or large total explained 
variance (close to 100%) explain most of the variation in the data. With simple 
models residual variance falls to zero with few components. If this is not the case, 
it means that there may be a large amount of noise in the data. Alternatively, it may 
also mean that the data structure is too complex to be accounted for by only few 
components. Variables with small residual variance and large explained variance for 
a particular component are well explained by the relevant model. Variables with 
large residual variance for all or the few fi rst components have a small or moderate 
relationship with other variables. If some variables have much larger residual variance 
than other variables for all components or the fi rst components, they may be 
excluded in a new calculation. This may produce a model that is easier to interpret. 
Calibration variance is based on fi tting the calibration data to the model. Validation 
variance is computed by testing the model on data not used in building the 
model. 
Outliers may sometimes account for large residual variance. An outlier is a 
sample which looks so different from the others that either it is not well described 
by the model or it infl uences the model too much. In practice, true spectral outliers 
are fi rst considered to be samples whose spectral characteristics are not represented 
within the specifi ed sample set. At least one of the model components may focus 
on trying to describe only this particular sample, even if this is irrelevant to the more 
important structure present in the other samples. In PCA, outliers can be detected 
by using different plots or analysis. For example, score plots show sample patterns 
according to one or two components. It is easy to identify a sample lying far away 
from the others. Such a sample is likely to be an outlier. Residuals measure how 
well samples or variables fi t the model determined by the components. A sample 
with a high residual is poorly described by the model, which, however, fi ts the other 
samples quite well. Such a sample does not fi t among the samples well described by 
the model and can be considered an outlier. Deciding to exclude or retain such 
spectral outliers solely on mathematical criteria is not necessarily correct. The outlier 
may be considered as not part of the group intended for use as a calibration set. 
However, such outliers may indicate additional characteristics not taken into account 
during initial sample selection. 
4.2.3.9 Pattern Recognition 
Pattern recognition can be classifi ed according to the distinction between supervised 
and unsupervised techniques. Unsupervised methods such as cluster analysis classify 
data without calibration and are based solely on the collected sample data. Supervised 
classifi cation uses the spectral data and some class membership information. 
Therefore, mathematical models are computed in a fi rst step with a calibration set 
containing spectra and class information. This model is then applied to predict new 
sample classes. Feature extraction methods such as PCA or wavelet compression 
are often applied before cluster analysis. PCA is valuable for extracting features or 
visualizing the data set. Another benefi t of PCA is to reduce the number of 
wavelengths. Many clustering algorithms are in use. Nonhierarchical methods 
include Gaussian mixture models, K means, and fuzzy C means, each of which can 
be subdivided into hard and soft clustering methods. For example, hard clustering 
by K means would declare one given item as one class membership, whereas soft 
CHEMOMETRICS 397

398 PROCESS ANALYTICAL TECHNOLOGY 
clustering by fuzzy C means would assign fractional membership degrees of each 
cluster. 
Current methods for supervised pattern recognition are numerous. Typical linear 
methods are linear discriminant analysis (LDA) based on distance calculation, soft 
independent modeling of class analogy (SIMCA), which emphasizes similarities 
within a class, and PLS discriminant analysis (PLS - DA), which performs regression 
between spectra and class memberships. More advanced methods are based on 
nonlinear techniques, such as neural networks. Parametric versus nonparametric 
computations is a further distinction. In parametric techniques such as LDA, statistical 
parameters of normal sample distribution are used in the decision rules. Such 
restrictions do not infl uence nonparametric methods such as SIMCA, which perform 
more effi ciently on NIR data collections. 
4.2.3.10 SIMCA Classifi cation 
Classifi cation is useful when the response considered is a category variable that can 
be interpreted in terms of several classes to which a sample may belong. The main 
goal of classifi cation is to assign new samples reliably to preexisting classes, but 
classifi cation results can also be used as a diagnostic tool to identify the most important 
variables to retain in the model or to fi nd outliers. Applications include predicting 
whether a pharmaceutical product meets specifi ed quality requirements, in 
which case the result is simply a binary response, or more generally testing or con- 
fi rming the identity of a substance. PCA and discriminant analysis are techniques 
that have found extensive use in NIR analysis for that purpose. SIMCA, a multivariate 
technique optimized for NIR data analysis, combines PCA models for each 
defi ned class in the training set. The approach can be applied to a more general class 
of problems than simple classifi cation, for example, identifi cation. PCA is performed 
on a data set requiring qualitative analysis. After computing the PCs, scores are 
calculated and used to perform qualitative analysis by surrounding each region of 
multidimensional space containing each group ’ s scores with a surface. Conceptually, 
this enclosing surface can be modeled as an ellipsoid and distances to these clouds 
are computed using Mahalanobis metrics based on the PC scores. 
In practice, the optimal number of PCs should be chosen for each class model 
separately, according to a suitable validation scheme. Each model is checked for 
possible outliers and improved as much as possible, like any PCA model. Before 
using the models to predict class membership for new samples, the specifi city of the 
models should be verifi ed, that is, checked for class overlap and adequate interclass 
distance. Once each class has been modeled, and class overlap is not excessive, new 
samples can be fi tted to each model. Unknown samples are then compared to the 
class models, and assigned to classes according to their similarity to the training 
samples. The modeling stage implies that enough samples have been identifi ed as 
members of each class to allow reliable models to be built. Accurate sample description 
also requires a suffi cient number of variables. Actual classifi cation is based on 
statistical tests performed on Mahalanobis distances between sample and model. 
With each unknown sample, all variable values are computed using the model scores 
and loadings, before being compared to the actual values. The residuals are then 
combined into a measure of the object - to - model distance. The scores are also used 
to measure the sample ’ s distance from the model center, known as leverage. Finally, 

both object - to - model distance and leverage are taken into account to decide the 
classes to which the sample belongs. Any sample belonging to a class should have 
a small distance to the class model. 
4.2.3.11 Regression 
Historically, one motivation for performing NIRS was to develop fast and noninvasive 
quantitative analysis. To achieve this, it is not suffi cient to extract PCs from the 
data. There must be a regression method relating these PCs to the constituent, 
analyte, or physical property for which the calibration is performed. Regression 
concerns all methods attempting to fi t a model to the observed data. The fi tted 
model may be used to describe the relationship between two groups of variables or 
to predict values of unknown samples. If X and Y are the two data matrices involved 
in regression, the purpose is to compute a Y = f ( X ) model, which tries to explain, 
or predict, the variations in the Y variable(s) from those in the X variable(s). The 
link between X and Y is explored through a common set of samples from which 
both X and Y values have been collected and are clearly known. Building a regression 
model involves collecting variable values for selected samples and fi tting a 
mathematical relationship to the corresponding spectral data. For example, spectroscopic 
measurements are performed on solutions with known concentrations of a 
given compound. Regression is used to relate the concentration to the spectrum. 
Once the regression model is build, the unknown concentration for new samples 
can be indirectly predicted using the spectroscopic measurements as predictors. The 
advantage appears obvious if one considers noninvasive and nondestructive measurement 
by NIR. If the concentration is diffi cult or expensive to measure directly, 
spectroscopic analysis offers an alternative and much cheaper method of determination. 
Thus the indication for using regression as a predictive tool is its potential for 
performing fast and low - cost measurement as a substitute for more expensive or 
time - consuming alternatives. 
The data may require appropriate preprocessing before a regression can be built 
and used. The calibration step must be followed by a validation step, that is, the 
model must be checked for its effi ciency in predicting independent data. Once the 
number of components has been selected based on the calibration and validation 
variances, the model is diagnosed by interpreting the loading and score plots (for 
PCR and PLS), the loading weight plots (for PLS), and B coeffi cients (PCR), and 
examined for the prediction of new data. 
4.2.3.12 Multiple Linear Regression 
Classic univariate regression uses a single predictor, which is usually insuffi cient to 
model a property in complex samples. Multivariate regression takes into account 
several predictive variables simultaneously for increased accuracy. The purpose of 
a multivariate regression model is to extract relevant information from the available 
data. Observed data usually contains some noise and may also include irrelevant 
information. Noise can be considered as random data variation due to experimental 
error. It may also represent observed variation due to factors not initially included 
in the model. Further, the measured data may carry irrelevant information that has 
little or nothing to do with the attribute modeled. For instance, NIR absorbance 
CHEMOMETRICS 399

400 PROCESS ANALYTICAL TECHNOLOGY 
spectra may contain information relative to solvents, processing path, instrument 
status, light probes, and the like, in addition to the analyte concentration for measurement. 
A good regression model should be able to pick out only the relevant 
information while leaving irrelevant variation aside. 
Multiple linear regression is a method based on ordinary least - squares regression. 
It involves matrix inversion which leads rapidly to collinearity issues if the variables 
are not linearly independent. In MLR, all X variables participate in the model 
independently of each other, and their covariations are not taken into account. X 
variance is not meaningful in this context. Variable independence is also an essential 
precondition. Further, to perform the inversion, MLR requires more samples than 
predictors and no missing values in the data table. If the data table complies with 
these conditions, MLR will approximate the response values by linear combination 
of predictor values, yielding regression coeffi cients known as B coeffi cients. Other 
results are predicted Y values, residuals with error measures, and ANOVA. It is 
noteworthy that MLR is the only multivariate method for which formal statistical 
tests of signifi cance for regression coeffi cients are available. To evaluate the goodness 
of the model, diagnostic tools are associated with the regression coeffi cients. 
The standard error is an estimate of the precision of a given coeffi cient. A Student ’ s 
t value can be computed and compared to a reference t distribution, which in turn 
indicates a signifi cance level or p value. It shows the probability of a t value equal 
to or larger than the observed value if the true value of the regression coeffi cient 
were 0. Predicted Y values are computed for each sample by applying the model 
equation with the estimated B coeffi cients to the observed X values. For each 
sample, the residual is the difference between the observed Y value and predicted 
Y value. The only relevant measure of how well the MLR model performs is provided 
by the Y variances. Residual Y variance is the variance of the Y residuals. It 
expresses how much variation remains in the observed response after the modeled 
part is removed. It is an overall measure of misfi t, that is, the error made when fi tted 
Y values are computed as a function of X values. 
4.2.3.13 PCR and PLS Regression 
Multivariate regression is better suited to fi t a relationship between spectroscopic 
data and the variable to estimate. PC regression (PCR) is a two - step procedure 
which fi rst decomposes the X matrix by PCA, then fi ts an MLR model, using the 
PCs instead of the raw data as predictors in the regression step. MLR and PCR 
model one Y variable at a time, while a PCR model using all PCs gives the same 
solution as MLR. Partial least squares or projection to latent structures (PLS) will 
model both X and Y matrices simultaneously to fi nd the latent or hidden variables 
in X that will best predict the latent or hidden variables in Y . The difference between 
PCR and PLS lies in the algorithm. PLS components are similar to PCs and are also 
referred to as PCs. PLS1 deals with only one response variable at a time (like MLR 
and PCR). PLS2 handles several responses simultaneously. PCR and PLS are projection 
methods, like PCA. Model components are extracted in such a way that most 
information is carried by the fi rst PC, then the second PC, and so on. At a certain 
point, the variation modeled by any new PC becomes mostly noise. Residual variances 
are used to determine the optimal number of PCs modeling useful information 
while avoiding overfi tting. PLS uses both independent and dependent variables 

to fi nd the regression model. It switches iteratively between X and Y . PLS usually 
needs fewer PCs than PCR to reach the optimal solution because the focus is on 
the dependent variables. Results of PCR modeling are given as scores, loadings, 
predicted Y values, residuals, error measures, and B coeffi cients. Results of PLS 
modeling are given as T scores and U scores, P loadings and Q loadings, loading 
weights, predicted Y values, residuals, and error measures. 
The PLS scores are interpreted in the same way as PCA scores since they are 
the sample coordinates along the model components. The additional feature in PLS 
is that two different sets of components are considered, summarizing variations in 
the X space or Y space. PLS loadings express the relatedness of each X and Y variable 
to the model component. T scores are the coordinates of data points located 
in the X space that describe the part of structure in X which is most predictive for 
Y. U scores summarize the part of structure in Y which is explained by X along a 
given model component. The relationship between T scores and U scores is a model 
of the relationship between X and Y along a specifi c component, and it can be 
visualized for diagnostic purposes. It follows that loadings will be interpreted differently 
in the X space and Y space. P loadings are similar to PCA loadings. They 
express how much each X variable contributes to a specifi c model component. 
Directions determined by the projections of X variables are used to interpret the 
meaning of the location of a projected data point on a T score plot, in terms of 
variations in X. Q loadings express the direct relationship between Y variables and 
T scores. Thus, the directions determined by the projections of Y variables by means 
of Q loadings can be used to interpret the meaning of the location of a projected 
data point on a T score plot in terms of sample variation in Y . When plotted on a 
single graph, P and Q loadings make it possible to interpret T scores by considering 
variations in both X and Y . In contrast to PCA loadings, PLS loadings are not normalized, 
so that P loadings and Q loadings do not share a common scale. Thus, only 
their directions can be interpreted, not their lengths. The residuals should be randomly 
distributed and free from systematic trends. The most useful residual plots 
are Y residuals versus predicted Y , and Y residuals versus scores plots. 
Where there is more than one Y variable, PLS2 is designed for interpretation of 
all the variables simultaneously. It is often argued that PLS1 or PCR are better 
predictors; this is usually true if there are strong nonlinearities in the data, in which 
case modeling each Y variable separately according to its own nonlinear features 
might perform better than trying to build a common model for all Y ’ s. On the other 
hand, if the Y variables are somewhat noisy, but strongly correlated, PLS2 will model 
the whole information by excluding more noise. 
As in PCA, outliers may infl uence modeling and should be detected. In regression, 
there are many ways a sample can be defi ned as an outlier. It may be an outlier 
according to X variables only or to Y variables only, or to both. It may also not be 
an outlier for either separate set of variables but become an outlier for ( X, Y ) 
regression. 
4.2.3.14 Regression Practice in NIRS 
Calibration is the fi tting stage: The main data set, containing only the calibration 
samples set, is used to compute model parameters such as PCs, regression coeffi - 
cients, and the like. The models must be validated to get an idea of how well a 
CHEMOMETRICS 401

402 PROCESS ANALYTICAL TECHNOLOGY 
regression model performs if used to predict new, unknown samples. A test set 
consisting of samples with known variable values is used. Only the X values are fed 
into the model, from which response values are predicted and compared to the 
known, true response values. The model is validated if prediction residuals are low. 
Validating a model means checking how well the model performs on real new data. 
As a regression model is usually made to perform predictions on future unknown 
samples, validation must estimate the uncertainty of prediction. If the uncertainty 
is reasonably low, the model can be considered usable. The steps required for complete 
modeling are illustrated in Figure 20 . And in Figure 21 the end result obtained 
from a typical chemometric package performing calibration and validation is 
shown. 
Independent (external) test - set validation and cross - validation are the most 
current methods of estimating prediction error. External test - set validation is based 
on testing the model on a subset of available samples, which will not be involved in 
FIGURE 20 Typical working fl owchart of the development of a PLS calibration with NIR 
spectra. 
Independant 
validation set 
NIR scanning 
Conventional 
analysis 
Calibration 
Cross 
validation 
Test set 
validation 
Calibration set 
NIR scanning 
Conventional 
analysis 
Method 
development 
Optimization 
Routine use 
Test 
FIGURE 21 Typical output of a PLS calibration and validation. 
0.5 1 1.5 2 2.5 3 3.5 4 0 
0.5
1 
1.5
2 
2.5
3 
3.5
4 
y = 0.97x + 0.063 
0.5 1 1.5 2 2.5 3 3.5 4 0 
0.5
1 
1.5
2 
2.5
3 
3.5
4 
y = 1x + 0.0023 
SEC = 0.04 
Bias = -8.75 e-005 
R=0.999 
SEP=0.06 
Bias=0.008 
R=0.999 
Calibration Validation

computing the model components. For example, the global data table is split into 
two subsets. The calibration set contains all samples used to compute the model 
components, using both X and Y values. The test set contains all the remaining 
samples, for which X values are fed into the model once a new component has been 
computed. The predicted Y values are then compared to the observed Y values, 
yielding a prediction residual that can be used to compute a validation residual 
variance, or a measure of the uncertainty of future predictions, called root mean 
square error of prediction (RMSEP). This value is calculated for each modeled 
response and indicates the average uncertainty that can be expected when predicting 
Y values for new samples, expressed in the same units as the Y variables. Table 
3 gives a formula for RMSEP. The results of future predictions can then be presented 
as the predicted values ± RMSEP. This measure is valid provided that the new 
samples are similar to those used for calibration. Otherwise, the prediction error 
might be much higher. No assumption about statistical error distribution is made 
for modeling. As a consequence, prediction error cannot be given as a proper statistical 
interval estimate (e.g., twice the standard deviation, etc.). RMSEP is a practical 
average prediction error and if both calibration and validation sample sets are 
representative of future samples to predict, it is a good error estimate. 
With cross - validation, the same samples are used both for model estimation and 
testing. Cross - validation represents an alternative way of utilizing samples for validation 
if their number is small or moderate. The method consists in leaving out a 
few samples from the calibration data set and calibrating the model on the remaining 
data points. The left - out sample values are then predicted and the corresponding 
prediction residuals computed. The process is repeated with another subset of the 
calibration set, and so on until every object has been left out once. All prediction 
residuals are combined to compute the validation residual variance and RMSEP. 
Full cross - validation leaves out only one sample at a time. Segmented cross - validation 
leaves out a whole group of samples at a time. Segmented cross - validation is 
faster, but segment selection requires some care since it should feature unique 
information. For example, samples which can be considered as replicates of each 
other should not be present in different segments. 
In the case of an independent test set the fi le should contain 20 – 40% of the full 
data. The calibration and test set must cover the same population of samples as 
TABLE 3 Descriptors Used to Estimate Performance of Calibration 
.y 
j
m 
j y y 
m 
= 
. ( ) 
. 
= . 1 
2 
1 
cov , . ( )= 
. ( ) . . . ( ) 
. 
= . y y 
y y y y 
m 
j
m 
j j 1 
1 
R 
y y 
y y 
= . ( ) 
. 
cov , 
. . 
SEC = 
. . ( ) 
. . 
= .j
m 
j j y y 
m q 
1 
2 
1 
SEP = 
. . ( ) = .j
n 
j j y y 
n 
1 
2 
Bias = 
. . ( ) = .j
n 
j j y y 
n 
1 
where y j = reference value for the sample j 
. yj = predicted value for the sample j 
m = number of samples in calibration 
n = number of samples in validation 
q = PC number 
CHEMOMETRICS 403

404 PROCESS ANALYTICAL TECHNOLOGY 
possible. Replicate measurements should never be present in both the calibration 
and test sets. This risk exists with a random selection which is proposed in current 
NIR software. If it is the simplest way to select a test set, it leaves the selection to 
the computer. Manual selection is recommended since it gives full control over the 
selection of a test set. 
Multiple regression programs also calculate auxiliary statistics, designed to help 
decide how well the calibration fi ts the data, and how well it can be expected to 
predict future samples. For example, two of these statistics are the standard error 
of calibration (SEC) and the multiple correlation coeffi cient ( R ). The SEC (also 
called standard error of estimate, or residual standard deviation) and the multiple 
correlation coeffi cient indicate how well the calibration equation fi ts the data. Their 
formulas are given in Table 3 . 
The SEC has the same units as the dependent variables, and it refl ects the differences 
between the instrumental value for the analyte of interest and the reference 
laboratory value. It expresses the modeling error and cannot be used to 
estimate future prediction errors. In the absence of instrumental error, SEC is only 
the measure of reference laboratory error. It is an indication of whether calculation 
using the calibration equation will be suffi ciently accurate for the purposes for which 
it has been generated. In practice, SEC is less accurate than the reference laboratory, 
even in the absence of instrumental error, if the wavelength range used as independent 
variables in the calibration does not account for all interference in the samples 
or if other physical phenomena are present. The detection limit and sensitivity, or 
signal - to - noise, defi nes instrument performance for a specifi c NIR application. 
Detection limits can usually be approximated for any NIRS method as equal to 
three times the SEC for the specifi ed application. Sensitivity for quantitative NIR 
methods can typically be evaluated as the slope of the calibration line for the concentration 
of the examined analyte ( y axis), versus the change in optical response 
( x axis), between samples of varying concentration. Sensitivity, from a purely instrumental 
point of view, is expressed as a signal - to - noise or peak height ratio for a 
particular compound versus peak - to - peak noise at some absorbance (usually zero). 
However, in a practical sense, the above considerations do not really matter in NIR 
spectroscopy. This is due to the fact that the applications developed for practical 
NIR are mostly based on empirical calibration methods and that the calibration is 
specifi c to the problem of interest. It is a characteristic of calibration equations using 
multivariate mathematical techniques to compensate for the common variations 
found in noisy chemical samples and imperfect instrumental measurements. This is 
why properly calculated and validated NIR calibration models are robust and work 
extremely well. 
The multiple correlation coeffi cient R is a dimensionless measure of how well 
the calibration fi ts the data. R can have values between . 1 and +1, but in a calibration 
situation only positive values exist. A value close to zero indicates that the calibration 
fails to relate the spectra to the reference values. As the correlation coeffi cient 
increases, the spectra become better and better predictors of the reference values. 
Because the multiple correlation coeffi cient is dimensionless, it is a useful way of 
comparing data or results with different units, and that are diffi cult to compare in 
other ways. However, its value gives no indication of how well the calibration equation 
can be expected to perform on future samples. 

4.2.3.15 Some Pitfalls 
The fi rst step before any multivariate data analysis is sampling. The selection or 
preparation of a set of calibration samples is a critical issue. For example, the analyst 
must collect or prepare samples which span the complete range of constituent concentrations 
as evenly distributed as possible. It is usual for random sample selection 
to cause the models to fi t most closely to the mean concentration samples. Samples 
at high or low concentration levels will not infl uence the slope and intercept in the 
case of multivariate regression. Ideally, even concentration distribution will allow 
the model to minimize the residuals at the extremes and at the center with relatively 
equal weighting. Calibration sets must not only uniformly cover an entire constituent 
range, they must also be composed of a correctly distributed number of sample 
types. For example, ideal calibration sets are composed of more than 10 – 15 samples 
per analytical term. These samples ideally have widely varying composition evenly 
distributed across the calibration range. Last but not least, spectroscopic measurements 
should be performed under as identical conditions as possible between calibration 
samples and routine samples. A recurrent diffi culty is the effect of moisture 
within a solid or powdered sample. The presence or absence of water in a sample 
infl uences the extent of hydrogen bonding, which affects both band position and 
width in the complete NIR domain. If a calibration model is developed from samples 
featuring a wide range of the component of interest but a small range in moisture, 
it will only be useful for samples with that narrow moisture range. Such a calibration 
may not be robust enough for routine application. To summarize, each calibration 
problem involves slightly different aspects. 
Paying attention to fi ner details may be obvious, as unexpected error sources can 
affect the quality of computed models. Errors occur, for example, in sample preparation 
and measurement. Parameters of potential infl uence include temperature differences 
in samples or the instrument while recording the data, calibration sample 
instability, instrument noise and drift, changes in instrument wavelength setting, 
nonlinearity, stray light effects, particle size differences, concentration - dependent 
color differences, residual solvent interaction, and nonhomogeneous samples. The 
reference method may not measure the same components as the spectroscopic 
method. Controlling for these aspects may sound overly rigorous and may initially 
dampen enthusiasm for NIRS. However, multicomponent problems are inherently 
complex and require the management of several variables simultaneously in order 
to develop a usable calibration. The ultimate goal of successful calibration is to calculate 
a mathematical model with the calibration samples which is most sensitive 
to changes in the modeled parameter and least sensitive to all other noncalibration 
related factors, such as physical, chemical, and instrumental variables. Every case 
must be evaluated in terms of the chemical and physical data carried by the calibration 
samples and the information which the analyst wishes to obtain. 
4.2.3.16 Example Analytical Applications of NIRS 
Assuming NIR spectra have been recorded, once a set of pharmaceutical samples 
has been analyzed with high precision by some analytical reference method, the 
concentration of an analyte of interest clearly determined or the identity of a given 
CHEMOMETRICS 405

406 PROCESS ANALYTICAL TECHNOLOGY 
compound confi rmed, the sample data can then be fi led into a training set to generate 
a calibration for subsequent predictions. To form a usable training set, the 
samples should evenly span the concentration range for the analyte of interest or 
the expected variations in quality. An obvious pitfall is to develop calibrations that 
only use sample sets with uneven constituent distributions or too narrow variations 
of the attribute of interest. In that case the model will most closely fi t the dominating 
samples in the calibration set. The calibration will be highly weighted to the 
mean value and behave poorly against variations in further samples. Conversely, an 
ideally evenly distributed calibration set will weight the calibration model equally 
across the entire concentration range. A properly developed calibration model 
of this kind will perform most accurately with samples at high and low 
concentrations. 
Despite these sampling issues, which are important for NIR spectroscopists in 
the pharmaceutical industry, the potential of NIR applications remains intact. NIR 
is rapid and nondestructive, requiring little or no sample preparation. It can monitor 
concentrations of several chemical species and physical parameters simultaneously. 
Its speed and ease are a major bonus in many analytical or monitoring processes. 
All types of materials are concerned, from solid or liquid raw materials (e.g., excipients), 
active pharmaceutical ingredients (API), intermediate synthesis products, 
intermediate formulations in the form of powders, slurries, granulates, or pellets, up 
to fi nal dosage forms such as capsules, tablets, or lyophilized substances. NIR measurements 
can be performed in close or direct contact with the sample, in both the 
laboratory or directly online or inline in the production plant, in order to obtain 
analytical information rapidly and save time. Transmittance measurements have 
become an alternative to the conventional refl ectance spectroscopy of pharmaceuticals. 
The important difference is that transmittance NIR samples a volume, whereas 
refl ectance NIR merely samples the surface of solid samples. This has the advantage 
of more representative values for less homogenous samples like tablets or capsules. 
On the other hand, more attention has to be paid to sample presentation with 
respect to stray light and light scattering. 
Traditionally, pharmaceutical excipients have been characterized in the laboratory 
by viscosity, pH in dispersion/solution, water content, ash content, constituent 
amounts, particle size, and so on. In reality, there are also other subtle properties 
that are not covered by these parameters but that nevertheless affect the properties 
of a drug formulation. Many of these variations can be extracted from NIR refl ectance 
spectra in addition to identity testing by other methods (Figure 22 ). The raw 
material spectra and results from other analytical methods can then be combined 
to predict the impact of raw material quality variations on fi nal batch quality. This 
conformity test by NIR can be achieved by recording numerous raw material 
batches of known quality and calculating an envelope of acceptability around the 
mean spectrum. Conform raw material is only qualifi ed when it lies within the 
threshold values of the envelope at each wavelength. This test would fail poor materials 
with high impurities and high water levels. 
Most pharmaceutical production is performed in batches, both when synthesizing 
active compounds and manufacturing pharmaceutical formulations. NIR can 
measure the fi nal state in batch production in terms of spectral similarity, using 
SIMCA, spectral correlation, or spectral distance. Individual batch development can 
also be monitored using PCA. Several parameters of the same product can be 

determined quantitatively using PLS models. A current basis for comparison includes 
processing time as a Y variable in a PLS model based on several batches considered 
good for release. Limits for an accepted batch can then be calculated and compared 
with online NIR measurements during the batch process. 
Drying of drug substances is an important step in the production of active materials. 
It is usually performed in an agitated vacuum dryer, carefully controlled to 
ensure a suitably dry product and to minimize the risk of overdrying which could 
damage the high cost material. Some drug substances are hydrates, which present 
special requirements for dryer control. If the waters of crystallization are removed, 
the batch may need reprocessing or may even be lost. The traditional means of 
monitoring and controlling drying operations is to use the slight changes in temperature 
within the dryer to estimate an endpoint and then to sample the volatile 
content for laboratory analysis. There is often no distinct temperature endpoint. 
Some products will also retain residual solvent within the crystals. Sampling vacuum 
dryers is diffi cult as the vacuum often has to be released, and the laboratory analysis 
can take several hours. The use of an NIR fi ber - optic probe, inserted directly into 
the plant scale dryer, can provide real - time analysis of dryer contents and can thus 
FIGURE 22 Excipient library with the result of a PCA analysis for identity testing. 
Lactose 
Cellulose 
Starch 
Opadry 
MgSt 
SDS
CHEMOMETRICS 407

408 PROCESS ANALYTICAL TECHNOLOGY 
be considered as a valuable alternative. Differences in physical and chemical properties 
between polymorphs of an active compound (e.g., solubility, dissolution rate, 
chemical reactivity, resistance to degradation, bioavailability) are highly signifi cant 
for the pharmaceutical industry. Effective methods are required not only to control 
the content in the active compound but also to detect and determine the undesirable 
form. NIRS can be used in some circumstances to confi rm a low percentage of 
undesirable crystalline state in the amorphous polymorph of a compound. The 
underlying methods are both qualitative (e.g., classifi cation) and quantitative. The 
simplicity, expeditiousness, and reliability of the NIR method make it a promising 
tool for controlling the polymorphic purity of the amorphous phase. 
4.2.3.17 Conclusion 
In many pharmaceutical companies, quality control departments already use NIRS 
to identify formulations. Figure 23 illustrates a PLS calibration for the active content 
determination in a low - dose tablet. Once identity testing is passed, it is straightforward 
to consider as a next step the determination of active content in intact tablets. 
Thus, qualitative and quantitative analysis can be performed by acquiring a single 
NIR spectrum per sample. Two analytical techniques are replaced by one — nondestructive 
— NIR measurement. For this purpose near - infrared spectroscopy is a fast 
and powerful alternative to traditional analysis, which only remains necessary as 
reference analytics. 
FIGURE 23 Quantifi cation of the API content in a tablet and overview of the corresponding 
PLS regression. 
1000 1500 2000 2500 
-2 
-1
0
1
2 
x 10-3 Spectra 
Wavelengths 
Absorbance 
0 5 10 15 20 
0 
10 
20 
30 
40 
50 
Factors 
Cross validation 
0 2 4 6 8 
0
2
4
6
8 
Validation 
Predicted property 
Reference property 
PLS 
Rval 
Bias 
SEP(C) 
SEP 
SEC 
PC number 13.0000 
0.1114 
0.9985 
0.1699 
0.1695 
-0.0123 
0.9962 
Statistics 
SECV 
Rcal 

The viability of analytical methods obtained by combining vibrational spectroscopy 
— mainly NIRS — and chemometrics, once validated, can be proved on numerous 
industrial examples. However, dominant application fi eld of chemometrics is 
not the pharmaceutical industry, and most accumulated experience is found in the 
agricultural and food industries. Best practice in chemometrics is independent of 
the analytical problem or working fi eld. To solve complex pharmaceutical problems 
by using a large band of spectroscopic methods and at the same time optimizing 
sample handling, multivariate data analysis algorithms are highly recommended. 
Thus, education of chemists, pharmacists, and analysts with regards to chemometrics 
remains a requirement to catch from the beginning the usefulness of multivariate 
data analysis applied to spectroscopic data. 
BIBLIOGRAPHY 
Section 4.2.1 
Bakeev , K. A. , Ed . (2005), Process Analytical Technology , Blackwell , London . 
Giodici , P. ( 2003 ), Applied Data Mining. Statistical Methods for Business and Industry , Wiley , 
Hoboken, NJ . 
Initiative for Pharmaceutical cGMPs for the 21st Century , FDA , Washington, DC . 
Martin , A. ( 1993 ), Physical Pharmacy , 4th ed., Lippincot Williams & Wilkins . Philadelphia, 
PA . 
Section 4.2.2 
Colthup , N. B. , Daly , L. H. , and Wiberley , S. E. ( 1990 ), Introduction to Infrared and Raman 
Spectroscopy , 3rd ed., Academic . 
The Proceedings of NIR - 97, J. Near - Infrared Spectros ., 6 (1 – 4), NIR Publications, 1998. 
Rouessac F. , and Rouessac , A. ( 2000 ), Chemical Analysis. Modern Instrumentation Methods 
and Techniques , Wiley - VCH , New York . 
Siesler , H. W. , Ozaki , Y. , Kawata , S. , and Heise , H. M. Eds (2002), Near - Infrared Spectroscopy. 
Principles, Instruments, Applications , Wiley - VCH , New York . 
Williams , P. , and Norris , K. ( 2001 ), Near - Infrared Technology in the Agriculture and Food 
Industries , 2nd ed. American Association of Cereal Chemists , St Paul, MN . 
Section 4.2.3 
Brereton , R. G. ( 2003 ), Chemometrics: Data Analysis for the Laboratory and Chemical Plant , 
Wiley , Hoboken, NJ . 
Kramer , R. ( 1998 ), Chemometric Techniques for Quantitative Analysis , Marcel Dekker , 
New York . 
Manly , B. F. J. ( 2000 ), Multivariate Statistical Methods. A Primer , 2nd ed., Chapman & 
Hall/CRC , New York . 
Mark , H. ( 1991 ), Principles and Practice of Spectroscopic Calibration , Wiley , New York . 
Martens , H. , and Martens , M. ( 2001 ), Multivariate Analysis of Quality: An Introduction , Wiley , 
New York . 
Massart , D. L. , Vendeginste , B. G. M. , Buydens , L. M. C. , De Jong , S. , Lewi , P. J. , and Smeyers - 
Verbeke , J. ( 1997 / 1998 ), Handbook of Chemometrics and Qualimetrics, Parts A and B , 
Elsevier , Amsterdam . 
BIBLIOGRAPHY 409

410 PROCESS ANALYTICAL TECHNOLOGY 
Naes , T. , Isaksson , T. , Fearn , T. , and Davies , T. ( 2002 ), A User - Friendly Guide to Multivariate 
Calibration and Classifi cation , NIR Publications , Chichester . 
Peer Review Articles 
Chalus , P. , Roggo , Y. , Walter , S. , and Ulmschneider , M. ( 2005 ), Near - infrared determination 
of active substance content in intact low - dosage tablets , Talanta , 66 , 1294 – 1302 . 
Inschneider , A. ( 2002 ), Validierung von chemometrischen Methoden am Beispiel multivariater 
Datenanalyse in der Nah - Infrarot Spektroskopie , in Handbuch der Validierung , 
Kromidas , S. Ed., Wiley-VCH, Weinheim, pp. 302 – 307. 
Roggo , Y. , Edmond , A. , Chalus , P. , and Ulmschneider , M. ( 2005a ), Infrared imaging for qualitative 
analysis of pharmaceutical solid forms and trouble shooting , Anal. Chim. Acta , 535 , 
79 – 87 . 
Roggo , Y. , Jent , N. , Edmond , A. , Chalus , P. , and Ulmschneider , M. ( 2005b ), Characterizing 
process effects on pharmaceutical solid forms using near - infrared spectroscopy and infrared 
imaging , Eur. J. Pharm. Biopharm ., 61 ( 1 – 2 ), 100 – 110 . 
Roggo , Y. , Roeseler , C. , and Ulmschneider , M. ( 2004 ), Near - infrared spectroscopy for qualitative 
comparison of pharmaceutical batches , J. Pharm. Biomed. Anal ., 36 , 777 – 786 . 
Russell , F. , and Ulmschneider , M. ( 2004 ), Dissolution testing by means of NIR transmittance 
spectroscopy , Am. Pharm. Rev ., July/August. 
Sukowski , L. , and Ulmschneider , M. ( 2005 ), In - line process analytical technology on qualitative 
NIR modelling: An innovative approach for the pharmaceutical quality control , 
PharmInd , 67 ( 7 ), 830 – 835 . 
Ulmschneider , M. , Barth , G. , Reder , B. , V o gel , A. , and Schilling , D. ( 2000a ), A transferable 
basic library for the identifi cation of active substances using near - infrared spectroscopy , 
Pharm. Ind ., 62 ( 4 ), 301 – 304 . 
Ulmschneider , M. , Barth , G. , and Trenka , E. ( 2000b ), Building transferable cluster calibrations 
for the identifi cation of different solid excipients with near - infrared spectroscopy , Pharm. 
Ind ., 62 ( 5 ), 374 – 376 . 
Ulmschneider , M. , and P e nigault , E. ( 2000a ), Assessing the transfer of quantitative NIR - calibrations 
from a spectrometer to another one , Analusis , 28 , 83 – 87 . 
Ulmschneider , M. , and P e nigault , E. ( 2000b ), Direct identifi cation of key - intermediates in 
containers using Fourier - Transform near - infrared spectroscopy through the protective 
polyethylene primary packaging , Analusis , 28 , 136 – 140 . 
Ulmschneider , M. , and P e nigault , E. ( 2000c ), Non - invasive confi rmation of the identity of 
tablets by near - infrared spectroscopy , Analusis , 28 , 336 – 346 . 
Ulmschneider , M. , Wunenburger , A. , and P e nigault , E. ( 1999 ), Using near - infrared spectroscopy 
for the non invasive identifi cation of fi ve pharmaceutical active substances in sealed 
vials , Analusis , 27 , 854 – 856 . 

411 
4.3 
CHEMICAL IMAGING AND 
CHEMOMETRICS: USEFUL TOOLS FOR 
PROCESS ANALYTICAL TECHNOLOGY 
Yves Roggo and Michel Ulmschneider 
F. Hoffmann - La Roche Ltd, Basel, Switzerland 
Contents 
4.3.1 Introduction 
4.3.2 Defi nition of Hyperspectral Imaging 
4.3.3 Hyperspectral Image Acquisition and Instrumentation 
4.3.3.1 Principles of Hyperspectral Image Acquisition 
4.3.3.2 Spectroscopic Instrumentation 
4.3.4 Chemometric for Imaging 
4.3.4.1 Data Preprocessing 
4.3.4.2 Pixel Classifi cation 
4.3.5 Practical At - Line and Offl ine Chemical Imaging 
4.3.5.1 Practical Tools for Distribution Map Analysis 
4.3.5.2 Wavelength Selection and Chemical Interpretation 
4.3.5.3 Unsupervised Classifi cation for Process Troubleshooting 
4.3.5.4 Supervised Classifi cation, NIR Imaging, and Process Development 
4.3.5.5 Further Developments: Online Analysis by Hyperspectral Imaging 
4.3.6 Conclusions 
References 
4.3.1 INTRODUCTION 
Chemical compound homogeneity is an important issue for pharmaceutical solid 
forms. A classical spectrometer [1 – 3] integrates spatial information. However, use 
of a mean spectrum on a surface can be a drawback in solid - form analysis. For 
example, in the pharmaceutical industry, it is important to map the distribution of 
active ingredients and excipients in a tablet as this reveals physical interaction 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

412 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
between components and helps to solve homogeneity issues — hence the increasing 
number of spectroscopic imaging studies on the visualization of chemical component 
homogeneity [4 – 9] . 
Vibrational hyperspectral imaging (chemical imaging) is a recent development 
that combines the chemical information from spectroscopy with spatial information. 
In principle, it is possible to collect hyperspectral images with single - point detectors, 
that is, classical mapping with microscopes. However, array detectors measure all 
pixels simultaneously, reduce recording time, provide uniform background, and 
improve the signal - to - noise ratio [4] . A complete spectrum is acquired for each 
pixel, meaning that a hyperspectral data set is in fact a data cube, that is, a three - 
dimensional (3D) matrix. 
This chapter begins by defi ning hyperspectral imaging, then presents details of 
instrumentation and image formation, and concludes by describing some chemometric 
image analysis tools with the help of pharmaceutical examples. 
4.3.2 DEFINITION OF HYPERSPECTRAL IMAGING 
This chapter considers three types of digital image. The fi rst is binary (or black and 
white), where the pixel value can be either 0 or 1. The second is monochromatic 
(e.g., gray scale), which can be presented as a two - dimensional (2D) X . Y array 
describing the distribution of light intensity, where X and Y are the numbers of 
digitization steps (i.e., pixels) along the two spatial directions. Each pixel has a value 
between 0 (black) and 255 (white). The third is color, which can be described in the 
red – green – blue (RGB) space, for example, with three planes (for red, green, and 
blue, respectively) as a 3D X . Y . 3 matrix. Each pixel has a value between 0 and 
255 for red, green, and blue, generating 255 3 possible colors for such images (known 
as 24 - bit images). 
By analogy, a hyperspectral data set is defi ned by at least 50 planes; an absorbance 
map is acquired for each wavelength in the spectral range. If the wavelengths 
number less than 50, the term multispectral imaging is used. 
With chemical imaging, a new type of data structure needs to be analyzed. Chemical 
imaging experiments yield a 3D X . Y . . matrix or data cube, where X and Y 
are the spatial dimensions and . the spectral dimension. One spectrum per pixel is 
recorded and selection of a wavelength will show an absorbance picture of the 
sample [10] (Figure 1 ). 
The data cube combines spectral and spatial information and therefore includes 
the requisite statistics for spectral classifi cations. However, new chemometric 
strategies have to be applied to interpret chemical imaging results. 
4.3.3 HYPERSPECTRAL IMAGE ACQUISITION 
AND INSTRUMENTATION 
4.3.3.1 Principles of Hyperspectral Image Acquisition 
This section discusses three modes of acquisition: mapping, array detection, and 
fi ber bundles. 

Mapping Historically, mapping [11] was the fi rst method used to acquire hyperspectral 
data cubes, in particular with Raman spectroscopy and infrared (IR) microscopy. 
The image is created pixel by pixel in a “ step - and - acquire ” mode: A spectrum 
is measured at one point of the sample, and then the sample moves to the next 
measurement position and another spectrum is acquired. The process is iterative for 
all positions in the area that defi ne the image. 
The drawback is the measurement time, which depends on the number of pixels. 
Spectrometer manufacturers therefore developed line mapping, in which samples 
are scanned line by line, thereby reducing the acquisition time. These devices were 
developed for Raman, IR, or near - IR (NIR) spectroscopy (with diode array 
detectors). However, due to the moving stage, this kind of imaging principle is only 
suitable for at - line applications. 
Focal Plane Array Detectors Focal plane array (FPA) optical detectors are composed 
of several thousand individual detector elements forming a matrix of pixels. 
As their name indicates, they are placed in the spectrometer ’ s focal plane. They can 
be manufactured to be sensitive to ultraviolet (UV), visible (Vis), NIR, or IR radiation. 
Recent developments in optics have produced cooled and uncooled FPAs with 
different numbers of pixels from 64 . 64 up to 1024 . 1024 and different spectral 
ranges of detection (from 1000 to 12,000 nm). The different types of FPA include 
those built with indium antimonide (InSb), platinum silicide (PtSi), indium gallium 
arsenide (InGaAs), and mercury cadmium telluride (HgCdTe) [12] . The mercury 
cadmium telluride (MCT) detector has become the dominant IR FPA through its 
entire range coverage. 
Fiber Bundles Fiber bundles are used for Raman imaging. Several optical fi bers 
are grouped together, each analyzing a specifi c sample area [13] . A 3D data cube is 
FIGURE 1 Data structure of chemical image. 
x 
y 
Hyperspectral 
= 
Single wavelengthi mage 
image 
data cube 
Spectra 
HYPERSPECTRAL IMAGE ACQUISITION AND INSTRUMENTATION 413

414 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
collected using a 2D charge - coupled device (CCD) detector. At the collection point 
the fi bers are arranged in a circular pattern or square array in order to analyze 
defi ned surfaces on the samples (Figure 2 ). At the detection end, the fi bers are 
aligned in order to detect the signal: The fi rst CCD detector dimension is used for 
spatial information and the second for the spectral region. The technical diffi culty 
is to assign the signal position on the detector end with the pixel position in the 
image. This technique ’ s main advantage is fast image acquisition. 
4.3.3.2 Spectroscopic Instrumentation 
The imaging detectors, whether for point mapping, line scanning, or array detection, 
can be coupled with different types of spectrometers. Instrument types are classifi ed 
by wavelength selection modality into imaging Fourier transform (FT) and tunable 
fi lter (TF) spectrometers, both of which are presented below, and dispersive 
spectrometers. FT imaging systems are classical laboratory instruments while TF 
spectrometers are compact and robust systems for chemical imaging. 
FT Spectrometers FT spectrometers (Figure 3 ) differ from scanning spectrometers 
by the fact that the recorded signal is an interferogram [14] (see Chapter 6.2 ). They 
can be coupled to a microscope or macrochamber with an FPA detector. FT chemical 
imaging systems (CISs) are available for Raman, NIR, and IR spectroscopy. 
However, they can only be considered as research instruments. For example, most 
IR imaging systems are FT spectrometers coupled to microscopes. This type of 
spectrometer allows the acquisition of spectra in refl ection, attenuated total 
refl ection (ATR), or transmission mode. 
TF Systems A TF is a device whose spectral transmission can be controlled by 
applying a voltage or acoustic signal. There are two main TF devices: acousto - optical 
TF (AOTF), based on diffraction, and liquid crystal TF (LCTF), based on birefringence. 
An AOTF is a transparent crystal in which an ultrasonic wave fi eld is created, 
FIGURE 2 Principle of fi ber bundles. ( Adapted from A. D. Gift, J. Ma, K. S. Haber, B. L. 
McClain, and D. Ben - Amotz, Journal of Raman Spectroscopy , 30, 757 – 765, 1999 .) 
Detection Collection 
Fiber 1 
Fiber n 
Fiber bundles 
Spatial detection

that is, the selected wavelength is a function of fi eld intensity. An LCTF is built using 
several Lyot fi lters (Figure 4 ): The incoming light is fi rst polarized for one Lyot fi lter 
element, then the birefringent crystal introduces a phase difference ( . ) between the 
rays of light, and the light fi nally passes through the second “ analyzer ” polarizer, 
which selects the transmitted wavelength. The disadvantage of an LCTF compared 
to an AOTF is in speed: AOTFs scan in microseconds, LCTFs in milliseconds [15] . 
With TFs, samples can be scanned wavelength by wavelength. Single - wavelength 
images are then grouped into a data cube. Specifi c wavelengths (e.g., specifi c for 
particular chemical compounds) can be selected to reduce acquisition time. Commercial 
TF devices (Figure 5 ) are available for NIR and Raman imaging. TF imaging 
systems appear suitable for process analytical technology (PAT) applications in that 
they can be installed online due to their fast acquisition times and simple and robust 
mode of image acquisition. 
FIGURE 3 IR microscope linked to FPA detector. 
Spectrometer 
Microscope FPA detector 
Sample 
MCT array 
IR wavelengths 
FIGURE 4 Lyot fi lter: LCTF element. ( Adapted from N. Gat, Proceedings SPIE , 4056, 
50 – 64, 2000 .) 
Analyzer Polarizer 
Liquid crystal 
Quartz 
Light 
Glass 
Glass 
N Lyot filters 
Selection of one 
specific 
wavelength 
Variable 
HYPERSPECTRAL IMAGE ACQUISITION AND INSTRUMENTATION 415

416 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
4.3.4 CHEMOMETRIC FOR IMAGING 
Hyperspectral imaging systems generate a large amount of data which has to be 
processed in order to fi nd the relevant information. Extraction methods must therefore 
be developed to display images of chemical compound homogeneity ( “ distribution 
maps ” ). The main steps in data cube analysis (Figure 6 ), presented in sequence 
below, begin with data preprocessing (to improve information quality), proceed with 
classifi cation (to identify the main chemical compounds for each pixel and display 
the distribution maps), and terminate in the deployment of tools for distribution 
map analysis. 
4.3.4.1 Data Preprocessing 
Preprocessing enhances chemical information and removes noise and scattering 
effects [14] . Its specifi city resides in the fact that data cubes can be preprocessed in 
both the wavelength and spatial dimensions. 
Classical spectral preprocessing, described in Section 4.3.1 , comprises normalization, 
smoothing, and baseline correction. Some CIS provide intensity as refl ectance. 
The data are therefore converted to absorbance. Another type of spectral preprocessing 
can be added based on prior chemical knowledge: wavelength selection. The 
less informative spectral ranges are removed in order to reduce computation time 
and improve convergence of chemometric algorithms. 
Image preprocessing techniques are also useful. Thus bad pixels (e.g., those 
without signals or outliers) can be removed and several data cubes grouped together 
for simultaneous analysis in order to simplify comparison. The hyperspectral image 
can also be masked to select only the regions of interest. Manuals describe other 
image preprocessing methods, such as spatial smoothing, contrast enhancement, 
deblurring, and fi ltering [16] . 
FIGURE 5 Acquisition by LCTF NIR imaging system. ( Adapted from N. Lewis, J. 
Schoppelrei, E. Lee, and L. Kidder, in Process Analytical Technology , Editor K. A. Bakeev 
Blackwell, London., 2005, pp. 187 – 225 .) 
S S 
Tunable filters 
FPA detector 
Optical parts (objectives) 
Sources 
Samples 
Wavelength selection

4.3.4.2 Pixel Classifi cation 
After data preprocessing, classifi cation is a critical step in distribution map extraction. 
Various algorithms can be applied. Although an exhaustive list is beyond the 
scope of this handbook, methods can be grouped to give an overview of data cube 
processing solutions (Figure 6 ). Chemometric techniques are grouped by dichotomy 
according the following criteria: univariate or multivariate analysis, strong or weak 
N - way methods, multivariate curve resolution or pattern recognition techniques, 
and supervised or unsupervised classifi cations. 
Univariate and Multivariate Methods Univariate methods belong to classical 
spectroscopy: Depending on the chemical structure of the compound, a specifi c 
wavelength is selected to compute the distribution map. Peak height, area, and ratio 
between two peaks are used to display a false color picture in which the lowest value 
is displayed in black or blue and the highest in white or red. Univariate methods 
are the simplest ways of obtaining distribution maps. However, selection of compound 
- specifi c wavelengths can be diffi cult, especially when samples are complex 
mixtures. The advantage of hyperspectral imaging is that full spectra are available 
in the data cube. Drawing on all the information contained in the data sets rather 
than on a few wavelengths improves chemical map extraction. The techniques used 
in this case are termed multivariate. Multivariate image analysis (MIA) based on 
chemometrics generally improves distribution map extraction. Several types of MIA 
FIGURE 6 Image analysis workfl ow: processing chain. 
I— Data pretreatement 
1. Spectral preprocessing 
e.g., smoothing nomalization, baseline correction, spectral rang selection... 
2. Image preprocessing 
e.g., bad pixel correction, selection of regions of interest, image concatenation... 
II— Pixel classification 
Method selection 
Multivariate methods 
Univariate methods 
e.g., Peak integration... 
Weak N-way methods 
(unfold methods) 
Strong N-way methods 
e.g., PARAFAC, Tucker 3... 
Pattern recognition methods 
Multivariate curve resolution methods 
e.g., MCR-ALS, Simplisma... 
Unsuervised classifications 
e.g., Fuzzy C means, K means... 
Distribution map 
Supervised classifications 
e.g., PLS-DA, SIMCA... 
III— Distribution map analysis 
1. Pixel distribution histogram 
e.g., mean, standard deviation of the pixel values... 
2. Particle statitistics 
e.g., number, size, spatial distribution... 
3. Computation of quantitative values 
e.g., water content... 
CHEMOMETRIC FOR IMAGING 417

418 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
methods are presented below: strong N - way, multivariate curve resolution, and 
pattern recognition methods. 
Multivariate Image Analysis: Strong and Weak Multiway Methods Strong and 
weak N - way methods analyze 3D and 2D matrices, respectively. Hyperspectral data 
cube structure is described using chemometric vocabulary [17] . A two - way matrix, 
such as a classical NIR spectroscopy data set, has two modes: O object (matrix lines) 
and V variables (matrix columns). Hyperspectral data cubes possess two object 
modes and one variable mode and can be written as an OOV data array because 
of their two spatial directions. 
Strong Multiway Methods Strong multiway methods analyze data cubes directly, 
without any matrix rearrangement, whereas weak multiway methods require a prior 
unfolding step. Examples of strong multiway techniques used for hyperspectral 
imaging analysis include parallel factor analysis (PARAFAC) and Tucker 3 [three - 
way principal - component analysis (PCA)] [18] . Their main advantage is that they 
take into account the correlations between image pixels in the OO modes: Unfold 
methods do not use pixel spatial proximity, resulting in neglect of some data cube 
information. 
Even if strong N - way methods are used to reduce image noise, compress data, 
and improve data cube visualization, weak multiway methods are more often used 
as they facilitate classifi cation using classical single - point spectra. 
Classical chemometric methods, that is, the classifi cation and regression presented 
in Section 4.3.1 , are also applied to hyperspectral images. However, X . Y . 
. matrices have to be unfolded into ( X . Y ) . . matrices before processing. In other 
words, the three - way OOV array is unfolded into a classical two - way OV matrix. 
Weak Multiway Methods Figure 7 shows the three steps in weak N - way analysis: 
Unfold the data cube, perform the selected chemometric methods, and refold the 
matrix in order to display distribution maps. Weak N - way analysis comprises two 
main variants: 
Multivariate curve resolution (MCR) has been in common use for 30 years in 
the analytical chemistry community [high - performance liquid chromatography 
(HPLC), FTIR, UV, NIR, and Raman]. It refers to self - modeling mixture 
analysis. Of the various methods, simple - to - use interactive self - modeling 
mixture analysis (SIMPLISMA) [19] and MCR alternating least squares 
(MCR - ALS) [20] appear the most successful with hyperspectral imaging. Their 
aim is to extract the chemical compound natures (so - called pure spectra) and 
concentration profi les from multicomponent systems. Their main advantage is 
that they are calibration free, that is, no prior knowledge is required. 
Pattern recognition can be classifi ed according to several parameters. Below we 
discuss only the supervised/unsupervised dichotomy because it represents two 
different ways of analyzing hyperspectral data cubes. Unsupervised methods 
(cluster analysis) classify image pixels without calibration and with spectra 
only, in contrast to supervised classifi cations. Feature extraction methods [21] 
such as PCA or wavelet compression are often applied before cluster analysis. 

PCA is used to extract features and visualize data. The most important PCA 
application is the reduction in the number of wavelengths. 
Examples of nonhierarchical clustering [22] methods include Gaussian mixture 
models, K means, and fuzzy C means. They can be subdivided into hard and soft 
clustering methods. Hard classifi cation methods such as K means assign pixels to 
membership of only one cluster whereas soft classifi cations such as fuzzy C means 
assign degrees of fractional membership in each cluster. 
Supervised classifi cations use spectral and class membership information [23] . 
Mathematical models are fi rst computed with a calibration set containing spectral 
and class information. This model is then applied to predict new sample classes. 
Supervised pattern recognition algorithms fall into three main categories of contrasting 
pairs. The fi rst comprises methods based on discrimination [e.g., linear discriminant 
analysis (LDA)] and those emphasizing similarity within a class [e.g., soft 
independent modeling of class analogy (SIMCA)]. The second comprises linear and 
nonlinear methods such as neural methods. The third comprises parametric and 
nonparametric computations. In parametric techniques such as LDA, the decision 
rules use statistical parameters of normal sample distribution. In summary, classical 
supervised classifi cation methods are distance - based LDA, SIMCA, and PLS discriminant 
analysis (PLS - DA), which performs regression between spectra and class 
memberships. 
Reference spectra choice is critical when applying supervised pattern recognition 
methods. The fi rst solution is to use pure compound spectra as references. The 
drawback is that mixture spectra in data cubes often differ from the reference 
spectra. Applying the model may therefore give wrong results. The second solution, 
suitable in a few studies, is to select image pixels where only one compound is 
present in order to obtain the calibration sets. 
FIGURE 7 Application of weak N - way methods: ( a ) preprocessing; ( b ) unfold matrix 
analysis; ( c ) refolding. Multivariate curve resolution and pattern recognition techniques. 
. 
D 
Spectral data 
matrix 
3x .2 y 
(a) 
Spectral 
data 
unfolding 
(b) 
. 
y
y 
x x x 
= 
.
Loadings 
Scores 
Score 
data 
refolding 
3x . 2y 
PC
y
y 
x x x 
(c) 
Hyperspectral 
data cube 
data cube 
Scores 
CHEMOMETRIC FOR IMAGING 419

420 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
4.3.5 PRACTICAL AT - LINE AND OFFLINE CHEMICAL IMAGING 
Having reviewed the methods, we will now illustrate their application to pharmaceutical 
situations, with particular respect to univariate methods, PCA, and supervised 
classifi cation (PLS - DA). 
4.3.5.1 Practical Tools for Distribution Map Analysis 
Image Comparison A solution for distribution map analysis is to compare images. 
References such as pure compounds or original samples can be included in the 
image. For example, references and samples can be measured simultaneously if the 
fi eld of view is large enough; otherwise two data cubes can be concatenated, that is, 
grouped together. After the distribution map has been extracted, it can be readily 
interpreted simply by image comparison. In the example shown in Figure 8 , the aim 
was to detect counterfeits. The original samples were compared with the suspected 
counterfeit. After PCA extraction, the differences between the two groups were 
clearly detected. The counterfeits had no active pharmaceutical ingredient (API) 
and the excipients were not identical. A self - calibrating comparison was then 
performed with NIR imaging for fast counterfeit detection. 
Pixel Distribution and Particle Size Determination Quantitative parameters can 
also be computed to analyze a distribution map. The fi rst tool (Figure 9 ) is to display 
the pixel histogram and calculate classical statistics such as mean, standard 
FIGURE 8 Self - calibrating image comparison for counterfeit identifi cation: score images. 
White: higher score. Black: lower score. 
0 10 20 30 40 50 60 70 80 
0
5 
10 
15 
20 
25 
30 
mm 
mm 
mm 
0 10 20 30 40 50 60 70 80 
0
5 
10 
15 
20 
25 
30 
mm 
Original Counterfeit 
PC1: Excipient-OH PC2: API

PRACTICAL AT-LINE AND OFFLINE CHEMICAL IMAGING 421 
deviation, and the parameters of normal distribution, that is, skew and kurtosis. For 
example, a higher mean may be explained by a higher content of a chemical 
compound, and a lower standard deviation may be synonymous with greater homogeneity. 
The second tool is to compute particle size and obtain information about 
spatial particle distribution (Figure 10 ). Several quantitative parameters are 
obtained, such as particle number, particle size, and percentage area covered by the 
particles. Thus the particle distribution statistics provide a measure of sample 
homogeneity. 
4.3.5.2 Wavelength Selection and Chemical Interpretation 
The common aim of the Raman and NIR spectroscopy examples below was to 
display the API and excipient localization on the tablet surface in order to assess 
distribution homogeneity and characterize the solid - state properties of the API. 
FIGURE 9 Image and associated pixel histogram (white: higher absorbance, black: lower 
absorbance). 
Pixels 
50 100 150 200 250 300 
50 
100 
150 
200 
250 
-6 -5 -4 -3 -2 -1 0 1 
0 
500 
1000 
1500 
2000 
2500 
3000 
Numbers 
of pixels 
(a) (b) 
FIGURE 10 Particle size computation with imaging techniques: ( a ) powder NIR image 
(particles are in black); ( b ) statistics. 
mm 
0 0.5 1 1.5 2 2.5 
0 
0.5
1 
1.5
2 
mm 
(a) (b) 
Particle size statistics 
- Number of particles: 209 
- Percentage area covered: 8.7 
- Mean area: 0.02 mm2 
- Area standard deviation: 0.06 mm2 
Particle distribution statistics 
- Mean nearest-neighbordistance: 1.35 mm 
- Center of mass coordinates: X=1.5 ; Y=1

422 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
Example 1 API Mapping and Raman Spectroscopy 
Method 
A Raman microscope (Renishaw, 785 - nm laser, spectral range 800 – 100 cm . 1 ) with a 
line - mapping detector (21 pixels/line) was used to analyze a solid dosage form. The 
image size was 105 . 88 pixels, that is, 325 . m . 270 . m, and acquisition time was 
about 40 min. Spectra were smoothed and normalized. Peak heights were determined 
for the three main compounds — API, lactose, and cellulose (Figure 11 ) — in 
order to create distribution maps. 
Results and Discussion 
Raman bands are narrow and specifi c peak selection is simple. One of the simplest 
classifi cation methods, that is, univariate peak height, could thus be successfully 
applied. The drawbacks of Raman spectroscopy are long acquisition time and the 
effect of line scanning in the images (Figure 11 ). Another issue is fl uorescence. Some 
chemical compounds fl uoresce (e.g., cellulose), obscuring the relevant chemical 
signal and making the distribution map more diffi cult to extract. 
Thus Raman imaging is a useful tool for detecting small API particles on the 
surface of pharmaceutical solid forms. It may even be the most suitable chemical 
imaging technique for API mapping due to its low spatial resolution (up to 0.5 . m/ 
pixel) and the polymorphism of the spectral information. 
Example 2 Tablet Reconstruction by NIR Imaging 
Method 
The tablet was cut lengthwise using a trimmer to leave a plane surface with the 
tablet coating removed (Figure 12 ). Sample and references were analyzed using 
a chemical imaging NIR spectrometer (Sapphire, Malvern) with the following acquisition 
parameters: detector size 320 . 256 array, spectral range 1100 – 2450 nm, and 
spatial resolution 40 . m/pixel. Acquisition time was about 5 min. 
Results and Discussion 
After the second derivative, wavelengths could be selected giving contrast images 
and displaying the localization of mannitol, API, and crospovidone (Figures 13 and 
14 ) with the NIR images obtained at specifi c wavelengths. In this example, chemometrics 
can help defi ne concentration maps by being able to use the information 
present in the whole spectra and not only at specifi c wavelengths. For example, the 
MCR - ALS technique was able to extract fi ve compounds whereas wavelength selection 
could only display three components due to peak overlapping in the NIR range. 
However, simpler methods are also fast and easy to use. Because the aim was to 
localize the API, the wavelength selection method provided the expected results. 
Images were displayed at specifi c wavelengths (peak - height method). The single - 
wavelength images were then binarized, in an operation similar to the transformation 
of a gray - scale image into a black - and - white image. A color image resulted (Figure 
14 ). The red channel was associated with the API, the blue channel with mannitol, 
and the green channel with crospovidone. Tablet reconstruction was then possible. 
This highlights the main advantage of the spectroscopic technique: the large area of 
analysis, meaning that the images are more representative of the sample. 

PRACTICAL AT-LINE AND OFFLINE CHEMICAL IMAGING 423 
FIGURE 11 Raman images at specifi c wavelengths and reference spectra (image size 
325 . m . 270 . m; white: higher absorbance, black: lower absorbance). 
API spectra 
Lactose spectra 
Cellulose spectra 
API 
276 cm-1 
0.1 
0.2 
0.3 
0.4 
0.5 
0.6 
0.7 
0.8 
0.9 
Cellulose 
436 cm-1 
Lactose 
473 cm-1 
Raman reference spectra Raman images 
Counts Counts Counts 
Raman shift (cm–1) 
Raman shift (cm–1) 
Raman shift (cm–1) 
700 600 500 400 300 200 
700 600 500 400 300 200 
700 600 500 400 300 200 
2000 
4000 
6000 
8000 
10000
0 
200 
400 
600 
800 
1000 
1200 
400 
600 
800 
1000 
1200
0 
FIGURE 12 Sample and cutting holder. 

424 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
FIGURE 13 Images at component - specifi c wavelengths and second derivative spectra of 
high - absorption pixel (image size: 1.02 cm . 1.3 cm; white: higher absorbance, black: lower 
absorbance). 
Mannitol 
2080 nm 
API 
2260 nm 
Crospovidone 
1930 nm 
1500 2000 2500 
1500 2000 2500 
1000 1500 2000 2500 
Wavelength (nm) 
(a) (b) 
FIGURE 14 Tablet reconstruction. Gray (normally red): API. Dark gray (normally blue): 
mannitol. Light gray (normally green): crospovidone. Black: other. Image size 1.02 cm . 
1.3 cm. 

PRACTICAL AT-LINE AND OFFLINE CHEMICAL IMAGING 425 
4.3.5.3 Unsupervised Classifi cation for Process Troubleshooting 
The aim of the example was to compare tablets with respect to “ good ” and “ bad ” 
dissolution properties using IR imaging and to identify the root causes of bad 
dissolution [24] . 
Methodology Six samples were analyzed: three passed the dissolution test (good 
samples) and three failed (bad samples). Several ingredients were used as references: 
Avicel (cellulose), the API, magnesium stearate, and poloxamer. Dissolution 
was tested after NIR imaging measurement to obtain the wet chemical references. 
An Equinox 55 spectrometer coupled to a Hyperion 3000 microscope equipped 
with a 64 . 64 MCT FPA detector (Bruker, Ettlingen, Germany; www.brukeroptics. 
com ) was used to acquire IR spectra between 3900 and 900 cm . 1 at 16 cm . 1 resolution 
(i.e., 376 data points) under N 2 purge. The pixel - grouping binning function was 
applied to improve the signal, and a 16 . 16 pixel image was fi nally acquired. The 
number of scans was 20 and the area of analysis per FPA measurement was 270 . m 
. 270 . m. Two FPA images were recorded per tablet, that is, a 32 . 16 pixel image 
per tablet. The three data cubes of the good samples were concatenated in the Y 
dimension to obtain a 32 . 48 image. The same 32 . 48 image was produced for the 
bad samples. The two sets of good and bad samples were then concatenated in the 
X dimension to obtain a 64 . 48 image. The fi nal data cube was a 64 . 48 . 376 
matrix. 
Results and Discussion In the PCA results (Figures 15 and 16 ), loadings were 
interpreted using raw material reference spectra. The fi rst loading was attributed to 
poloxamer, the second to magnesium stearate (Figure 16 ), the third to the API, and 
the fourth to Avicel. However, the other ingredients could also have contributed to 
the loadings because PCA loadings are not pure component spectra. PCA extracts 
orthogonal signals, which is not the case for reference spectra. In particular, the PC3 
loading attributed to the API could have been contaminated by other components, 
especially magnesium stearate. This confi rms the drawback of feature extraction 
methods. 
FIGURE 15 PCA score images: PC2: magnesium stearate; PC3: API; PC4: Avicel; PC1: 
poloxamer, not discriminant and not displayed (white: higher absorbance, black: lower 
absorbance). 
Pixels 
Pixels 
10 20 30 40 50 60 
10 
20 
30 
40 
50 
60 
Pixels 
Pixels 
10 20 30 40 50 60 
10 
20 
30 
40 
50 
60 
Pixels 
Pixels 
10 20 30 40 50 60 
10 
20 
30 
40 
50 
60 
PC3 
Good samples Bad samples Good samples Bad samples Good samples Bad samples 
PC2 PC4 
Sample 1 
Sample 2 
Sample 3 
Sample 5 
Sample 6 
Sample 7 
Sample 4 Sample 8

426 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
This PCA image study showed the differences between the two data sets, that is, 
it separated the good and bad dissolution samples. The main differences were due 
to the distribution of magnesium stearate and the API. No differences were observed 
in the spatial distribution of poloxamer or Avicel. Magnesium stearate is hydrophobic, 
thus protecting the tablet core from moisture and hence slowing dissolution. 
When a sample had more active ingredient on the surface, the dissolution properties 
were increased. 
Peak height is a classical method of interpreting IR spectra. Its main advantage 
is in the selection of specifi c wavelengths. The disadvantage is that the bands often 
overlap and fi nding a specifi c band becomes problematic. PCA solves the problem 
of wavelength selection. Its main advantage is that it reduces the number of 
variables, that is, the number of images to analyze. The disadvantage lies in the 
interpretation of the loadings, which differ from pure substance spectra. Interpretation 
can be complicated by the fact that several chemical species may contribute to 
one loading. 
4.3.5.4 Supervised Classifi cation, NIR Imaging, and Process Development 
The aim of this study was to use NIR imaging to solve granulation issues in new 
formulation development. Undesired powder agglomerations developed during the 
granulation step (Figure 17 ). Imaging was applied to characterize the agglomeration 
structure. 
Method The sample contained starch, API, Avicel, crospovidone, and sodium 
lauryl sulfate. Sample and references were analyzed in triplicate by NIR imaging 
(Spectral Dimensions, 20 coadds, spectral range 1100 – 2450 nm). Full image size was 
320 . 256 pixels or 4.1 . 3.3 mm. The NIR images were interpreted, and sample raw 
materials mapped, using PLS classifi cation with fi ve loadings (based on the reference 
spectra for starch, API, Avicel, crospovidone, and sodium lauryl sulfate). 
Results and Discussion The PLS model identifi ed all fi ve chemical species. PLS 
multivariate analysis showed (Figure 18 ) that the core contained Avicel and API 
and that the periphery contained starch and crospovidone. The solution to the 
granulation issue was to add a premixing step to avoid agglomeration. NIR imaging 
proved useful for improving process understanding. 
FIGURE 16 PCA loadings and comparison with reference spectra: ( a ) PC2, magnesium 
stearate; ( b ) PC3, active ingredient; ( c ) PC4, Avicel. 
PC4 
Avicel 
PC3 
-
- 
PC2 
MgSt Active 
(a) (b) (c) 
loadings Reference

PRACTICAL AT-LINE AND OFFLINE CHEMICAL IMAGING 427 
FIGURE 17 Powder agglomeration: visible picture. 
FIGURE 18 PLS classifi cation image (three replicates and fi ve chemical compounds). 
Simulation of granulation issues. 
Avicel 
Sodium 
lauryl sulfate 
API 
Starch 
Crosspovidone 
Image 1 Image 2 Image 3 
0 20 40 60 80 100 120 
mm

428 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
The advantage of supervised classifi cation is that it avoids a wavelength selection 
step or the interpretation of PCA loading. It also rapidly extracts several chemical 
components. These methods provide accurate results provided the sample spectra 
are similar to the reference spectra. In our case the powder agglomeration was 
heterogeneous and the layers had a high content of excipient, making it possible to 
apply supervised classifi cation. 
4.3.5.5 Further Developments: Online Analysis by Hyperspectral Imaging 
The two main online applications of hyperspectral imaging are blending endpoint 
determination (Figure 19 ) and capsule control (Figure 20 ). Other applications may 
be possible, such as content uniformity, determination of dissolution properties, and 
water quantifi cation, but are not described in this chapter. 
In blending, the main imaging advantage over single - point spectroscopy is the 
ability to analyze a larger area. We studied blends of three compounds (Figure 19 ), 
selecting single wavelengths for Avicel, API, and lactose. Only Avicel was present 
in the sample at the beginning of the experiment. Avicel and API were detected 
after three blender rotations; the two maps were complementary. All compounds in 
the mixture were homogeneous after 27 rotations. Timed distribution map analysis 
is thus a blending monitoring tool. Our example shows the IR images. However, 
NIR imaging has proved highly successful in the literature [25] . 
In capsule control, NIR radiation penetrates the shell, enabling the fi lling to be 
checked. The capsules in our example (Figure 20 ) were blue and opaque. However, 
FIGURE 19 IR imaging for blend monitoring (r = blender rotation; white: higher 
absorbance, black: lower absorbance). 
Active Lactose Avicel
661r 27r 7r 3r 0 
661r 27r 7r 3r 0 
661r 27r 7r 3r 0 
Blending process

NIR imaging was able to detect empty capsules and monitor fi llings, including pellet 
appearance. 
4.3.6 CONCLUSIONS 
Hyperspectral imaging provides information that is spatial and spectral and both 
qualitative and quantitative. It can map chemical compound distribution and determine 
particle size. Such information cannot be obtained using classical spectroscopy. 
We have applied the method to the solution of quality control and process problems 
affecting pharmaceutical tablets: dissolution, polymorph distribution, moisture 
content determination, API localization and characterization, content uniformity, 
blending, and granulation. The choice of imaging technique is based on several 
criteria: spatial and spectral resolution, measurement time, and wavelength 
range. Tables 1 and 2 summarize the main advantages and disadvantages of the 
different methods. 
FIGURE 20 Capsule control by NIR imaging (image size 33 mm . 41 mm). Simulation of 
empty capsule in blister. 
TABLE 1 Instrumentation Comparison 
Parameters FT System Dispersive 
Liquid Crystal 
Tunable Filter 
Spectroscopy NIR, IR, Raman Raman, NIR Raman, NIR 
Image acquisition Mapping, FPA detector Mapping FPA detector 
Measurement mode Refl exion, transmission, 
ATR 
Refl exion, transmission Refl exion 
Image size Microscopic, macroscopic Microscopic Microscopic, 
macroscopic 
Acquisition time Slow Slow Fast 
CONCLUSIONS 429

430 USEFUL TOOLS FOR PROCESS ANALYTICAL TECHNOLOGY 
TABLE 2 Comparison of Chemical Imaging Methods 
Parameters Raman Imaging Infrared Imaging Near - Infrared Imaging 
Instrumentation 
robustness 
LCTF: +++ 
other: + 
+ +++ 
Spectral information 
specifi city 
+++ +++ ++ 
Constituant map 
extration 
Easy Easy Need chemometrics 
Pharmaceutical 
application 
examples 
Polymorph screening, 
small API particle 
detection 
Unknown particle 
identifi cation 
Tablet reconstruction, 
blend homogeneity, 
process 
troubleshooting, 
counterfeit 
identifi cation 
Spectral imaging is a complex and multidisciplinary fi eld. The introduction of new 
FPAs is making it increasingly powerful and attractive. It has proven potential in 
qualitative pharmaceutical analysis and can be used when spatial information 
becomes relevant for an analytical application. Even if online applications and regulatory 
method validation require further study, the potential contribution of imaging 
to quality control and PAT needs no further demonstration. 
ACKNOWLEDGMENTS 
We would like to thank Anton Fischer (QC Manager, Hoffmann - La Roche) and 
Rolf Altermatt (Section Head, Hoffmann - La Roche) for providing the resources 
that made this study possible and Christelle Gendrin (Ph.D. student, Hoffmann - La 
Roche) for her greatly appreciated help. 
REFERENCES 
1. Roggo , Y. , Roeseler , C. , and Ulmschneider , M. ( 2004 ), Near infrared spectroscopy 
for qualitative comparison of pharmaceutical batches , J. Pharm. Biomed. Anal. , 36 , 
777 – 786 . 
2. Chalus , P. , Roggo , Y. , Walter , S. , and Ulmschneider , M. ( 2005 ), Near - infrared determination 
of active substance content in intact low - dosage tablets , Talanta , 66 , 1294 – 1302 . 
3. Roggo , Y. , Duponchel , L. , and Huvenne , J. - P. ( 2003 ), Comparison of supervised pattern 
recognition methods with McNemar ’ s statistical test: Application to qualitative analysis 
of sugar beet by near - infrared spectroscopy , Anal. Chim. Acta , 477 , 187 – 200 . 
4. El - Hagrasy , A. S. , Morris , H. R. , D ’ Amico , F. , Lodder R. A. , and Drennen , J. K. ( 2001 ), 
Near - infrared spectroscopy and imaging for the monitoring of powder blend homogeneity 
, J. Pharm. Sci. , 90 , 1298 – 1307 . 
5. Beleites , C. , Steiner , G. , Sowa , M. G. , Baumgartner , R. , Sobottka , S. , Schackert , G. , and 
Salzer , R. ( 2005 ), Classifi cation of human gliomas by infrared imaging spectroscopy and 
chemometric image processing , Vibrat. Spectrosc. , 38 , 143 – 149 . 

6. Chalmers , J. M. , Everall , N. J. , Schaeberle , M. D. , Levin , I. W. , Neil Lewis , E. , Kidder , L. 
H. , Wilson , J. , and Crocombe , R. ( 2002 ), FT - IR imaging of polymers: An industrial 
appraisal , Vibrat. Spectrosc. , 30 , 43 – 52 . 
7. Juan , A. D. , Tauler , R. , Dyson , R. , Marcolli , C. , Rault , M. , and Maeder , M. ( 2004 ), Spectroscopic 
imaging and chemometrics: A powerful combination for global and local sample 
analysis , TrAC Trends Anal. Chem. , 23 , 70 – 79 . 
8. Lewis , E. N. , Schoppelrei , J. , and Lee , E. ( 2004 ), Near - infrared chemical imaging and the 
PAT initiative , Spectroscopy , 19 , 26 – 36 . 
9. Lewis , E. N. , Lee , E. , and Kidder , L. H. ( 2004 ), Combining imaging and spectroscopy: 
Solving problems with near infrared chemical imaging , Microscopy Today , 12 , 8 – 12 . 
10. Lewis , N. , Schoppelrei , J. , Lee , E. , and Kidder , L. ( 2005 ), Near - infrared chemical imaging 
as a process analytical tool , in Bakeev , K. A. , Ed., Process Analytical Technology , 
Blackwell , London , pp. 187 – 225 . 
11. Lewi , P. J. ( 2005 ), Spectral mapping, a personal and historical account of an adventure in 
multivariate data analysis , Chemomet. Intell. Lab. Syst. , 77 , 215 – 223 . 
12. Tran , C. D. ( 2003 ), Infrared multispectral imaging: Principles and instrumentation , Appl. 
Spectrosc. Rev. , 38 , 133 – 153 . 
13. Gift , A. D. , Ma , J. , Haber , K. S. , McClain , B. L. , and Ben - Amotz , D. ( 1999 ), Near - infrared 
Raman imaging microscope based on fi ber - bundle image compression , J. Raman 
Spectrosc. , 30 , 757 – 765 . 
14. Burns , D. A. , Ciurczak , E. W. ( 2001 ), Handbook of Near - Infrared Analysis , 2nd ed. , rev. 
and expanded, CRC Press, New York . 
15. Gat , N. ( 2000 ), Imaging spectroscopy using tunable fi lters: A review , Proc. SPIE , 4056 , 
50 – 64 . 
16. Russ , J. ( 2002 ), The Image Processing Handbook , 4th ed. , CRC Press , London . 
17. Huang , J. , Wium , H. , Qvist , K. B. , and Esbensen , K. H. ( 2003 ), Multi - way methods in image 
analysis — Relationships and applications , Chemomet. Intell. Lab. Syst. , 66 , 141 – 158 . 
18. Bro , R. ( 2003 ), Multivariate calibration: What is in chemometrics for the analytical 
chemist? Anal. Chim. Acta , 500 , 185 – 194 . 
19. De Braekeleer , K. , and Massart , D. L. ( 1997 ), Evaluation of the orthogonal projection 
approach (OPA) and the SIMPLISMA approach on the Windig standard spectral data 
sets , Chemomet. Intell. Lab. Syst. , 39 , 127 – 141 . 
20. Tauler , R. ( 1995 ), Multivariate curve resolution applied to second order data , Chemomet. 
Intell. Lab. Syst. , 30 , 133 – 146 . 
21. Naes , T. , and Martens , H. ( 1991 ), Multivariate Calibration , Wiley , New York . 
22. Massart , D. L. , Vandeginste , B. G. M. , Buydens , L. M. C. , Jong , S. D. , Lewi , P. J. , and 
Smeyers - Verbeke , J. ( 1997 ), Handbook of Chemometrics and Qualimetrics: Part A, Data 
Handling in Science and Technology, 20A , Elsevier, Amsterdam. 
23. Massart , D. L. , Vandeginste , B. G. M. , Deming , S. M. , Michotte , Y. , and Kaufman , L. ( 2003 ), 
Chemometrics: A Textbook, Data Handling in Science and Technology , 2, Elsevier, 
Amsterdam . 
24. Roggo , Y. , Edmond , A. , Chalus , P. , and Ulmschneider , M. ( 2005 ), Infrared hyperspectral 
imaging for qualitative analysis of pharmaceutical solid forms , Anal. Chim. Acta , 535 , 
79 – 87 . 
25. Lyon , R. C. , Lester , D. S. , Lewis , E. N. , Lee , E. , Yu , L. X. , Jefferson , E. H. , and Hussain , 
A. S. ( 2002 ), Near - infrared spectral imaging for quality assurance of pharmaceutical 
products: Analysis of tablets to assess powder blend homogeneity , AAPS PharmSciTech 
[Electronic Resource], 3 , E17 . 
REFERENCES 431


SECTION 5 
PERSONNEL 


435 
5.1 
PERSONNEL TRAINING IN 
PHARMACEUTICAL 
MANUFACTURING 
David A. Gallup , Katherine V. Domenick , and Marge Gillis 
Training and Communications Group, Inc., Berwyn, Pennsylvania 
Contents 
5.1.1 Overview 
5.1.1.1 Background 
5.1.1.2 Training Requirements 
5.1.1.3 Good Training Practices and Pharmaceutical Manufacturing 
5.1.1.4 What Is Competency - Based Training? 
5.1.1.5 Why Is Competency - Based Training So Important? 
5.1.2 Developing a Training Plan: Strategy to Ensure Training Compliance in a Pharmaceutical 
Manufacturing Facility 
Section 1 Training Organization 
Section 2 Training Programs 
Section 3 Training Development 
Section 4 Training Implementation 
Section 5 Training Recordkeeping 
References 
5.1.1 OVERVIEW 
5.1.1.1 Background 
By law and by ethical commitment, companies that manufacture pharmaceutical 
products must make sure that what they produce is safe and effective. Ensuring that 
pharmaceutical manufacturing personnel possess the competencies necessary 
to perform their jobs correctly and effi ciently is critical to a safe and successful 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

436 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
manufacturing process. Developing required skills, providing discrete knowledge, 
and instilling an ethical and responsible approach to work are critical to training in 
an environment centered on good manufacturing practices. 
This chapter includes the following types of pharmaceutical manufacturing 
organizations: 
• Active pharmaceutical ingredient (API) or bulk manufacturers 
• Biotechnology manufacturers 
• Traditional manufacturers, solid and liquid dose 
• Vaccine manufacturers 
This chapter reviews the requirements for training personnel in a pharmaceutical 
manufacturing environment; it then focuses on how to develop and implement a 
training strategy that ensures pharmaceutical manufacturers are in compliance with 
the mandated Code of Federal Regulations (CFR), the current Food and Drug 
Administration (FDA) quality systems draft guidance document and good training 
practices (GTPs). 
5.1.1.2 Training Requirements 
The training requirements for personnel working in a pharmaceutical manufacturing 
environment are specifi ed in (21 CFR) 211.25 [1] : 
(a) Each person engaged in the manufacture, processing, packing or holding of a drug 
product shall have education, training, and experience, or any combination thereof, to 
enable that person to perform the assigned functions. Training shall be in the particular 
operations that the employee performs and in current good manufacturing practices 
(including the current good manufacturing practice regulations in this chapter and 
written procedures required by these regulations) as they relate to the employee ’ s 
functions. Training in current good manufacturing practice shall be conducted by quali- 
fi ed individuals on a continuing basis and with suffi cient frequency to assure that 
employees remain familiar with cGMP requirements applicable to them. 
(b) Each person responsible for supervising the manufacture, processing, packing, or 
holding of a drug product shall have the education, training, and experience, or any 
combination thereof, to perform assigned functions in such a manner as to provide 
assurance that the drug product has the safety, identity, strength, quality, and purity 
that it purports or is represented to possess. 
(c) There shall be an adequate number of qualifi ed personnel to perform and supervise 
the manufacture, processing, packing, or holding of each drug product. 
In summary, the Code of Federal Regulations states that pharmaceutical manufacturing 
personnel must be trained as follows: 
1. To do their specifi c jobs 
2. In current good manufacturing practices (cGMPs) 
3. On written procedures (required in 21 CFR 211 : 80) 
The regulation also goes on to specify that training must fulfi ll the following 
conditions: 

1. Be provided by qualifi ed individuals 
2. Be conducted on a frequent, continuing basis 
Finally, the code states that manufacturing supervisors must be trained in the same 
manner as their employees. 
In essence, the code specifi es that pharmaceutical manufacturing companies must 
train their employees using qualifi ed trainers; it also states that supervisors must 
have the same training as “ qualifi ed personnel. ” What the regulations do not say 
is how to conduct this training. 
According to John W. Levchuck, then of the FDA, “ The FDA has not published 
a guideline establishing acceptable procedures for personnel training, nor is a guideline 
being planned . Neither has the FDA specifi ed strict training requirements ” [2] . 
Levchuck ’ s statement may be on its way to being updated. The latest pharmaceutical 
manufacturing training direction comes from the FDA ’ s draft guidance document 
“ Guidance for Industry: Quality Systems Approach to Pharmaceutical Current 
Good Manufacturing Practice Regulations ” [3] . Published in 2004 this guidance 
document states in its develop personnel section: 
Under a quality system, senior management is expected to support a problem - solving 
and communicative organizational culture. Managers are expected to encourage communication 
by creating an environment that values employee suggestions and acts on 
suggestions for improvement. Management is also expected to develop cross - cutting 
groups to share ideas to improve procedures and processes. 
In the quality system, it is recommended that personnel be qualifi ed to do the operations 
that are assigned to them in accordance with the nature of, and potential risk to 
quality presented by, their operational activities. Under a quality system, managers are 
expected to defi ne appropriate qualifi cations for each position to help ensure individuals 
are assigned appropriate responsibilities. Personnel should also understand the 
impact of their activities on the product and the customer (this quality systems parameter 
is also found in the CGMP regulations, which identify specifi c qualifi cations, i.e., 
education, training, and experience or any combination thereof; see 21 CFR 211.25(a) 
& (b)). Under a quality system, continued training is critical to ensure that the employees 
remain profi cient in their operational functions and in their understanding of 
CGMP regulations. Typical quality systems training would address the policies, processes, 
procedures, and written instructions related to operational activities, the product/ 
service, the quality system, and the desired work culture (e.g., team building, communication, 
change, behavior). 
Under a quality system (and the CGMP regulations), training is expected to focus on 
both the employees ’ specifi c job functions and the related CGMP regulatory 
requirements. 
Under a quality system, managers are expected to establish training programs that 
include the following: 
• Evaluation of training needs 
• Provision of training to satisfy these needs 
• Evaluation of effectiveness of training 
• Documentation of training and/or re - training 
OVERVIEW 437

438 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
When operating in a robust quality system environment, it is important that supervisory 
managers ensure that skills gained from training be incorporated into day - to - day 
performance. 
The quality systems draft guidance document reinforces the training requirements 
contained in 21 CFR 211.25 and goes on to add several key items. These 
include an evaluation of training needs, an evaluation of training effectiveness, and 
a documentation requirement. Once again, however, the document does not describe 
how training is to be designed, conducted, or evaluated. In addition, the guidance 
document does not provide details on the specifi c training information to be collected, 
how it should be stored, or how long the documentation records should be 
retained. 
Managers have started to view training as a business - critical component in achieving 
improved performance and compliance with regulations, but at the same time 
they may have questions about how to direct training efforts. In the absence of fi rm 
guidelines for training, many in the industry have interpreted FDA commentary and 
audit results as supporting a competency - based approach to training, with validated 
and reliable training programs that produce measurable performance outcomes. 
Work responsibilities and tasks are specifi ed in standard operating procedures, 
guidelines, batch records, employee directives, and protocol. Training must ensure 
that all employees know of the existence of all these documents, how to access them, 
and how they are used to direct work. Further, “ qualifi ed personnel ” must demonstrate 
that they have read and understand these documents and can perform work 
as directed by them. 
5.1.1.3 Good Training Practices and Pharmaceutical Manufacturing 
Good training practices and a competency - based approach to training may fi rst have 
been introduced to the general manufacturing industry in 1961 with the publication 
of a booklet by Robert Mager [4] . In this work, Mager tells educators and trainers 
how to follow a systematic approach to develop training materials and programs so 
students and trainees know what competencies they should possess after training. 
The idea of developing systematic, competency - based training in the pharmaceutical 
industry has its roots in an article published in 1982 by Ronald Tetzlaff [5] , then 
of the FDA, which builds on the work done by Mager and others in the training 
fi eld by explaining how a systematic approach to training program design is the 
best way to build effective, consistent training for employees in pharmaceutical 
manufacturing. 
5.1.1.4 What Is Competency - Based Training? 
Competency - based training is training aimed at ensuring that certain competencies 
(skills and knowledge) are achieved. Competency - based training measures trainees ’ 
mastery of materials through a test or skill demonstration or both to ensure that 
trainees have acquired competencies. For example, a competency - based program 
for tablet operators may be designed to ensure that operators can follow all steps 
in equipment start - up, operation, shut - down, and troubleshooting. The test to ensure 
that trainees acquired thoses competencies during training might include labeling 

DEVELOPING A TRAINING PLAN 439 
a diagram of the equipment and performing a demonstration of how to start up, 
operate, shut down, and troubleshoot the equipment. 
5.1.1.5 Why Is Competency - Based Training So Important? 
Competency - based training is important because it fulfi lls FDA training guidelines 
while at the same time meeting business needs. Businesses need a trained, competent 
workforce capable of manufacturing products effectively and effi ciently. How 
effectively does pharmaceutical manufacturing respond to the challenge? According 
to one report, “ A lack of trained and experienced technical and production staff will 
have an impact on more than half of the world ’ s biopharmaceutical developers and 
contract manufacturers in the next fi ve years, and impact their ability to meet 
demand, according to a survey of 100 international biopharmaceutical manufacturers 
and contract manufacturing organizations ” [6] . 
From a compliance perspective, interest in training is also starting to rise: “ The 
U.S. FDA is paying more attention to drug manufacturers ’ training programs under 
its quality system approach to inspections. While personnel training is not the hottest 
compliance topic in pharmaceutical GMP literature, it is a frequent U.S. FDA investigator 
observation, is appearing on a growing percentage of warning letters, and 
has been involved in some of the highest profi le FDA regulator actions in recent 
years ” [7] . So competency - based training makes sense from both business needs and 
compliance requirements. A well - thought - out, effi ciently executed training plan can 
help ensure pharmaceutical companies not only meet the intent of regulations but 
also show a return on money invested in training. 
5.1.2 DEVELOPING A TRAINING PLAN: STRATEGY TO ENSURE 
TRAINING COMPLIANCE IN A PHARMACEUTICAL 
MANUFACTURING FACILITY * 
Ensuring training compliance within a pharmaceutical manufacturing facility 
requires the facility to establish and implement a comprehensive training plan. 
Set up and executed properly, this plan will encompass and meet all of the training 
requirements of: 
• 21 CFR 210.25 
• “ Guidance for Industry: Quality Systems Approach to Pharmaceutical Current 
Good Manufacturing Practice Regulations ” 
• Good training practices 
Managers in many organizations have an idea of what they want their training to 
look like, but few have a well - defi ned, written training plan to help them get there. 
A training plan is a roadmap that guides and integrates the training process throughout 
the facility. This plan helps ensure a facility accomplishes its goals consistently, 
according to its standards of quality, production, cost, and safety. The plan should 
* The material in this section was originally published in 1999 in the PDA Journal [8] . The information 
is updated and expanded and examples of the material described are presented. 

440 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
carry the full weight and authority of other corporate directives with the highest 
levels of management support for training as an essential component in the manufacture 
of safe and effective health care products. 
Why develop a training plan rather than a standard operating procedure (SOP) 
that covers training? Some managers believe that such a procedure is “ required ” 
[9] . Remember, though, that developing a training SOP — or any SOP — mandates 
training on that SOP. Ideally, the training plan should be more comprehensive than 
what is normally captured in an SOP. The training plan is a document that guides 
the development of training programs and materials and the way those programs 
and materials are implemented. A training plan includes training policy and goes 
beyond a procedure. And because it is not a procedure, it need not be trained. 
A training plan might look like the schematic in Figure 1 . The plan has fi ve 
sections. 
Note : The facility should have a master training plan that covers the entire site. 
Each department should develop its own training plan, including the training needed 
to qualify personnel for all job functions. The departmental plan should provide a 
schedule of training and retraining to ensure the availability of a suffi cient number 
of qualifi ed personnel to perform all job functions at all times. 
A checklist for ensuring all components of a training plan are identifi ed is shown 
in Figure 2 . 
Section 1 Training Organization 
This section of the training plan calls for the development of a training philosophy 
and mission statement. These provide the overall foundation for a strong training 
FIGURE 1 Training plan schematic. 
Section 5. 
Training 
recordkeeping 
Section 4. 
Training 
implementation 
Section 3. 
Training 
development 
Section 2. 
Training 
programs 
Section 1. 
Training 
organization 
Training 
plan 

DEVELOPING A TRAINING PLAN 441 
organization. In addition, a training organization must be established with persons 
designated to support the training effort throughout the facility. 
1.1 Develop a Training Philosophy Once the decision is made to develop a 
training plan, those involved need to formulate a training philosophy. Most organizations 
have developed philosophies around production, quality, cost, safety, and 
other aspects of their operation. Ask an employee what the company ’ s philosophy 
is on meeting production quotas, and he or she can probably sum it up in a few 
words. But how many organizations have developed a well - articulated training philosophy? 
The answer is very few. 
A training philosophy that acknowledges training as a critical component in 
achieving corporate business objectives is a key element in establishing the credibility 
of training in an organization. A training philosophy represents an organization ’ s 
commitment to training. It tells employees, suppliers, customers, and other stakeholders 
the organization is committed to ensuring its people have the skills and 
knowledge needed to compete in an increasingly competitive environment. A training 
philosophy should acknowledge that training is conducted not only to meet 
regulatory requirements but also because properly trained personnel contribute to 
the quality of the operation and fi nancial success of the company. Figure 3 provides 
an example of a training philosophy. 
1.2 Develop a Training Mission Statement The training philosophy provides the 
basis for the training mission statement. The mission statement identifi es the overall 
goals of organizational training and emphasizes the importance of establishing 
personnel standards of performance. It also positions training as critical to quality 
FIGURE 2 Training plan checklist. 
1. Training organization 
1.1. Develop a training philosophy 
1.2. Develop a training mission statement 
1.3. Develop a training organization 
1.4. Develop a training support network 
2. Training programs 
2.1. Identify all training programs/personnel to be trained 
2.2. Establish personal qualification pathways 
2.3. Training program materials 
3. Training development 
3.1. Establish a training program design model 
3.2. Ensure development of valid training programs 
3.3. Establish a change control methodology 
4. Training implementation 
4.1. Develop an implementation schedule 
4.2. Establish a training failure response process 
5. Training recordkeeping 
5.1. Establish a recordkeeping system 
5.2. Training record requirements 

442 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
outcomes. Representatives from every department should participate in drafting a 
company training mission statement to ensure it refl ects their interests. A sample 
training mission statement is presented in Figure 4 . 
Once the training philosophy and mission statement are developed, they need to 
be communicated to all employees to solicit feedback and support. The communication 
should focus on the direction the company is taking and the role managers will 
play in ensuring training is carried out effectively in the organization. A clearly 
worded philosophy and mission statement set the groundwork for building a plan 
that integrates training into all areas of personnel development. 
1.3 Develop a Training Organization If it has not already been set up, a training 
organization needs to be established by the corporate or facility management team. 
FIGURE 3 Sample training philosophy. 
We are committed to a strong training organization as supported by our mission 
statement. Our training philosophy includes the following elements: 
• Commitment to preparing all employees to do their jobs competently and efficiently 
to produce products that consistently meet or exceed customer requirements for 
safety and efficacy 
• Commitment to continuous training of all employees throughout the organization so 
they may reach their full human potential 
FIGURE 4 Sample training mission statement. 
Our training mission statement is based on our training philosophy. We will… 
• Support all corporate objectives in the business plan by providing excellent training 
programs to all employees 
• Evaluate the efficacy of those programs on an ongoing basis 
• Grow and develop our employees in all areas of their lives, including interpersonal 
and technical areas 
• Commit to support the organization as a top-tier growth company that can 
adapt to change in an ongoing regulated environment 
• Provide ongoing training to ensure all personnel are in compliance with corporate 
and regulatory guidelines 
• Provide resources for employee growth while meeting the organization’s quality, 
productivity, safety, and financial goals

DEVELOPING A TRAINING PLAN 443 
Our experience has shown that training is most effective when the training organization 
or department is a facility function — on the same level as other major departments 
within the manufacturing facility. A typical organization chart for the training 
department within a pharmaceutical manufacturing facility might look like the one 
in Figure 5 . 
Roles and responsibilities should be defi ned in the plan as follows. Depending 
upon the size of the facility, some of the roles may be combined. 
Training and Development Manager Since the goal of training is to ensure that 
personnel are capable of performing their jobs to produce consistent, quality products, 
the training function must support all other functions and be supported by 
them. The training and development manager must have lines of communication 
with all other functions, including quality, safety, production, information systems, 
and laboratory. Although many organizations agree on this in principle, the practice 
is likely to be different. 
The training and development manager should direct, monitor, and build support 
for the training effort throughout the organization. The position should meet regularly 
with site management to review the training plan of action and the progress 
to date and ensure the training effort supports business and compliance goals. He 
or she also directs training coordinators and assists in communicating and developing 
training plans and outcome assessments for the departments to ensure they are 
implemented effectively. 
This function should extend to departmental training to ensure that systems exist 
in each department to meet specifi c training requirements. The training and development 
manager should develop guidelines to be followed by each department for 
complying with the corporate training policy. These guidelines will identify the training 
coordinators, instructional designers, departmental trainers, and subject matter 
experts (SMEs), as well as persons responsible for scheduling and recording participation 
in training. 
The training and development manager should be responsible for at least the 
following: 
FIGURE 5 Sample training department structure. 
and 
d 
d

444 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
• Developing, maintaining, and administering the facility training plan 
• Identifying training needs for each department to support site business objectives 
and compliance efforts 
• Developing training plans and schedules to meet organizational training 
needs 
• Reviewing training plans and schedules to ensure support for training from all 
departments 
• Establishing metrics to measure training effectiveness 
• Directing and monitoring training coordinators in executing training plans 
Safety Trainer(s) Safety trainers are those individuals that provide safety training 
to the manufacturing employees. The safety trainers are responsible for at least the 
following: 
• Working with instructional designers to develop safety training programs 
• Presenting required safety training to appropriate departments, such as: 
Lockout/tagout 
Confi ned space entry 
Hearing protection 
Respiratory protection 
Bloodborne pathogens 
• Identifying ongoing safety training needs 
Management Trainer(s) Management trainers are individuals that provide “ soft 
skill ” training programs to the other employees at the pharmaceutical manufacturing 
site. The management trainers are responsible for at least the following: 
• Working with instructional designers to develop management development 
programs 
• Presenting management development programs, such as: 
Leadership 
Problem solving/decision making 
Coaching and counseling 
Time management 
GXP Trainer(s) GXP trainers are primarily responsible for presenting training 
programs on current regulations within the pharmaceutical manufacturing industry. 
The GXP trainers are responsible for at least the following: 
• Working with instructional designers to develop GXP training programs 
• Presenting initial GXP training during new employee orientation 
• Presenting ongoing GXP topics in: 
Good manufacturing practices 
Good documentation practices 
Good laboratory practices 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 

DEVELOPING A TRAINING PLAN 445 
Instructional Designers The instructional designers develop training programs 
as directed by the training coordinators. Instructional designers serve as a liaison 
between the training coordinators, training and development manager, departmental 
supervisors, and SMEs. Instructional designers are responsible for at least the 
following: 
• Identifying appropriate SMEs for training projects, in coordination with department 
training coordinators 
• Developing competency - based training programs according to established 
principles of instructional systems design 
• Developing evaluations and assessment measures 
• Developing train - the - trainer programs for department trainers 
1.4 Develop a Training Support Network To function properly, a training department 
needs to establish a network for training support within each functional 
department of the manufacturing facility. Within each functional department, three 
roles should be identifi ed to help support the training effort. Sometimes these roles 
can be performed by one person; sometimes the roles are performed by individuals 
within the department. Typically, the roles are not full - time positions. Figure 6 provides 
a list of departments and the training support roles necessary for each department; 
you may want to add to the list. Then, each support role is described. 
Training Coordinators Training coordinators are the liaison between functional 
departments and the training department. They are responsible for communicating 
with departmental training needs to the training and development manager. Training 
coordinators also serve as the liaison between departmental SMEs, departmental 
trainers, and instructional designers. They work to identify training needs and 
develop training programs that support the plans set out by the training and development 
manager with site management approval. They are also responsible for 
building support within the departments for executing training plans. Training 
FIGURE 6 Departments/training support roles. 
Department 
• Administration and finance 
• Buildings and grounds 
• Facilities and engineering 
• Human resources 
• Information systems 
• Manufacturing 
• Regulatory compliance 
• Research and development 
• Safety and security 
• Shipping and receiving 
Departmental Trainers Subject Matter Experts Training Coordinator(s) 

446 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
coordinators may serve more than one department within a facility. Training coordinators 
are responsible for at least the following: 
• Consulting with department management to identify training needs 
• Communicating departmental training needs and progress of training to the 
training and development manager 
• Developing departmental training plans and specifying training programs 
as directed by the training and development manager and department 
management 
• Directing instructional designers in developing training programs, evaluations, 
and assessments 
Subject Matter Experts Subject matter experts are individuals who possess a broad 
base of knowledge about the area where they work. These individuals are typically 
employees who have been with the company and department a signifi cant amount 
of time — typically fi ve years or more. Subject matter experts are responsible for at 
least the following: 
• Working with training coordinators to identify departmental training needs 
• Working with instructional designers to design and develop training 
programs 
• Working with departmental trainers to implement training programs 
Departmental Trainers The departmental trainers should be selected for their 
skill and interest in training and ability to perform their jobs. Some SMEs may 
become departmental trainers — but only with the stipulation that they want to learn 
how to become a trainer. Departmental trainers are responsible for at least the 
following: 
• Working with instructional designers to design and develop training 
programs 
• Training new hires and incumbents on job - specifi c skills and regulatory issues 
In order to meet the 21 CFR 211.25 requirements, departmental trainers must complete 
a train - the - trainer program. An outline of topics to be covered in a typical 
train - the - trainer program is shown in Figure 7 . 
Section 2 Training Programs 
This section includes a listing of all personnel to be trained as well as a listing of 
suggested training programs that need to be provided at the manufacturing facility. 
In addition, employee qualifi cation pathways are discussed and illustrated. 
2.1 Identify All Training Programs/Personnel to be Trained Figure 8 provides 
a sample list of personnel who need to be trained and the training programs they 
need to work effectively and effi ciently at a pharmaceutical manufacturing facility; 
you will want to add to the list. 

DEVELOPING A TRAINING PLAN 447 
FIGURE 7 Sample train - the - trainer program outline. 
1. Introductions 
2. Program objectives 
3. Trainer responsibilities 
4. Training process: trainee characteristics 
5. Training process: adult learning theory 
6. Training process: what makes successful training? 
7. What makes a successful trainer? 
8. Training techniques 
9. Plan 
10. Prepare 
11. Present 
12. Feedback and coaching 
13. Using an OJT checklist 
14. Exercise 1— OJT with SOP 
15. Exercise 2— OJT checklist 
Review and wrap up 
2.2 Establish Personal Qualifi cation Pathways Good business practice, as well 
as CFR 21 211.25, mandates that each employee possess the education, training, and 
experience to enable him or her to perform assigned functions in a safe and effective 
manner. The training plan must contain a policy allowing personnel to demonstrate 
their qualifi cations. This should cover training and supervisory personnel as 
well as those directly involved with operations. 
A qualifi cation pathway should include all of the training programs necessary to 
qualify employees in job functions. A sample personal qualifi cation pathway for a 
compression technician is shown in Figure 9 . 
2.3 Training Program Materials While there are basically three methods for 
delivering training available — classroom led, self - instruction, or a combination of 
the two — training materials should include: 
1. A method of imparting knowledge or desired attitudes. This type of training 
typically occurs in a classroom - led or self - instructional system (paper or computer 
based). 
2. A method of assessing if that knowledge or attitude transfer took place, that 
is some form of assessment. 

448 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
If the training requires employees to be competent in a particular task or skill, then 
these materials must also be developed: 
3. A structured checklist that allows different trainers to demonstrate the same 
way of doing a particular task to each trainee. 
4. An evaluation checklist that can be used by a trainer (it is suggested that the 
trainer who provided the initial training not do the competency evaluation) 
to determine if a trainee is competent in performing the assigned skill. 
Frequency of Training In general, providing refresher or retraining on a routine 
basis is not necessary. Some training programs are mandated to be repeated, or 
refreshed, usually by federal, state, or local government, that is, some safety training. 
FIGURE 8 Training programs/personnel to be trained. 
Management Training Programs 
Training Programs 
p i h s r e d a e L
Coaching and counseling 
Problem solving/decision making 
Coaching and counseling 
Specialty Training Programs 
Training Programs 
Bloodborne pathogens 
Conducting effective 
investigations 
Appropriate personnel 
Drug-labeling regulations 
Electronic signatures and batch 
records 
Appropriate personnel 
FDA inspections 
Hazcom 
SOP writing 
Surveillance monitoring 
Working with contract research 
organizations 
Appropriate personnel 
Validation concepts 
Supervisors and above 
Supervisors and above 
Supervisors and above 
Supervisors and above 
Personnel to be Trained 
Appropriate personnel 
Appropriate personnel 
Appropriate personnel 
Appropriate personnel 
Appropriate personnel 
Appropriate personnel 
Appropriate personnel 
Personnel to be Trained

DEVELOPING A TRAINING PLAN 449 
Other training programs may be mandated to be repeated by a company — such as 
sexual harassment. Retraining may be necessary when a person has been absent 
from a job for a period of time due to illness, pregnancy, other job assignment, and 
so on. In that case, the company may require retraining after a certain period of 
missed job function. In addition, retraining or training needs may be indicated when 
situations such as the following occur: 
• Out - of - specifi cation product (quality issues) 
• Excessive waste 
• Decrease in production 
Periodic audits of work performance may identify training needs as well. However, 
our experience has been that retraining is typically identifi ed through situations 
similar to those mentioned above. 
Section 3 Training Development 
Employees need job - specifi c training in environment, SOPs, safety, GMP regulations 
and awareness, and technical skills. The training plan should ensure a systematic 
approach to develop and implement competency - based training for all programs 
and materials. 
3.1 Establish a Training Program Design Model Effective competency - based 
training is the result of applying a systematic process to training program design. 
This process involves following certain well - defi ned steps to develop a training 
program that meets both the trainee ’ s and the organization ’ s needs. The most reliable 
method of training and qualifying people in the safe and effective performance 
of work is using an instructional system design model. One instructional systems 
FIGURE 9 Sample personal qualifi cation pathway. 
r o t a r e p o n o i s s e r p m o C : e l t i t b o J g n i r u t c a f u n a M : t n e m t r a p e D 
t r a t S g n i n i a r T s e i t u D b o J k s a T b o J 
Date 
Training Completion 
Date 
) R B M ( d r o c e r h c t a b r e t s a M n o i t a t n e m u c o d g n i t e l p m o C 
g n i n a e l c t n e m p i u q e r o n i M s e r u d e c o r p g n i n a e l c g n i m r o f r e P 
s t r a p e l b a h c a t e d f o g n i n a e l c r o j a M 
s r e t n e c k r o w f o g n i n a e l c r o j a M 
e t t e F c i s a b — t n e m p i u q e g n i t a r e p O 
n a i l l i K 
y o t r u o C 
y r o e h t n o i s s e r p m o C d e c n a v d a — t n e m p i u q e g n i t a r e p O
operating support equipment 
r e t s e t s s e n d r a h f o g n i n a e l c d n a e s U 
r e t e m o r c i m f o g n i n a e l c d n a e s U 
s r o t c e t e d l a t e m f o g n i n a e l c d n a e s U 
s r e k c e h c h g i e w f o e s U 
s t e l l a p o t s k c a s g n i r r e f s n a r T s l a i r e t a m g n i t r o p s n a r T 
s e t o t f o g n i l d n a h d n a e s U 
g n i t a o c r o f s k c a s / s e t o t g n i g a t S
Use and cleaning of friability tester

450 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
design model, the training program design model (TPDM) developed by Gallup 
and Griffi n, is shown in Figure 10 . 
3.2 Ensure Development of Valid Training Programs Validity may be defi ned 
as something that accomplishes that which it purports to accomplish. A training 
program is considered valid when it accurately imparts the competencies required 
to do a job. How can a training program be validated? The way to ensure program 
validation is to follow a training program design module and the following 
process: 
1. Conduct a task analysis that identifi es discrete bits of knowledge and specifi c 
job competencies; have SMEs sign off on completed task analyses indicating 
that all discrete bits of knowledge and specifi c tasks required to complete a 
job have been identifi ed. 
2. Write measurable performance objectives that state exactly what a trainee is 
to know or do after training has been completed; have SMEs sign off on com- 
FIGURE 10 Training program design model. 

DEVELOPING A TRAINING PLAN 451 
pleted performance objectives indicating that they agree that the level of 
stated performance is acceptable. 
3. Develop assessment instruments that match up with the performance objectives; 
have SMEs sign off on assessments indicating they agree that the instruments 
adequately measure trainee competency. 
Assess trainee ’ s knowledge and skills using the approved assessment instruments; 
ensure that assessments are conducted in the same manner for all employee 
groups. 
3.3 Develop a Change Control Policy and Procedure A change control policy 
helps ensure personnel are prepared to carry out new and revised policies and procedures 
effectively. An effective change control policy should include directions for 
regular reviews and revision of training materials and programs as well as scheduling 
training for new and incumbent employees on all pertinent SOPs. The plan must 
also include a procedure for responding to developments in operations, processes, 
and documentation. 
Section 4 Training Implementation 
4.1 Develop an Implementation Schedule The plan must include a schedule to 
ensure that training begins at initial employment and is ongoing. The schedule 
should include time frames for completing job qualifi cation training and address 
training required for changes in process and new or increased performance expectations. 
Developing a training implementation schedule can be accomplished by 
adding to the matrix developed in Section 2 , “ Identify All Training programs/Personnel 
to be Trained. ” This revised matrix is shown in Figure 11 . 
4.2 Establish a Training Failure Response Process The training plan should 
describe a process for dealing with personnel who do not pass qualifying evaluations. 
It should list steps for isolating the cause of the failure and differentiate between 
discrepancies in training program design and inappropriate candidates for the job. 
It should include publishing the protocol for dealing with test failure. A typical 
failure response for a task - based process is shown in Figure 8 . After a person completes 
a training program covering the knowledge to perform a specifi c task, an 
assessment is given. If that assessment is passed at a predetermined score, the person 
goes on to complete the skill - based portion of the training. Once the skill portion 
is completed, the person is observed as they perform the skill. If they perform it 
correctly, they are certifi ed or qualifi ed to perform the specifi c process operation. If 
they do not pass the skill demonstration check, they must repeat the training again. 
The number of times they are allowed to repeat the training and assessment should 
be part of the training failure response process. Figure 12 illustrates a basic failure 
response process. 
Section 5 Training Recordkeeping 
The FDA does not specify training documentation/recordkeeping requirements in 
the Code of Federal Regulations , but it has made them an industry requirement by 
virtue of precedent - setting industry demands. 

452 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
FIGURE 11 Sample implementation schedule. 
Specialty Training Programs 
Training Programs Personnel to be Trained Training Scheduled 
Bloodborne pathogens 
Conducting effective 
investigations 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Appropriate personnel As needed 
Drug-labeling regulations 
Electronic signatures and batch 
records 
FDA inspections 
m o c z a H
SOP writing 
Surveillance monitoring 
Working with contract research 
organizations 
Validation concepts 
FIGURE 12 Basic failure response process. 
Complete knowledgebased 
training 
Take written 
assessment 
Pass 
assessment 
? 
Complete 
skill-based 
training 
Demonstrate 
competency 
Compentency 
demonstrated 
? 
No 
Yes 
No 
Qualified 
or 
certified 
Yes 
Discussions with training managers at pharmaceutical manufacturing companies 
audited by the FDA suggest that it is through documentation/recordkeeping that 
the FDA “ backs into ” an audit of a company ’ s training. This typically happens in 
one of three ways: 
1. Auditors ask to see a SOP and observe an employee perform the procedure. 
Depending on the employee ’ s performance, the auditors may ask to review 
the documents or records of the observed employee ’ s training. 

DEVELOPING A TRAINING PLAN 453 
2. Through investigation of production records, auditors spot deviations or out - 
of - specifi cation issues. This in turn may lead to an examination of an operator ’ s 
work and qualifi cations for the job. The auditors may ask how or what additional 
training was conducted and request the records that prove training was 
provided. 
3. Auditors identify processes or work practices that appear to be performed 
incorrectly during a routine tour of a manufacturing facility. The auditors may 
ask to see the training materials or training records associated with individuals 
working in the area. 
Further, training managers note that if the training department can produce training 
documentation quickly and in an organized fashion, auditors are more likely to view 
the overall training effort as effective. At the same time, when solid training documentation 
cannot be produced quickly, a more in - depth review of the training 
system may occur [10] . As a result of these requirements to produce training records, 
it is critical that an effi cient training recordkeeping system be in place at the pharmaceutical 
manufacturing facility. 
5.1 Establishing a Recordkeeping System Any documentation system may 
meet the requirements of “ produce . quickly and present in an organized 
fashion. ” 
Three types of methods are commonly used for training documentation/recordkeeping 
in the pharmaceutical manufacturing industry: paper systems, electronic 
systems with paper backup, and stand - alone electronic systems. Some pharmaceutical 
companies use a paper - based documentation system. Electronic systems with 
paper backup, also known as learning management systems (LMSs), range from 
elaborate programs such as IsoTrain, SAP, Plateau, and Registrar to departmental 
databases created in Microsoft Access. Paper backup typically includes at least sign - 
in sheets and sometimes evaluation instruments. An LMS is essentially a database 
made up of multiple tables to store discrete units of information. These might 
include employee information, course information, evaluation data, and training 
programs or materials. These tables can be merged and queried to produce specifi c 
reports. Report capabilities allow users to compare, analyze, question, and project 
needs. 
5.2 Training Record Requirements The following basic information should be 
included in any training recordkeeping system: 
• Employee name 
• Employee identifi cation 
• Personal qualifi cation pathway, including training programs that need to be 
completed and targeted completion dates 
• Results of training programs completed 
In addition, the training program materials, including trainee guides and assessment 
instruments, should be readily accessible — either electronically or in a paper - based 
fi ling system. 

454 PERSONNEL TRAINING IN PHARMACEUTICAL MANUFACTURING 
REFERENCES 
1. U.S. Food and Drug Administration (FDA) ( 2003 ), Current good manufacturing practice 
for fi nished pharmaceuticals , Code of Federal Regulations , Title 21, Part 211.25, FDA, 
Rockuille, MD. 
2. Levchuck, J. W. (1991), Training for GMPs , J. Parenter. Sci. Technol. , 45 ( 6 ), 270 – 275 . 
3. U.S. Food and Drug Administration (FDA) ( 2004 ), Guidance for industry: Quality systems 
approach to pharmaceutical good manufacturing practice regulations , FDA, Rockuille, 
MD. 
4. Mager , R. ( 1962 ), Preparing Instructional Objectives , Center for Effective Performance , 
Atlanta, GA . 
5. Tetzlaff , R. ( 1982 ), A systematic approach to GMP training , Pharm. Technol. , 6 ( 11 ), 
42 – 51 . 
6. Langer , E. S. ( 2004 ), Training day for International Biopharma Manufacturers , Contract- 
Pharma , 6 ( 2 ), 46 . 
7. Morris , W. ( 2006 ), Personnel training: A growing compliance concern , PDA Newslett. , 
XLII ( 4 ), 1 – 38 . 
8. Gallup , D. A. , Beauchemin , K. A. , and Gillis , M. ( 1999 ), A comprehensive approach to 
compliance training in a pharmaceutical manufacturing facility , PDA J. S. Technol. , 53 ( 4 ), 
163 – 167 . 
9. Vesper , J. L. ( 2000 ), Defi ning your GMP training program with a training procedure , 
BioPharm. , 13 ( 7 ), 28 – 32 . 
10. Gallup , D. A. , Beauchemin , K. A. , Gillis , M. , Altopedi , D. and Manor , J. ( 2003 ), Selecting 
a training documentation/recordkeeping system in a pharmaceutical manufacturing environment 
, PDA J. Sci. Technol. , 57 ( 1 ), 49 – 55 . 

CONTAMINATION AND 
CONTAMINATION CONTROL 
SECTION 6


457 
6.1 
ORIGIN OF CONTAMINATION 
Denise Bohrer 
Universidade Federal de Santa Maria, Santa Maria, Brazil 
Contents 
6.1.1 Introduction 
6.1.2 Endogenous Contaminants 
6.1.2.1 Raw Materials 
6.1.2.2 Additives 
6.1.2.3 Decomposition of Formulation Constituents 
6.1.3 Exogenous Impurities 
6.1.3.1 Residual Solvents 
6.1.3.2 Containers 
6.1.3.3 Delivery Systems 
6.1.3.4 Particulate Matter 
6.1.4 Concluding Remarks 
References 
6.1.1 INTRODUCTION 
There are many ways to approach the issue of contamination. Although the origin 
of contaminants in pharmaceutical products can easily be identifi ed, defi nitions on 
this theme can be looked at from different angles. 
In a fi rst approach, the defi nition of contamination in pharmaceutical products 
can be referred to in terms of related substances and process contaminants: While 
related substances are structurally related to the drug substance, process contaminants 
are introduced during manufacturing or handling procedures. These two categories 
include all types of contaminants but do not defi ne them. 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

458 ORIGIN OF CONTAMINATION 
The U.S. Pharmacopeia and National Formulary (USP - NF 27) [1] presents defi nitions 
of terms related to contaminants in impurities in offi cial articles < 1086 > and 
ordinary impurities < 466 > . In these monographs, impurities are classifi ed into inorganic, 
organic, biochemical, isomeric, and polymeric, and defi nitions are given in 
terms of foreign substances, toxic impurities, concomitant components, signal impurities, 
and ordinary impurities. According to these defi nitions, foreign substances are 
not the consequence of the synthesis or preparation of the drug, yet they are introduced 
by contamination or adulteration. Concomitant components are characteristic 
of bulk chemicals and are not considered impurities in the pharmacopeial sense. 
Ordinary impurities are species in bulk chemicals that are innocuous and do not 
present signifi cant undesirable biological activity in the given amounts. 
Different from foreign impurities, which theoretically are not present and, therefore, 
are not included when monograph tests and assays are selected, toxic impurities 
and signal impurities may arise from the synthesis, preparation, or degradation of 
compendial articles and, differently from ordinary impurities , may present undesirable 
biological activity, even as minor components. In this context, the following 
topics should be taken into account: 
• Source of a drug substance: natural, synthetic, biotechnological 
• Ratio impurity/drug substance and its toxicological effects 
• Pharmacology of the impurity 
Other factors that should also be taken into account when limits for impurity 
levels in bulk substances are to be set are: 
• Route of administration 
• Dose 
• Target population 
• Duration of the therapy 
Not included in process contaminants , but equally an external source, are those 
impurities arising during packaging, during drug reconstitution or admixing, and 
also during the administration to the patient. Very important under this category is 
the particulate matter in intravenous formulations, mainly infusion solutions due to 
the large volume administered. 
Figure 1 presents an overview of impurity sources, both endogenous and exogenous, 
considering the steps at which these can occur. It also summarizes how “ the 
origin of contamination ” will be addressed in this chapter. It is divided into endogenous 
and exogenous contaminants. The fi rst part, endogenous contaminants, is 
devoted to water since it is the main raw material, if not in the drug form at least 
in any step of its preparation. Following water, the contamination arising with raw 
materials will be treated, considering the presence of concomitants (natural impurities 
of raw materials) and by - products, synthesized along with the drug. Also important 
while dealing with endogenous contaminants are those formed through 
inadvertent drug decomposition. Exogenous source of contaminants — aid materials 
used during drug or formulation preparation, containers, and delivery systems — will 
also be considered. 

ENDOGENOUS CONTAMINANTS 459 
FIGURE 1 Schematic representation of an overview on impurity sources. 
Step 
Manufacturing 
Storing 
Handling 
Procedure 
Raw material 
isolation 
Formulation 
preparation 
Sterilization 
Opening 
Admixing 
Administration 
Raw material 
synthesis 
Additives 
Concomitants 
Containers 
Containers Decomposition 
Decomposition 
Particulate 
Tubing lines 
Particulate 
Tubing lines 
By-products 
Aid materials 
Solvents 
Chemicals 
Aid materials 
Solvents 
Chemicals 
Endogenous 
impurites 
Exogenous 
impurites 
6.1.2 ENDOGENOUS CONTAMINANTS 
Endogenous contaminants can already be present as concomitants in raw substances, 
whether formed during drug synthesis as a by - product or arising from 
reactions among formulation constituents. The generation of contaminants in formulations 
after processing and packaging is generally promoted by external agents 
such as light, ultraviolet (UV) radiation, heat, or air, causing unexpected reactions 
between constituents. 
Pharmacopeial compendia establish a couple of tests that should be carried out 
with raw material batches to determine their level of impurities. Concomitants 
usually detected are heavy metals, chloride, or sulfated ash. The presence of by - 
products varies from raw material to raw material. Examples can be both innocuous 
substances such as monohydrogen phosphate, which is an impurity in dihydrogen 
phosphate salts and glycine, which is an impurity in alanine, and toxic substances, 
such as the classical enantiomer S of thalidomide. 
6.1.2.1 Raw Materials 
Raw materials employed in the pharmaceutical industry may have two different 
origins. They are either naturally occurring substances or synthesized drugs. Among 
the natural products are active ingredients from plant extracts or animals, chemicals, 

460 ORIGIN OF CONTAMINATION 
and biotechnological products. As vast as the variety of different raw materials is 
the array of impurities. This section will deal with the impurities already present in 
raw materials as concomitants or those arising during drug synthesis. Concomitants 
include the usual impurities considered by pharmacopeial compendia, such as heavy 
metals and arsenic, for which limits are prescribed in all monographs. By - products 
arising during drug synthesis may or may not be a problem when present as 
contaminants. 
Water Water is the primary raw material in pharmaceutical formulations. It is the 
most used vehicle since it is the major component of the human body. For many 
products, it is the main component, and, even in those containing non - water - soluble 
substances, water must be present. Depending on the product and the form of 
administration, lipophilic drugs are prepared as water – oil emulsions. 
The amount and level of contaminants or impurities in water for pharmaceutical 
purpose depend on its use. Since water is used in all industries and scientifi c work, 
international and national standard authorities have established water quality 
parameters for all types of applications. Health - related water standards are given 
by organizations such as the World Health Organization (WHO) [2] , the Environmental 
Protection Agency (EPA) [3] , and the American Society for Testing and 
Materials Standards (ASTM) [4] in the United States and by pharmacopeial compendia 
when the aim is specifi cally related to water for pharmaceutical products for 
human and veterinary consumption. 
Standards for water quality are similar among pharmacopeias (USP [1] , BP [5] , 
DAB [6] , Ph. Eur. [7] , IP [8] ), with only a few minor differences in the allowed levels 
for chemical contaminants. Pharmacopeias classify water in three categories: puri- 
fi ed water, highly purifi ed water , and water for injection. Whatever the process for 
water purifi cation, the raw material is always drinking water. Drinking water parameters, 
given by governmental regulatory agencies, include a large number of chemicals. 
They comprise not only naturally occurring substances but also a series of 
chemicals that may be present from anthropogenic activities in natural waters such 
as benzene, ethylene diamine tetraacetic acid (EDTA), and cadmium. Table 1 shows 
the guidelines for drinking water quality adopted by WHO [2] . 
The water purifi cation process adopted by pharmaceutical industries must be 
able to furnish water with the quality parameters presented in Table 2 . 
The ability to achieve a guideline value depends on a number of factors, 
including: 
• Concentration of the chemical in the raw water 
• Nature of the raw water 
• Treatment processes 
Although there are no levels for them, pesticides, for example, should not be present 
in water for pharmaceutical purposes. 
There are treatments available for the pharmaceutical industry for purifying 
water, which should be chosen according to the degree of purity necessary. They are 
listed in Table 3 according to their degree of complexity: the higher the ranking, the 
more complex the process. 

ENDOGENOUS CONTAMINANTS 461 
TABLE 1 WHO Guideline Values for Chemicals in 
Drinking Water and EPA National Drinking Water 
Standards [2, 3] 
Contaminant Guideline Value (mg/L) 
Primary Standards 
Acrylamide 0.0005 
Alachlor 0.002 
Antimony 0.006 
Arsenic 0.010 
Asbestos (fi bers > 10 . m) 7 million fi bers per liter 
Atrazine 0.003 
Barium 2 
Benzene 0.005 
Benzo( a )pyrene (PAHs) 0.0002 
Beryllium 0.004 
Boron 0.5 
Bromate 0.010 
Cadmium 0.005 
Carbofuran 0.04 
Carbon tetrachloride 0.005 
Chloramines (as Cl 2 ) 4.0 
Chlordane 0.002 
Chlorine (as Cl 2 ) 4.0 
Chlorine dioxide (as ClO 2 ) 0.8 
Chlorite 1.0 
Chlorobenzene 0.1 
Chromium (total) 0.1 
Copper 1.3 
Cyanide (as free cyanide) 0.2 
2,4 - DB 0.07 
Dalapon 0.2 
1,2 - Dibromo - 3chloropropane 
(DBCP) 
0.0002 
o - Dichlorobenzene 0.6 
p - Dichlorobenzene 0.075 
1,2 - Dichloroethane 0.005 
1,1 - Dichloroethylene 0.007 
Cis - 1,2 - Dichloroethylene 0.07 
Trans - 1,2 - Dichloroethylene 0.1 
Dichloromethane 0.005 
1,2 - Dichloropropane 0.005 
Di(2 - ethylhexyl) adipate 0.4 
Di(2 - ethylhexyl) phthalate 0.006 
Dinoseb 0.007 
Dioxin (2,3,7,8 - TCDD) 0.00000003 
Diquat 0.02 
EDTA 0.6 
Endothall 0.1 
Endrin 0.002 
Epichlorohydrin 0.0004 
Ethylbenzene 0.7 
Ethylene dibromide 0.00005 

462 ORIGIN OF CONTAMINATION 
Contaminant Guideline Value (mg/L) 
Fluoride 4.0 
Formaldehyde 0.9 
Glyphosate 0.7 
Haloacetic acids (HAA5) 0.060 
Heptachlor 0.004 
Heptachlor epoxide 0.0002 
Hexachlorobenzene 0.001 
Hexachlorocyclopentadiene 0.05 
Lead 0.015 
Lindane 0.0002 
Mercury (inorganic) 0.002 
Methoxychlor 0.04 
Nickel 0.02 
Nitrate (measured as nitrogen) 10 
Nitrite (measured as nitrogen) 1 
Oxamyl (Vydate) 0.2 
Pentachlorophenol 0.001 
Picloran 0.5 
Polychlorinated biphenyls (PCBs) 0.0005 
Selenium 0.05 
Simazine 0.004 
Styrene 0.1 
Tetrachloroethylene 0.005 
Thallium 0.002 
Toluene 1 
Total trihalomethanes (TTHMs) 1 
Toxaphene 0.003 
2,4,5 - TP (Silvex) 0.05 
1,2,4 - Trichlorobenzene 0.07 
1,1,1 - Trichloroethane 0.2 
1,1,2 - Trichloroethane 0.005 
Trichloroethylene 0.005 
Uranium 0.030 
Vinyl chloride 0.002 
Xylenes (total) 10 
Secondary Standards 
Aluminum 0.05 to 0.2 
Chloride 250 
Copper 1.0 
Fluoride 2.0 
Foaming agents 0.5 
Iron 0.3 
Manganese 0.05 
Silver 0.10 
Sulfate 250 
Total dissolved solids 500 
Zinc 5 
TABLE 1 Continued

ENDOGENOUS CONTAMINANTS 463 
Table 4 presents the quality parameters of water obtained by the purifi cation 
processes listed in Table 3 . 
The effectiveness of each treatment in removing the contaminants listed in Tables 
1 and 2 are given in Table 5 [2] . 
Even when water complies with quality parameters as a raw material, it can 
present some impurities after being turned into a pharmaceutical product. Table 6 
presents the level of some contaminants found in water for injection (WFI). Since 
the raw material should have passed in the quality test, contaminants either were 
below the allowed concentration level or were introduced after packaging. Contaminants 
introduced after packaging most likely originate from the packaging materials. 
Section 6.1.3.2 discusses containers as sources of contamination. 
In summary, water can be a source of contaminants. If the raw material (drinking 
water) complies with the quality parameters established by authorities, contaminants 
still present can be eliminated by usual water purifi cation processes available 
to the pharmaceutical industry. While distillation and reverse osmosis provide water 
with the quality specifi cations for purifi ed water and highly purifi ed water, WFI is 
generally obtained by membrane fi ltration (associated with another purifi cation 
process) not only because of chemical contamination but mainly because of sterility 
requirements. 
TABLE 2 Water Quality for Pharmaceutical Purposes (Pharmacopeial Standards) 
Parameter 
Water Grade 
Purifi ed Water Highly Purifi ed Water Water for Injection 
Conductivity (at 20 ° C) 
. S cm . 1 
4.3 1.1 1.1 
Total organic carbon 
(TOC) (mg/L) 
0.5 0.5 0.5 
Nitrates (ppm) 0.2 0.2 0.2 
Aluminum ( . g/L) 10 10 10 
Heavy metals (ppm) a 0.1 0.1 0.1 
Chloride Pass/fail — 0.5 ppm 
Sulfate Pass/fail — Pass/fail 
Ammonium (ppm) 0.2 — 0.2 
Calcium/magnesium Pass/fail — Pass/fail 
Residue evaporation 
mg/100 mL 
1 — 0.4 (volume . 10 mL) 
0.3 (volume > 10 mL) 
a Measured as lead. 
— = not informed. 
TABLE 3 Ranking of Complexity of Water Treatment 
Processes for Chemicals 
Ranking Process 
1 Distillation 
2 Ion exchange 
3 Reverse osmosis 
4 Membrane fi ltration 

TABLE 4 Quality Parameters of Water after Different Purifi cation Treatments 
Purifi cation Process 
Conductivity 
(at 20 ° C) 
. S cm . 1 
Total 
Organic 
Carbon 
(mg/L) 
Nitrates 
(ppm) 
Aluminum 
(. g/L) 
Heavy 
Metals 
(ppm)a 
Ammonium 
(ppm) 
Calcium/ 
Magnesium 
(mg/L) 
Residue Evap. 
mg/100 mL 
(as silicate) 
None (tap water example) 
240 10 < 10 
< 200 1 1 35 1 
Distillation (single) 10.2 0.03 — — 0.5 – 1 0.01 1 – 3 0.5 – 1 
Distillation (double) 2.1 0.06 — — 0.1 – 0.8 0.01 0.3 – 0.1 0.1 – 0.7 
Ion exchange 2 – 30 — — — < 0.01 — 
— 
1 
Reverse osmosis 10 – 25 0.03 — — < 0.04 0.4 1.6 0.1 
Membran fi 
ltration 0.056 0.01 — — < 0.01 
< 0.01 — 
< 0.01 
a Measured as lead — not informed. 
464

ENDOGENOUS CONTAMINANTS 465 
Concomitants Concomitants can be considered impurities present in naturally 
occurring, nonsynthesized raw materials. They may either present toxic effects, as 
with arsenic, or be as harmless as chloride ions. An overview of usual concomitants 
and their limits cited in pharmacopeial compendia are listed in Table 7 . 
The presence of concomitants in pharmaceutical products, although inevitable, 
may not exactly be a problem. The presence of magnesium in calcium salts is very 
TABLE 5 Effectiveness of Each Treatment in Removing the Contaminants 
Contaminants 
Treatment 
Distillation (%) Ion Exchange 
Reverse Osmosis 
(%) 
Membran 
Filtration (%) 
Ions > 70 > 80% Monovalent > 95 
Polyvalent > 97 
> 80 
Organic compounds > 80 Not effi cient > 99 > 80 
Particles > 80 Not effi cient > 99 > 99 
Source : From refs. 2 and 9 . 
TABLE 6 Contaminants Found in Water for Injection 
Parameter Sample Content ( . g/L) Reference 
Aluminum Sterile water, Abbott < 5 10 
Sterile water, McGaw < 5 10 
Sterile water, Travenol < 5 10 
Aqua ad injectabilia, Braun, 50 mL 1 11 
Arsenic Water for injection, EMS 39.3 12 
Water for injection, Geyer 30.9 12 
Zinc Sterile water 13.9 13 
Silicate Sterile water, Geyer 280 ± 13 14 
TABLE 7 Main Inorganic Impurities Listed in Pharmacopeial Compendia and Their 
Limits 
Impurity Limit Impurity Limit 
Sulfated ash 0.01 – 1% Heavy metals 1 – 50 ppm 
Chloride 10 – 500 ppm Arsenic 1 – 4 ppm 
Sulfate 50 – 400 ppm Aluminum 0.2 – 1 ppm 
0.1 – 0.6% Cadmium 5 – 10 ppm 
Fluoride 3 ppm Chromium 0.05–10 ppm 
Bromide 50 ppm Copper 0.1 ppm 
Oxalate 100 – 350 ppm Iron 2 – 100 ppm 
Phosphate 25 – 400 ppm Lead 0.1 – 50 ppm 
Sulfi te 15 ppm Nickel 0.2 – 1 ppm 
Ammonium 200 ppm Silver 250 ppm 
Zinc 10 – 30 ppm 
Note : Observed: Some impurities are related to certain products only. For example, the limit for silver 
is in cysplatin and sulfi te in sugars. 

466 ORIGIN OF CONTAMINATION 
common due to the similarity between these cations, both alkaline earth metals with 
a very similar chemical behavior. The consequence is that most raw materials containing 
calcium also contain small amounts of magnesium. The same occurs between 
sodium and potassium, so that the limit for potassium as an impurity in sodium 
chloride (according to the BP monograph) is 550 ppm or 0.55 mg potassium per 
gram sodium salt. 
If species such as sodium, potassium, chloride, and sulfate are tolerable in a fairly 
large range, other toxic species such as arsenic and lead have narrower limits. For 
arsenic and lead, the limits in pharmacopeial compendia are generally set between 
1 and 10 ppm; in fact, these limits are high for such harmful species. 
Studies showing the presence of impurities in raw materials are not very frequent, 
mainly because the quality of raw materials is certifi ed by guarantee bulletins, and 
products are supposed to be within the quality attested by the certifi cate. 
Two studies with raw materials used to prepare parenteral formulations were 
carried out to show their content of aluminum and arsenic [15, 16] . It is possible to 
see in Figures 2 and 3 that aluminum and arsenic were present in all investigated 
raw materials. There were also different levels of contamination among the substances. 
While salts such as NaCl and KCl presented low aluminum contaminations, 
phosphates, gluconate, and also citric acid were relatively contaminated. The authors 
attributed this difference to the affi nity of aluminum to the latter substances. Arsenic 
showed a more uniform distribution of contamination. With the exception of the 
amino acid tyrosine, the arsenic level in all substances was below 1 . g/g, not exceeding 
the limits prescribed by pharmacopeias. 
Since analyses to check the presence of impurities in pharmaceutical 
products are generally carried out with fi nished products, it is not possible to 
attribute the presence of contaminants to the raw materials. Unless the raw 
material itself is checked, many other sources can aggregate impurities to the fi nal 
product. 
FIGURE 2 Aluminium present as impurity in substances used in parenteral nutrition 
[15] . 

ENDOGENOUS CONTAMINANTS 467 
By - products By - products are perhaps the most diffi cult impurities to summarize 
since each drug has its own by - products that may appear as impurities. They are 
synthesized along with active ingredients and are generally diffi cult to separate, 
owing to their similarity to the drug of interest. Most are isomeric species, differing 
from each other by the presence of only a small group or just by the position of a 
hydrogen atom. Even more diffi cult to separate and, therefore, purify are chiral 
isomers. Modern drugs that contain only one chiral isomer as the active ingredient 
are yet more diffi cult. Pharmacopeias usually present in the monographs the by - 
products that can exist as impurity of an active ingredient. 
6.1.2.2 Additives 
Additives are all formulation constituents other than the active ingredient. Although 
additives could be classifi ed into excipients and vehicles (excipients for solid preparations 
and vehicles for liquid ones), there are several other agents used in pharmaceutical 
formulations with specifi c functions such as preservatives, sweeteners, 
coatings, colorants, antioxidants, surfactants, emulsifying agents, and fl avors. Since 
they comprise a vast amount of products, this section will deal with additives for 
compounding pharmaceutical products for internal use only [17, 18] . 
Table 8 lists additives and the possible impurities they might present. The limits 
for these impurities were taken from manufacturers ’ specifi cation bulletins. To compound 
Table 8 , products of pharmacopeial grade (USP, BP, Ph. Eur., DAB) were 
selected from qualifi ed suppliers (Merck, Aldrich, Sigma, Fluka, EMS, Riedel 
deHa e n) . 
6.1.2.3 Decomposition of Formulation Constituents 
The most relevant endogenous contamination arising from the decomposition of 
formulation constituents is the generation of peroxides in vitamins, amino acids, and 
lipid emulsions by action of light and air, that is oxygen. 
FIGURE 3 Arsenic level in substances used in formulations for parenteral nutrition [16] . 
As (.g/g) 
Tyr 
Leu 
Na2HCO3 
NaH2PO4 
KH2PO4 K2HPO4 
MgCl2 
CaCl2
MgSO4 
Na2HPO4 
NaCl 
gluconate 
KCl 
NaAc 
Om Cystine 
Cys 
Arg Val Ser His lle 
Thr 
Gly 
Met 
Phe
Asn Asp 
Pro 
Glu 
Ala Lys N-acetyl-Tyr 
sorbitol 
ascorbic ac. 
Vit. B2 
Vit. B5 
folic ac. 
glucose 
mannitol heparin 
malic ac. xylitol 
2.0 
1.6 
1.2 
0.8 
0.4
0

468 ORIGIN OF CONTAMINATION 
TABLE 8 Additives for Pharmaceutical Formulations and Impurities and Their Limits 
According to Manufacturers (Merck, Aldrich, Sigma, Fluka, EMS, Riedel deHa e n) 
Additive Impurity Limit 
Solvents 
Ethanol Acetone and isopropyl alcohol . 0.01% 
Acetaldehyde and acetal . 0.001% 
Benzene . 0.0002% 
Methanol . 0.01% 
Total of other impurities . 0.03% 
Glycerol Chloride . 0.0010% 
Sulfate . 0.0010% 
Halogen compounds (as Cl) . 0.0030% 
Heavy metals . 0.0005% 
Arsenic . 0.0001% 
Calcium . 0.0001% 
Cadmium . 0.0001% 
Mercury . 0.0001% 
Ammonium . 0.0005% 
Lead . 0.0002% 
1,2,4 - Butantriol . 0.2% 
Residual solvents, class 3 a . 0.5% 
Aldehydes . 10 ppm 
Sulfated ash . 0.01% 
Polyethylene glycol Dioxane 10 ppm max 
Ethylene glycol and diethylene 
glycol 
0.4% max 
Ethylene glycol (class 2) 620 ppm max 
Ethylene oxide 1 ppm max 
Formaldehyde (HCHO) 30 ppm max 
Heavy metals 20 ppm max 
Sulfated ash 0.2% max 
Preservatives 
Benzyl alcohol Peroxide value a . 5 
Benzaldehye . 0.15% 
Cyclohexylmethanol . 0.10% 
Benzene . 0.0002% 
Chlorobenzene . 0.01% 
Toluene . 0.01% 
Benzoic acid Sulfate . 0.02% 
Heavy metals . 0.001% 
Arsenic . 0.0001% 
Cadmium . 0.0010% 
Copper . 0.0010% 
Mercury . 0.0001% 
Lead . 0.0005% 
Zinc . 0.0010% 
Halogen compounds (as Cl) . 0.01% 
Toluene . 890 ppm 

ENDOGENOUS CONTAMINANTS 469 
Additive Impurity Limit 
Chlorobutanol Chloride . 0.01% 
Chloroform . 60 ppm 
Residual solvents, class 3 a . 0.5% 
Sulfated ash . 0.1% 
Methylparaben Heavy metals . 0.001% 
Arsenic . 0.0003% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0005% 
Zinc . 0.001% 
Methanol . 0.3% 
Sulfated ash . 0.05% 
Methylparaben sodium salt Chloride . 0.03% 
Sulfate . 0.03% 
Heavy metals . 0.001% 
Arsenic . 0.0003% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0005% 
Zinc . 0.001% 
Propylparaben sodium salt Chloride . 0.03% 
Sulfate . 0.03% 
Heavy metals . 0.001% 
Arsenic . 0.0003% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0005% 
Zinc . 0.001% 
1 - Propyl alcohol . 0.5% 
Sulfated ash 34 – 36% 
Propylparaben Heavy metals . 0.001% 
Arsenic . 0.0003% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0005% 
Zinc . 0.001% 
Residual solvents, class 3 a . 0.5% 
Sulfated ash . 0.05% 
Potassium sorbate Arsenic . 0.0003% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0002% 
Zinc . 0.001% 
Aldehydes (as acetaldehyde) . 0.15% 
Residual solvents, class 3 a < 0.5% 
TABLE 8 Continued

470 ORIGIN OF CONTAMINATION 
Additive Impurity Limit 
Sodium benzoate Chloride . 0.02% 
Sulfate . 0.01% 
Total chlorine . 0.03% 
Heavy metals . 0.001% 
Arsenic . 0.0001% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0002% 
Zinc . 0.001% 
Sorbic acid Heavy metals . 0.0010% 
Arsenic . 0.0003% 
Cadmium . 0.001% 
Copper . 0.001% 
Mercury . 0.0001% 
Lead . 0.0002% 
Zinc . 0.001% 
Aldehydes (as acetaldehyde) . 0.15% 
Residual solvents, class 3 a . 0.5% 
Sulfated ash . 0.2% 
Antioxidants 
Ascorbic acid Heavy metals . 0.001% 
Ign. residue . 0.05% (as 
SO4 ) 
Chloride . 50 mg/kg 
Sulfate . 20 mg/kg 
Copper . 5 mg/kg 
Iron . 2 mg/kg 
l (+) - Ascorbyl palmitate Heavy metals . 0.001% 
Arsenic . 0.0003% 
Copper . 0.0025% 
Lead . 0.001% 
Zinc . 0.0025% 
Sulfated ash . 0.1% 
Residual solvents, class 3 a . 0.5% 
Butyl hydroxyanisole Arsenic 0.0003% max 
Heavy metals 0.001% max 
Lead 0.0005% max 
Mercury 0.0001% max 
3 - tert - Butyl - 4 - methoxyphenol 10% max 
Hydroquinone 0.2% max 
Sulfated ash 0.01% max 
Butyl hydroxytoluene Arsenic 0.0003% max 
Heavy metals 0.001% max 
Lead 0.0005% max 
Mercury 0.0001% max 
Residual solvents, class 2 (MeOH) a 0.2% max 
Sulfated ash 0.002% max 
TABLE 8 Continued

ENDOGENOUS CONTAMINANTS 471 
Additive Impurity Limit 
Formaldehyde sulfoxylate 
sodium 
Iron 0.0025% max 
Sodium sulfi te 5.0% max 
Residual solvents, class 2 (MeOH) a 0.3% max 
Phosphoric acid Volatile acids (as CH 3 COOH) . 0.001% 
Chloride . 0.0005% 
Fluoride . 0.0010% 
Nitrate . 0.0003% 
Phosphite and hypophosphite (as 
H3 PO 3 ) 
. 0.02% 
Sulfate . 0.005% 
Heavy metals . 0.001% 
Arsenic . 0.0002% 
Cadmium . 0.00010% 
Copper . 0.002% 
Iron . 0.005% 
Mercury . 0.0001% 
Potassium . 0.005% 
Sodium . 0.03% 
Lead . 0.0010% 
Zinc . 0.002% 
Sodium bisulfi te Arsenic 0.001% max 
Heavy metals 0.003% max 
Iron 0.005% max 
Sodium metabisulfi te Chloride . 0.01% 
Heavy metals . 0.001% 
Thiosulfate . 0.02% 
Arsenic . 0.0002% 
Iron . 0.001% 
Mercury . 0.0001% 
Lead . 0.0005% 
Selenium . 0.0006% 
Sodium thiosulfate Sulfate and sulfi te (as SO 4 ) . 0.2% 
Heavy metals . 0.001% 
Tocopherol Heavy metals . 0.001% 
Arsenic . 0.0003% 
Copper . 0.0025% 
Mercury . 0.0001% 
Lead . 0.0005% 
Zinc . 0.0025% 
Methanol . 3000 ppm 
Pyridine . 200 ppm 
Toluene . 890 ppm 
Sulfated ash . 0.1% 
Vehicles 
Cellulose powder Ether - soluble substances . 0.15% 
Water - soluble substances . 1.0% 
Heavy metals . 0.001% 
Sulfated ash . 0.3% 
TABLE 8 Continued

472 ORIGIN OF CONTAMINATION 
Additive Impurity Limit 
Gelatine Sulfur dioxide (SO 2 ) . 0.004% 
Heavy metals . 0.001% 
Arsenic . 0.00008% 
Chromium . 0.001% 
Iron . 0.003% 
Zinc . 0.003% 
Peroxide (as H 2 O 2 ) . 0.001% 
Ash . 2.0% 
Lactose Heavy metals . 0.0005% 
Arsenic . 0.0001% 
Copper . 0.0025% 
Lead . 0.00005% 
Zinc . 0.0025% 
Sulfated ash . 0.1% 
Starch Reducing matter (as maltose) max 0.7% 
Sulfated ash max 0.4% 
Sorbitol Chloride . 0.002% 
Sulfate . 0.006% 
Heavy metals . 0.0005% 
Arsenic . 0.00013% 
Nickel . 0.0001% 
Lead . 0.00005% 
Related substances (mannitol) . 2.0% 
Reducing sugars after hydrolysis/ 
total sugar (as glucose) 
. 0.5% 
Reducing sugars (as glucose) . 0.11% 
Sulfated ash . 0.02% 
Sucrose Chloride . 0.0035% 
Sulfate . 0.005% 
Sulfi te (as SO 2 ) . 0.0010% 
Heavy metals . 0.0005% 
Arsenic . 0.0001% 
Lead . 0.00005% 
Sulfated ash . 0.02% 
Residual solvents, class 3 a . 0.5% 
Talc Heavy metals . 0.004% 
Aluminum . 2.0% 
Arsenic . 0.0003% 
Calcium . 0.9% 
Iron . 0.25% 
Lead . 0.0005% 
Asbestos (according to Ph. Eur.) Not 
detectable 
Zinc oxide Chloride . 0.005% 
Sulfate . 0.02% 
Arsenic . 0.0005% 
Cadmium . 0.001% 
Iron . 0.001% 
Lead . 0.005% 
TABLE 8 Continued

ENDOGENOUS CONTAMINANTS 473 
Additive Impurity Limit 
Chelating Agent s 
EDTA Heavy metals . 0.001% 
Calcium . 0.001% 
Iron . 0.001% 
Magnesium . 5 ppm 
Nitrilotriacetic acid . 0.1% 
Sulfated ash . 0.1% 
Buffers 
Boric Acid Sulfate . 0.04% 
Heavy metals . 0.0015% 
Disodium hydrogen 
phosphate 
Chloride . 0.02% 
Fluoride . 0.001% 
Sulfate . 0.05% 
Heavy metals . 0.001% 
Arsenic . 0.0002% 
Cadmium . 0.0001% 
Iron . 0.002% 
Mercury . 0.0001% 
Lead . 0.0004% 
Sodium dihydrogen 
phosphate 
Chloride . 0.005% 
Fluoride . 0.001% 
Hydrogen phosphate (HPO 4 ) . 0.5% 
Sulfate . 0.01% 
Heavy metals . 0.0005% 
Arsenic . 0.0002% 
Cadmium . 0.0001% 
Iron . 0.001% 
Lead . 0.0004% 
Mercury . 0.0001% 
a Residual solvents see Section 9.1.3.1; peroxide value: mmol peroxide/L. 
TABLE 8 Continued 
Lipid emulsions are essential components of parenteral nutrition. However, due 
to the amount of polyunsaturated fatty acids (PUFA), it is possible that chemical 
degradation occurs, forming hydroperoxides. 
The exposure of lipid emulsions to room light and spotlight simulating phototherapy 
conditions in a neonatal intensive care unit for 24 h increases the level of 
hydroperoxides in these formulations by up to 60 times [19] . Lipid admixtures stored 
in ethyl vinyl acetate (EVA) bags presented a signifi cant formation of lipid hydroperoxide. 
After one month under stressful conditions (gas - permeable container, 
40 ° C), the peroxide value (PV = millimoles of peroxide per liter) was 450 times 
higher compared with controls (nitrogen - overlaid glass bottles) [20] . The results 
presented in Table 10 show that in contrast to lipids stored in closed glass containers 
gassed with nitrogen, signifi cant and time - dependent peroxide formation occurred 
in all - in - one (AIO) plastic bags. Moreover, the lower PV of the lipid stored in V90 
bags (polypropylene – polyamide 7 : 3, double layered) compared with EVA bags 

474 ORIGIN OF CONTAMINATION 
(single layered) is indicative of the decreased oxygen permeability of the double - 
layered material and therefore the stabilization of the PV, although the increase of 
the PV occurred slowly in Intralipid samples stored in V90 bags. 
Differences were found in the PV of Intralipid (LCT = long - chain triglyceride) 
and Lipofundin (MCT = medium - chain triglyceride) samples. Initial PV in Intralipid 
(0.02 mmol/L) was lower than in Lipofundin (0.10 mmol/L), whereas in the latter 
the formation rate of PV was slower over time. This behavior is due to the double 
PUFA content of Intralipid. 
Polyunsaturated fatty acids are oxidized by enzymatic and nonenzymatic pathways. 
Nonenzymatic oxidation is a free - radical mediated peroxidation. It is a chain 
reaction providing a continuous supply of free radicals that initiate further peroxidation. 
The whole process can be depicted as follows [21] : 
RH X R XH + > + • • 
R O ROO 2 
• • + > 
ROO RH ROOH R • • + > + 
The reaction is initiated by any existing free radical (X . ), by light, or by metal ions 
that react with lipids (RH) generating peroxy radicals (ROO . ) and fi nally lipid 
hydroperoxides (ROOH). Besides hydroperoxides, malondialdehyde (MDA), 
ethane, and pentane can also be formed if PUFA with three or more double bonds 
and if . - 3 and . - 6 PUFA are present, respectively [14] . The formation of pentane 
(2 . mol/L) and MDA (10 . mol/L), along with hydroperoxides was observed in a 
lipid emulsion (Intralipid 10%) [22] . Malondialdehyde was also found in parenteral 
nutrition admixtures for newborn infants, in 12 samples analyzed, MDA levels 
varied from 1632 to 14,679 nmol/L [23] . Pironi and colleagues [24] also found MDA 
in fat emulsions. They compared three 20% lipid emulsions containing different 
amounts of PUFA and . - tocopherol, 24 hours after admixing them to make AIO 
solutions. They concluded that lipid peroxide generation is directly related to PUFA 
content and inversely related to the . - tocopherol/PUFA ratio of the emulsion. 
Moreover, MDA was also formed in these samples, presenting an increased concentration 
in the soybean - oil - containing emulsions when compared with the olive oil 
emulsion. Table 10 shows the PV and MDA measured in these samples. 
Peroxide formation has also been attributed to a prooxidant action of vitamin E 
on PUFA, as vitamin E is another component present in lipid emulsions [19, 25] . 
Lipid oxidation occurring despite high concentrations of vitamin E (tocopherol) 
may appear surprising since the latter is generally regarded as the most effi cient 
chain - breaking, lipid - soluble antioxidant [26] . Studies have revealed, however, that 
for vitamin E to be an effi cient antioxidant in an isolated lipoprotein emulsion, it 
requires a suitable “ co - antioxidant, ” such as vitamin C [27, 28] . The co - antioxidant 
eliminates the vitamin E radical formed in the interaction of the vitamin with 
initiating radical oxidants, which can have an adverse effect by promoting lipid 
peroxidation. 
Through the measurement of triglyceride hydroperoxides in Intralipid samples 
stored in glass and plastic syringes exposed to direct light or after being wrapped 
with aluminum foil, it was possible to demonstrate that Intralipid is highly oxidizable 

ENDOGENOUS CONTAMINANTS 475 
TABLE 9 Peroxidation of Intralipid 20% and Lipofundin MCT 20% during Storage: 
Infl uence of Container Material, Light Exposure, and Temperature 
Exposure, 
Condition, Time 
(day) 
Nutrimix 2/3 
Bag (EVA) 20 – 
27 ° C Daylight a 
Nutrimix 2/3 Bag 
(EVA) 20 – 27 ° C 
Light Protetion 
3 - chamber 
Bag (V 90) 
20 – 27 ° C 
Daylight a 
Control sample 
Closed Glass 
Bottle 20 – 27 ° C 
Daylight a 
Intralipid 20% 
1 0.02 ± 0.006 c 
5 0.47 ± 0.016 b 0.06 ± 0.004 b,c 0.33 ± 0.011 
8 0.02 ± 0.005 
9 0.84 ± 0.010 0.05 ± 0.007 0.50 ± 0.017 
14 1.46 ± 0.020 0.17 ± 0.002 0.60 ± 0.014 
15 0.02 ± 0.001 
19 1.87 ± 0.027 0.27 ± 0.017 0.76 ± 0.018 
22 2.48 ± 0.038 0.40 ± 0.013 0.92 ± 0.025 
28 0.02 ± 0.005 
29 2.93 ± 0.052 0.52 ± 0.023 1.24 ± 0.032 
Lipofundin MCT 20% 
1 0.11 ± 0.006 c 
5 0.57 ± 0.018 b 0.19 ± 0.017 d,e 0.64 ± 0.014 
8 0.10 ± 0.018 
9 0.88 ± 0.024 0.35 ± 0.002 0.69 ± 0.010 
14 1.31 ± 0.048 0.54 ± 0.025 0.64 ± 0.021 
15 0.08 ± 0.003 
19 1.59 ± 0.064 0.61 ± 0.018 0.67 ± 0.036 
22 1.99 ± 0.076 0.84 ± 0.007 0.70 ± 0.054 
28 0.10 ± 0.032 
29 2.48 ± 0.040 0.99 ± 0.044 0.69 ± 0.032 
Source : Form ref. 20 . 
a,b P < 0.001 (Nutrimix 2/3 with light protection). 
c,d P < 0.002 (Nutrimix 2/3 with light protection versus control samples). 
TABLE 10 Lipid Peroxide (LPX) and MDA Concentration in Lipid Emulsions 
Emulsions 
LPX ( . M/L) MDA ( . M/L) 
Bottles 
( n = 6) 
No fat 
PN 
( n = 6) 
AIO–T0 
( n = 6) 
AIO – T24 
( n = 6) 
Bottles 
( n = 6) 
No fat 
PN 
( n = 6) 
AIO – T0 
( n = 6) 
AIO–T24 
( n = 6) 
Soybean oil 28.8 ± 29.9 < 2 < 2 11.3 ± 15.0 8.8 ± 3.1 4.9 ± 0.8 5.0 ± 0.3 6.9 ± 1.0 
Soybean oil 
+ MCT 
1.7 ± 0.5 < 2 < 2 5.2 ± 8.8 3.7 ± 0.4 3.4 ± 0.4 4.1 ± 0.3 7.8 ± 1.8 
Olive oil 4.1 ± 2.8 < 2 < 2 5.5 ± 6.4 0.7 ± 0.1 5.4 ± 0.3 5.3 ± 0.3 3.3 ± 0.4 
Source : From ref. 24 . 
Note : Values are mean ± standard deviation of six bags (one sample per bag, analyzed in triplicate). 
AIO, all - in - one mixture immediately after the addition of lipid emulsion; AIO – T24, all in one mixture 24 h after the 
addition of lipid emulsion. 

476 ORIGIN OF CONTAMINATION 
under routine clinical conditions, in spite of its vitamin E content ( . 60 . mol/L 
. - tocopherol and 20 . mol/L each of the . and . isomers). 
The problem is particularly evident for neonates undergoing phototherapy. 
Besides the fact that lipids are used extensively to supply premature babies with 
calories, bags and extension sets are exposed for longer periods of time to the 
intense radiation of spotlights due to the low infusion rates of neonatal solutions. 
Figure 4 depicts the evolution of triglyceride hydroperoxide formation in a 20% 
Intralipid emulsion under different conditions of light exposure. It shows that lipid 
oxidation occurs despite the high concentration of vitamin E present in the samples. 
Decomposition can be prevented by wrapping containers and tubing sets in aluminum 
foil or by adding ascorbate to the infusate. 
Lipid peroxides are also able to react with other components of parenteral nutrition 
admixtures (trace elements), causing a drop in pH with the subsequent potential 
for physical – chemical instability [29] . Table 11 shows the peroxide value and the 
pH drop in a pure lipid emulsion and a lipid - containing AlO admixture stored in 
EVA bags under different conditions of temperature and light exposure in the presence 
and absence of trace elements. 
Peroxide formation has also been observed in multivitamin solutions for parenteral 
nutrition. Lavoie and co - workers [30] have studied the action of light, air, and 
composition on the stability of multivitamin formulations, and also total parenteral 
nutrition (TPN) admixtures containing and not containing vitamins and fatty acids. 
They analyzed the generation of peroxide in multivitamin solutions and in TPN for 
adults and neonates. The analysis of multivitamin solutions for enteral use revealed 
the presence of peroxides at the initial opening of the bottle. The levels were higher 
in Poly - Vi - Sol (vitamin A, Vitamin D, and vitamin C, vitamin B 1 , ribofl avin, and 
FIGURE 4 Kinetics of accumulation of triglyceride hydroperoxide in Intralipid exposed to 
phototherapy light under conditions that mimic the situation in an NICU (neonatal intensive 
care unit). A solution of 20% Intralipid was dispensed into a glass reservoir containing plastic 
tubing (line 3), with aluminum foil (line 4), or a plastic reservoir after supplementation with 
1 mmol/L sodium ascorbate (line 5) and exposed to phototherapy light for up to 24 h. For all 
samples the light fl uxes measured were 11.7 – 25.9 . W/cm 2 per nanometer for 24 h. At the time 
points indicated, aliquots were withdrawn, extracted, and their hexane extracts analyzed for 
the presence of triglyceride hydroperoxides and . - tocopherol as described in the methods 
section [19] . 

ENDOGENOUS CONTAMINANTS 477 
niacinamide) than in Tri - Vi - Sol (vitamin A, vitamin D, and vitamin C) (Fig. 5 ) 
[30] . 
The authors attributed the difference in the peroxide content of the preparations 
to ribofl avin. This vitamin catalyses the photoinduced reaction between ascorbate 
and oxygen, which leads to hydrogen peroxide generation. The drop in hydrogen 
peroxide concentration over time observed with Poly - Vi - Sol is explained by the 
transformation of hydrogen peroxide into a more reactive species, which in turn 
could react with other components of the formulation. In contrast, the increase in 
catalase - resistant peroxides in the Tri - Vi - Sol preparation suggests a greater contribution 
of air in the peroxide generation. In comparison to the formulation for parenteral 
use, the enteral multivitamin presented levels of 100 times higher, where 
catalase - resistant peroxides represented the bulk of peroxides. To study the effect 
of air and light on the generation of peroxides in multivitamins for TPN, the authors 
TABLE 11 Peroxide Value and pH Drop in Pure Lipid Emulsion and AIO Admixture 
Stored in EVA Bags under Different Conditions of Temperature and Light Exposition in 
Presence and Absence of Trace Elements 
Sample Conditions Parameters 
Without Trace 
Elements 
With Trace 
Elements 
AIO 2 – 8 ° C PV (mmol/L) 0.04 0.19 
Light protected 29 days pH drop 0.01 0.02 
AIO 20 – 30 ° C PV (mmol/L) 0.52 1.92 
Light exposed 29 days pH drop 0.03 0.11 
Intralipid 20% 40 ° C PV (mmol/L) 2.77 18.04 
Light protected 14 days pH drop 0.77 1.54 
Note : PV: peroxide value. 
FIGURE 5 Total peroxide (circles), catalase - resistant peroxides (squares), and H 2 O 2 as the 
difference (triangles) were measured in three lots of oral multivitamins without ribofl avin 
(Tri - Vi - Sol) and three lots of oral multivitamins with 0.6 mg of ribofl avin (Poli - Vi - Sol) by 
time after the initial opening of the bottle. Compared with the preparation without ribofl avin, 
the level of H 2 O 2 was initially higher ( P < 0.01) in (Poli - Vi - Sol), and that level dropped over 
time ( P < 0.05). In both multivitamin preparations, the levels of catalase - resistant and total 
peroxide rose ( P < 0.05) until day 8. Data represent the mean ± standard error of the mean. 
Variations too small relative to the symbol are not shown [30] . 
0 
0 
5 10 
10 
15 20 
20 
25 
30 
40 
50 
60 
70 
80 
0 
10 
20 
30 
40 
50 
60 
70 
80 
Days 
0 5 10 15 20 25 
Days 
Peroxide (mM)

478 ORIGIN OF CONTAMINATION 
limited the experiment to fat - free TPN since lipids are also able to generate peroxides 
[31] . In neonate solutions, the peroxide concentration ranged from 190 to 
300 . mol/L in sets unprotected from light in contrast to the values for sets protected 
from light, which ranged from 60 to 130 . mol/L. In the adult formulation, peroxide 
reached less than one - tenth of the concentration measured in the neonatal preparation 
(Fig. 6 ). The authors attributed the difference to the nutrient composition, as 
the adult formulation was four times lower in vitamins and higher in fi nal amino 
acid and glucose concentrations. Amino acids and glucose contributed to a decrease 
in the concentration of peroxides [32] . 
It is possible to see in Figure 6 that protecting solutions from light and air 
decreases peroxide generation in these formulations. However, because of the high 
peroxide concentration in neonatal solutions, protection from air during the infusion 
is not effi cient. 
The action of vitamins in the presence of lipids was also investigated by these 
researchers [33] . The generation of peroxide was compared between TPN preparations 
left at dark and at daylight for 6 h. Four different formulations containing and 
not containing vitamins and lipids were included in the experiment. The results, 
presented in Figure 7 allowed for the conclusion that lipids had a signifi cant 
but minor additive effect compared with vitamins in generating peroxides. 
Contamination of TPN by air during compounding accounted for the photoinduced 
generation of peroxide in TPN solutions. It was, however, more effective to protect 
the solution from light exposure than to avoid contact with oxygen. 
Since ascorbate reduces photooxidation of lipid emulsions and multivitamin 
preparations (see Figure 4 ) [19] , Lavoie et al. [34] studied the formation of oxidative 
by - products of vitamin C in multivitamins exposed to light. They found that the loss 
of ascorbic acid in photoexposed multivitamin preparations was associated with the 
generation of products other than dehydroascorbate and 2,3 - diketogulonic acid, 
which are the usual products of vitamin C oxidation. The authors showed that 
hydrogen peroxide at concentrations found in TPN solutions induced the transformation 
of dehydroascorbate into new, biologically active compounds that had the 
potential to affect lipid metabolism. They believe that these species have peroxide 
and aldehyde functions [35] . 
Since air (oxygen) is one of the factors responsible for peroxide generation in 
lipid emulsions and vitamin solutions, Balet et al. [36] compared multilayered versus 
single - layered (EVA) bags in terms of oxidation of parenteral nutrition solutions. 
They measured PV, . - tocopherol, and ascorbic acid at the moment of admixing and 
after 6 and 14 days in 24 parenteral solutions. Admixtures in multilayered bags 
showed less oxidation than in EVA bags; no important difference was observed in 
. - tocopherol content, but just after 6 days storage, ascorbic acid and dehydroascorbic 
acid disappeared in the EVA bags. 
Sodium metabisulfi te is an antioxidant agent widely used in pharmaceutical 
preparations to reduce or prevent oxidation. There are some studies, however, that 
have shown that metabisulfi te, under specifi c conditions, may have indirect oxidant 
properties. Baker et al. [37] demonstrated that sulfi te propofol emulsion, but not 
EDTA propofol emulsion, underwent chemical changes during a simulated intravenous 
infusion. Compounds were identifi ed as propofol oxidation products. The 
increase of propofol oxidation products demonstrated that sulfi te from metabisulfi te 
created a strong oxidant environment when air was introduced. Lavoie et al. [38]

ENDOGENOUS CONTAMINANTS 479 
FIGURE 6 ( a ) Peroxide concentration in a neonatal fat - free solution of parenteral nutrition 
run through four different infusion sets: with or without air inlet, protected and unprotected 
from light (lux). Photoprotection was associated with a signifi cantly lower peroxide 
content (indicated by * ; P < 0.05). Protection from light resulted in a greater prevention 
against the generation of peroxides than did protection from air contamination. ( b ) Peroxide 
concentration in an adult solution of fat - free parenteral nutrition run through four different 
infusion sets: with or without air inlet, protected and unprotected from light (lux). The concentration 
of peroxides is 10 - fold lower than in the neonatal solution. Protection from light 
and air was not associated with an overall effect on peroxide generation. Protection from 
light and air resulted in a signifi cant decrease in peroxide generation at 20 and 24 h of infusion 
(indicated by * ; P < 0.05), whereas photoprotection alone resulted in a signifi cant difference 
at 24 h (indicated by +; P < 0.05). Results are expressed as mean ± standard error of 
the mean (SEM); n = 3. TBH = tert - butylhydroperoxide [31] . 
Light/air 
Light/no air 
No light/no air 
No light/air 
Light/air 
Light/no air 
No light/no air 
No light/air 
300 
200 
100 
15 
10
5
0 
4 8 20 24 
Duration of incubation (h) 
(b) 
(a) 
Duration of incubation (h) 
Peroxides (.mol/L eq. TBH) Peroxides (.mol/L eq. TBH) 
0 
2 4 8 20 24 
* 
* 
* 
* 
* 
* 
* 
†

480 ORIGIN OF CONTAMINATION 
showed that sulfi te has been able to cause the oxidation of lipids. The reaction of 
sulfi te with oxygen may lead to several sulfi te - derived oxidant species [39] . 
Besides the oxidation to sulfate ( SO SO e 3 2
4 2
. . . > +2 ), sulfi te may undergo a 
one - electron oxidation leading to the formation of the reactive sulfi te radical 
( SO SO e 3 3 
2. •. . > + ), which reacts with oxygen forming two strong oxidant species: 
sulfi te peroxyl ( SO O SO OO 3 2 3 
•. •. + > ) and sulfate radical ( SO O SO 3 2 4 
•. •. + > ). 
These species and the sulfi te radical itself react as oxidants, turning sodium metabisulfi 
te into a prooxidant agent, depending on the circumstances. 
Peroxidation and free - radical formation should be considered as important 
aspects of pharmaceutical stability and quality of parenteral nutriton and intravenous 
drugs. Peroxidation and free - radical formation depend on environmental 
factors, such as storage conditions and container material, but are also infl uenced 
by formulation components or additives such as tocopherols and metabisulfi te. 
Since the generation of these harmful species occurs generally at the time of use, 
manufacturing quality controls fail in demonstrating their existence. 
6.1.3 EXOGENOUS IMPURITIES 
6.1.3.1 Residual Solvents 
Residual solvents are organic volatile chemicals that remain in active substances, 
excipients, and other pharmaceutical products after processing. In spite of their toxic 
properties, solvents play an important role in the production of pharmaceuticals, 
during the synthesis, separation, or purifi cation, and their use cannot be avoided. 
Solvents in this category do not include those used as excipients. 
FIGURE 7 Infl uence of a lipid emulsion and daylight on peroxide levels in freshly 
prepared solutions of parenteral nutrition containing multivitamins (PN + MVI and PN + 
Lipid + MVI) . (PN = parenteral nutrition; MVI = multi vitamin preparation.) The data represent 
the mean ± SEM, n = 3; the variations are not depicted because of their small size relative 
to the symbols. The peroxide content rose signifi cantly over time ( P < 0.001), and 
exposure to daylight had a signifi cant effect on peroxide generation ( P < 0.001) [33] . 
0 2 4 6 
0 
100 
200 
300 
400 
Peroxides (mM eq. TBH) 
PN + Lipids + MVI, protected from light 
PN + Lipids + MVI, exposed to light 
PN + MVI, protected from light 
PN + MVI, exposed to light 
Time (h)

EXOGENOUS IMPURITIES 481 
Residual solvents are organic volatile chemicals that were not completely removed 
by practical manufacturing techniques and may, therefore, be contaminants in pharmaceutical 
products. 
The International Conference on Harmonisation of Technical Requirements for 
Registration of Pharmaceuticals for Human Use (ICH) [40] has adopted impurities 
guidelines for residual solvents that prescribe limits for the amount of residual solvents 
allowed. 
Residual solvents are divided into three classes. Class 1 solvents are those known 
to cause toxic effects and should be avoided in the production of active substances 
and excipients. Class 2 solvents present less severe toxicity than class 1, and class 3 
solvents have such low toxic potential that exposure limits are not necessary. Table 
12 presents the general characteristics of the solvents included in each class, and 
Table 13 lists the solvents and their concentration limit in pharmaceutical 
products. 
6.1.3.2 Containers 
The primary function of packaging is to provide adequate protection. Pharmacopeial 
compendia have established requirements for containers based on drug form 
characteristics. Thus, while for capsules and tablets the requirements are generally 
related to the design of the container (e.g., tight, well closed), for injectables, and 
ophthalmic, and inhalation products, materials of construction are also addressed, 
as compatibility is a very important issue for this kind of dosage. Considering the 
purpose of this chapter, only those packaging materials whose interaction with the 
formulation is a factor will be focused on here. 
Packaging components for pharmaceuticals are basically made of glass and polymeric 
materials such as plastics and rubbers. 1 In spite of this simple classifi cation, 
glass, plastic, and rubber are not elementary materials but rather complex 
mixtures. 
The evaluation of the chemical stability of a packaging component depends on 
the likelihood of packaging component – dosage form interaction and is usually 
TABLE 12 General Characteristics of Residual Solvents 
Class Action Characteristics 
1 Solvents to be avoided Strongly suspected human carcinogens and 
environmental hazards 
2 Solvents to be limited Nongenotoxic animal carcinogens or 
possible causative agents of other 
irreversible toxicity such as neurotoxicity 
or teratogenicity 
3 Solvents with low toxic potential No health - based exposure limit is needed; 
solvents with permitted daily exposure of 
50 mg or more per day 
Source : From ref. 40 . 
1 Although metallic packaging is also used for pharmaceutical products, its use is restricted to blisters ’ 
supports, which have no contact with the formulation, and tubes for ointments. 

482 ORIGIN OF CONTAMINATION 
TABLE 13 Solvents Included in Each Class and Their Concentration Limit in 
Pharmaceutical Products 
Solvent 
Concentration 
Limit (ppm) Solvent 
Concentration 
Limit (ppm) 
Class 1 Class 3 
2 Acetic acid — 
Benzene 4 Acetone — 
1,2 - Dichloroethane 5 Anisole — 
1,1 - Dichloroethane 8 1 - Butanol — 
1,1,1- Trichloroethane 1500 2-Butanol — 
Class 2 
Butyl acetate — 
Acetonitrile 410 
tert - Butyl methyl ether — 
Chlorobenzene 360 
Cumene — 
Chloroform 60 
Dimethyl sulfoxide — 
Cyclohexane 3880 
Ethanol — 
1,2 - Dichloroethene 1870 
Ethyl acetate — 
Dichloromethane 600 
Ethyl ather — 
1,2 - Dimethoxyethane 100 
Ethyl formate — 
N,N - Dimethylacetamide 1090 
Formic acid — 
N,N - Dimethylformamide 880 
Heptane — 
1,4 - Dioxan 380 
Isobutyl acetate — 
2 - Ethoxyethanol 160 
Isopropyl acetate — 
Ethylene glycol 620 
Methyl acetate — 
Formamide 220 
3 - Methyl - 1 - Butanol — 
Hexane 290 
Methyl ethyl ketone — 
Methanol 3000 
2 - Methyl - 1 - propanol — 
2 - Methoxyethanol 50 
Pentane — 
Methyl butyl ketone 50 
1 - Pentanol — 
Methylcyclohexane 1180 
1 - Propanol — 
N - methylpyrrolidone 4840 
2 - Propanol — 
Nitromethane 50 
Propyl acetate — 
Pyridine 200 
Tetrahydrofuran — 
Sulfolane 160 
Tetralin 100 
Toluene 890 
1,1,2 - Trichloroethene 80 
Xylene 2170 
Source : From ref. 40 . 
— = not limited. 
carried out by exposing a sample of the packaging component to an appropriate 
solvent at elevated temperatures. The resulting extract should be analyzed for 
extractables. Thus, for glass samples the released alkalinity should be evaluated, 
whereas for plastics and rubbers the amount of extractables, as well as their nature, 
are to be determined. The elevated temperature has the purpose of increasing the 
extraction rate and simulating in a short period of time a longer exposure time at 
room temperature. The solvent used for the extraction test should present the same 

EXOGENOUS IMPURITIES 483 
propensity to interact with the packaging material as the dosage form. Although it 
is desirable that the dosage form itself be used in the test, pharmacopeial compendia 
prescribe tests with standard solvents such as purifi ed water, drug vehicle, and isopropyl 
alcohol. 
Even when impurities and degradation profi les of a drug substance have been 
established and containers comply with guidelines, some unexpected drug – 
container interactions can occur during the sterilization procedure or shelf life. 
This section includes the most important container materials: glass, plastic, and 
elastomers and addresses the contamination issues for injectable formulations. 
Glass Containers Glass containers for packaging pharmaceutical preparations 
should meet stability specifi cations, which are related to the presence of certain 
components in the structure of the glass and to the formulation type and use. 
According to pharmacopeial prescriptions, four different types of glass may be used 
for storing pharmaceutical preparations. Table 14 presents the classifi cation of glass 
containers in accordance with the product ’ s nature and pharmaceutical usage. 
Though mostly made of silica, the different glass types are obtained by adding 
or subtracting certain glass components. Table 15 lists the principal components of 
pharmaceutical glasses, based on the criteria presented in Table 14 . Type I glass must 
have at least 10% boron oxide and a higher concentration of aluminum oxide than 
the ordinary soda – lime glass to improve its resistance. The reduced amount of 
sodium oxide lessens the solubility of type I glass in water. Type II glass is made of 
TABLE 14 Glass Classifi cation According to Pharmacopeial Prescription and Their 
Pharmaceutical Usages 
Glass Type General Description Uses 
I Highly resistant borosilicate 
glass 
Parenteral preparations 
II Treated soda – lime glass Acidic or neutral parenteral preparations 
III Soda – lime glass Not for parenteral preparations 
NP Soda – lime glass Oral or topic use 
TABLE 15 Typical Composition of Glass Containers According to the Pharmacopeial 
Classifi cation 
Component Type I, Borosilicate (%) Types II, III, IV, Soda – Lime (%) 
SiO 2 70 73 
B 2 O 3 10 — 
Na 2 O 9 14 
Al 2 O 3 6 2 
BaO 2 — 
K 2 O 1 — 
CaO 1 7 
MgO 0.5 4 
ZnO 0.5 — 
Source : From Corning, Life Science Catalogue: Technical information Description of Glasses Used in 
Corning Labware. 

484 ORIGIN OF CONTAMINATION 
common soda – lime glass after a dealkalinization process on the inside surface at 
the time of manufacture. In this process, known as sulfur treatment, sulfur dioxide 
is introduced as newly formed bottles pass from the forming machine, and it reacts 
at the surface of the glass to form sodium sulfate, according to the equation: 
SO O Na O Na SO 2 2 2 2 4 (glass) + + .>. 1
2 
The container is delivered with a haze of sodium sulfate, which should be rinsed 
away before fi lling. 
Stability test for glass containers prescribed by pharmacopeial compendia is 
limited to the action of water on the glass surface to measure the alkalinity released 
(water attack). After autoclaving the container fi lled with water at 121 ° C for 60 min, 
100 mL of the resulting extract is titrated with 0.02 N H 2 SO 4 . The volume of the 
acidic titrant should not be greater than a specifi c value based on the container ’ s 
capacity. Briefl y, types I and II glass containers with a capacity of less than 10 mL 
should consume less than 2.0 mL titrant, whereas for type III glass containers the 
limit is 20 mL. For containers with a capacity of between 20 and 500 mL, the titrant 
consumption should be from 0.8 to 0.2 mL for types I and II glass and from 8.0 to 
2.2 mL for type III glass. For non - parenteral (NP) glass, no limit of titrant is 
established. 
The released alkalinity is, however, not enough to show the actual stability of a 
glass container. Glass is not as inert as it appears, and water is not the only substance 
that attacks the glass surface. In spite of this, the only test prescribed at USP for 
toxic impurities in glass is arsenic ( < 661 > containers, USP 27); type I glass containers 
should be assayed according to the water attack procedure and present not more 
than 0.1 ppm As. Though not listed as a glass constituent, As may be used as a fi ning 
agent in the glass industry (see below) and, therefore, may be present in the glass 
structure. 
Studies have shown that even type I glasses are not completely inert. Bohrer 
et al. [41] showed that the hydrolytic resistance of glass containers is diminished in 
the presence of some substances commonly present in infusion solutions. Table 16 
presents the released alkalinity and also the concentration of some glass constituents 
(silicate, sodium, boron, aluminum) in the aqueous extract obtained after the 
hydrolytic resistance test. 2 These data show that, despite complying with pharmacopeial 
prescription, glass can release its constituents by action of species in the solutions 
and even in pure water. Basic solutions of bicarbonate and gluconate presented 
the highest level of glass constituents extracted, confi rming the potential of basic 
solutions to attack and dissolve the glass network. Glucose and citric acid interacted 
with the glass surface, selectively extracting not only Al but also Cu and Pb (traces 
also present in the glass). This specifi c action of citrate and glucose could be related 
to their metal - complexing ability. 
The presence of arsenic as a contaminant in pharmaceutical products is a pharmacopeial 
concern. Practically all monographs present limits for arsenic, ranging 
from 0.1 to 3 ppm. Bohrer et al. have shown that arsenic is a ubiquitous contaminant 
2 The hydrolytic resistance test described in this chapter is the “ powdered glass test ” and not the “ water 
attack test ” as described above; see Appendix < 661 > , USP 27 [1] . 

TABLE 16 Volume H 2 SO 4 Solution Consumed in Titration of Solutions before and after Heating in Contact with Glass Mass for 30 min at 
121 ° C and Species Extracted from Glass during the USP Powdered Glass Test Carried out with Clear Ampoules in Presence of Different 
Solutions 
Sample 
Volume 0.02 
N 
H 2 SO 4 
(mL) Species Concentration (mg/L) 
± 
SD ( n 
= 3) 
Before 
Heating 
After 
Heating Na Silicate Borate Al Cu Pb 
NaCl 
a 
0.56 
± 
0.01 — 
8.30 
± 
2.01 2.22 
± 
0.11 0.15 
± 
0.01 0.09 
± 
0.02 0.03 
± 
0.01 
KCl 
a 
0.69 
± 
0.01 14.42 
± 
1.51 9.96 
± 
1.71 1.58 
± 
0.03 0.19 
± 
0.01 0.14 
± 
0.08 0.04 
± 
0.02 
CaCl 2 
a 
0.45 
± 
0.02 13.41 
± 
1.33 7.31 
± 
2.08 2.00 
± 
0.11 0.18 
± 
0.01 0.06 
± 
0.01 0.02 
± 
0.01 
MgCl 2 
a 
0.13 
± 
0.01 15.25 
± 
1.64 6.54 
± 
0.53 3.25 
± 
0.21 0.16 
± 
0.01 0.09 
± 
0.03 0.02 
± 
0.01 
Sodium gluconate 2.68 
± 
0.22 1.83 ± 0.05 — 
246.0 
± 
18.7 1.91 
± 
0.09 2.04 
± 
0.02 0.49 
± 
0.07 0.05 
± 
0.01 
Sodium hydrogen 
phosphate 
a 
10.22 
± 
0.91 — 
10.21 
± 
3.41 2.66 
± 
0.09 4.93 
± 
0.12 0.05 
± 
0.01 0.05 
± 
0.02 
Potassium hydrogen 
phosphate 
a 
8.78 
± 
0.95 18.38 
± 
2.34 45.32 
± 
3.76 1.73 
± 
0.06 4.30 
± 
0.07 0.13 
± 
0.04 0.04 
± 
0.02 
Sodium bicarbonate 24.3 
± 
0.41 12.68 
± 
1.03 — 
271.7 
± 
23.2 5.98 
± 
0.14 5.22 
± 
0.24 0.88 
± 
0.09 0.06 
± 
0.03 
Citric acid 
a 
1.92 
± 
0.04 19.82 ± 1.90 16.46 
± 
0.87 1.44 
± 
0.10 6.36 
± 
0.53 0.49 
± 
0.03 0.10 
± 
0.05 
Glucose 
a 
1.08 
± 
0.05 13.86 
± 
1.22 11.10 
± 
0.88 3.18 
± 
0.17 6.34 
± 
0.33 0.28 
± 
0.01 0.15 
± 
0.03 
Note : 
Concentration of the solutions: 
0.01 mol/L [41] . 
a pH 
. 
7
. 
485

486 ORIGIN OF CONTAMINATION 
in raw materials [16] (see discussion of water in Section 9.1.2.1), but its presence in 
fi nal products occurs in concentrations much higher than in the corresponding raw 
substances [12] . According to the authors, the main source of arsenic in injectable 
formulations is glass containers [16] . As mentioned above, glass can contain arsenic 
because arsenic oxide(III) is a fi ning agent to improve glass transparency [42] , a 
feature especially important for solutions for intravenous administration that must 
be subjected to visual inspection of the content before use [43] . Arsenic oxide(III) 
is supposed to react with potassium nitrate in the glass melt to release oxygen and 
nitrogen oxides. These gases form large bubbles that rapidly rise to the surface, stirring 
the bath and sweeping small bubbles formed by the decomposition of batch 
materials. Table 17 shows the arsenic levels measured by these authors in solutions 
for parenteral nutrition. Among these products, sodium bicarbonate and calcium 
gluconate are the most contaminated, presenting levels even higher than the limit 
of 0.1 mg/L for arsenic in infusion solutions. 
Other elements such as chromium, barium, and zinc have also been found in 
solutions for parenteral nutrition, and, though not stated in the studies, the origin 
of these elements is probably the glass packaging. Since glass is an inorganic material, 
inorganic species may arise as contaminants from glass containers. Exceptions 
could be Zn and Ba, which, besides being present in type I glasses, are components 
of plastic additives as well. Table 18 summarizes inorganic elements as contaminants 
in different infusion solutions and their respective concentrations. 
Data on Al are being presented separately because of its special behavior and 
toxicity. Since the discovery in 1976 by Alfrey and co - workers [48] that Al can accumulate 
in patients with reduced renal function causing toxic manifestations including 
neurological diseases, much has been done to reveal and reduce aluminum 
contamination sources. Nowadays, Al is also considered toxic for patients with 
normal renal function who receive parenteral nutrition, mainly preterm infants [49] . 
Recognized sources of Al include water used for dialysis, solutions for total parenteral 
nutrition, and oral aluminum - containing compounds. Though additives and raw 
materials are sources as well, the most important source of Al in solutions for parenteral 
nutrition is the storage of parenteral preparations in glass containers. Figure 
8 shows the Al level in different kinds of solutions stored in plastic and glass containers 
for a period of three months. While not more than 50 . g/L Al was leached 
from polyethylene (PE) containers, over 3000 . g/L Al was found in solutions stored 
in glass containers , confi rming that glass is a valid and continual source of Al. 
The same authors also showed that though glass is the source, Al leaching depends 
on the nature of the substance in contact with its surface [51, 52] . In an experiment 
with glass and an ion exchanger containing Al attached, complexing agents and also 
amino acids were able to extract Al from both sources, but following an extraction 
yield related to the affi nity of the ligand for the metal. The results of this interaction, 
measured over 2 months, can be seen in Figure 9 . Based in this conclusion, the 
notable contamination by Al in gluconate and phosphate solutions can be attributed 
to the affi nity of these species with Al. They are able to selectively withdraw Al from 
the glass network when in contact with the glass containers (see Table 20 ). 
Figure 10 shows the result of one year of storage of different amino acids, components 
of parenteral solutions in type I glass containers. Not all the solutions were 
contaminated by Al, but only those whose amino acids presented affi nity with this 
element. 

EXOGENOUS IMPURITIES 487 
TABLE 17 Arsenic Species Present as Contaminant in Commercial Parenteral Solutions a 
Product 
Total As 
(. g/L) 
As(V) 
(. g/L ± RSD) 
As(III) 
(. g/L ± RSD) 
KCl 19.1% (10) 41.3 41.3 ± 1.6 n.d. 
KCl 10% (4) 23.6 23.6 ± 0.8 n.d. 
NaCl 20% (10) 43.8 40.9 ± 5.7 2.9 ± 0.2 
NaCl 20% (4) 15.9 15.9 ± 3.2 n.d. 
Sodium acetate 2 meq/mL (1) 46.1 41.8 ± 2.0 4.3 ± 0.6 
Sodium phosphate 0.5 mol/L (10) 36.7 36.7 ± 2.5 n.d. 
Sodium bicarbonate 8.4% (5) 248.6 227.8 ± 0.5 20.8 ± 2.0 
Sodium bicarbonate 8.4% (9) 198.3 147.2 ± 1.2 51.1 ± 0.8 
Calcium gluconate 10% (11) 73.1 46.8 ± 4.1 26.3 ± 2.3 
Calcium gluconate 10% (9) 92.7 72.8 ± 0.5 19.9 ± 1.9 
Calcium gluconate 10% (4) 239.6 176 ± 7.2 63.6 ± 2.0 
Magnesium sulfate 50% (11) 15.7 15.7 ± 0.2 5.7 ± 0.2 
Magnesium sulfate 50% (10) 33.5 12.2 ± 1.2 21.3 ± 1.8 
Magnesium sulfate 50% (9) 53.8 39.9 ± 5.6 13.9 ± 3.0 
Glucose 25% (6) 21.8 16.5 ± 2.6 5.3 ± 0.6 
Glucose 25% (6) 18.6 14.0 + 2.2 4.6 ± 0.7 
Vitamins (Tiaminose b ) 103.7 86.1 ± 0.9 17.6 ± 3.2 
Vitamins (Dextrovitase c ) 10.2 10.2 ± 0.2 3.8 ± 0.1 
Vitamins (Frutovena d ) 61.7 21.5 ± 1.7 40.2 ± 6.7 
Amino acids 10% (3) 2.5 2.5 ± 0.1 n.d. 
Amino acids 10% (1) 15.4 15.4 ± 2.3 n.d. 
Amino acids 10% (2) 41.0 41.0 ± 0.9 n.d. 
Amino acids 8% (2) 21.1 16.7 ± 1.2 4.4 ± 0.4 
Amino acids 8% (2) 94.7 87.9 ± 5.7 6.8 ± 1.0 
Lipid emulsion 10% (2) 0.9 0.9 ± 0.3 n.d. 
Lipid emulsion 20% (2) 1.7 1.7 ± 0.6 n.d. 
Heparin 5000 UI/mL (7) 56.7 17.2 ± 2.8 39.5 ± 2.3 
Heparin 5000 UI/mL (8) 79.4 79.4 ± 4.9 23.3 ± 1.2 
a Number in parentheses refer to product brands: (1) B. Braun, (2) Fresenius, (3) Baxter, (4) Halex Istar, 
(5) JP, (6) Merck, (7) EMS, (8) Elkins Sinn, (9)Ariston, (10) Geyer, (11) Hipolabor. RSD, relative standard 
deviation; n.d., not detected (below LOD). 
b Glucose 3 g, ascorbic acid 0.25 g, thiamin chloridrate 0.015 g in 10 mL. 
c Glucose 2 g, ascorbic acid 2 g, pyridoxine chloridrate 20 mg, nicotinamide 30 mg, ribofl avin in 20 mL. 
d Frutose 5 g, ascorbic acid 1 g, pyridoxine chloridrate 20 mg, sodium pantothenate 10 mg, ribofl avin 4 mg 
in 20 mL. 
Source : Form ref. 16 . 
Table 20 presents the Al level in parenteral formulations, collected by different 
authors in different countries. Not all of them mention glass packaging as responsible 
for Al contamination, but it seems to have been the major source. These data 
also confi rm that the nature of the formulation components plays a key role in 
selective Al leaching. 
These results show that signifi cant variations in the contaminant and level of 
contamination depend on the formulation constituents, brand, and mostly the nature 
of the packaging material. 

488 
TABLE 18 Metallic Contaminants Found in Commercial Solutions for Parenteral Nutrition 
Sample 
Element Concentration ( . g/L) 
Ref. 
Zn 
Cr 
Fe 
Pb 
As 
Mn 
Ba 
Sn 
Ge 
Cu 
Cd 
TPN bag 9.1 — — — — — — — — — — 13 
Amino acids 60 – 4,70 n.d. 
— — — — — — — — — 44 
l - Cysteine HCl 
32,000 – 86,000 110 – 230 — — — — — — — — — 44 
NaCl, 
KCl, 
Na acetate 350 – 560 20 – 230 — — — — — — — — — 44 
Ca gluconate 280 – 2,380 — — — — — — — — — — 44 
Phosphate 910 – 2,330 390 – 440 — — — — — — — — — 44 
Ca gluconate 47 – 244 — 237 – 655 — — — — — — — — 45 
TPN 233 – 703 — 84 — — — — — — — — 45 
Standard adult TPN 
formula, Baxter 
— 86.8 — 0.6 288.0 309.9 11.0 0.5 5.5 — — 46 
Standard adult TPN 
formula, Abbott 
— 25.8 — 1.1 65.0 109.9 16.0 0.4 5.5 — — 46 
Standard adult TPN 
renal formula, Abbott 
— 99.8 — 0.6 298.0 259.9 28.0 2.2 28.9 — — 46 
Standard adult TPN 
formula 
— 139.8 — 0.4 65.9 299.9 22.0 1.4 15.9 — — 46 
Standard adult TPN 
formula, Baxter 
— 15.8 — 0.7 61.0 47.9 73.0 0.6 9.2 — — 46 
Standard adult amino 
acid solution, 
Fresenius 
131.2 — — 4.99 — — — — — 2.96 0.35 47 
Standard adult amino 
acid solution, B. 
Braun 
88.92 — — 16.80 — — — — — 6.66 4.37 47 
Standard adult amino 
acid solution, Baxter 
1.39 — — 4.39 — — — — — 40.81 0.71 47 
Note : 
n.d. 
= not detected, 
— = not measured. 

EXOGENOUS IMPURITIES 489 
FIGURE 8 Aluminum extracted from type II glass containers and from polyethylene containers 
by action of NaCl, KCl, albumin, glucose, heparin, HCl, and NaOH solutions after 30 
and 60 days storage at room temperature. The three different albumins are: A, bovine 
(Merck); B, bovine (Reagen), and C, egg (Sigma) [50] . 
800 
600 
400 
200 
0 
800 
600 
400 
200 
0
NaCl 10% 
NaCl 10% 
KCl 10% 
KCl 10% 
Albumin A 20% 
Albumin A 20% 
Albumin B 20% 
Albumin B 20% 
Albumin C 20% 
Albumin C 20% 
Glucose 10% 
Glucose 10% 
Heparin 5000 UI/mL 
Heparin 5000 UI/mL 
HCL 1 mol/L 
HCL 1 mol/L 
NaOH 1 mol/L 
NaOH 1 mol/L 
Polyethylene containers 
Glass containers 
Al (.g/L) 
Al (.g/L) 
Blank After 30 days After 60 days 4035 .g/L 
2570 .g/L 
Plastic Containers Plastics are organic and polymeric in nature. A polymer is a 
large molecule built up by the repetition of small and simple chemical units. The 
repeated unit of the polymer is usually equivalent or nearly equivalent to the 
monomer or the starting material from which the polymer is formed. The structural 
units of the polymers most used to manufacture plastic containers, along with their 
uses for pharmaceutical purposes, are shown in Table 21 . 
Modern polymer technology is founded on catalysis, and catalytic methods are 
extensively used in the production of plastics. Catalysts, since they only catalyse 
reactions, do not count as polymer constituents but may be present as impurities in 
the polymeric material. Table 22 lists the usual catalysts used for the polymerization 
of the polymers mentioned above, which can be found as contaminants in formulations 
stored in plastic containers. 

490 ORIGIN OF CONTAMINATION 
Other contaminants that can originate from plastic containers are the additives 
necessary to turn the raw polymer into adequate containers. While PE may be used 
without any additive, the other plastics are virtually useless alone but are converted 
into highly serviceable products by combining them with other substances or materials. 
The additives most commonly found in plastics used for pharmaceutical products 
are antioxidants, heat stabilizers, lubricants, plasticizers, fi llers, and colorants. These 
additives can be in liquid, solid, or fi ne particle forms and are used in amounts 
varying from less than 1% to more than 50% of the plastic mass. The additives 
necessary for each of the selected types of polymers are described in Table 23 . 
Additives authorized for use in plastics for pharmaceuticals, along with their 
limits in the polymer mass (according to BP and Ph. Eur.), are summarized in 
Table 24 . 
Since additives, together with the polymers, provide a large variety of substances, 
and the leachability of such components cannot be predicted a priori, 
FIGURE 9 Aluminum extracted from glass particles (18 mesh) and Al - form exchanger 
(18 mesh) as function of time by action of some amino acids and complexing agents. Concentration 
of ligands: 0.05 mol/L [51] . 
6.0 
4.5 
1.5
0 20 40 60 
3.0 
Time (days) 
Al (mmol l–1) Al (mmol l–1) Glass 
EDTA 
NTA 
Citr 
Oxal 
Orn 
Glu 
N-acetyl-tyr 
Arg, Ser 
Asp 
Lys 
water 
EDTA 
NTA 
Citr 
Oxal 
Orn 
Asp 
N-acetyl-tyr 
Arg, Ser, Tyr 
Glu 
Lys 
water 
0 6 12 18 24 30 
Time (days) 
1.8 
1.2 
0.6 
Resin 
FIGURE 10 Aluminum leached from glass containers by amino acids as a function of time 
at room temperature. Amino acid concentration: 0.028 mol/L [53] . 
1200 
900 
600 
300 
0 
0 100 200 300 400 
Cys 
Cystine 
Glu 
Asp 
Arg 
Ala Lys Gly 
Al (.g/L) 
Time (days)

EXOGENOUS IMPURITIES 491 
TABLE 19 Aluminum Present as Contaminant in Commercial Parenteral Solutions, and 
Al Present in Container Materials 
Product 
Al Solution ± SD 
(. g/L) Container 
Al Container 
(%) 
NaCl 20% 149 ± 10 Glass ampoule 1.43 
13 ± 4 Polyethylene 0.04 
KCl 10% 68 ± 6 Glass ampoule 1.25 
23 ± 5 Polyethylene 0.06 
Magnesium sulfate 50% 560 ± 85 Glass ampoule 1.25 
380 ± 288 Glass ampoule 1.43 
Sodium acetate 2 meq/mL 45 ± 7 Glass ampoule 2.14 
17 ± 8 Polyethylene 0.05 
Potassium phosphate 2 meq/mL 988 ± 76 Glass ampoule 1.98 
1325 ± 142 Glass ampoule 2.45 
Sodium phosphate 0.5 M 933 ± 88 Glass ampoule 1.65 
879 ± 203 Glass ampoule 2.05 
Calcium gluconate 10% 5621 ± 1165 Glass ampoule 1.51 
5960 ± 62 Glass ampoule 2.21 
Sodium bicarbonate 8.4% 833 ± 141 Glass bottle 0.99 
922 ± 102 Glass bottle 1.03 
Oligoelements a 1129 ± 33 Glass ampoule 2.14 
Oligoelements b 1854 ± 744 Glass ampoule 2.21 
Amino acids 10% 164 ± 6 Glass bottle 0.82 
Rubber closure 3.91 
Amino acids 10% 116 ± 30 Glass bottle 0.76 
Rubber closure 4.23 
Amino acids 10% 93 ± 23 Glass bottle 0.84 
Rubber closure 5.34 
Amino acids 10% 65 ± 13 Glass bottle 0.89 
Rubber closure 5.70 
Amino acids 10% 23 ± 8 Plastic bag 0.01 
Glucose 50% 13 ± 1 Polyethylene 0.04 
293 ± 14 Glass ampoule 1.15 
Glucose 25% 9 ± 3 Polyethylene 0.08 
370 ± 23 Glass ampoule 1.87 
Albumin 20% 644 ± 58 Glass fl ask 0.67 
Rubber closure 4.06 
149 ± 23 Glass fl ask 0.66 
Rubber closure 3.99 
Heparin 5000 UI/mL 732 ± 23 Glass ampoule 2.88 
738 ± 54 Glass ampoule 3.03 
Note : All solutions were within their guaranteed period of shelf life. 
Source : From ref. 52 . 
Suppliers : Abbott, Ariston, Aster, Baxter, Behring, B.Braun, Darrow, Fresenius, Fujisawa, Gayer, Halex 
Istar, Hypofarma, Santisa, Roche, Zenalb. 
a Composition: 22.0 mg ZnSO 4 , 6.3 mg CuSO 4 , 2.46 mg MnSO 4 , 102.5 mcg CrCl 3 per ampoule. 
b Composition: 8.8 mg ZnSO 4 , 1.60 mg CuSO 4 , 123.04 mcg MnSO 4 20.50 mcg CrCl 3 per ampoule. 

492 ORIGIN OF CONTAMINATION 
TABLE 20 Aluminium Present in Commercial Solutions for Parenteral Administration 
Product Specifi cation Brand 
Al concentration 
. g/L Reference 
Electrolytes 
NaCl 10% Abbott 3 10 
10% Commercial Polfa 14 54 
5.85% Braun 1 11 
1 M Kabi 22 11 
10% Abbott 43 ± 8 * 15 
10% Elkins Sinn 78 ± 7 * 15 
KCl 15.0% Commercial Polfa 18 54 
— Abbott 4 (5 – 11) 10 
— Elkins Sinn 3 10 
7.45% Braun < 0.6 11 
7.45% Kabi 12 11 
7.45% Braun 33 ± 5 * 15 
10% Abbott 97 ± 8 * 15 
10% Aster 23 ± 5 * 15 
Ca chloride — Elkins Sinn 15 (12 – 19) 10 
— Abbott 5 10 
0.5 M Boehringer 27 11 
1 N Kabi 224 11 
Na acetate 2 meq/ml Braun 17 ± 8 * 15 
K acetate — Abbott < 5 10 
Mg sulphate 50% Abbott 5 10 
50% Braun 606 11 
— Geyer 560 ± 85 * 15 
— Ariston 380 ± 288 * 15 
Na phosphate — Invenex 2236 (2026 – 2370) 10 
0.5 M Abbott 933 ± 88 * 15 
0.5 M Geyer 430 ± 177 * 15 
K phosphate — Invenex 92 10 
— Abbott 2189 (2069 – 2301) 10 
— Braun 188 11 
— Kabi 2826 11 
2 meq/l Fresenius 1021 ± 188 * 15 
2 meq/l Braun 988 ± 76 * 15 
2 meq/l Aster 332 ± 27 * 15 
Ca gluconate 10% Elkins Sinn 3973 (1095 – 5565) 10 
10% Lyphomed 2245 (2000 – 2586) 10 
10% Braun 4734 11 
10% Fresenius 6549 11 
10% Pharma Hameln 4421 11 
10% Commercial Polfa 4400 54 
10% Elkins Sinn 3987 ± 993 * 15 
10% Braun 4530 ± 1072 * 15 
10% Ariston 5960 ± 62 * 15 
10% Halex Istar 6781 ± 1842 * 15 

EXOGENOUS IMPURITIES 493 
TABLE 20 Continued 
Product Specifi cation Brand 
Al concentration 
. g/L Reference 
Trace elements (TE) 
Zinc chloride — Abbott 99 (81 – 123) 10 
TE Tracitrans Fresenius 994 11 
Pediatric TE Ped - el Pharmacia 3000 54 
Pediatric TE Peditrace Pharmacia 130 54 
Pediatric TE Ped - element Darrow 1423 ± 68 * 15 
TE Ad - element Darrow 3574 ± 237 * 15 
TE Tracitrans Fresenius 5712 ± 988 * 15 
Amino acids 
Freamine 8.5% McGaw 12 (5 – 24) 10 
Travasol 10% Travenol 7 (6 – 8) 10 
HepatAmine 8% McGaw 22 10 
Aminoplasmal 
10% 
Braun 55 11 
Aminopaed 10% Kabi 38 11 
Aminosteril 8% Fresenius 17 11 
Primene 10% Clintec 120 54 
Aminomel 
12.5% 
Clintec 121 54 
Vaminolact 6.5% Pharmacia 30 54 
Aminoplasmal 
L10 
Braun 160 ± 48 * 15 
Pediamino 10% Braun 116 ± 30 * 15 
Aminoped 10% Fresenius 195 ± 27 * 15 
Nefroamino 
AEH 
Braun 272 ± 66 * 15 
Prim e ne 10% Baxter 65 ± 13 * 15 
Carbohydrates 
Glucosteril 70% Fresenius 9 11 
Glucose 40% Fresenius 20 11 
Glucose 50% Schiwa/Hormonchemie 18 11 
Glucose 20% Schiwa/Hormonchemie 3 11 
Glucose 10% Schiwa/Hormonchemie < 0.6 11 
Dextrose (10, 20, 
50%) 
Abbott < 5 10 
Dextrose (10, 
50%) 
McGaw < 5 10 
Dextrose (5, 10, 
50%) 
Travenol < 5 10 
Glucose 20% Commercial Polfa 16 54 
Glucose 25% Darrow 9 ± 3 * 15 
Equiplex 9 ± 2 * 15 
Glucose 50% Fresenius 15 ± 3 * 15 
Darrow 17 ± 3 * 15 
J.P. Ind. Farm 8 ± 2 * 15 
B. Braum Lab. 13 ± 1 * 15 
Ariston 11 ± 4 * 15 

494 ORIGIN OF CONTAMINATION 
Product Specifi cation Brand 
Al concentration 
. g/L Reference 
Lipids 
Intralipid 10% Kabi Vitrum < 5 10 
Intralipid 20% Kabi Vitrum < 5 10 
Intralipid 10% Pfrimmer Kabi 5 11 
Intralipid 20% Pfrimmer Kabi 7 11 
Lipofundin 20% Braun 35 11 
Lipofundin 20% Braun 14 11 
Intralipid 20% Pharmacia 30 54 
Lipofundin 20% Braun 180 54 
Lipofundin 10% Braun 29 ± 6 * 15 
Lipofundin 20% Braun 34 ± 7 * 15 
Vitamins 
Vitamin C 
500 mg 
EMS 3443 ± 233 * 15 
Vitamin B12 
1mg 
Bunker 92 ± 23 * 15 
B complex Hipolabor 1089 ± 127 * 15 
B complex Ariston 709 ± 65 * 15 
Multivitamin 
MVI 
ICN 588 ± 63 * 15 
Others 
Albumin 20% — 190.4 55 
Albumin 20% Behring 235 ± 12 * 15 
Heparin 
1000 U/mL 
— 211.7 55 
Heparin 
5000 U/mL 
Fujisawa Inc. 732 ± 23 * 15 
Heparin 
5000 U/mL 
Crist a lia 72 ± 6 * 15 
— : not informed. 
* mean value ± standard deviation of three samples of the same lot. 
TABLE 20 Continued 
pharmacopeial compendia have determined procedures to investigate biological 
and physical – chemical properties of plastics. In general, tests do not include the 
measurement of the polymeric material or the additives themselves but the biological 
reactivity (toxicity screening) of their extracts. Extracts are generally obtained 
by autoclaving the plastic container fi lled with water at 121 ° C for 30 min or at 100 ° C 
for 2 h. Only containers for parenteral and ophthalmic preparations have a more 
controlled regulation since their risk of toxicity is more signifi cant, and therefore 
assays for plastic additives are foreseen in some pharmacopeial compendia [5] . 
Regulatory guidelines for plastics for pharmaceutical containers also set limits 
for impurities other than additives. These are inorganic species (metallic and nonmetallic) 
that can also be present in plastic containers, and their determination 
serves as a quality criterion for the plastic material. They may or may not be related 

EXOGENOUS IMPURITIES 495 
TABLE 21 Polymeric Materials, Their Structural Units, and Uses in Pharmaceutical 
Products 
Polymer Monomers Uses in Biotechnological Products 
Polyethylene (PE) Ethylene Dry dosage forms 
Noninjectable aqueous solutions 
Intravenous aqueous infusions 
Polypropylene (PP) Propylene Dry dosage forms 
Noninjectable aqueous solutions 
Intravenous aqueous infusions 
Polyvinyl chloride 
(PVC) 
Vinyl chloride Dry dosage forms 
Noninjectable aqueous solutions 
Intravenous aqueous infusions 
Blood and blood components 
Tubing for blood 
Polyethylene vinyl 
acetate (EVA) 
Ethylene and vinyl acetate Intravenous aqueous infusions 
Tubing for parenteral nutrition 
preparations 
Polyethylene 
terephthalate (PET) 
Terephthalic acid or 
dimethyl terephthalate 
and ethylene glycol 
Dry oral dosage forms 
Liquid oral dosage forms 
Source : Data from refs. 5 (BP) and 56 (FDA). 
TABLE 22 Catalysts Used for Polymerization Reaction of Plastics Materials Listed in 
Table 21 
Polymer Catalysts 
Polyethylene (PE) TiCl 4 + Al(C 2 H 5 ) 3 ; CrO 3 /SiO 2 , MoO 3 /Al 2 O 3 
Polypropylene (PP) . - TiCl 3 + Al(C 2 H 5 ) 3 
Polyvinyl chloride (PVC) — 
Polyethylene vinyl acetate (EVA) — 
Polyethylene terephthalate (PET) — 
Source : From ref. 57 . 
TABLE 23 Overview on Additives Commonly Used in Plastic Manufacturing 
Polymer 
Additives 
Antioxidant 
Heat 
Stabilizer Lubricant Plasticizer Filler Colorant 
PE . — . — . . 
PP . — . — . — 
PVC — . . . . . 
EVA . — . — . . 
PET — — . — . . 
Source : From ref. 58 . 

496 ORIGIN OF CONTAMINATION 
TABLE 24 Acceptable Additives Allowed in Plastics for Pharmaceuticals According to 
the BP and Ph. Eur. and Their Limits 
Additive 
Number Additive Name Polymer Limit (%) 
Plasticizer 
1 Di(2 - ethylhexyl)phthalate PVC 40 
2 Zinc octanoate PVC 1 
3 N.N. - diacylethylenediamines PVC 1 
4 Epoxidized soya oil PVC 10 
5 Epoxidized linseed oil PVC 10 
Antioxidants 
7 Butylhydroxytoluene PE, PP, 
EVA 
0.125 
8 Ethylene bis[3,3 - bis[3 - (1,1 - dimethylethyl) 
- 4 - hydroxyphenyl]butanoate] 
PE, PP 0.3 
9 Pentaerythrityl tetrakis[3 - (3,5 - di -tert - 
butyl - 4 - hydroxyphenyl)propionate] 
PE, PP 
EVA 
0.3 
0.2 
10 2,2 . ,2 . ,6,6 . ,6 . - hexa - tert - butyl - 4,4 . ,4 . 
[(2,4,6 - trimethyl - 1,3,5 - benzenetriyl) 
- trismethylene]triphenol 
PE, PP 
EVA 
0.3 
0.2 
11 Octadecyl 3 - (3,5 - di - tert - butyl - 4 
- hydroxyphenyl)propionate 
PE, PP 
EVA 
0.3 
0.2 
12 tri(2,4 - di - tert - Butylphenyl) phosphite PE, PP 
EVA 
0.3 
0.2 
13 1,3,5 - tris(3,5 - di - tert - Butyl - 4 - hydroxybenzyl) 
- s - triazine - 2,4,6(1 H ,3 H ,5 H ) - trione 
PE, PP 0.3 
14 2,2 . - bis(octadecyloxy) - 5,5 . - spirobi 
[1,3,2 - dioxaphosphinane] 
PE, PP 0.3 
15 Dioctadecyl disulfi de PE, PP 0.3 
16 Didodecyl 3,3 . - thiodipropionate PE, PP 0.3 
17 Dioctadecyl 3,3 . - thiodipropionate PE, PP 0.3 
18 a 2,4 - bis(1,1 - Dimethylethyl)phenyl 
biphenyl - 4,4 . - diyldiphosphonite 
2,4 - bis(1,1 - Dimethylethyl)phenyl 
biphenyl - 3,4 . - diyldiphosphonite 
2,4 - bis(1,1 - Dimethylethyl)phenyl 
biphenyl - 3,3 . - diyldiphosphonite 
2,4 - bis(1,1 - Dimethylethyl)phenyl 
biphenyl - 4 - ylphosphonite 
2,4 - bis(1,1 - Dimethylethyl)phenyl phosphite 
2,4 - bis(1,1 - Dimethylethyl)phenyl 4 . - [bis[2,4 - bis 
(1,1 - Dimethylethyl)phenoxy]phosphanyl] 
biphenyl - 4 - ylphosphonate 
2,4 - bis(1,1 - Dimethylethyl)phenol 
PE, PP 0.1 
Lubricants and Fillers 
19 Stearic acid 
20 Oleamide EVA 0.5 
21 Erucamide EVA 0.5 
22 a Copolymer of dimethyl butanedioate 
and 1 - (2 - hydroxyethyl) - 2,2,6, 
6 - tetramethylpiperidin - 4 - ol 
PE, PP 0.3 

EXOGENOUS IMPURITIES 497 
Additive 
Number Additive Name Polymer Limit (%) 
23 Hydrotalcite PE, PP 0.5 
24 Alkanamides PE, PP 0.5 
25 Alkenamides PE, PP 0.5 
26 Sodium silico - aluminate PE, PP 0.5 
27 Silica PE, PP 0.5 
28 Sodium benzoate PE, PP 0.5 
29 Fatty acid esters PE, PP 0.5 
30 Trisodium phosphate PE, PP 0.5 
31 Liquid paraffi n PE, PP 0.5 
32 Zinc oxide PE, PP 0.5 
33 Magnesium oxide PE, PP 0.5 
34 Calcium stearate PE, PP 0.5 
35 Zinc stearate PE, PP 0.5 
36 Talc PP 0.5 
Colorants 
37 Titanium dioxide PE, PP 4 
38 Ultramarine blue PVC n.i. 
Note : PE and PP for parenteral and ophthalmic preparations may contain at most three oxidants; n.i.: 
not informed. 
a Only for nonparenteral preparations. 
TABLE 24 Continued 
to any of the additives listed in Table 24 . For example, the investigation of aluminum 
in PE and polypropylene (PP) is related to the use of aluminum - based catalysts for 
the attainment of the polymers; barium as a contaminant in a plastic container is 
indicative of the presence of a barium soap stabilizer, possibly used in plastic compounding. 
The tests, including chemicals such as sulfated ash residue and heavy 
metals, are listed in Table 25 with the respective limits for each species in the different 
plastic materials. 
While not mandatory from regulatory guidelines, much research has been carried 
out to investigate the extractability of plastic additives in contact with a variety of 
pharmaceutical formulations, mainly those for parenteral use. The research concentrates 
on the extractability of plasticizer phthalates, mainly di - 2 - ethylhexylphthalate 
(DEHP) from polyvinyl chloride (PVC) into the blood, blood components, and 
infusion solutions. The purpose for these studies lies in its, up to now, controversial 
hazardous effects on humans. The amount of additive necessary to turn rigid PVC 
into a fl exible material (40% m/m) and the absence of chemical bonds between the 
polymer and the plasticizer make it a potentially extractable species. 
Limits for DEHP in formulations stored in PVC bags are not set in pharmacopeial 
compendia. The BP prescribes a standard test whose results should show that 
the amount of DEHP extracted by action of ether on PVC bags does not exceed 
40% of the polymer mass. This test is valid for PVC for all uses: Dry dosage forms, 
noninjectable aqueous solutions, intravenous aqueous infusions, blood and blood 
components, and tubing used in sets for blood and blood components. On the other 

498 ORIGIN OF CONTAMINATION 
TABLE 25 Inorganic Impurities and Their Limits in Extracts of Plastic Containers for 
Pharmaceuticals 
Impurity Number Impurity Polymer Use Limit 
1 Heavy metals PE, PP 
PVC 
Injectable, ophthalmic 
Aqueous parenteral, blood, 
tubing 
2.5 ppm 
50 ppm 
2 Sulfated ash PE, PP 
EVA 
PET 
Injectable, ophthalmic 
Parenteral container and 
tubing 
Oral forms 
0.2%, 1% 
1.2% 
0.5% 
3 Aluminum PE, PP 
PET 
Injectable, ophthalmic 
Oral forms 
1 ppm 
1 ppm 
4 Chromium PE, PP Injectable, ophthalmic 0.05 ppm 
5 Titanium PE, PP 
PET 
Injectable, ophthalmic 
Oral forms 
1 ppm 
1 ppm 
6 Vanadium PE, PP injectable, ophthalmic 0.1 ppm 
Zinc PE, PP 
PVC 
PET 
Injectable, ophthalmic 
Aqueous parenteral, blood 
Oral forms 
1 ppm 
0.2% 
1 ppm 
7 Zirconium PE Injectable, ophthalmic 0.1 ppm 
8 Barium PVC 
PET 
Aqueous parenteral, blood, 
tubing 
Oral forms 
5 ppm 
1 ppm 
9 Cadmium PVC Aqueous parenteral blood, 
tubing 
0.6 ppm 
10 Calcium PVC Aqueous parenteral, blood 0.07% 
11 Tin PVC Aqueous parenteral, blood, 
tubing 
20 ppm 
12 Ammonium PVC Blood 2 ppm 
13 Chloride PVC Blood 0.4 ppm 
14 Antimonium PET Oral forms 1 ppm 
15 Cobalt PET Oral forms 1 ppm 
16 Germanium PET Oral forms 1 ppm 
17 Manganese PET Oral forms 1 ppm 
Source : Adapted from Ph. Eur [ref. 7 ]. 
hand, the Food and Drug Administration (FDA) launched a nonregulatory publication 
entitled “ Safety Assessment of Di - 2 - ethylhexylphthalate (DEHP) Released 
from PVC Medical Devices ” [59] , where concerns about its toxicity are discussed 
in detail. This publication intends to provide risk managers with the information 
necessary to make decisions about the safety of DEHP exposure from medical 
devices. Conclusions of the study were reached by calculating the dose of DEHP 
received by patients undergoing different medical procedures. For the pharmaceutical 
industry, what is truly relevant are concerns about the presence of DEHP in 
intravenous (IV) infusion fl uids and drugs and in parenteral nutrition because of 
their storage in plastic bags and the delivery of these preparations through PVC 
tubing sets. The study concluded that there is little to no risk posed by patient exposure 
to the amount of DEHP released from PVC IV bags following the infusion of 
crystalloid fl uids (e.g., normal saline, Ringer ’ s lactate). There is a small risk, however, 

EXOGENOUS IMPURITIES 499 
3 The tolerable intake values for DEHP are outlined in the International Organization for Standardization 
ISO/DIS 10993 - 17 standard, “ Method for the Establishment of Allowable Limits for Leachable 
Substances. ” 
posed by the exposure to the amount of DEHP released from PVC bags used to 
store and administer drugs that require a pharmaceutical vehicle for solubilization. 
The dose of DEHP received by adult patients receiving TPN is estimated to be less 
than the tolerable intake (TI), 3 suggesting that there is little need for concern about 
DEHP - mediated effects in these patients. The dose of DEHP received by neonates 
undergoing TPN is uncertain. Depending on the data used to derive the TI/dose 
ratio, neonates may be at an increased risk of DEHP - mediated adverse effects. 
In fact, studies have shown that, depending on the constituents in the formulation, 
the extractable DEHP can be much higher than the expected. Our knowledge of 
the nature of the polymer and its additives, as well as of standard migration, is not 
great enough to predict container – formulation interactions. Migration is a unique 
phenomenon, and predictions based only on the structure and physical or chemical 
properties may not be effective for solving interaction problems. Table 26 shows the 
level of DEHP found in different kinds of parenteral formulations stored in PVC 
bags. It is clear that the amount of DEHP leached into the formulation depends on 
the nature of its constituents. While in saline solutions the amount did not exceed 
TABLE 26 Concentration of DEHP in Infusion Solutions stored in Plastic 
containers 
Sample Volume (mL) Time of storage DEHP ( . g/L) Ref. 
0.9% saline 1000 5 months 
14 months 
7 
24 
60 
0.9% saline 500 0.5 month 
3 months 
13 
< 4 
60 
0.9% saline 100 6 months 8 60 
0.9% saline 1000 — 8 60 
5% glucose 1000 5 months 
12 months 
4 
4 
60 
5% glucose 500 0.5 month 
3 months 
34 
7 
60 
ACO bag 100 12 months 7 61 
Travenol bag 100 12 months 5 61 
Multivitamins 48 h (45 ° C) 21 62 
TPN — 3.85% lipid 2200 – 24 h 330 63 
TPN — 3.85% lipid 2200 – 1 week (4 ° C) 450 63 
TPN — 2.50% lipid 2200 – 24 h 230 63 
TPN — 2.50% lipid 2200 – 1 week (4 ° C) 260 63 
TPN — 1.85% lipid 650 24 h 300 63 
TPN — 1.85% lipid 650 – 1 week (4 ° C) 360 63 
TPN — 1.00% lipid 800 – 24 h 240 63 
TPN — 1.00% lipid 800 – 1 week (4 ° C) 270 63 
5% glucose – 5% paclitaxel 1 h 6,600 64 
5% glucose – 5% paclitaxel 2 h 18,500 64 
5% glucose – 10% paclitaxel 1 h 29,400 64 
5% glucose – 10% paclitaxel 2 h 56,600 64 

500 ORIGIN OF CONTAMINATION 
24 . g/L in 14 months, in oleaginous vehicles, it reached over 300 . g/L in 2 h, which 
corroborates with FDA studies. Due to the lipophilic character of DEHP, its migration 
process from the packaging into hydrophilic preparations is limited to a random 
dissolution on the contacting surface. On the other hand, in lipophilic preparations 
such as lipid emulsions, DEHP is encouraged to leave the polymer surface through 
a dissolution process into the lipophilic milieu. 
There are drug products whose interaction with PVC bags and infusion sets are 
so high that they must include labeling precautions for use with PVC containers. 
These drugs include antineoplastics such as paclitaxel, docetaxel, tacrolimus, and 
teniposide, and others such as ciprofl oxacin, cefoperazone sodium, fl uconazole, metronidazole 
HCl, cimetidine, and propofol [64, 65] . 
Although differences in the methods of the studies shown in Table 26 (time of 
contact, volume of solution, temperature, etc.) do not allow for direct comparisons, 
the data presented suggest that the leachability of DEHP is highly dependent on 
the nature of the pharmaceutical formulation. The question of migration of DEHP 
into solutions is relevant not only to lipophilic vehicles but to drugs such as paclitaxel 
and etoposide, which are able to extract much more DEHP than simple lipophilic 
lipidic emulsions do. 
Attempts to substitute PVC containers have shown, however, that non - PVC 
containers may also contain plasticizer components resulting in the release of some 
DEHP into the preparation. A study conducted by Sautou - Miranda and co - workers 
showed that [65] leaches rapidly from coextruded and triple - layered IV tubing into 
etoposide infusion solution (see details of the results under Section 9.1.3.3). 
Although most studies focus on DEHP, it is not the only additive that can be 
leached from plastic materials. Scalia et al. [66] investigated the migration of 
antioxidants from polyolefi nic plastics into oleaginous vehicles. The study concentrated 
on two antioxidants, Irganox 1010 and Irgafos 168, used in PE and PP polymers. 
4 The study was carried out by dipping plastic sheets containing 0.15% of 
antioxidants into a mixture of fi ve oils (caprylic, capric triglyceride, cyclomethicone, 
dicaprylyl ether, isohexadecane, and C 12 – 15 alkyl benzoate) commonly used as pharmaceutical 
vehicles, and storing the oil mixture in bottles manufactured with the 
polymers. The results showed that the amount of antioxidant leached into the oleaginous 
mixture varied remarkably in relation to the polymeric material, and 
decreases in the order EP > RACO > PP > high - density polyethylene (HDPE) 
(RACO and EP are ethylene – propylene copolymers). Table 27 shows the percentage 
of antioxidants left in the polyolefi n bottles fi lled with the mixture and stored 
at 25 ° C for 1 year. The authors concluded that the migration of both antioxidants 
is related to the polymer crystallinity and structure: The higher the crystallinity, the 
lower the additive release. They suggested that PP and HDPE are satisfactory for 
the manufacture of oil. 
Elastomeric Closures Closures for pharmaceutical products are generally made 
of polymeric materials, which may be of either a synthetic or natural origin. While 
brittle closures such as screw caps are made of conventional thermoplastics with a 
single composition, elastomeric closures are made of complex mixtures of many 
4 Pentaerythrityl tetrakis(3,5 - di - tert - butyl - 4 - hydroxyphenyl)propionate (Irganox 1010) and tris(2,4 - di - 
tert - butylphenyl)phosphite (Irgafos 168), additives 9 and 12 in Table 25 . 

EXOGENOUS IMPURITIES 501 
ingredients. Because elastomers or rubbers can be molded into an almost limitless 
variety of permanent shapes and are easily penetrated by needles and reseal after 
needle withdrawal, they are primary closures for parenteral vials and for preparations 
intended for repeated use. 
The polymeric materials usually used to manufacture rigid closures are practically 
the same as those seen under plastic containers (Section 6.1.3.2 ). The same 
impurities are therefore to be expected in these packaging components. On the 
other hand, though made of polymeric materials, elastomeric closures present a 
different structure. In the manufacture of rubber, elastomer, the chief component, 
is combined with other chemicals to produce a material with specifi c properties that 
meet target needs, such as its above - mentioned ability to reseal on repeated use. 
Table 28 lists the common elastomers used in the pharmaceutical industry and their 
monomeric structures. 
The substances listed in Table 28 correspond to the basic structure of elastomeric 
closures. The other components in rubber formulations are curing or vulcanizing 
agents, accelerators, activators, antidegradants, plasticizers, fi llers, and pigments. The 
most common additives used to compound rubber for the pharmaceutical industry 
are listed in Table 29 . The amount of each component may vary from rubber to 
rubber, and, depending on the component, the amount can reach more than 50% 
of the total mass of a formulation. While accelerators are used in amounts of around 
1%, fi llers may make up more than 50% of the formulation mass. 
Table 30 shows four typical pharmaceutical rubber formulations based on natural, 
halobutyl, ethylenepropylenediene (EPDM), and silicone elastomers. 
The presence of such a variety of components makes elastomeric closures a 
potential source of contaminants. Polymeric materials present a reasonable inertness, 
however, the possibility of the other components reaching the drug formulation 
should be considered. Moreover, these ingredients, though intended for 
pharmaceutical purposes and therefore meeting pharmacopeial requirements, may 
still present their own impurities as contaminants. Carbon black is usually an impure 
material, and it can contain polynuclear aromatic hydrocarbons [68] , which can be 
extracted into the packaged drug. Clays contain metallic impurities that can also be 
extracted into the drug formulation or even react with a drug constituent. 
Pharmacopeial monographs do not set limits for additives, as they do for plastic 
polymers, and rubber closure (BP, Appendix XIX, USP . 381 . ) tests are limited to 
TABLE 27 Percentage of Antioxidants Left in Polyolefi n 
Bottles Filled with Mixture and Stored at 25 ° C for 1 year 
Polyolefi n Bottles 
Antioxidant Left (%) 
Irganox 1010 Irgafos 168 
HDPE 92.7 ± 5.7 47.8 ± 5.2 
PP 76.3 ± 7.1 47.0 ± 4.1 
RACO 57.9 ± 6.3 36.4 ± 4.7 
EP 0 0 
Source : From ref. 66 . 
EP: ethylene - propulene amorphous copolymer blend; RACO: ethylene 
-co - propylene random copolymer. 

TABLE 28 Common Elastomers Used in the Pharmaceutical Industry 
Common Name Chemical Name Structure 
Butyl rubber Poly(isobutylene - isoprene) 
C CH ( ) 2 
CH3 
CH3 
CH2 C
CH3 
CH CH2 
n 
50 
Halobutyl rubber Halogenated poly(isobutylene - isoprene) 
C CH2 
CH3 
CH3 
CH C
CH3 
CH CH2 
n 
65 
X 
X = Cl or Br 
( ) 
Ethylene – propylene rubber Poly(ethylene - propylene) 
CH2 CH2 CH2 
CH3 
CH 
n 
( )
3 
502

Common Name Chemical Name Structure 
Ethylene – propylene – diene rubber Poly(ethylene - propylene - diene) 
CH2 CH2 CH2 
CH3 
CH 
n 
15
( ) ) ( 
5 
diene 
Silicone rubber Polydimethylsiloxane 
CH3 
CH3 n 
Si O 
Urethane rubber Adipic acid – ethylene glycol polyester 
HO CH2 O CH2 C
O 
C
O 
CH2 O 
n
OH ( )
2 
( )
4 
( )
2 
Fluoroelastomers Poly(tetrafl 
uoroethylene) 
C C
F F
F F n 
Natural rubber 
cis - 1,4 - Polyisoprene 
C CH2 
CH3 
CH CH2 
n 
503

Common Name Chemical Name Structure 
Polyisoprene rubber 
cis - 1,4 - Polyisoprene 
C CH2 
CH3 
CH CH2 
n 
Neoprene rubber Polychloroprene 
C CH2 
Cl 
CH CH2 
n 
Styrene butadiene rubber Poly(butadiene - styrene) 
CH CH2 CH CH2 
n 
CH2 CH 
C6H5 
( )
4 
Nitrile rubber Poly(butadiene - acrylonitrile) 
( ) CH CH2 CH CH2 
n 
CH2 CH 
CN 
( )
5 2 
Polybutadiene Polybutadiene 
CH CH2 CH CH2 
n 
Source : 
From ref. 
67 . 
TABLE 28 Continued 
504

TABLE 29 Ingredients Other Than Elastomers Used to Compound Rubbers for Pharmaceutical Industry 
Curing Agents Accelerators Activators Antidegradants Plasticizers Fillers Pigments 
Sulfur Hexamethylene 
tetramine 
Zinc 
Oxide Amines Paraffi 
nic waxes Carbon Black Carbon black 
Sulfur - containing 
chemicals 
Ditiocarbamates Stearic acid Hindered phenols Silicon oil Alumnum silicates (clays) 
Titanium dioxide 
Peroxides Sulfonamides 
Waxes Paraffi 
nic oils Magnesium silicate (talc) Iron oxide 
Cadmium oxide Thiuram 
Naphthenic oils Barium sulfate Chromium oxide 
Magnesium oxides Thiazole 
Organic phosphates Zinc oxide Organic dyes 
Zinc oxide 
Silica 
Source : 
From ref. 
67 . 
505

506 ORIGIN OF CONTAMINATION 
showing the content of sulfur, sulfated ash, volatile sulfi des, and extractable zinc, 
ammonium, and heavy metals. 
Extractability tests prescribed by other regulatory agencies [FDA, Parenteral 
Drug Association (PDA)] for closures for drug packaging [69, 70] are also limited 
to the amount of extractable residues or tests to evaluate the in vivo reaction of the 
extractable residue when the material fails in the in vitro tests. 
The tests, however, are neither qualitative (in the sense of showing which substances 
can be extracted) nor specifi cally quantitative since they are conceived to 
show only the total amount of extractable as a residue. USP . 381 . “ Elastomeric 
Closures for Injection, ” for example, recommends the calculation of the weight of 
the residue after evaporating the solvent (purifi ed water, drug vehicle, or isopropyl 
alcohol) used for extraction. Tests in vivo are recommended only when the material 
does not meet the requirements of the in vitro tests. 
These tests, however, are unable to show further contamination in the fi nal 
product since the extractability also depends on the interaction between the container 
and the formulation constituents. In spite of possible interactions, very little 
is known about the leachability of closure constituents through the direct contact 
of closures with formulations. 
Mennermaa and colleagues [71] determined the composition of three different 
types of rubber stoppers used to seal parenteral solutions. After immersing the stoppers 
in 0.9% NaCl solution and autoclaving at 121 ° C for 15 min, the analysis of the 
aqueous extract was carried out by proton - induced X - ray emission. Table 31 presents 
the elemental composition of the rubber stoppers investigated in micrograms/ 
gram dry weight. The concentration of zinc varied from sample to sample by a factor 
of 40,000 (5 – 20,000 ppm). Titanium, Fe, and Br were present in all stoppers, and even 
Pb (2 ppm) was found in one extract. 
TABLE 30 Composition of Four Typical Rubbers Used as Closures for Pharmaceutical Formulations 
Based on Natural, Halobutyl, EPDM, and Silicone Elastomers 
Ingredient 
Rubber Type 
Red Rubber Gray Rubber Gray Rubber Black Rubber 
Elastomer Natural rubber Halobutyl rubber Dimethylpolysiloxane 
polymer 
EPDM 
Filler Alumnum silicate Alumnum silicate Silica Carbon black 
Plasticizer Paraffi nic oil Naphthenic oil — Naphthenic oil 
Pigment Iron oxide Titanium dioxide 
Carbon black 
Carbon black — 
Activator Zinc oxide 
Stearic acid 
— — Zinc oxide 
Stearic acid 
Accelerator Thiuram 
Thiazole 
Thiuram — Thiuram 
Zinc 
dithiocarbamate 
Antidegradant Butylated 
hydroxytoluene 
Butylated 
hydroxytoluene 
— — 
Curing agent Sulfur Zinc oxide 2,4 - Dichlorobenzoyl 
peroxide 
Sulfur 
Source : From ref. 67 . 

EXOGENOUS IMPURITIES 507 
Schoenmakers et al. [72] analyzed two representative commercial rubbers by gas 
chromatography – mass spectrometry (GC – MS) and detected more than 100 different 
compounds. The rubbers, mixtures of isobutylene and isoprene, were analyzed 
after being cryogenically grinded and submitted to two different extraction procedures: 
a Sohxlet extraction with a series of solvents and a static - headspace extraction, 
which entailed placing the sample in a 20 - mL sealed vial in an oven at 110 ° C 
for 5, 20, or 50 min. Although these are not the conditions to which pharmaceutical 
products are submitted, the results may give an idea of which compounds could be 
expected from these materials. Residual monomers, isobutylene in the dimeric or 
tetrameric form, and compounds derived from the scission of the polymeric chain 
were found in the extracts. Table 32 presents an overview of the nature of the compounds 
identifi ed in the headspace and Soxhlet extracts of the polymers. While the 
liquid - phase extraction was able to extract less volatile compounds, the headspace 
technique was able to show the presence of compounds with low molecular mass 
TABLE 31 Elemental Composition of Rubber Stoppers Based on Element 
Concentration in Aqueous Extract 
Element 
Amount ( . g/g) 
Sample 1, Old 
Formulation of 
Bromolutyl 
Siliconized 
Rubber 
Sample 2, New 
Formulation of 
Bromolutyl 
Siliconized Rubber 
Sample 3, New 
Formulation of 
Chlorolutyl 
Siliconized Rubber 
Ti 500 1,600 200 
Fe 7,500 8,000 4,500 
Cu 10 10 — 
Zn 50 5 20,000 
Br 100 7,500 50 
Pb 2 — — 
Source : From ref. 71 . 
TABLE 32 Nature of Compounds Identifi ed by GC - MS in Headspace and Soxhlet 
Extracts of Two Rubbers under Study 
Compound Identifi ed in Extracts Obtained with 
Headspace Soxhlet 
Alkanes C 5 and higher C 9 – C 30 
Oligomers Tmax (elution) = 235 ° C Tmax (elution) = 280 ° C 
Aromatics 92 < MW < 132 168 < MW < 182 
Fatty acids No 228 < MW < 352 
Esters No Yes 
2,6 - di - tert - Butyl - p - cresol Yes Yes 
2,6 - di - tert - Butyl - p - benzoquinone Yes Yes 
Ketone MW = 198 MW = 198 
Source : From ref. 72 . 
Note : MW, molecular weight. 

508 ORIGIN OF CONTAMINATION 
in the extract. A complete screening of the extracted substances can be seen in 
Figure 11 , where the graphics depict the abundance of the different classes of compounds 
detected in the extracts of different solvents for both rubbers. Not only 
oligomers, but also phthalates, phenols, and acidic compounds, were found in the 
extracts, mainly from rubber P1. 
Jenke [73] studied the extractability of aniline, diphenylguanidine, dedenzylamine, 
and triisopropanolamine from a synthetic polyisoprene rubber similar to the 
material used in pharmaceutical applications. Rubber samples were autoclaved 
(121 ° C) in contact with water or NaCl 0.9% solution for 1 h. Table 33 presents the 
concentration of each compound in solution after the extraction procedure using 
2 g rubber material. Extraction profi les ranged between 1.64 and 3.73 mg/L, with the 
exception of diphenylguanidine, whose extraction yield reached 11.76 mg/L. 
6.1.3.3 Delivery Systems 
Pharmaceutical formulations also have direct contact with plastic materials through 
delivery sets, transfer tubing, and devices, as well as during the phases of their production, 
via gaskets, fi lters, and transportation. 
Even though there is no specifi c pharmacopeial prescription for testing these 
materials, they should present the same profi le stipulated for container and closures. 
In the case of tubing lines, which are also made of elastomeric materials, an 
extractability test for elastomeric closures is recommended, mainly for those made 
of PVC. 
It is more diffi cult to generalize about the leachability of plastic tubing materials 
than it is for plastic containers or closures themselves since it depends not only on 
the nature of the plastic material but also on the fl ow rate, temperature, solvent, and 
time of contact with the tubing. 
Studies conducted by different authors on the release of chemical substances 
from medical devices, mainly those used for infusing solutions, show that these are 
potential sources of contamination for pharmaceutical formulations. One of the 
most studied is diethylhexyl phthalate, the same plasticizer found in PVC infusion 
bags to give fl exibility. The same concerns about the use of PVC bags for the storage 
of lipids or lipophilic formulations are valid for tubing. 
Table 34 presents the amount of DEHP leached from typical PVC infusion lines 
from a delivery set of 2.25 m kept at 27 ° C. A sample volume from 8 to 140 mL was 
perfused in 24 h through the lines and collected for analysis. It was possible to see 
that while amino acids did not promote the migration of DEHP from PVC tubing, 
TABLE 33 Accumulation of Extractable from Synthetic 
Polyisoprene Rubber after Autoclaving for 1 h 
Compound Concentration in Solution (mg/L) 
Aniline 1.64 
Diphenylguanidine 11.76 
Dedenzylamine 2.12 
Triisopropanolamine 3.73 
Source : From ref. 73 . 

EXOGENOUS IMPURITIES 509 
FIGURE 11 Infl uence of the solvent used for the Soxhlet extractions of two different commercial 
rubber closures (P1) and (P2) on the area of the chromatographic peaks of different 
classes of compounds detected in GC – MS [72] . 
16 . 107 
14 . 107 
12 . 107 
10 . 107 
8 . 107 
6 . 107 
4 . 107 
2 . 107 
Area 
0 
18 . 107 
16 . 107 
14 . 107 
12 . 107 
10 . 107 
8 . 107 
6 . 107 
4 . 107 
2 . 107 
Area 
0
Acid 
Acid 
Acid 
Acid 
Acid 
Phenol 
Phenol 
Phenol 
Phenol 
Phthalate 
Phthalate 
Phthalate 
Phthalate 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Oligomer 
Alkane 
Alkane 
Alkane 
Alkane 
Alkane 
Alkane 
Alkane 
Toluene 
Chloroform 
Acetone 
Isopropanol 
(a) 
(b) 
Nature of compounds 
Nature of compounds (x10) 
(x10) 
(x10) (x10) 
(x10)
(x100) 
the lipophilic lipid emulsion and propofol promoted the extraction of large amounts 
of DEHP into the preparations. 
An investigation of DEHP extractability from PVC tubing into a large variety 
of pharmaceutical formulations used for different purposes was carried out by 

510 ORIGIN OF CONTAMINATION 
Haishima and colleagues [75] . The authors divided the IV drugs into fi ve groups 
according to the properties of their active ingredients and additives. Group 1 
included drugs that were practically insoluble or insoluble in water and contained 
additives such as surfactants, oils, glycerine, ethanol, or benzyl alcohol. Group 2 
included drugs equally insoluble in water but soluble in acidic or basic solutions. 
Formulations slightly soluble in water were placed in group 3, and formulations very 
soluble or freely soluble in water were classifi ed in groups 4 and 5. The difference 
between group 4 and 5 was the presence of additives suspected to induce DEHP 
migration in group 4. Tests were carried out by placing the drug formulations into 
PVC tubing of 10 cm in length with 2.13 mm inner diameter for 1 h at room temperature 
under shaking. As shown in Table 35 , products included in group 1 promoted 
the release of large amounts of DEHP, with the exception of insulin and 
dinoprost, probably due to the absence of oleaginous excipients. No signifi cant 
DEHP release was observed by most of the other drugs assigned to groups 2 
through 5, with the exception of human serum albumin and antithrombin III (both 
group 4) and phenytoin solution (group 2), probably due to the presence of propylene 
glycol and ethanol in the formulation. These results confi rm the affi nity of 
DEHP to lipophilic media. 
Due to their utilization modus, many studies conducted with tubing materials are 
essentially kinetic, where tubing length, time of contact, fl ow rate, and temperature 
are important parameters. 
Kambia et al. [76] studied the kinetics of DEHP migration into TPN solutions 
from PVC tubing. They quantifi ed the amount of DEHP leached into two types of 
emulsions 24 h after preparation and storage at 4 ° C. The authors concluded that the 
extraction depended on the lipid content of the formulations and the fl ow rates. 
Figure 12 shows the DEHP concentration time course obtained from bags and 
outlet tubing during 10 – 11 h simulated infusion of TPN after 24 - h storage at 4 ° C. 
To overcome the problem of DEHP leaching into parenteral infusions containing 
lipophilic components, a triple - layered tubing material has been used. This type of 
tubing is made of a PVC outer layer, while its innermost layer, which comes into 
contact with the drug solution, is made of inert PE. In spite of this arrangement, it 
has been shown that, depending on the preparation, DEHP from the PVC outer 
layer may be released in the infusion solution. 
Sautou - Miranda et al. [77] showed that DEHP is rapidly leached from PVC, 
coextruded, and also triple - layered IV tubing by action of an etoposide infusion 
TABLE 34 DEHP Loaded in Infusion Solutions after Exposure to PVC Infusion Lines 
Sample 
Volume 
(mL) 
Number of 
Samples DEHP ( . g/mL) 
Range DEHP 
(. g/mL) 
Amino acid – glucose 140 06 0.31 ± 0.56 0.0 – 1.05 
Lipid emulsion 24 10 422.78 ± 47.39 329.15 – 490.40 
Midazolam 24 03 0.90 ± 0.38 0.55 – 1.30 
Propofol infusion 10 10 654.87 ± 96.49 423.85 – 736.10 
Fentanyl 28.8 10 3.63 ± 0.92 1.95 – 5.05 
Imipenem 8 03 0 . 0.10 – 0.05 
Source : From ref. 74 . 

TABLE 35 DEHP Released from PVC Tubing into Intravenous Drug Preparations after 1 h Contact at Room Temperature 
Active Ingredient Concentration Additives DEHP ( . g/L) SD ( . g/L) 
Group 1 
Cyclosporin 500 
. g/mL Polyoxythene castor oil, 
ethanol 
27,363.9 384.8 
Tacrolymus hydrate 10 
. g/mL Absolute ethanol, 
HCO - 60 4,091.9 31.9 
Propofol 10 mg/mL Soybean oil, 
concentrated 
glycerin, egg yolk lecithin, 
edetate 
19,451.2 852.5 
Flurbiprofen axetil 10 mg/mL Soybean oil, 
egg yolk 
lecithin, concentrated 
glycerin 
17,838.5 821.6 
Vitamins — fat soluble Whole amounts of 
Sorbita was mixed with 
PN - Twin no. 
2 (2.2 L) 
Sodium citrate, 
sodium 
pyrosulfi te, sodium 
thioglycollate, 
HCO - 60, 
benzyl alcohol, polysorbate 
80 
1,157.1 5.1 
Menatetrenone 5 mg/mL Aminoethylsulfonic acid, 
sesame oil, soybean 
lecithin, d - sorbitol, 
concentrated glycerin 
8,457.5 62.9 
Insulin human 40 units/mL Concentrated glycerin, 
m - cresol 
281.6 6.0 
Dinoprost 2 mg/L — 185.8 17.3 
Miconazole 1 mg/L HCO - 60 30,098.3 423.3 
Diazepam 5 mg/L Propylene glycol, 
ethanol, 
benzyl alcohol, sodium 
benzoate, benzoic acid 
2,008.8 257.6 
Prednisolone sodium succinate 10 mg/L Sodium carbonate, 
sodium 
hydrogen phosphate, 
sodium dihydrogen 
phosphate 
915.6 182.3 
511

Active Ingredient Concentration Additives DEHP ( . g/L) SD ( . g/L) 
Prednisolone sodium succinate 1 mg/L Sodium carbonate, 
sodium 
hydrogen phosphate, 
sodium dihydrogen 
phosphate 
407.1 2.4 
Group 2 
Famotidine 20 mg/L 
l - Aspartic acid, 
d - manitol 166.0 0.9 
Droperidol 2.5 mg/mL p - Oxymethyl benzoate, 
p - oxypropyl benzoate 
171.0 0.6 
Droperidol 50 
. g/mL p - Oxymethyl benzoate, 
p - oxypropyl benzoate 
167.4 24.6 
Sivelestat sodium hydrate 1 mg/L 
d - Mannitol, 
sodium 
hydroxide, propylene 
glycol, ethanol 
885.7 10.6 
Phenytoin 50 mg/L 
Methotrexate 0.2 mg/mL Sodium chloride, 
sodium 
hydroxide 
372.8 6.8 
Haloperidol 5 mg/mL Glucose, 
lactic acid, 
sodium 
hydroxide 
50.6 2.5 
Epinephrine 0.25 mg/mL Chorobutanol, 
sodium 
hydrogen sulfi te, 
hydrochloric acid, sodium 
chloride 
290.3 24.6 
Group 3 
Methilergometrine maleate 
0.2 mg/mL — 462.7 4.2 
Vecuronium bromide 2 mg/mL 
d - Mannitol 192.7 1.5 
Panipenem betamipron 5 mg/mL — 237.0 1.2 
Mynocycline hydrochloride 1 mg/mL — 150.0 8.9 
Nicardipine hydrochloride 
0.1 mg/mL 
d - Sorbitol 211.6 24.0 
Bromhexine hydrochloride 2 mg/mL Glucose 174.9 23.7 
Ceftazidime 10 mg/mL Sodium carbonate 301.0 0.5 
Fluconazole 1 mg/mL — 210.5 0.2 
Aspoxicillin 50 mg/mL Sodium chloride 296.7 2.6 
Carbazochrome sodium sulfonate 0.05 mg/mL Sodium hydrogen sulfi te, 
d - sorbitol, 
propylene glycol 
246.1 3.0 
TABLE 35 Continued 
512

Active Ingredient Concentration Additives DEHP ( . g/L) SD ( . g/L) 
Group 4 
Oxytocin 0.01 units/mL Chlorobutanol 423.1 0.8 
Hydroxyzine hydrochloride 0.05 mg/mL Benzyl alcohol 430.8 144.4 
Ranitidine hydrochloride 0.1 mg/mL Phenol 197.9 29.5 
Human immunoglobulin G 50 mg/mL 
d - Sorbitol 243.9 14.3 
Panthenol 250 mg/mL Benzyl alcohol 412.1 18.2 
Human serum albumin 250 mg/mL Sodium 
N - acetyl tryptophan, 
sodium caprylate, sodium 
hydrogen carbonate 
10,080.8 84.1 
Human antithrombin III 
25 units/mL Sodium chloride, 
sodium 
citrate, d - mannitol 
2,008.2 21.8 
Nitroglycerin 0.5 mg/mL 
d - Mannitol 267.6 8.9 
Sulpyrine 2.5 mg/mL Benzyl alcohol 302.8 3.8 
Erythromycin lactobionate 
2.5 mg/mL Benzyl alcohol 92.2 0.7 
Clindamycin phosphate 3 mg/mL Benzyl alcohol 274.9 4.0 
Group 5 
Imipenem cilastatin sodium 5 mg/mL Sodium hydrogen carbonate 205.1 1.6 
5% glucose 50 mg/mL — 284.6 4.8 
Ferric oxide, 
saccharated 
0.4 mg/mL — 244.5 5.5 
Maltose, 
sodium chloride, 
magnesium chloride, potassium 
dihydrogenphosphate, sodium 
acetate 
— — 262.8 5.0 
Atropine sulfate 0.5 mg/mL — 200.7 5.1 
Ampicillin sodium 10 mg/mL — 262.3 6.8 
Aminophyline 0.5 mg/mL Ethylenediamine 301.1 4.0 
Fosfomycin sodium 20 mg/mL Glucose solution 289.6 6.7 
Calcium gluconate 
85 mg/mL — 179.4 4.3 
Cefazolun sodium hydrate 
10 mg/mL — 215.1 0.9 
Amino acids, 
electrolytes 
— Sodium hydrogen sulfi 
te 328.5 5.0 
Suxamethonium chloride 2 mg/mL — 228.6 2.1 
Ioversol 320 mg/mL — 404.0 79.5 
l - Isoprenaline hydrochloride 1 
. g/mL Sodium hydrogen sulfi te, l - 
cysteine hydrochloride 
326.3 8.6 
Source : 
From ref. 
75 . 
513

514 ORIGIN OF CONTAMINATION 
solution. Table 36 presents the concentration of DEHP leached from tubing into 
etoposide infusion solutions as a function of time of exposure for two tubing lengths. 
The results showed that fast and considerable leaching of DEHP occurred, even 
when the PVC component had no direct contact with the solution. DEHP leaching 
was greatest for the tubing made only of PVC, but it also occurred with the coextruded 
(PE + PVC) and triple - layered tubes, despite their claims to prevent such 
leaching. The authors concluded that either DEHP was present in the internal PE 
layer or it migrated rapidly through the other polymer layers. 
The kinetics of DEHP extraction from triple - layered tubing was studied by these 
authors by varying tube length, fl ow rate, and drug concentration. Figure 13 shows 
FIGURE 12 Comparative kinetics of DEHP leachability during simulated infusion of TPN 
[76] . (a) Kinetics of DEHP leachability during simulated infusion of TPN 24 hours after 
reconstitution of the preparation (n = 2 bag). Formula 1: infusion volume of 2200 mL (fl ow 
rate 177 mL/h, lipid concentration 3.85%). (b) Kinetics of DEHP leachability during simulated 
infusion of TPN 24 hours after reconstitution of the preparation (n = 2 bag). Formula 
2: infusion volume of 650 mL (fl ow rate 46 mL/h, lipid concentration 1.85%). 
Concentration (ng/mL) Concentration (ng/mL) 
1200 
(a) 
(b) 
800 
400
0 
0 0.25 1 2 4 8 11 
0 0.25 1 2 4 8 11 
Time (h) 
Time (h) 
In infusion bag From outlet tubing 
0 
300 
600 
900 
1200 
1500 
1800

EXOGENOUS IMPURITIES 515 
the amount of DEHP leached with the volume of solution infused. As could be 
expected, higher drug concentrations and tubing lengths increased the amount of 
DEHP released. Similarly, the slower the fl ow rate was, the higher the migration of 
the additive . 
As mentioned under plastic containers (Section 6.1.3.2 ), DEHP is not the only 
additive that can be leached from plastic materials. Jenke et al. [78] studied the 
composition of extractables from plastic tubing used in pharmaceutical production 
facilities. Eight tubing materials made of silicone and neoprene rubber were characterized 
by their extractable substances. The authors investigated not only organic 
but also inorganic extractables. Extracts were obtained by static experiments using 
both water and ethanol. The tubing was cut and autoclaved in water at 121 ° C for 
1 h, or lengths of tubing were fi lled with 100% ethanol and kept at 55 ° C for 24 h. 
Metal determination was carried out by inductively coupled plasma – atomic emission 
spectroscopy (ICP - AES), which included 29 elements. Besides the elements 
listed in Table 37 , Be, Co, Cr, Cd, Se, V, Ge, Pb, and Bi were also measured but not 
detected since they either presented a concentration not different from the blank 
sample or below the lowest quantity determinable (LQD). It is important to mention 
that the ICP - AES is not a very sensitive technique. LQD values for elements 
ranging between 0.01 and 0.1 mg/L are relatively high levels for contaminants such 
as Cd or Be. This means that these elements could possibly be present in the extracts 
but not detected using this technique. In general, all tested tubing materials presented 
metals that were extracted either by water or by ethanol, although water, 
due to its greater capability to interact with metallic ions, promoted higher extraction 
yields. All tubing materials contained extractable Ca, Mg, Zn, and B. Elevated 
TABLE 36 Concentration of DEHP Leached from Tubing into Etoposide Infusion Solutions as a 
Function of the Time of Exposure 
Tubing Length 
and Type 
Mean ± SD DEHP Concentration ( . g/mL) 
0 h 1 h 2 h 3 h 4 h 6 h 
25 cm 
PE b < 0.5 < 0.5 < 0.5 < 0.5 < 0.5 < 0.5 
PVC b,c < 0.5 22.59 ± 4.21 35.05 ± 2.51 49.88 ± 4.78 58.57 ± 1.03 73.51 ± 3.51 
Triple layered d < 0.5 18.92 ± 2.23 31.49 ± 3.78 44.19 ± 5.38 54.83 ± 4.28 61.99 ± 1.25 
PVC e < 0.5 19.93 ± 1.90 33.01 ± 1.87 46.82 ± 1.98 55.46 ± 3.15 66.10 ± 1.23 
Coextruded f < 0.5 18.64 ± 1.38 28.77 ± 2.33 39.37 ± 3.47 48.13 ± 2.21 54.60 ± 2.53 
50 cm 
PE b < 0.5 < 0.5 < 0.5 < 0.5 < 0.5 < 0.5 
PVC b,c < 0.5 51.67 ± 3.63 82.00 ± 2.65 112.69 ± 4.89 131.67 ± 3.65 155.22 ± 2.35 
Triple layered d < 0.5 39.85 ± 0.49 59.73 ± 3.04 81.51 ± 1.57 85.01 ± 2.17 98.72 ± 2.33 
PVC e < 0.5 45.38 ± 2.08 73.27 ± 0.96 94.65 ± 0.88 117.53 ± 3.43 143.41 ± 11.39 
Coextruded f < 0.5 39.98 ± 0.74 59.61 ± 1.49 78.03 ± 0.52 82.47 ± 3.56 107.83 ± 9.68 
Source : From ref. 77 . 
b Manufactured by Vycon, Ecouan, France. 
c Tubing length 80 cm. 
d Consisted of an outer layer of PVC, a middle layer of EVA, and an inner layer of PE. 
e Manufactured by Cair, Civrieux d ’ Azergues, France. 
f PVC and PE tubing. 
DEHP: diethyhexylphthalate; PE: polyethylene; PVC: polyvinyl chloride. 

516 ORIGIN OF CONTAMINATION 
FIGURE 13 ( a ) Cumulative amount of DEHP leached from 50 - and 80 - cm PVC - only 
tubing after infusion of 0.4 - mg/mL etoposide solution using various fl ow rates. ( b ) Cumulative 
amount of DEHP leached from 50 - and 80 - cm triple - layered tubing after infusion of 0.2 and 
0.4 mg/mL etoposide solution with a fl ow rate of 30 mL/h. ( c ) Amount of DEHP leached from 
50 - and 80 - cm triple - layered tubing after infusion of 0.4 - mg/mL etoposide solution with a 
fl ow rate of 30 mL/h [77] . 
(a) 
(b) 
(c) 
concentrations of other metals were also extracted from samples 3 and 5, probably 
due to the tubing ’ s reinforced embedded metallic wire. 
The authors also measured silicon and both organic and inorganic carbon. While 
inorganic carbon is indicative of the presence of carbonates (earth carbonates as 
additives), organic carbon is related to organic extractables. As could be expected, 

TABLE 37 Levels of Total Carbon (Organic and Inorganic), and Silicone and Metals Extracted of the Tubing Materials with Water or Ethanol 
by Heating at 121 ° C for 1 h 
Tubing Material 
Total Inorganic 
Carbon ( . g/g) 
Total Organic 
Carbon ( . g/g) 
Si ( . g/g) 
Ethanolic 
Extract 
Si ( . g/g) 
Aqueous 
Extract 
Metal Concentration in 
Aqueous Extract 
< 0.5 
. g/mL 0.5 – 1.0 
. g/mL 
> 1 
. g/mL 
Silicone 1 0 14.0 765 101 B, 
Mg, 
Zn 
Silicone 2 0.9 38.1 1360 < 0.2 Ca B, 
Mg, 
Mn 
Mo, Ti, Zr. 
Sn, Zn, Sb, 
Li, Ag, Ni 
Silicone 3 (embedded wire) 2.3 250 1860 66.0 Ca, 
Ba, 
Mn, 
Mg, 
Al, Cu 
— 
B
, 
Fe
, 
Zn
, 
Silicone 4 0.2 34.0 1120 130 Ca, 
B, 
Fe, 
Mg, 
Al, 
Zn 
— 
— 
Silicone 5 (embedded wire) 0 49.9 1300 87.3 Ca, 
B, 
Mn, 
Fe, 
Mg, 
Zn, Sb 
— 
— 
Silicone 6 0 46.9 739 < 0.2 Mn, Fe, Mg, Zn, 
Cu, Sb, Ni 
B Ca 
Santoprene 1 10.1 16.6 — 
a < 0.2 Ca, 
Ba, 
B, 
Mg, 
Zn, 
— — 
Santoprene 2 4.6 13.5 Not tested 
0 Ca, 
Mo, 
Ti, 
Zr, 
Mn, Zn, Li, Ag, 
Ni 
— 
Mg 
Source : 
From ref. 
78 . 
a Tube disintegration. 
517

518 ORIGIN OF CONTAMINATION 
Santoprene extracts contained higher concentrations of inorganic carbon and little, 
if any, extractable silicon. On the other hand, the amount of silicone in extracts of 
silicone polymers was very high. From Table 37 , it is possible to see that these materials, 
though all made of silicone, are very different each other. While about 100 . g/g 
Si was found in the aqueous extract from sample 1, no measurable Si was detected 
in the same extracts from samples 2 and 6. As could be expected, silicone extractables 
were much higher in ethanolic extracts, being that 739 . g/g Si from sample 6 
was the lowest amount leached. 
The analysis of aqueous and ethanolic extracts by techniques such as liquid 
chromatography – mass spectrometry (LC – MS) and CG – MS showed that all silicon 
materials contained essentially the same peaks, that is, the same compounds were 
extracted. The distribution, however, varied from material to material. The primary 
organic extractables from silicone tubing were homologous series of silicone oligomers 
with the structural formula [(CH 3 ) 2 SiO] n . Direct matches with an MS spectral 
library allowed the identifi cation of oligomers with n = 5 to n = 25. However, while 
in sample 3, the peaks were at the lower range of n , the oligomer distribution for 
sample 2 was shifted toward higher molecular weights, which are less soluble and 
therefore less extractable. Figure 14 illustrates the differences in the ethanolic 
extracts of materials 2 and 3. 
Results of Santoprene extracts are quite different from those of silicone. GC – MS 
chromatograms for water extracts of Santoprene tubing show peaks associated with 
C8 and C9 acids, phthalates, and other organic substances. Figure 15 shows the 
GC – MS chromatograms of static ethanol extracts of Santoprene tubing materials, 
and Table 38 lists the compounds identifi ed in Figure 14 , based on matches of MS 
spectrum peaks. 
6.1.3.4 Particulate Matter 
Particulate contamination can be classifi ed as intrinsic or extrinsic depending on its 
origin. Intrinsic contaminants arise from the manufacturing, packaging, transport, 
and storage of solutions; extrinsic contaminants are introduced or generated during 
drug reconstitution and administration to the patient. 
The presence of particulate matter, while undesirable in any pharmaceutical 
product, is truly a problem in intravenous and ophthalmic preparations. Most 
common particulates in intravenous preparations are glass fragments, from the 
opening of glass ampoules, particles from rubber stoppers and intravenous equipment, 
and particles from plastic syringes, providing that the manipulation of such 
solutions is carried out in a controlled clean area. Air in a controlled area, in the 
immediate proximity of exposed sterilized containers/closures and fi lling/closing 
operations is appropriate when it has a particle count of no more than 3520 in a 
size range of 0.5 . m per cubic meter. This corresponds to a Class 100 air cleanliness 
level and is the regular environmental condition for manipulating sterile 
preparations. 
Pharmacopoeial compendia prescribe the examination of particulate contamination 
for injections and infusion solutions and consider particulate contamination as 
the presence of extraneous mobile undissolved particles, other than gas bubbles, 
unintentionally present in the solution. They limit the number of particles according 
to their size and the volume of the preparation. 

EXOGENOUS IMPURITIES 519 
Generally, particulate matter is eliminated by fi ltration, but the word elimination 
or even the concept of being substantially free of , as appear in some compendia, are 
levels practically impossible to reach. In fact, complying with pharmacopeia prescription 
is limited to a removal of particles above a certain size, since the elimination 
of particles becomes increasingly more diffi cult the smaller the particle is. 
Elimination diffi culty seemingly increases in an exponential fashion as the size 
decreases. 
Pharmacopeial count limits for particulates in parenteral solutions is given in 
Table 39 . The limit depends on the method used for the determination and also on 
the volume of the sample. Two different procedures for the determination are generally 
proposed: light obscuration particle count test (LO) and microscopic particle 
count test (M), since neither is applicable to all kinds of samples. 
It is possible to display the relationship between size and number of particles 
graphically. Groves [79] reported a logarithmic relationship between the number of 
FIGURE 14 GC – MS chromatograms of the static ethanol extracts of silicone tubing materials 
(underivatised). The chromatograms from all the silicone materials were similar (same 
peaks but different relative sizes), and thus chromatograms from two different silicones are 
shown. A majority of the peaks are attributable to silicone oligomers. Peaks associated with 
cyclic oligomers are identifi ed by the number of repeating units, n : e.g., [CH 3 ) 2 SiO] n . Peaks 
denoted with * produced exact compound matches versus a library of mass spectra, while 
peaks denoted by # produced a library match to the right compound class (cyclic oligomer) 
but wrong specifi c oligomer. Small peaks at 7.95, 10.32, and 11.93 min were linked to 5 – 7 
member linear silicone oligomers. IS = internal standard (dimethyl phthalate) [78] . 

520 ORIGIN OF CONTAMINATION 
particles and their diameter in a solution. Thus, using the limits prescribed by BP, 
the graphical representation would be as shown in Figure 16 . 
The author, however, considers that if the distribution number/size of particles 
follows a random pattern, graphically represented in Figure 16 , it cannot be considered 
contamination. On the other hand, if the solution no longer contains random 
identities but rather is dominated by particles such as starch grains from a stopper 
composition, glass shards, from a delaminating container, skin fragments, or clothing 
fi bers, a nonrandom behavior would occur and be defi ned as “ contamination. ” In 
other words, only positive deviation of the log - log particle size distribution curve, 
constructed using, for example, pharmacopoeial limits, would be considered contamination. 
Figure 17 graphically displays the particulate matter found in 5% dextrose 
solutions. The smooth curve represents the BP specifi cation and the two others 
represent the samples. One product is within the specifi cation (2G) and the other 
is not (2AL). According to the author, while in the 2G sample carbon black particles 
with sizes between 1 and 3 . m were identifi ed, in the 2AL sample, besides carbon 
black (1 – 5 . m), some other particles such as starch grains, lacquer fl akes, and rubber 
were found [80] . 
FIGURE 15 GC – MS chromatograms of the static ethanol extracts of Santoprene tubing 
materials (underivatized). The chromatograms from these two Santoprene materials were 
quite different from those of the silicone materials (Figure 14 ). IS = internal standard 
(dimethyl phthalate). See Table 39 for the tentative peak identifi cations. Only those peaks 
with recorded spectral library matches are noted for each sample, although retention times 
and patterns may suggest some additional peak identifi cations [78] . 
Abundance 
Abundance 
750000 
700000 
650000 
600000 
550000 
500000 
450000 
400000 
350000 
300000 
250000 
200000 
150000 
100000 
50000 
Material 8 
Material 7 
IS 
IS 
3
3 
5 
9
10
13 
24 – 26 
27 
27 
Time 
1000000 
900000 
800000 
700000 
600000 
500000 
400000 
300000 
200000 
100000 
Time 
6 8 10 12 14 16 18 20 22 24 26 28 30 
6 8 10 12 14 16 18 
18 13 
20 
20 
22 24 26 28 30

EXOGENOUS IMPURITIES 521 
TABLE 38 Peak Identifi cation in Ethanol Extracts of Santoprene Tubing Samples 7 
and 8 
Peak Number Tentative Compound Identifi cation Present in Material 
3 2,4 - di - t - Butylphenol 7, 8 
5 4 - (1,1,3,3 - Tetramethylbutyl)phenol 8 
6 Isomer of octyl phenol (TMS) 8 
7 Hexadecane 7 
8 Isomer of octyl phenol (TMS) 8 
9 4 - Methyl - 6 - tert - octyl phenol 8 
10 Isomer of decyl phenol 7, 8 
11 Isomer of nonyl phenol (TMS) 8 
12 Heptadecane 7 
13 Isomer of undecyl phenol 7, 8 
14 Octadecane 7 
15 Isomer of decyl phenol (TMS) 8 
16 Nonadecane 7 
17 Isomer of undecyl phenol (TMS) 7, 8 
18 Hexadecanoic acid, ethyl ester 7 
19 Cyclohexadecane, heneicosane 7 
20 Octadecanoic acid, ethyl ester 7 
21 Docosane 7 
22 Tetratriacontane, 9 - methyl - nonadecane 7, 8 
23 Tetracosane 7 
24 Pentacosane 7, 8 
25 Hexacosane, nonadecane 7, 8 
26 Heptacosane 8 
27 Irganox 1076 7, 8 
Source : From ref. 78 . 
TABLE 39 Limits for Particulate Contamination in Infusion Solutions Established by 
Pharmacopeial Compendia 
Compendia Volume 
Particle Size 
(. m) LO Limit M Limit 
BP SVP . 10 6000/container 3000/container 
. 25 600/container 300/container 
LVP . 10 25/mL 12/mL 
. 25 3/mL 2/mL 
USP SVP . 10 6000/container — 
. 25 600/container — 
LVP . 10 25/mL 12/mL 
. 25 3/mL 2/mL 
IP SVP, LVP — One or more particles in 
more than one 
container 
— 
Note : The limits are related to the method used for the determination: LO = light obscuration particle 
count test, M = microscopic particle count test, IP = International Pharmacopeia. 

522 ORIGIN OF CONTAMINATION 
FIGURE 16 Coordinate plot of log (number of particles oversize per milliliter) versus log 
(particle diameter) for BP specifi cations. 
Number of particles oversize (mL–1) 
Diameter, mm 
1500 
1000 
500 
2 4 6 
FIGURE 17 Cumulative particle size distribution in 5% dextrose solutions for injection: 
( ) product 2AL packed in 500 - mL glass containers with lacquer - coated rubber stopper; 
( ) product 2G packed in 500 - mL plastic bags; (smooth curve), BP specifi cation [80] . 
Number of particles oversize (mL–1) 
Diameter, mm 
1500 
1000 
500 
2 4 6 
.0

EXOGENOUS IMPURITIES 523 
Techniques for counting particulate matter allow a discrimination of the particles 
by sizes even lower than the pharmacopoeial limits, and though not prescribed, it is 
also possible to characterize particulate matter by its composition. Table 41 displays 
the amount of particles found in infusion solutions classifi ed according to size. 
Since the results correspond to different studies, they are displayed in different 
fashions. Nevertheless, all solutions present particulate matter as contaminants, and 
it is possible to observe the exponential behavior of the relationship between the 
number and size of the particles. 
Foroni et al. [81] measured and characterized inert particles both in individual 
solutions and in the fi nal ternary mixture prepared by the sterile transfer technique. 
The results are presented in Table 40 and also in Figure 18 . The difference between 
the number of particles greater than 2 and 5 . m is higher than 50. The histograms 
in Figure 18 show the number of particles of the individual components (Figure 18 a ) 
and the number of particles after transferring the components into an EVA bag 
(Figure 18 b ). The results show that the contamination of 10% KCl, 30% dextrose, 
amino acids, and lipid emulsion is signifi cantly higher than of the other components 
(Figure 18 a ). The results also show that the concentration of particles in the components 
packaged in glass ampoules, KCl and disodium phosphate is also higher 
than in the other components. The identifi cation of contaminants was also carried 
out, where particles composed of silica, aluminum, and sodium were identifi ed in a 
KCl solution, and particles containing sulfur and silicate were identifi ed in all bags. 
The authors consider the presence of sulfur to be related to rubber particles and 
attribute the presence of silicate to contact with talcum from gloves used by manipulators 
during the manufacturing of the bags. 
Kamiya et al. [83] evaluated particulate contamination in 199 samples of admixed 
and un - admixed parenteral nutrition solution bags from 10 hospitals in Japan. Seven 
samples were used as controls since they had not been mixed with ampoules or vials 
(un - admixed samples). Size and number of particles were measured using a particle 
counter, and the identifi cation of elements was carried out by scanning electron 
microscopy coupled to energy dispersion spectroscopy. The authors collected the 
residual volume of the samples (10 – 60 mL) after their usage. The results are presented 
in Table 40 . 
Figure 19 shows the scanning electron micrograph of two types of particles 
on the fi lter after fi ltration (0.22 . m membrane fi lter) of 50 mL of an admixture 
containing 1700 mL glucose/electrolytes/amino acids plus one vial of vitamin complex 
(Maltamin), two glass ampoules of 1 mL panthol (Pantol), two glass ampoules of 
20 mL sodium chloride (Conclyte - Na), two glass ampoules of 2 mL metoclopramide 
(Primperan), and two glass ampoules of 10 mL potassium l - aspartate (Aspara K). 
Figure 19 also shows the identifi cation of these particles. Their composition suggests 
that they are particles of glass (Figure 19 a ) and particles of rubber (Figure 19 b ). 
A similar study was carried out by Ball et al. [82] in New Zealand and the United 
Kingdom. The authors analyzed 20 samples of adult and 20 samples of the pediatric 
PN admixtures, collecting the fi rst and second fractions drawn from the infusion 
sets. The number of particles greater than 5 . m was 50 times higher than the number 
of particles greater than 40 . m in all solutions (Table 40 ). The analysis of the particulate 
matter allowed for a characterization of particles such as rubber and glass 
fragments (Table 41 ). 

TABLE 40 Mean (Range) of Particles Found in Mixed and Unadmixed Parenteral Nutrition Solutions Classifi ed According to Their Size 
Sample 
Volume 
(mL) 
n 
Particle Size 
Reference 
> 1.3 
. m 
> 5 
. m 
> 10 
. m 
> 25 
. m 
> 50 
. m 
Mean Range Mean Range Mean Range Mean Range Mean Range 
30 % Dextrose 25 3 452 
± 127 n.d. 
a 
— 
— 
— 
— 
— 
— 
— 
81 
50 % Dextrose 25 3 2831 
± 278 — 
± 90 — 
— 
— 
— 
— 
— 
81 
Amino acids 25 3 3715 
± 184 — 
± 58 — 
— 
— 
— 
— 
— 
81 
All - in - one solution (adult) 
b 
1 20 3.47 
± 1.24 1.40 
± 0.73 0.96 
± 0.45 1.06 
± 0.39 — 
— 
82 
Two - in - one (pediatric) 
b 
1 20 7.59 
± 2.56 11.77 
± 7.38 1.43 
± 1.09 0.38 
± 0.26 — 
— 
82 
Lipid emulsion (syringe 
packed) 
1 20 16.72 
± 10.85 8.98 
± 4.63 1.27 
± 1.23 0.76 
± 0.34 — 
— 
82 
Unadmixed samples 10 – 60 7 62.7 8 – 146 1.70 0 – 4 0.4 0 – 2 — — 0 
83 
Admixed samples 
c 
10 – 60 192 960.9 30 – 9539 42.8 0 – 587 6.4 1 – 146 — — 0.09 0 – 1 83 
Solutions mixed using 1 – 3 
glass ampoules 
c 
10 – 60 29 862.1 30 – 5707 31.3 0 – 176 4.4 0 – 24 — — 0.1 0 – 1 83 
Solutions mixed during 
4 – 13 glass ampoules 
c 
10 – 60 63 1163.4 142 – 9539 66.2 4 – 587 10.6 1 – 146 — — 0.08 0 – 1 83 
a n.d. 
= not detected. 
b First milliliters collected. 
c residual volume collected. 
524

EXOGENOUS IMPURITIES 525 
FIGURE 18 ( a ) Total number of particles in each component of the admixture analyzed 
by fi ltration - observation method. ( b ) Number of particles per milliliter of each component 
analyzed by fi ltration - observation method. 
EVA bag 
KCL 10% 
Phosphate 
Electrolytes 
Trace elements 
30% dextrose 
50% dextrose 
Aminoacids 
Lipid emulsion 
KCL 10% 
Phosphate 
Electrolytes 
Trace elements 
30% dextrose 
50% dextrose 
Aminoacids 
Lipid emulsion 
EVA bag 
KCL 10% 
Phosphate 
Electrolytes 
Trace elements 
30% dextrose 
50% dextrose 
Aminoacids 
Lipid emulsion 
KCL 10% 
Phosphate 
Electrolytes 
Trace elements 
30% dextrose 
50% dextrose 
Aminoacids 
Lipid emulsion 
Total number of particles . 5 .m Total number of particles . 10 .m 
Number of particles . 10 mm/mL of component Number of particles . 5 mm/mL of component 
Components 
Components Components 
Components 
0 10000 20000 30000 0 10000 20000 30000 
0 100 200 300 400 500 0 100 200 300 400 500 
(a) 
(b) 
FIGURE 19 Identifi cation of two types of particles by scanning electron microscopy 
coupled to energy dispersion spectroscopy: ( a ) suggests glass particles; ( b ) suggests rubber 
particles [83] . 
(a) (b) 
Counts 
Counts

526 ORIGIN OF CONTAMINATION 
Roseman et al. [84] used scanning electron microscopy to analyze fl akes found 
in solutions stored in glass ampoules (type I glass ampoules). They observed that 
all fl akes had similar characteristics (colorless, platelike, < 1 . m in thickness), but 
their sizes ranged from a few micrometers to a 100 . m in length. The elemental 
analysis of four fl akes gave the results presented in Table 42 . Because the fl akes 
presented the same elements of glass, the authors broke a glass ampoule and analyzed 
the fragments (results also presented in Table 42 ). Since the composition of 
the fl akes and glass fragments were similar to each other, the authors believed that 
the fl akes originated in the ampoule due to an attack on the glass surface by some 
chemical substance. It has been shown that, besides strong alkali or hydrofl uoric 
acid, other substances such as EDTA and other organic acids such as citric acid, as 
well as phosphate salts are able to attack glass (see Section 9.1.3.2). The existence 
of such particulates in glass ampoules was attributed to a delamination process that 
could occur just above the bottom of the ampoule, a region of thermal stress because 
of the intense heat used to form the bottom of the container. 
The major problem associated with particulate contamination in infusion fl uids 
is not related to the composition itself (since they are mostly pieces of the container 
and therefore innocuous elements) but to the potential of each particle to cause 
TABLE 41 Elements in Particulate Matter Found in PN Solutions 
Most Frequent 
Elements 
Frequency of 
Appearance Probable Origin Reference 
Oxygen 1 Glass; rubber 82, 83 
Silicon 1 Glass; rubber 81, 82, 83 
Carbon 3 Rubber 82, 83 
Aluminum 1 Glass 81, 82, 83 
Magnesium 2 Talc 81, 82 
Sodium 2 Glass, rubber 81, 83 
Chlorine 3 Rubber 83 
Potassium 3 Glass 83 
Note : Frequency of appearance: 1 = high; 2 = medium; 3 = low. 
TABLE 42 Elemental Analysis of Flakes Found in Solution Stored Glass Ampoule and 
of Glass Fragments Resulting of a Broken Glass Ampoule 
Element 
Composition (%) 
Flake 1 Flake 2 Flake 3 Flake 4 Broken glass a 
Silicon 25 – 30 35 – 45 30 – 35 25 – 30 30 – 40 
Alumnum 10 – 12 3 – 5 8 – 10 10 – 11 2 – 5 
Potassium 4 – 6 n.d. b 3 – 5 3 – 4 n.d. 
Calcium 1 – 2 n.d. n.d. 1 – 2 n.d. 
Boron 3 – 8 2 – 6 3 – 8 Present 4 – 10 
Sodium 1 – 3 1 – 2 1 – 3 1 – 2 4 – 6 
Source : From ref. 84 . 
Note : All ampoules were of type I glass. 
a Mean of four samples; n.d. = not detected. 

adverse reactions. The probability of an adverse reaction occurring is proportional 
to the number (and size) of particles introduced in the circulatory system. Turco et 
al. [85] suggest that responses resulting from the infusion of particles include physical 
occlusion, infl ammatory responses, neoplastic responses, and antigenic responses. 
Animal studies have shown that the distribution of particles in tissues is related to 
their diameter. Particles larger than 8 . m are trapped in the lung capillaries, 3 – 6 - . m 
particles are lodged in the spleen and hepatic lymph nodes, and 1 - . m particles in 
the liver [86] . 
6.1.4 CONCLUDING REMARKS 
The safety of drug therapy is closely related to the quality of drugs. All steps in drug 
processing may contribute to increasing the presence of foreign species in the fi nal 
product. The requirements with respect to active ingredients, excipients, residual 
solvents, containers, and closures, although set up guidelines, are not a guarantee of 
products free of contaminants. 
Although pharmacopeial biological reactivity tests for containers are good indicators 
of the toxicity of extractables, there are species that do not cause an acute 
toxic reaction but rather a chronic reaction, as is the case of phthalates (DEHP) 
and metallic species such as aluminum. Moreover, there are species that are extractable 
from packaging materials only by action of formulation constituents and therefore 
are not present in the extracts obtained by conventional pharmacopeial 
compendia tests. 
The degradation of formulations due to lipid and vitamins peroxidation is a 
contamination problem not foreseen in pharmacopeial compendia since it is the 
exposure of the solutions to air and ambient light that induce peroxide 
generation. 
Particulate contamination has been found in PN solutions and other intravenous 
drugs and fl uids. Administration of particles through infusion solutions can result 
in adverse effects. The probability of these effects to occur increases proportionally 
with the amount of fl uid administrated. 
The presence of contaminants, though undesirable in all kind of drug formulations, 
is really critical in that intent for parenteral administration. Patients who 
require intensive or prolonged parenteral therapy, the immunocompromised, and 
neonates and infants might have increased susceptibility to the detrimental effect 
of contaminants. 
REFERENCES 
1. U.S. Pharmacopeia ( 2005 ), USP 27, U.S. Pharmacopeial Convention, Rockville, MD. 
2. World Health Organization (WHO) ( 2004 ), Guidelines for Drinking - water Quality , Vol. 
1 : Recommendations , WHO , Geneva . 
3. Water programs, Environmental Protection Agency, National Interim Primary Drinking 
Water Regulations, Code of Federal Regulations , Part 141, ( 1985 ). 
4. ASTM International ( 2005 ), ASTM Book of Standards , Vol. 11.02: Water and Environmental 
Technology: Water (II) , ASTM International, West Conshohocken, PA. 
REFERENCES 527

528 ORIGIN OF CONTAMINATION 
5. BP ( 2003 ), British Pharmacopoeia Commission, London. 
6. Deutsche Arzneibuch (DAB) ( 2005 ), Deutschen Apotheker Verlag, Stuttgart. 
7. European Pharmacopoeia (Ph.Eur.) ( 2004 ), 4th ed., European Directorate for the Quality 
of Medicines (EDQM), Ph.Eur., Strasbourg. 
8. The International Pharmacopoeia (IP) ( 2003 ), 3rd ed., World Health Organization, 
Geneva. 
9. K u ster , F. W. , and Thiel , A. ( 1985 ), Rechentafeln f u r die Chemische Analytik , 103 Aufl age, 
Walter de Gruyter, Berlin, p. 258 . 
10. Koo , W. W. , Kaplan , L. A. , Horn , J. , Tsang , R. , and Steichen , J. ( 1986 ), Aluminum in parenteral 
nutrition solutions — sources and possible alternatives , J. Parenteral Enteral Nutr. , 
10 , 591 – 595 . 
11. Recknagel , S. , Br a tter , P. , Chrissafi dou , A. , Gramm , H. - J. , Kotwas , J. , and R o sick , U. ( 1994 ), 
Parenteral aluminum loading in critical care medicine. Part I: Aluminum content of infusion 
solutions and solutions for parenteral nutrition , Infusionsther. Transfusionsmed., 21, 
266 – 273 . 
12. Bohrer , D. , do Nascimento , P. C. , Becker , E. , de Carvalho L. M. , and Dessuy , M. ( 2005 ), 
Arsenic species in solutions for parenteral nutrition , J. Parenteral Enteral Nutr. , 29 , 1 – 7 . 
13. PluhatorMurton , M. M. , Fedorak , R. N. , Audette , P. J. , Marriage , B. J. , and Yatscoff , R. W. 
( 1996 ), Extent of trace - element contamination from simulated compounding of total 
parenteral nutrient solutions , Am. J. Health - Syst. Pharm. , 53 , 2299 – 2303 . 
14. Bohrer , D. , Bortoluzzi , F. , Nascimento , P. C. , Carvalho , L. M. , and Ramirez , A. G. ( 2008 ), 
Silicate release from glass for pharmaceutical preparations, Int. J. Pharm. , in press. 
15. Bohrer , D. , do Nascimento , P. C. , Binotto , R. , Becker , E. , and Pomblum , S. G. ( 2002 ), 
Contribution of the raw material to the aluminum contamination in parenterals , J. Parenteral 
Enteral Nutr. , 26 , 382 – 388 . 
16. Bohrer , D. , Becker , E. , do Nascimento , P. C. , de Carvalho L. M. , and Marques , M. S. ( 2006 ), 
Arsenic release from glass containers by action of intravenous nutrition formulation 
constituents , Int. J. Pharm. , 315 , 24 – 29 . 
17. Thompson , J. E. , and Davidow , L. ( 2003 ), A Practical Guide to Contemporary Pharmacy 
Practice , Lippincott Williams & Wilkins , Philadelphia . 
18. Wade , A. , and Weller , P. J. ( 1994 ), Handbook of Excipients , 2nd ed., Pharmaceutical Press , 
London . 
19. Neuzil , J. , Darlow , B. A. , Inder , T. E. , Sluis , K. B. , Winterbourn , C. C. , and Stocker , R. 
( 1995 ), Oxidation of parenteral lipid emulsion by ambient and phototherapy lights: 
Potential toxicity of routine parenteral feeding , J. Pediatr. , 126 , 785 – 790 . 
20. Steger , P. J. K. , and M u hlebach , S. F. ( 1997 ), In vitro oxidation of IV lipid emulsions in 
different All - in - One admixture bags assessed by an iodometric assay and gas - liquid 
chromatography , Nutrition , 13 , 133 – 140 . 
21. Murray , R. K. , Granner , D. K. , Mayes , P. A. , and Rodwe , V. W. ( 2000 ), Harper ’ s Biochemistry 
, 25th ed., McGraw - Hill , New York , p. 169 . 
22. Helbock , H. J. , Motchnik , P. A. , and Ames , B. N. ( 1993 ), Toxic hydroperoxides in intravenous 
lipid emulsions used in preterm infants , Pediatrics , 91 , 83 – 87 . 
23. Picaud , J. C. , Steghens , J. P. , Auxenfans , C. , Barbieux , A. , Laborie , S. , and Claris , O. ( 2004 ), 
Lipid peroxidation assessment by malondialdehyde measurement in parenteral nutrition 
solutions for newborn infants: A pilot study , Acta Paediatr. , 93 , 241 – 245 . 
24. Pironi , L. , Guidetti , M. , Zolezzi , C. , Fasano , M. C. , Paganelli , F. , Merli , C. , Bersani , G. , 
Pizzoferrato , A. , and Miglioni , M. ( 2003 ), Peroxidation potential of lipid emulsions after 
compounding in all - in - one solutions , Nutrition , 19 , 784 – 788 . 

25. Steger , P. J. K. , and M u hlebach , S. F. ( 1998 ), Lipid peroxidatoin of IV lipid emulsions in 
TPN bags: The infl uence of tocopherols , Nutrition , 14 , 179 – 185 . 
26. Burton , G. W. , and Ingold , K. U. ( 1986 ), Vitamin E as an in vitro and in vivo antioxidant , 
Ann. NY Acad. Sci. , 570 , 7 – 22 . 
27. Bowry , V. W. , Ingold , K. U. , and Stocker , R. ( 1992 ), Vitamin E in human low - density lipoprotein. 
When and how this antioxidant becomes a pro - oxidant , Biochem. J. , 288 , 
341 – 344 . 
28. Bowry , V. W. , and Stocker , R. ( 1993 ), Tocopherol - mediated peroxidation. The pro - oxidant 
effect of Vitamin E on the radical - initiated oxidation of human low - density lipoprotein , 
J. Am. Chem. Soc. , 115 , 6029 – 6040 . 
29. Steger , P. J. K. , and M u hlebach , S. F. ( 2000 ), Lipid peroxidatoin of intravenous lipid emulsions 
and all - in - one admixtures in total parenteral nutrition bags: The infl uence of trace 
elements , J. Parenteral Enteral Nutr. , 24 , 37 – 41 . 
30. Laborie , S. , Lavoie , J - C. , Rouleau , T. , and Chessex , P. ( 2002 ), Multivitamin solutions for 
enteral supplementation: A source of peroxides , Nutrition , 18 , 470 – 473 . 
31. Laborie , S. , Lavoie , J - C. , Pineaut , M. , and Chessex , P. ( 2000 ), Contribution of multivitamins, 
air and light in the generation of peroxides in adult and neonatal parenteral nutrition 
solutions , Ann. Pharmacother. , 34 , 440 – 445 . 
32. Helbock , H. J. , Motchnick , P. A. , and Ames , B. N. ( 1993 ), Toxic hydroperoxides in intravenous 
lipid emulsions used in preterm infants , Pediatrics , 91 , 83 – 88 . 
33. Lavoie , J - C. , B e langer , S. , Spalinger , M. , and Chessex , P. ( 1997 ), Admixture of a multivitamin 
preparation to parenteral nutrition: The major contributor to in vitro generation 
of peroxides , Pediatrics , 99 , 61 – 70 . 
34. Lavoie , J. C. , Chessex , P. , Rouleau , T. , Migneault , D. , and Comte , B. ( 2004 ), Light - 
induced byproducts of vitamin C in multivitamin solutions , Clin. Chem. , 50 , 135 – 
140 . 
35. Knafo , L. , Chessex , P. , Rouleau , T. , and Lavoie , J - C. ( 2005 ), Association between hydrogen 
peroxide - dependent byproducts of ascorbic acid and increased hepatic acetyl - CoA carboxylase 
activity , Clin. Chem. , 51 , 1462 – 1471 . 
36. Balet , A. , Cardona , D. , Jane , S. , Molins - Pujol , A. M. , Sanchez Quesada , J. L. , Gich , I. , and 
Mangues , M. A. ( 2004 ), Effects of multilayered bags vs ethylvinyl - acetate bags on oxidation 
of parenteral nutrition , J. Parenteral Enteral Nutr. , 28 , 85 – 91 . 
37. Baker , M. , Gregerson , M. S. , Martin , S. M. , and Buettner , G. R. ( 2003 ), Free radical and 
drug oxidation products in an intensive care unit sedative: Propofol with sulfi te , Crit. Care 
Med. , 31 , 787 – 792 . 
38. Lavoie , J - C. , Lachance , C. , and Chessex , P. ( 1994 ), Antiperoxide activity of sodium metabisulfi 
te. A double - edged sword , Biochem. Pharmacol. , 47 , 871 – 876 . 
39. Hayon , E. , Treinin , A. , and Wilf , J. ( 1972 ), Electronic spectra, photochemistry, and autoxidation: 
Mechanism of the sulfi te - bisulfi te - pyrosulfi te systems. The SO2
. , SO3
. , SO4
. 
radicals , J. Am. Chem. Soc. , 94 , 47 – 57 . 
40. International Conference on Harmonization (ICH) ( 1997 ), Guideline on residual solvents, 
ICH, Geneva. 
41. Bohrer , D. , do Nascimento , P. C. , Becker , E. , Bortoluzzi , F. , depoi , F. , and de Carvalho , L. 
M. ( 2004 ), Critical evaluation of the standard hydrolytic resistance test for glasses used 
for containers for blood and parenteral formulations , PDA J. Pharm. Sci. Technol. , 58 , 
96 – 105 . 
42. Scholze , H. (1988), Glas, Natur, Struktur und Eigenschaften , 3rd ed., Springer , Heidelberg , 
pp. 128 , 324 . 
REFERENCES 529

530 ORIGIN OF CONTAMINATION 
43. Bacon , F. R. ( 1986 ), Glass containers for parenterals , in Avis , K. E. , Lachman , L. , and 
Lieberman , H. A. , Eds., Pharmaceutical Dosages Forms: Parenteral Medications , Vol 2 , 
Marcel Dekker , New York , pp. 55 – 110 . 
44. Hak , E. B. , Storm , M. C. , and Helms , R. A. ( 1998 ), Chromium and zinc contamination of 
parenteral nutrient solution components commonly used in infants and children , Am. J. 
Health - Syst. Pharm. , 55 , 150 – 154 . 
45. Borg , C. , Constant , H. , Fusselier , M. , and Aulagner , G. ( 1994 ), Zinc, copper and 
iron in total parenteral nutrition mixtures: A contamination study , Nutrition , 13 , 
325 – 326 . 
46. Buchman , A. L. , Neely , M. , Grossie Jr, B. , Truong , L. , Lykissa , E. , and Ahn , C. ( 2001 ), 
Organ heavy - metal accumulation during parenteral nutrition is associated with pathologic 
abnormalities in rats , Nutrition , 17 , 600 – 606 . 
47. do Nascimento , P. C. , Marques , M. S. , Hilgemann , M. , de Carvalho L. M. , Bohrer , D. , 
Pomblum , S. G. , and Schirmer , S. ( 2006 ), Simultaneous determination of cadmium, copper, 
lead, and zinc in amino acid parenteral nutrition solutions by anodic stripping voltammetry 
and sample digestion by Uvirradiation , Anal. Lett. , 39 , 1 – 14 . 
48. Alfrey , A. C. , Le Gendre , G. R. , and Kaenhy , W. D. ( 1976 ), The dialysis encefalophathy 
syndrome, possible aluminum intoxication , N. Eng. J. Med. , 294 , 184 – 188 . 
49. Driscoll , W. R. , Cummings , J. J. , and Zorn , W. ( 1997 ), Aluminum toxicity in preterm infants , 
N. Eng. J. Med. , 337 , 1090 – 1091 . 
50. Bohrer , D. , do Nascimento , P. C. , Binotto , R. , and Pomblum , S. G. ( 2001 ), Infl uence of the 
glass packing on the contamination of pharmaceutical products by aluminium. Part I: 
Salts, glucose, heparin and albumin , J. Trace Elem. Med. Biol. , 15 , 95 – 101 . 
51. Bohrer , D. , do Nascimento , P. C. , Martins , P. , and Binotto , R. ( 2002 ), Availability of aluminum 
from glass and an Al - form exchanger in presence of complexing agents and amino 
acids , Anal. Chim. Acta , 459 , 267 – 276 . 
52. Bohrer , D. , do Nascimento , P. C. , Binotto , R. , and Becker , E. (2003), Infl uence of the glass 
packing on the contamination of pharmaceutical products by aluminium. Part III: Interaction 
container - chemicals during heating for sterilisation , J. Trace Elements Med. Biol. , 17 , 
107 – 115 . 
53. Bohrer , D. , do Nascimento , P. C. , Binotto , R. , and Carlesso , R. ( 2001 ), Infl uence of the 
glass packing on the contamination of pharmaceutical products by aluminium. Part II: 
Amino acids for parenteral nutrition , J. Trace Elements Med. Biol. , 15 , 103 – 108 . 
54. Popinska , K. , Kierkus , J. , Lyszkowska , M. , Socha , J. , Pietraszek , E. , Kmiotek , W. , and 
Ksiazky , J. ( 1999 ), Aluminum contamination of parenteral nutrition additives, amino acid 
solutions, and lipid emulsions , Nutrition , 15 , 683 – 686 . 
55. Baydar , T. , Aydin , A. , Duru , S. , Isimer , A. , and Sahin , G. ( 1997 ), Aluminum in enteral 
nutrition formulas and parenteral solutions , Clin. Toxicol. , 35 , 277 – 281 . 
56. Guidance for industry, container closure systems for packaging human drugs and biologics, 
chemistry, manufacturing and control documentation, Food and Drug Administration, 
1999 . 
57. Ullmann ’ s Encyclopedia of Industrial Chemistry , Vol. A5, 5th ed., VCH, Weinheim, 1993 , 
p. 334 . 
58. Solomon , D. D. , Jurgens , R. W. , and Wong , K. L. ( 1986 ), Plastic containers for parenterals , 
in Avis , K. E. , Lachman , L. , and Lieberman , H. A. , Eds., Pharmaceutical Dosages Forms: 
Parenteral Medications , Vol. 2 , Marcel Dekker , New York , pp. 111 – 153 . 
59. U.S. Food and Drug Administration (FDA) ( 1999 ), Safety assessment of di - 2 - ethylhexylphthalate 
(DEHP) released from PVC medical devices, FDA, Center for Devices and 
Radiological Health, Rockville, MD. 

60. Arabin , A. , and O stelius , J. ( 1980 ), Determination by electron - capture gas chromatography 
of mono(2 - ethylhexyl) phthalate and di(2 - ethylhexyl) phthalate in intravenous solutions 
stored in poly(vinyl chloride) bags , J. Chromatogra. B , 193 , 405 – 412 . 
61. Arabin , A. , Jacobsson , S. , Hagman , A. , and O stelius , J. ( 1986 ), Studies on contamination 
of intravenous solutions from poly(vinyl chloride) bags with dynamic headspace gas 
chromatography – mass spectrometry and gradient liquid chromatography diode array 
techniques , Int. J. Pharm. , 28 , 211 – 218 . 
62. Faouzi , M. A. , Khalfi , F. , Dine , T. , Luyckx , M. , Brunet , C. , Gressier , B. , Goudaliez , F. , Cazin , 
M. , Kablan , J. , Belabed , A. , and Cazin , J. C. ( 1999 ), Stability, compatibility and plasticizer 
extraction of quinine injection added to infusion solutions and stored in polyvinyl chloride 
(PVC) containers , J. Pharm. Biomed. Anal. , 21 , 923 – 930 . 
63. Kambia , K. , Dine , T. , Gressier , B. , Bah , S. , Germe , A. - F. , Luyckx , M. , Brunet , C. , Michaud , 
L. , and Gottrand , F. ( 2003 ), Evaluation of childhood exposure to di(2 - ethylhexyl) 
phthalate from perfusion kits during long - term parenteral nutrition , Int. J. Pharm. , 262 , 
83 – 91 . 
64. Allwood , M. C. , and Martin , H. ( 1996 ), The extraction of diethylhexylphthalate (DEHP) 
from polyvinyl chloride components of intravenous infusion containers and administration 
sets by paclitaxel injection , Int. J. Pharm. , 127 , 65 – 71 . 
65. Sautou - Miranda , V. , Brigas , F. , Vanheerswynghels , S. , and Chopineau , J. ( 1999 ), Compatibility 
of paclitaxel in 5% glucose solutions with ECOFLAC low - density polyethylene 
containers - stability under different storage conditions , Int. J. Pharm. , 178 , 77 – 82 . 
66. Marcato , B. , Guerra , S. , Vianello , M. , and Scalia , S. ( 2003 ), Migration of antioxidants additives 
from various polyolefi nic plastics into oleaginous vehicles , Int. J. Pharm. , 257 , 
217 – 225 . 
67. Smith , E. J. , and Nash , R. J. ( 1986 ), Elastomeric closures for parenterals , in Avis , K. E. , 
Lachman , L. , and Lieberman , H. A. , Eds., Pharmaceutical Dosages Forms: Parenteral 
Medications , Vol. 2 , Marcel Dekker , New York , pp. 155 – 215 . 
68. Accardi - Dey , A. , and Gschwend , P. M. ( 2003 ), Reinterpreting literature sorption data 
considering both absorption into organic carbon and adsorption onto black carbon , 
Environ. Sci. Technol. , 37 , 99 – 106 . 
69. U.S. Food and Drug Administration (FDA) ( 1999 ), Guidance for industry, container 
closure systems for packaging human drugs and biologics, FDA, Rockville, MD. 
70. Parenteral Drug Association (PDA) ( 1998 ), Pharmaceutical package integrity, Technical 
Report 27, PDA Bethesda, MD. 
71. Mannermaa , J. P. , R a is a nen , J. , Hyv o nen - Dabek , M. , Spring , E. , and Yliruusi , J. ( 1994 ), 
Use of proton - induced X - ray emission (PIXE) analysis in the evaluation of large volume 
parenteral rubber stoppers , Int. J. Pharm. , 103 , 125 – 129 . 
72. Delaunay - Bertoncini , N. , van der Wielen , F. W. M. , De Voogt , P. , Erlandsson , B. , and 
Schoenmakers , P. J. ( 2004 ), Analysis of low - molar - mass materials in commercial rubber 
samples by Soxhlet and headspace extractions followed by GC - MS analysis , J. Pharm. 
Biomed. Anal. , 35 , 1059 – 1073 . 
73. Jenke , D. R. ( 1997 ), Utilization of extraction profi les to estimate the accumulation of 
extractables from polymeric materials , J. Appl. Polym. Sci. , 63 , 843 – 848 . 
74. Loff , S. , Kabs , F. , Witt , K. , Sartoris , J. , Mandl , B. , Niessen , K. H. , and Waag , K. L. ( 2000 ), 
Polyvynilchloride infusion lines expose infants to large amounts of toxic plasticizers , 
J. Pediatr. Surg. , 35 , 1775 – 1781 . 
75. Haishima , Y. , Seshimo , F. , Higuchi , T. , Yamazaki , H. , Hasegawa , C. , Izumi , S. , Makino , T. , 
Nakahashi , K. , Ito , R. , Inoue , K. , Yoshimura , Y. , Saito , K. , Yagami , T. , Tsuchiya , T. , and 
Nakasawa , H. ( 2005 ), Development of a simple method for predicting the levels of DEHP 
REFERENCES 531

532 ORIGIN OF CONTAMINATION 
migrated from PVC medical devices into pharmaceutical solutions , Int. J. Pharm. , 298 , 
126 – 142 . 
76. Kambia , K. , Dine , T. , Gressier , B. , Germe , A. - F. , Luyckx , M. , Brunet , C. , Michaud , L. , and 
Gottrand , F. ( 2001 ), High - performance liquid chromatographic method for the determination 
of DEHP in total parenteral nutrition and in plasma , J. Chromatogr. B , 755 , 
297 – 303 . 
77. Boithias , S - B. , Sautou - Miranda , V. , Bourdeaux , D. , Tramier , V. , Boyer , A. , and Chopineau , 
J. ( 2005 ), Leaching of DEHP from multilayer tubing into etoposide infusion solutions , 
Am. J. Health - Syst. Pharm. , 62 , 182 – 188 . 
78. Jenke , D. R. , Story , J. , and Lalani , R. ( 2006 ), Estractables/leachables from plastic tubing 
used in product manufacturing , Int. J. Pharm. , 315 , 75 – 92 . 
79. Groves , M. J. (1991), Particulate contamination in parenterals: Current issues , Boll. Chim. 
Farma. , 130 , 347 – 354 . 
80. Bikhazi , A. B. , Shiatis , J. A. , and Haddad , A. F. ( 1977 ), Quantitative estimation of particulate 
matter in pharmaceutical preparations intended for intravenous administration , 
J. Pharm. Sci. , 66 , 181 – 186 . 
81. Foroni , L. A. , Rochat, M. H. , Trouiller , P. , and Calop , J. Y. (1993), Particle contamination 
in a ternary nutritional admixture , J. Parenteral Sci. Technol. , 47 , 311 – 314 . 
82. Ball , P. A. , Bethune , K. , Fox , J. , Ledger , R. , and Barnett , M. ( 2001 ), Particulate contamination 
in parenteral nutrition solutions: Still a cause for concern? Nutrition , 17 , 926 – 929 . 
83. Oie , S. , and Kamiya , A. ( 2005 ), Particulate and microbial contamination in in - use admixed 
parenteral nutrition solutions , Biol. Pharm. Bull. , 28 , 2268 – 2270 . 
84. Roseman , T. J. , Brown , J. A. , and Scothorn , W. W. ( 1976 ), Glass for parenteral products: 
A surface view using the scanning electron microscope , J. Pharm. Sci. , 65 , 22 – 29 . 
85. Turco , S. J. , and Davis , N. M. ( 1971 ), Detrimental effects of particulate matter on the pulmonary 
circulation , J. Am. Med. Assoc. , 217 , 81 – 82 . 
86. Hearse , D. J. , Sonmez , B. , Saldanha , C. , Braimbridge , M. V. , Maxwell , M. P. , and Erol , C. 
( 1986 ), Particle - induced coronary vasoconstriction in the rat heart: Pharmacological 
investigation of underlying mechanisms , Thorac. Cardiovasc. Surg. , 34 , 316 – 325 . 

533 
6.2 
QUANTITATION OF MARKERS 
FOR GRAM - NEGATIVE AND 
GRAM - POSITIVE ENDOTOXINS 
IN WORK ENVIRONMENT AND 
AS CONTAMINANTS IN 
PHARMACEUTICAL PRODUCTS 
USING GAS CHROMATOGAPHY – 
TANDEM MASS SPECTROMETRY 
Alvin Fox 
University of South Carolina, Columbia, South Carolina 
Contents 
6.2.1 Introduction 
6.2.2 Analysis of GC – MS/MS Markers for LPS and PG and Applications 
6.2.3 Operational Parameters 
6.2.4 Concluding Remarks 
References 
6.2.1 INTRODUCTION 
Endotoxins are bacterial cell envelope constituents that, when present in pharmaceutical 
products, cause pyrogenic reactions sometimes resulting in lethality. The 
toxicity of endotoxins is directly related to their chemical composition. However, 
the viability of the organism is irrelevant since endotoxin derived from dead or live 
microbes is equally active. The classical endotoxin is lipopolysaccharide (LPS). 
However, peptidoglycan (PG) also displays endotoxin - like activities. LPS is found 
only in gram - negative bacterial outer membranes, while PG is present in the cell 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

534 QUANTITATION OF MARKERS 
wall of both gram - positive and gram - negative bacteria. The Limulus amoebocyte 
lysate (LAL) test is widely used as a biological assay for LPS levels but has poor 
sensitivity for detection of PG. Substances are found in LPS and PG [3 - hydroxy 
fatty acids (3 - OH FAs) and muramic acid (Mur), respectively] that are rarely found 
elsewhere in nature and serve as chemical markers. Both markers for LPS and PG 
can be detected using gas chromatography – tandem mass spectrometry (GC – MS/ 
MS) and this latter technology is the focus of this chapter. 
A recently published book provides an excellent survey of issues that relate to 
contamination with endotoxins (present in both viable and nonviable bacteria), 
their released cell wall constituents, and also viable bacteria in the pharmaceutical 
industry [1] . It is important to know both the content of the work environment (e.g., 
indoor air) and the pharmaceutical products themselves. The former provides information 
on possible sources of microbial contamination and the latter the purity of 
the fi nal commercial product (or precursors in various stages in its preparation). In 
some cases it is vital to know the actual bacterial species involved in contamination; 
culture - based methods are standard microbiological techniques which were the 
focus of Jimenez [1] and thus will not be discussed further. Any contamination (e.g., 
with endotoxins), regardless of the species of origin, is of utmost of importance 
(e.g., in determining the safety of a batch of antibiotics to be administered intravenously). 
This is determined optimally by non - culture - based methods. 
Endotoxic reactions result from infl ammatory responses to both culturable and 
nonculturable organisms. Thus, traditional culture - based methods often signifi cantly 
underestimate the risk of endotoxin present. Culture - and non - culture - based monitoring 
methods differ greatly in their characteristics. Importantly, results from 
culture - based methods are strongly affected by variability in culture conditions (e.g., 
media selected, culture time, and temperature), making it diffi cult to standardize 
protocols for quantitation of contamination and only detecting those organisms that 
grow under the conditions selected. However, the use of non - culture - based procedures 
certainly does not eliminate the need for information obtained by culture. 
Real - time polymerase chain reaction (PCR) methods have become the primary 
alternative to bacterial culture, measuring the levels of characteristic genes originating 
from particular species (whether culturable or not). PCR has proven to be highly 
amenable in the clinical microbiology laboratory, but its utility in more diverse 
environmental or pharmaceutical products can be more problematic. Of course, 
PCR will not detect an endotoxin. PCR enzymatically amplifi es a genetic region, 
with a characteristic deoxyribonucleic acid (DNA) sequence; a fl anking set of 
primers (short pieces of complementary DNA) focuses the enzyme on the DNA 
region of interest. Classical PCR involves detection of a PCR product by electrophoretic 
mobility on a gel, which is time consuming. Real - time PCR is distinct from 
classical PCR in that electrophoresis is avoided and the PCR product is detected 
simply by an increase in fl uorescence; this can be performed with or without prior 
culture. This approach is of great utility, but there can be variation depending on 
the sample matrix (e.g., in some cases total inhibition of the reaction by metal ions 
or humic acids). Assessment of total bioload (e.g., by measuring the levels of chemical 
markers for bacterial endotoxins by GC – MS/MS) is much less subject to matrix - 
based variation and is more readily standardized. 
An overview of the relative advantages (and disadvantages) of modern non - 
culture - based methods to each other and versus culture was also provided by 

ANALYSIS OF GC–MS/MS MARKERS FOR LPS AND PG AND APPLICATIONS 535 
Jimenez [1] , including discussing measurement of adenosine triphosphate (ATP) 
levels, microscopy/fl uorescent - activated cell sorting, molecular biology (e.g., PCR 
and microarrays), and immunology - based methods. The endotoxin section focused 
on the LAL bioassay method, which is still the most common method for measuring 
such contamination; modern alternatives, including bioassays for cytokines and 
analytical chemical methods (GC – MS), were also discussed. The LAL assay measures 
biological activity which can vary with slight changes in structure among 
endotoxins affecting determination of levels; this is not the case for chemical analysis 
[2] . Furthermore biological assays, including the LAL assay, often suffer from 
false positives from cross - reactivity with other contaminants. As mentioned above, 
the LAL assay only detects the gram - negative endotoxin LPS, not the gram - positive 
endotoxin PG. 
Chemical assays for endotoxin are considerably more selective than the LAL 
assay but until recently lacked the sensitivity required for routine detection of gram - 
positive and gram - negative endotoxins. This relates to the introduction of advanced 
but still user - friendly GC – MS/MS instrument which are rapidly replacing GC – MS 
for trace analysis in the pharmaceutical industry and elsewhere. This chapter will 
thus focus on GC – MS/MS methodology for detection of 3 - OH FAs (markers for 
LPS) and Mur (a marker for PG), which are each well established [3 – 8] . Ergosterol 
is also used as a marker for fungal contamination using GC – MS/MS [9] . 
Endotoxicity results from the interaction of a bacterial cell envelope component 
(e.g., LPS or PG with a cell surface receptor constituting part of the nonspecifi c 
immune system, (i.e., a toll - like receptor on white blood cells). This results in the 
production of cytokines [e.g., interleukin 1 (IL - 1) or tumor necrosis factor (TNF)] 
as part of an intracellular enzyme cascade which can cause severe tissue injury. 
Bioassays or immunoassays can be used to detect such reactions respectively. As 
noted above the most widely used bioassay is the LAL assay. A lysate of amoebocytes 
of the horseshoe crab ( Limulus ) contains an enzymatic clotting cascade which 
is activated by extremely low levels of LPS (nanogram levels or lower). There are 
variants of this assay that can detect PG, but they are not as widely used. As noted 
above, other bioassays employ cultured cell lines that respond to LPS or PG, respectively. 
Unfortunately bioassays are highly amenable to false positives (from the 
presence of cross - reactive substances) or false negatives from inhibition (by contaminants 
present in the sample) [10] . A detailed discussion of these assays is 
beyond the scope of this chapter and has been reviewed elsewhere [1] . 
6.2.2 ANALYSIS OF GC – MS / MS MARKERS FOR LPS AND 
PG AND APPLICATIONS 
For groups experienced in trace analysis of chemical markers for bacteria, currently 
samples are analyzed almost exclusively by GC – MS/MS. GC – MS/MS assays are 
now well established and a wide range of clinical and environmental samples have 
been analyzed. However, application in the pharmaceutical industry requires further 
evaluation. For example, Mur is released by hydrolysis and analyzed as an alditol 
acetate; 3 - OH FAs after methanolysis are converted to methyl trimethylsilyl derivative. 
Detailed analytical procedures have been described elsewhere for Mur [3, 4, 7, 
8, 11] and for 3 - OH FAs [2, 5, 12, 13] . The compounds of interest contain active 

536 QUANTITATION OF MARKERS 
groups (e.g., OH or CO 2 H) that interact with GC columns. Derivatization is necessary 
to convert the samples into a suitable form for analysis. Prederivatization and 
postderivatization clean - up yields simple chromatograms. Quantitation employs 
stable - isotope - labeled forms of the markers (e.g., 13 C - labeled Mur derived from 
labeled blue - green algae). 
The 3 - OH FAs have had great utility in the determination of LPS levels in indoor 
air. However, in tissues and body fl uids it has been determined that 3 - OH FAs are 
naturally present at low levels as products of mammalian metabolism (mitochondrial 
fatty acid . oxidation). Due to this background GC – MS/MS for 3 - OH FAs is 
not recommended as a general marker to determine trace LPS levels in clinical 
samples [14] . However, in certain situations the assessment of 3 - OH FAs has been 
successfully used, for example, in the diagnosis of chronic peridontitis [15] . There is 
great potential for the utility of 3 - OH FAs as markers for LPS contamination in 
pharmaceutical products, where often the background matrix would be anticipated 
to be much less complex. 
Chemical analysis of Mur levels has proved effective in both clinical and environmental 
samples, since Mur is not synthesized by eukaryotic cells. For example, 
it is readily detected in infected human body fl uids, for example, synovial fl uids from 
patients with staphylococcal arthritis and spinal fl uids from those with pneumococcal 
pneumonia [7, 16] . However, the most widely used method for its analysis, as an 
alditol acetate, is time consuming. A large number of derivatives have been tested 
in order to develop a simpler alternative. Unfortunately the limit of detection for 
these alternative approaches has not been optimal [17, 18] . 
The marker monomer is chemically converted into a volatile form suitable of 
passage through a gas chromatograph. The samples then pass into the tandem mass 
spectrometer where the ionized molecules are detected. GC – MS/MS employs GC 
separation coupled with the exquisite selectivity of MS/MS. In MS (monitoring/ 
quantitation mode), background peaks are screened out. MS/MS screens out background 
a second time, thus dramatically lowering the detection limit. Alternatively, 
in the identifi cation mode (MS/MS), a chemical fi ngerprint of the compound of 
interest allows defi nitive identifi cation. The analysis of Mur serves as an example: 
natural 12 C muramic acid is fi rst released from PG polymers (present as a minor 
component in a complex sample matrix) by hydrolysis. Conversion of 12 C muramic 
acid to a volatile derivative, muramicitol lactam pentaacetate molecular weight 
(MW) 445, is essential for GC – MS/MS analysis. 
Maximal selectivity, in GC – MS analysis, requires selected ion monitoring (SIM). 
In SIM, one or more prominent ions characteristic of a given compound are monitored 
exclusively, thereby ignoring background or contaminant ions. Similarly, 
maximum selectivity in GC – MS/MS analysis involves twofold SIM, or multiple - 
reaction monitoring (MRM). In quadropole mass spectrometers, the fi rst stage of 
MRM involves selective transmission of a molecular ion from the fi rst mass spectrometer 
to the collision cell. This instrumental clean - up removes all nonidentical 
molecular weight contaminants produced from the initial ionization. The selected 
ion is then fragmented by collision with an inert gas (e.g., argon). In the second mass 
spectrometer the selected fragment ion is monitored. If either sensitivity or specifi city 
is adversely affected, then the limit of detection is affected. GC – MS/MS provides 
much greater specifi city than GC – MS. Using GC – MS it was impossible to reliably 
visually differentiate chromatograms that contained the lowest concentrations of 

ANALYSIS OF GC–MS/MS MARKERS FOR LPS AND PG AND APPLICATIONS 537 
Mur from negative controls (including plants and fungi). GC – MS/MS chromatograms 
of dust were always readily differentiated from controls. However, these 
analyses were clearly performed at the current limits of sensitivity for GC – MS/MS. 
Current developments in MS technology may lead to dramatically improved sensitivity 
by GC – MS/MS. 
Tandem mass spectrometry consists of two stages. For example, in Mur analysis, 
in the fi rst stage, the molecule is isolated essentially intact with a MW of 403 due 
to the loss of a ketene (loss of 42). Coeluting molecules of different MW are largely 
but not totally eliminated; that is, only molecules with a MW of 403 are permitted 
to pass into the next stage. In the second stage, molecules with MW of 403 are 
broken into characteristic fragments including one that contains the original lactam 
with a MW of 198. Thus, in the second stage, only molecules with a MW of 198 are 
detected. Thus, for a second time, background molecules of different MW that 
coelute are essentially eliminated. 
The use of a stable - isotope - labeled ( 13 C) analog of muramic acid, in each sample, 
verifi es that there is no false - negative result. This assures that muramic acid is not 
lost during the sample preparation or hidden within the background in the instrumental 
analysis. Although 12 C muramic acid and 13 C muramic acid have the same 
retention time on GC analysis, they can be discriminated in the tandem mass spectrometer. 
GC – MS/MS analysis of 13 C muramic acid is identical to natural ( 12 C) 
muramic acid. However, the MWs in the fi rst and second stages are correspondingly 
higher, 412 and 205, respectively. Thus in the tandem mass spectrometer, it is possible 
to monitor two separate windows simultaneously, one for 13 C muramic acid (top 
window) and one for natural muramic acid (bottom window). Accurate quantitation 
is readily accomplished by comparing the ratio of the area of 13 C versus 12 C muramic 
acid (see Figure 1 ). 
A chemical fi ngerprint is generated consisting of scission products, in the MS/MS 
instrument, characteristic of the compound of interest. The parent molecule has a 
MW of 403. In each case the same major fragments are observed: masses 361, 301, 
FIGURE 1 GC – MS/MS monitoring chromatogram. The upper window depicts the internal 
standard ( 13 C muramic acid) and the lower natural ( 12 C) muramic acid isolated from dust. 
The peak areas in the two separate windows are normalized relative to the highest peak in 
that window. Medical samples appear similar 
Retention time 
Abundance 
0 
100
0 
100 
%
% 
20 22 24 26 28 30 32 34 36 38 40 
412.4 204.9 
403.4 197.9

538 QUANTITATION OF MARKERS 
258, 240, 198, 156, and 138: 361 results from the loss of a ketene (loss of 42) and 301 
from a subsequent loss of acetic acid (loss of 60). Breakage at C4 – C5 (loss of 145) 
would generate 258 and further loss of acetic acid (loss of 60) would generate 
198. Loss of ketene or acetic acid from 198 results in 156 and 138, respectively (see 
Figure 2 ). 
6.2.3 OPERATIONAL PARAMETERS 
Every marketed product has a level of endotoxin tolerated based on the minimum 
pyrogenic dose and amount of the drug to be administered as per Food and 
Drug Administration (FDA) guidelines [19] . However, there are none for the 
more advanced chemical assays described here. Indeed there are only a few 
highly specialized university laboratories that currently have experience in trace 
chemical analysis of LPS and PG. There are no commercial testing laboratories. 
Simplifi cation and automation will allow more widespread availability of these 
methods. 
At the current time it is recommended that 1 – 10 mg of solid sample be analyzed. 
In environments that are not heavily contaminated with airborne dust it takes 
several days for sample collection. In an unoccupied or minimally occupied room 
around 1 mg of dust needed for analysis can be collected in 72 h. In a heavily occupied 
room the levels of dust increase dramatically and collection takes around 6 – 8 h. 
FIGURE 2 GC - MS/MS chemical fi ngerprint (product ion spectrum) of ( a ) standard 
muramic acid (2 ng total in sample) and ( b ) muramic acid isolated from dust. Medical samples 
appear similar. 
Abundance 
Mas/charge ratio 
100 
10 
20 
30 
40 
50 
60 
70 
80 
90 
198.1 
361.0 
0 . 5 8 3 1 . 8 3 1 258.1 
156.1 
240.1 
168.1 
301.1 181.1 
326.1 
403.1 
0 
404.0 47
0 
10 
20 
30 
40 
198.1 
258.2 385.2 
361.2 273.1 
138.1
156.2
168.1 
333.2 240.2 301.1 
367.1

Chemically inert membranes (e.g., Tefl on) are preferably used for dust collection 
since they are not affected by heating in sulfuric acid, the fi rst stage of the chemical 
analysis. 
Large (milligram - to - gram) quantities of dust can readily be collected from air 
conditioners or surfaces. The concentration of dust in air varies from micrograms 
to milligrams pero cubic meter. Low concentrations are more typical of offi ce buildings 
and laboratories, while the high range is more typical of dusty environments 
such as barns or chicken houses. Since commonly used pumps collect a few 
liters per minute, except in grossly contaminated environments, only microgram 
amounts of dust can be collected without extended air - sampling periods. Chemical 
markers of interest are only one component of bacteria and the bacteria constitute 
only a small fraction of the dust. Mur is present at between 5 and 50 ng/mg (corresponding 
to approximately 50 – 500 ng PG/mg dust) in normal house and air conditioning 
dust. Somewhat higher levels have been found in airborne dust (over 
100 ng Mur/mg). LPS is present at between 500 and 5000 ng/mg dust. Solids (e.g., 
pharmaceutical products) are more simply analyzed, generally without any sample 
pretreatment. Generally, for the optimal level of sensitivity, around 10 mg should be 
analyzed. 
As noted above, sensitive and specifi c GC – MS/MS methods for the determination 
of 3 - OH FAs and Mur have been developed. MS is an alternative to the classical 
LAL assay for determination of LPS, while no other regulated approach exists for 
PG assessment. These chemical methods are reproducible and provide quantitative, 
accurate determination of microbial biocontamination. At the present time mass 
spectrometric measurement of LPS and PG have matured suffi ciently to be used 
for routine assessment of air quality. Numerous products of medical and environmental 
origin have been analyzed. However, use for assessment of pharmaceutical 
products remains limited. 
The precision of MS assays is in the range typical of most clinical assays (i.e., 
under 5 – 15%). The best choice of internal standard is the stable - isotope - labeled 
form (preferably 13 C) of the compound of interest (e.g., . - hydroxy myristic acid or 
muramic acid). Specifi c trace detection of chemical markers in complex matrices 
requires appropriate negative controls. Procedures are often described that do not 
employ the mass spectrometer and false positives are often reported. The mere 
analysis of blank fi lters or water blanks is not satisfactory since chemical noise 
contributed by the sample is much greater and is not accounted for with this form 
of control. 
As noted above the predominant technique for measuring endotoxin levels is the 
LAL assay. This assay involves assessing activation of a clotting cascade in amoebocyte 
lysates (from the horseshoe crab). The LAL primarily detects LPS, and the 
sensitivity of detection of PG by LAL is extremely poor. The LAL bioassay and MS 
measurements of LPS sometimes correlate poorly. LAL measures biological activity 
while MS measures total quantity. Thus the two techniques are not strictly comparable. 
However, some differences (between results with LAL versus MS) may relate 
to the superior specifi city of tandem mass spectrometry. It is worthy of note that 
GC – MS/MS can provide some information on the population of gram - negative 
bacteria present since distribution of hydroxy FAs varies among bacterial species. 
The 2 - and 3 - OH fatty acids with 10 – 18 carbon atoms are all common in organic 
dust. 
OPERATIONAL PARAMETERS 539

540 QUANTITATION OF MARKERS 
6.2.4 CONCLUDING REMARKS 
User - friendly commercial GC – MS/MS instruments have been available since the 
mid - 1990s. Unfortunately there has been limited development in instruments that 
perform automated processing of samples for GC – MS/MS analysis, although a 
prototype automated derivatization has been built [20] . The pharmaceutical industry 
is well versed in the use of other types of mass spectrometry detection methods 
for high - throughput drug analysis [e.g., liquid chromatography – tandem mass spectrometry 
(LC - MS/MS) which eliminates the necessity for posthydrolysis derivatization 
by employing electrospray ionization mass spectrometry]. Unfortunately there 
has been limited interest in applying these methods for detection of LPS or PG 
markers [21] . Thus analysis of markers for endotoxins remains technically demanding 
and this has inhibited the widespread use of these techniques outside of a few 
specialist laboratories. However, there is great potential in the application of GC – 
MS/MS methods for assessing contamination with gram - positive and gram - negative 
bacterial endotoxins in the pharmaceutical industry. Tandem mass spectrometers 
are still expensive. However, such instruments are widely available in the pharmaceutical 
industry for general drug analysis. The most modern machines are run by 
Windows - based PC programs. Thus they can be readily operated by individuals, 
after appropriate training, with little prior experience of mass spectrometry or 
indeed analytical chemistry. 
ACKNOWLEDGMENT 
The research described in this chapter was supported by Philip Morris USA, Inc., 
and Philip Morris International. 
REFERENCES 
1. Jimenez , L. ( 2004 ), Microbial Contamination Control in the Pharmaceutical Industry , 
Marcel Dekker , New York . 
2. Saraf , A. , Larsson , L. , Burge , H. , and Milton , D. ( 1997 ), Quantifi cation of ergosterol and 
3 - hydroxy fatty acids in settled house dust by gas chromatography – mass spectrometry: 
Comparison with fungal culture and determination of endotoxin by a Limulus amoebocyte 
lysate assay , Appl. Environ. Microbiol ., 63 , 2554 – 2559 . 
3. Fox , A. , Wright , L. , and Fox , K. ( 1995 ), Gas chromatography tandem mass spectrometry 
for trace detection of muramic acid, a peptidoglycan chemical marker in organic dust , 
J. Microbiol. Meth ., 22 , 11 – 26 . 
4. Fox , A. , Krahmer M. , and Harrleson , D. ( 1996 ), Monitoring muramic acid in air (after 
alditol acetate derivatization) using a gas chromatograph – ion trap tandem mass spectrometer 
, J. Microbiol. Meth ., 27 , 129 – 138 . 
5. Saraf , A. , and Larsson , L. ( 1996 ), Use of gas chromatography ion trap tandem mass spectrometer 
for the determination of microorganisms in organic dust , J. Mass Spectrom ., 31 , 
389 – 396 . 
6. Larsson , L. , and Saraf , A. ( 1997 ), Use of gas chromatography – ion trap tandem mass 
spectrometry for the detection and characterization of microorganisms in complex 
samples , Mol. Biotechnol ., 7 , 279 – 287 . 

7. Kozar , M. , Krahmer , M. T ., Fox , A. , and Gray , B. M. ( 2000 ), Failure to detect muramic acid 
in normal rat tissues but detection in cerebrospinal fl uid from patients with pneumococcal 
meningitis , Infect. Immun ., 68 , 4688 – 4698 . 
8. Kozar , M. , Laman , J. D. , and Fox , A. ( 2002 ), Muramic acid is not generally present in 
human spleen as determined by gas chromatography – tandem mass spectrometry , Infect. 
Immun ., 70 , 741 – 748 . 
9. Sebastian , A. , and Larsson , L. ( 2003 ), Characterization of the microbial community in 
indoor environments: A chemical - analytical approach , Appl. Environ. Microbiol ., 69 , 
3103 – 3109 . 
10. Kaneko , T. , Goldman , W. E. , Mellroth , P. , Steiner , H. , Fucase K. , Kusumoto , S. , Harley , W. , 
Fox , A. , Golenbock , D. , and Silverman , N. ( 2004 ), Monomeric and polymeric Gram - 
negative peptidoglycan but not purifi ed LPS stimulate the Drosophila IMD pathway , 
Immun ., 20 , 1 – 20 . 
11. Fox , A. , Harley , W. , Feigley , C. , Salzberg , D. , Toole , C. , Sebastian , A. , and Larsson , L. 
( 2005 ), Large particles are responsible for elevated bacterial marker levels in school air 
upon occupation , J. Environ. Monit ., 7 , 450 – 456 . 
12. Mielniczuk , Z. , Mielniczuk , E. , and Larsson , L. ( 1993 ), Gas chromatography – mass spectrometry 
methods for analysis of 2 - and 3 - hydroxylated fatty acids: Application for 
endotoxin measurement , J. Microbiol. Meth ., 17 , 91 – 102 . 
13. Mielniczuk , Z. , Mielniczuk , E. , and Larsson , L. ( 1995 ), Determination of muramic acid in 
organic dust by gas chromatography – mass spectrometry , J. Chromatogr. B , 670 , 167 – 
172 . 
14. Sponazr , B. , Norin , E. , Midvedt , T. , and Larrson , L. ( 2002 ), Limitations in the use of 3 - 
hydroxy fatty acids to determine endotoxin in mammalian tissues , J. Microbiol. Meth ., 50 , 
283 – 289 . 
15. Ferrando , R. , Szponar , B. , S a nchez , A. , Larsson , L. , and Vaero - Guill e n , P. L. ( 2005 ), 3 - 
Hydroxy fatty acids in saliva as diagnostic markers in chronic peridontitis , J. Microbiol. 
Meth ., 62 , 285 – 291 . 
16. Fox , A. , Fox , K. , Christensson , B. , Krahmer , M. , and Harrelson , D. ( 1995 ), Absolute identifi 
cation of muramic acid at trace levels in human septic fl uids in vivo and absence in 
aseptic fl uids , Infect. Immun ., 64 , 3911 – 3955 . 
17. Kozar , M. , and Fox , A. ( 2002 ), Analysis of a stable halogenated derivative of muramic 
acid by gas chromatography – negative ion chemical ionization tandem mass spectrometry , 
J. Chromatogr ., 946 , 229 – 238 . 
18. Sebastian , A. , Harley , W. , Fox , A. , and Larsson , L. ( 2004 ), Evaluation of the methyl ester 
O - methyl acetate derivative of muramic acid for the determination of peptidoglycan in 
environmental samples by ion - trap GC - MS - MS , J. Environ. Monit ., 6 , 1 – 6 . 
19. U.S. Department of Health and Human Services (DHHS) ( 1987 ), FDA guidelines on 
validation of the Limulus amoebocyte lysate test as an end product test for human and 
animal parenteral drugs, biological products and medical devices, DHHS, Rockville, 
MD. 
20. Steinberg , P. , and Fox , A. ( 1999 ), Automated derivatization instrument: Preparation of 
alditol acetates for analysis of bacterial carbohydrates using gas chromatography – mass 
spectrometry , Anal. Chem ., 71 , 1914 – 1917 . 
21. Shahgholi , M. , Ohorodnik , S. , Callahan , J. , and Fox , A. ( 1997 ), Trace detection of underivatized 
muramic acid in environmental dust samples by microcolumn liquid chromatography 
– electrospray tandem mass spectrometry , Anal. Chem ., 69 , 1956 – 1960 . 
REFERENCES 541


543 
6.3 
MICROBIOLOGY OF NONSTERILE 
PHARMACEUTICAL 
MANUFACTURING 
Ranga Velagaleti 
BASF Corporation, Florham Park 
Contents 
6.3.1 Introduction 
6.3.2 Global Regulations and Regulatory Guidances Relevant to Microbial Bioburden 
Control during Nonsterile Manufacturing 
6.3.2.1 GMPs for Finished Products and Components of Finished Products 
6.3.2.2 GMPs for Active Pharmaceutical Ingredients 
6.3.2.3 GMPs for Bulk Pharmaceutical Excipients 
6.3.3 Pharmacopeial Guidelines Relevant to Microbiology of Nonsterile Manufacturing 
6.3.3.1 United States Pharmacopeia 
6.3.3.2 European Pharmacopoeia 
6.3.3.3 Japanese Pharmacopoeia 
6.3.4 Current Regulatory Expectations for Microbial Bioburden Control during Nonsterile 
Manufacturing 
6.3.5 Industry Perspective on Microbial Bioburden Control for Nonsterile Pharmaceutical 
Manufacturing 
6.3.5.1 Survey Results 
6.3.5.2 Recommendations 
6.3.6 Microbial Bioburden Control during Shelf Life of Pharmaceutical Products 
6.3.7 Summary and Conclusions 
References 
6.3.1 INTRODUCTION 
In order to comply with general microbiological quality attributes for nonsterile 
pharmaceuticals and established monograph or drug application specifi cations for 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

544 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
raw materials and fi nished products, manufacturers need to follow the good manufacturing 
practice (GMP) regulations, with specifi c reference to contamination 
control stated in various sections of GMPs. For example, GMP regulations promulgated 
by the U.S. Food and Drug Administration (FDA) for fi nished products in the 
Code of Federal Regulations state (21 CFR Part 211.113): “ Appropriate written 
procedures, designed to prevent objectionable microorganisms in drug products not 
required to be sterile, shall be established and followed ” [1] . To meet this regulatory 
requirement, the microbial bioburden of nonsterile pharmaceutical raw materials 
and fi nished - product manufacturing environment (air, walls, and fl oors), equipment 
used for manufacturing, as well as the fi nished raw materials and drug products 
should be monitored and controlled to ensure that they have acceptable amount of 
total microbial bioburden and are free of objectionable microorganisms. Historically, 
the microbial contamination of nonsterile pharmaceuticals has been a concern, 
with a number of products recalled due to microbial contamination with undesirable 
microorganisms due to their suspected health hazards. As early as 1979, Dunnigan 
of the Bureau of Medicines at the FDA, commenting on the acceptable levels and 
contamination of microbiological contamination in drug products, expressed his 
concern that topical preparations contaminated with gram - negative organisms are 
a probable moderate to serious health hazard [2] . 
The U.S. Pharmacopeia (USP), in a recent revision of general chapter < 1111 > 
(585), “ Microbiological Examination of Nonsterile Products: Acceptance Criteria 
for Pharmaceutical Preparations and Substances for Pharmaceutical Use, ” stated: 
“ The presence of certain microorganisms in nonsterile preparations may have the 
potential to reduce or even inactivate the therapeutic activity of the product and 
has a potential to adversely affect the health of the patient. Manufacturers have 
therefore to ensure low bioburden of fi nished dosage forms by implementing current 
guidelines on GMPs during the manufacture, storage and distribution of pharmaceutical 
preparations ” [3] . 
Pharmaceutical manufacturers and their trade associations have been proactive 
in addressing the concerns related to microbial contamination in nonsterile manufacturing 
environment. In an article in the March 1997 issue of Pharmaceutical 
Technology , the Pharmaceutical Research and Manufacturers Association of 
America (PhRMA) Environmental Monitoring Work Group provided a realistic 
assessment of the need for understanding and controlling microbial bioburden 
contributed by raw materials, primary packaging components, the manufacturing 
process, and the manufacturing environment. The work group emphasized in this 
article that, except for sterile products, most pharmaceutical products are neither 
intended to be nor represented as being sterile, but because they are administered 
to people who are ill and in a weakened state, the microbial content should be 
minimized [4] . The Australian Pharmaceutical Manufacturers ’ Association (APMA) 
published a guideline on nonsterile pharmaceuticals for human use emphasizing the 
need for control of microbial contamination and the importance of adhering to 
GMPs to control microbiological levels [5] . 
Besides GMP regulations for fi nished dosage forms (which also include regulatory 
requirements for inactive ingredients, active ingredients, containers, and 
closures used for fi nished products) published by the FDA [1] , the International 
Conference on Harmonization (ICH) [6] (GMPs for active ingredients), authorities 
in Australia [7, 8] (pharmaceutical products), as well as the World Health Organization 
(WHO) [9] (pharmaceutical products), International Pharmaceutical 

GLOBAL REGULATIONS AND REGULATORY GUIDANCES 545 
Excipients Council (IPEC) [10] (excipients), and the USP [11] have also published 
GMP guidelines. These guidelines require manufacture of pharmaceutical products 
with acceptable microbial bioburden and free of objectionable microorganisms. 
This chapter discusses regulatory aspects of microbiology of nonsterile pharmaceutical 
manufacturing, including manufacturing environment, raw materials, and 
fi nished products. In addition, the microbiological bioburden control of pharmaceutical 
raw materials and products is reviewed, with particular reference to the type 
of fi nished dosage forms. Assessment of microbial bioburden requirements for nonsterile 
pharmaceutical manufacturing by the pharmaceutical industry is discussed. 
The microbiology of the pharmaceutical manufacturing environment for production 
of sterile pharmaceuticals or microbiological quality expectations for sterile 
products is not discussed in this chapter. 
6.3.2 GLOBAL REGULATIONS AND REGULATORY GUIDANCES 
RELEVANT TO MICROBIAL BIOBURDEN CONTROL DURING 
NONSTERILE MANUFACTURING 
GMP regulations for fi nished pharmaceutical drugs and components of fi nished 
drugs promulgated by the FDA [1] , GMP guidelines for active pharmaceutical 
ingredients (Q7A) from the ICH [6] , and GMP guidelines for bulk pharmaceutical 
excipients published by the USP [11] are discussed to illustrate the GMP requirements 
for pharmaceutical fi nished drug products, active (drug substances) and inactive 
(excipients) components of the drug products, respectively. In the context of 
this discussion on GMP regulations relevant to the microbiology of nonsterile 
manufacturing, contamination is defi ned as “ the undesired introduction of impurities 
of a chemical or microbiological nature, or of foreign matter, into or on to a 
raw material, intermediate, or API during production, sampling, packaging, or 
repackaging, storage, or transport ” [6] . A similar defi nition of contamination applies 
with reference to drug products. Microbiological bioburden refers to the level and 
type of microorganisms that may be present. The drug product, drug product components, 
manufacturing equipment, or environment can be considered contaminated 
when bioburden exceeds established specifi cations or acceptance criteria. 
6.3.2.1 GMP s for Finished Products and Components of Finished Products 
The FDA enforces the implementation of current good manufacturing practice 
(cGMP, GMP) regulations (21 CFR Parts 210 and 211) [1] for drug products manufactured 
in the United States (also applicable to drugs and drug components manufactured 
abroad and imported into the United Sates), under the Food, Drug, and 
Cosmetic Act (FD & C Act). Under FD & C Act 501(a)(2)(B), failure to comply with 
any GMP regulation shall render such drug to be adulterated and such drug as well 
as the person who is responsible for failure to comply shall be subjected to regulatory 
action [1] . 
It is important to understand the regulatory defi nitions of fi nished drug product, 
active ingredient, and inactive ingredient [1] in the context of the discussions on the 
subject presented here. Under GMP regulations 21 CFR Part 210.3(b)(4), drug 
product is defi ned as a fi nished dosage form, for example, tablets, capsule, and solution, 
that contains an active ingredient generally, but not necessarily, in association 

546 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
with inactive ingredients. The term also includes a fi nished dosage form that does 
not contain an active ingredient but is intended to be used as a placebo. Placebo is 
used especially during safety and effi cacy studies during drug development and in 
preclinical animal studies or human clinical trials. Active ingredient (also known as 
active drug substance) means any component that is intended to furnish pharmacological 
activity or other direct effect on the diagnosis, cure, mitigation, treatment, 
or prevention of disease or to affect the structure or any function of the body of 
humans or other animals [21 CFR Part 210.3(b)(7)]. Inactive ingredient (also known 
as excipient) means any component other than an active ingredient [21 CFR Part 
210.3(b)(8)]. 
One of the important aspects of GMP regulations is the quality control during 
manufacturing to ensure that the drug products are free of contamination from 
objectionable microorganisms. GMPs [1] require that the personnel engaged in drug 
product manufacture wear protective apparel such as head, face, hand, and arm 
covering (garbing) as necessary to protect drug products from contamination [Part 
211, Section 211.28(a)]. The regulations also require that the fl ow of components, 
drug product containers, closures, labeling, in - process materials, and drug products 
through the building or buildings be designed to prevent contamination [211.42(b)]. 
Contamination prevention also includes designation of separate or defi ned areas or 
such other control systems for the manufacturing operations as are necessary to 
prevent contamination [211.42(c)]. Appropriate equipment is required for adequate 
control over microorganisms [211.46(b)]. Adequate exhaust systems or other systems 
adequate to control contaminants are required where air contamination is likely to 
occur during production [211.46(c)]. Water is used in many areas of pharmaceutical 
manufacturing operations. GMPs require [211.48(a)] that incoming potable water 
into the manufacturing plant should comply with the U.S. Environmental Protection 
Agency ’ s (EPA ’ s) Primary Drinking Water Regulations (40 CFR Part 141), where 
control of microbiological contamination in potable water is defi ned. GMPs require 
written procedures be designed to prevent contamination of equipment, components, 
drug product containers, closures, packaging, and labeling materials or drug 
products and that such procedures be followed [21 CFR Part 211.56(c)]. Equipment 
cleaning and maintenance are also required to prevent contamination [21 CFR Part 
211.67(a)]. Components and drug product containers and closures should at all 
times be handled and stored in a manner to prevent contamination [21 CFR Part 
211.80(b)]. GMPs also require testing and release as described in 21 CFR Part 
211.84(d)(6), which states that each lot of component, drug product container, or 
closure that is liable to microbiological contamination that is objectionable in view 
of its intended use should be subjected to microbiological tests before use. Appropriate 
written procedures, designed to prevent objectionable microor ganisms in 
drug products not required to be sterile, should be established and followed [21 
CFR Part 211.113(a)]. GMP regulations in 21 CFR Part 211.165(b) require appropriate 
laboratory testing be performed, as necessary, of each batch of drug product 
required to be free of objectionable microorganisms. 
6.3.2.2 GMP s for Active Pharmaceutical Ingredients 
The ICH Q7A guidance for GMPs for active pharmaceutical ingredients (APIs) [6] 
illustrates GMP expectations for microbial control during API production. The ICH 

GLOBAL REGULATIONS AND REGULATORY GUIDANCES 547 
guideline specifi es that where microbiological specifi cations have been established 
for intermediates during API production, or for an API, facilities should be designed 
to limit exposure to objectionable microbiological contaminants, as appropriate 
(ICH Q7A Section 4.1). Adequate ventilation, air fi ltration, and exhaust systems 
have been suggested to minimize risk of contamination due to microorganisms 
(Q7A Section 4.2). If tighter microbiological specifi cations for water used for API 
manufacture are called for, appropriate specifi cations for total microbial counts, 
objectionable microorganisms, and/or endotoxins should be established (Q7A 
Section 4.2). The guideline also states that where the manufacturer of a nonsterile 
API either intends or claims that it is suitable for use for further processing to 
produce a sterile drug product, water used in the fi nal isolation and purifi cation 
steps for such API should be monitored and controlled for total microbial counts, 
objectionable microorganisms, and endotoxins (Q7A Section 4.3). The guideline 
recommends cleaning of equipment assigned for continuous production or campaign 
production of successive batches of the same intermediate or API at appropriate 
intervals to prevent buildup and carryover of objectionable levels of 
microorganisms (Q7A Sections 5.2 and 8.5). Under the laboratory controls section 
of the guideline, ICH recommends that if the API has a specifi cation for microbiological 
purity, appropriate action limits for total microbial counts, objectionable 
organisms, and endotoxins should be established and met (Q7A Section 11.1). 
Appropriate microbiological specifi cation tests should be conducted on each batch 
of intermediate or API where microbial quality is specifi ed (Q7A Section 11.2). The 
guidelines specify that equipment cleaning/sanitation studies should address microbiological 
and endotoxin contamination for those processes where there is a need 
to reduce total microbiological count or endotoxins in the API, where a nonsterile 
API is used to manufacture sterile products (Q7A Section 12.7). For APIs manufactured 
by cell culture, fermentation techniques, raw materials used such as media, 
and buffer components may have potential for microbiological contamination. For 
example, in such cases, depending on the intended use of the API or intermediate, 
control of bioburden, viral contamination, and/or endotoxins during manufacturing 
and monitoring of the process at appropriate stages may be necessary (Q7A Section 
18.1). When the quality of the API can be affected by microbial contamination, 
manipulations using open vessels should be performed in a biosafety cabinet or 
similarly controlled environment (Q7A Section 18.3). 
6.3.2.3 GMP s for Bulk Pharmaceutical Excipients 
The USP general chapter < 1078 > outlines the GMP guidelines for bulk pharmaceutical 
excipients [11] . The guideline states that building and facilities should be 
designed so that operations performed within do not contribute to an actual or 
potential contamination of the excipient. An effective and regular cleaning program 
of equipment used is recommended to remove product residues and dirt, which may 
also contain microorganisms and act as a source of contamination. Further, the 
guideline states that all equipment that has been in contact with contaminated material 
should be thoroughly cleaned and disinfected before coming in contact with an 
excipient. A controlled environment may be necessary to avoid microbial contamination. 
Potable water used for production of excipients should be compliant with 
chemical and microbiological standards, including freedom from pathogenic 

548 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
organisms. Purifi ed water is also used for production of pharmaceutical excipients, 
and the systems used to produce purifi ed water from potable water (deionizers, 
ultrafi ltration, or reverse - osmosis systems) may have potential for microbial growth. 
Appropriate specifi cations for chemical and microbiological quality of water used 
for pharmaceutical production and periodic testing to demonstrate compliance with 
specifi cations are recommended. When excipient product specifi cations require it 
to be endotoxin or pyrogen - free, and water is used for production of that excipient, 
validation of the purifi ed water systems to produce endotoxin and pyrogen - free 
water is required. 
6.3.3 PHARMACOPEIAL GUIDELINES RELEVANT TO 
MICROBIOLOGY OF NONSTERILE MANUFACTURING 
United States, European, and Japanese pharmacopeia have described general 
requirements, specifi cations, and tests for monitoring microbial bioburden in 
nonsterile pharmaceutical products. Although there are minor differences in the 
expectations of the pharmacopeia, the general principles for microbial bioburden 
monitoring remain similar as described below. 
6.3.3.1 United States Pharmacopeia 
In the general chapter “ Microbiological Examination of Nonsterile Products ” [3] , 
the USP provides guidance for microbial examination of various nonsterile pharmaceutical 
dosage forms (Table 1 ). Enumeration of the total aerobic counts and 
total yeasts and molds in products are determined using procedures stated in USP 
general chapter < 61 > [12] , and tests for specifi ed microorganisms ( Staphylococcus 
aureus , and Pseudomonas aeruginosa, Escherichia coli, Salmonella sp., and Candida 
albicans ) are performed using procedures stated in USP general chapter < 62 > [13] . 
Acceptance criteria for microbiological quality of nonsterile dosage forms for 
various routes of administration are also provided. In addition, in pharmacopeial 
monographs, microbiological specifi cations are described where applicable in monographs 
for excipients, actives, and drug products. Microbiological best laboratory 
practices are described in general chapter < 1117 > [14] , which provides guidance on 
implementing good laboratory practice standards when examining various sample 
matrixes for microbial bioburden in the laboratory. Alternative microbiological 
testing methods other than those specifi ed in USP can be used provided such 
methods are appropriately validated. The procedures for validation of alternative 
microbiological methods are described in USP general chapter < 1223 > [15] . 
The USP also provides guidance on reduced microbial limits testing for product 
release and stability evaluation when drug products have reduced water activity 
[ratio of vapor pressure of H 2 O in product ( P ) to vapor pressure of pure H 2 O ( Po ) 
at the same temperature] well below 0.75. Water is required for microbial growth 
in the pharmaceutical products [16] . However, the more resistant microorganisms, 
including spore - forming Clostridium sp., Bacillus sp., Salmonella sp., and fi lamentous 
fungi, which may not proliferate in the drug product with low water activity, 
may persist in the dormant state in the product. For example, a water activity of 
0.61 is required for the growth of the fungus Xeromuces bisporus , a water activity 

of 0.62 is required for the growth of the yeast Zygosaccharomyces rouxii , while the 
majority of other fungi and yeasts require water activity higher than 0.75 for growth. 
Water activity of 0.75 is adequate for the growth of the bacterium Halobacterium 
halobium , but the majority of other bacteria require higher water activity for growth. 
The guidance also provides for a microbial limit testing strategy for pharmaceutical 
products for various routes of administration based on water activity. For example, 
for compressed tablets and liquid - fi lled capsules with representative water activities 
of 0.36 and 0.30, respectively, reduced testing is recommended, while for topical 
creams and nasal inhalants, with water activity of 0.97 and 0.99, respectively, testing 
for total aerobic counts, total yeasts and molds, S. aureus , and P. aeruginosa is recommended. 
Antimicrobial preservatives are added to nonsterile pharmaceutical dosage 
forms to protect them from microbial growth or from microorganisms that are 
introduced inadvertently during or subsequent to the manufacturing process. USP 
general chapter < 51 > describes how to perform antimicrobial effectiveness testing 
[17] . In this chapter, it is emphasized that antimicrobial preservatives should not be 
used as a substitute to GMPs or solely to reduce the viable microbial population of 
a nonsterile product. It is recommended that antimicrobial effectiveness of preservatives 
must be demonstrated for dosage forms such as multiple - dose topical and 
TABLE 1 Acceptance Criteria for Microbial Bioburden Control of Nonsterile Dosage 
Forms 
Route of 
Administration 
Total Aerobic 
Count Limit 
(CFU/g, CFU/mL) 
Total Combined 
Yeasts and Molds 
Limit (CFU/g, 
CFU/mL) 
Specifi ed 
Microorganism/s 
(in 1 g or 1 mL) 
Nonaqueous 
preparations for 
oral use 
1000 100 Absence of E. coli 
Aqueous 
preparations for 
oral use 
100 10 Absence of E. coli 
Rectal use 1000 100 
Oromucosal use 
(gingival, 
cutaneous, nasal, 
auricular) 
100 10 Absence of S. aureus 
and P. aeruginosa 
Vaginal use 100 10 Absence of S. aureus, 
P. aeruginosa , and 
C. albicans 
Transdermal patches 
(one patch 
including adhesive 
layer and backing) 
100 10 Absence of S. aureus 
and P. aeruginosa 
Inhalation use (other 
than liquid 
preparations for 
nebulization) 
100 10 Absence of S. aureus, 
P. aeruginosa , and 
bile - tolerant gram - 
negative bacteria 
Source : From ref. 3 . 
PHARMACOPEIAL GUIDELINES RELEVANT TO MICROBIOLOGY 549

550 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
oral dosage forms, ophthalmic, otic, nasal, and irrigation as described in the USP 
general chapter on pharmaceutical dosage forms [18] . 
The USP also provides guidance on water for pharmaceutical purposes [19] . The 
guidance points out that the major exogenous source of microbial contamination of 
bulk pharmaceuticals is the source water. Source water (potable water) should meet 
quality attributes of drinking water, in which coliform levels are controlled. Other 
microorganisms present in the incoming water, although not objectionable in nature, 
may compromise subsequent purifi cation steps. Microorganisms present in source 
water may absorb to carbon beds, deionized resin beds, and fi lter membranes (used 
in the processing of potable water to purifi ed water) and initiate the formation of 
a biofi lm. Microorganisms in a biofi lm are adapted to the prevailing low - nutrient 
environment. Microorganisms in the biofi lm may also move downstream when they 
are shed from the existing biofi lm and carried to other areas of the water system. 
A detailed account of biofi lm formation in the pipelines and methods for detection 
and quantitation of biofi lms is provided by Olson (1997) [20] . Microbial contamination 
may also come from unprotected vents, faulty air fi lters, ruptured disks in 
the water system, and backfl ow from contaminated outlets, among others. Several 
categories of microorganisms may be problematic to the pharmaceutical water 
systems and as a result to the pharmaceutical products, which are manufactured 
using water from these systems. The microorganisms may include opportunistic or 
overt pathogens, nonpathogenic indicators of potentially undetected pathogens, or 
microorganisms that could be resistant to a preservative in a drug product or which 
can degrade an active or inactive ingredient in a drug product. Defi ning alert and 
action levels of microorganisms in water systems and monitoring these levels by 
microbiological testing at scheduled intervals can serve as an early warning that will 
allow remedial actions to occur and prevent a system from producing water unfi t 
for pharmaceutical use. For purifi ed water, for example, maximum action levels 
are considered to be 100 colony - forming units of microorganisms per milliliter 
(CFUs/mL) [19] . 
6.3.3.2 European Pharmacopoeia 
For pharmaceutical products, the European Pharmacopoeia describes sampling 
plans and tests for quantitative enumeration of bacteria and fungi that can grow 
under aerobic conditions [21] . Enterobacteria and certain other gram - negative 
bacteria, E. coli, Salmonella sp., P. aeruginosa , and S. aureus and Clostridia and counts 
for Clostridium perfringens [22] are described under tests for specifi ed microorganisms. 
In the general chapter on microbiological quality of pharmaceutical preparations 
[23] , pharmaceutical dosage forms are classifi ed into three categories: Category 
1 is for sterile pharmaceutical preparations, which is not the subject of this chapter, 
and categories 2 and 3 describe requirements for various nonsterile dosage forms. 
For example, for category 2, preparations for topical use, testing for total aerobic 
count and fungi, enterobacteria and other gram - negative bacteria, P. aeruginosa , and 
S. aureus is performed. Under category 3A, preparations for oral rectal administration, 
testing for total viable aerobic count and E. coli is performed, whereas under 
category 3B, where raw materials of natural origin are included in the drug products, 
testing for total viable aerobic count, nitrobacteria and other gram - negative bacteria, 
Salmonella sp., E. coli , and S. aureus is performed. Microbiological specifi cations 

are detailed wherever applicable under each monograph. Antimicrobial preservative 
effectiveness tests are recommended, where antimicrobial preservatives are 
used, such as for multidose containers where there is a potential for microbiological 
contamination, or aqueous or topical preparations, which also have potential for 
microbial contamination with high moisture content in these dosage forms. Under 
the general chapter on effi cacy of antimicrobial preservation [24] , both the rationale 
for using antimicrobial preservatives and the tests for demonstrating effectiveness 
of antimicrobial preservatives are described. 
6.3.3.3 Japanese Pharmacopoeia 
In the general chapter on microbial attributes of nonsterile pharmaceutical products, 
the guidance suggests that the presence of microbial contaminants in nonsterile 
products [25] can reduce or inactivate the therapeutic activity of the product and 
has the potential to adversely effect the health of the patients and recommends 
manufacturers to ensure that contamination levels are as low as possible for fi nished 
dosage forms. Microbial enumeration limits for raw materials (total aerobic microbial 
count and total combined yeasts and molds count) and fi nished dosage forms 
are described. For inhalation, nasal, and topical routes of administration, tests 
for total aerobic microbial count and total combined and yeast and mold count, 
P. aeruginosa , and S. aureus are recommended. For vaginal preparations, testing for 
total aerobic microbial count and total combined yeast and mold count, E. coli, S. 
aureus , and C. albicans is recommended, while for oral liquids and solids, testing for 
total aerobic microbial count and total combined and yeast and mold count and 
E. coli is recommended. The microbiological assessment of preservatives is required 
when preservatives are used in a pharmaceutical product to control microbial bioburden. 
The test microorganisms and methods for evaluating the effi cacy of the 
preservative in pharmaceutical products are described in the general chapter on 
preservative effectiveness tests [26] . 
6.3.4 CURRENT REGULATORY EXPECTATIONS FOR MICROBIAL 
BIOBURDEN CONTROL DURING NONSTERILE MANUFACTURING 
Regulatory expectations for microbial bioburden for nonsterile pharmaceutical 
products are reviewed using the FDA guide to inspections of microbiological quality 
control laboratories [2] , purifi ed water systems [27] , topical products [28] , and oral 
solutions and suspensions [29] . 
In the guide to inspections of microbiological quality control laboratories [2] , a 
number of problems associated with microbiological contamination of topical drug 
products, nasal solutions, and inhalation products have been emphasized. In that 
guidance, Dunnigan of the FDA is quoted as saying that topical preparations contaminated 
with gram - negative organisms are a probable moderate to serious health 
hazard. 
In the guide to inspections of topical products [28] , it is indicated that water 
deionizers are usually excellent breeding areas for microorganisms, where fl ow rates, 
temperature, surface area of resin beds, and microbial quality of the feed water 
all infl uence microbial growth. Since topical products (e.g., creams, ointments) 
CURRENT REGULATORY EXPECTATIONS 551

552 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
generally contain purifi ed water, the microbial integrity of the product is contingent 
on the microbial bioburden of purifi ed water. The guidance also points out that in 
assessing the signifi cance of microbial contamination of the product, both the identifi 
cation of isolated microorganisms and the number of organisms found are signifi 
cant. When high numbers of nonpathogenic microorganisms are present, they 
may affect product effi cacy and/or physical and chemical stability. Topical creams 
may tend to separate on standing. For topical products, data should be available to 
show con tinued effectiveness of preservatives throughout the product ’ s shelf life. 
Purifi ed water is normally used for the manufacture of nonsterile pharmaceutical 
products. The guide to inspections of high - purity water systems [27] primarily discusses 
microbiological aspects and suggests that an action limit of over 100 CFU/mL 
for a purifi ed water system is unacceptable. The purpose of establishing an action 
limit is to assure that the water system is under control. Purifi ed water used in the 
manufacture of drug products should be free of “ objectionable microorganisms, ” 
which are defi ned as organisms that when present can cause infections when the 
drug product is used as directed. The defi nition can also include organisms capable 
of growth in the drug product. Microorganisms may exist in water systems as either 
free fl oating or attached to the walls of pipes and tanks (biofi lms). Microorganisms 
continuously slough off from biofi lms and contribute to movement of microorganisms 
upstream or downstream in the water systems. The guidance emphasizes that 
establishing the level of contamination allowed in high - purity water systems used 
in the manufacture of a nonsterile product requires an understanding of the use of 
the product, the formulation (preservative system), and the manufacturing process. 
In the manufacture of antacids, which do not have an effective preservative system, 
an action limit below 100 CFU/mL is required. The inspection guide also points out 
that equipment used in purifi ed water systems may be liable to contamination. 
Pumps used in water systems, if not continuously used, may have still (stagnant) 
water, which is the site of microbial growth. Pseudomonas sp. contamination reported 
in the water system was attributed to a pump which was operational only periodically. 
Deadlegs and fi ttings are a source of microbial bioburden in water systems. 
Threaded fi ttings should not be used in pharmaceutical water systems, as they are 
a source of microbial growth. Filters are another source of contamination, as 
microbes tend to accumulate and grow on their surface. While a 0.2 - . m point - of - use 
fi lter can mask the level of microbiological contamination, it will not necessarily 
stop endotoxin contamination, since the toxins can pass through the fi lter. Filters 
should be frequently changed to prevent contamination. 
In a water system, contamination by Pseudomonas sp. was found after FDA 
testing and the same species was also found in a topical steroid product after FDA 
testing [27] . The inspection resulted in product recall and issuance of a warning 
letter. The root cause for microbial bioburden was found to be a one - way water 
system that employ ultraviolet (UV) light to control microbiological contamination, 
where the UV light is turned on only when water is needed, and when the UV lights 
are not on, sterilization is ineffective and the standing water is a source of microbial 
growth. A fl exible hose used in this water system that was diffi cult to sanitize may 
also have contributed to the observed microbial bioburden. 
In the guide to inspection of oral solutions and suspensions [29] , it is stated 
that in some oral liquids microbiological contamination can present signifi cant 
health hazard. For instance, microbiological contamination with gram - negative 

microorganisms is objectionable in some oral liquids, especially those used in infants 
and immunocompromised patients. In oral liquids such as antacids, contamination 
with Pseudomonas sp. is objectionable. In general, contamination of any preparations 
with gram - negative organisms is not desirable. Such contamination may suggest 
a defi cient manufacturing process, inadequate preservative system, and potential 
use of contaminated raw materials. 
6.3.5 INDUSTRY PERSPECTIVE ON MICROBIAL BIOBURDEN 
CONTROL FOR NONSTERILE PHARMACEUTICAL MANUFACTURING 
The PhRMA Environmental Monitoring Work Group published an article in the 
March 1997 issue of Pharmaceutical Technology on microbiological monitoring of 
environmental conditions for nonsterile manufacturing [4] . This publication is a 
compilation of survey results of nonsterile manufacturing facilities within the United 
States and recommendations based on the survey results. In this section, a summary 
of the survey results and recommendations published in this article are reviewed 
with a few comments from this author based on experiences in this area. 
6.3.5.1 Survey Results 
Most manufacturers conducted the monitoring of nonsterile pharmaceutical manufacturing 
environments. Companies producing topicals, liquids, and aerosols tend to 
have more extensive programs than those companies manufacturing tablets or other 
solid oral dosage forms. Among the nonsterile manufacturing environments monitored 
were air, processing equipment, and water. Air quality was monitored using 
centrifugal air samplers, settling plates, and so on. Product contact surfaces of equipment 
were monitored using swabs or contact plates, with swabs being preferred 
because of better access of irregular surfaces. Since swabs used are presterilized, 
they are less likely to contaminate the surface being sampled, whereas contact plates 
which contain nutrient agar media may encourage growth of microorganisms after 
sampling, unless the sampled area is thoroughly cleaned with ethanol or other sterilizing 
agent. However, contact plates are easier to use on plain fl at surfaces and also 
help provide an accurate assessment of in situ microbial status. 
Companies with monitoring programs have established alert and action limits. 
Alert levels of microorganisms may indicate a change from normal operating procedures 
but may not require any corrective action, whereas action levels may suggest 
a signifi cant change from normal operating procedures and require corrective 
actions to control further proliferation of microbes. 
6.3.5.2 Recommendations 
Specifi c Recommendations The work group recommended that the routine 
monitoring of microbiological activity in nonsterile environments should not be 
mandatory but should be determined by the type of nonsterile dosage form manufactured. 
In general, cleaning and sanitization procedures, good maintenance schedules, 
well - defi ned process control programs, raw material quality, facility design and 
control, and training of personnel involved in various aspects of manufacturing can 
INDUSTRY PERSPECTIVE ON MICROBIAL BIOBURDEN CONTROL 553

554 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
contribute to microbial bioburden control. A defi ned monitoring program can 
provide data on historical trends and such data can help identify deviations, root - 
cause analysis, troubleshooting, and early and effective actions. Among nonsterile 
dosage forms, inhalation products are of particular concern, followed by liquids and 
topical formulations, with a lower priority given to solid oral dosage forms. When 
monitoring programs are instituted, sampling should include those areas that are 
most likely to be contaminated, such as product contact surfaces of processing 
equipment, ventilation systems, process gases and purifi ed water, and water systems. 
Sampling frequency can be based on historical trends and types of dosage forms 
manufactured, focusing in particular on those products more likely to be susceptible 
to microbial contamination. Manufacturing process, cleaning, and utilities validations 
should include microbial sampling to ensure microbial bioburden control. 
Holding times for process steps such as coating solutions and wet granulation should 
be determined by sampling and testing these matrices for microbial bioburden 
during validation and subsequent manufacturing. 
General Recommendations Complying with GMP regulations specifi ed in 21 CFR 
Part 211 [1] is the most effective way of controlling microbial bioburden in a nonsterile 
manufacturing environment as (as described under Section 6.3.2.1 ). Appropriate 
written procedures should be established to control microbial bioburden and 
prevent objectionable microorganisms. Microbiologists should be well trained in 
sampling and testing of the pharmaceutical manufacturing environment, so that 
microbial bioburden is not introduced into product or environmental samples during 
sampling and testing. Identifi cation of isolates should be undertaken as applicable 
when action levels (or above acceptance criteria) are seen during quantitation of 
microbial bioburden. 
6.3.6 MICROBIAL BIOBURDEN CONTROL DURING SHELF LIFE OF 
PHARMACEUTICAL PRODUCTS 
GMPs require that the stability of pharmaceutical products be tested to evaluate 
the product integrity throughout the shelf life of the product [1, 6, 11] . The ICH 
guidance Q1A (R2), stability testing of drug substances and drug products [30] , 
emphasizes that the stability testing should cover, as appropriate, the physical, 
chemical, biological, and microbiological attributes. In addition, for drug products, 
if antimicrobial preservatives are used, preservative content should also be 
investigated. 
6.3.7 SUMMARY AND CONCLUSIONS 
In nonsterile pharmaceutical manufacturing, control of microbial bioburden through 
appropriate design of manufacturing facilities and implementation of GMPs 
during manufacturing would help control microbial bioburden in pharmaceutical 
raw materials and fi nished products. The GMPs for fi nished product and its components 
(active and inactive ingredients) are discussed in this chapter to highlight 
expectations for contamination control during manufacturing. In support of GMP 

expectations for contamination control and testing required to establish such control, 
pharmacopeia in the United States, Europe, and Japan provide test methods and 
procedures for determination of microbial bioburden, which are also summarized 
in this chapter. Manufacturers of pharmaceutical products through their trade associations 
have taken a proactive approach to microbial bioburden control in nonsterile 
manufacturing by publishing results from surveys of member companies and 
have made general recommendations for implementation of microbial bioburden 
control by member companies. A brief summary of the industry work group survey 
results on current industry practices and general recommendations for implementation 
is presented. The requirement for microbial bioburden monitoring throughout 
the shelf life of the products is also emphasized. 
REFERENCES 
1. U.S. Food and Drug Administration (FDA) ( 2007 ), Current good manufacturing practice 
regulations, 21 CFR Parts 210 and 211, FDA, Rockville, MD. 
2. U.S. Food and Drug Administration (FDA) ( 1993 ), Guide to inspections of microbiological 
pharmaceutical quality control laboratories, FDA, Rockville, MD. 
3. U.S. Pharmacopeia (USP) ( 2007 ), < 1111 > Microbiological examination of nonsterile 
products: Acceptance criteria for pharmaceutical preparations and substances for pharmaceutical 
use, U.S. Pharmacopeial Convention, Rockville, MD. 
4. Pharmaceutical Research and Manufacturers Association of America (PhRMA) ( 1997, 
March ), Microbiological monitoring of environmental conditions for nonsterile manufacturing 
, Pharm. Technol. , 58 – 74 . 
5. Australian Pharmaceutical Manufacturers ’ Association (APMA) ( 1990 ), The Control of 
Microbial Contamination in Nonsterile Pharmaceutical Products for Human Use , APMA , 
Canberra, Australia. 
6. International Conference on Harmonization (ICH) ( 2001 ), Guidance for industry. Q7A 
good manufacturing practice guidance for active pharmaceutical ingredients, ICH, 
Brussels, Belgium. 
7. Therapeutic Goods Administration Laboratories of Australia (TGAL) ( 1990 ), Guidelines 
for Assessing the Results of Microbiological Tests on Nonsterile Pharmaceuticals for 
Human Use , TGAL , Canberra, Australia . 
8. Tang , S. ( 1998 ), Microbial limits reviewed — Basis for unique Australian regulatory 
requirements for microbial quality of nonsterile pharmaceuticals, PDA , J. Pharm. Sci. 
Technol ., 52 , 100 – 109 . 
9. World Health Organization (WHO) ( 2003 ), Good manufacturing practices for pharmaceutical 
products, main principles, WHO Technical Report Series No. 908, WHO, Geneva, 
Switzerland. 
10. International Pharmaceutical Excipients Council (IPEC) ( 2006 ), Good Manufacturing 
Practices , IPEC , Washington DC . 
11. U.S. Pharmacopeia (USP) ( 2006 ), Good manufacturing practices for bulk pharmaceutical 
excipients, general chapter < 1078 > , U.S. Pharmacopeial Convention, Rockville, MD. 
12. U.S. Pharmacopeia (USP) ( 2007 ), Microbiological examination of nonsterile products: 
Microbial enumeration tests < 61 > , U.S. Pharmacopeial Convention, Rockville, MD. 
13. U.S. Pharmacopeia (USP) ( 2007 ), Microbiological examination of nonsterile products: 
Tests for specifi ed microorganisms < 62 > , U.S. Pharmacopeial Convention, Rockville, 
MD. 
REFERENCES 555

556 MICROBIOLOGY OF NONSTERILE PHARMACEUTICAL MANUFACTURING 
14. U.S. Pharmacopeia (USP) ( 2007 ), Microbiological best laboratory practices, general 
chapter < 1117 > , U.S. Pharmacopeial Convention, Rockville, MD. 
15. U.S. Pharmacopeia (USP) ( 2007 ), Validation of alternative microbiological methods, 
general chapter < 1223 > , U.S. Pharmacopeial Convention, Rockville, MD. 
16. U.S. Pharmacopeia (USP) ( 2007 ), Application of water activity determination to nonsterile 
pharmaceutical products, general chapter < 1112 > , U.S. Pharmacopeial Convention, 
Rockville, MD. 
17. U.S. Pharmacopeia (USP) ( 2007 ), Antimicrobial effectiveness testing, general chapter 
< 51 > , U.S. Pharmacopeial Convention, Rockville, MD. 
18. U.S. Pharmacopeia (USP) ( 2007 ), Pharmaceutical dosage forms, general chapter < 1151 > , 
U.S. Pharmacopeial Convention, Rockville, MD. 
19. U.S. Pharmacopeia (USP) ( 2007 ), Water for pharmaceutical purposes, general chapter 
< 1231 > , U.S. Pharmacopeial Convention, Rockville, MD. 
20. Olson , P. W. ( 1997 ), Biofi lms in the pipeline and in the patient , J. Pharm. Sci. Technol ., 
51 ( 6 ), 252 – 261 . 
21. European Pharmacopoeia (Eur. Ph.) ( 2005 ), Microbiological examination of nonsterile 
products (total viable aerobic count), general chapter 2.6.12, Council of Europe, 
Strasbourg Cedex, France. 
22. European Pharmacopoeia (Eur. Ph.) ( 2005 ), Microbiological examination of nonsterile 
products (tests for specifi ed microorganisms), general chapter 2.6.13, Council of Europe, 
Strasbourg Cedex, France. 
23. European Pharmacopoeia (Eur. Ph.) ( 2005 ), Microbiological quality of pharmaceutical 
preparations, general chapter 5.1.4, Council of Europe, Strasbourg Cedex, France. 
24. European Pharmacopoeia (Eur. Ph.) ( 2005 ), Effi cacy of antimicrobial preservation, 
general chapter 5.1.3, Council of Europe, Strasbourg Cedex, France. 
25. Japanese Pharmacopoeia (JP) ( 2001 ), Microbiological attributes of nonsterile pharmaceutical 
products, general information chapter 7, Society of Japanese Pharmacopoeia, 
Tokyo. 
26. Japanese Pharmacopoeia (JP) ( 2001 ), Preservatives — Effectiveness tests, general information 
chapter 12, Society of Japanese Pharmacopoeia, Tokyo. 
27. U.S. Food and Drug Administration (FDA) ( 1993 ), Guide to inspections of high purity 
water systems, FDA, Rockville, MD. 
28. U.S. Food and Drug Administration (FDA) ( 2000 ), Guide to inspections of topical drug 
products, FDA, Rockville, MD. 
29. U.S. Food and Drug Administration (FDA) ( 1994 ), Guide to inspection of oral solutions 
and suspensions, FDA, Rockville, MD. 
30. International Conference on Harmonization (ICH) ( 2003 ), Guidance for industry, Q1A 
(R2) stability testing of new drug substances and products, ICH, Brussels, Belgium. 

DRUG STABILITY 
SECTION 7


559 
7.1 
STABILITY AND SHELF LIFE OF 
PHARMACEUTICAL PRODUCTS 
Ranga Velagaleti 
BASF Corporation, Florham Park, New Jersey 
Contents 
7.1.1 Introduction 
7.1.2 Stability Requirements in GMP Regulations and Guidelines 
7.1.2.1 Finished Products 
7.1.2.2 Excipients 
7.1.2.3 Active Pharmaceutical Ingredients 
7.1.3 Stability Requirements for Excipients 
7.1.4 Stability Requirements for Drug Substances (APIs) 
7.1.4.1 Selection of Batches and Container Closure System 
7.1.4.2 Storage Conditions and Testing Frequency 
7.1.4.3 Stress Studies and Stability - Indicating Methods for Analysis of API 
Stability 
7.1.4.4 Evaluation of Stability Results 
7.1.4.5 Stability Commitment 
7.1.4.6 Storage Statement and Labeling 
7.1.5 Stability Requirements for Drug Products 
7.1.5.1 Selection of Batches and Container Closure System 
7.1.5.2 Bracketing and Matrixing 
7.1.5.3 Storage Conditions and Testing Frequency 
7.1.5.4 Stress Studies and Stability - Indicating Methods for Analysis of Drug Product 
Stability 
7.1.5.5 Evaluation of Stability Results 
7.1.5.6 Stability Commitment 
7.1.5.7 Storage Statement and Labeling 
7.1.6 Photostability Studies 
7.1.6.1 Photostability of APIs 
7.1.6.2 Photostability of Drug Products 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

560 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
7.1.7 Evaluation of Stability Data and Shelf Life Determination 
7.1.7.1 Shelf Life Estimation for Drug Substances or Drug Products Intended for 
Room Temperature Storage 
7.1.7.2 Shelf Life Estimation for Drug Substances or Drug Products Intended for 
Storage in a Refrigerator 
7.1.7.3 Shelf Life Estimation for Drug Substances or Drug Products Intended for 
Storage in a Freezer 
7.1.7.4 Shelf Life Estimation for Drug Substances or Drug Products Intended for 
Storage below . 20 ° C 
7.1.8 Assignment of Climatic Zones and Recommended Storage Conditions 
7.1.9 Stability Testing Parameters for Different Dosage Forms 
7.1.9.1 Stability Tests Common to All Dosage Forms 
7.1.9.2 Stability Tests Specifi c to Dosage Forms 
7.1.10 Summary and Conclusions 
References 
7.1.1 INTRODUCTION 
Pharmaceutical drug products (e.g., tablets, capsules, creams, injectables, and inhalation 
products) and the components in a drug product, that is, the active pharmaceutical 
ingredient (API, drug substance), and the inactive ingredients (excipients) should 
be stable in the drug product for the proposed shelf life duration of the drug product 
and the proposed duration of the shelf life of the individual components. Shelf life 
(retest or expiration date) is the time period during which excipients, APIs, and drug 
products are expected to remain within approved shelf life specifi cation, provided 
that they are stored under the conditions defi ned on the container label. Stability 
and storage conditions for excipients, APIs, and drug products are determined by 
evaluation of their quality parameters with time under the infl uence of a variety of 
environmental factors such as temperature, humidity, and light. 
Stability applies to chemical, physical, microbiological, therapeutic, and toxicological 
properties. The U.S. Pharmacopeia (USP) [1] defi ned stability as the extent 
to which a product retains, within specifi ed limits and throughout its period of 
storage and use (i.e., its shelf life), the same properties and characteristics that it 
possessed at the time of manufacture. For example, chemical stability implies that 
each active ingredient in a drug product retains its chemical integrity and label 
potency within the specifi ed limits during the shelf life. Physical stability is assured 
if the original physical properties, including appearance, palatability, uniformity, 
dissolution, and suspendability, are retained during the shelf life. Microbiological 
stability is demonstrated if the sterility or resistance to microbial growth is retained 
according to the specifi ed requirements and antimicrobial agents that are present 
retain effectiveness within the specifi ed limits. The therapeutic effect of the drug 
should remain unchanged and no signifi cant increase in toxicity should occur [1] 
during the shelf life if the product is stable. 
Good manufacturing practice (GMP) regulations and guidelines allude to the 
importance of stability and shelf life of pharmaceutical products. The U.S. Food and 

Drug Administration ’ s (FDA ’ s) GMP regulations for fi nished products require stability 
testing of drug products to assess the stability characteristics, the results of 
which are used for determining appropriate storage conditions and expiration dates 
[2] . The FDA ’ s GMP guidance for APIs requires that a documented, on - going testing 
program should be established to monitor the stability characteristics of APIs and 
the results should be used to confi rm the appropriate storage conditions and retest 
or expiry dates [3] . The USP guidance on GMPs for bulk pharmaceutical excipients 
suggests that there should be a documented testing program designed to assess the 
stability characteristics of excipients and the results of such stability testing should 
be used in determining appropriate storage conditions and their reevaluation 
(retest) or expiration date [4] . The GMP guideline for pharmaceutical products from 
the World Health Organization (WHO) [5] alludes to the stability of fi nished pharmaceutical 
products and starting materials. 
In addition to GMP requirements, several guidelines are published specifi cally 
to address pharmaceutical stability. Organizations with international mandates such 
as the WHO [6] and International Conference on Harmonization [ICH, comprising 
principally the European Union (EU), Japan, and the United States] [7] have published 
stability guidelines. The stability guidance of the WHO [6] defi nes stability 
data package requirements for pharmaceutical products. In addition to addressing 
APIs and drug products, the WHO guidance also emphasizes that the stability of 
excipients that may contain or form reactive degradation products should be considered 
for stability testing. The ICH Q1A (R2) guidance published by the FDA [7] 
states that the purpose of stability testing is to provide evidence on how the quality 
of a drug substance or drug product varies with time under the infl uence of a variety 
of environmental factors such as temperature, humidity, and light and to establish 
a retest period for the drug substance or a shelf life for the drug product and recommend 
storage conditions. The ICH has published a comprehensive series of stability 
guidances covering different aspects of stability: for example, ICH Q1A (R2), Stability 
Testing of New Drug Substances and Products [7] ; ICH Q1B, Photostability 
Testing of New Drug Substances and Products [8] ; ICH Q1C, Stability Testing of 
New Dosage Forms [9] ; ICH Q1D, Bracketing and Matrixing Designs for Stability 
Testing of New Drug Substances and Products [10] ; and ICH Q1E, Evaluation of 
Stability Data [11] . Some of these guidelines have been adopted and published by 
the FDA (see References). 
In this chapter, the stability and shelf life of pharmaceutical excipients, APIs, and 
fi nished drug products are reviewed, with information derived from GMP regulations 
and guidelines as well as technical and regulatory guidance documents on 
stability. The importance of photostability studies for stability - indicating method 
development and for appropriate container closure selection to assure stability of 
the marketed product through shelf life is discussed. The evaluation and assessment 
of stability results for the determination of shelf life are discussed. The use of validated 
stability - indicating analytical methods for assessment of the potency of drug 
substance and the amount of degradation products and impurities through the shelf 
life is emphasized. Recommended storage conditions for various climatic zones are 
discussed. The testing parameters/attributes for evaluation of stability of various 
drug dosage forms are outlined. (The mechanisms and pathways of degradation of 
pharmaceutical products are not discussed in this chapter.) 
INTRODUCTION 561

562 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
7.1.2 STABILITY REQUIREMENTS IN GMP 
REGULATIONS/GUIDELINES 
7.1.2.1 Finished Products 
GMPs require assessment of stability characteristics of the pharmaceutical drug 
product, and the principles of such assessment are outlined in 21 CFR Part 211, 
Section 211.166 (a)(b) [2] , and summarized in this section. The drug product stability 
samples should be stored in the same container closure as that in which the drug 
product is marketed. Appropriate storage conditions for samples designated for 
stability evaluation and reliable, specifi c, and meaningful test methods should be 
used. If a drug product is reconstituted at the time of dispensing as directed on the 
label, testing of the material for compliance with stability specifi cations is required 
pre - and post - reconstitution. Results from stability testing should be used to determine 
the appropriate storage conditions and expiration dates. The product stability 
estimates should be valid based on statistical evaluation of each stability attribute. 
Shelf life, expiration date, and storage conditions assigned for a drug product should 
be supported by long - term stability storage and testing of an adequate number of 
batches. When drug applications are submitted to the FDA, long - term storage stability 
studies may be in progress and data may not be available for the full proposed 
shelf life duration to support shelf life, expiration dating, or storage conditions. In 
such cases, tentative expiration dates can be established based on the accelerated 
stability studies, which are designed to increase the rate of chemical degradation or 
physical change by using exaggerated temperature and humidity conditions. In addition, 
other basic stability information on known physical and chemical stability 
attributes of excipients and APIs and container closure systems is also relevant 
for shelf life determination. Tentative expiration dates thus established must be 
supported by long - term storage stability studies. 
7.1.2.2 Excipients 
The GMPs for bulk pharmaceutical excipients [4] emphasize that the stability of 
excipients is an important contributing factor to the stability of the fi nished dosage 
form. Changes in raw materials used or changes in manufacturing procedures of 
excipients may affect their stability. In addition, the container closure packaging for 
excipients may vary widely to include but not be limited to metal and plastic drums, 
plastic bottles, and tank cars, which can affect the stability of the product inside. The 
stability characteristics of the excipient should be assessed using appropriate storage 
conditions and storage intervals and the testing should be performed to determine 
compliance with the established stability specifi cations. Results of such testing 
should be used to determine appropriate storage conditions and reevaluation or 
retest dates. Samples should be examined at or before retest date to ensure that the 
material is still in compliance with the specifi cation and thus is suitable for its 
intended use. The excipient stability testing program should be ongoing and should 
be documented through a stability protocol that will include number of lots tested 
per year, sample size, test intervals, storage conditions (e.g., temperature, humidity), 
and test methods that are stability indicating. The materials used for container closures 
for stability sample storage (if the containers and closures are smaller than 

STABILITY REQUIREMENTS FOR EXCIPIENTS 563 
those used for the marketed product) should be similar to the packaging materials 
used for marketed products. Storage time and conditions should support the conditions 
that are likely to prevail during the designated shelf life of the product. A 
model approach can be used for excipients that are available in different grades 
such as various molecular weights of a polymer or different monomer ratios or 
mixtures or blends of other excipients. Stability data available on model products 
can be used to determine the theoretical stability for similar products. In addition, 
for excipients that have been on the market for a long time, historical data may be 
used to assign the shelf life, retest dates, and storage conditions [4] . 
7.1.2.3 Active Pharmaceutical Ingredients 
GMPs for APIs [3] require a documented stability storage and testing program to 
monitor the stability characteristics of APIs against established stability specifi cations. 
The results from stability studies are used to establish storage conditions and 
retest or expiry dates. For APIs, retest rather than expiration date terminology is 
commonly used. The GMP guideline also emphasizes that the assay procedures (for 
drug substance and degradation products) in stability samples should be validated 
and it should be stability indicating. A stability - indicating assay is defi ned as “ a validated 
quantitative analytical procedure that can detect the changes with time in the 
pertinent properties of the drug substance and measures the active ingredient accurately 
without interference from degradation products, process impurities or other 
potential impurities ” [ 12 ; Section III C ] . 
The GMP API guideline also emphasizes that stability samples should be stored 
in containers and closures (the sum of packaging components that together contain 
and protect API or drug product dosage forms) that simulate the market container 
and should be consistent with storage conditions specifi ed in the FDA stability guidance 
[7] . The GMP guideline requires that the fi rst three commercial production 
batches should be placed on the stability program to confi rm the retest or expiry 
date. If the API is expected to remain stable for at least two years, fewer than three 
batches are allowed. The stability program should be ongoing with at least one batch 
of API manufactured each year tested annually to confi rm stability, with testing 
done more frequently for APIs with shorter shelf lives. 
The WHO GMPs for pharmaceutical products [5] require that quality control 
evaluate the stability of fi nished pharmaceutical products and, when necessary, of 
starting materials and intermediate products. The guideline requires that a written 
stability program be developed and implemented to generate the stability data. 
Expiration dates and shelf life specifi cations should be based on stability data generated 
under defi ned storage conditions. The WHO guideline also specifi es that stability 
should be determined prior to marketing and following any signifi cant change 
in, for example, process, equipment, and packaging materials. 
7.1.3 STABILITY REQUIREMENTS FOR EXCIPIENTS 
In the absence of specifi c stability guidance from the FDA and ICH (and limited 
guidance from the WHO) for excipients, the USP GMP guidance for bulk pharmaceutical 
excipients [4] is used to review stability requirements for excipients. A 

564 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
description of stability requirements for excipients is provided in Section 7.1.2.2 . 
The principles underlying the evaluation of stability data and determination of shelf 
life (retest dates) for excipients is similar to that described for APIs and drug 
products in Section 7.1.7 . 
7.1.4 STABILITY REQUIREMENTS FOR DRUG SUBSTANCES ( API s ) 
The guidance for industry Q1A (R2) [7] defi nes requirements for API stability data 
package submission with drug applications in the EU, Japan, and the United 
States. 
7.1.4.1 Selection of Batches and Container Closure System 
Stability data should be generated on at least three primary batches, which should 
be manufactured to a minimum of pilot scale by the same synthetic route and 
manufacturing process as the production batches. The quality of the API placed on 
a formal stability program should be similar to the quality of the material to be 
made on a commercial production scale. The container closure system must be the 
same or simulate the packaging proposed for storage and distribution of marketed 
product. 
7.1.4.2 Storage Conditions and Testing Frequency 
The guidance ICH Q1A (R2) [7] provides information on storage conditions and 
testing frequency for APIs under four intended storage conditions: (1) general cases, 
(2) Intended for storage in a refrigerator, (3) Intended for storage in a freezer, and 
(4) drug products intended for storage below . 20 ° C (for this storage no specifi c 
guidance is provided except to be treated on case - by - case basis). The storage conditions 
are summarized in Table 1 . 
Marketed API Intended for Room Temperature Storage Conditions The storage 
conditions defi ned in the guidance document [7] as a general case has no defi nition 
assigned in the guidance. In this chapter, the author is interpreting this term as 
requiring room temperature storage (vs. storage in refrigerator and freezer) conditions 
for the marketed API. The storage conditions (temperature and humidity) and 
lengths of storage for stability studies should be suffi cient to cover the following 
stages after manufacture: storage, shipment, and subsequent use. For submission 
of drug application, the data for long - term storage should cover a minimum of 12 
months on three primary batches. The testing should, however, continue for the 
duration of the proposed shelf life and retest period. The accelerated and, when 
necessary, intermediate (storage condition designed to moderately increase rate of 
chemical degradation or physical change) storage conditions should be carried out 
for 6 months. In this scenario, long - term storage can be conducted at either at 
25 ° C ± 2 ° C; 60% relative humidity (RH) ± 5% RH or 30 ° C ± 2 ° C, 65% RH ± 5% 
RH. If the long - term stability study is conducted at 30 ° C ± 2 ° C, 65% RH ± 5% RH, 
no intermediate condition storage is required. If long - term storage studies are conducted 
at 25 ° C ± 2 ° C, 60% RH ± 5% RH and a signifi cant change occurs in the test 

TABLE 1 Storage Conditions for Stability Evaluation of APIs 
Stability Study Type Stability Storage Conditions 
Minimum Time Period Covered 
by Data at Submission (months) 
Marketed API Intended for Room Temperature (General Case) [7] Storage Conditions 
Long term 25 ° C ± 2 ° C, 60% RH ± 5% 
RH or 30 ° C ± 2 ° C, 65% 
RH ± 5% RH 
12 
Intermediate 30 ° C ± 2 ° C, 65% RH ± 5% 
RH 
6 
Accelerated 40 ° C ± 2 ° C, 75% RH ± 5% 
RH 
6 
Marketed API Intended for Storage in Refrigerator 
Long term 5 ° C ± 3 ° C 12 
Accelerated 25 ° C ± 2 ° C, 60% RH ± 5% 
RH 
6 
Marketed API Intended for Storage in Freezer 
Long term . 20 ° C ± 5 ° C 12 
Source : From ref. 7 . 
result at any time during the 6 months testing at the accelerated storage condition 
(40 ° C ± 2 ° C, 75% RH ± 5% RH), testing for established stability specifi cations 
at the intermediate storage condition (30 ° C ± 2 ° C, 65% RH ± 5% RH) should be 
conducted and evaluated against signifi cant change criteria. The guidance document 
[7] defi nes signifi cant change for an API as failure to meet its specifi cation. If intermediate 
storage condition and testing become a necessity, a minimum of 6 months 
of data from this study should be submitted with the application. 
The sampling frequency and testing in long - term stability studies should be 
targeted to generate data suffi cient to establish a stability profi le for the API. The 
guidance [7] recommends testing every 3 months over the fi rst year, 6 months over 
the second year, and annually thereafter throughout the retest period under long - 
term storage condition. For a 6 - month accelerated storage stability condition, sampling 
at 0, 3, and 6 months is recommended. When signifi cant change to established 
test specifi cation occurs under accelerated storage condition, sampling at time 0, 6, 
9, and 12 is recommended for a 12 - month intermediate storage condition. 
Marketed API Intended for Storage in Refrigerator For an API intended for 
storage in a refrigerator (5 ° C ± 3 ° C), if a signifi cant change occurs between three 
and six months testing at the accelerated storage condition of 25 ° C ± 2 ° C, 60% 
RH ± 5% RH, the proposed retest period should be based on the real - time data 
available at the long - term storage condition of 5 ° C ± 3 ° C [7] . On the other hand, if 
a signifi cant change occurs within three months at the accelerated storage condition, 
the effect of short - term excursions outside the label storage condition during shipping 
or handling should be discussed. When signifi cant change occurs during the 
fi rst three months storage at accelerated condition, consideration of further storage 
or testing at this condition is not necessary. However, testing on a single batch of 
API for a period shorter than three months, with more frequent sampling during 
STABILITY REQUIREMENTS FOR DRUG SUBSTANCES (APIs) 565

566 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
three months may be necessary to narrow the period in which the API can be 
demonstrated to be stable [7] . 
Marketed API Intended for Storage in Freezer For an API with intended storage 
in a freezer ( . 20 ° C ± 5 ° C), the retest period should be based on the real - time data 
available at the long - term storage condition of . 20 ° C ± 5 ° C [7] . Since no accelerated 
storage condition for a freezer - stored API is proposed [7] , storage and testing of a 
single batch at elevated temperatures of 5 ° C ± 3 ° C or 25 ° C ± 2 ° C for an appropriate 
time period should be considered to understand the effect of short - term excursions 
outside the label storage condition during shipping or handling. Other storage conditions 
may be considered with appropriate justifi cation for their selection. 
7.1.4.3 Stress Studies and Stability - Indicating Methods for Analysis of 
API Stability 
Appropriate physical, chemical, biological, and microbiological attributes of the API 
that are likely to be susceptible to change during storage and are likely to infl uence 
the quality, safety, and effi cacy should be tested. The analytical testing performed 
to evaluate the stability of the API should be stability indicating as defi ned above 
under GMPs for APIs [ 3 ; Section 11.5 ]. The stability - indicating assay accurately 
measures the active ingredient without interference from degradation products, 
process im purities, excipients, or other potential impurities. In order to develop a 
stability - indicating method, it is necessary to conduct stress studies on an API. Stress 
testing is carried out on a single batch of the drug substance and the stresses can 
include acid (e.g., 0.2 N HCl) and base (e.g., 0.2 N NaOH) hydrolysis, temperature 
(10 ° C increments above the accelerated stability storage temperature of 40 ° C, e.g., 
50 ° C, 60 ° C), humidity (75% RH or greater), photolysis (see Section 7.1.6 ), and oxidation 
(e.g., 10% H 2 O) of the API [7] . Hydrolysis is a chemical transformation 
process whereby an organic molecule RX reacts with water (H 2 O), resulting in 
direct replacement of X by OH. Among various stresses stated above, hydrolysis 
could be the most common stress leading to degradation of an API since amides, 
esters, and salts of weak acids and strong bases are well known to hydrolyze [13] . 
Photolysis is the process whereby chemicals are altered directly as a result of irradiation 
or indirectly through interaction with products of irradiation [13] . Carbonyl, 
nitroaromatic, N - oxide function, C . C double bond, weak C . H bond, suplfi des, 
alkenes, polyenes, and phenols are functional groups that may react with light [14] . 
Oxidation reactions depend on several factors, including temperature, light, pH, 
oxygen concentration, impurities including metal ions (e.g., cupric, ferric), and the 
oxidizable component of the molecule [15] . 
The stress studies should demonstrate that impurities and degradants from the 
active ingredient do not interfere with the quantitation of the API [12] . Stress testing 
of the API, in addition to validating the stability - indicating power of the analytical 
method, can also help establish the degradation pathways and the intrinsic stability 
of the molecule [7] . 
7.1.4.4 Evaluation of API Stability Results 
When the stability data for all three batches on stability at various sampling intervals 
show little variability from initiation of the stability study, the API can be considered 

stable and no statistical analysis of the data are required. Where the data show API 
degradation with time, the time at which the 95%, one - sided confi dence limit for 
the mean curve intersects the acceptance criteria should be defi ned. When the 
analysis shows small batch - to - batch variation, combining the data into an overall 
estimate is recommended by applying appropriate statistical tests to the slopes of 
regression lines and zero time intercepts for the individual batches. The overall 
retest period should be based on the minimum time a batch can be expected to 
remain within acceptance criteria. Appropriate statistical methods should be considered 
to test the goodness of the fi t of the data on all batches and where applicable 
combined batches to the assumed degradation line or curve [7] . The stability evaluation 
should cover API content, levels of degradation products, and other appropriate 
attributes. 
7.1.4.5 Stability Commitment 
When the submission includes long - term stability studies on three batches of API 
covering the proposed retest period, a postapproval commitment is unnecessary. If, 
on the other hand, when at the time of approval long - term stability data for the 
primary batches do not cover the proposed retest period granted, a commitment 
should be made to continue the stability studies postapproval to fi rmly establish the 
retest period. If the submission includes data from stability studies on fewer than 
three production batches, a commitment should be made to continue these studies 
through the proposed retest period and to place additional production batches for 
a total of three and generate data through the proposed retest period. If the submission 
does not include stability data on the production batches, a commitment should 
be made to place the fi rst three production batches on long - term stability studies 
through the proposed retest period. 
7.1.4.6 Storage Statement and Labeling 
A storage statement for the fi nished API should be prepared in compliance with 
the national and regional requirements. The established retest period, which is supported 
by the stability data, should be displayed on the certifi cate of analysis (COA) 
and as appropriate on the container label also. 
7.1.5 STABILITY REQUIREMENTS FOR DRUG PRODUCTS 
The guidance for industry Q1A (R2) [7] defi nes requirements for stability data 
package submission with drug applications in the EU, Japan, and the United States 
for drug products. Information on the chemistry of the API molecule and its degradation 
behavior from the stability studies of the API (see Section 7.1.4 ) should 
be useful in designing the stability program for the drug product, since one of the 
key measures of shelf life determination of the drug product is the stability of the 
API in the drug product formulation. 
7.1.5.1 Selection of Batches and Container Closure System 
Stability data should be generated on at least three primary batches of the drug 
product, with the manufacturing process simulating the production batches, with 
STABILITY REQUIREMENTS FOR DRUG PRODUCTS 567

568 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
formulation and quality specifi cations similar to those batches intended for marketing. 
The three batches of the drug product manufactured should use different batches 
of the drug substance where possible, at least two of three batches should be pilot 
scale (batch manufactured by a procedure fully representative of and simulating 
that applied to a full production - scale batch), and the third batch can be of the 
smaller size when justifi ed [7] . 
The drug product for the stability studies should be packaged in the same container 
closure system as proposed for marketing of the drug product, and each 
individual strength and container size of the proposed packaging confi guration 
should be placed on stability, unless bracketing and matrixing designs are used in 
compliance with ICH guidance for stability testing [10] . 
7.1.5.2 Bracketing and Matrixing 
During the design of stability studies, bracketing and matrixing [10] may be used to 
achieve reduced testing while at the same time generating enough stability data for 
evaluation of shelf life. 
In bracketing, the design may include reduction in storage and sampling of 
dosage strengths or container closure confi guration. For example, in a three - batch 
stability study with dosage strengths of 50, 75, and 100 mg in 15 - , 100 - , and 150 - mL 
high - density polyethylene (HDPE) containers, testing for 50 - and 100 - mg strengths 
in 15 - and 150 - mL container sizes may be adequate with no testing proposed for 
the 75 - mg strength. 
Matrixing design [10] may involve elimination of some stability sample pull time 
points to achieve reduced testing strategy. For example, a one - half reduction in time 
points eliminates one in every two time points from full study design, and one - third 
reduction eliminates one in every three time points. However, such a scenario must 
include full testing at initial, 12 - month, and fi nal time points under a 36 - month shelf 
life study [10] . 
7.1.5.3 Storage Conditions and Testing Frequency 
The drug product guidance Q1A (R2) [7] provides information on storage conditions 
and testing frequency for the drug products under six intended storage conditions: 
(1) general case (room temperature), (2) drug products packaged in 
impermeable containers, (3) drug products packaged in semipermeable containers, 
(4) drug products intended for storage in a refrigerator, (5) drug products intended 
for storage in a freezer, and (6) drug products intended for storage below . 20 ° C 
(for this storage no specifi c guidance is provided except to be treated on a case - by - 
case basis). The storage conditions are summarized in Table 2 . 
Marketed Drug Product Intended for Room Temperature Storage Conditions The 
storage condition defi ned in the guidance document [7] as a general case has no 
defi nition assigned in the guidance and in this chapter is interpreted as room temperature 
storage (other than refrigerator and freezer) condition for the marketed 
drug product. The storage conditions (temperature and humidity) and lengths of 
storage for stability studies should be suffi cient to cover the following stages after 
manufacture: storage, shipment, and subsequent use. 

For submission of drug application, the data for long - term storage should cover 
a minimum of 12 months on at least three primary batches. The primary batches are 
those used in a formal stability study from which the stability data are derived and 
submitted in a registration application for the purpose of establishing a retest period 
or shelf life/expiration date. The testing should, however, continue for the duration 
of the proposed shelf life and retest period. The accelerated and, when necessary, 
intermediate storage conditions should be carried out for 6 months. In this scenario, 
long - term storage can be conducted at either 25 ° C ± 2 ° C, 60% RH ± 5% RH 
or 30 ° C ± 2 ° C, 65% RH ± 5% RH. If the long - term stability study is conducted at 
30 ° C ± 2 ° C, 65% RH ± 5% RH, no intermediate condition storage is required. If 
long - term storage studies are conducted at 25 ° C ± 2 ° C, 60% RH ± 5% RH and a 
signifi cant change occurs in the test result at any time during the 6 months testing 
at the accelerated storage condition (40 ° C ± 2 ° C, 75% RH ± 5% RH), testing for 
established stability specifi cations at the intermediate storage condition (30 ° C ± 2 ° C, 
65% RH ± 5% RH) should be conducted and evaluated against signifi cant change 
criteria (Table 2 ). 
TABLE 2 Storage Conditions for Stability Evaluation of Drug Products 
Stability Study Type Stability Storage Conditions 
Minimum Time Period Covered 
by Data at Submission (months) 
Marketed Drug Product Intended for Room Temperature Storage Conditions 
Long term 25 ° C ± 2 ° C, 60% RH ± 5% 
RH or 30 ° C ± 2 ° C, 65% 
RH ± 5% RH 
12 
12 
Intermediate 30 ° C ± 2 ° C, 65% RH ± 5% 
RH 
6 
Accelerated 40 ° C ± 2 ° C, 75% RH ± 5% 
RH 
6 
Marketed Drug Product Packaged in Semipermeable Containers 
Long term 25 ° C ± 2 ° C, 40% RH ± 5% 
RH or 30 ° C ± 2 ° C, 35% 
RH ± 5% RH 
12 
Intermediate 30 ° C ± 2 ° C, 65% RH ± 5% 
RH 
6 
Accelerated 40 ° C ± 2 ° C, no more than 
25% RH 
6 
Marketed Drug Product Intended for Storage in Refrigerator 
Long term 5 ° C ± 3 ° C 12 
Accelerated 25 ° C ± 2 ° C, 60% RH ± 5% 
RH 
6 
Marketed API Intended for Storage in Freezer 
Long term . 20 ° C ± 5 ° C 12 
Source : From ref. 7 . 
STABILITY REQUIREMENTS FOR DRUG PRODUCTS 569

570 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
The guidance document [7] defi nes signifi cant change for drug product as one 
or more of the following: (1) a 5% change in assay from initial value or failure 
to meet the acceptance criteria for potency when using biological or immunological 
procedures; (2) any degradation product exceeding its acceptance criterion; 
(3) failure to meet the acceptance criteria for appearance, physical attributes, and 
functionality test (e.g., color, phase separation, resuspendability, caking, hardness, 
and dose delivery per actuation; under accelerated storage conditions, some changes 
in physical attributes may be expected, e.g., softening of suppositories and melting 
and possible phase separation in of creams); (4) failure to meet acceptance criterion 
for pH; and (5) failure to meet acceptance criteria for dissolution for 12 dosage 
units. 
If intermediate storage and testing become a necessity, a minimum of 6 months 
of data from this study should be submitted with the application. 
The sampling frequency and testing in long - term stability studies should be targeted 
to generate data suffi cient to establish a stability profi le of the drug product. 
For long - term storage conditions, the guidance [7] recommends testing every 3 
months over the fi rst year, 6 months over the second year, and annually thereafter 
through the proposed shelf life for drug products when the proposed shelf life is at 
least 12 months. 
For a 6 - month accelerated storage stability condition, sampling at 0, 3, and 6 
months is recommended. When signifi cant changes to established test specifi cations 
are likely to occur under the accelerated storage condition, increased testing is 
required with inclusion of a fourth sampling point. When signifi cant changes to 
established test specifi cations occur under the accelerated storage condition, testing 
at the intermediate storage condition for 12 months with sampling at time 0, 6, 9, 
and 12 is recommended. 
Drug Products Packaged in Impermeable Containers For drug products packaged 
in impermeable containers that provide a permanent barrier, moisture or 
solvent loss is not a concern and for such products stability studies can be conducted 
under any controlled or ambient humidity conditions. 
Drug Products Packaged in Semipermeable Containers Stability studies for 
aqueous - based drug products packaged in semipermeable containers (containers 
that allow the passage of solvent, usually water, while preventing solute loss) should 
be conducted under conditions of low relative humidity and temperatures specifi ed 
in Table 2 . Stability attributes such as potential water loss and physical, chemical, 
biological, and microbiological stability should be evaluated. 
If long - term storage studies are conducted at 25 ° C ± 2 ° C, 40% RH ± 5% RH and 
a signifi cant change other than water loss occurs during the six months testing at 
the accelerated storage condition (40 ° C ± 2 ° C, 75% RH ± 5% RH), testing for 
established stability specifi cations at the intermediate storage condition (30 ° C ± 2 ° C, 
65% RH ± 5% RH) should be conducted to evaluate the effect of 30 ° C (Table 2 ). 
While a signifi cant change in water loss (5% water loss from initial value) alone 
under accelerated storage conditions need not prompt testing of samples under the 
intermediate storage condition, water loss through the proposed shelf life should be 
monitored to ensure that the drug product has no signifi cant water loss during long - 
term storage at 25 ° C ± 2 ° C, 40% RH ± 5% RH. 

Marketed Drug Product Intended for Storage in Refrigerator For a drug product 
intended for storage in a refrigerator, if a signifi cant change occurs between three 
and six months testing at the accelerated storage condition of 25 ° C ± 2 ° C, 60% RH 
± 5% RH, the proposed retest period should be based on the real - time data available 
at the long - term storage condition of 5 ° C ± 3 ° C. On the other hand, if a signifi - 
cant change occurs within three months at the accelerated storage condition, the 
effect of short - term excursions outside the label storage condition during shipping 
or handling should be discussed. When signifi cant change occurs during the fi rst 
three months storage at the accelerated condition, consideration of further storage 
or testing is not necessary. However, testing on a single batch of drug product for a 
period shorter than three months with more frequent sampling during the three 
months may be necessary to narrow the period in which the drug product can be 
demonstrated to be stable. 
Marketed Drug Products Intended for Storage in Freezer For a drug product with 
intended storage in a freezer, the retest period should be based on the real - time 
data available at the long - term storage condition of . 20 ° C ± 5 ° C. Since no accelerated 
storage condition for a freezer - stored API is proposed, storage and testing of 
a single batch at elevated temperatures of 5 ° C ± 3°C or 25°C ± 2°C for an appropriate 
time period should be considered to understand the effect of short - term excursions 
outside the label storage condition during shipping or handling. Other storage 
conditions may be considered with appropriate justifi cation for selection. 
7.1.5.4 Stress Studies and Stability - Indicating Methods for Analysis of 
Drug Product Stability 
Appropriate physical, chemical, biological, and microbiological attributes, antimicrobial 
preservative and antioxidant content, and dosage functionality test (e.g., 
dose delivery system) of the drug product that are likely to be susceptible to change 
during storage and are likely to infl uence quality, safety, and effi cacy should be 
tested. The analytical testing performed to evaluate the stability of the drug product 
should be validated and stability indicating [12] . The stability - indicating assay accurately 
measures the active ingredient in drug product without interference from 
degradation products, process impurities, excipients, or other potential impurities. 
In order to develop a stability - indicating method, it is necessary to conduct stress 
studies. Stress testing (studies undertaken to elucidate the intrinsic stability of the 
drug substance) can be carried out on the drug product similar to that described for 
APIs in Section 7.1.4 . Degradation information obtained from stress studies for the 
active ingredient in the drug product should demonstrate the specifi city of the assay, 
that is, that impurities and degradants from the active ingredient and drug product 
excipients do not interfere with the quantitation of the active ingredient in the drug 
product [12] . 
7.1.5.5 Evaluation of Stability Results 
Data on stability of the drug products should be presented for all testing intervals 
and evaluated with physical, chemical, and microbiological, microbial preservative 
effectiveness, antioxidant effectiveness, and functionality tests as appropriate to the 
STABILITY REQUIREMENTS FOR DRUG PRODUCTS 571

572 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
drug product dosage form and all established specifi cations for the attributes being 
considered. 
When the stability data for all three batches on stability at various sampling 
intervals show little variability from initiation of the stability study, the drug product 
can be considered stable and no statistical analysis of the data is required. Where 
the data show change in quantitative attributes (API and degradation amounts, 
dissolution rates), the time at which the 95% one - sided confi dence limit for the 
mean curve intersects the acceptance criteria should be defi ned. When the analysis 
shows small batch - to - batch variation, combining the data into an overall estimate 
is recommended by applying appropriate statistical tests to the slopes of regression 
lines and zero - time intercepts for the individual batches. The overall shelf life should 
be based on the minimum time during the stability study a drug product can be 
expected to remain within acceptance criteria. Appropriate statistical methods 
should be considered to test the goodness of fi t of the data on all batches and where 
applicable combined batches to the assumed degradation line or curve [7] . The stability 
evaluation should cover the levels of degradation products and other appropriate 
attributes. The mass balance, that is, whether the addition of assay value and 
degradation products add up to 100% of the initial value, should be considered 
taking into account the margin of analytical error of the method used. Details on 
stability evaluation are discussed in Section 7.1.7 . 
7.1.5.6 Stability Commitment 
When the submission includes long - term stability studies on three batches of drug 
product covering the proposed shelf life, a postapproval commitment is unnecessary. 
If, on the other hand, when at the time of approval long - term stability data for the 
primary batches of drug product do not cover the proposed shelf life granted, a 
commitment should be made to continue the long - term stability studies (and six - 
month accelerated stability study if not performed), postapproval, to fi rmly establish 
the shelf life. If the submission includes data from stability studies on fewer than 
three production batches (batches manufactured at production scale by using production 
equipment in a production facility as specifi ed in the application), a commitment 
to continue long - term studies through proposed shelf life and accelerated 
studies for six months for a total of three batches should be made. If the submission 
does not include stability data on the production batches, a commitment should be 
made to place the fi rst three production batches on long - term stability studies 
through the proposed shelf life and accelerated studies for six months. When signifi 
cant change occurs in the accelerated stability study on the primary batches 
(requiring a study under intermediate conditions), stability of commitment batches 
can be conducted at either the accelerated or intermediate condition. On the other 
hand, if a signifi cant change occurs for the commitment batches under the accelerated 
condition, testing at the intermediate storage condition is also required for 
commitment batches. A formal stability commitment protocol should be in place 
for primary and commitment batches [7] . 
7.1.5.7 Storage Statement and Labeling 
A storage statement for the fi nished drug product should be established in compliance 
with the national and regional requirements, which should be based on the 

conclusions derived from stability evaluation. For drug products that cannot tolerate 
freezing, specifi c instructions should be provided. The established expiration date 
(shelf life) should be displayed on the COA and as appropriate on the container 
label [7] . 
7.1.6 PHOTOSTABILITY STUDIES 
The stability guidance for drug substances and drug products [7] suggests that photostability 
testing should be an integral part of stress testing. The guidance on photostability 
testing [8] suggests that intrinsic stability of new drug substances and drug 
products should be evaluated to demonstrate that light exposure does not result in 
unacceptable change. The guidance [8] also recommends a systematic approach to 
stability testing, including on drug substance by direct exposure, and tests on the 
exposed drug product outside the the immediate (primary) pack. In the immediate 
(primary) pack, the packaging material is in contact with the drug substance or drug 
product and also includes any label associated with the primary pack. If change after 
exposure of drug product outside the immediate pack is not acceptable, testing on 
the immediate pack should be conducted. If change in the immediate pack exposed 
to light is not acceptable, photostability testing on the marketing pack (combination 
of immediate pack and other secondary packaging such as carton) should be conducted. 
If the change in the marketed pack exposed to light is not acceptable, the 
packaging confi guration should be redesigned or the product may require reformulation 
and photostability testing performed under the redesigned scenario. 
The photostability guidance [8] provides recommendations for the light sources 
to which the API or drug product should be exposed. Two light exposure options 
are suggested. Option 1 includes light output similar to internationally recognized 
standards for outdoor light (D65) or indoor indirect light (ID 65), defi ned in the 
Internation Organization for Standardization (ISO) 10977. The light sources covered 
under option 1 are the artifi cial daylight fl uorescent lamp with a combination of 
ultraviolet (UV) and visible outputs, xenon, or metal halide lamps. In option 2, it is 
recommended that the sample should be exposed to both the cool white fl uorescent 
and near - UV lamp. The cool white fl uorescent lamp should comply with the outputs 
specifi ed in ISO 10977. It is recommended that the near - UV fl uorescent lamp having 
a spectral distribution from 320 to 400 nm with a maximum energy emission between 
350 and 370 nm be used with a signifi cant portion of UV in both bands, that is, 320 – 
360 and 360 – 400 nm. For confi rmatory studies, the samples should be exposed to 
light providing an overall illumination of not less than 1.2 million lux hours and an 
integrated near - UV energy of not less than 200 Wh/m 2 . Use of a validated chemical 
actinometric system is recommended to ensure that the samples are exposed to 
desired light exposure by exposing actinometer solutions to light side by side with 
the drug substance or drug product samples. Light exposure also should be measured 
using calibrated radiometers or lux meters. Dark controls (covered with an 
aluminum foil to protect from light) should also be placed side by side with the 
light - exposed samples. The guidance provides important references on chemical 
actinometers [16] , interlaboratory studies by industry and FDA [17] , and perspectives 
from FDA scientists [18] on photostability testing of pharmaceutical products. 
In addition, a reference on forced degradation testing (including but not limited to 
light stress) of ibuprofen bulk drug and tablets by Farmer et al. [19] illustrates the 
PHOTOSTABILITY STUDIES 573

574 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
differences in degradation of ibuprofen as neat API and ibuprofen in association 
with excipients in drug product. A reference is also provided to illustrate how different 
light sources and intensities effect the degradation of a light - sensitive compound 
valerophenone [20] . 
7.1.6.1 Photostability of API s 
Photostability studies of a drug substance can provide an overall evaluation of the 
sensitivity of the drug substance to light. Because the exposure levels under options 
1 and 2 described above can facilitate worst - case exposure (forced degradation or 
light stress) of a drug substance, extensive degradation is likely to be seen, especially 
if the drug substance has functional groups that are sensitive to photooxidation. If 
degradation of APIs occurs during photostability study, that may present an opportunity 
for development of a stability - indicating method. Photostability of the drug 
substance may be carried out on its neat form or in solutions with samples placed 
in chemically inert and transparent containers for exposure (e.g., quartz glass). 
Depending on the known photosensitivity of the drug substance (i.e., based on the 
functional groups that are susceptible to oxidation or known absorbance since 
compounds in the 290 – 800 - nm range are likely to undergo degradation [13] ), the 
intensity of the light exposure and length exposed can be controlled to achieve the 
desired degradation in a forced degradation study. It is also important to partition 
the temperature effects on degradation of the drug substance from light effects since 
elevated temperatures seen under light exposure may infl uence degradation independent 
of light exposure. Appropriate temperature controls should be used. 
Changes in physical state such as melting, sublimation, and evaporation should be 
minimized. Use of quartz glass containers are recommended for photostability 
studies. An appropriate amount of solid APIs can be spread in a container with 
thickness of approximately 3 mm. Liquids can be exposed in appropriate containers 
that are closed tightly. Appropriate dark controls (containers covered with aluminum 
foil and placed side by side along the exposed samples) should be used. Analysis 
of samples should include both the exposed and dark controls and the analysis 
should be performed concomitantly. Parameters examined can include physical 
properties such as appearance and color of solids and, if a solution is exposed, 
the clarity and color of solution. The assay and degradation products should be 
evaluated using validated stability - indicating methods that clearly separate the 
degradation products and impurities from the drug substance in the sample 
chromatograms. 
The confi rmatory studies with drug substance should be designed to identify 
precautionary measures needed to protect the API from light during formulation 
or manufacturing of the drug product, especially if the drug substance is found to 
degrade under light exposure. These studies also should help in the design of 
primary and secondary packaging material for commercial packing of the API. 
7.1.6.2 Photostability of Drug Products 
Drug product should be exposed to light conditions as stated above. The testing 
should be done in a sequential manner with fully exposed drug product and as 
necessary followed by testing the product in the immediate pack and then in the 

marketing pack [8] . Results of the photostability study should demonstrate that the 
drug product is adequately protected from light during the shelf life. The testing 
should include one batch during the developmental phase followed by confi rmation 
of the photostability characteristics using one batch. Additional two batches should 
be conducted if the results in the confi rmatory study with one batch are similar to 
the developmental study. Confi rmatory studies should establish the photostability 
characteristics of the drug product under standardized conditions. Results from 
these studies should help in (1) identifi cation of precautionary measures needed 
during manufacture and packaging, (2) container closure design for protection from 
light, and (3) storage conditions and light protection required during shelf life of 
the marketed product. While keeping the light exposure constant during the exposure 
period, care should be taken to control the temperature to ensure that the 
degradation products observed are light exposure related. Physical characteristics 
of the samples under test conditions should be monitored to ensure that changes in 
the physical state of the drug product are minimized. Dark controls (unexposed 
controls) should be tested side by side with the light - exposed samples. Maximum 
light exposure under direct exposure for solid oral dosage forms such as tablets and 
capsules can be achieved by spreading them in a single layer. Quartz glass containers 
are recommended as containers for direct exposure. Based on the results obtained 
during direct exposure of the drug product, if immediate and marketing pack exposure 
becomes necessary, samples should be placed in such a manner as to facilitate 
uniform exposure. Key physical parameters to be examined should include appearance, 
clarity and color of solutions, dissolution, and disintegration. Quantitation of 
drug substance, degradation products of drug substance (known and unknown), 
impurities, and excipients present in the samples of drug product from photostability 
studies should be performed using validated stability - indicating methods. It is 
important to sample representative exposed or dark control samples to make a 
realistic assessment of the photostability of the various dosage forms. Solutions, 
suspensions, and creams should be examined to ensure that there is no settling or 
partitioning of phases and that uniform and homogenous samples are analyzed. 
7.1.7 EVALUATION OF STABILITY DATA AND 
SHELF LIFE DETERMINATION 
Requirements for stability data evaluation and shelf life determination are described 
in the guidance for industry — ICH Q1E [11] , which covers evaluation of stability 
data that should be submitted in registration applications for new molecular entities 
and associated drug products. It also provides recommendations for the establishment 
of retest periods and shelf lives for drug substances and drug products intended 
to be stored at room temperature, in refrigerator or freezer, and below . 20 ° C. In 
addition to the guidance document, studies by Grimm [21, 22] should be reviewed 
which laid the foundation for testing and evaluation of stability data. 
7.1.7.1 Shelf Life Estimation for Drug Substances or Drug Products Intended 
for Room Temperature Storage 
The stability data evaluation for drug substance or drug product intended for 
storage at room temperature should include evaluation of any signifi cant change at 
EVALUATION OF STABILITY DATA AND SHELF LIFE DETERMINATION 575

576 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
the various time points during accelerated stability study and at intermediate storage 
conditions when they are used and evaluation of long - term stability data. Different 
scenarios for retest period and shelf life determination are presented based on the 
results observed under accelerated, intermediate, and long - term storage conditions 
[11] , as described below. 
Long-Term and Accelerated Stability Data Show Little or No Change over Time 
and Little or No Variability The drug substance or drug product can be considered 
stable if the long - term and accelerated stability data show little or no change 
or little no variability over time. In this scenario, no statistical analysis is required 
but a justifi cation should include a discussion of the pattern of change or variability 
or lack of change or variability, supportive accelerated stability data, mass balance 
of drug substance in samples, and/or other supporting data. The retest period or 
shelf life in this scenario can be up to twice as long as but not more than 12 months 
beyond the period covered by long - term storage stability data. However, extrapolation 
of the retest period or shelf life beyond the period covered by long - term data 
can be proposed. 
Long-Term and Accelerated Stability Data Show Change over Time and/or 
Variability In this scenario, statistical analysis of the long - term data can be useful 
in establishing a retest period or shelf life. When there are differences, for example, 
among batches or among dosage strengths, container sizes, and/or fi ll or any 
combination thereof that preclude combining the data, the proposed retest date or 
shelf life should not exceed the shortest period supported by long - term studies of 
any batch, other factors, or combinations of factors. However, when differences 
are attributed, for example, to strength of the dosage form, different shelf lives 
can be assigned to different strengths within the dosage form of the drug product, 
with appropriate discussion of the data generated. If the statistical analysis is 
performed on long - term data and the statistical and other relevant data are supportive, 
it can be appropriate to extrapolate and propose a retest period or shelf 
life up to twice as long but not more than 12 months beyond the period covered by 
long - term storage stability data. When long - term stability data are not amenable to 
statistical analysis, the proposed retest period or shelf life can be up to one and a 
half as long as but not more than 6 months beyond the period covered by long - term 
data. 
Accelerated Stability Data Show Signifi cant Change When a signifi cant change 
occurs at the accelerated storage condition, stability results at the intermediate 
condition and the long - tem storage condition will determine the retest period or 
shelf life. The guidance [11] suggests two exceptions to the observations at accelerated 
conditions: (1) physical change such as softening of a suppository designed to 
melt at 37 ° C if the melting point is clearly demonstrated and (2) failure to meet 
acceptance criteria for dissolution for 12 units of gelatin capsules or gel - coated 
tablets if the failure can be unequivocally attributed to gelatin cross - linking. These 
exceptions do not apply to physical change as exemplifi ed by phase separation of 
a semisolid dosage form, creams, or other occurrences at accelerated conditions, 
and if such a change occurs, testing at the intermediate condition should be 
performed. 

When no signifi cant change is observed at the intermediate condition, extrapolation 
beyond the period covered by long - term data can be proposed when statistical 
analysis data and relevant supporting data back up such extrapolation. The proposed 
retest period or shelf life can be up to one and a half as long as but not more 
than six months beyond the period covered by long - term data. When long - term 
stability data are not amenable to statistical analysis but the relevant supportive 
data are available, the proposed retest period can be up to three months beyond 
the period covered by long - term data. 
In a scenario where signifi cant change occurs at the intermediate condition, the 
retest period or shelf life shorter than the period covered by long - term data is 
appropriate but should not exceed the period covered by long - term data. 
7.1.7.2 Shelf Life Estimation for Drug Substances or Drug Products Intended 
for Storage in a Refrigerator 
Long-Term and Accelerated Stability Data Show Little Change over Time and/or 
Variability Under this scenario, the proposed retest period or shelf life can be up 
to one and a half times as long as but not more than six months beyond the period 
covered by long - term data without the support of statistical analysis. 
Long-Term and Accelerated Stability Data Show Change over Time and/or 
Variability Under this scenario, the proposed retest period and shelf life can 
be up to one and a half times as long as but not more than six months beyond the 
period covered by long - term stability data when backed by statistical analysis and 
supporting data. On the other hand, the proposed retest period and shelf life can 
only be up to three months beyond the period covered by long - term stability data 
if the long - term data are amenable to statistical analysis but the statistical analysis 
is not performed or the long - term data are not amenable to statistical analysis but 
data supporting the proposed retest date and shelf life are available. 
Accelerated Stability Data Show Signifi cant Change or Variability Under this 
scenario, if signifi cant change occurs during accelerated stability study between 
three and six months, the proposed retest period or shelf life should be based on 
the long - term data, and extrapolation is not considered appropriate, but the retest 
period or shelf life shorter than that supported by long - term stability data is considered 
appropriate. In the case of variability in the long - term stability data, verifi cation 
of the proposed retest period or shelf life by statistical analysis is considered 
appropriate. On the other hand, if signifi cant change occurs during accelerated stability 
study within the fi rst three months, the proposed retest period or shelf life 
should be based on the long - term data, and extrapolation is not considered appropriate, 
but the retest period or shelf life shorter than that supported by long - term 
stability data is considered appropriate. Conducting a stability study at accelerated 
conditions for shorter than three months should be considered under this scenario 
at least for one batch with more frequent sampling and analysis. 
7.1.7.3 Shelf Life Estimation for Drug Substances or Drug Products Intended 
for Storage in a Freezer 
In this scenario, the retest period or shelf life should be based on long - term data. 
EVALUATION OF STABILITY DATA AND SHELF LIFE DETERMINATION 577

578 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
7.1.7.4 Shelf Life Estimation for Drug Substances or Drug Products Intended 
for Storage below - 20 ° C 
In this scenario, the retest period or shelf life should be based on long - term data. 
7.1.8 ASSIGNMENT OF CLIMATIC ZONES AND RECOMMENDED 
STORAGE CONDITIONS 
In June 2004, the FDA issued a guidance document to the industry on stability data 
packaging for registration applications in climatic zones III and IV [23] (adopted 
from the ICH guidance Q1F). The guidance provided an update to the stability 
storage conditions for climatic zones I and II and provided new guidance for climatic 
zones III and IV based on recommendations provided in the WHO stability guidance 
[24] . The WHO guidance [24] described stability testing recommendations and 
storage conditions for all four climatic zones, considering Grimm ’ s [25] recommendation 
for climatic zones III and IV. The QIF guidance [23] clarifi ed that the long - 
term storage condition for climatic zones I and II is 25 ° C ± 2°C, 60% RH ± 5% RH, 
and the intermediate storage condition for climatic zones I and II is 30 ° C ± 2 ° C, 
65% RH ± 5% RH. It also stated [23] that the storage condition of 30 ° C ± 2 ° C, 65% 
RH ± 5% can also be a suitable alternative for the long - term storage condition of 
25 ° C ± 2 ° C, 60% RH ± 5% RH for climatic zones I and II, in which case no intermediate 
condition is required. For climatic zones III and IV, the Q1F guidance [23] 
suggested using 30 ° C ± 2 ° C, 65% RH ± 5% (for products intended to be stored at 
room temperature; general case stated in the guidance) as the long - term storage 
condition (12 months) with accelerated storage at 40 ° C ± 2 ° C, 75% RH ± 5% 
(6 months), and further no intermediate storage condition was recommended for 
climatic zones III and IV. For aqueous - based drug products packaged in semipermeable 
containers, the guidance [23] recommended 30 ° C ± 2 ° C, 35% RH ± 5% RH 
for long - term (12 months) and 40 ° C ± 2 ° C and not more than 25% RH ± 5% RH 
for accelerated storage (6 months). The WHO Expert Committee on Specifi cations 
for Pharmaceutical Preparations in its fortieth meeting, held at Geneva, Switzerland, 
in October 2005, recommended to split the climatic zone IV into zone IVA with 
long - term storage condition remaining at 30 ° C, 65% RH, whereas for climatic zone 
IVB, the long - term storage condition recommended was 30 ° C, 75% RH [26] . In 
Annex 1 of working document QAS/06.179 (2006) [26] , the WHO defi ned the fi ve 
climatic zones and proposed long - term testing conditions for each zone, which are 
summarized in Table 3 . Following the WHO guidance [26] publication in 2006, the 
FDA Q1F [23] was withdrawn in 2006, with future revision of Q1F possible. 
7.1.9 STABILITY TESTING PARAMETERS FOR DIFFERENT 
DOSAGE FORMS 
The testing parameters published by the WHO as Annex 2 to working document 
QAS/06.179 in 2006 [27] is the basis for the information summarized in this section 
for various dosage forms. Although the parameters listed here provide guidance for 
various dosage forms, pharmacopeial and those approved by regulatory authorities 

TABLE 3 Defi nition of Climatic Zones and Recommended Long - Term Stability 
Conditions 
Climatic Zone Defi nition 
Criteria [Mean 
Annual Temperatures 
Measured in Open 
Air ( ° C) and Mean 
Annual Partial Vapor 
Pressure (hPa)] 
Long - Term Testing 
Condition [Temperature 
( ° C) and RH] 
I Temperate climate . 15 ° C, . 11 hPa 21 ° C, 45% RH 
II Subtropical and 
mediterranean 
climate 
> 15 – 22 ° C, > 11 – 18 hPa 25 ° C, 60% RH 
III Hot and dry climate > 22 ° C, . 15 hPa 30 ° C, 35% RH 
IVA Hot and humid 
climate 
> 22 ° C, > 15 – 27 hPa 30 ° C, 65% RH 
IVB Hot and very humid 
climate 
> 22 ° C, > 27 hPa 30 ° C, 75% RH 
Source : From ref. 26 . 
in drug applications should be considered in applying the information described 
below for various dosage forms. 
7.1.9.1 Stability Tests Common to All Dosage Forms 
Appearance, assay, and degradation products, preservative, and antioxidant content 
as applicable should be considered for all dosage forms. The microbial bioburden 
of sterile dosage forms must be controlled and tested in compliance with pharmacopeial 
and/or internal specifi cations. The microbial bioburden of nonsterile dosage 
forms should be controlled, with appropriate sampling and testing. 
7.1.9.2 Stability Tests Specifi c to Dosage Forms 
Tablets Dissolution (or disintegration, if justifi ed), water content, hardness/ 
friability. 
Hard Gelatin Capsules Brittleness, dissolution (or disintegration, if justifi ed), 
water content, and microbial bioburden. 
Soft Gelatin Capsules Dissolution (or disintegration, if justifi ed), microbial bioburden, 
pH, leakage, and pellicle formation. 
Emulsions Phase separation, pH, viscosity, microbial bioburden, mean size and 
distribution of dispersed globules. 
Oral Solutions and Suspensions Formation of precipitate, clarity for solutions, 
pH, viscosity, microbial bioburden, extractables, and polymorphic conversion 
when applicable. Additional tests for suspensions include redispersability, 
rheological properties, mean size, and distribution of particles. 
Powders for Oral Solutions and Suspensions Water content, reconstitution time, 
and reconstituted solutions and suspensions should be tested as above for oral 
solutions and suspensions. 
STABILITY TESTING PARAMETERS FOR DIFFERENT DOSAGE FORMS 579

580 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
Topical, Ophthalmic, and Otic Preparations Clarity, homogeneity, pH, resuspendability 
(for lotions), consistency, viscosity, microbial bioburden, and water 
loss should be tested. For ophthalmic and otic products additional attributes 
should include sterility, particulate matter, and extractables. 
Suppositories Softening range, dissolution at 37 ° C. 
Transdermal Patches In vitro release rates, leakage, microbial bioburden/ 
sterility, and peel and adhesive forces. 
Metered - Dose Inhalers and Nasal Aerosols Content uniformity, aerodynamic 
particle size distribution, microscopic evaluation, water content, leak rate, 
microbial bioburden, valve delivery, extractables, leachables from plastic and 
elastomeric components. 
Nasal Sprays Clarity, microbial bioburden, pH, particulate matter, unit spray 
medication content uniformity, droplet and/or particle size distribution, weight 
loss, pump delivery, microscopic evaluation of suspensions, particulate matter, 
extractables, leachables from plastic and elastomeric components of container 
closure and pump. 
Small - Volume Parenterals Color, clarity of solutions, particulate matter, pH, 
sterility, endotoxins. Powders for injection solutions include clarity, color, 
reconstitution time and water content, pH, sterility, endotoxins/pyrogens, and 
particulate matter. Suspensions for injection should include additional particle 
size distribution, redispersability, and rheological properties. Emulsion for 
injection should include phase separation, viscosity, mean size, and distribution 
of dispersed globules. 
Large - Volume Parenterals Color, clarity, particulate matter, pH, sterility, endotoxin/
pyrogen, and volume. 
7.1.10 SUMMARY AND CONCLUSIONS 
Studies on the stability of pharmaceutical products provide information on how the 
quality of excipients, APIs, and drug products varies with time under the infl uence 
of various environmental factors such as temperature, humidity, and light and help 
determine shelf life and recommended storage conditions for the life cycle of products. 
Good manufacturing practice regulations and GMP guidelines require that 
stability studies on pharmaceutical products be conducted and shelf life be determined 
based on the results from stability studies. Accelerated stability studies at 
elevated temperature and humidity and long - term stability studies at more moderate 
temperature and humidity conditions are performed during drug development 
and approval process to predict the shelf life of pharmaceutical products. The proposed 
shelf life is then confi rmed by performing long - term studies for the duration 
of the shelf life or longer. Stress studies using acid, base, temperature, oxidation, and 
light stresses are conducted to predict the degradation products that may be formed 
during the accelerated and/or long - term studies and to develop stability - indicating 
methods required for stability evaluation. Photostability and temperature studies 
also help determine the packaging confi guration as well as make recommendations 
for storage conditions during the shelf life. For a realistic assessment of shelf life in 
diverse climates that exist around the world, fi ve climatic zones are identifi ed and 

the long - term storage condition for shelf life determination in these climatic zones 
is based on humidity and temperature conditions likely to prevail in those climatic 
zones. Quality parameters for evaluation of stability of pharmaceutical products 
depend on the chemical nature of the active ingredient being studied as well as the 
type of dosage form of the drug product. 
REFERENCES 
1. U.S. Pharmacopeia ( 2006 ), Stability considerations in dispensing practices, general chapter 
. 1191 . , U. S. Pharmacopeial Convention, Rockville, MD. 
2. U.S. Food and Drug Administration (FDA) (2007) Current good manufacturing practice 
regulations, 21 CFR Parts 210 and 211, FDA, Rockville, MD. 
3. U.S. Food and Drug Administration (FDA) ( 2001 ), Guidance for industry, Q7A Good 
manufacturing practice guidance for active pharmaceutical ingredients, FDA, Rockville, 
MD. 
4. U.S. Pharmacopeia ( 2007 ), Good manufacturing practices for bulk pharmaceutical excipients, 
general chapter . 1078 . , U.S. Pharmacopeial Convention, Rockville, MD. 
5. World Health Organization (WHO) ( 2003 ), Good manufacturing practices for 
pharmaceutical products: Main principles, WHO Technical Report Series 908, WHO, 
Geneva. 
6. World Health Organization (WHO) ( 2006 ), Stability testing of active substances and 
pharmaceutical products, Working Document QAS/06.179, WHO, Geneva. 
7. U.S. Food and Drug Administration (FDA) ( 2003 ), Guidance for industry, ICH Q1A 
(R2), Stability testing of new drug substances and products, FDA, Rockville, MD. 
8. U.S. Food and Drug Administration (FDA) ( 1997 ), Guidance for industry, ICH Q1B, 
Photostability testing of new drug substances and products, FDA, Rockville, MD. 
9. International Conference on Harmonization (ICH) ( 1997 ), Guidance for industry, ICH 
Q1C, Stability testing of new dosage forms, ICH, Brussels. 
10. International Conference on Harmonization (ICH) ( 2003 ), Guidance for industry, ICH 
Q1D, Bracketing and matrixing designs for stability of drug substances and drug products, 
ICH, Brussels. 
11. U.S. Food and Drug Administration (FDA) ( 2004 ), Guidance for industry, ICH Q1E, 
Evaluation of stability data, FDA, Rockville, MD. 
12. U.S. Food and Drug Administration (FDA) ( 2000 ), Guidance for industry, Analytical 
procedures and methods validation — Chemistry, manufacturing, and controls documentation, 
FDA, Rockville, MD. 
13. Lymann , W. J. , Reehl , W. F. , and Rosenblatt , D. H. ( 1982 ), Handbook of Chemical Property 
Estimation Methods , McGraw - Hill , New York . 
14. Albini , A. , and Fasani , E. ( 1998 ), Photochemistry of Drugs: An Overview of Practical 
Problems, Drugs, Photochemistry and Photostability , The Royal Society of Chemistry , 
London . 
15. Carstensen , J. T. ( 1990 ), Drug Stability — Principles and Practices , Marcel Dekker , 
New York . 
16. Drew , H. D. , Brower , J. F. , Juhl , W. E. , and Thornton , L. K. ( 1998 ), Quinine photochemistry: 
A proposed chemical actinometer system to monitor UV exposure in photostability 
studies of pharmaceutical drug substances and drug products , Pharmacopeial Forum , 
24 ( 3 ), 6334 . 
REFERENCES 581

582 STABILITY AND SHELF LIFE OF PHARMACEUTICAL PRODUCTS 
17. Drew , H. D. , Thornton , L. K. , Juhl , W. E. , and Brower , J. F. ( 1998 ), An FDA/PhRMA 
interlaboratory study of the International Conference on Harmonization ’ s proposed 
photostability testing procedures and guidelines , Pharmacopeial Forum , 24 ( 3 ), 6317 . 
18. Sager , N. , Baum , R. G. , Wolters , R. J. , and Layloff , T. ( 1998 ), Photostability studies of 
pharmaceutical products , Pharmacopeial Forum , 24 ( 3 ), 6331 . 
19. Farmer , S. , anderson , P. , Burns , P. K. , and Velagaleti , R. ( 2002 , May), Forced degradation 
of ibuprofen in bulk drug and tablets and determination of specifi city, selectivity, and the 
stability - indicating nature of the USP ibuprofen assay method , Pharm. Technol. , 28 – 42 . 
20. Farmer , S. , McCauslin , L. , Burns , P. K. , and Velagaleti , R. ( 2004 , Aug.), Photosensitivity of 
internal standard valerophenone used in USP ibuprofen bulk drug and tablet assay and 
its effect on quantitation of ibuprofen and its impurities , Pharm. Technol. , 68 – 74 . 
21. Grimm , W. ( 1985 ), Storage conditions for stability testing — Long term testing and stress 
tests (part I) , Drugs Made in Germany , 28 , 196 – 202 . 
22. Grimm W. ( 1986 ), Storage conditions for stability testing — Long term testing and stress 
tests (part II) , Drugs Made in Germany , 29 , 39 – 47 . 
23. U.S. Food and Drug Administration (FDA) ( 2004 ), Guidance for industry, ICH Q1F, 
Stability data package for registration applications in climatic zones III and IV, FDA, 
Rockville, MD. 
24. World Health Organization (WHO) ( 2001 ), Stability testing of pharmaceutical products 
containing well established drug substances in conventional dosage form, WHO Technical 
Report Series 863, Annex 5, WHO, Geneva. 
25. Grimm W. ( 1998 ), Extension of International Conference on Harmonization tripartite 
guideline for stability testing of new drug substances and products to countries of climatic 
zones III and IV , Drug Dev. Ind. Pharm. , 24 , 313 – 325 . 
26. World Health Organization (WHO) ( 2006 ), Assignment of climatic zones and recommended 
storage conditions, Working Document QAS/06.179, Annex 1, WHO, Geneva. 
27. World Health Organization (WHO) ( 2006 ), Testing parameters (for dosage forms), 
Working Document QAS/06.179, Annex 2, WHO, Geneva. 

583 
7.2 
DRUG STABILITY 
Nazario D. Ramirez - Beltran, 1 Harry Rodriguez, 2 and 
L. Antonio Estevez 1 
1University of Puerto Rico, Mayag u ez, Puerto Rico 
2 Cordis LLC, a Johnson & Johnson Company, San German, Puerto Rico 
Contents 
7.2.1 Overview of Stability 
7.2.1.1 Introduction 
7.2.1.2 Regulatory Requirements 
7.2.1.3 Stability and Shelf Life 
7.2.1.4 Short - and Long - Term Stability Studies 
7.2.1.5 Statistical Considerations 
7.2.2 Design of Stability Studies 
7.2.2.1 Introduction 
7.2.2.2 Basic Design Considerations 
7.2.2.3 Design of Stability Studies 
7.2.3 Long - Term Stability Analysis 
7.2.3.1 Introduction 
7.2.3.2 Drug Shelf Life for Single Batch 
7.2.3.3 Drug Shelf Life for Multiple Batches 
7.2.3.4 Shelf Life Estimation for Multiple Factors 
7.2.4 Short - Term Stability Analysis 
7.2.4.1 Introduction 
7.2.4.2 Chemical Reaction Kinetics 
7.2.4.3 Degradation Data Evaluation at Accelerated Conditions 
7.2.4.4 Preliminary Shelf Life Calculation from Stressed Data 
7.2.5 Concluding Remarks 
Appendix Computer Programs 
References 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

584 DRUG STABILITY 
7.2.1 OVERVIEW OF STABILITY 
7.2.1.1 Introduction 
Every drug product on the market requires the expiration date to be specifi ed on 
the immediate container label [1] . Regulations from the U.S. Food and Drugs 
Administration (FDA) and the International Conference on Harmonization (ICH) 
require that pharmaceutical companies present factual evidence supporting the 
shelf life for either existing or new products. The presence of the right amount of 
the active ingredients in a pharmaceutical formulation is extremely important for 
the drug product to be an effective medicament. A rigorous protocol is usually 
implemented to measure the amount of the active ingredient through time for a 
given drug product and the collected data are analyzed to estimate the shelf life, 
which leads to the calculation of the expiration date. Thus, the expiration date provides 
the consumer the confi dence that the drug product will retain its identity, 
strength, quality, and purity throughout the expiration period. 
The main purpose of this chapter is to describe a set of practical tools to design 
a stability study as well as to discuss and illustrate how to determine the shelf life 
for a given drug product. This chapter is organized as follows. This section presents 
a general overview of the stability studies and considers the current regulatory 
requirements. The stability study not only characterizes the degradation of the active 
ingredient for a drug product over time but also provides the basis to establish 
the shelf - life period. This section will also present a brief description of short - 
and long - term stability studies and the important statistical requirements and 
issues stated in the FDA and ICH guidelines. Section 7.2.2 presents the fundamentals 
for designing a stability study. The design of a stability study intends to establish 
a procedure to extract reliable data to determine the shelf life, which is based on 
testing a limited number of batches of a drug product and will be applicable to all 
future batches of the drug product that will be manufactured under similar conditions. 
Designing long - term stability studies usually involves the determination of 
the following factors: number of batches, strength, packaging confi gurations, and 
storage conditions. Essentially, the stability process design includes the application 
of matrixing and bracketing design. Section 7.2.3 presents a description of the conventional 
procedure to determine the shelf life, which is established based on long - 
term stability studies that are conducted under normal storage conditions. Calculation 
of a drug product shelf life for single and multiple batches will be discussed and 
illustrated by examples. A set of computer programs are included in the Appendix 
to carry out the stability calculations. Section 7.2.4 discusses the accelerated stability 
testing procedure, which is often used to provide a tentative shelf life. At the early 
stage of a drug product development, the primary goal of the stability studies is to 
determine the rates of chemical and physical reactions and their relationships with 
storage conditions such as temperature, moisture, light, and others. A short - term 
stability study is conducted under stressed conditions to increase the rate of chemical 
and physical degradation of the drug product. This section will discuss the chemical 
reaction kinetics and the statistical procedure to establish a preliminary shelf 
life. This section will cover the regression techniques to estimate the parameters of 
the Arrhenius equation. 

7.2.1.2 Regulatory Requirements 
The FDA requires pharmaceutical industries to establish a stability testing program 
for each drug product [1 – 3] . Matthews [4] reviews the regulatory situation in Europe. 
The purpose of the stability program is to design the appropriate procedure to 
extract the assay data and calculate the shelf life to determine the expiration date 
of a drug product. The program should be described in a written protocol including 
all the requirements established by the regulation. The FDA establishes that the 
requirements to assess stability are the following [1] : 
1. A sample size and test intervals for each attribute to be tested should be speci- 
fi ed based on statistical rationale. 
2. The storage conditions for the samples should be specifi ed. The storage 
conditions should be the same as those specifi ed in the drug product 
labeling. 
3. All test methods used for each of the drug product attributes need to be quali- 
fi ed to demonstrate adequacy and reliability. 
4. The stability samples should be stored in the same packaging confi guration 
used for the fi nal drug product to be marketed. 
5. If the drug product needs to be reconstituted, testing of the drug product 
before and after reconstitution should be provided. 
6. The number of batches to be used for the stability program for a new drug 
product is at least three. Different batches of the drug substance should be 
used. 
7. The pharmaceutical industry should maintain records for all data, protocols, 
and reports related to the stability program. 
7.2.1.3 Stability and Shelf Life 
The shelf life is the period of time for which the drug product is assured to maintain 
its identity, strength, quality, and purity when stored at the conditions specifi ed on 
the labeling. The expiration date marks the end of the shelf life period. The shelf 
life and the expiration date for a drug product are determined by carrying out a 
stability study. The stability of the drug product is assessed by testing all the attributes 
required to release the product to the market. The tests are carried out at 
specifi ed time intervals for a determined period of time. The expiration date is 
usually included on the label of the drug product. 
7.2.1.4 Short - and Long - Term Stability Studies 
The short - term stability study is conducted over extreme environmental conditions 
to increase the rate of chemical degradation. The data obtained from short - term 
stability studies are normally used to evaluate longer term chemical effects at nonextreme 
storage conditions, but they are also helpful to assess the effect of short - 
term excursions outside the recommended storage conditions that might occur 
during shipping. However, the results from short - term stability studies cannot always 
OVERVIEW OF STABILITY 585

586 DRUG STABILITY 
be used to predict physical changes of the drug product. The recommended testing 
frequency should be a minimum of three time points, including the initial and fi nal 
time points, for a six - month study. If signifi cant changes are expected, additional or 
more frequent testing should be conducted as well as stability testing at intermediate 
conditions. The actual extreme and intermediate conditions depend on the recommended 
storage conditions for drug product, for example, room temperature or 
under refrigeration. A signifi cant change occurs if one or more of the following 
events take place: 
1. A 5% change in assay from its initial value or failure to meet the acceptance 
criteria for potency when using biological or immunological procedures. 
2. Presence of any impurity at a level exceeding its acceptance criterion. 
3. Failure to meet an acceptance criterion for appearance, physical attribute, or 
functionality test (e.g., color, phase separation, resuspendibility, caking, hardness, 
or dose delivery per actuation). However, some changes in physical 
attributes (e.g., softening of suppositories, melting of creams) may be expected 
under accelerated conditions. 
4. Failure to meet the acceptance criterion for pH. 
5. Failure to meet the acceptance criteria for dissolution of 12 dosage units. 
The long - term stability study is performed under the recommended storage conditions 
and packaging confi guration to be used for marketing the drug product and 
during the shelf life period, which is displayed on the product label. The recommended 
testing frequency should be every three months over the fi rst year, every 
six months over the second year, and annually thereafter. The general storage conditions 
used for long - term, intermediate, and short - term stability studies for climatic 
regions I and II, which includes Europe, Japan, and the United States [5, 6] , are 
shown in Table 1 . Alternate storage conditions for long - term, intermediate, or short - 
term stability study could be used if justifi ed. 
TABLE 1 Storage Conditions for Stability Studies 
Study Condition Storage Conditions 
Minimum Time Period Covered 
at Submission (Months) 
Product Stored at Room Temperature 
Long Term 25 ° C ± 2 ° C, 60% RH ± 5% RH 12 
Intermediate 30 ° C ± 2 ° C, 60% RH ± 5% RH 6 
Accelerated 40 ° C ± 2 ° C, 75% RH ± 5% RH 6 
Product Stored in a Refrigerator 
Long Term 5 ° C ± 3 ° C 12 
Accelerated 25 ° C ± 2 ° C, 60% RH ± 5% RH 6 
Product Stored in a Freezer 
Long Term . 20 ° C ± 5 ° C 12 
Note : RH, relative humidity. 

7.2.1.5 Statistical Considerations 
It is recommended to use an appropriate statistical method to analyze the data 
generated during a stability study. The purpose of using statistics is to establish, with 
a high degree of confi dence, the shelf life, that is, the period during which a quantitative 
attribute of the drug product remains within acceptance criteria for all future 
batches manufactured, packaged, and stored under similar conditions. Stability 
studies are expensive and time consuming and statistics can certainly help in this 
aspect. Statistical design principles can be applied to reduce the amount of testing 
required [7] . 
The shelf life for a single batch is usually computed based on regression techniques. 
An appropriate approach to shelf life estimation when using regression 
analysis is by calculating the earliest time at which the 95% confi dence limit for the 
mean intersects the proposed acceptance criterion [8] . A detailed description of 
shelf life calculations is provided in Sections 7.2.3 and 7.2.4 . 
The analysis of covariance is the conventional statistical tool to determine whether 
or not the degradation lines from several batches belong to a single population. If 
the degradation lines are statistically different, the minimum criterion is used to 
determine the shelf life for the current and future batches. According to the minimum 
criterion, the shelf life of future batches corresponds to the shortest of all shelf lives 
from the tested batches. On the other hand, if the batches belong to a single population, 
regression analysis is used to develop a single degradation line based on pooled 
data of the tested batches. 
A short - term stability study is conducted under stressed conditions to increase 
the rate of chemical and physical degradation of the drug product. The classical 
statistical procedure to establish a preliminary shelf life is based on regression techniques, 
which are used to estimate the parameters of the Arrhenius equation. 
A systematic approach for stability data evaluation should be used to determine 
whether extrapolation beyond the time period covered in the stability study for the 
long - term data is appropriate to calculate the shelf life period. The approach consists 
of evaluating any signifi cant change at the accelerated condition and, if applicable, 
at the intermediate condition in order to follow a guideline on how to establish the 
shelf life period by extrapolation. The required relevant supporting data and the 
commitment to place stability batches to be tested until the end of the extrapolated 
shelf life should be provided to the regulatory agency. The relevant supporting data 
include satisfactory long - term data from development batches manufactured with 
close related formulation in a small scale and stored in a packaging confi guration 
similar to the primary stability batches. The data evaluation approach is executed 
as follows [8] . 
Product to be Stored at Room Temperature For products to be stored at room 
temperatures the extent of extrapolation will depend on the evaluation of the results 
from the short - term, intermediate, and long - term data. If no signifi cant change at 
the accelerated (short - term) condition is observed, then the long - term data are 
evaluated for change and variability: 
• Long - Term Data with Small or No Change over Time and Small or No 
Variability In this case a statistical analysis may be considered unnecessary 
OVERVIEW OF STABILITY 587

588 DRUG STABILITY 
but a justifi cation should be provided. Extrapolation can be proposed and 
the resulting shelf life can be up to twice the period covered in the stability 
study for the long - term data but not more than 12 months beyond the period 
tested. 
• Long - Term Data with Change over Time and Variability In this case the statistical 
analysis of the long - term data can be useful to determine the shelf life 
of the drug product. If there is statistical difference among batches, among 
factors, or among factor combinations, it is not possible to perform data pooling 
and the shelf life will correspond to the shortest of all shelf lives obtained from 
all batches, factors, or factor combinations. If the statistical difference is related 
to a particular factor (e.g., strength or packaging), different shelf lives can be 
assigned to each level of that factor. The extent of extrapolation depends on 
whether statistical analysis is applicable to the stability data: 
1. Data for which statistical analysis does not apply: The proposed shelf life 
can be up to 1.5 times but should not be more than 6 months beyond the 
period covered in the stability study for the long - term data. This will require 
relevant supporting data to show that the drug product will meet all attribute 
acceptance criteria by the end of the proposed shelf life period. 
2. Data for which statistical analysis does apply: When statistical analysis is 
applicable to the long - term data and it is not performed, a justifi cation is 
required and the extent of extrapolation is the same as when statistical 
analysis does not apply. If the statistical analysis is performed, the extent of 
extrapolation can be up to twice but not more than 12 months beyond the 
period covered in the stability study for the long - term data. 
If signifi cant change at any time during the 6 - month period of the accelerated (short - 
term) condition is observed, then stability testing at an intermediate condition is 
required and the extent of extrapolation will depend on the outcome of both the 
long - term and intermediate conditions: 
• Data with No Signifi cant Change at Intermediate Conditions In this case 
extrapolation of the long - term data can be proposed and the extent of extrapolation 
will depend on whether a statistical analysis is applicable. 
Data for which statistical analysis does not apply: The proposed shelf life can be 
up to 3 months beyond the period covered in the stability study for the long - term 
data. This will require relevant supporting data to show that the drug product 
will meet all the attribute acceptance criteria by the end of the proposed shelf 
life period. 
Data for which statistical analysis does apply: When statistical analysis is 
applicable to the long - term data and it is not performed, a justifi cation is 
required and the extent of extrapolation is the same as when statistical analysis 
does not apply. If the statistical analysis is performed, the extent of extrapolation 
can be up to 6 months beyond the period covered in the stability 
study for the long - term data. This will require relevant supporting data to 
show that the drug product will meet all the attribute acceptance criteria by 
the end of the proposed shelf life period. 
0 
0 

• Data with Signifi cant Change at Intermediate Conditions In this case extrapolation 
of the long - term data is not allowed and the shelf life should not exceed 
the period covered by the long - term stability study. 
Product to be Stored In a Refrigerator For products to be stored in a refrigerator 
the same data evaluation approach but with more restrictive extent of extrapolation 
as for product to be stored at room temperature will be used unless otherwise speci- 
fi ed below. 
If no signifi cant change at the the accelerated (short - term) condition is observed, 
then the long - term data is evaluated for change and variability: 
• Long - Term Data with Small or No Change over Time and Small or No 
Variability Extrapolation can be proposed and the resulted shelf life can be 
up to 1.5 times but should not be more than six months beyond the period 
covered in the stability study for the long - term data. 
• Long - Term Data with Change over Time and Variability The extent of 
extrapolation depends on whether statistical analysis is applicable to the stability 
data: 
Data for which statistical analysis does not apply: The proposed shelf life can 
be up to three months beyond the period covered in the stability study for 
the long - term data. 
Data for which statistical analysis does apply: When statistical analysis applies 
and is not performed, a justifi cation is required and the extent of extrapolation 
is the same as when statistical analysis does not apply. If the statistical 
analysis is performed, the extent of extra polation can be up to 1.5 times but 
not more than six months beyond the period covered in the stability study 
for the long - term data. 
If signifi cant change at any time during the six - month period of the accelerated 
(short - term) condition is observed, then extrapolation is not considered appropriate. 
If the long - term data show variability, the proposed shelf life can be verifi ed by 
statistical analysis. 
Product to be Stored in a Freezer For a drug product to be stored in a freezer the 
shelf life is determined from the long - term stability data and no extrapolation is 
allowed. 
7.2.2 DESIGN OF STABILITY STUDIES 
7.2.2.1 Introduction 
The FDA guidelines establish, fi rst, that the stability study protocol must describe 
how the stability study is designed and carried out and, second, the statistical methods 
to be used to analyze the data. The design of a stability study is aimed at establishing 
an expiration dating period. The expiration dating period is based on testing on a 
DESIGN OF STABILITY STUDIES 589 
0 
0 

590 DRUG STABILITY 
limited number of batches of a drug product, and the results of these tests are applied 
to future batches of the drug product manufactured under similar conditions. Therefore, 
the study design should reduce the bias and identify and control any expected 
or unexpected source of variations. An appropriate stability design should provide 
the highest accuracy and precision of the established shelf life. 
In this section, the fundamentals for designing a stability study are described. 
A stability study will characterize the degradation of the active ingredient 
with respect to time and is required to determine the shelf life period used to 
calculate the expiration date of a drug product. The shelf life is the maximum 
allowable period of time for the drug product to be stored in its fi nal packaging 
while maintaining the therapeutic amount of the active pharmaceutical ingredient 
(API). The expiration date marks the end of the shelf life period. The expiration 
date is calculated by adding the shelf life to the manufacturing end date of the drug 
product. The determined shelf life is applicable to all future batches of the drug 
product manufactured under similar conditions as the batches used in the stability 
study. 
The design of a stability study should be based on knowledge of the behavior 
and properties of the drug substance obtained from stability studies on the drug 
substance and data generated during clinical formulation studies. The stability study 
is performed by establishing and executing a protocol. The stability study protocol 
should specify all the aspects to be considered (e.g., sample size, test methods, and 
acceptance criteria) while carrying out the stability study. It is important to comply 
with all FDA (or ICH) regulations when performing a stability study. 
Designing a stability study is based on a factorial design of experiments where a 
systemic procedure is used to determine the effect on the response variable of 
various factors and factor combinations. A linear model is used to represent the 
relationship between the factors and factor combinations with the response variable. 
Once the experimental design is established, the assays are conducted and stability 
data are saved to fi nally estimate the shelf life period. 
Typically, the following factors are involved in the design of stability studies: 
number of batches, strength, package confi guration, and storage conditions. The 
typical response variable of the stability study is the API amount. However, any 
other response variables (drug product attributes) that are susceptible to change 
during storage and are likely to infl uence quality, safety, and effi cacy of the drug 
product (e.g., physical appearance, sterility, drug release rate, impurities, and degradation 
products) should also be considered. The testing procedure should include, 
as appropriate, the physical, chemical, biological, and microbiological attributes, 
preservative contents (e.g., antioxidant or antimicrobial preservatives), and functionality 
tests (e.g., for a dose delivery system). 
Section 7.2.2 describes practical guidelines to design stability studies that are 
aligned with the guidance for industries provided by the FDA [9] . A different 
approach can be used depending on the characteristics of the drug product and with 
an appropriate scientifi c justifi cation. 
7.2.2.2 Basic Design Considerations 
When designing a stability study, the following aspects should be taken into 
account: 

Preliminary Stability Data During the drug product design phase, stability 
data are generated to choose the fi nal package confi guration and storage 
conditions and characterize the product. Such data are helpful in the development 
of a good stability study design that can offer as much information as 
possible with as few data as possible. Also, the preliminary stability data may 
provide scientifi c justifi cation to the selection of the type of stability study 
design. 
Drug Product Attributes All attributes that may affect the quality of the drug 
product have to be included in design of a stability study. These attributes have 
to be tested at every time period and, thus, the number of samples needed at 
each time should be set accordingly. 
Drug Product Specifi cations The specifi cation for each attribute of the drug 
product is required to defi ne the acceptance criteria of the stability study. 
Test Methods All test methods used in the stability study should be previously 
qualifi ed. 
Preliminary Testing Data Any data generated during the developmental and 
clinical trial phases of the drug may be useful to determine the expected 
manufacturing process variability. The process variability is an important 
factor to defi ne the sampling plan to be used in the stability study. 
Design Factors Identifi cation of the design factors is crucial when choosing the 
design of the stability study. It is important to identify the most relevant design 
factors that may affect the stability of the drug product during storage. Failure 
to identify a relevant design factor may cause a signifi cant delay in the stability 
study completion and submission to the regulatory agency, for example, the 
FDA. 
Full Design versus Reduced Design A full design requires testing the drug 
product for factor combinations and at all time periods. The full design provides 
the largest amount of information to determine the shelf life of a drug 
product. However, the number of required combinations increases exponentially 
with the number of factors. On the other hand, a reduced design requires 
drug product testing only at a fraction of factor combinations. The reduced 
design requires less testing effort but entails the potential risk to result in a 
shorter shelf life than the one obtained from a full design due to the reduced 
amount of data collected. 
7.2.2.3 Design of Stability Studies 
Full Stability Study Design Assume that a stability study includes three factors: 
batch, strength (denoted S1, S2, and S3 in what follows), and packaging size (denoted 
P1, P2, and P3). Each value of a factor is normally referred to as level . Therefore, if 
there are four packaging sizes, the factor called packaging size has four levels. It 
should be noted that the FDA requires at least three batches and consequently the 
batch factor will have three levels. If the other two factors have three levels each, 
then the required number of experiments is 27. Table 2 shows the 27 experiments 
that should be performed at each point in time. The number of combinations, C , can 
be readily calculated by multiplying the number of levels of each factor: 
DESIGN OF STABILITY STUDIES 591

592 DRUG STABILITY 
C n = . . . . LF LF LF LF 1 2 3  
where LF 1 = number of levels of factor 1 
LF 2 = number of levels of factor 2 
LF 3 = number of levels of factor 3 
LF n = number of levels of factor n 
Once the factor combinations are defi ned, the number of samples required for 
the study can be determined. Assuming that the only testing required for the stability 
study design specifi ed in Table 2 is amount of drug and that the test requires one 
package of the drug product, the amount of samples required for the study is 27 
packages per time period. A drug product package for each of the 27 combinations 
has to be included in the study for each time period. 
For a proposed shelf life of 36 months, testing of the drug product has to be 
carried out at the following time periods ( t ): 0, 3, 6, 9, 12, 18, 24, and 36 months. 
Considering the full design shown in Table 2 , a total of 216 samples are needed to 
perform all required testing. The fi rst 27 samples are tested at the beginning of the 
study to obtain data for t = 0. All other samples are placed in an environmental 
chamber that will maintain the temperature and relative humidity as specifi ed in 
the stability study protocol. At each testing time, 27 additional samples are pulled 
from the chamber for testing, as shown in Table 3 . The data analysis for stability 
studies will be discussed in section 7.2.3 . 
Reduced Stability Study Designs There are occasions where it is not possible to 
obtain the total number of required samples to perform a full stability study design 
or it is simply desired to reduce the sampling requirement; this is done by selecting 
a reduced stability study design. To perform a reduced stability study design, some 
assumptions are needed which should be fully justifi ed; the justifi cations should be 
provided to the regulatory agency. Usually, a reduced design applies when a preliminary 
stability study is performed. 
Bracketing and matrixing are the two reduced designs recommended by the FDA 
[9] . Each of these methods applies to different situations. Using both of them simultaneously 
may reduce the ability of the study to determine the shelf life since factor 
combinations can be confounded due to the aliasing effect [10] . 
Bracketing The bracketing design consists in testing only two levels: the highest 
and lowest values of the factor. Bracketing requires showing that the selected levels 
are the extremes of the factor range. If the stability of the extreme levels is different, 
TABLE 2 Factor Combinations for Three Factors with Three Levels Each 
Batch 
S1 S2 S3 
P1 P2 P3 P1 P2 P3 P1 P2 P3 
1 C1 C2 C3 C10 C11 C12 C19 C20 C21 
2 C4 C5 C6 C13 C14 C15 C22 C23 C24 
3 C7 C8 C9 C16 C17 C18 C25 C26 C27 

it is expected that the stability of any intermediate level should be higher than the 
stability of the least stable extreme. Table 4 shows the bracketing design for the full 
design example presented in Table 2 . 
Note that bracketing was not applied to the batch factor because the FDA regulation 
requires testing at least three batches to determine a drug product shelf life. 
Even so, the sampling required for the bracketing design was reduced substantially. 
The sample size required per time period is 12, a small number when compared to 
TABLE 3 Testing Schedule for Full Stability Study Design 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T T T T T T T T 
C2 T T T T T T T T 
C3 T T T T T T T T 
C4 T T T T T T T T 
C5 T T T T T T T T 
C6 T T T T T T T T 
C7 T T T T T T T T 
C8 T T T T T T T T 
C9 T T T T T T T T 
C10 T T T T T T T T 
C11 T T T T T T T T 
C12 T T T T T T T T 
C13 T T T T T T T T 
C14 T T T T T T T T 
C15 T T T T T T T T 
C16 T T T T T T T T 
C17 T T T T T T T T 
C18 T T T T T T T T 
C19 T T T T T T T T 
C20 T T T T T T T T 
C21 T T T T T T T T 
C22 T T T T T T T T 
C23 T T T T T T T T 
C24 T T T T T T T T 
C25 T T T T T T T T 
C26 T T T T T T T T 
C27 T T T T T T T T 
Note : T = sample tested. 
TABLE 4 Factor Combinations for Bracketing of Three Factors with Three Levels 
Batch 
S1 S2 S3 
P1 P2 P3 P1 P2 P3 P1 P2 P3 
1 C1 — C3 — — — C19 — C21 
2 C4 — C6 — — — C22 — C24 
3 C7 — C9 — — — C15 — C27 
DESIGN OF STABILITY STUDIES 593

594 DRUG STABILITY 
the 27 samples required for the full design. As a result, the amount of samples 
required for the whole stability study is 8 . 12, or 96, whereas the full - design study 
required 216 samples. The execution procedure of stability testing for both a complete 
and a reduced stability study is the same, as shown in Table 5 . 
Matrixing The matrixing design consists in selecting a fraction of the total number 
of possible combinations included in the full design. At each time, a different fraction 
will be tested. The main assumption is that the stability of each fraction represents 
the stability of all factor combinations at that particular time. The matrixing 
design can be applied across the factors or across the time points of the stability 
design. The degree of reduction (e.g., one - half or one - third) from a full design 
depends on the quantity of factors to be considered. The larger the number of 
factors and levels included in the full design, the larger the degree of reduction that 
can be implemented. 
Applying a matrixing design on time points only, all factor combinations 
(full factorial design) should be tested at the initial and fi nal points in time, while a 
TABLE 5 Testing Schedule for Bracketing Stability Study Design 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T T T T T T T T 
C2 — — — — — — — — 
C3 T T T T T T T T 
C4 T T T T T T T T 
C5 — — — — — — — — 
C6 T T T T T T T T 
C7 T T T T T T T T 
C8 — — — — — — — — 
C9 T T T T T T T T 
C10 — — — — — — — — 
C11 — — — — — — — — 
C12 — — — — — — — — 
C13 — — — — — — — — 
C14 — — — — — — — — 
C15 — — — — — — — — 
C16 — — — — — — — — 
C17 — — — — — — — — 
C18 — — — — — — — — 
C19 T T T T T T T T 
C20 — — — — — — — — 
C21 T T T T T T T T 
C22 T T T T T T T T 
C23 — — — — — — — — 
C24 T T T T T T T T 
C25 T T T T T T T T 
C26 — — — — — — — — 
C27 T T T T T T T T 
Note : T = sample tested. 

fraction of the full factorial design is tested at intermediate points in time. If the full 
long - term stability data required for the proposed shelf life will not be available for 
review before submission to the regulatory agency, the full factorial design should 
be tested at 12 months and at the last scheduled point in time prior to submission. 
Therefore, the FDA regulations require testing full factorial design at 0, 12, and 36 
months for a matrixing design. In addition, data from at least three time points, 
including initial, should be available for each factor combination through the fi rst 
12 months of the study. For matrixing at accelerated storage conditions, testing 
should occur at a minimum of three time points, including initial and fi nal, for each 
factor combination. 
The matrixing design should be as balanced as possible so that each factor combination 
is tested to the same extent over the duration of the study. A matrixing 
design is applicable when the supporting data indicate predictable product stability 
and small variability. A statistical justifi cation for using a matrixing design could be 
based on evaluating the proposed matrixing design with respect to its power to 
detect differences in the degradation rates among the factors or the level of accuracy 
on estimating the shelf life. 
A matrixing design performed across factors other than time points generally has 
less precision in shelf - life estimation and yields a shorter shelf life than the corresponding 
full design due to the confounding and aliasing effects [10] . Matrixing 
design may have insuffi cient power to detect main factors or factor interaction 
effects. Thus an excessive reduction in the number of factor combinations may 
produce an unreliable estimation of the shelf life due to the missing factor combinations. 
On the other hand, a matrixing design on time points would often have similar 
ability as the full design to detect differences in rates of change among factors and 
to establish a reliable shelf life. This is because full testing of all factor combinations 
would still be performed at both the initial and the last time points. For illustration 
purposes, matrixing design will be implemented over time points, factors, and in both 
time points and factors. The reader should be wise when choosing a design for a 
particular stability study and should have in mind that supporting data and justifi cations 
have to be provided. 
Matrixing in Time Points When matrixing in time points design, a fraction of the 
combinations is selected using the fractional factorial design procedure. Note that 
the full design presented above is equivalent to a 3 k factorial design, where k is the 
number of factors and 3 is the number of levels of each factor. The total number of 
combinations for three factors and three levels is 3 3 , or 27. Assuming that a one - third 
reduction for the matrixing design is desired, the required sample size per time 
period is 18 compared to 27 for the full design. As a result, the number of samples 
needed for the whole stability study is 5 . 18 + 3 . 27, or 171, whereas the full 
design required 216 samples. Table 6 shows the schedule to implement a matrixing 
design over time points for 36 months of stability study. This table shows a full design 
on times 0, 12, and 36 and two - thirds on the remaining times. The sampling shown 
in Table 6 has been designed to maintain a proper balance. If bracketing was 
applied to one of the factors (either strength or packaging size) or the design has 
only two levels on one of those factors, the full experiment design will be reduced, 
as shown in Table 7 . Table 8 shows a two - thirds matrixing design with three strength 
DESIGN OF STABILITY STUDIES 595

596 DRUG STABILITY 
levels and two packaging size levels. Thus, 18 factor combinations are scheduled for 
time equal to 0, 12, and 36 months and 12 combinations for the remaining time 
points. 
For this matrixing design, the total number of samples needed for the whole stability 
study is 3 . 18 + 5 . 12, or 114, compared to 216 samples needed for the full 
3 k design. Furthermore, for the case where both strength and packaging size have 
TABLE 6 Testing Schedule for Matrixing in Time Point Stability Study Design 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T — T T T — T T 
C2 T T — T T T — T 
C3 T T T — T T T T 
C4 T T — T T T — T 
C5 T T T — T T T T 
C6 T — T T T — T T 
C7 T T T — T T T T 
C8 T — T T T — T T 
C9 T T — T T T — T 
C10 T T T — T T T T 
C11 T — T T T — T T 
C12 T T — T T T — T 
C13 T — T T T — T T 
C14 T T — T T T — T 
C15 T T T — T T T T 
C16 T T — T T T — T 
C17 T T T — T T T T 
C18 T — T T T — T T 
C19 T T — T T T — T 
C20 T T T — T T T T 
C21 T — T T T — T T 
C22 T T T — T T T T 
C23 T — T T T — T T 
C24 T T — T T T — T 
C25 T — T T T — T T 
C26 T T — T T T — T 
C27 T T T — T T T T 
Note : T = sample tested. 
TABLE 7 Factor Combination Considering Two Levels 
in Packaging 
Batch 
S1 S2 S3 
P1 P2 P1 P2 P1 P2 
1 C1 C2 C7 C8 C13 C14 
2 C3 C4 C9 C10 C15 C16 
3 C5 C6 C11 C12 C17 C18 

only two levels or bracketing is applied to both factors, the full experiment design 
is reduced, as shown in Table 9 . In this case, the total number of factor combinations 
is 12 rather than 27. Table 10 shows the design when bracketing and matrixing are 
implemented. Bracketing is applied to factors while matrixing is applied to time 
points. Table 10 shows that there are 12 factor combinations for times 0, 12, and 36 
and 8 factor combinations on the remaining time points. Thus, for this matrixing 
design, the number of samples required for the whole stability study is 3 . 12 + 5 . 
8, or 76, compared to 216 samples needed for the full 3 k design. 
Matrixing in Factors For the matrixing - in - factors design, the factor combinations 
are eliminated in a systematic way as shown in Table 11 . As a result, not all factor 
combinations are tested during the stability study. Such a design can be used when 
the factor combination eliminated exhibits a similar behavior as the other factor 
TABLE 8 Testing Schedule for Matrixing in Time Points (Two Levels for Packaging Size) 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T — T T T — T T 
C2 T T T — T T T T 
C3 T T — T T T — T 
C4 T — T T T — T T 
C5 T T T — T T T T 
C6 T T — T T T — T 
C7 T T T — T T T T 
C8 T T — T T T — T 
C9 T — T T T — T T 
C10 T T T — T T T T 
C11 T T — T T T — T 
C12 T — T T T — T T 
C13 T T — T T T — T 
C14 T — T T T — T T 
C15 T T T — T T T T 
C16 T T — T T T — T 
C17 T — T T T — T T 
C18 T T T — T T T T 
Note : T = sample tested. 
TABLE 9 Factor Combination Considering Two Levels in 
Packaging Size and Strength 
Batch 
S1 S2 
P1 P2 P1 P2 
1 C1 C2 C7 C8 
2 C3 C4 C9 C10 
3 C5 C6 C11 C12 
DESIGN OF STABILITY STUDIES 597

598 DRUG STABILITY 
combinations within the study. The implementation of this matrixing design is shown 
in Table 12 . All factor combinations are tested at times 0, 12, and 36 months, and 
just 18 factor combinations at the remaining times. For this matrixing design, the 
number of samples needed for the whole study is 3 . 27 + 18 . 5, or 171, while 216 
samples were needed for the full 3 k design. 
Matrixing in Factors and Time Points Matrixing in both factors and time points 
is a combination of the two previously mentioned matrixing designs and the schedule 
is shown in Table 13 . In this matrixing design, all factor combinations are tested 
at times 0, 12, and 36 and fractional factorial at the remaining times. The total 
number of experiments to be conducted in this stability study is 3 . 27 + 5 . 12, or 
141, while 216 samples were required for the full 3 k design. 
7.2.3 LONG - TERM STABILITY ANALYSIS 
7.2.3.1 Introduction 
This section considers the stability analysis of the long - term studies. The purpose of 
this section is to provide a set of fundamental statistical tools to calculate the shelf 
life for single and multiple packages and strengths. The statistical methods will be 
TABLE 10 Testing Schedule for Matrixing in Time Points (Two Levels for Packaging 
Size and Strength) 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T — T T T — T T 
C2 T T T — T T T T 
C3 T T — T T T — T 
C4 T — T T T — T T 
C5 T T T — T T T T 
C6 T T — T T T — T 
C7 T T — T T T — T 
C8 T — T T T — T T 
C9 T T T — T T T T 
C10 T T — T T T — T 
C11 T — T T T — T T 
C12 T T T — T T T T 
Note : T = sample tested. 
TABLE 11 Matrixing in Factor Combination Elimination 
Batch 
S1 S2 S3 
P1 P2 P3 P1 P2 P3 P1 P2 P3 
1 — C2 C3 C10 C11 — C19 — C21 
2 C4 — C6 — C14 C15 C22 C23 — 
3 C7 C8 — C16 — C18 — C26 C27 

LONG-TERM STABILITY ANALYSIS 599 
described fi rst under the context of a general application and then numerical examples 
will be described step by step to illustrate particular applications. A set of 
computer programs are provided in the Appendix to facilitate the implementation 
of the stability analysis procedures. Chen et al. [11] pointed out that the stability 
analysis usually consists of three steps. The fi rst step is to collect the assay results at 
several time intervals for the sample stored under appropriate conditions. The 
second step is to select the appropriate model to describe the relationship between 
assay results and sampling times, and the third step is to establish the expiration 
dating period. The fi rst step was described in Section 7.2.2 and the second and third 
steps are described in Section 7.2.3 . 
7.2.3.2 Drug Shelf Life for Single Batch 
Calculation of the shelf life for a drug product in a single package type will be 
illustrated assuming that assay results are obtained from a single batch. The FDA 
guideline establishes that the expiration dating period for a drug product consists 
TABLE 12 Testing Schedule for Matrixing in Factor Stability Study Design 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T — — — T — — T 
C2 T T T T T T T T 
C3 T T T T T T T T 
C4 T T T T T T T T 
C5 T — — — T — — T 
C6 T T T T T T T T 
C7 T T T T T T T T 
C8 T T T T T T T T 
C9 T — — — T — — T 
C10 T T T T T T T T 
C11 T T T T T T T T 
C12 T — — — T — — T 
C13 T — — — T — — T 
C14 T T T T T T T T 
C15 T T T T T T T T 
C16 T T T T T T T T 
C17 T — — — T — — T 
C18 T T T T T T T T 
C19 T T T T T T T T 
C20 T — — — T — — T 
C21 T T T T T T T T 
C22 T T T T T T T T 
C23 T T T T T T T T 
C24 T — — — T — — T 
C25 T — — — T — — T 
C26 T T T T T T T T 
C27 T T T T T T T T 
Note : T = sample tested. 

600 DRUG STABILITY 
of determining the time at which the 95% one - sided lower confi dence interval for 
the mean degradation curve intersects the lower acceptable specifi cation limit, 
which is usually adopted by the FDA as 90% of the label claim (LC). Assuming 
that concentration of a drug product decreases linearly with time and it can be 
expressed as 
y x i n i i i = + + = . . . 1, . . . , (1) 
where y i is the percentage of the label claim (assay result) at a given time x i for the 
i th sample, x i is the time at which the i th sample was analyzed, and . and . are the 
regression parameters. The coeffi cient . represents the percentage label claim when 
x i = 0 and is usually known as the batch effect, while coeffi cient . is known as the 
degradation rate, and the product . x i is the stability loss over time. It is assumed 
that the random variable . i follows a normal distribution with zero mean and a 
constant variance . 2 and n is the total number of samples. In Section 7.2.4 , the 
TABLE 13 Testing Schedule for Matrixing in Factor and Time Point Stability 
Study Design 
Factor Combination 
By Time in Months 
0 3 6 9 12 18 24 36 
C1 T — T T T T T T 
C2 T T — T T — T T 
C3 T — — — T — — T 
C4 T — — — T — — T 
C5 T T T — T T — T 
C6 T — T T T T T T 
C7 T T T — T T — T 
C8 T — — — T — — T 
C9 T T — T T — T T 
C10 T T — T T — T T 
C11 T — — — T — — T 
C12 T — T T T T T T 
C13 T T T — T T — T 
C14 T — T T T T T T 
C15 T — — — T — — T 
C16 T — — — T — — T 
C17 T T — T T — T T 
C18 T T T — T T — T 
C19 T — — — T — — T 
C20 T — T T T T T T 
C21 T T — T T — T T 
C22 T — T T T T T T 
C23 T — — — T — — T 
C24 T T T — T T — T 
C25 T T — T T — T T 
C26 T T T — T T — T 
C27 T — — — T — — T 
Note : T = sample tested. 

LONG-TERM STABILITY ANALYSIS 601 
various degradation kinetic models will be presented. In this context, Equation (1) 
describes a zero - order degradation. 
The drug expiration date for a single batch exhibits a 95% confi dence that the 
average drug characteristic of the dosage units in the batch is within specifi cations 
up to the end of the expiration date. The 95% one - sided lower confi dence bounds 
for the mean degradation line is shown in Figure 1 . 
The 95% confi dence interval can be expressed by the probability statement 
P t t n . ( ) = . 0 05 2 0 95 . . , 
(2) 
where t 0.05, n . 2 is the upper fi ve percentile of the t student distribution with n . 2 
degrees of freedom. Assuming that the percentage of the label claim (%LC) follows 
a normal distribution, the statistic t is defi ned as 
t 
z 
x m 
y x n x x S 
n m 
xx = = 
. + ( ) [ ] + . ( ) [ ] 
. ( ) [ ] 2 
2 2 1 
2 
. . . .
. SSE 2 
(3) 
where 
S x x y y xx 
i 
n 
i 
n 
i = . ( ) = . ( ) 
= = 
. . 
1 
2 
1 
2 SSE . 
(4) 
where SSE is the sum - of - squares error, z is a random variable that follows the 
standard normal distribution, .m 2 
is a random variable that follows the chi - square 
distribution with m degrees of freedom, is the estimated %LC at time x , a and b 
are the estimates of the parameters . and . , respectively, and x. is the average sampling 
time. 
The t statistic was used to eliminate the unknown parameter . 2 . Thus, Equation 
(3) can be written as 
FIGURE 1 Graphical representation of shelf life. 
Percent label 
claim 
100 
90 
1 2 3 4 5 6 7 
Shelf life 
Time (Months) 
95% one-sided 
lower confidence 
bound for mean 
Degradation line 
y

602 DRUG STABILITY 
t 
y x 
Sy 
= 
. + . ( ) 
. 
. . 
(5) 
S 
n 
x x
S n y 
xx 
. 
( ) = + 
. ... 
...
. 
MSE MSE= 
SSE 1 
2 
2 
(6) 
where MSE is the mean - square error. 
Thus, Equation (2) can be now written as 
P
y x 
S 
t 
y 
n 
. ( ) 
. 
. 
. 
. . 
. .. . 
.. .
= . 
. . 
0 05 2 0 95 , 
(7) 
and after arranging terms, Equation (7) can be expressed as 
P y t S x n y . . . . .+ ( )= . 0 05 2 0 95 , . . 
(8) 
Therefore, the 95% one - sided lower confi dence bound is 
L x y t S a bx t 
n 
x x 
S n y n 
xx 
( ) 
( ) 
. = . = . . + 
. ... 
...
. 0 05 2 0 05 2 
2 1 
, MSE . . , . . 
(9) 
The points where L ( x ) intersects the acceptable lower specifi cation limit . can 
be obtained by fi nding the roots of the following equation: . . L ( x ) = 0. It should 
be pointed out that this equation can also be written in the form 
f x a bx t 
n 
x x 
S n 
xx 
( ) [ ( )] 
( ) 
. = . + . + 
. ... 
... 
= . . 2 
0 05 2 
2 
2 
0 , MSE 
1 
(10) 
The quadratic equation (10) has two roots and the shelf life is obtained by computing 
the root of Equation (10) that is smaller than a reference point, which is 
defi ned as 
x 
b 
b ref = 
. 
.
0 
1
. 
(11) 
A numerical example is presented below to illustrate, step by step, the shelf - life 
calculation procedure. 
Example 1 Shelf Life for Single Batch The assay results for a batch of a drug 
product are given in Table 14 . Based on these results, determine the shelf life for 
this batch. 
TABLE 14 Single - Batch Assay Results (%LC) for Example 1 
Sampling time, months 0 3 6 9 12 18 24 36 
Percent label claim 103.5 97.3 97.2 94.2 94.8 94.1 91.3 88.7 

LONG-TERM STABILITY ANALYSIS 603 
Solution First, regression techniques are used to estimate the expected degradation 
line [12] . This yields 
. . . y x = . 99 64 0 3335 
It is well known that conventional statistical computer programs such as 
MINITAB, STATGRAPHICS, or SAS will conduct model fi tting and provide the 
following additional information: t 0.05, n . 2 = t 0.05,6 = 1.943, MSE = 4.1665, n = 8, x. = 13.5, 
and S xx = 1008. For the reader ’ s convenience, a computer program to perform model 
regression fi tting and to compute all the statistics required to determine the shelf 
life is given in the Appendix . 
Thus, the 95% one - sided lower confi dence bound for the mean degradation rate 
is 
L x x 
x 
( ) . . . . 
( .) = . . + .
... 
... 
99 64 0 3335 1 943 4 1665 
1
8 
13 5 
1008 
2 
Consequently, the shelf life can be obtained by using the equation 
f x x 
x 
( ) [ ( . . )] . ( . ) 
( .) = . . . + 
. . 90 99 64 0 3335 1 943 4 1665 
1
8 
13 5 
1008 
2 2 
2 
.. 
... = 0 
(12) 
To compute the shelf life, we compute fi rst the reference point as 
x 
b 
b ref = 
. 
. 
= 
. 
= 0 
1 
99 64 90 
0 3335 
28 90 
. .
. 
. 
Therefore, the shelf life is the root smaller than 28.90. A simple and practical 
tool to compute the roots of Equation (12) is perhaps solving the following 
equivalent problem. Find x such that it minimizes the absolute value of f ( x ). 
This root is obtained by using the quasi - Newton line search (QNLS) algorithm [13] . 
The computer program requires an initial point and we recommend using the 
value 
x x d ( ) 0 = . ref (13) 
where d is a positive value that should be explored by using a trial - and - error method 
until the program converges to a local minimum, usually the range of d is between 
0 and 10. For instance, when d = 0, the QNLS method converges to x R = 23.6989. 
The shelf life x L is the integer part of the root x R and, in this case, the expiration 
dating period of this batch is x L = 23 months since f ( x R ) = 0, and x R < x ref . The QNLS 
method is implemented in MATLAB software and the computer program is also 
given in the Appendix . 

604 DRUG STABILITY 
7.2.3.3 Drug Shelf Life for Multiple Batches 
Test for Poolability of Batches The FDA guidelines establish that at least three 
batches must be used to determine the batch - to - batch variability. Thus, if the statistical 
procedure shows evidence that the three batches belong to the same population, 
a single shelf life for all batches will be determined by pooling data from all batches. 
The FDA guidelines also established that batch similarity of the degradation lines 
can be assessed by the equality of slopes and the equality of intercepts of individual 
batches. 
Assuming that the degradation decreases linearly with time (zero - order degradation), 
it can be represented by the following model, which is analogous to Equation 
(1) : 
y x i I j n ij i i ij ij i = + + = = . . . 1 1 , . . . , , . . . , (14) 
where y ij is the assay result (%LC) of the i th batch for a drug product sampled at 
time x ij , n i is the number of sampling times for the i th batch, I is the total number 
of batches, . i 
and . i 
are the intercept and slope of the degradation line for the i th 
batch, respectively, and . ij is assumed to be a random variable with zero mean and 
constant variance. It is worth mentioning that . i is considered as the batch effect 
and . i as the degradation rate for the i th batch. 
The FDA guidelines indicate that the tests for the equality slopes and the 
equality of intercepts should be performed at the 0.25 level of signifi cance, as 
was suggested by Bancroft [14] . It would be desirable to test whether or not the 
batches belong to a single population and if that is the case to derive a model 
with a single intercept and slope for the entire population of batches. Thus, a 
poolability test is implemented to determine if a single population should be used. 
The poolability test is implemented in two steps: (1) testing for equality of 
slopes and (2) testing for equality of intercepts. The conventional procedure to 
test these hypotheses is accomplished by using the analysis of covariance for a 
completely randomized design [15, 16] . An alternative method that can easily be 
generalized for studying multi factor is using a regression model with indicator 
variables [17] . 
Analysis of Covariance (ANCOVA) for Testing Similarity of Slopes The fi rst step 
is determining whether or not the degradation rates for all batches behave in a 
similar fashion. The following hypothesis will be tested: 
H i j i I j I i j 0 1 1 : , . . . . .= . = = for all , , , . . . , (15) 
where I is the number of batches. 
To develop the F statistics and be able to test the above hypothesis, it is required to 
compute different sum of squares and cross products for %LC and also for sampling 
times [15, 16, 18] . The aggregated sum of squares of the sampling times is defi ned as 
Z S i xx 
i 
I 
xx = 
= .
1 
( ) 
(16) 

LONG-TERM STABILITY ANALYSIS 605 
where 
S i x x x 
x
n 
x x x 
xx 
j
n 
ij i 
j
n 
ij 
i
i 
i 
j
n 
ij i 
i i 
i 
( ) ( ) = . = . 
= = 
= = 
=
. . 
.
1 
2 
1 
2 
2 
1 
i 
i 
i i 
x
n
i
i
i 
(17) 
The aggregated sum of squares of %LC is defi ned as 
Z S i yy 
i 
I 
yy = 
= .
1 
( ) 
(18) 
where 
S i y y y 
y
n 
y y y 
yy 
j
n 
ij i 
j
n 
ij 
i
i 
i 
j
n 
ij i 
i i 
i 
( ) ( ) = . = . 
= = 
= = 
=
. . 
.
1 
2 
1 
2 
2 
1 
i 
i 
i i 
y
n
i
i
i 
(19) 
The aggregated sum of cross products between the %LC and sampling times is 
given as 
Z S i xy 
i 
I 
xy = 
= .
1 
( ) 
(20) 
where 
S i x x y y x y 
x y 
n xy 
j
n 
ij i ij i 
j
n 
ij ij 
i i 
i 
i i 
( ) ( )( ) = . . = . 
= = 
. . 
1 1 
i i 
i i 
(21) 
Thus, the sum - of - squares error based on the aggregated sums is computed as 
SSEa yy 
xy 
xx 
Z 
Z
Z 
= . 
2 
(22) 
The aggregated sum - of - squares error for each batch, which is also needed, is 
calculated as 
SSE SSE = 
= .i 
I 
i 
1 
( ) 
(23) 
where 
SSE( ) ( ) 
( ) 
( ) 
i S i 
S i 
S i yy 
xy 
xx 
= . 
2 
(24) 

606 DRUG STABILITY 
The mean - square errors due to slope contribution and the total mean - square 
errors are defi ned as 
MS 
SSE SSE 
slope = 
. 
. 
a 
I 1 
(25) 
MSE 
SSE 
where = 
. 
= 
= .
N I 
N n 
i 
I 
i 2 1 
(26) 
Finally, the F statistics to test hypothesis (15) can be written as 
F(slope) 
MS 
MSE 
slope = 
(27) 
It should be noted that hypothesis (15) is rejected when F (slope) > F 0.25, I . 1, N . 2 I , 
where F 0.25, I . 1, N . 2 I is the upper percentile of the F distribution with I . 1, and N . 2 I 
degrees of freedom. If hypothesis (15) is not rejected (i.e., the slopes are similar), it 
can be proceeded with the next step: to test whether or not the intercepts of the 
involved batches are similar. The computational procedure for this is described 
next. 
ANCOVA for Testing Intercept Similarities The second step is to test whether or 
not the intercepts for the individual degradation lines are equal given that the degradation 
lines from the considered batches have similar slopes. The hypothesis of 
interest can be written as 
H ij i I j I i j 0 1 1 : . .= . = = for all , . . . , , . . . , (28) 
Since differences in slopes were not identifi ed, model (14) reduces to a model 
with a single degradation rate as follows: 
y x i I j n ij i ij ij i = + + = = . . . 1 1 , . . . , , . . . , (29) 
The intercepts can be decomposed into two elements: the common intercept and 
the batch effect. That is, . i = . + . i , where . . . = . xii is the common intercept 
and . i is the batch effect; . is the expected value of y ij , and xii is the average of x ij 
and was defi ned by Equation (32) . It should be noted that the deviation of each 
intercept from the common intercept is called the batch effect. Thus, model (29) can 
be written as 
y x x ij i ij ij = + + . + . . . . ( )ii (30) 
A model without a batch effect is compared with a model that includes the batch 
effect to be able to measure the intercept effect. Equation (30) includes the batch 
effect and will be called the complete model. The model with no batch effect will 
be called a reduced model and can be expressed as

LONG-TERM STABILITY ANALYSIS 607 
y x x i I j n ij ij ij i = + . + = = . . . ( )ii 1 1 , . . . , , . . . , (31) 
The sums of squares, cross products for totals, and errors for the reduced model 
are computed as [12, 18] 
R x x x 
x
N 
x 
N xx 
i 
I 
j
n 
ij 
i 
I 
j
n 
ij 
i 
I i i 
= . = . = 
= = = = = 
.. .. . 
1 1 
2 
1 1 
2 
2 
1 
1 
( )ii 
ii 
ii 
j
n 
ij 
i
x x Nx 
= .
= 
1 
ii ii 
(32) 
R y y y 
y
N 
y 
N yy 
i 
I 
j
n 
ij 
i 
I 
j
n 
ij 
i 
I i i 
= . = . = 
= = = = = 
.. .. . 
1 1 
2 
1 1 
2 
2 
1 
1 
( )ii 
ii 
ii 
j
n 
ij 
i
y y Ny 
= .
= 
1 
ii ii 
(33) 
R x x y y x y 
x y 
xy 
i 
I 
j
n 
ij ij 
i 
I 
j
n 
ij ij 
i i 
= . . = . 
= = = = 
.. .. 
1 1 1 1 
( )( ) ii ii 
ii ii 
N 
(34) 
where N was defi ned in Equation (26) . 
The sum of squares for the complete model can be computed as 
T nx x 
x
n 
x
N xx 
i 
I 
i i 
i 
I 
i
i 
= . = . 
= = 
. . 
1 
2 
1 
2 2 
( ) i ii 
i ii 
(35) 
T ny y 
y
n 
y
N yy 
i 
I 
i i 
i 
I 
i
i 
= . = . 
= = 
. . 
1 
2 
1 
2 2 
( ) i ii 
i ii 
(36) 
T nx x y y 
x y 
n 
x y 
N xy 
i 
I 
i i i 
i 
I 
i i 
i 
= . . = . 
= = 
. . 
1 1 
( )( ) i ii i ii 
i i ii ii 
(37) 
E x x R T xx 
i 
I 
j
n 
ij i xx xx 
i 
= . = . 
= = 
.. 
1 1 
2 ( )i 
(38) 
E y y R T yy 
i 
I 
j
n 
ij i yy yy 
i 
= . = . 
= = 
.. 
1 1 
2 ( )i 
(39) 
E x x y y R T xy 
i 
I 
j
n 
ij i ij i xy xy 
i 
= . . = . 
= = 
.. 
1 1
( )( ) i i 
(40) 
The sum - of - squares errors for the reduced model can be computed as 
SSEr= . R 
R
R yy 
xy 
xx 
2 
(41) 
The sum - of - squares errors for the complete model can be computed as

608 DRUG STABILITY 
SSEc= . E 
E
E yy 
xy 
xx 
2 
(42) 
The mean - square errors for the intercept effect and the mean - square errors for 
the complete model can be written as 
MSE 
SSE SSE 
MSE 
SSE 
int 
r c 
c 
c = 
.
. 
= 
. . I NI 1 1 
(43) 
Finally, the F statistics to measure whether or not the intercepts are different 
from batch to batch is given as 
F(int) 
MSE 
MSE
int 
c 
= 
(44) 
The hypothesis (28) is rejected if F (int) > F 0.25, I . 1, N . I . 1 , where F 0.25, I . 1, N . I . 1 is the 
upper percentile of the F distribution with I . 1 and N . I . 1 degrees of freedom. 
Thus, if the null hypotheses (15) and (28) are not rejected at the 0.25 level of signifi 
cance, the batches can be considered to come from a single population or pool 
and a single shelf life is computed based on the studied batches. If that is the case, 
model (14) reduces to the expression 
y x i I j n ij ij ij i = + + = = . . . 1 1 , . . . , , . . . , (45) 
where . and . are the common intercept and slope, respectively, for model (45). 
The approach to estimate the shelf life of a single batch can be applied to the 
pooled stability data from all batches. Thus, the shelf life is obtained by fi nding the 
minimum root of Equation (10) , where a and b are the estimates of . and . from 
model (45), respectively, x. is the average sampling time, and S xx is defi ned by Equation 
(4) and considering for all sampling times. 
If hypothesis (15) is not rejected and hypothesis (28) is rejected, model (14) 
reduces to the expression 
y x i I j n ij i ij ij i = + + = = . . . 1 1 , . . . , , . . , . (46) 
Assuming a common slope with different intercepts for different batches, the 
expiration dating period is computed for each individual batch. The minimum of 
the expiration dating periods of the individual batches is the expiration dating 
period of the drug product. 
A third alternative may occur if hypothesis (15) is rejected; it is concluded in this 
case that the batches do not belong to a single population and the shelf life should 
be computed for each individual batch. The minimum of the expiration periods of 
the individual batches is the shelf life of the drug product. 
Example 2 Slopes and Intercepts Are the Same This example illustrates the 
application of the poolability test for batch similarity. Assay data from three batches, 
expressed as %LC, are shown in Table 15 . Determine whether or not the three 
batches belong to a single population. 

LONG-TERM STABILITY ANALYSIS 609 
TABLE 15 Assay Results (%LC) for Example 2 
Sampling time, months 0 3 6 9 12 18 24 36 
Batch 1 102.4 98.1 99.2 97.5 95.0 96.1 — — 
Batch 2 101.1 101.2 99.0 97.2 96.4 95.5 94.3 — 
Batch 3 104.1 102.1 99.5 98.1 95.7 94.1 94.0 93.5 
Solution The poolability test is implemented in two steps. 
Step 1. The fi rst test will determine whether or not the slopes from the three batches 
are similar. From Table 15 the following information can be extracted: I = 3, n 1 = 6, 
n 2 = 7, n 3 = 8, and N = 21. The fi rst hypothesis to be tested is 
H0 1 2 3 : . . . . = = = (47) 
Using Equations (16) – (21) , the sum of squares and cross products for %LC and 
sampling time are computed as follows: 
S x 
x 
xx 
j 
j ( ) . . . 1 
6 
3 6 18 
48
6 
210 
1 
6 
1
2 1
2 
2 2 2 
2 
= . = + + + . = 
= .
i 
S x 
x 
xx 
j 
j ( ) . . . . 2 
7 
3 6 24 
72
7 
429 43 
1 
7 
2
2 2
2 
2 2 2 
2 
= . = + + + . = 
= .
i 
S x 
x 
xx 
j 
j ( ) . . . 3 
8 
3 6 36 
108 
8 
1008 
1 
8 
3
2 3
2 
2 2 2 
2 
= . = + + + . = 
= .
i 
S y 
y 
yy 
j 
j ( ) . . . . . . 
. 
. 1 
6 
102 4 98 1 96 1 
588 3 
6 
33 6 
1 
6 
1
2 1
2 
2 2 2 
2 
= . = + + + . = 
= .
i 5 
S y 
y 
yy 
j 
j ( ) . . . . . . 
. 
. 2 
7 
101 1 101 2 94 3 
684 7 
7 
43 
1 
7 
2
2 2
2 
2 2 2 
2 
= . = + + + . = 
= .
i 75 
S y 
y 
yy 
j 
j ( ) . . . . . . 
. 
3 
8 
104 1 102 1 93 5 
781 1 
8 
111 
1 
8 
3
2 3
2 
2 2 2 
2 
= . = + + + . = 
= .
i .98 
S x y 
x y 
xy 
j 
j j ( ) ( . ) ( . ) ... ( . ) 1 
6 
3 98 1 6 99 2 18 96 1 
48 
1 
6 
1 1 
1 1 = . = + + + . 
= .
i i ( .) 
. 
588 3 
6 
69 6 = . 
S x y 
x y 
xy 
j 
j j ( ) ( . ) ( ) . . . ( . ) 
( 
2 
7 
3 101 2 6 99 24 94 3 
72 
1 
7 
2 2 
2 2 = . = + + + . 
= .
i i 684 7 
7 
131 23 
. ) 
. = . 
S x y 
x y 
xy 
j 
j j ( ) ( . ) ( . ) . . . ( . ) 3 
8 
3 102 1 6 99 5 36 93 5 
1 
1 
8 
3 3 
3 3 = . = + + + . 
= .
i i 08 781 1 
8 
294 45 
( .) 
. = . 

610 DRUG STABILITY 
The total sums of squares are computed as 
Z S i xx 
i 
xx = = + + = 
= .
1 
3 
210 429 43 1008 1647 43 ( ) . . 
Z S i yy 
i 
yy = = + + = 
= .
1 
3 
33 65 43 75 111 98 189 38 ( ) . . . . 
Z S i xy 
i 
xy = =. . . =. 
= .
1 
3 
69 6 131 23 294 45 495 28 ( ) . . . . 
Using Equation (22) , the aggregated sum - of - squares error is computed as 
SSEa= . = . = Z 
Z
Z yy 
xy 
xx 
2 2 
189 38 
495 28 
1647 43 
40 48 . 
. 
. 
. 
The sum - of - squares errors for each batch is computed using Equation (24) : 
SSE(1)= . = . = S 
S
S yy 
xy 
xx 
( ) 
( ) 
( ) 
. 
. 
. 1 
1
1 
33 65 
69 6 
210 
10 58 
2 2 
SSE(2)= . = . = S 
S
S yy 
xy 
xx 
( ) 
( ) 
( ) 
. 
.
. 
. 2 
2
2 
43 75 
131 23 
429 43 
3 65 
2 2 
SSE(3)= . = . = S 
S
S yy 
xy 
xx 
( ) 
( ) 
( ) 
. 
. 
. 3 
3
3 
111 98 
294 45 
1008 
25 97 
2 2 
The aggregate sum - of - squares error is obtained using Equation (23) : 
SSE SSE = = + + = 
= .i 
i 
1 
3 
10 58 3 65 25 97 40 2 ( ) . . . . 
Thus, the mean - square error due to the slope is computed using Equation (25) : 
MS 
SSE SSE 
slope 
a = 
. 
. 
= 
. 
. 
= 
I 1 
40 48 40 2 
3 1 
0 14 
. . 
. 
The mean - square error for the regression model (14) is 
MSE 
SSE = 
. 
= 
. 
= 
N I 2 
40 2 
21 2 3 
2 68 
.
( ) 
. 

LONG-TERM STABILITY ANALYSIS 611 
Finally, the F statistic for testing equal degradation rate [hypothesis (47)] is 
F( 
.
. 
. slope) 
MS 
MSE 
slope = = = 0 14 
2 68 
0 052 
The critical value to test for equality on slopes is F 0.25, I . 1, N . 2 I = F 0.25,2,15 = 1.52. Since 
F (slope) < F 0.25,2,15 , hypothesis (47) for equal degradation rate cannot be rejected at 
the 0.25 level of signifi cance. It is concluded that there is no signifi cant difference 
among the slopes of model (14), that is, the model represented by Equation (14) 
reduces to model (29). 
Step 2. The second step consists in testing for equality of intercepts among the 
batches. To derive the appropriate statistic for testing the underlying hypothesis, it 
is required to compute the following sum and cross products for the %LC and the 
sampling time. The hypothesis to be tested is 
H0 1 2 3 : . . . . = = = (48) 
The sum and cross products are computed by using equations (32) – (40) : 
R x 
x
N xx 
i j
n 
ij 
i = . = + + + . = 
= = 
.. 
1 
3 
1 
2 
2 
2 2 2 
2 
3 6 36 
228 
21 
1754 5 ii . . . . 
R y 
y
N yy 
i j
n 
ij 
i = . = + + + . 
= = 
.. 
1 
3 
1 
2 
2 
2 2 2 
2 
102 4 98 1 93 5 
2054 1 
21 
ii . . . . . . 
. = 189 97 . 
R xy 
x y 
N xy 
i j
n 
ij ij 
i = . = + + + 
= = 
.. 
1 
3 
1 
3 98 1 6 99 2 36 93 5 ii ii ( . ) ( . ) ... ( . ). =. 228 2054 1 
21 
503 06 
( .) 
. 
T 
x
n 
x
N xx 
i 
i
i 
= . = + + . = 
= .
1 
3 2 2 2 2 2 2 48
6 
72
7 
108 
8 
228 
21 
107 14 i ii . 
T 
y
n 
y
N yy 
i 
i
i 
= . = + + . = 
= .
1 
3 2 2 2 2 2 2 588 3 
6 
684 7 
7 
781 1 
8 
2054 1 
21 
0 i ii . . . . 
.59 
T 
x y 
n 
x y 
N xy 
i 
i i 
i 
= . = + + 
= .
1 
3 48 588 3 
6 
72 684 7 
7 
108 781 1 i i ii ii ( .) ( . ) ( .) 
8 
228 2054 1 
21 
7 78 . =. ( .) 
. 
E x x R T xx 
i j
n 
ij i xx xx 
i 
= . = . = . = 
= = 
.. 
1 
3 
1 
1754 5 107 14 1647 36 ( ) . . . i 
E y y R T yy 
i j
n 
ij i yy yy 
i 
= . = . = . = 
= = 
.. 
1 
3 
1 
2 189 97 0 59 189 38 ( ) . . . i 

612 DRUG STABILITY 
E x x y y R T xy 
i j
n 
ij i ij i xy xy 
i 
= . . = . =. + =. 
= = 
.. 
1 
3 
1 
503 06 7 78 49 ( )( ) . . i i 5 28 . 
The sum - of - squares errors for the reduced and complete models are given by 
Equations (41) and (42) , respectively: 
SSEr= . = . = R 
R
R yy 
xy 
xx 
2 2 
189 97 
503 06 
1754 5 
45 73 . 
. 
. 
. 
SSEc= . = . = E 
E
E yy 
xy 
xx 
2 2 
189 38 
495 28 
1647 36 
40 47 . 
. 
. 
. 
The mean - square errors for measuring intercept effects are calculated using Equation 
(43) as 
MSE 
SSE SSE 
int 
r c = 
.
. 
= 
.
. 
= 
I 1 
45 73 40 47 
3 1 
2 63 
. . 
. 
MSE 
SSE 
c 
c = 
. . 
= 
. . 
= 
N I 1 
40 47 
21 3 1 
2 38 
. 
. 
Finally the F statistic for testing the equality of intercepts is given by Equation 
(44) and computed as 
F( 
.
. 
. int) 
MSE 
MSE
int 
c 
= = = 2 63 
2 38 
1 105 
The critical value to test the equality of the intercept is F 0.25, I . 1, N . I . 1 = F 0.25,2,17 = 
1.51. Since F (int) < F 0.25,2,17 , there is not enough evidence to reject the null hypothesis 
expressed by Equation (47) at the 25% of level of signifi cance. It can be concluded 
that the intercepts are not statistically different. 
In summary, the null hypotheses for equality of slopes and intercepts are not 
rejected at the 0.25 signifi cance level and, consequently, all batches are considered 
from the same population. Therefore, the expiration dating period can be computed 
by using model (45). 
After pooling the data from the three batches, the regression line can be written 
as 
. . . y a bx x ij ij ij = + = . 100 93 0 2867 
The regression subroutine also provides the following calculations: MSE = 2.407, 
x. = 10.8571, and S xx = 1.754.6. Thus, to determine the shelf life, the following equation 
can be used:

LONG-TERM STABILITY ANALYSIS 613 
f x x 
x 
( ) [ ( . . )] . ( . ) 
( . ) = . . . + 
. 
90 100 93 0 2867 1 7291 2 407 
1 
21 
10 86 
17 
2 2 
2 
54 6 
0 
. 
... 
... 
= 
To fi nd the shelf life, the root of the above equation has to be found that is smaller 
than the reference value; the reference value is given by 
x 
a 
b ref = 
.
. 
= 
. 
= 90 100 93 90 
0 2867 
38 11 
.
. 
. 
The expiration dating period must be smaller than 38.11 months. Using the initial 
point x (0) = x ref . d = 38.11 . 8, the QNLS algorithm converges to x R = 32.801. 
Therefore, the expiration dating period for the underlying production batches is 32 
months. 
Minimum Approach for Multiple Batches When the hypothesis for equality of 
slopes is rejected at the 0.25 signifi cance level, the minimum approach should be 
implemented. This is because the degradation lines of individual batches cannot be 
considered the same since they have different degradation rates. In this situation 
the FDA guideline establishes that the overall expiration dating period has to 
ensure that the product will remain within acceptable limits regardless of the batch 
from which it comes. Thus, the shelf life for each batch is calculated and the expiration 
dating period is based on the lowest of all shelf lives. Mathematically, this can 
be expressed as 
min , , L L { ( ) . . . , ( )} x Kx k 1 (49) 
where the x L ( i ) is the shelf life of the i th batch, and i is the total number of batches. 
Since the minimum of all the expiration dating periods is the shortest shelf 
life among all batches, this estimate will provide a 95% confi dence that the 
strength of the drug product will remain above the acceptable lower specifi cation 
limit. 
Example 3 Slopes and Intercepts Are Different A stability study provides the 
assay results shown in Table 16 . Based on this information, determine the shelf life 
for this drug product. 
TABLE 16 Assay Results (%LC) for Example 3 
Sampling time, months 0 3 6 9 12 18 24 
Batch 1 99.2 97.1 96.1 95.2 93.8 93.1 92.4 
Batch 2 98.7 97 96.2 95.1 94.2 93.3 — 
Batch 3 102.5 98.9 97.1 95.6 94.1 93.1 — 

614 DRUG STABILITY 
Solution The conventional approach to determine the shelf life will require to test 
whether or not the slopes and the intercepts of the straight lines associated with the 
degradation rate of the considered batches are the same. ANCOVA is also used to 
test equalities on both slopes and intercepts. 
The strategy consists of testing fi rst the null hypothesis of equalities of slopes, 
which is given by Equation (14) . To determine whether or not this hypothesis is 
rejected, the statistics defi ned by Equation (27) is computed: 
F( 
.
. 
. slope) 
MS(slope) 
MSE 
= = = 4 0507 
0 8032 
5 04 
The null hypothesis, expressed by Equation (15) , is rejected at the 0.25 level of 
signifi cance because the critical value ( F 0.25,2,13 = 1.55) of this test is smaller than 
F (slope). It is concluded that the slopes of the degradation lines of these batches 
are different and consequently the minimum approach applies. Thus, the shelf life 
for each batch is computed and the one that exhibits the minimum shelf life is 
applied to all manufactured batches. 
The shelf life for the fi rst batch is obtained by fi nding the root which is smaller 
than the reference point of the following equation: 
[ ( . . )] . (. ) 
( . ) 
. 
90 98 0461 0 2898 2 015 0 6689 
1
7 
10 28 
429 43 
2 2 
2 
. . . + .
. x 
x 
.. 
... 
= 0 
The reference point of this equation is 
x 
a 
b ref = 
. 
. 
= 
. 
= 
. 98 0461 90 
0 2698 
29 82 
. 
. 
. 
Using as an initial point x (0) = x ref . 8, the root is x R (1) = 24.93 and the shelf life 
for the fi rst batch is x L (1) = 24 months. 
Using data for the second batch, the following equation is derived: 
[ ( . . )] . (. ) 
( ) 
90 98 1157 0 2957 2 1318 0 2328 
1
6 
8 
210 
2 2 
2 
. . . + 
. ... 
... 
= x 
x 
0 
The reference point for this equation is x ref = 27.44 and the initial point to accomplish 
convergence is x ref . 5. The corresponding root is x R (2) = 23.44 and the shelf 
life for batch 2 is x L (2) = 23 months. 
The associated equation for the third batch is 
[ ( . . )] . (. ) 
( ) 
90 100 91 0 5033 2 1318 1 5415 
1
6 
8 
210 
0 2 2 
2 
. . . + .
... 
... 
= x 
x 
The reference point for this equation is x ref = 21.67 and the initial point to accomplish 
convergence is x ref . 5. The root for this equation is x R (3) = 17.59 and the shelf 
life for batch 3 is x L (3) = 17 months. 

LONG-TERM STABILITY ANALYSIS 615 
Therefore, the shelf life that should be applied to future batches is 
min{24 23 17 months. } = 17 
Example 4 Equality of Slopes and Different Intercepts A stability study provides 
the assay results shown in Table 17 . Based on this information, determine the shelf 
life for this drug product. 
Solution The ANCOVA method is used for testing whether or not the batches 
have a common slope. The statistics that determines whether or not this hypothesis 
is rejected is given by Equation (27) and yields 
F( 
.
. 
. slope) 
MS(slope) 
MSE 
= = = 0 9891 
2 7134 
0 3645 
The null hypothesis, expressed by Equation (15) , cannot be rejected at the 0.25 
signifi cance level because the critical value ( F 0.25,2,17 = 1.51) of this test is larger than 
F (slope). It is concluded that the slopes of the degradation lines of these batches 
are similar. 
It is necessary to test whether or not the intercepts of batches are similar. The 
statistic to test intercept similarities is given by Equations (43) and (44) and provides 
the following results: 
MSE 
SSE SSE 
MSE 
SSE 
int 
r c 
c 
c = .
. 
= . = = 
. . 
= 
I NI 1 
172 19 48 10 
2 
62 04 
1 
48 1 . . 
. 
. 0 
19 
F( 
. 
. 
. int) 
MSE 
MSE
int 
c 
= = = 62 04 
2 53 
24 51 
The null hypothesis, expressed by Equation (28) , is rejected at the 0.25 signifi - 
cance level because the critical value ( F 0.25,2,19 = 1.49) of this test is larger than F (int). 
It is concluded that the intercepts are different at the 0.25 signifi cance level and the 
shelf life is estimated for each batch and the lowest value is applied to all the manufactured 
batches. 
TABLE 17 Assay Results (%LC) for Example 4 
Sampling time, months 0 3 6 9 12 18 24 36 
Batch 1 98.4 96.1 94.2 93.5 90 89.1 89.2 87.3 
Batch 2 99.1 97.2 96.3 95.2 93.4 91.5 90.3 — 
Batch 3 104.1 102.1 99.5 98.1 95.7 94.1 94.0 93.5 

616 DRUG STABILITY 
The model that describes the degradation line is given in Equation (46) , where 
I = 3, n 1 = 8, n 2 = 7, and n 3 = 8. To estimate the common slope and the three different 
intercepts of model (46), it has to be expanded as follows: 
y x 
y x 
y x 
y x 
11 1 11 11 
12 1 12 12 
18 1 18 18 
21 2 21 
= + + 
= + + 
= + + 
= + + 
. . . 
. . . 
. . . 
. . 
 
. 
. . . 
. . . 
. . . 
. . 
21 
22 2 22 22 
27 2 27 27 
31 3 31 31 
32 3 
y x 
y x 
y x 
y 
= + + 
= + + 
= + + 
= + 
 
x 
y x
32 32 
38 3 38 38 
+ 
= + +
. 
. . . 
 
(50) 
A matrix representation of model (50) is given by the expression 
y Xb e = + (51) 
where the elements of y, X, b , and e are given in Table 18 . 
Regression techniques are used to estimate the parameter of model (51). Thus 
the intercepts for batches 1, 2, and 3 are a 1 = 96.369, a 2 = 97.871 and a 3 = 101.78, 
respectively; the common slope is b = . 0.3069. All p values of the regression models 
are close to zero, indicating that the intercept of the batches are highly signifi cant 
and the coeffi cient of multiple determination is very high, R 2 = 0.9996, indicating 
very good model fi tting. 
TABLE 18 Description of the Matrix and Vectors of 
Model (50) 
Vector y Matrix X Vector b Vector e 
y 11 . 1 . 11 0 x 11 . 1 . 11 
y 12 . 2 . 12 0 x 12 . 2 . 12 
 . 3    . 3  
y 18 . . 18 0 x 18 . . 18 
y 21 0 . 21 0 x 21 . 21 
y 22 0 . 22 0 x 22 . 22 
      
y 27 0 . 27 0 x 27 . 27 
y 31 0 . 31 1 x 31 . 31 
y 32 0 . 32 1 x 32 . 32 
      
y 38 0 . 38 1 x 38 . 38 

LONG-TERM STABILITY ANALYSIS 617 
The expiration dating period will be estimated by solving the corresponding 
quadratic equation (10) for each batch. The major diffi culty of using the quadratic 
equation is the estimation of the MSE for each batch. The MSE estimate for 
each batch may be computed by performing a regression analysis for each batch; 
however, this approach may not be correct because there is no guarantee that each 
regression provides the same slope. Thus, we recommend implementing the following 
approach. Use results from the regression analysis of model (51) and compute 
the residuals for the entire model. Extract the corresponding errors for each batch 
and compute the SSE for each batch using Equation (4) . Use this result to calculate 
the MSE for each batch using Equation (6) and the sample size for the corresponding 
batch. In addition, extract the corresponding values of x ij for each batch and 
compute x. , and S xx using Equation (4) . Results from the described procedure are 
shown in Table 19 . 
The quadratic equation for batch 1 can be written as 
[ ( . . )] . (. ) 
( .) 
90 96 369 0 3069 1 9432 3 1476 
1
8 
13 5 
1008 
2 2 
2 
. . . + .
... 
. x 
x 
.. = 0 
The reference point for this equation is x ref = 20.7492 and the initial point to 
accomplish convergence is x ref . 6. The root for this equation is x R (1) = 16.62 and 
the shelf life for batch 1 is x L (1) = 16 months. 
Following a similar procedure the quadratic equation associated with batch 2 is 
[ ( . . )] . (. ) 
( . ) 
. 
90 97 871 0 3069 2 015 0 6061 
1
7 
10 2857 
429 43 
2 2 
2 
. . . + .
x 
x ... 
... 
= 0 
The reference point for this equation is x ref = 25.6451 and the initial point to 
accomplish convergence is x ref . 6. The root for this equation is x R (2) = 22.1394 and 
the shelf life for batch 2 is x L (2) = 22 months. 
The corresponding quadratic equation for batch 3 is 
[ ( . . )] . (. ) 
. ) 
90 101 78 0 3069 1 9432 4 3646 
1
8 
13 5 
1008 
2 2 
2 
. . . +( . ... 
. x 
x 
.. 
= 0 
The reference point for this equation is x ref = 38.38 and the initial point to accomplish 
convergence is x ref . 5. The root for this equation is x R (3) = 30.0519 and the 
shelf life for batch 3 is x L (3) = 30 months. 
Therefore, the shelf life that should be applied to all manufactured batches is 
min months { } 16 22 30 16 = 
TABLE 19 Information Required for Quadratic Equation 
Batch SSE MSE x. S xx n i 
1 18.8860 3.1476 13.5 1008 8 
2 3.0331 0.6061 10.28 429.43 7 
3 26.187 4.3646 13.5 1008 8 

618 DRUG STABILITY 
7.2.3.4 Shelf Life Estimation for Multiple Factors 
Most drug products are manufactured with more than one strength and are marketed 
in more than one package and consequently the stability analyses must be carried 
out for every combination of package and strength. For instance, suppose that a drug 
product is available in two strengths and four containers sizes. Thus, eight sets of data 
from the 2 . 4 strength size combinations can be analyzed and eight separate shelf 
lives should be estimated to calculate the shelf life for the drug product. 
A general analysis - of - covariance model for a stability design with several batches 
and packages can be expressed as 
y x i I j J k n ijk ij ij ijk ijk = + + = = = . . . 1 1 1 , . . . , , . . . , , . . . , (52) 
where y ijk is the response from the k th time point of the i th batch and the j th 
package of a drug product, x ijk is the sampling time at which the y ijk response 
was obtained, . ij 
is the intercept for the i th batch and the j th package, . ij is the degradation 
rate of the i th batch and the j th package, and . ijk is the random error 
assumed to be independently and normally distributed with zero mean and constant 
variance . 2 . 
Chen et al. [11] show that there are 16 different models of Equation (52) ; 
however, these models reduce to 9. The general procedure consists in identifying 
the class of model that is associated to a given assay information. Once the model 
is determined, the appropriate procedure is implemented to determine the shelf life 
of the drug product. Chen et al. [11] established the procedure for estimating the 
shelf life. 
Identifi cation of Analysis of Covariance Model A general procedure, based on 
regression analysis, to identify the analysis - of - covariance model that applies to a 
given set of assay results to determine the shelf life is introduced here. We call this 
procedure the regression model with indicator variables for testing poolability of 
batches and packages. 
To introduce the fundamental concepts and facilitate the understanding, we consider 
fi rst a simple case, second a general model, and fi nally the methodology as an 
example. 
Suppose we have three batches and the corresponding indicator variables model 
is defi ned as 
y u u x u u e ij ij ij = + + + + + + . . . . 0 1 1 2 2 1 1 2 2 2 ( ) . . (53) 
where y ij is the assay result from the j th time point of the i th batch of a drug product, 
x ij is the time at which the assay sample y ij was obtained, u 1 and u 2 are indicator 
variables that assign a binary code to each factor combination, . 0 is part of the 
intercept of the regression line, and . 1 is part of the slope of the regression line. It 
should be noted that the intercept of model (53) is . 0 + . 1 u 1 + . 2 u 2 and the slope is 
given by . 1 + . 1 u 2 + . 2 u 2 . The parameters . and . are estimated from assay data; e ij 
is the random error assumed to be independently and normally distributed with 
zero mean and constant variance . 2 . 

LONG-TERM STABILITY ANALYSIS 619 
Let B1, B2, and B3 be codes that represent batches 1, 2, and 3, respectively. The 
indicator variables are used to indicate at which batch (level of factor) is assigned 
the response variable. For instance, a response variable associated with B1 is represented 
by u 1 = 1 and u 2 = 0. Similarly, a response variable from B2 is represented 
by u 1 = 0 and u 2 = 1. Also a response variable from B3 is represented by u 1 = 0 and 
u 2 = 0. Thus, the values of u 1 and u 2 defi ne in a unique manner the factor combination. 
The indicator variables for this model are summarized in Table 20 . 
The response variable from batch 1 can be written as follows [after replacing the 
values of indicator variables in model (53)]: 
y x e 
y x e 
y n
11 0 1 11 1 1 11 
12 0 1 12 1 1 12 
1 0 1
= + + + + 
= + + + + 
= + 
. . . 
. . . 
. . 
( ) 
( ) 
.
. 
 
1 1 1 1 1 
21 0 2 21 1 2 21 
22 0 2 22
1 1 + + + 
= + + + + 
= + +
x e 
y x e 
y x
n n ( ) 
( ) 
( 
. 
. . . 
. . . 
. 
. 
1 2 22 
2 0 2 2 1 2 2 
31 0 31 1 31 
32
2 2 2 
+ + 
= + + + + 
= + + 
= 
.
.
) 
( )
e 
y x e 
y x e 
y
n n n 

. . . 
. . 
. . 
. . 
0 32 1 32 
3 0 3 1 3 3 3 3 
+ + 
= + + 
x e 
y x e n n n 
 
(54) 
where n i is the number of sampling times in the i th batch. 
It should be noted that the system of linear equations expressed by (54) represents 
the response variable from the three batches. The required condition for the 
three batches to have the same intercept is that . 1 = . 2 = 0. The three batches will 
have the same slope if and only if . 1 = . 2 = 0. Thus, the problem for testing poolability 
reduces to fi t the regression model (54) and test the following hypotheses: h 1 : . 1 
= . 2 = 0 and h 2 : . 1 = . 2 = 0. Therefore, if the null hypothesis h 2 is not rejected at the 
0.25 signifi cance level, it implies that the slopes of the three batches are the same, 
that is, . 1 = . 2 = . 3 = . . Similarly, if the null hypothesis h 1 is not rejected, the intercepts 
of the three batches are the same, that is, . 1 = . 2 = . 3 = . . If that is the case, 
the shelf life is determined by a model with a single intercept and a single slope. 
This procedure was applied to data from Examples 2 and 3 and the results from 
the regression analysis are given in Table 21 . 
TABLE 20 Binary Codifi cation for Indicator Variables 
Batch u 1 u 2 Assay Result 
B1 1 0 y 1 j 
B2 0 1 y 2 j 
B3 0 0 y 3 j 

620 DRUG STABILITY 
The indicator variables model to perform a poolability test for two factors (packages 
and batches) can be expressed as follows: 
y x u u u v v 
v x 
ijk ijk r r 
s s ijk 
= + + + + + + + 
+ + + 
. . . . . . . 
. 
0 1 1 1 2 2 1 1 2 2 . . . 
. . . (. . . . 
. . 
1 1 2 2 1 1 
2 2 11 1 1 12 1 2 
u u u v 
v v u v u v 
r r 
s s 
+ + + + 
+ + + + + + 
. . . 
. . . . . . . . 
+ + .rs r s ijk u v e ) 
(55) 
where i = 1, . . . , I, j = 1, . . . , J , and k = 1, . . . , n ij and I is the number of batches, 
J is the number of packages, n ij is the sample size of the i th batch and the j th package, 
and u i and v j are indicator variables that are defi ned as shown in Table 22 . 
The fi rst column of Table 22 shows the factor combinations. For instance, B1P1 
represents the fi rst batch and the fi rst package, B1P2 indicates the fi rst batch and 
the second package, and BIPJ represents the last batch and the last package. The 
subscripts r and s are defi ned as follows: r = I . 1, s = J . 1. 
Testing for poolability requires fi tting the regression model with indirect variables 
to the assay data and testing four possible hypotheses. Table 23 shows the 
hypotheses to be tested, the corresponding interpretation to whether or not the 
hypothesis is accepted or rejected, the code of the model, and the model established 
by the FDA [8, 11] . In Table 23 the letters A and R represent whether the considered 
hypotheses are accepted or rejected, respectively. 
Rules for Determining Shelf Life Once the analysis - of - covariance model 
has been identifi ed, a set of rules for computing the shelf life must be implemented. 
This section describes the rules to follow to determine the expiration dating period 
for each of the nine representative models described in the previous section 
[8, 11] : 
TABLE 21 Testing for Poolability of Batches 
Example 2: Equality of Slopes and 
Intercepts a 
Example 3: Slopes and Intercepts Are 
Different b 
Parameter Estimate t Statistic p Value Parameter Estimate t Statistic p Value 
. 0 101.58 112.20 0.0000 . 0 100.91 163.99 0.0000 
. 1 . 0.87 . 0.61 0.5513 . 1 . 2.86 . 3.44 0.0044 
. 2 . 0.62 . 0.46 0.6543 . 2 . 2.79 . 3.21 0.0068 
. 1 . 0.29 . 5.66 0.0000 . 1 . 0.50 . 8.14 0.0000 
. 1 . 0.04 . 0.32 0.7559 . 1 0.23 3.09 0.0085 
. 2 . 0.01 . 0.14 0.8883 . 2 0.21 2.37 0.0337 
Example 2: 
a h 1 : . 1 = . 2 = 0. P Values of this table show that this hypothesis is not rejected at the 0.25 signifi cance 
level and batches have common intercepts. 
h 2 : . 1 = . 2 = 0. This hypothesis is not rejected at the 0.25 signifi cance level and batches have common 
slopes. 
Example 3: 
b h 1 : . 1 = . 2 = 0. This hypothesis is rejected at the 0.25 signifi cance level and batches have different 
intercepts. 
h 2 : . 1 = . 2 = 0. This hypothesis is rejected at the 0.25 signifi cance level and batches have different 
slopes. 

LONG-TERM STABILITY ANALYSIS 621 
1. C0: The expiration dating period is computed separately for each batch and 
package combination. The minimum shelf life of the individual batches (at 
least three batches) from the same package is the shelf life of that package. 
2. C1: The data from the individual batches are combined within each package. 
For each package: 
• M1: Assuming a common slope with different intercepts for different batches, 
the shelf lives are computed for individual batches. The lowest of all shelf 
lives of the individual batches is the expiration dating period of that 
package. 
• M3: Assuming a common slope and a common intercept for all batches, a 
single shelf life is computed as the expiration dating period of that package. 
3. C2: The data from individual packages are combined within the batch: 
• M2: For each batch, assuming a common slope with different intercepts for 
different packages, the shelf lives are computed for individual packages. The 
lowest shelf life of the individual batches from the same package is used as 
the expiration dating period of that package. 
• M4: For each batch, assuming a common slope and a common intercept for 
all packages, a single shelf life is computed. The lowest expiration dating 
period of the individual batches is used as the shelf life for every package. 
4. C3: The data for all packages and batches are combined: 
• M5: Assuming different intercepts but a common slope for all batches and 
package combinations. The lowest shelf life of the individual batches from 
the same package is used as the expiration dating period of that package. 
• M6: Assuming a common intercept for all batches with a common slope for 
all batch and package combinations, a single expiration dating period is 
computed for each package to be used as the expiration dating period of 
that package. 
TABLE 22 Binary Code for Factor Combinations 
Factor Combinations u1 u2 . ur v1 v2 . vr yijk 
B1P1 1 0 . 0 1 0 . 0 y11k 
B2P1 0 1 . 0 1 0 . 0 y21k 
          
BrP1 0 0 . 1 1 0 . 0 yr1k 
BIP1 0 0 . 0 1 0 . 0 yI1k 
B1P2 1 0 . 0 0 1 . 0 y12k 
B2P2 0 1 . 0 0 1 . 0 y22k 
          
BrP2 0 0 . 1 0 1 . 0 yr2k 
BIP2 0 0 . 0 0 1 . 0 yI2k 
          
B1PJ 1 0 . 0 0 0 . 1 y1Jk 
B2PJ 0 1 . 0 0 0 . 1 y2Jk 
          
BrPJ 0 0 . 1 0 0 . 1 yrJk 
BIPJ 0 0 . 0 0 0 . 1 yIJk

622 DRUG STABILITY 
TABLE 23 Procedure for Identifying Analysis of Covariance Model 
Code Hypotheses A R Interpretation ANCOVA Model 
1 C0: M0a H1 : .1 = .2 = . . . 
= .r = 0 
X Intercepts of all batches are 
the same, 
.1j = .2j = . . . = .Ij = .j 
yijk = .j + .ij x ijk 
+ eijk 
2 C1: M1 H2 : .1 = .2 = . . . 
= .r = 0, 
.11 = .12 = . . . 
= .rs = 0 
X Slopes of all batches are the 
same, 
.1j = .2j = . . . = .Ij = .j 
yijk = .ij + .j x ijk 
+ eijk 
2 C1: M3 H1 and H2 X All batches have the same 
intercept and the same 
slope, 
.1j = .2j = . . . = .Ij = .j 
.1j = .2j = . . . = .Ij = .j 
yijk = .j + .j x ijk 
+ eijk 
1 C0: M0b H3 : .1 = .2 = . . . 
= .s = 0 
X Intercepts of all packages are 
the same, 
.i 1 = .i 2 = . . . = .iJ = .i 
yijk = .i + .ij x ijk 
+ eijk 
3 C2: M2 H4 : .1 = .2 = . . . 
= .s = 0, 
.11 = .12 = . . . 
= .rs = 0 
X Slopes of all packages are the 
same, 
.i 1 = .i 2 = . . . = .iJ = .i 
yijk = .ij + .i x ijk 
+ eijk 
3 C2: M4 H3 and H4 X Intercepts and slopes of all 
packages are the same, 
.i 1 = .i 2 = . . . = .iJ = .i 
.i 1 = .i 2 = . . . = .iJ = .i 
yijk = .j + .i x ijk 
+ eijk 
3 C3: M5 H2 and H4 X All batches and all packages 
will have a common slope, 
.1j = .2j = . . . = .Ij = .j 
.i 1 = .i 2 = . . . = .iJ = .i 
.i = .j = . 
yijk = .ij + .xijk 
+ eijk 
4 C3: M6 H1 , H2 , and H4 X All batches and packages 
have common slope and the 
intercepts of all batches are 
the same, 
.1j = .2j = . . . = .Ij = .j 
.1j = .2j = . . . = .Ij = .j 
.i 1 = .i 2 = . . . = .iJ = .i 
.i = .j = . 
yijk = .j + .xijk 
+ eijk 
4 C3: M7 H2 , H3 , and H4 X All batches and packages 
have common slope and the 
intercepts of all packages 
are the same, 
.i 1 = .i 2 = . . . = .iJ = .i 
.1j = .2j = . . . = .Ij = .j 
.i 1 = .i 2 = . . . = .iJ = .i 
.i = .j = . 
yijk = .i + .xijk 
+ eijk

LONG-TERM STABILITY ANALYSIS 623 
• M7: Assuming a common intercept for all packages and a common slope for 
all batches and package combinations, the expiration dating periods are 
computed for each batch. The minimum of the individual batches is used as 
the expiration dating period for every package. 
• M8: Assuming a common slope and a common intercept for all packages and 
batch combinations, a single expiration dating period is computed to be used 
as the shelf life for every package. 
Example 5 Multiple - Factor Stability Study A well - known example introduced 
by Shao and Chow [19] is used to illustrate the application of shelf life calculations 
for a multifactor case. A stability study was conducted on a 300 - mg tablet of a drug 
product to establish the shelf life for each of the two types of packages used for this 
product: bottle and blister. The results are shown in Table 24 . Each type of package 
includes fi ve batches. The tablets were tested for potency at 0, 3, 6, 9, 12, and 18 
months. Determine the shelf life based on these stability data. 
TABLE 24 Assay Results (%LC) 
Package Batch 
By Sampling Time in Months 
0 3 6 9 12 18 
Bottle 1 104.8 102.5 101.5 102.4 99.4 96.5 
2 103.9 101.9 103.2 99.6 100.2 98.8 
3 103.5 102.1 101.9 100.3 99.2 101.0 
4 101.5 100.3 101.1 100.6 100.7 98.4 
5 106.1 104.3 101.5 101.1 99.4 98.2 
Blister 1 102.0 101.6 100.9 101.1 101.7 97.1 
2 104.7 101.3 103.8 99.8 98.9 97.1 
3 102.5 102.3 100.0 101.7 99.0 100.9 
4 100.1 101.8 101.4 99.9 99.2 97.4 
5 105.2 104.1 102.4 100.2 99.6 97.5 
Code Hypotheses A R Interpretation ANCOVA Model 
4 C8: M8 H1 , H2 , H3 , and 
H4 
X Slopes and intercepts of all 
batches and packages are 
the same, 
.1j = .2j = . . . = .Ij = .j 
.i 1 = .i 2 = . . . = .iJ = .i 
.i = .j = . 
.1j = .2j = . . . = .Ij = .j 
.i 1 = .i 2 = . . . = .iJ = .i 
.i = .j = . 
yijk = . + .xijk 
+ eijk 
1 C0 H1 , H2 , H3 , and 
H4 
X Intercepts and slopes of 
batches and packages are 
different 
yijk = .ii + .ij x ijk 
+ eijk 
TABLE 23 Continued

624 DRUG STABILITY 
Solution Based on these data, we can extract the following information to build 
the regression model with indicator variables: I = 5 batches, J = 2 packages, r = 4, 
s = 1, and n = 6 sampling times for all batches: 0, 3, 6, 9, 12, and 18 months. The 
indicator variables are shown in Table 25 and the indicator variables model for this 
case is 
y x u u u u v 
x u u u 
ijk ijk 
ijk 
= + + + + + + 
+ + + 
. . . . . . . 0 1 1 1 2 2 3 3 4 4 1 1 
1 1 2 2 3 (. . .3
3 4 4 11 11 11 
21 2 1 31 3 1 4 4 1 
+ + + 
+ + + +
. . u v uv 
u v u v u v eijk 
. 
. . . ) 
(56) 
where subscript i refers to the batch, subscript j to the package type, and subscript 
k to the sampling time. 
A conventional computer package for regression analysis was used to estimate 
the parameters for the model described by Equation (56) and the results are summarized 
in Table 26 . 
TABLE 25 Defi nition of Indicator Variables 
Factor Combinations u 1 u 2 u 3 u 4 v 1 y ijk 
B1P1 1 0 0 0 1 y 11 k 
B2P1 0 1 0 0 1 y 21 k 
B3P1 0 0 1 0 1 y 31 k 
B4P1 0 0 0 1 1 y 41 k 
B5P1 0 0 0 0 1 y 51 k 
B1P2 1 0 0 0 0 y 12 k 
B2P2 0 1 0 0 0 y 22 k 
B3P2 0 0 1 0 0 y 32 k 
B4P2 0 0 0 1 0 y 42 k 
B5P2 0 0 0 0 0 y 52 k 
TABLE 26 Parameter Estimation for Indicator Variables 
Model 
Parameter Estimate t Statistics p Value 
. 0 104.95 185.76 0.0000 
. 1 . 1.63 . 2.23 0.0306 
. 2 . 1.33 . 1.83 0.0747 
. 3 . 2.83 . 3.88 0.0003 
. 4 . 3.64 . 4.99 0.0000 
. 1 0.44 0.96 0.3408 
. 1 . 0.43 . 6.91 0.0000 
. 1 0.15 1.73 0.0901 
. 2 0.07 0.88 0.3823 
. 3 0.31 3.62 0.0007 
. 4 0.26 3.01 0.0043 
. 1 . 0.01 . 0.18 0.8557 
. 11 . 0.06 . 0.65 0.5161 
. 21 0.05 0.56 0.5770 
. 31 . 0.02 . 0.21 0.8313 
. 41 0.04 0.43 0.6660 

LONG-TERM STABILITY ANALYSIS 625 
P Values of Table 26 show the hypothesis H 3 : . = 0 and H 4 : . 1 = . 11 = . 21 = . 31 
= . 41 = 0 are not rejected at the 0.25 signifi cance level and, according to Table 23 , 
the intercepts and slopes of all packages are the same, that is, . i 1 = . i 2 = . i and 
. i 1 = . i 2 = . i 
. Therefore, the analysis - of - covariance model that should be applied to 
the assay results has the code 3 C2: M4 and has the form y ijk = . i + . i 
x ijk + e ijk . 
Since there is no difference between the packages, a simple analysis - of - 
covariance model is suitable. The rules for computing the expiration dating period 
is given in Section 7.2.3.3 and indicate that a shelf life should be computed for each 
batch and the minimum criterion is used to determine the expiration dating period 
for the underlying batches of the drug product. Thus, regression model fi tting should 
be conducted using model 3 C2: M4 and the results are used to develop the quadratic 
equation. 
The corresponding quadratic equation to determine the shelf life for the fi rst 
batch is 
[ ( . . )] . (. ) 
( ) 
90 103 54 0 3233 1 8125 1 48 
1 
12 
8 
420 
0 2 2 
2 
. . . + .
... 
... 
= x 
x 
The reference point for this equation is x ref = 41.89 and the initial point to accomplish 
convergence is x ref . 7. The root for this equation is x R (1) = 33.25 and the shelf 
life for batch 1 is x L (1) = 33 months. Since the sampling times are the same for all 
batches, the following values remain invariant for all batches: n = 12, x. = 8, S xx = 
420, and t 0.05,10 = 1.8125. The values required by the quadratic equation and the 
associated shelf life for each batch are given in Table 27 . The computer program to 
perform this calculation is given in the Appendix . 
Therefore, the shelf life that should be applied to current and future batches is 
min{33 32 54 49 31 months } = 31 
This approach provides a conservative estimate of the overall shelf life because 
it provides more than 95% confi dence for all batches except the batch from which 
it is estimated. It is worth mentioning that the minimum approach has been highly 
criticized by several researchers. For instance, Chow and Shao [20] pointed out that 
the minimum approach lacks statistical justifi cations. Ruberg and Stegeman [21] and 
Ruberg and Hsu [22] described the drawbacks of this methodology. However, Chen 
et al. [11] show that the FDA procedure performs reasonably well for data from a 
typical stability design with three batches. The FDA procedure is based on the 
assumption that batch effects are fi xed. If the analysis shows that batch - to - batch 
differences are small, it is advantageous to combine all the data to obtain one overall 
estimate. If the analysis shows evidence of batch - to - batch differences, then the FDA 
TABLE 27 Required Parameters to Compute Shelf Life 
batch a b MSE x ref x R Shelf Life 
2 103.84 . 0.3429 1.47 40.37 32.49 32 
3 102.34 . 0.1429 1.24 86.4 54.26 54 
4 101.54 . 0.1671 0.74 69.02 49.84 49 
5 105.17 . 0.4426 0.45 34.28 31.08 31 

626 DRUG STABILITY 
7.2.4 SHORT - TERM STABILITY ANALYSIS 
7.2.4.1 Introduction 
Assuring acceptable stability of drugs remains a challenge to industry. As mentioned 
in Section 7.2.2.1 the stability concept involves the classical aspect of API degradation 
as well as the presence of degradation products at levels that represent a risk to 
the patient [23] . In general, two stability studies are conducted to ensure that the 
market drug product is under the required specifi cations: short-term and long-term 
studies. A short - term stability study is an accelerated stability testing study, that is, a 
study under stressed storage conditions. The main goal of accelerated stability testing 
is not only to determine the chemical reaction kinetics but also to establish a tentative 
expiration date under the environment of marketing storage conditions. 
The stability of a drug product depends on the storage conditions, temperature 
and relative humidity, at which the product is exposed during its shelf life period. 
The effect of the environment on the degradation depends on the packaging con- 
fi guration used and the chemical characteristics of the drug product. The degradation 
of a drug product is mainly caused by the chemical reaction of the API with 
the excipients or with species in the environment (atmospheric oxygen and humidity) 
causing a decrease of the assay results over time. During the whole shelf life 
period, the API of a drug product is degrading at a certain rate. Other quality attributes 
(e.g., physical appearance, sterility, drug release rate, impurities) may also be 
changing with time, affecting the functionality, effi cacy, or purity of the drug product. 
The magnitude of these changes during the expiration dating period should in no 
way represent a risk to the patient. 
The rate of a chemical reaction is a function of temperature and concentration 
of the reactants present. Therefore, to induce a change in the reaction rate, a change 
in temperature or reactant concentration must be caused. The temperature can be 
readily changed by varying the storage temperature of the drug product. The reactants 
in a drug product are the API, the excipients, and the environment air and 
humidity. Of these, only the environment air and humidity can be changed because 
the API and the excipients within the drug product are specifi ed and cannot be 
changed. 
It is of interest for the pharmaceutical industry to know the degradation of its 
drug products at accelerated conditions to assess degradation for longer term storage 
at nonextreme conditions and short excursions outside of the recommended storage 
uses the minimum of all estimates obtained from individual batches based on the 
premise that the overall expiration dating period may depend on the minimum time 
a batch may be expected to remain within acceptable limits. It also shows that the 
smaller is the experimental variance, the higher is the expiration period. Thus, the 
FDA exhibits conservatism when the number of batches is large with small batch 
variabilities. The FDA established that the 0.25 signifi cance level is used to compensate 
for the expected low power of the design due to the relatively limited sample 
size in a typical formal stability study [8] . We supported the FDA arguments for the 
0.25 signifi cance level and this handbook outlines the procedure approved by the 
FDA. 

conditions that might occur during shipping. The short - term stability study is the 
method to evaluate degradation of a drug product at accelerated conditions. As an 
example of this, Gil - Alegre et al. [24] studied the degradation kinetics of mitonafi de, 
an antineoplastic agent, at temperatures between 60 ° C and 90 ° C at 10 ° C intervals. 
This is a very stable drug and, thus, rather high temperatures had to be used to 
have detectable concentration changes. Normally, accelerated testing is carried out 
at constant temperature. However, ramping temperatures have been used in nonisothermal 
testing [25] . This approach may lead to better prediction of the low - 
temperature kinetic constant and thus expiration dating periods [23] . 
7.2.4.2 Chemical Reaction Kinetics 
To understand how degradation data are treated, it is convenient to mention the 
basics of chemical reaction kinetics. The principles of chemical reaction engineering 
can be found in any reaction engineering or reactor design textbook [26] . A chemical 
reaction is the process whereby one or more components are transformed into one 
or more different components. The rate of reaction is the velocity at which the 
component(s) are being transformed in a chemical reaction. For the chemical 
reaction 
A B E + > (57) 
the reaction rate can be expressed as 
dC 
dt 
kn m 
n m = . + [ [ A] B] 
(58) 
where C is the molar concentration of the component to be studied, that is, [A] or 
[B], the brackets represent the concentration of the reactants, and k is the rate 
constant, also known as the specifi c reaction rate. The rate of reaction is positive if 
it refers to a product and negative if it refers to a reactant. The reaction order is the 
sum of the exponents ( n + m ) in Equation (58) . According to this, a reaction of 
order zero is represented with the equation 
dC 
dt 
k = . 0 
(59) 
which can be integrated to give 
C C kt = . 0 0 (60) 
where C is the molar concentration at any time t, C 0 is the concentration at time 
t = 0, k 0 is the rate constant of the zero - order reaction, and t is the time. For a fi rst - 
order reaction, the rate can be expressed as 
dC 
dt 
k C = . 1 
(61) 
SHORT-TERM STABILITY ANALYSIS 627

628 DRUG STABILITY 
which integrates to 
ln or 
C 
C 
k t C C C e k t 
0 
1 0 0 
1 ( )= . = . . 
(62) 
The temperature has a strong effect on the rate constant; this effect is represented 
by the Arrhenius equation: 
k Ae E RT = . ( ) (63) 
where A is the frequency factor and has the same units of k, E is the activation 
energy, R is the gas constant, and T is absolute temperature in kelvin. The logarithmic 
expression of the Arrhenius equation is 
ln ln k 
E 
RT 
A = . + 
(64) 
7.2.4.3 Degradation Data Evaluation at Accelerated Conditions 
Chemical reaction kinetics can be used to evaluate degradation data at accelerated 
conditions and predict the drug product assay at normal conditions for periods 
longer than the proposed shelf life. This is applicable to limited cases because the 
reaction kinetics is often complex for drug products. The following example illustrates 
the procedure to follow to calculate the API concentration with time for a 
drug product stored at normal conditions when such data are not yet available. 
Example 6 Estimation of Degradation from Accelerated Data: First - Order 
Case Consider the degradation data shown in Table 28 , obtained at normal, intermediate, 
and accelerated storage conditions. All assay values are expressed as a percentage 
of the label claim (%LC). Ignoring the fi rst two columns, that is, using only 
the data at higher temperatures (30 ° C and higher), estimate the assay values at 25 ° C 
as a function of time. [ Note: The data in this example are artifi cial and will be used 
only for demonstration purposes. The normal - conditions data are displayed in the fi rst 
two columns (25 ° C). Usually these data are not available when a new drug application 
(NDA) is submitted and that is why the procedure being presented here is 
important.] 
TABLE 28 Assay Degradation Data for Drug Product 
t (months) Assay at 25 ° C t (months) Assay at 30 ° C Assay at 40 ° C Assay at 50 ° C 
0 99.9 0 99.9 99.9 99.9 
3 99.4 2 99.4 98.0 95.6 
6 98.2 4 98.7 95.9 91.2 
9 97.8 6 97.4 95.1 87.4 
12 97.4 
18 95.6 
24 94.5 
36 91.7 

Solution The analysis can be done assuming that the reaction is fi rst order (this is 
usually the case in real life) or zero order (simpler analysis with no signifi cant error 
for degradation of less than 10%). Both cases are presented here. For fi rst - order 
kinetics [refer to Equation (61) ], the fi rst step is to obtain the natural logarithm of 
the concentrations, which is done in Table 29 . Then, these values are graphed as a 
function of time where straight lines should be obtained as shown in Figure 2 . The 
slope of each line corresponds to the value of k 1 , which is shown in the bottom part 
of Table 29 . Next, a plot of ln( k 1 ) versus 1/ T is prepared with the three values of k 1 , 
one for each temperature as shown in Figure 3 . Note that absolute temperature 
should be used here. From Equation (64) , the value of k 1 can be extrapolated to T 
= 25 ° C (1/ T = 0.003354 K . 1 ) to then predict the drug product assay as a function of 
time. The values of slope and intercept are shown in the inset of Figure 3 ; the calculation 
is 
lnk25 8237 4 0 003354 21 639 5 990 = . . + = . . . . . 
k25 
1 0 00250 = . . month 
Once the specifi c reaction rate is calculated, the drug product assay can be predicted 
using Equation (62) . The results are shown in Table 30 along with the actual data 
from Table 28 . 
TABLE 29 Degradation Data Treated as First - Order Kinetics 
t (months) ln(assay) at 30 ° C ln(assay) at 40 ° C ln(assay) at 50 ° C 
0 4.60 4.60 4.60 
2 4.59 4.58 4.55 
4 4.58 4.55 4.50 
6 4.57 4.55 4.46 
k 1 0.00415 0.00847 0.02224 
ln( k 1 ) . 5.483 . 4.771 . 3.798 
1/T 0.00330 0.00319 0.00309 
FIGURE 2 Degradation data treated as fi rst - order kinetics. 
30°C 
40°C 
50°C 
1 2 3 4 5 6 7 8 
In (Assay) 
Time (months) 
4.62 
4.60 
4.58 
4.56 
4.54 
4.52 
4.50 
4.48 
4.46 
SHORT-TERM STABILITY ANALYSIS 629

630 DRUG STABILITY 
TABLE 30 Predicted Assay Values Using Arrhenius and 
First - Order Kinetics 
t (months) Assay Predicted Assay 
0 99.0 100.0 
3 98.5 98.3 
6 97.3 97.5 
9 96.9 96.8 
12 96.5 96.1 
18 94.7 94.6 
24 93.6 93.2 
36 90.8 90.5 
FIGURE 3 Arrhenius plot for fi rst - order degradation. 
–3.0 
–3.5 
–4.0 
–4.5 
–5.0 
–5.5 
–6.0 
1/T 
0.00305 0.00310 0.00315 0.00320 0.00325 0.00330 0.00335 
y = –8237.4x + 21.639 
R2 = 0.9885 
In (k
1
) 
Example 7 Estimation of Degradation from Accelerated Data: Zero - Order 
Case Repeat Example 6 assuming that the degradation is of zero order. 
Solution For zero - order kinetics [refer to Equation (60) ], the original concentrations 
can be graphed as a function of time. Straight lines can be drawn through the 
points even if the actual degradation order is 1 because the degradation is small 
(less than 10%), as seen in Figure 4 . The slope of each line corresponds to the value 
of k 0 . These values are shown in Table 31 . Next, a plot of ln( k 0 ) versus 1/ T is prepared 
with the three values of k 0 , one for each temperature, which is shown in Figure 5 . 
As in Example 6 , the value of k 0 can be extrapolated to T = 25°C (1/ T = 0.003354 
K . 1 ) using Equation (64) to then predict the drug product assay as a function of 
time. The values of slope and intercept are shown in the inset of Figure 5 ; the calculation 
is 
lnk25 7974 8 0 003354 25 369 1 379 = . . + = . . . . . 
k25 
1 0 2519 = . . month 
Once the specifi c reaction rate is calculated, the drug product assay can be predicted 
using Equation (60) . The results are shown in Table 32 along with the actual data 
from Table 28 . 

FIGURE 4 Degradation data treated as zero - order kinetics. 
30°C 
40°C 
50°C 
100 
98 
96 
94 
92 
90 
88 
86 
1 0 2 3 4 5 6 7 
Time (months) 
Assay (%) 
TABLE 31 Degradation Data Treated as Zero - Order 
Kinetics 
30 ° C 40 ° C 50 ° C 
k 0 0.410 0.825 2.095 
ln( k 0 ) . 0.892 . 0.192 0.740 
1/ T 0.00330 0.00319 0.00309 
FIGURE 5 Arrhenius plot for zero - order degradation. 
1.2 
0.8 
0.4
0 
–0.4 
–0.8 
–1.2 
0.00305 0.00310 0.00315 0.00320 0.00325 0.00330 0.00335 
In (k0) 
1/T 
y = –7974.8x + 25.369 
R2 = 0.9899 
TABLE 32 Predicted Assay Values Using Arrhenius and 
Zero - Order Kinetics 
t (months) Assay Predicted Assay 
0 99.0 100.0 
3 98.5 99.2 
6 97.3 98.5 
9 96.9 97.7 
12 96.5 97.0 
18 94.7 95.5 
24 93.6 94.0 
36 90.8 90.9 
Shelf life 41 39 
SHORT-TERM STABILITY ANALYSIS 631

632 DRUG STABILITY 
Comparison of Models To compare the predictions of both models (fi rst - and 
zero - order kinetics), Table 33 and Figure 6 have been prepared. As expected, both 
predictions give quite similar results. The fi rst - order model gives slightly better 
results, suggesting that the degradation is actually of fi rst order. Both predictions 
yield similar results because degradation reactions are quite slow and therefore the 
time frame of these measurements is very small compared, say, to the half - life of 
the degradation (or the time it takes for the assay to reach the value of 50% LC). 
Although the discussion of half - life calculation is beyond the scope of this handbook, 
suffi ce it to say that based on the results of Examples 6 and 7 , the half - life for 
fi rst - order kinetics is about 23 years and for zero - order kinetics (probably less accurate) 
is 17 years. These values are about 40 times the time frame used in the accelerated 
(short - term) tests of 6 months (0.5 year). Thus, any kinetic model will work 
reasonably well. 
7.2.4.4 Preliminary Shelf Life Calculation from Stressed Data 
Preliminary shelf life can be calculated based on accelerated results. Regression 
techniques described in Section 7.2.3 applies to the estimation assays obtained from 
TABLE 33 Actual and Predicted Assay Values 
t (months) Assay 
Predicted Assay 
First Order Zero Order 
0 99.9 100.0 100.0 
3 99.4 99.3 99.2 
6 98.2 98.5 98.5 
9 97.8 97.8 97.7 
12 97.4 97.0 97.0 
18 95.6 95.6 95.5 
24 94.5 94.2 94.0 
36 91.7 91.4 90.9 
FIGURE 6 Actual and predicted assay values. 
90 
92 
94 
96 
98 
100 
0 5 10 15 20 25 30 35 40 
Time (months) 
Assay 
Actual assay 
Predicted, first order 
Predicted, zero order

accelerated data as illustrated in Examples 6 and 7 . The procedure is the same as 
outlined in Section 7.2.3 , that is, to develop the quadratic equation and determine 
the shelf life for each of the observed and predicted assay results. 
The quadratic equation to determine the shelf life for the observed assay is 
[ ( . . )] . (. ) 
( .) 
90 99 884 0 2275 1 9432 0 0408 
1
8 
13 5 
1008 
2 2 
2 
. . . + .
... 
. x 
x 
.. 
= 0 
The reference point for this equation is x ref = 43.44 and the initial point to accomplish 
convergence is x ref . 1.5. The root for this equation is x R (obs) = 41.78 and the 
shelf life for observed assay results is x L (obs) = 41 months. This value will be used 
for comparison purposes. 
The quadratic equation for the predicted assay assuming a fi rst - order chemical 
reaction applies is 
[ ( . . )] . (. ) 
( .) 
90 99 95 0 2393 1 9432 0 0032 
1
8 
13 5 
1008 
2 2 
2 
. . . + .
... 
.. 
x 
x 
. = 0 
The reference point for this equation is x ref = 41.60 and the initial point to accomplish 
convergence is x ref . 0.4. The root for this equation is x R (fi rst) = 41.17 and the 
shelf life for the predicted assay assuming that the chemical reaction is fi rst order 
is x L (fi rst) = 41 months. 
Now, assuming that the chemical reaction is of zero order, the quadratic equation 
for the predicted assay is 
[ ( . . )] . (. ) 
( .) 
90 99 99 0 2515 1 9432 0 0013 
1
8 
13 5 
1008 
2 2 
2 
. . . + .... 
.. 
x 
x 
. = 0 
The reference point for this equation is x ref = 39.74 and the initial point to accomplish 
convergence is x ref . 0.3. The root for this equation is x R (zero) = 39.49 and the 
shelf life for the predicted assay assuming that the chemical reaction is zero order 
is x L (zero) = 39 months. 
In summary, the shelf life from the actual data is 41 months, that estimated from 
fi rst - order kinetics is also 41 months, and the one predicted from zero - order kinetics 
is 39 months. All three values are similar. However, the results reveal that the actual 
degradation reaction of the drug product is more likely to be of fi rst order. 
7.2.5 Concluding Remarks 
A succinct description of the basics of chemical reaction engineering has been 
presented and its application to the estimation of shelf life has been outlined 
through examples. These techniques are of crucial importance in NDAs to regulatory 
agencies such as the FDA. Normally, at the time a new drug application is 
submitted, not enough data at low temperature are available since long - term studies 
take years. The tools presented here are the alternative approved by the FDA 
and ICH. 
SHORT-TERM STABILITY ANALYSIS 633

634 DRUG STABILITY 
APPENDIX COMPUTER PROGRAMS 
This Appendix contains four computer programs in MatLab that can be used by the 
reader to perform typical calculations related to shelf - life estimations. 
This program computes the degradation line for single batch - and - save results to 
be input into the program called shelf life estimation: 
% Regression Analysis for a single batch 
clear, clc 
close all 
batch=3; 
i=3; % the number of the batch to be computed 
n1(1)=7; 
n1(2)=6; 
n1(3)=6; 
N=n1(i); 
x=[0 3 6 9 12 18 24 36 
0 3 6 9 12 18 24 36 
0 3 6 9 12 18 24 36]; 
y=[99.2 97.1 96.1 95.2 93.8 93.1 92.4 0 
98.7 97 96.2 95.1 94.2 93.3 0 0 
102.5 98.9 97.1 95.6 94.1 93.1 0 0]; 
x1=[x(i,1:n1(i))]; 
y1=[y(i,1:n1(i))]; 
x1=x1.; 
X=[ones(N,1) x1]; 
Y=y1.; 
b=inv(X.*X)*X.*Y; 
y_est=X*b; 
e=Y-y_est; 
SSE=e.*e; 
SST=Y.*Y-N*(mean(Y).2); 
SSR=SST-SSE; 
t=tinv(0.95,N-2) 
R2=SSR/SST 
MSE=SSE/(N-2) 
av_x=mean(x1) 
Sxx=x1.*x1-N*(av_x).2 
vec=[b. N MSE av_x Sxx] 
save res vec 
This program computes the degradation line after pooling data from several 
batches that have common slope and intercept: 
.
% Regression Analysis for pooled data from several batches 
clear, clc 
close all 
batch=3;

x=[0 3 6 9 12 18 24 36 
0 3 6 9 12 18 24 36 
0 3 6 9 12 18 24 36]; 
y=[102.4 98.1 99.2 97.5 95 96.1 95.2 94.3 
101.1 101.2 99 97.2 96.4 95.5 94.3 94.8 
104.1 102.1 99.5 98.1 95.7 94.1 94 93.5]; 
n(1)=6; 
n(2)=7; 
n(3)=8; 
x1=[x(1,1:n(1)) x(2,1:n(2)) x(3,1:n(3))]; 
y1=[y(1,1:n(1)) y(2,1:n(2)) y(3,1:n(3))]; 
x1=x1.; 
N=sum(n); 
X=[ones(N,1) x1]; 
Y=y1.; 
b=inv(X.*X)*X.*Y; 
y_est=X*b; 
e=Y-y_est; 
SSE=e.*e; 
SST=Y.*Y-N*(mean(Y).2); 
SSR=SST-SSE; 
R2=SSR/SST 
MSE=SSE/(N-2) 
av_x=mean(x1) 
Sxx=x1.*x1-N*(av_x).2 
vec=[b. N MSE av_x Sxx] 
save res vec 
This program computes the shelf life for data that have been input in either 
program 1 or 2: 
clear, clc 
close all 
delete it * % para borrar archivos anteriores display( .1: 
f = (90 -(b0 - b1 *x)).2-t.2*MSE*[1/n + (x - av_x) .2/Sxx] .) 
flag_fun=input(.Enter the number of the function .) 
load res 
b=vec(1:2); 
n=vec(3); 
MSE=vec(4); 
av_x=vec(5); 
Sxx=vec(6) 
d=-5; % expiration date 
reference=(b(1) - 90)/( -b(2)) 
if flag_fun==1 
x0=(b(1) - 90)/( -b(2))+d; % expiration date lower value 
end 
options = optimset( .LargeScale.,.off.); 
APPENDIX COMPUTER PROGRAMS 635

636 DRUG STABILITY 
cont=0; 
save contador cont 
[x,fval,exitflag,output] = fminunc(@(x) 
obj_fun_expiration_date(x,flag_fun),x0,options) 
This routine is required by program 3, shelf life estimation: 
function f = obj_fun_expiration_date(x,flag_fun) 
load contador 
cont=cont+1 
save contador cont 
load res 
b=vec(1:2); 
n=vec(3); 
MSE=vec(4); 
av_x=vec(5); 
Sxx=vec(6); 
t=tinv(0.95,n-2); 
if flag_fun==1 
f = abs((90 -(b(1)+b(2)*x)).2-(t.2)*MSE*(1/n+(x-av_x) 
.2/Sxx)); 
end 
save([.it. num2str(cont)], .f.,.x.,.flag_fun.) 
This program performs a statistical test for determining whether or not the slopes 
and/or intercepts of the degradation lines from several batches are equal: 
% ANCOVA 
clear, clc 
close all 
batch=3; 
x=[0 3 6 9 12 18 24 36 
0 3 6 9 12 18 24 36 
0 3 6 9 12 18 24 36]; 
y=[99.2 97.1 96.1 95.2 93.8 93.1 92.4 0 
98.7 97 96.2 95.1 94.2 93.3 0 0 
102.5 98.9 97.1 95.6 94.1 93.1 0 0]; 
n(1)=7; 
n(2)=6; 
n(3)=6; 
N=sum(n); 
%%%%%%%%%%%%%%%%%%% 
% testing slopes 
%%%%%%%%%%%%%%%%%% 
for i=1:batch 
sy=0; 
sx=0; 
sxy=0; 

for j=1:n(i) 
sy=sy+y(i,j) .2; 
sx=sx+x(i,j) .2; 
sxy=sxy+x(i,j) *y(i,j); 
end 
Syy(i)=sy -(sum(y(i,1:n(i)))).2/length(y(i,1:n(i))); 
Sxx(i)=sx -(sum(x(i,1:n(i)))).2/length(x(i,1:n(i))); 
Sxy(i)=sxy -(sum(y(i,1:n(i))))*(sum(x(i,1:n(i))))/ 
length(y(i,1:n(i))); 
end 
SxxW=sum(Sxx); 
SyyW=sum(Syy); 
SxyW=sum(Sxy); 
SSEW=SyyW-SxyW.2/SxxW; 
for i=1:batch 
sse(i)=Syy(i) -Sxy(i).2/Sxx(i); 
end 
SSE=sum(sse); 
SS_slope=SSEW - SSE; 
k=batch-1; 
MS_slope=SS_slope/(batch-1); 
MSE=SSE/(N-2*batch); 
F_slope=MS_slope/MSE; 
F_slope 
df_num=batch-1 
df_den=N-2*batch 
F_cri=finv(0.75,df_num,df_den) 
%%%%%%%%%% 
% testing intercepts 
%%%%%%%%%% 
sy=0; 
sy1=0; 
sx=0; 
sx1=0; 
sxy=0; 
for i=1:batch 
for j=1:n(i) 
sy=sy+y(i,j) .2; 
sx=sx+x(i,j) .2; 
sxy=sxy+x(i,j) *y(i,j); 
end 
sy1=sy1+sum(y(i,1:n(i))); 
sx1=sx1+sum(x(i,1:n(i))); 
end 
sy; 
SY1=sy1.2/N; 
SYY=sy-SY1; 
sx; 
APPENDIX COMPUTER PROGRAMS 637

638 DRUG STABILITY 
SX1=sx1.2/N; 
SXX=sx-SX1; 
sxy; 
SXY1=sx1*sy1/N; 
SXY=sxy-SXY1; 
tx=0; 
ty=0; 
txy=0; 
for i=1:batch 
ty=ty+(sum(y(i,1:n(i)))) .2/length(y(i,1:n(i))); 
tx=tx+(sum(x(i,1:n(i)))) .2/length(x(i,1:n(i))); 
txy=txy+(sum(x(i,1:n(i)))) *(sum(y(i,1:n(i))))/ 
length(y(i,1:n(i))); 
end 
my=sy1.2/N; 
mx=sx1.2/N; 
mxy=sx1*sy1/N; 
Tyy=ty - my 
Txx=tx - mx 
Txy=txy - mxy 
Eyy=SYY-Tyy 
Exx=SXX-Txx 
Exy=SXY-Txy 
SSEr=SYY-SXY.2/SXX 
SSEc=Eyy-Exy.2/Exx 
SSb0=SSEr - SSEc 
MSb0 = SSb0/(batch -1) 
MSEc=SSEc/(N-batch-1) 
F_b0=MSb0/MSEc 
df_num=batch-1 
df_den=N-batch-1 
F_cri_inter=finv(0.75,df_num,df_den) 
REFERENCES 
1. Expiration Dating, Code of Federal Regulations , Title 21, Vol. 4, Part 211, 
21CFR211.137. 
2. American Medical Association (AMA) ( 2001 ), Report 1 of the Council on Scientifi c 
Affairs (A - 01), Pharmaceutical expiration dates, available: http://www.ama - assn.org/ama/ 
pub/category/print/13652.html , accessed on Feb. 19, 2007. 
3. Bennett , B. , and Cole , G. , Eds. ( 2003 ), Pharmaceutical Production. An Engineering 
Guide , Institution of Chemical Engineers (IChemE) , Rugby, Warwickshire , United 
Kingdom . 
4. Matthews , B. R. ( 1999 ), Regulatory aspects of stability testing in Europe , Drug. Dev. Ind. 
Pharm. , 25 , 831 – 856 . 
5. Dietz , R. , Feilner , K. , Gerst , F. , and Grimm , W. ( 1993 ), Drug stability testing: Classifi cation 
of countries according to climatic zone , Drugs Made in Germany , 36 , 99 – 103 . 

6. U.S. Food and Drug Administration (FDA) ( 2001 ), Stability testing of new drug substances 
and products, ICH Q1A, FDA, Washington, DC. 
7. Fairweather , W. R. , Lin , T. Y. , and Kelly , R. ( 1995 ), Regulatory, design, and analysis of 
complex stability studies, U.S. Food and Drug Administration , J. Pharm. Sci. , 84 , 
1322 – 1326 . 
8. U.S. Food and Drug Administration (FDA) ( 2004 ), Evaluation of stability data, ICH Q1E, 
FDA, Washington, DC. 
9. U.S. Food and Drug Administration (FDA) ( 2003 ), Bracketing and matrixing designs for 
stability testing of new drug substances and products, ICH Q1D, FDA, Washington, 
DC. 
10. Montgomery , D. G. ( 2004 ), Design and Analysis of Experiments , 6th ed., Wiley , New 
York . 
11. Chen , J. J. , Ahn , H. , and Tsong , Y. ( 1997 ), Shelf - life estimation for multifactor stability 
studies , Drug Inform. J. , 31 , 573 – 587 . 
12. Montgomery , D. C. , Peck , E. A. , and Vining , G. G. , ( 2007 ), Introduction to Linear Regression 
Analysis , 3rd ed., Wiley , New York. 
13. Reklaitis , G. V. , Ravidran , A. , and Ragsdell , K. M. ( 1983 ), Engineering Optimization, 
Methods and Applications , Wiley , New York . 
14. Bancroft , T. A. ( 1964 ), Analysis of inference for incompletely specifi ed models involving 
the use of preliminary test(s) of signifi cance , Biometrics , 20 , 427 – 442 . 
15. Snedecor , G. W. , and Cochran , W. G. ( 1989 ), Statistical Methods , 8th ed., Iowa State University 
Press , Ames, IA . 
16. Wang , S. G. , and Chow , S. C. ( 1993 ), Advanced Linear Model: Theory and Applications , 
Marcel Dekker , New York . 
17. Ramirez - Beltran , N. D. , and Olivares , L. ( 1998 ), Drug Shelf - life Estimation Using Clustering 
Techniques. Proceeding of the Computing Research Conference CRC ’ 98 , edited by the 
University of Puerto Rico, Electrical and Computer Engineering Department. Mayaguez 
Puerto Rico, pp. 70 – 73 . 
18. Chow , S. - C. , and Liu , J. - P. ( 1995 ), Statistical Design and Analysis in Pharmaceutical Science , 
Marcel Dekker , New York . 
19. Shao , J. , and Chow , S - Ch. ( 1994 ), Statistical inference in stability analysis , Biometrics , 
50 ( 3 ), 753 – 763 . 
20. Chow , S. C. , and Shao , J. ( 1991 ), Estimating drug shelf - life with random batches , Biometrics 
, 47 , 1071 – 1079 . 
21. Ruberg , S. , and Stegeman , J. W. ( 1991 ), Pooling data for stability studies: Testing the equality 
of batch degradation slopes , Biometrics , 47 , 1059 – 1069 . 
22. Ruberg , S. , and Hsu , J. ( 1992 ), Multiple comparison procedures for pooling batches in 
stability studies , Technometrics , 34 , 465 – 472 . 
23. Waterman , K. C. , and Adami , R. C. ( 2005 ), Accelerated aging: Prediction of chemical 
stability of pharmaceuticals , Int. J. Pharm. , 293 , 101 – 125 . 
24. Gil - Alegre , M. E. , Bernabeu , J. A. , Camacho , M. A. , and Torres - Su a rez , A. I. ( 2001 ), Statistical 
evaluation for stability studies under stress storage conditions , Il Farm. , 56 , 
877 – 883 . 
25. Oliva , A. , Llabr e s , M. , and Fari n a , B. ( 2006 ), Data analysis of kinetic modeling used in 
drug stability studies: Isothermal versus nonisothermal assays , Pharm. Res. , 23 , 
2595 – 2602 . 
26. Levenspiel , O. ( 1998 ), Chemical Reaction Engineering , 3rd ed., Wiley , Ney York. 
REFERENCES 639


641 
7.3 
EFFECT OF PACKAGING ON 
STABILITY OF DRUGS AND 
DRUG PRODUCTS 
Emmanuel O. Akala 
School of Pharmacy, Howard University, Washington, DC 
Contents 
7.3.1 Introduction 
7.3.1.1 Rationale for Concern with Stability of Drugs and Drug Products 
7.3.2 Factors Infl uencing Stability of Drugs and Drug Products 
7.3.2.1 Moisture, Hydrolysis, and pH 
7.3.2.2 Oxygen and Oxidation 
7.3.2.3 Light 
7.3.2.4 Temperature 
7.3.2.5 Microbes 
7.3.2.6 Active Pharmaceutical Ingredients and Excipients 
7.3.3 Drug Packaging and Packaging Materials 
7.3.3.1 Introduction 
7.3.3.2 Materials of Fabrication of Packaging Components 
7.3.4 Effect of Packaging on Drug Product Stability 
7.3.4.1 Introduction 
7.3.4.2 Solid Dosage Forms 
7.3.4.3 Nonsterile Liquid Dosage Forms 
7.3.4.4 Sterile Liquid Dosage Forms 
7.3.4.5 Container Extractables and Leachables and Drug Product Stability 
7.3.4.6 Biotechnological Products 
7.3.4.7 A Look into the Future of the Effects of Packaging on Stability of Drug 
Products 
References 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

642 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
7.3.1 INTRODUCTION 
Stability is an essential attribute of drug products as evidenced by continuing 
national and international governmental interventions in the form of regulations. It 
is required that stability testing should demonstrate physical and chemical stability 
of a drug and its product(s) at a variety of environmental conditions such as 
temperature, humidity, and light [1] . Moreover, International Conference on Harmonization 
(ICH) guidelines provide requirements for stability studies [2, 3] . 
Pharmaceutical and biopharmaceutical companies are also making frantic efforts to 
improve on the development and manufacture of drug products with adequate stability 
profi les. All these activities (by governments and pharmaceutical and biopharmaceutical 
companies) are for the following purposes [4] : to deliver drug products 
of high quality (effi cacious and safe) to the fi nal consumer (the patient), to meet 
the legal requirements stipulated by the regulatory agencies, to provide the assurance 
to the manufacturer that the drug products will perform as intended, and to 
serve as a baseline for future drug development efforts. 
7.3.1.1 Rationale for Concern with Stability of Drugs and Drug Products 
Loss of Potency The effectiveness of any drug product (dosage form) is a function 
of its ability to deliver the required amount of the active therapeutic moiety to the 
biophase (site of action) for the intended length of time to achieve the purpose of 
therapy. Loss of the active ingredient in the dosage form, consequent upon chemical 
degradation, may result in poor performance of the drug product (i.e., the serum 
drug concentration may not reach the minimum effective concentration to elicit the 
desired therapeutic response). The potency of any drug product is not expected to 
remain constant ad infi nitum ; however it is expected that the quality of the drug 
product [the stability of the active pharmaceutical ingredients (APIs) among other 
attributes] is maintained until it either gets to the patients or reaches the expiration 
date, which is defi ned as the date placed on the immediate container label of the 
drug product that designates the date through which the product is expected to 
remain within specifi cations [5] . The specifi cations (the period within which the 
effi cacy, safety, and esthetics of the drug product can be assured) are often defi ned 
in terms of the shelf life of the drug product. When the chemical degradation of the 
APIs is found to be the major contributor to the degradation process of the drug 
product, the shelf life is considered the time that elapses before the drug content 
falls below 90% of the label claim, with the assumption that the drug product is 
stored as directed on the label [4, 6] . 
Toxic Decomposition Products In situations where the content of the API in a 
drug product is well above 90% as indicated above, the formation of toxic degradation 
products within the shelf life (which may cause untoward effects to the patients) 
may warrant the reassignment of a different expiration date or recall of the drug 
product in question. Consequently, the pharmaceutical industry is often concerned 
with both the amount as well as the nature of the degradation products. The formation 
of toxic products is particularly problematic with protein drugs which may 
maintain therapeutic activity after deliberate modifi cation or pertubation of molecular 
structure in a domain removed from that associated with therapeutic activity 

but can result in acquired immunogenicity [4] . A discussion of the examples of 
products of drug degradation which are more toxic than the original (parent) drugs 
has been given by Guillory and Poust [7] . It has been reported that the concerns for 
the potential increase in the amount of toxic degradation products informed the 
reluctance of the regulatory agencies to approve the design of new drug products 
with stability overages [4] . 
Poor Bioavailability The bioavailability of a drug product is customarily defi ned 
in terms of the amount of API delivered to the blood (plasma concentration) and 
the rate at which it is delivered. For routes of administration other than intravascular, 
the extent of absorption is very important in that all the drug administered may 
not have a chance to get to the plasma and then get carried to the biophase. One 
of the most important problems of drug absorption and oral bioavailability is when 
the drug is not delivered from the drug product over an appropriate time frame (in 
solution form) to the sites in the gastrointestinal (GI) tract where it is well absorbed 
(i.e., problem of getting the drug into solution). This problem can be caused by 
changes in the quality attributes of the drug product on storage (especially the dissolution 
rate), though the potency is still acceptable and no toxic degradation 
product has been formed. The changes in quality attribute can be traced to changes 
in the physicochemical properties of the excipients with aging. The result is ineffectiveness 
of the drug product on administration. 
Microbial Contamination The microbiological quality of all drug products is 
receiving a lot of attention in today ’ s drug product design, unlike in the past, when 
the concern was for drug products that must remain sterile: parenteral and ophthalmic 
products. The control of total bioburden and the exclusion of pathogenic 
microbes are considered desirable [4] . For proper maintenance of the microbiological 
quality of drug products the quality of the raw materials and the manufacturing 
facility must be controlled and assured. Further, in assessing the suitability for use 
of the container closure system, protection is considered to be one of the most 
important considerations. Thus the container closure system is expected to protect 
the drug product from causes of degradation such as light, temperature, loss of 
solvent, oxygen, water vapor, and microbial contamination [8] . Package integrity 
must not be compromised from the manufacturer through the distribution channels 
and to the fi nal consumer: the patient. 
Changes in Physical Appearance of Drug Product Some of the physical changes 
may affect the effi cacy of the drug products, while in others the potency may not 
change with the physical change but may affect pharmaceutical elegance such that 
patient acceptability suffers. Loss of an emulsifying agent may lead to breaking of 
an emulsion (phase separation) and caking of a suspension will lead to failure in 
resuspending the suspension on shaking. These changes will affect dose uniformity 
in the prescribed dosage regimen. Mottling of tablets may not affect the potency of 
the drug product, but such tablets will not be acceptable to patients and may affect 
compliance with dosage regimen. Moreover drug – excipient interactions may lead 
to changes in the physical appearance of the drug product as exemplifi ed by the 
interaction between lactose and the amino functional group in drug which often 
INTRODUCTION 643

644 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
leads to the formation of yellow color on the surface of tablets with aging. Though 
the potency is still intact, such tablets will not appeal to patients. 
Loss of Functional Stability It is one thing to design a very effective drug formulation; 
it is another important thing for it to be delivered properly by the container 
closure system during the shelf life of the drug product. Thus the performance of 
the container closure system for drug products (defi ned as the ability to function in 
the manner for which it was designed, i.e. ability to deliver the dosage form in the 
amount or at the rate described in the package insert [8] ) is an important part of 
drug product stability. Reports have shown that early transdermal patches had 
problems of skin adhesion with aging (they showed loss of adhesion with the tendency 
to fall off the patient ’ s skin) [4] . Other drug products for which consistency 
in drug delivery should be monitored over the entire shelf life are as follows: pre- 
fi lled syringes, a metered tube, a dropper, a spray bottle, a powder inhaler, and a 
metered - dose inhaler. 
7.3.2 FACTORS INFLUENCING STABILITY OF DRUGS AND 
DRUG PRODUCTS 
The preceding section has elaborated on the importance of the stability of drugs 
and drug products. The general principle in the design and manufacture of 
drug products is that quality (defi ned as the physical, chemical, microbiological, 
biological, bioavailability, and stability attributes that a drug product should maintain 
if it is to be deemed suitable for therapeutic and diagnostic use [8] ) should be 
built into all the processes or stages of drug manufacture rather than be inspected 
into the fi nal product ready for distribution. Consequently, knowledge of the factors 
that can promote all forms of drug and drug product instability such as chemical 
degradation (formation of new chemical entities), physical degradation (drug loss 
without the production of distinctly different chemical products), and biological 
degradation (the most important is microbial degradation, though cognizance should 
be taken of the fact that nonmicrobiological organisms such as ants and rats may 
be important) during the shelf life of the drug product is very important. The awareness 
of these factors will help not only in the preformulation and formulation stages 
of drug development but also in the selection and appraisal of the protective 
capability of the appropriate container closure system for the drug product in 
question. 
7.3.2.1 Moisture, Hydrolysis, and p H 
Liquid Dosage Forms Most hydrolytic degradations take place in the presence of 
moisture. Thus hydrolysis together with the infl uence of pH on hydrolysis as a route 
of degradation of drugs is important for all types of dosage forms: liquid, solid, 
semisolid, and gases. Recognition of the chemical groups which are susceptible to 
hydrolysis will aid the drug formulation scientist to determine a priori the paths to 
take to achieve drug products of maximum stability. Examples of such chemical 
groups are as follows: ester linkages (acetyl salicylic acid, procaine, teracaine, and 
physostigmine), amides (cinchocaine, ergometrine, and chloramphenicol), lactams 

FACTORS INFLUENCING STABILITY OF DRUGS AND DRUG PRODUCTS 645 
(penicillins, cephalosporins, nitrazepam, and chlordiazepoxide), imides (glutethimide 
and ethosuximide), and lactones (pilocarpine and spironolactone) [7, 9] . The 
rate at which hydrolysis proceeds in each of the chemical groups highlighted above 
is a function of the chemical environment of the chemical group within the drug 
molecule. Substituent groups can exert electronic, steric, or hydrogen - bonding 
effects known to affect the susceptibility of the chemical groups to hydrolytic degradation. 
In fact, modifi cation of drug chemical structure to control drug stability 
(though caution should be taken not to alter therapeutic effi cacy) using appropriate 
substituents has been used to solve stability problems. The concept of Hammett 
linear free - energy relationship for the effects of substituents on the rates of aromatic 
side - chain reactions such as hydrolysis of esters has been reported and the principle 
has been used to produce and select the best substituents for allylbarbituric acids 
of optimum stability [9, 10] . 
While laboratory experiments together with the knowledge of chemical kinetics 
can be used to formulate a solution drug product at a pH of optimum stability (using 
appropriate buffers) for a drug whose hydrolytic degradation is catalyzed by hydrogen 
ion (specifi c acid catalysis) or hydroxyl ion (specifi c base catalysis), the known 
infl uence of buffer components (called general acid – base catalysis) in buffered drug 
products should be considered. Failure to evaluate complete stability of the drug by 
determining the catalytic coeffi cients for specifi c acid catalysis and base catalysis 
and the catalytic coeffi cients of buffer components will result in poor determination 
of the conditions for maximum stability of the drugs in solution. Furthermore, the 
ionization of the drug should be taken into account in the determination of the 
pH – rate profi le for a complete determination of the conditions of maximum stability. 
Other strategies that have been used to stabilize drug products in solution 
involve the use nonaqueous solvents such as mono - and polyhydric alcohols (e.g., 
ethanol glycerin, and propyleneglycol) which can change the dielectric constant of 
the system. Micellar solubilization by surfactants and reduction in solubility (as 
exemplifi ed by increase in the stability of penicillin in procaine penicillin in the 
presence of additives such as citrates, dextrose, sorbitol, and gluconate) are capable 
of reducing hydrolysis of drugs in solution [9] . 
The infl uence of packaging on the stability of drugs and drug products will be 
examined in greater detail later in this chapter, but suffi ce it to state here that the 
permeability of the packaging materials to gases and liquids should be considered 
for all types of dosage forms since it can lead to decomposition of the drug products 
due to oxidation and hydrolysis. In the case of solution, leachables from container 
closure systems can alter the predetermined stability of liquid dosage forms. For 
example, leaching of dioctyl phthalate, a plasticizer used in polyvinyl chloride (PVC) 
plastics, into intravenous solutions containing surfactants has been reported [11, 12] . 
The potential adverse effect of the leached compound is twofold: on the safety of 
the patient and on the stability of the drug administered intravenously. Under the 
consideration for the suitability for the intended use of any proposed packaging 
system, as indicated by the U.S. Food and Drug Administration (FDA), compatibility 
(packaging components are regarded compatible with the dosage form if there is 
no suffi cient interaction to cause unacceptable changes in the quality of either the 
dosage form or the packaging component, i.e., leachable - induced degradation, precipitation, 
or changes in pH) and safety (packaging components should be constructed 
of materials that will not leach harmful or undesirable amounts of substances 

646 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
to which a patient will be exposed when being treated with the drug product) are 
given prominent positions [8] . It is indicated clearly that for a drug product such as 
an injection, inhalation, ophthalmic, or transdermal, a comprehensive study should 
be carried out: extraction studies on the packaging component to determine which 
chemical species (and their concentrations) may migrate into the dosage form and 
a toxicological evaluation of those substances which are extracted to determine the 
safe level of exposure via the label - specifi ed route of administration [8] . When the 
leached compound is an electrolyte, it can affect the ionic strength of the solution 
and hence the rate of degradation of the drug, which can be predicted by the 
Br o nsted – Bjerrum model [9]. 
Liquid dosage forms which are disperse systems (colloidal, i.e., microspheres, 
nanoparticles, and micelles; suspensions; and emulsions) often contain preservatives 
which are methyl, ethyl, propyl, and butyl esters of para - hydroxybenzoic acid in 
various combinations. A typical example is the antacid suspensions which have high 
pH values which make the esters of the preservatives susceptible to hydrolysis. One 
way to circumvent this problem is to use several preservatives in combination with 
the hope that some quantities of the preservatives will remain to prevent the suspension 
from microbial attack. A report showing the assay of the four esters and 
the parent acid (one of the decomposition products) in drug products in which all 
the preservatives were used has been given [13] . 
Semisolid Dosage Forms The nature of the base (vehicle) used for the fabrication 
of semisolid dosage forms affects their hydrolytic stability. Increased degradation 
of benzylpenicillin sodium in hydrogels of various natural and semisynthetic polymers 
has been reported [14] . Also at pH 6 in Carbopol hydrogels, the percentage 
of undecomposed pilocarpine at equilibrium is a function of the apparent viscosity 
of the medium [15] . 
Solid Dosage Forms Moisture can affect the chemical stability as well as the physical 
stability of solid dosage forms. Reports have shown that the amount of moisture 
adsorbed by tablets in blister packages increased with increasing humidity and it 
resulted in decreased mechanical strength of the tablets [16] . Storage of prednisone 
and erythromycin tablets in moisture - permeable packaging changed drug release 
from the tablets [17, 18] . Drug release from enteric - coated and sugar - coated tablets 
was more susceptible to the effect of humidity than that from fi lm - coated tablets. 
For example, storage of sugar - coated tablets changed the disintegration time, leading 
to increased or decreased dissolution rate [19] . Storage of two chloramphenicol 
capsules at high humidity prolonged the disintegration time and decreased the 
release rate [20] . Further, decrease in drug release from ampicillin capsules during 
storage at high humidity has been reported; it was attributed to the agglomeration 
of drug particles caused by moisture [21] . Gelatin - coated acetaminophen tablets 
exhibited a marked decrease in dissolution rate during storage at high humidity 
[30 ° C, 80% relative humidity (RH), and the effect was moderated by the addition 
of pancreatin to the dissolution medium. Pancreatin was believed to act by cleaving 
the cross - linked gelatin [22] . 
When solid dosage forms such as tablets adsorb moisture, drug present on the 
surface will be dissolved (if it is soluble). The drug in solution on the surface of the 
tablet will be subject to hydrolytic decomposition, and the process will be infl uenced 

FACTORS INFLUENCING STABILITY OF DRUGS AND DRUG PRODUCTS 647 
by the pH of the solution. It has been reported that an increase in the water vapor 
pressure signifi cantly increases the decomposition of aminosalicylic acid [23] . A 
study on water permeation through rubber closures of injection vials indicates that 
rubber closures with a low permeability are often able to take signifi cant amounts 
of water [24] . Thus freeze - dried products, which are highly hygroscopic and highly 
reactive with water, must be adequately protected against uptake of water in order 
to ensure the chemical stability of the active ingredient in the vial. 
7.3.2.2 Oxygen and Oxidation 
Semisolid and Solid Dosage Forms Whether the oxidative degradation process is 
by auto - oxidation (an uncatalyzed reaction which proceeds slowly under the infl uence 
of molecular oxygen) or by chain processes which involve initiation, propagation, 
and termination reactions, molecular oxygen is very important. Consequently, 
every effort should be made to determine and control oxygen concentration in an 
aqueous solution. Florence and Attwood [9] have given an oxidation scheme involving 
a chain reaction. The description of the scheme is in order to provide an appreciation 
of the infl uence of the interaction of molecular oxygen, components of the 
solution and even leachables from the packaging materials on oxidation. It is believed 
that initiation can take place through the free radicals formed by organic compounds 
consequent upon the action of light, heat, or transition metals (e.g., copper and iron) 
present in trace amounts in the buffer. The free radicals readily combine with 
molecular oxygen in the propagation step to form a peroxy radical which then 
abstracts hydrogen from a molecule of organic compound to form a hydroperoxide 
and then create another free radical. The termination occurs when the free radicals 
are destroyed by inhibitors or by side reactions which break the chain. 
The presence of certain functional groups is known to make some drugs highly 
susceptible to oxidative degradation [7, 9] . The phenol functional group present in 
steroids is sensitive to oxidation. The ether group is susceptible to oxidation as 
reported for econazole nitrate and miconazole nitrate. Catecholamines such as 
dopamine and isoproterenol are susceptible to oxidation. Phenothiaxines possess a 
thioether functional group which is oxidized to sulfoxide in the presence of water. 
Many drug molecules possess carbon – carbon double bonds which can be easily 
attacked by peroxyl radicals, leading to oxidative degradation. In fact, the double 
bond is highly susceptible to singlet oxygen (which is highly oxidizing); the singlet 
oxygen is believed to form from the ground state of oxygen called triplet oxygen 
when excited by light. Consequently, the general term oxidation is more than mere 
exposure of a susceptible drug molecule to oxygen but also exposure to conditions 
that favor oxidation such as photolysis [6] . Amphotericin is a polyene antibiotic with 
seven conjugated double bonds. It is oxidized by peroxy radicals with a loss of activity 
and aggregation [25] . Addition of peroxyl radicals to simvastatin can result in 
oxidation with the formation of polymeric peroxides [26] . Carboxylic acid is another 
group highly susceptible to oxidation: The rate of oxidation of ascorbic acid depends 
on oxygen concentration [27] . 
Literature on pharmaceutical sciences is replete with oxidative degradation of 
drugs in solution. 5 - Aminosalicylic acid undergoes oxidation and the product of 
oxidation forms polymeric compounds [28] . Further, morphine has been reported 
to be oxidatively degraded in solution [29] as well as hydrocortisone [30] . 

648 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
Semisolid and Solid Dosage Forms Since most oxidative reactions occur in solution, 
there is not much effort to investigate the effect of oxidative degradation in 
solid dosage forms. Auto - oxidation of tetrazepam tablets has been described [31] . 
7.3.2.3 Light 
Photochemical decomposition is an important route of chemical degradation of 
drugs. Generalization on or a priori prediction of the effect of light on drug is very 
diffi cult because of the strong dependence of the decomposition on the spectral 
properties of drugs as well as the spectral distribution of the source of light [6] . 
Nevertheless, certain trends have been reported which might provide guidance on 
the effect of light on certain drug molecules so that efforts can be made at the early 
stage of drug development to circumvent the problems. Molecules with saturated 
bonds are not capable of interacting with visible or near - ultraviolet light. Those 
molecules with . electrons [aromatic hydrocarbons (nitroaromatic and aryl halide 
functional groups), heterocyclic aromatic compounds, aldehydes, ketones, sulfi des, 
alkenes, and polyenes] absorb light throughout the wavelength range of visible or 
near - ultraviolet light and are very susceptible to photochemical decomposition. 
Sunlight can potentially affect drugs capable of absorbing light at wavelengths 
below 280 nm, and those capable of absorbing light strongly at wavelengths greater 
than 400 nm can be potentially degraded by light in both sunlight and room light 
[7, 32] . 
The importance of the excipients used in the development of drug products 
vis - a - vis photochemical decomposition has been stressed [9] . Photochemical degradation 
of drugs may occur by the direct interaction of the drug with the light of a 
particular wavelength (primary photochemical reaction) or the excipeints may 
absorb light radiation and then transfer the energy to the drug molecule (which 
cannot absorb the incident light) to cause decomposition. Such excipients are called 
photosensitizers. In fact, the ICH Harmonized Tripartite Guideline on Stability 
Testing (Photostability Testing of New Drug Substances and Products) [33] recognized 
the role that excipients in fi nished drug products can play on photostability 
and recommends that photostability testing should be carried out as follows: tests 
on drug substance, tests on the exposed drug product outside of the immediate pack, 
and, if necessary, tests on the drug product in the immediate pack and tests on the 
drug product in the marketing pack. As indicated earlier, oxidative degradation may 
be initiated by light; thus all the factors affecting the stability of drugs and drug 
products do not occur in isolation. Stabilization of drugs against photochemical 
decomposition often involves the use of colored containers (amber glass excludes 
light of wavelength less than 470 nm and can give a good measure of protection to 
drugs sensitive to ultraviolet light) and storage in the dark. Incorporation of ultraviolet 
absorbers in polymer fi lm for tablet coating is another method of protecting 
drugs from photochemical degradation [34] . 
Liquid Dosage Forms Sodium nitropruside in aqueous solution for injection will 
remain stable for up to one year if protected from light; however, its shelf life is 
about 4 h when exposed to normal room light [35] . It has been reported that uric 
acid increases the photostability of sulfathiazole sodium in solutions [36] . Further 
dl - methionine increased the photostability of ascorbic acid in solution [37] . The 

FACTORS INFLUENCING STABILITY OF DRUGS AND DRUG PRODUCTS 649 
infl uence of light on the stability of molsidomine in infusion fl uids has been reported 
[38] . It has been shown that the photochemical degradation of nifedipine is a function 
of light intensity (light source: high - pressure mercury lamp, sunlight, and fl uorescent 
lamp) at room temperature: The amount of photodegraded nifedipine was 
proportional to the number of incident photons [39] . Stabilization of the photochemical 
degradation of daunorubicin in solution (at 290 – 700 nm) by the addition 
of various colorants has been reported: scarlet GN, amaranth, ponceau 6 R, and 
tartrazine [40] . 
Solid Dosage Forms A complex relationship was found between the discoloration 
rate of sulfi somide in tablets irradiated with a mercury lamp in comparison with 
ultraviolet light intensity [41] . The functional attributes of capsules such as drug 
release rate can change following the interaction of dyes and gelatin in capsule shell 
under the infl uence of light. The changes were reported to be enhanced by the 
interaction of light and humidity [42] . The solid - state photodecomposition of the 
drug, tretinoin tocoferil, has been shown to be highly temperature dependent. It was 
found that the rate constant for the photodegradation of tretinoin tocoferil has an 
Arrhenius - type temperature dependence [43] . A report has shown the stabilization 
of indomethacin coloration by the addition of titanium dioxide to gelatin capsule 
shells [44] . Moreover, incorporation of synthetic iron oxides resulted in the stabilization 
of tablets of nifedipine and sorivudine against photodegradation [45] . 
7.3.2.4 Temperature 
The temperature dependence of the rate constant for a chemical reaction is often 
expressed using the Arrhenius equation: 
k Ae E RT = . a / (1) 
where k is the kinetic rate constant, A is the preexponential factor, E a is the activation 
energy, R is the ideal gas constant, and T is the temperature. As predicted by 
the equation, increasing the temperature increases the rate of reaction. Consequently, 
temperature is of considerable importance in consideration of the stability 
of drugs and drug products. Further, if the Arrhenius plot is linear, the degradation 
rates obtained at some temperature levels can be used to predict the rate of degradation 
at other temperatures. This equation is strictly valid for reactions that take 
place in solution, and it assumes there is no change in the degradation mechanism 
or the order of reaction; nevertheless, it has been used to predict the stability of 
dosage forms. For solid dosage forms, the effect of temperature on stability is complicated 
because of the effect of temperature on the excipients and the likely change 
in the properties of the excipients which may affect the stability of active ingredient. 
Though the Arrhenius - type equation has been used to model the dependence of 
reaction rate on temperature for solid dosage forms, the activation energy obtained 
could not be expected to have the same meaning as the one obtained with the solution 
of the drug [9] . 
Where the decomposition shows an approach to equilibrium, as found for vitamin 
E tablets [9, 46] , the equilibrium concentrations of products of degradation and 
reactants are obtained at a series of temperatures. Then the logarithm equilibrium 

650 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
constant K is plotted against the reciprocal of temperature according to the van ’ t 
Hoff equation to model the dependence of temperature on drug breakdown: 
ln const k 
H 
RT 
= . + 
. 
(2) 
Liquid Dosage Forms The arrhenius equation has been used to model the discoloration 
of a liquid multisulfa preparation and the decomposition of liquid multivitamin 
preparations [47 – 49] . The degradation of codeine sulfate in solutions at 
several temperatures has been reported [50] . Further, the infl uence of pH on the 
Arrhenius plots for the hydrolysis of ciclosidomine has been studied [51] . The stability 
of a clofi bride emulsion for oral administration was reduced (aggregation of the 
emulsion) when the storage temperature was increased from 25 ° C to 40 ° C. Moreover, 
storage at 4 ° C resulted in phase separation of the same emulsion [52] . It is 
believed that the stability of a vaccine can be predicted by estimating the loss of 
antigenicity during long periods of storage at different temperatures through accelerated 
stability studies. Though under certain conditions and for some vaccine 
products the Arrhenius equation can be used to make a prediction of real storage 
stability under accelerated stability data, at high temperatures many degradation 
processes of biomolecules are complicated by unfolding or other conformational or 
structural changes. Below the freezing point, the degradation rate can suddenly 
increase or decrease, showing no linear correlation with temperature [53] . Further 
the degradation rate constant k is not the only factor determining the residual antigenicity 
P t of a vaccine. The time t during which a vaccine is stored at a given temperature 
and the initial antigenicity of the vaccine P 0 also have an infl uence which 
can be expressed by the relationship [53, 54] 
k 
P P 
t 
t = 
. 0 
(3) 
Semisolid Dosage Forms Heating easily brings about phase changes in semisolid 
drug products; consequently it is not feasible to use high temperatures to study the 
kinetics of degradation and hence the prediction of stability. Thus stability is often 
evaluated at the temperature in which the formulation is stored. This practice often 
takes a long time and stability problems may not be detected until the studies have 
been in progress for several months or years [55] . Storage temperature can bring 
about changes in the hardening of suppositories such that the time required for the 
suppository to melt is affected. It has been shown that the hardening effect increased 
with increased temperature up to 25 ° C but decreased at higher temperatures owing 
to partial melting of the suppository base [6] . 
Solid Dosage Forms The decomposition of aspirin in tablets made in a microcrystalline 
cellulose base was modeled by fi rst - order kinetics and rate constants obtained 
adhered to an Arrhenius equation [56] . A tablet formulation containing polyvinylpyrrolidone 
(a disintegrant) showed a change in drug release when the temperature 
of storage was increased from 23 to 65 ° C [57] . Changes in the dissolution rate of 
hydrochlorothiazide tablets at room temperature have been estimated from changes 

FACTORS INFLUENCING STABILITY OF DRUGS AND DRUG PRODUCTS 651 
observed at 37, 50, and 80 ° C [58] . A report of the interaction of certain furoic acids 
when tableted with microcrystalline cellulose to cause the formation of carbon 
monoxide has been given [59] . The interaction was fast at 55 ° C and caused the 
tablets to crumble but was less pronounced at room temperature. 
7.3.2.5 Microbes 
Microbial stability is one of the factors affecting drug product reliability. Under 
design and interpretation of stability studies, the Center for Drugs and Biologics 
[5] has elaborated on the microbial quality of drug products. It is clearly indicated 
that drug products containing preservatives to control microbial contamination 
should have the preservative content monitored at least at the beginning and end 
of the projected expiration dating period of the product. Also adequacy of the preservative 
system under conditions of use for multiuse containers should be considered. 
Further, nonsterile products that require control of microbial quality and 
that do not contain preservatives should be tested at specifi c intervals throughout 
the projected dating period according to the release specifi cation for bioburden. 
Topical preparations should also be tested for the presence of pathogens that may 
be identifi ed as potentially harmful. These statements are also reinforced under the 
consideration of the suitability of a packaging system: The container closure system 
should provide the dosage form with adequate protection from factors that can 
cause degradation; microbial contamination is one of the factors listed [8] . 
The U.S. Pharmacopeia/National Formulary USP 28/NF23 [1] refers to the stability 
of a drug product as the extent to which a product remains within specifi ed limits 
and throughout its period of storage and uses the maintenance of the same properties 
and characteristics that it possessed at the time of manufacture. Microbial stability 
(defi ned as the retention of sterility or resistance to microbial growth according 
to the specifi ed requirements; antimicrobial agents that are present retain effectiveness 
within the specifi ed limits) is one of the main types of stability recognized by 
USP 28/NF23. Thus in addition to efforts to stabilize drug products against chemical 
and physical decompositions brought about by environmental factors already discussed, 
liquid and semisolid drug products must be protected against microbial 
attack where necessary. The antimicrobial preservatives added to pharmaceutical 
products should be evaluated during stability studies by either chemical methods 
or microbial challenge test. The microbial challenge test is recommended at the last 
time point during stability testing. 
Liquid Dosage Forms Both sterile products (consideration for the maintenance 
of sterility and preservative effi cacy changes) and nonsterile products (consideration 
for the proliferation of microorganisms) are important. Ophthalmic and parenteral 
liquid products are often prepared in a sterile condition; nevertheless 
appropriative preservatives are often added so that they can remain sterile during 
the period of storage, distribution, and use. Further, other drug products that are 
not required to be sterile and therefore are not sterilized during their manufacture 
and may contain ingredients which can make then susceptible to microbial growth 
are also protected using antimicrobial preservatives [60] . Examples are aqueous 
products such as emulsions and suspensions. The use of methyl, ethyl, propyl, and 
butyl esters of para - hydroxybezoic acid in various combinations in drug products 

652 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
has been mentioned previously [13] . The desire to allow certain amounts of the 
preservatives to remain in the suspension for preservation often informs the use of 
the preservatives in combination. 
Semisolid Dosage Forms The FDA requires that all ophthalmic ointments should 
be sterile. It is also required that a suitable preservative or mixture of preservatives 
to prevent the growth of microorganisms must be added to ophthalmic ointments 
that are packaged in multiple - use containers; chlorobutanol and methyl - and propyl - 
para - hydroxybezoic acid are often used as preservatives in ointments [61] . Since the 
sterilization of fi nished ophthalmic ointment is fraught with diffi culty that centers 
on the stability of the components, aseptic methods of processing are normally used 
to achieve the sterility requirements [62] . Ointments for topical applications, though 
not required to be sterile, are expected to meet acceptable standards for microbial 
content [1] . Creams and gels are often formulated to contain appropriate antimicrobial 
preservatives, as the high water content favors the growth of microbes. 
Solid Dosage Forms At fi rst sight one will expect solid dosage forms to be free 
of microbial contamination. However, reports have shown that tablet contamination 
could arise from the raw materials. Studies were carried out on the effects of the 
production, environment, method of production, and microbial quality of the starting 
materials on microbial loading during the various stages of tablet production 
[63] . High levels of microbial contamination were seen during the wet granulation 
process, which reduced signifi cantly during the drying process. 
7.3.2.6 Active Pharmaceutical Ingredients and Excipients 
Physicochemical Properties of APIs 
Hygroscopicity Hygroscopicity is the amount of moisture absorbed by a powder 
when it is exposed to an atmosphere of a known relative humidity. It is part of the 
preformulation program and its importance lies in the use of the information gained 
to decide whether or not a particular salt of the drug in question could be used for 
a dosage form design. Report has shown that fl urazepam is used as a monosulfate 
rather than the disulfate which however has many other desirable characteristics: It 
is so hygroscopic that it will remove water from a hard - shell capsule and then render 
it very brittle [64] . Knowledge of the hygroscopicity of the APIs can aid in the selection 
of packaging materials. 
Crystalline State or Amorphous State and Polymorphism Solid drug particles occur 
as pure crystalline substances of defi nite indentifi able shape or as amorphous particles 
without defi nite structure. The energy required for a molecule of a drug to 
leave the lattice in a crystal is very much greater than the energy required in an 
amorphous powder. Drugs in the crystalline state have low reactivity. A linear relationship 
has been found between the solid - state degradation rate constant of various 
vitamin A derivatives and the inverse of the melting point [65] . Moreover, results 
of studies on the relationships between degradation rate and crystallinity of . - 
lactam antibiotics such as cefazolin indicate that a drug with low crystallinity tends 
to have decreased chemical stability [66] . 

FACTORS INFLUENCING STABILITY OF DRUGS AND DRUG PRODUCTS 653 
Polymorphism is the phenomenon by which a solid particle may exist in different 
crystalline forms called polymorphs. Medicinal compounds form different types of 
crystals, depending on the conditions (temperature, solvent, time) under which 
crystallization is induced. The molecules of the drug exhibit different space – lattice 
arrangements in the crystalline form from one polymorph to another. It should be 
emphasized that polymorphism exists only in the solid state. Only one form of a 
pure drug substance is stable at a given temperature and pressure with the other 
forms, called metastable forms, converting in time to the stable crystalline form. It 
is common for a metastable form of a medicinal agent to change form even when 
present in a completed pharmaceutical preparation, although the time required for 
a complete change may exceed the normal shelf life of the product itself. Although 
the drug is chemically indistinguishable in each form, polymorphic forms differ signifi 
cantly with respect to a number of physical properties such as density, melting 
point, solubility, stability, and dissolution characteristics, which are of prime importance 
to the proper development of the dosage forms containing the drug. Solid - 
state hydrolysis of carbamazepine from needle - shaped crystals with a high crystalline 
order was found faster than that of beam - shaped and prismatic forms and the reactivity 
of carbamazepine to light is strongly dependent on the crystalline form of the 
drug [67, 68] , and a similar report has been given for the photodegradation of 
furosemide [69] . 
Modifi cation of Chemical Structure of Drug The use of a Hammett linear free - 
energy relationship to investigate the effects of substituents on the rates of aromatic 
side - chain reactions such as hydrolysis of esters has been alluded to earlier vis - a - vis 
attainment of optimum stability [9, 10] . Degradation of erythromycin under acidic 
pH conditions is inhibited by substituting a methoxy group for the C - 6 hydroxyl as 
found for the acid stability of clathromycin, which is 340 times greater than that of 
erythromycin [70] . 
Vapor Pressure Active pharmaceutical ingredients with suffi ciently high vapor 
pressures can become lost by their volatilization through the containers such that 
the stability and content uniformity suffer. Further such compounds can interact 
with other drug molecules and packaging components [71] . Nitroglycerin is notorious 
for this behavior and special packaging materials are needed for dispensing 
sublingual nitroglycerin tablets. When unstabilized nitroglycerin sublingual tablets 
were stored in enclosed glass containers, the high volatility of the drug gave rise to 
redistribution of nitroglycerin among stored tablets with concomitant deterioration 
in the uniformity of the tablets on storage [72] . 
Pharmaceutical Excipients Drugs are rarely administered to humans as pure 
chemical compounds. What is given is a preparation containing the drug and the 
materials of formulation called excipients. These excipients are added for various 
purposes to ensure adequate performance of the drug products. However, some of 
the excipients can exert deleterious effects on the stability of drug products. The 
degradation of codeine has been reported to be susceptible to the infl uence of 
buffer: Its hydrolytic rate constant in phosphate buffer at pH 7.0 is about 20 times 
faster than in buffered solution at the same pH [9] . Further, various phosphate 
species were found to enhance the degradation of bezypenicillin [73] , cefadroxil 

654 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
[74] , carbenicillin [75] , and spironolactone [76] . Some drugs may not be capable of 
absorbing light directly to undergo photolysis, but the excipients present in the formulation 
may absorb light radiation and transfer the absorbed energy to the drug 
to cause degradation [9] . Furthermore, transition metals such as copper and iron 
which are present in trace amounts in buffer are capable of initiating the chain 
process in oxidative degradation. Preservatives may affect the stability of dosage 
forms. Preservation of zinc insulin (in a multidose vial) with phenol can cause physical 
instability of the suspension. In the case ophthalmic pharmaceutical products, 
all raw materials used in the preparation should be of the highest quality available. 
Consequently, the practice by some drug manufacturers is to compound all ophthalmic 
drugs using water for injection. 
Among other factors, the stability of APIs in the ointment base is of utmost 
importance in the selection of the appropriate base [60] . Studies have shown that 
the stability of hydrocortisone in ointment fabricated using polyethylene glycol base 
is very poor [77, 78] . Semisolid products often turn yellow or brown with age because 
of oxidative decompositions involving the base (especially natural fats and oils) used 
in the formulation of the products [55] . Extensive oxidation of natural fatty materials, 
termed rancidifi cation, is associated with the development of a disagreeable 
odor. Various phase transitions, crystallizations, and transesterifi cation reactions of 
suppository bases are believed to be responsible for the hardening of suppository 
on storage, which may adversely affect the quality of the suppository, such as poor 
availability of the active ingredients [6] . Moreover, aspirin was found to decompose 
in polyethylene glycols used as components of suppository bases; the decomposition 
was attributed to a transesterifi cation reaction which gave rise to salicylic acid and 
acetylated polyethylene glycol [79] . Degradation of aspirin was also found when 
cocoa butter was used as the suppository base [80] . Polyethylene glycol esters of 
indomethacin were found in stored suppositories fabricated with polyethylene 
glycol bases [81] . 
Reports have shown that magnesium strearate (a lubricant) accelerated the discoloration 
of tablets containing amines and lactose [82] . Excipients such as binders 
(e.g., povidone) and disintegrants such as crosspovidone containing phenolic impurities 
can impact negatively the photostability of tablets by participating in the 
free - radical reaction [83] . Enhancement of the oxidation of phenylbutazone by dyes 
via the production of singlet oxygen that participates in chain reactions has been 
reported [84] . Excipients with high moisture content can increase the decomposition 
of drugs by increasing the amount of moisture associated with the drug. The percentage 
decomposition of aminosalicylic acid increased with increase in the water vapor 
pressure [85] . 
7.3.3 DRUG PACKAGING AND PACKAGING MATERIALS 
7.3.3.1 Introduction 
A packaging system, often called a container closure system, comprises packaging 
components that all together contain and protect the drug product [8] . It is customary 
to refer to two types of drug product packaging components: primary packaging 
component (the package components that are directly in contact with the drug 

product) and the outer packaging components (usually a series of enclosures) often 
referred to as secondary packaging components, which include cartons, corrugated 
shippers, and pallets [86] . Contrary to the general belief that, with the possible 
exception of labeling, outer packaging components do not require any special consideration 
when applied to drug products, reports have shown that the outer package 
can extend the shelf life of drug products considerably: A blister pack exposed 
naked at 37 ° C, 90% RH had a shelf life of 21 days (suggesting that a blister might 
not be suitable for packaging the product). However, when the same blister pack 
was put in a carton and the carton in turn placed in a display outer overwrap and 
stored under the same conditions in the company ’ s warehouse, the product was still 
in specifi cation after 6 years [87] . 
The traditional defi nition of pharmaceutical packaging as a system capable of 
containing the drug product such that it remains safe and effi cacious within its shelf 
life has shifted to a defi nition that recognizes the importance of packaging in drug 
product performance. Consequently, a pharmaceutical packaging has been defi ned 
as the combination of components necessary to contain, preserve, protect, and 
deliver a safe, effi cacious drug product [87] . In fact, this defi nition echoes the statement 
in the guidance for industry Container Closure Systems for Packaging Human 
Drugs and Biologics [8] that the container closure system must be suitable for the 
intended use (it should adequately protect the dosage form; it should be compatible 
with the dosage form, and it should be composed of materials that are considered 
safe for use with the dosage form and the route of administration). It goes further 
to indicate that if the packaging system has a performance feature in addition to 
containing the product, the assembled container closure system should be shown to 
function properly. Thus drug products such as prefi lled syringes, transdermal patches, 
metered - dose inhalers, and nasal sprays all contain formulations whose successful 
delivery to the fi nal consumer (the patient) depends on the proper functioning of 
the packaging system. The requirements for protective, compatibility, safety, and 
performance characteristics of packaging systems vis - a - vis suitability for intended 
use vary from one type of dosage form and from one route of administration to 
another and a table that serves as a guide to the pharmaceutical industry on the 
suitability considerations for common classes of drug products has been provided 
elsewhere [8] . 
7.3.3.2 Materials of Fabrication of Packaging Components 
The science of packaging materials vis - a - vis pharmaceutical and cosmetic products 
is very broad. The selection of a package often begins with a determination of the 
product ’ s physical and chemical characteristics, its protective needs, and its marketing 
requirements. The materials selected must have the following characteristics: 
They must protect the preparation from various hazards (physical, climatic, chemical, 
and biological) [87] ; they must not be reactive with the product (the current 
good manufacturing practice (CGMP) regulations [88] ); they must not impart to 
the product tastes or odors; they must be nontoxic; they must be FDA approved 
(it should be indicated that the FDA approves only the materials used in the container 
and not the containers as such and an FDA publication lists substances generally 
recognized as safe (GRAS)]; they must meet applicable tamper resistance 
DRUG PACKAGING AND PACKAGING MATERIALS 655

656 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
requirements; and they must be adaptable to commonly employed high - speed packaging 
equipment [89] . 
Glass Containers Glass is commonly used in pharmaceutical packaging because 
it possesses superior protective qualities, it is economical, and containers are readily 
available in a variety of sizes and shapes. It is essentially chemically inert, impermeable, 
strong, and rigid and has FDA clearance. Glass does not deteriorate with age, 
and with a proper closure system, it provides an excellent barrier against practically 
every element except light. Colored glass, especially amber, can give protection 
against light when it is required. The major disadvantages of glass as a packaging 
material are its fragility and its weight. 
The USP/NF [1] recognizes four major types of glass containers: Type I borosilicate 
glass containers are used for preparations that are intended for parenteral 
administration. Type II glass containers which are made of soda - lime glass that is 
suitably dealkanized are usually employed for packaging acidic and neutral parenteral 
preparations (type I glass may also be used for this purpose, but it is used for 
packaging alkaline parenteral preparations). Type III soda – lime glass containers are 
not used for parenteral preparations except if stability data indicate otherwise. The 
containers of the fourth category (type NP glass) are intended for packaging nonparenteral 
articles (i.e., those intended for oral and topical use). 
Plastic Containers Plastics in packaging have proved useful for a number of 
reasons: The ease with which they can be formed, their high quality, and the freedom 
of design to which they lend themselves; plastic containers are extremely resistant 
to breakage and thus offer safety to consumers along with reduction of breakage 
losses at all levels of distribution and use [89] . Plastic - coated glass bottles are used 
in aerosol containers to protect from fl ying glass in the event of glass shattering; 
further, plastic coating around the neck of the container serves to absorb some of 
the shock from the crimping operation and decreases the danger of breaking around 
the neck. 
Plastic containers for pharmaceutical products are primarily made from the following 
polymers: 
(a) Polyethylene is a good barrier against moisture but a relatively poor one 
against oxygen and other gases; most solvents do not attack polyethylene, and it is 
unaffected by strong acids and alkalis; both high - and low - density polyethylene are 
used and they are long - chain polymers. The density of polyethylene directly determines 
the four basic physical characteristics of the blow - molded container — 
stiffness, moisture – vapor transmission, stress cracking, and clarity or translucency — 
and these characteristics determine the suitability of polyethylene used in containers 
for packaging drugs. The determinants of suitability of polyethylene have been 
identifi ed by the USP/NF [1] as follows: oxygen and moisture permeability, modulus 
of elasticity, melt index, environmental stress crack resistance, and degree of crystallinity 
after molding). 
(b) Polypropylene polymers are long - chain polymers synthesized from propylene 
or propylene and other olefi ns under controlled conditions of heat and pressure 
with the aid of catalysts. Polypropylene does not stresscrack under any conditions. 
Except for hot aromatic or halogenated solvents, which soften it, this polymer has 

good resistance to almost all types of chemicals, including strong acids, alkalis, and 
most organic materials. Its high melting point makes it suitable for boilable packages 
and for sterilizable products. Lack of clarity is still a drawback, but improvement is 
possible with the construction of thinner walls. Polypropylene is an excellent gas 
and vapor barrier. Its resistance to permeation is equivalent to or slightly better 
than that of high - density or linear polyethylene, and it is superior to low - density or 
branched polyethylene [1, 89] . One of the biggest disadvantages of polypropylene 
is its brittleness at low temperatures. 
(c) Polyethylene terephthalate (PET) is a condensation polymer typically 
formed by the reaction of terephthalic acid or dimethyl terephthalate with ethylene 
glycol in the presence of a catalyst. PETG resins are high - molecular - weight polymers 
prepared by the condensation of ethylene glycol with dimethyl terephthalate 
or terephthalic acid and 15 – 34 mole % of 1,4 - cyclohexanedimethanol. PET and 
PETG resins and other ingredients used in the fabrication of these bottles conform 
to requirements in the applicable sections of the Code of Federal Regulations , Title 
21, regarding use in contact with food and alcoholic beverages: PET and PETG 
bottles that are interchangeably suitable for packaging liquid oral dosage forms [1] . 
Other polymers that have been used or proposed for use in packaging of drug 
products include polycarbonate, polyvinyl chloride, and polystyrene and to a lesser 
extent polymethyl methacrylate, polyethylene terephthalate, polytrifl uoroethylene, 
and polyamides [89] . 
Over the years efforts have been made to understand the various parameters to 
consider in the choice of plastics for packaging of drug products. Drug – plastic considerations 
have been divided into fi ve separate categories: permeation, leaching, 
sorption, chemical reaction, and alteration in the physical properties of plastics or 
products [89] . 
Metals Metals are used as collapsible tubes and in aerosol containers. The most 
common metals in use are tin, aluminum, and lead. Tin is the most expensive, while 
lead is the cheapest. Laminates of tin - coated lead provide the appearance and oxidation 
resistance of straight tin at lower prices [89] . Tin is the most chemically inert 
of all collapsible tube metals. It offers a good appearance and compatibility with a 
wide range of products. Aluminum tubes provide the attractiveness of tin at relatively 
lower cost. Lead has the lowest cost of all tube metals and is widely used for 
nonfood products such as adhesives. However, with internal linings, lead tubes are 
used for such products as fl uoride toothpaste. If the product is not compatible with 
bare metal, the interior can be fl ushed with wax - type formulations or with resin 
solutions. 
Aluminum is used as the material of construction for some aerosol drug products. 
Also a three - piece tin - plated steel container fi nds use in topical pharmaceutical 
aerosols, and to decrease the compatibility problems, an internal organic coating is 
often used [90] . 
Rubber Stoppers, cap liners, and bulbs for dropper assemblies used in the pharmaceutical 
industry are made form rubber. The rubber stopper is used primarily for 
multiple - dose vials and disposable syringes. The rubber polymers most commonly 
used are natural, neoprene, and butyl rubber. As certain performance expectations 
DRUG PACKAGING AND PACKAGING MATERIALS 657

658 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
are expected from rubber, certain ingredients are commonly found in a rubber 
closure: for example, rubber vulcanizing agent, accelerator/activator, extended fi ller, 
reinforced fi ller, softener/plasticizer, antioxidant, pigment, and waxes. The complexity 
of rubber necessitates caution, especially when used in packaging parenteral 
products: When the rubber stopper comes in contact with parenteral solution, it may 
absorb active ingredient, antibacterial preservative, or other materials, and one or 
more ingredients of the rubber may be extracted into the liquid [89] . 
7.3.4 EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 
7.3.4.1 Introduction 
The purpose of stability testing is to investigate drug product (with and without 
container closure system) changes with time under the infl uence of various hazards 
such as temperature, humidity, and light and to establish a shelf life for the drug 
product and to recommend storage conditions. Generally, packaging suitability is 
based on four attributes: protection, safety, compatibility, and performance (function 
and/or drug delivery). Certain factors that must be considered in evaluating container 
closure systems are as follows: materials of construction of the container 
closure system, surface treatments and/or processing aids, dosage form active ingredients 
and excipients, sterilization and/or other related processing, and storage 
conditions. Guidelines are available in offi cial compendia for the evaluation of 
container closure system. First guidelines are given for the evaluation of the container 
or the package materials: physicochemical and biological tests to evaluate 
glass and plastic bottles, metal closures, elastomeric closures, fl exible and blister 
materials, syringe components, and aerosol packaging. Second, guidelines are available 
for stability studies (accelerated, short - term tests and long - term tests) to characterize 
the effects of the package component on the product. The general belief is 
that it is impossible to have a completely inert container closure system; thus the 
evaluation is often designed to identify, characterize, and monitor these interactions 
to achieve a safe, unadulterated, stable, and effi cacious drug product [86] . At the 
end of the evaluation of a container closure system in the pharmaceutical industry, 
a technical report is issued which indicates the approach, results, and the 
conclusion. 
7.3.4.2 Solid Dosage Forms 
A stability study of losartan/hydrochlorothiazide tablets was carried out in three 
stages [91] . First, a stress test (forced - degradation study) was carried out without 
immediate packaging by exposing the tablets to severe storage conditions (separately 
50 and 50 ° C, 80% RH). After four weeks it was concluded that the tablets 
are sensitive to moisture. Second, preliminary testing was carried out to select 
the packaging system. Two packaging systems with different barrier properties 
were studied: polyvinyl chloride 250 . m – polyethylene 25 . m – polyvinylidenechloride 
60 g/m 2 foil (PVC – PE – PVdC – Al blisters capable of protecting the tablets partially 
from water vapor and gases) and oriented polyamide 25 . m – aluminum foil 
45 . m – polyvinyl chloride 60 . m foil (OPA – AL – PVC – Al blisters capable of giving 

absolute protection to the tablets). Drug assay, impurities, disintergration, and 
appearance were monitored. Following six months storage at 40 ° C, 75% RH, it was 
concluded that the protection against moisture offered by PVC – PE – PVdC – Al blisters 
was not suffi cient and that losartan/hydrochlorothiazide tablets should be packaged 
in OPA – AL – PVC – Al blisters. Finally formal stability testing of the tablets 
packaged in OPA – AL – PVC – Al blisters was carried out. The parameters monitored 
were as follows: assay, impurities, dissolution, disintegration, hardness, water, 
appearance, and microbiological quality. The results obtained during 6 months of 
accelerated and 12 months of long - term stability testing showed that losartan/ 
hydrochlorothiazide tablets packaged in OPA – AL – PVC – Al blisters were 
chemically, physically, and microbiologically stable and a shelf life of 24 h was 
proposed. 
Studies were carried out to use the drug product and package characterization 
data (tablet equilibrium moisture content, degradation rate of unpacked product, 
and moisture barrier properties of the packages) to predict and subsequently confi rm 
the package material that provided adequate stability for a moisture - sensitive compound 
(PGE - 7762928) [92] . The physical and chemical stability [high - performance 
liquid chromatography (HPLC) assay] of the products were determined at 2, 4, 6, 
8, 12, and 24 weeks at ICH conditions. At 6 months at 40 ° C, 75% RH, the percentage 
of active ingredient was 84% in polyvinyl chloride blisters, 91% in cyclic olefi n 
blisters, 97% in aclar (polychlorotrifl uoroethylene) blisters, 100% in cold - form aluminum 
blisters, and 99% in a high - density polyethylene bottle with a foil induction 
seal. The stability results for the packaged product were fairly consistent with the 
predictions based on the moisture sensitivity of the product and moisture barrier 
properties of the respective package. Based on the goal of 90% assay at 6 months, 
cold - form aluminum and Aclar blister packaging and high - density polyethylene 
bottles with foil induction seals provide acceptable PGE - 7762928 tablet stability. 
Though inert atmosphere packaging is a common practice in the parenteral sector 
of the pharmaceutical industry, there are relatively few examples of solid dosage 
forms that are packaged under reduced oxygen levels. Model granule formulations 
which include a drug known to exhibit oxidative degradation were packaged in 
stoppered glass vials maintained at different head space oxygen concentrations and 
head space relative humidities and stored at 40 ° C. The oxidative degradation was 
quantifi ed as a function of time and the data showed the dependence of oxidative 
degradation on head space oxygen concentration, relative humidity, drug loading, 
and time. It was recommended that the use of oxygen scavengers in bottles as well 
as inert atmospheric packaging foil – foil blister lines could be options for achieving 
pharmaceutical packages with low oxygen concentrations [93] . 
The effect of packaging and storage conditions on the in vitro performance of 
carbamazepine (CBZ) tablets was studied. Tablets used were Tegretol and the 
Egyptian generic Tegral, both presented in PVC – aluminum strip seals inside a 
carton and the German product Finlepsin dispensed in bottles of 50 tablets. In vitro 
performance was assessed through dissolution testing while chemical stability of 
CBZ was assessed via HPLC [94] . Tegral tablets were not affected by the tested 
stress conditions. Tegral tablets packaged in strips at 50 or 60 ° C and 75% RH 
showed fast disintegration and dissolution. The effect of 40 ° C, 97% RH for 6 months 
was similar to 1 month storage at 40 ° C, 97% RH; the tablets hardened and dissolved 
less than fresh Tegral tablets. Removal of Tegral tablets from their original strips 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 659

660 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
resulted in only 7% dissolution in 60 min. For Finlepsin, the effect of 97% RH at 
40 ° C was more profound than 97% RH at 25 ° C, but both conditions caused a 
decrease in dissolution rate, the extent of which was found to be dependent on the 
tablet position in the bottle. However, all stressed tablets examined in this study 
showed no change in the chemical stability of CBZ tablets under all stressed 
conditions. 
The model involving mass transfer characteristics of packaging materials and 
chemical stability has been used to characterize the diffusion of water through 
polyvinyl chloride blister packaging. It was found that the effect of moisture sorption 
on tablet crushing strength was dependent on the formulation and that it is 
necessary to consider the characteristics of the formulation when selecting a package 
[95] . Information on the tablet moisture sensitivity of compressed tablets together 
with package moisture permeability has been used to develop a physical model 
which predicted the changes in the crushing strength of tablet under a variety of 
storage and packaging conditions. The predictions are described as being useful in 
establishing long - term stability testing protocols. They are also useful in predicting 
what tablets stored in blister packages might reasonably be expected to experience 
in the marketplace where short - and long - term oscillating conditions are likely to 
occur. The theoretical predictions and experimental results emphasize the importance 
of matching the dosage formulation characteristics to the package material 
and testing conditions so that a more rational selection of packaging material and 
testing protocols can be made [96] . 
It is known that moisture transfer through packaging materials can affect hard 
gelatin capsules in two ways: changes in the properties of the shell and changes in 
the properties of the materials loaded in the capsule (contents of the capsules). 
Studies on capsules have recently examined the changes in both dissolution rates 
and potency of two brands of amoxicillin capsules packaged in PVC blister package 
(PVC blister having a thickness of 0.27 mm and aluminum base fi lm of thickness 
27 . m) and laminated – type package (the laminated package consisting of three 
layers: a nitrocellulose lacquer outer layer, a soft aluminum fi lm intermediate layer, 
and a polyethylene inner layer). The two brands of amoxicillin stored outside the 
package at 76, 80, and 92% RH and inside the package at 92% RH showed no signifi 
cant changes in dissolution profi les. Storage at 80 and 92% RH of the capsules 
of the two brands outside their packages resulted in a signifi cant loss of amoxicillin 
potency. The laminated - type package afforded better protection compared with the 
PVC – aluminum blister package. Following 20 weeks storage at 92% RH at room 
temperature, only 6.4% loss in amoxicillin potency occurred for capsules packaged 
in the laminated package compared with 51.8% loss in the blister package [97] . 
The effects of two storage conditions (50 ° C, 50% RH and 40 ° C, 90% RH) on the 
properties (hardness, disintegration, tablet weight, dimensions, dissolution rate, and 
content of API) of two brands (A and B) of fi lm - coated erythromycin stearate 
tablets and one brand (C) of enteric - coated tablets of erythromycin base were 
examined. The tablets were stored in paper bags, plastic dispensing bottles, and glass 
bottles. Aside from an increase in hardness by storage at 40 ° C, 90% RH for brand 
C tablets and a decrease in disintegration time for tablets of brand A under the 
same conditions, little or no changes were seen in most of the physical properties. 
However, dissolution rates signifi cantly decreased for all the tablets by storage 
under the two conditions. Glass containers offered better protection for the tablets, 

and consequently, the tablets retained higher dissolution rates as compared to 
tablets stored in plastic or paper containers [98] . 
The sorption of nitroglycerin by thermoplastic polymers and the stability of 
molded nitroglycerin tablets in strip packaging were studied. The polymers investigated 
varied greatly in their affi nity for nitroglycerin, the order of decreasing affi nity 
being vinyls > low - density polyethylene > ionomers > high - density polyethylene. 
With the proper choice of packaging, molded nitroglycerin tablets stabilized with 
povidone maintained acceptable potency for up to two years at 26 ° C when strip 
packaged in unit - dosage form. Chemical decomposition by hydrolysis of nitroglycerin 
was also investigated. Povidone accelerated the decomposition of nitroglycerin; 
at high temperature, decomposition was a signifi cant factor in tablet stability for 
tablets containing povidone [99] . A liquid dispersion system of etodolac (20%) and 
Gelucire 44/14: d - . - tocopheryl polyethylene glycol 1000 succinate (TPGS) blend 
(80%) in different ratios was prepared. The capsule formulation was subjected to 
stability studies at different temperature humidity conditions based on ICH guidelines. 
Physical and chemical properties of the dispersion did not change during a 
period of storage at room temperature and at 4 ° C, 0% RH. However, the relative 
humidity and storage time exerted an effect on the dissolution behavior of etodolac. 
It was concluded that changes in dissolution behavior after storage under conditions 
of high humidity and temperature might be related to the formation of etodolac 
microcrystal and water absorption by the carrier during storage, thereby necessitating 
packaging in moisture - resistant packaging [100] . 
7.3.4.3 Nonsterile Liquid Dosage Forms 
The effect of packaging materials on the stability of ultraviolet (UV) A, UVB, and 
infrared sunscreen emulsions after storage in different types of packaging materials 
(glass and plastic fl asks; plastic and metallic tubes) has been reported [101] . The 
samples (emulsions containing benzophenone - 3, octyl methoxycinnnamate, and 
Phycocorail) were stored at 10, 25, 35, and 45 ° C and representative samples were 
analyzed after 2, 7, 30, 60, and 90 days. Data showed that sample emulsions stored 
at different temperatures had similar rheological behaviors within 3 months and 
there were no signifi cant changes in the physical and chemical stability of emulsions 
stored in different packaging materials. Glass and plastic packaging materials were 
found adequate for storing the products. 
Reports have shown that chloroquine binds to glass but not to certain plastics. 
This information is very important in laboratory studies where signifi cant reductions 
in chloroquine concentration may occur when the drug is prepared and stored in 
glass containers [102] . Further, buffered chloroquine diphosphate solutions of 
varying pH and concentration were stored in soda or borosilicate glass. Storage in 
soda glass showed a decrease in original drug concentration of up to 60 and 97% 
in test tubes and glass wools, respectively, Borosilicate glass did not show any 
binding. The highest binding recorded was at physiological pH. Thus borosilicate 
glass should be used when storing, assaying, or carrying out sensitivity tests for 
malaria in order to avoid loss of chloroquine from solution [103] . 
Oxygen is often used as a challenge to fi nd out whether a particular drug is likely 
is to be affected by oxidative breakdown. The procedure involves storing the drug 
product in solution in ampoules purged with oxygen and then comparing their rate 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 661

662 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
of breakdown with similar solutions stored under nitrogen. Drug formulations 
shown to be susceptible to oxidation can be stabilized by replacing the oxygen in 
the storage containers with nitrogen or carbon dioxide; antioxidant can be added 
or the choice of container should aim at avoiding the presence of metal ions [9] . In 
the case of drug products (solution, semisolid, and solid) that are photolabile, the 
usual practice is to store them in containers which exclude UV light. It is believed 
that exposure to light in this wavelength range is the most frequent cause of photodecomposition. 
Amber glass is very effective in eliminating or reducing photodegradation 
because it is capable of excluding light of less than 470 nm [9] . Photolabile 
drugs can also be stored in the dark. 
7.3.4.4 Sterile Liquid Dosage Forms 
The use of prepackaged syringes in the hospital setting is increasing due to the 
advantages offered by their use and the current interest in unit - dose medication. It 
is known that all the various dosages and combinations prescribed in the hospital 
are not commercially available; consequently many hospital pharmacies are developing 
or operating a centralized parenteral admixture program. Packaging of the 
drug products requires consideration of the sterility and particulate matter content 
of single - syringe injections as well as compatibility between drugs and with the 
syringe components. Though plastic syringe packaging has a number of advantages, 
contact between plastic containers and drugs can present a number of problems, 
such as leaching, sorption, permeation, chemical reactivity of the plastic polymer, 
and alteration in physical properties of the plastic. Studies have been carried out on 
a commonly prescribed drug combination (hydroxyzine hydrochloride, meperidine 
hydrochloride, and atropine sulfate) used as preanesthetic medication to compare 
possible differences in stability between mixtures stored in glass and plastic containers. 
The combination drugs were stored in both glass and plastic syringes at 25 and 
3 ° C for a 10 - day period. Analyses were performed at intervals throughout the time 
period utilizing visual examination, pH determination, UV absorption spectra, and 
gas chromatography. No signifi cant degradation of the syringe contents or appearance 
of additional constituents was detected in any of the admixture preparations. 
Thus the storage of such preparations in glass versus plastic syringes yielded no 
signifi cant differences in product stability [104] . 
Report has shown the infl uence of the type of container on ceftazidime stability 
in intravenous solutions. Polypropylene (PP) and PVC (100 mL each) bags and 
100 - mL glass bottles were fi lled with 5% dextrose or 0.9% sodium chloride solutions 
containing ceftazidine (40 mg/mL) and stored at 20 and 35 ° C. One - milliliter samples 
were taken from each container at 0 and 20 h and assayed. Pyridine (the main degradation 
product) levels increased during storage and were higher in PVC and PP 
bags than in glass bottles in both diluents. Solutions stored in PP bags showed better 
stability than in PVC bags. Glass bottles seem to be the better container for storing 
ceftazidime solutions [105] . Studies have been carried out to evaluate the effect of 
diluent type, storage conditions, and nature of package on the stability of reconstituted 
Parecoxib sodium for injection (PSI) [106] . Results showed that the PSI 
reconstituted with normal saline, bacteriostatic normal saline, 5% dextrose injection, 
and 5% dextrose injection and half normal saline met visual acceptance criteria and 
showed almost no degradation under storage conditions, and no signifi cant differ

ences were seen between storage in glass vials or PP/glass syringes. However, PSI 
reconstituted with lactated Ringer ’ s injection and lactated Ringer ’ s and 5% dextrose 
injection showed visual precipitation in many vials as confi rmed by HPLC 
assay values at all points. 
The stability of an ophthalmic solution formulation of unoprostone isopropyl 
(UI) (a prostaglandin - like compound marketed as Rescula for treatment of elevated 
intraocular pressure in patients with primary open - angle glaucoma or ocular hypertension) 
was investigated in two types of packaging materials: PP and low - density 
polyethylene (LDPE). The concentrations of UI and its degradation products were 
monitored as a function of time. It was found that the rate of disappearance of drug 
was faster for the formulation stored in LDPE bottles than that stored in PP bottles. 
Further studies indicated that the inferior stability observed with LDPE packaging 
was primarily due to the sorption of UI to the packaging material and to a lesser 
degree chemical degradation. The sorption was found to be temperature dependent: 
Lowering the temperature reduced the sorption, thus improving the shelf life of the 
product [107] . 
The USP has proposed large - globule - size limits to ensure the physical stability 
of lipid injectable emulsions, expressed as the percent far greater than 5 . m, or 
PFAT 5 
, not exceeding 0.05%. Recent studies showed that injectable emulsions packaged 
in newly introduced plastic containers exceed the proposed USP PFAT 5 
limits 
and subsequently become signifi cantly less stable during simulated syringe - based 
infusion. Although modest growth in large - diameter fat globules was observed for 
the glass - based lipids, they remained within proposed USP globule size limits 
throughout the study. Thus glass - based lipids seem to be a more stable dosage form 
and potentially a safer way to deliver via syringe infusion to critically ill neonates 
[108] . 
Considerable interest has been shown, in recent years, in the clinical use of intravenous 
nitroglycerin for the treatment of myocardial infarction and in open heart 
surgery. In its intravenous use loss of the drug to PVC bags has been identifi ed as 
a problem and has been characterized as a diffusionally controlled absorption 
process having a half - life for fractional absorption of 3.2 h at 30 ° C [109] . Moreover, 
in recent studies, administration set tubing has been implicated as the source for the 
observed loss of nitroglycerin from solution [110 – 112] . Quantitative studies have 
been reported on the loss of nitroglycerin to plastic intravenous tubing. It was found 
that in the short time periods the loss of nitroglycerin from normal saline solutions 
to PVC tubing could be treated as an adsorption process. The rate of adsorption 
was rapid and could be quantifi ed as an apparent fi rst - order process. The half - life 
of the loss was 2.6 min [113] . Recently a model describing the loss of nitroglycerin 
from the solution in the plastic bags as a rapid adsorption onto the plastic surface 
followed by partitioning into the plastic was proposed [114] . The mechanism of loss 
of nitroglycerin stored in plastic and glass containers was studied from an equilibrium 
and kinetic approach. Data showed that nitroglycerin was removed from 
aqueous solution by the plastic container material through an absorption process. 
Nitroglycerin loss from aqueous solution did not occur by hydrolysis since solutions 
stored in glass containers at pH 5.7 and 35 ° C retained potency for at least 48 h. 
Although experimental data supported the thesis that the uptake of nitroglycerin 
was migration of the drug into the plastic matrix, the fi nding did not exclude the 
possible adsorption of the drug on the surface: The rate of adsorption may be much 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 663

664 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
faster than the rate of absorption with the result that any adsorption may be 
obscured [115] . 
An investigation has been carried out to defi ne the kinetics and mechanism of 
the interaction between various drugs and plastic infusion bags and to develop criteria 
that may be used in the prediction of such interactions. The sorption behavior 
of warfarin sodium, various benzodiazepines, and other drugs with PVC infusion 
bags was examined. The sorption of some compounds by PP infusion bags was also 
investigated. The sorption kinetics fro warfarin and diazepam could be accounted 
for by a diffusional model in which the loss of drug is determined primarily by the 
diffusivity of the compound in the plastic matrix. The rate and extent of sorption of 
warfarin showed a dependence on pH which could be interpreted in terms of ionization 
of the drug: Only the un - ionized form was sorbed. A rank - order relationship 
was found between the initial rate of uptake by the plastic bag and the hexane – water 
partition coeffi cients of the compounds. Thus hexane – water partition coeffi cient 
may be useful for the prediction of interactions between a drug substance and PVC 
infusion bags. Sorption of the compounds by infusion bags made from PP was insignifi 
cant, except for the highly lipophilic medazepam. Thus PP bags may be safer to 
use than PVC bags [116] . 
Drugs to be administered via the parenteral route which are unstable in water 
are often formulated as solids for reconstitution. Such drugs may be presented in 
vials or in two - compartment cartridges with powder in one side and diluent in the 
other. Rubber closures are used as primary packaging material because of their 
unique properties, such as elasticity for piercing and self - sealing which is maintained 
under freeze - drying conditions. Further, vial seals made from butyl or halogenated 
butyl rubber show low moisture vapor and gas transmission rate. During storage 
rubber closures are in intimate contact with vial contents and are a potential source 
of product contamination: shedding of particles, migration of benzothiazoles, leaching 
of metal ions, antioxidants and oligomers, silicone oil, sulfur, and paraffi n wax. 
Reconstituted solutions in aqueous vehicles show haze formation which arises from 
precipitation in oversaturated solutions, polymerization of decomposed drugs, and 
release of nonpolar volatiles (saturated hydrocarbons, unchlorinated or chlorinated 
olefi ns, alkylbenzenes, and low - molecular - weight polydimethylsiloxanes) from 
rubber closures [117] . Studies have shown that high product surface area is a driving 
force for the adsorption of volatiles released from rubber stoppers (products with 
high surface areas show high levels of turbidity after reconstitution). Attempts to 
impede haze formation include the use of solubilizers and the use of closure materials 
such as bromobutyl closures instead of butyl and chlorobutyl rubber or Tefl on - 
lined rubber stoppers. Reduction of migration of volatiles can be achieved by 
exposure of rubber stoppers to heat or vacuum within the sophisticated washing 
cycles [118] . 
Studies have been carried out on intravenous containers (plastic and glass) and 
plastic administration sets (with and without inline fi lters) on diazepam availability. 
No visual incompatibilities were observed and the pH remained constant. Solutions 
stored in glass bottles and infused through plastic sets retained greater than 90% 
of initial potency after 4 h. Solutions stored in plastic burette chambers and administered 
through plastic sets lost greater than 38% of potency after 2 h. The loss 
occurred principally in the plastic chamber and increased with increasing drug concentration 
and time. Drug availability was not affected by the type of intravenous 

fl uid, pH, fl ow rate, or fi ltration. It was recommended that if diazepam were to be 
infused, it should be diluted to at least 1 : 10 dilution in glass intravenous bottles and 
administered through plastic administration sets not having plastic burette chambers. 
Inline fi lters (0.45 . m) may be used without loss of potency [119] . Similar 
studies were carried out by Morris [120] . It was found that diazepam injection, in 
dilutions of 10 mg/dL or greater, was visually compatible and chemically stable in 
dextrose 5% in water, Ringer ’ s injection, lactated Ringer ’ s injection, and 0.9% 
sodium chloride in glass intravenous bottles when administered within 24 h. However, 
the same type of study carried out in plastic intravenous bags instead of glass bottles 
showed that the solutions had a potency loss of greater than 24% within 30 min after 
dilution. The percentage of diazepam lost from solution increased with increase in 
concentration and time [121] . 
Injectable lyophilized drug products are often presented in vials. The rubber 
stopper forms a critical barrier: Protection of the product against moisture or oxygen 
is strongly dependent upon the quality and functioning of this barrier. In fact, it has 
been reported that rubber not only is a barrier against moisture but also can be a 
source of water [122] . Studies have been carried out to investigate the speed of water 
uptake, the saturation, and fi nally the permeability of different types of rubber closures 
with a view to understanding the fundamental mechanisms which determine 
the barrier quality and obtaining the easy method to discriminate between different 
rubbers. Thirteen - millimeter closures of the same dimensions were used: the Helvoet 
FM 157 (grey) and FM 257 (grey); the Pharma Gummi PH 701/45 (red), PH 21/50 
(red), and PH 4104/40 (grey); and the West Company W888 (grey). All rubbers were 
of the bromobutyl rubber type, with the exception of W888, which consists of a 
halobutyl/ polyisoprene blend. It was found that permeation through the closures 
is highly unpredictable owing to the concentration dependency of the diffusion 
coeffi cient D , which increases strongly with increasing RH. It was indicated that 
new types of rubbers could be evaluated with respect to their barrier properties by 
means of the absorption profi les under stress conditions such as 40 ° C 95% RH. Thus 
one does not require to make a selection on the basis of laborious permeation 
experiments [123] . 
7.3.4.5 Container Extractables and Leachables and Drug Product Stability 
Containers and drug delivery devices for pharmaceutical and biological products 
may release low - molecular - weight compounds into the product. These compounds 
are called extractables and leachables. The toxicological concern is that an extractable 
or leachable compound from the container may affect the effi cacy and biological 
safety of the drug - product. The drug - contacting packaging materials (synthetic 
polymeric formulations) may include antioxidants, colorants, slip agents, and 
plasticizers. 
The subject of extractables and leachables in drug products is an area of active 
discussion in the pharmaceutical industry. Further the regulatory agencies have 
issued guidances on this subject in recent years. The FDA guidance on container 
closure systems defi nes extractables and leachables as follows [124] : Extractables 
are compounds that can be extracted from elastomeric or plastic components of the 
container closure system when in the presence of a solvent. Leachables are compounds 
that leach into the formulation from elastomeric or plastic components of 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 665

666 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
the drug product container closure system. Thus the term extractable refers to the 
removal of a compound from a container closure system under extractive conditions 
by exposure to solvents under nonstandard or atypical product handling conditions 
using a variety of solvents and/or stress conditions that are not used in the manufacturing 
process of the drug product (strong organic solvents or acid/alkaline solutions 
as well as elevated conditions of temperature and pressure). Leachable 
describes the removal of a compound from a container closure system under nonstressful 
solubilizing conditions (exposure of drug substance to surfaces and conditions 
defi ned specifi cally by the manufacturing process) [125] . The difference 
between the two terms is one of process and not related to a fundamental chemical 
property of the compound. An extractable or leachable is generated by partitioning 
of a given compound between two phases: the solid (container closure system) and 
liquid (drug product for leachables or a solvent for extractables). 
It is necessary to test the extractables and/or leachables from each container 
or device followed by toxicological evaluation of the hazard potential. Toxicological 
risk assessment includes evaluating the toxicity of the packaging components 
and the specifi c extractable or leachable substance. Toxicological endpoints include 
single, repeated, and chronic dose toxicity, genotoxicity, carcinogenicity, immune 
sensitization, irritation, and blood compatibility [126] . The concern that the 
regulatory authorities (FDA) have regarding drug product leachables is directly 
related to the particular dosage form and the route of administration [8] . Of the 
highest concern are the inhalation drug products [orally inhaled and nasal 
drug products (OINDPs)] [124, 127, 128] , which include metered - dose inhalers 
(MDIs), dry powder inhalers (DPIs), inhalation solutions and sprays, and nasal 
sprays. The FDA considers leachables to be a safety concern for OINDPs. Leachables 
are also of concern for injections and injectable suspensions, sterile powders 
and powders for injection, ophthalmic solutions and suspensions, and transdermal 
ointments and patches. Other dosage forms such as liquid and solid oral are of less 
concern [8] . 
There are many challenges (mainly scientifi c) for pharmaceutical development 
programs and teams regarding leachables and extractables in terms of analysis and 
questions such as: At what levels are drug product leachables of safety concern? 
What safety qualifi cation processes can be applied to leachables? How do you 
develop control strategies (including specifi cations and acceptance criteria) for 
leachables? Can you control drug product leachables by controlling potential leachables 
(i.e., extratables)? [128] . A report has shown that The Product Quality Research 
Institute (PQRI) Leachables and Extractables Working Group proposed a complete 
pharmaceutical development process for OINDPs related to leachables and extractables 
[129] based on its own comprehensive scientifi c research and on some earlier 
proposals from the International Pharmaceutical Aerosol Consortium for Regulation 
and Science [130] . The PQRI recommendations also discuss selection criteria 
for OINDP container closure components which are designed to exclude potential 
leachables of known safety concern and minimize both the numbers and levels of 
other potential leachables. It is believed that these recommendations move beyond 
quality by testing paradigm and suggest a process by which quality is designed into 
OINDP container closure system components (quality - by - design paradigm) [128] . 
Extractable and leachable testing studies are recommended even if the containers 
or closures meet compendial suitability tests. 

Aside from the toxicological risks posed by leachables and extractables, another 
source of drug product quality problems associated with leachables is stability 
because leachables could react chemically and may interact with drug product. 
Some of the more common chemically reactive leachables include transistion metals, 
radical initiators or propagators, organoperoxides, and reactive nucleophiles or 
electrophiles such as amines and aldehydes [131] . The principal sources of metal 
leachables include glass packages and product - contacting equipment in the manufacturing 
process, especially unpassivated stainless steel equipment. Metal leachables 
can often cause degradation via oxidation in which the metal catalyzes the 
formation of a short - lived peroxy radical, which further reacts with the drug substance 
[132] . Transition metals such as Fe n + and Mn n + are usually involved and the 
oxidation occurs via a Fenton catalytic cycle [133] . Degradation was observed for a 
developmental intravenous dosage form where metal ions leaching out of glass vials 
and stainless steel catalyzed single - electron oxidation. Interestingly, the antioxidants 
in the formulation (thioglycerol, ascorbic acid, or sodium bisulfi te) were involved 
in the Fenton reactions leading to oxidation [134] . For a photosensitive product, the 
use of amber glass vials resulted in other stability problems: The higher levels of 
leachables from amber vials enhanced the metal - catalyzed drug oxidation [135] . 
Metal leachables in liquid formulations may result in the formation of insoluble 
complexes with the pharmaceutical active substance or other formulation ingredients. 
It has been reported that accelerated stability testing at higher temperatures 
might not be indicative of such problems since these precipitates tend to form and 
grow at lower temperatures in a nonlinear fashion [136] . Aluminum ions coming 
from USP type I glass and some plastic packages such as LDPE and rubber closures 
are the most prevalent source for the formation of particulates. The Al 3+ leaches out 
and accumulates in solution in levels from 45 ppb up to 6 ppm, depending on factors 
such as the presence of buffers, the solution pH, and autoclaving [137] . 
Organoperoxide radicals can be present in plastic packaging materials as a result 
of residual free radicals from dissociation of the peroxides formed during polymerization 
or melt processing of the plastic or from the long - lived radicals formed in 
the polymer as a result of gamma - induced radiolysis of the polymer chain. Studies 
on the migration of radiolysis products from plastics sterilized by gamma radiation 
or electron beam showed that smaller molecular weight fragments are formed by 
radiolysis of the plastic additives and residual oligomers [138, 139] . Very reactive 
species such as aldehydes have been reported to react directly with drug products 
possessing a nucleophilic site such as a primary or a secondary amine. Formaldehyde 
can be formed in minute amounts as a by - product of longer chain alcohol oxidation 
and from other sources such as acrylates. A sterile formulation containing an antistroke 
developmental product packaged in glass vials with rubber stoppers exhibited 
a formadehyde - adduct degradate in levels above 2% after 13 weeks at 30 ° C. 
The formaldehyde source was found to be in the rubber stopper (formed from a 
reinforcing agent used in the stopper) [140] . Strategies to mitigate the stability risk 
of leachables include a careful examination of mode of sterilization and packaging 
selection. Use of silica - coated glass vials, use of nitrogen head space in the package, 
use of plastics, and inclusion of a chelating agents will reduce the adverse effects of 
leachables on stability. 
Protein aggregation has recently attracted a lot of attention from the FDA due 
to a number of incidents in which stability problems were caused by container 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 667

668 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
leachables/extractables or changes in excipients, as suggested in the case of Eprex 
(epoetin alfa, Johnson & Johnson, New Brunswisk, NJ). Since aggregation can 
induce immunogenicity reactions with potentially severe consequences, the FDA is 
requ iring manufactures to pay more attention [141] . Evaluation of packaging materials 
for potential extractables and leachables is critical to guarantee the integrity 
of the drug product and assure compliance with the Code of Federal Regulations 
(CFR), Title 21, Part 211.65, which states that equipment should be constructed so 
that contact surfaces that contact components, in - process materials, or drug products 
should not be reactive, additive, or absorptive so as to alter the safety, quality, or 
purity of the drug product beyond offi cial or other established requirements [142] . 
The U.S. Food, Drug and Cosmetic (FD & C) Act also states that a drug device should 
be deemed to be adulterated if its container is composed in whole or in part of any 
poisonous or deleterious substances which may render the components injurious to 
health [143] . A review of regulatory and scientifi c considerations on testing for 
extractables and leachables has been provided [144] . To facilitate the evaluation of 
extractables (to detect, identify, and quantify organic extractables) mass spectometry 
was used recently. Using the Agilent G1888 Network Headspace/680N GC/5975 
inert Mass Selective Detector (MSD) system, several primary packaging materials 
for pharmaceuticals were evaluated. Data on two of the materials (HDPE bottle 
and a soft elastomer liner from the inside of the screw cap) were presented. A multiple 
head space extraction technique was used and the highest attainable amount 
of extractables that could ever be concentrated in the drug product was calculated. 
It was found that the analytical results obtained from the inert 5975 head space gas 
chromatogrophy/mass spectrometry (GC/MS) provided excellent sensitivity [145] . 
The effect of ammonium sulfate treatment on cerium oxide glass vials was 
assessed following exposure to ionizing radiation. The bulk chemical composition 
of irradiated cerium oxide glass remains unchanged despite a temporary browning 
effect. Stability against alkali leachables of the internal silica matrix was enhanced 
with ammonium sulfate treatment. With the exception of alumina (Al 2 O 3 and Na 2 O), 
irradiation sterilization has a limited effect on altering the surface chemistry of 
ammonium sulfate – treated cerium oxide glass [146] . 
7.3.4.6 Biotechnological Products 
Despite the efforts at delivering products of biotechnology (peptides and proteins) 
through novel drug delivery systems, injection still remains the main mode of delivery. 
The unit dose for many injectable biotechnological products is the single - dose 
vial and occasionally prefi lled syringes. The product is often provided either as a 
solution or as a lypholized cake to be reconstituted and injected using syringes. 
Packaging represents the fi rst line of defense for all formulated drug products, protecting 
the product from the outside world and vice versa. At the same time, the 
package must be fully compatible with the product [147] . Thus the requirements for 
product purity, activity, and shelf life dictate a high standard for injectable drug 
packaging. 
Packaging conditions have effects on the stability of protein products. Peptide 
and protein drugs are high - molecular - weight compounds with unique physicochemical 
properties. They are extremely sensitive to their microenvironment: heat, light, 
pH, chemical contaminants, and so on. Trace amounts of metals, plasticizers, and 

other materials from packaging may deactivate or denature therapeutic peptide and 
proteins. Further, peptides and proteins have a tendency to adsorb on to the surface 
of container closure systems, thereby removing virtually all active material from the 
drug formulation. Even when the drug desorbs back into the solution, the interaction 
could result in loss of potency. Lyophilized biopharmaceutical products can be 
affected by moisture if the seal is not adequate to prevent ingress of moisture into 
the container. Thus packaging can make and break fi nal formulation of lyophilized 
product. Vials that are not designed specifi cally for lyophilization (with convex 
rather than fl at bottoms) can make the lyophilization process less effi cient. Rubber 
closures can also hinder freeze drying if they do not permit adequate venting during 
sublimation. Stopper rubbers adsorb and desorb moisture at different rates and 
under storage conditions stoppers that were not properly dehydrated can release 
water into lyophilized product. 
Another source of instability of biopharmaceutical products traceable to packaging 
is silicone oil, which is commonly used to lubricate elastomeric stoppers during 
fi ll/fi nish to facilitate insertion of the stopper into the vial. Silicone oil is known to 
inactivate protein through nucleation of protein around oil droplets. The problem 
is being circumvented by using a fl uoroelastomer coating on stoppers to provide the 
needed lubricity in addition to chemical innertness, barrier protection, and safety 
[148] . The fl uoroelastomer fi lms reduce adsorption of drug on to the stopper, provide 
lubricity for proper vial sealing, and also reduce the possibility of extractables 
migrating from the rubber stopper into the product. The prefi lled syringes are 
becoming very popular in the injectable markets. Some of the challenges to overcome 
are the compatibility and stability issues that arise when dealing with biotechnology 
formulations. Biotechnological products can react with the oily form of 
silicone, which is used as a lubricant to coat the sliding components of the syringe. 
It is believed that the propensity for silicone to react with the formulation is a function 
of the concentration of the silicone in the syringe and its chemical activity. The 
chemical activity is determined by the number of terminal hydroxyl groups, which 
is greater the shorter the silicone polymer chain length. It has been reported that 
baking on the silicone (which involves heating the silicone - coated syringe to a specifi 
c temperature for an appropriate time) results in longer chains that are more 
closely adhered to the surfaces they coat. Thus the concentration of silicone in the 
syringe and its chemical reactivity are both reduced and the product ’ s stability is 
increased [149] . Another benefi t of baked - on silicone is that it reduces the frequency 
of the “ break - loose ” effect, which occurs during storage when the rubber closure 
inside the syringe barrel expands outward so that eventually it displaces the 
low - friction silicone coating and comes into direct contact with the inner glass 
surface. Another challenge is the prevention of undesirable pH change that sometimes 
occurs in liquids stored in prefi lled syringes. The shift in pH occurs because 
the USP type I glass used in prefi lled syringe manufacture is a borosilicate which 
must be subjected to various temperature changes during the glass tube production 
process. During storage, sodium ions are released into the product and increase the 
concentration of hydroxide ion. This problem is solved by spraying ammonium 
sulfate into the glass barrel before the tempering process in the formation of 
syringes begins. 
Factor VIII (FVIII) is an essential coagulation factor in the blood which serves 
as a cofactor in the complex blood - clotting cascade. A defi ciency in FVIII is the 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 669

670 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
cause of hemophilia (type A), a hereditary life - treatning bleeding disorder [150] . 
The effect of the container on FVIII stability was shown to be temperature dependent. 
At 4 – 8 ° C, 2 of the 15 constituted FVIII concentrates showed better stability 
in plastic containers and 2 concentrates showed better stability in glass containers; 
at 20 – 23 ° C most concentrates showed better stability in plastic contaners; and at 
37 ° C all concentrates showed equal or better stability in plastic containers [151] . It 
was concluded that stability studies comparing different types of containers should 
be conducted at the product storage temperature. Due to limited diffusion of molecules 
in a solid state, containers would not be expected to play a critical role in 
storing lyophilized FVIII products unless the air permeation through the container 
and/or container stoppers is signifi cantly different. This is because air has been 
shown to accelerate the inactivation of recombinant FVIII SQ not only in solution 
but also in a lyophilized state [152] . 
It is known that interaction of proteins with the surfaces of their containers is a 
potentially signifi cant problem in biotechnology. The amphiphatic nature of protein 
molecules results in their adsorption to a wide variety of surfaces and can result in 
both their loss and destabilization. Reports have shown that when adsorption of a 
protein/peptide drug to container occurs, the drug molecule exchanges solution 
interactions for surface interactions where the free energy of exchange is negative 
[153] . Similarly, studies have shown that some surfactants are able to reduce/eliminate 
protein/peptide drug adsorption to glass and PP. In the presence of surfactants, 
where the surfactant – surface interaction is greater than the surface – protein/peptide 
interaction, drug adsorption is reduced or eliminated. For protein/peptide adsorption 
onto glass, where an electrostatic interaction (interaction between the positively 
charged peptide/protein and the negatively charged glass surface) predominates, 
only the most hydrophobic surfactants (polysorbate 20 and benzakonium chloride) 
were signifi cantly effective to reduce adsorption to surfaces as demonstrated with 
salmon calcitonin and bovine serum albumin (BSA). The nonionic surfactant polysorbate 
20 proved to be the most effective in eliminating adsorption to both plastic 
and glass surfaces. The predominat mechanism of protein/peptide adsorptrion to PP 
(plastic surface) was by a hydrophobic/dehydration mechanism [154] . The effect of 
Poloxamer 407 (Pluronic F - 127), a nonionic surfactant, on the adsorption of granulocyte 
colony - stimulating factor (G - CSF) to PVC was assessed. It was found that 
Poloxamer 407 at a concentration of 0.05% w/w may show promise as a solvent 
additive with which to minimize G - CSF adsorption to PVC [155] . 
The amount of surface adsorption of a number of proteins ranging in molecular 
mass from 6.5 to 670 kDa and isoelectric point (pI) from 4.3 to 10.5 to several commonly 
used container surfaces (glass vials: either untreated, siliconized, sulfur 
treated or Purcoat treated; plastic vials: polyester + 0.3%, polyester 5 . 0, PP, and 
nylon). A 5 - mL volume of protein solution was added to each vial, yielding a 
surface - to - volume ratio of 2.4 cm 2 /mL. No correlation was found between the 
amount adsorbed and the molecular mass or isoelectric point, although glass surfaces 
appeared to bind more protein under the experimental conditions examined 
[156] . 
Patients receiving total parenteral nutrient solutions frequently require exogenous 
insulin to fully use the administered glucose. Consequently a study was conducted 
to determine the percentage of insulin adsorbed to the glass and PVC when 
added to a nutrient solution infusion system. The following parameters on insulin 

availability from parenteral nutrient solutions were investigated: sample time for 
infusion, insulin concentration, amino acid or polypeptide source, electrolytes and 
vitamins, inline fi lters, glass and PVC infusion containers, and human albumin. 
Results showed that basic solutions of amino acids and protein hydrolysates in 
dextrose with 30 units of insulin failed to deliver approximately 44 – 47% of the 
added insulin. Varying the concentration of insulin had a small but statistically signifi 
cant effect on the degree of insulin loss. The use of inline fi lters and PVC bags 
caused an even greater loss of insulin. The addition of albumin or electrolytes and 
vitamins decreased the insulin loss [157] . 
7.3.4.7 A Look into the Future of the Effects of Packaging on Stability of 
Drug Products 
New Drug Delivery Systems Reports have shown that the advent of new drug 
delivery systems for various routes of drug administration (e.g., oral, nasal, pulmonary, 
transdermal, needle free) and the development of new biotechnological drugs 
have resulted in the need for enhanced protection against such factors as moisture, 
light, oxygen, and mechanical forces as well as making packaging play a more integral 
role in drug delivery [158] . Novel drug delivery systems have necessitated the 
emergence of specialized packaging needs because not all of them can be packaged 
in bottles or standard blisters. More often than not, the packaging must be unit dose 
and must also become an integral part of the drug delivery technology, as drug 
product packaging and packaging design contribute signifi cantly to the stability, 
shelf life, and performance of the drug and the drug delivery system. In fact, the old 
equation showing the relationship between a new chemical entity (NCE), drug 
delivery (DD), and drug product (DP) (i.e., NCE + DD = DP) is now being replaced 
by NCE + DD + packaging = DP. Table 1 gives the considerations needed when 
designing packaging for new drug delivery systems [158] . 
One of the stability problems associated with transdermal drug delivery devices 
is the degradation of the contents of the devices: drugs, permeation enhancers, 
matrix materials, and other components incorporated in the devices. Degradation 
undesirably breaks down the components as well as causes discoloration and formation 
of odors within the pouched system. Devices that are susceptible to degradation 
cannot be stored for a reasonable amount of time, thus causing practical problems 
in their distribution. A solution to this problem involves the incorporation of an 
antioxidant such as BHT into the drug formulation of the transdermal drug delivery 
device [159] . Also a desiccant has been incorporated within the sealed pouch of the 
transdermal drug delivery device. For example, the Climara transdermal estradiol 
system is packaged and sold within a sealed pouch containing a water scavenger to 
protect against hydrolysis of estradiol [160] . 
The effects of extreme temperatures on drug delivery of two albuterol sulfate 
hydrofl uoroalkane (HFA) MDIs were evaluated. Three Proventil HFA and three 
Ventolin HFA MDIs were stored at room temperature and served as controls, while 
three of each product were placed in the trunk of a vehicle in Tuscon, Arizona. The 
temperature of the vehicle was monitored for six months. Product performance for 
each of the MDIs was evaluated at room temperature. An additional study was 
performed to investigate the performance of the two products when actuated at 4, 
22, 47, and 60 ° C. Within one week of being placed in the test vehicle, all MDI 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 671

672 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
TABLE 1 Packaging Needs and Considerations for New Drug Delivery Technologies 
Delivery 
Route 
Company Names 
and Systems Characteristics Nature 
Packaging Needs and 
Description 
Oral CIMA, OralSolv 
RP Scherer, Zydis 
Biovail, Flashdose 
EthyPharm, 
Flashtab 
Yamanouchi, 
wowTab 
Elan, FASTMelt 
Eurand, Ziplets 
Fast - dissolving 
tablets or 
orally 
disintegrating 
dosage Forms 
Hygroscopic 
(polysaccharide 
or protein - 
based, taste 
masked, fragile 
Protection from 
mechanical forces: 
stiff packaging 
Peelable opening 
Ultrahigh moisture 
barrier, e.g., Aclar 
or foil to maintain 
shelf life stability 
for 2 – 3 years 
Pulmonary Inhale/Inhance TM 
Aradigm/AERx ® 
Alkermes/AIR ® 
BatellePharma 
Therapeutics/ 
EHD pulmonary 
delivery 
Inhalable 
delivery deep 
into lungs, 
either powder 
or liquid 
Can be fi ne 
powder with 
high surface 
area or 
aqueous/ 
nonaqueous 
formulation 
Medium – high barrier 
protection 
Mechanical needs 
depend on delivery 
device design 
Sterile barrier 
Chemical inertness, 
especially for high 
liquid doses 
Clear barrier offers 
QA check for both 
producer and end 
user 
Transdermal 3M/Latitude TM 
Noven/DOT 
MatrixTM 
Alza/D - Trans ® 
Sustained 
delivery/ 
absorption 
through skin 
Depending on 
drug/formulation 
Protection of active 
from environmental 
aggrevates 
Clear barrier offers 
aesthetics 
Pouching as primary/ 
secondary package 
Barrier property on 
carrier web critical 
Chemical inertness 
Transmucosal 
delivery 
Cephalon, Actiq 
Atrix Labs, BEMA 
CIMA Labs, 
Dravescent 
Sustained 
delivery via 
mucosal layer 
Moisture 
senstivity 
Ultrahigh barrier 
moisture protection 
Clear material offers 
QA check 
Aesthetics 
canisters developed physical distortions. The Proventil HFA MDIs bulged at the 
base of the canister, while the Ventolin HFA MDIs bulged around the valve. These 
distortions were thought to be the result of an increase in vapor pressure brought 
about by the initial high temperatures. After exposure to extreme environmental 
temperatures, an increase in propellant - leak rate was observed with Proventil HFA 
and Ventolin HFA MDIs, but little to no change was found in particle size, dose per 

actuation, respirable mass, and nonrespirable mass. Despite the two formulations ’ 
tolerance to extreme temperatures, drug delivery was affected when the MDIs were 
tested at specifi c temperatures outside of the recommended usage [161] . One of the 
alternatives to the pressurized metered - dose inhaler (pMDI) is the breath - actuated 
DPI. Stable and predictable therapeutic responses require a consistent dose delivery 
from an inhaler throughout its life and consistency of doses from one inhaler to 
another. Recognizing this, specifi cations for inhaler dose uniformity have been 
defi ned by regulatory agencies, including the European Pharmacopoeia (EP) [162] 
and the FDA [163] . DPI inhaler design, especially the geometry of the mouth piece, 
is critical for patients to produce an airfl ow suffi cient to lift the drug from the dose 
chamber, break up the agglomerates in a turbulent air stream, and deliver a dose 
to the lungs as therapeutically effective fi ne particles [164, 165] ). The airfl ow generated 
by inhalation directly determines particle velocity and hence the ease with 
which particles are deagglomerated. The materials used in the construction of DPIs 
and the characteristics of the formulation affect electrostatic charge accumulation. 
Some formulations as well as inhaler materials accumulate and retain electrostatic 
charge more strongly than others, and this will affect both drug retention within 
these inhalers as well as delivered aerosol behavior [166] . 
Pulmozyme ® , recombinant deoxyrucleic acid (DNA) – derived human DNase I 
(rhDNase) has been formulated for local delivery to the lung by inhalation. rhDNase 
was fi lled originally in glass vials for clinical studies. A direct stability comparison 
study was made between rhDNase in glass vials and rhDNase fi lled into plastic 
ampoules using the blow - fi ll seal technology of Alp, which uses low - density polyethylene 
resins which are gas permeable. The problem of gas permeation was 
addressed by packing the plastic ampoules in a gas - impermeable foil pouch which 
can be fi lled with nitrogen. The rate of deamindation was similar for rhDNase stored 
at 2 – 8 ° C in foiled and unfoiled ampoules but substantially different from protein 
stored in glass vials. This difference in deamidation rate was attributed to the 0.5 - 
unit difference in pH of rhDNase in glass (due to leaching of ions, possibly sodium, 
from the glass surface) versus plastic ampoules. This result provided the rationale 
for the choice of container closure component for rhDNase aerosols [167] . 
Improvements in the design of pMDIs has facilitated the extension of pMDIs to 
administering macromolecules such as peptides and proteins. The belief is that the 
HFA environment within the pMDI is inert and essentially moisture free, thereby 
providing good stability for macromolecules. Further, advances in valve and actuator 
technology in the pMDI help ensure a consistent and effi cient delivery, an 
important consideration based on the cost and potency of biotechnological drug 
substances [168] . The 3M company has made improvement in container closure 
systems (CCSs) for biopharmaceuticals. One improvement reduces the potentially 
problematic deposition of the active drug within the CCS by introducing can coating. 
Further 3M is using a novel semipermeable membrane system component to 
improve the dosing reproducibility of suspension formulations [169] . The company 
is also developing new fast - fi ll, fast - empty valve designs to improve dose - to - dose 
uniformity as well as to avoid the need for patients to prime the system if only used 
intermittently [170 – 172] . 
Methods for Prediction of Effects of Packaging on Stability of Drug Products 
Drug stability studies often involve multiple batches to ensure that a product will 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 673

674 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
consistently remain within specifi cations for its entire expiration date. The studies 
usually involve the same drug products in similar packages or in multiple strengths. 
The belief now is that application of sound statistical design principles can reduce 
the amount of testing required. The principles stated in the FDA 1987 publication 
Guideline for Submitting Documentation for the Stability of Human Drugs and 
Biologics have been extended to setting expiration dating periods for more complex 
situations [173] . PP may be used as a substitute for glass in primary packaging for 
various drug products. However, the diffusion of water through the PP plastic wall 
often builds up the water content with time during the long - term storage. It is often 
the shelf life limiting parameter for chemically stable aqueous solutions in PP 
bottles. The water diffusion rates of the widely distributed X - ray contrast agent 
VisipaquesTM in PP bottles and magnetic resonance contrast agent OmniscanTM 
in PP syringes were evaluated with the objective of developing a mathematical 
method of estimating the rate constant for water diffusion through the PP wall as 
a function of the following variables: temperature, humidity, surface area, wall thickness, 
concentration of the active ingredient, and fi ll volume. The effect of the variables 
was estimated by partial least - squares regression. The predictive ability of the 
cross - validated models was good for the two active agents. The models were used 
to predict the shelf life for the relevant combinations of temperature and humidity 
for the four climatic zones. The method presents an opportunity to measure the 
effect of the variables infl uencing drug stability [174] . 
A multivariate model has been developed to identify factors infl uencing stability, 
to estimate shelf life, to select new batches for further stability testing, and to evaluate 
changes in new batches. The model was capable of predicting the degradation 
rate constant as a function of storage temperature, pH, product concentration, and 
container volume [175] . Other investigators have reported on the applications of 
statistical analysis in stability studies of drug products. The time dependency of 
product change has been examined by ordinary least - squares regression and variance 
analysis for testing the possibility of pooling batches to increase the precision 
in shelf life estimates [176] . The Bayesian approach has been used to quantify the 
uncertainty in predicting shelf life from the Arrhenius equation as a function of 
stability consisting of various error distributions [177] . Monte Carlo simulation has 
been used to test the power of the analysis of variance (ANOVA) on matrixing data 
and single data to estimate shelf life. The simulation showed that the large amount 
of data from a matrixing design gave more precise shelf life estimates [178] . 
Desiccant is frequently included in the packaging of moisture - sensitive products 
in order to maintain low relative humidity inside the package and hence protect 
the product from moisture. Sorption – desorption moisture transfer (SDMT) models 
have been successfully applied to predict moisture transfer between a solid product 
and a desiccant inside a closed package. Further, theoretical simulations extended 
the use of the SDMT model to take into account the moisture permeation properties 
of the package [179] . The SDMT model was also used to predict the effect of 
desiccant quantity, tablet quantity, and tablet initial moisture content on the relative 
humidity inside high - density polyethylene (HDPE) bottles containing a moisture - 
sensitive drug product, roxifi ban tablets. Desiccant quantity, tablet weight, and initial 
moisture content before packaging were found to have an effect on stability. The 
use of theoretical calculations utilizing the SDMT model was shown to be useful in 
the understanding of packaging requirements for the product. Calculated relative 

humidity data corroborated experimental fi ndings regarding the effect of the variables 
studied on tablet stability [180] . It is known that hard gelatin capsule brittleness 
is a function of moisture content. A study has been carried out to indicate that 
the brittleness of empty capsules occurred at humidities below 40%. Mexitil - loaded 
capsules exhibited a similar profi le. The SDMT model was employed to estimate 
the fi nal relative humidity for the Mexitil/gelatin capsule system and results were 
presented to demonstrate the general applicability of the SDMT model for predicting 
the incidence of brittleness problems and for the formulation to ensure the 
absence of brittleness [181] . 
The sorption of two weak acids (warfarin and thiopentone) and two weak bases 
(chlorpromazine and diltiazem) into PVC infusion bags was described by a constant 
partition model. PVC – water partition coeffi cients were obtained using three different 
methods: equilibrium values for sorption into PVC bags, the sorption versus pH 
relationship, and partition into PVC strips. The data were compared with similar 
values derived from a liquid – liquid partition system and different organic solvents 
(octanol, dichloromethane, carbon tetrachloride, and hexane). Octanol is the preferred 
solvent, and it is suggested that octanol – water partition data can be used to 
predict sorption behavior [182] . 
Reports have shown that changes in the hardness of tablets composed of lactose 
and cornstarch were due to variation of the moisture content. Further it was shown 
that the hardness of the tablets in moisture - semipermeable packages (such as strip 
packs and press - through packs with or without an overwrap fi lm) could be predicted 
by an iterative calculation procedure through a mathematical model based on the 
physicochemical properties of the tablets and the moisture permeabilities of the 
packaging materials [183] . In another study by the same research group, moisture 
and temperature were used to predict the shelf - life of packaged tablets. Cognizance 
was taken of the fl unctuations of temperature and relative humidity during prolonged 
storage. A sugar - coated tablet with a core containing ascorbic acid whose 
color is affected by both moisture content and ambient temperature was investigated. 
It was found that the color change of ascorbic acid core of the sugar - coated 
tablet was dependent on its moisture content and the ambient temperature and that 
the color change of the tablet in moisture - semipermeable packages could be predicted 
by an iterative calculation procedure using a mathematical model based on 
the kinetics of the color change and the moisture permeabilities of the packaging 
materials [184] . 
It is believed that a container/formulation couple is compatible if the magnitude 
of ingredient loss is within acceptable limits over the entire shelf life of the product. 
In this connection an approach for assessing the interaction of drug with PVC plastic 
infusion bags was developed. The approach correlates the partition coeffi cients and 
dissociation constant (when appropriate) of the solute, the physical dimensions of 
the container, and the solution pH with single parameters that dictate the shape 
of the sorption profi le. To determine the equilibrium sorption level of PVC containers, 
the fractional binding of a solute was correlated with its hexane – water and 
octanol – water partition coeffi cients. Calculations based on single partition coeffi - 
cients are believed to be less effective in terms of mimicking the behavior of PVC. 
To determine the sorption profi le (fractional binding with time), the partition coef- 
fi cients are related to the fractional binding at a particular time through a single 
parameter referred to as the sorption number. Equilibrium fractional binding and 
EFFECT OF PACKAGING ON DRUG PRODUCT STABILITY 675

676 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
sorption profi les for various drugs stored in PVC containers are generated with the 
models and agree well with reported behavior [185] . The relationship between sorption 
number (a parameter defi ning initial solute uptake by PVC infusion bags) and 
solute octanol – water partition coeffi cient was investigated by following the time 
course of the sorption of drugs by PVC infusion bags which was approximated using 
a diffusion model in which the plastic was assumed to act as an infi nite sink. The 
model was found to be suitable for estimation of storage times relevant to clinical 
usage and enabled the magnitude of the uptake in a specifi c time to be described 
by a single parameter referred to as the sorption number. This parameter was 
defi ned by the plastic infusion solution partition coeffi cient, the diffusion coeffi cient 
in the plastic, the fraction un - ionized in the solution, the volume of the infusion 
solution, and the surface area of the plastic. The sorption number can be extrapolated 
to allow prediction of the effects of time, plastic surface area, solution volume, 
and solution pH on fractional solute loss. A reasonable correlation was established 
between the logarithm of this parameter and the logarithm of the octanol – water 
partition coeffi cient of various solutes. The model allows the fraction of a solute 
remaining in a plastic infusion bag at a given storage time to be estimated from the 
octanol – water partition coeffi cient of the solute and other readily available data 
[186] . 
REFERENCES 
1. U.S. Pharmacopeia/National Formulary ( 2005 ), The Offi cial Compendia of Standards , 
USP 28/NF23, U.S. Pharmacopeial Convention , Rockuille, Maryland . 
2. International Conference on Harmonization (ICH) Steering Committee ( 1994 ), ICH 
harmonized tripartite guideline, Stability testing of new drug substances and products, 
ICH, Geneva. 
3. International Conference on Harmonization (ICH) Steering Committee ( 1998 ), Draft 
ICH harmonized tripartite guideline, Stability testing of new drug substances and products, 
ICH, Geneva. 
4. Rhodes , C. T. ( 2000 ), Reasons for stability testing , in Carstensen , J. T. , and Rhodes , C.T. 
Eds., Drug Stability, Principles and Practices , Marcel Dekker , New York , p. 11 . 
5. Center for Drugs and Biologics, U.S. Food and Drug Administration (FDA), Department 
of Health and Human Services ( 1987 ), Guideline for submitting documentation 
for the stability of human drugs and biologics, CDB, Washington, DC. 
6. Yoshioka , S. , and Stella , V. J. ( 2000 ), Stability of Drugs and Dosage Forms , Kluwer 
Academic , New York . 
7. Guillory , J. K. , and Poust , R. I. ( 1996 ), Chemical kinetics and drug stability , in Banker , 
G. S. , and Rhodes , C. T. , Eds., Modern Pharmaceutics , 3rd ed., Marcel Dekker , New York , 
p. 179 . 
8. Center for Drug Evaluation and Research and Center for Biologics Evaluation and 
Research, U.S. Food and Drug Administration (FDA), Department of Health and 
Human Services ( 1999 ), Guidance for industry: Container closure systems for packaging 
human drugs and biologics, FDA, Washington, DC. 
9. Florence , A. T. , and Attwood , D. ( 2006 ), Physicochemical Principles of Pharmacy , 
4th ed., Pharmaceutical Press , London , p. 94 . 
10. Cartensen , J. T. , Serenson , E. G. , and Vance , J. J. ( 1964 ), Use of Hammett graphs in 
stability programs , J. Pharm. Sci ., 53 , 1547 – 1548 . 

11. Moorhatch , P. , and Ciou , W. L. ( 1974 ), Interactions between drugs and plastic intravenous 
fl uid bags. II. Leaching of chemicals from bags containing various solvent media , 
Am. J. Hosp. Pharm ., 31 , 149 – 152 . 
12. Venkataramanan , R. , Burckart , G. J. , Ptachcinski , R. J. , Blaha , R. , Logue , L. W. , Bahnson , 
A. , Giam , C. , and Brady , J. E. ( 1986 ), Leaching of diethyl phthalate from polyvinyl chloride 
bags into intravenous cyclosporine solution , Am. J. Hosp. Pharm ., 43 , 2800 – 2802 . 
13. Scheiffer , G. W. , Palermo , P. J. , and Pollard - Walker , S. ( 1984 ), Simultaneous determination 
of methyl, ethyl, propyl, and butyl 4 - hydroxybenzoates and 4 - hydroxybenzoic acid 
in liquid antacid formulations by gas chromatography , J. Pharm. Sci ., 73 , 128 . 
14. Ullmann , E. , Thoma , K. , and Zelfel , G. ( 1963 ), The stability of sodium penicillin G in 
the presence of ionic surfactants, organic gel formers, and preservatives , Pharm. Acta. 
Helv ., 38 , 577 – 586 . 
15. Testa , B. , and Etter , J. C. ( 1975 ), Hydrolysis of pilocarpine in carbopol hydrogels , Can. 
J. Pharm. Sci ., 10 , 16 – 20 . 
16. Amidon , G. L. , and Middleton , K. R. , ( 1988 ), Accelerated physical stability testing and 
long - term predictions of changes in the crushing strength of tablets stored in blister 
packages , Int. J. Pharm ., 45 , 79 – 89 . 
17. Taborsky - Urdinola , C. J. , Gray , V. A. , and Grady , L. T. ( 1981 ), Effects of packaging and 
storage on the dissolution of model prednisone tablets , Am. J. Hosp. Pharm ., 38 , 
1322 – 1327 . 
18. Gouda , M. W. , Moustafa , M. A. , and Molokhia , A. M. ( 1980 ), Effect of storage conditions 
on erythromycin tablets marketed in Saudi Arabia , Int. J. Pharm ., 5 , 345 – 347 . 
19. Barrett , D. , and Fell , J. T. ( 1975 ), Effect of aging on physical properties of phenylbutazone 
tablets , J. Pharm. Sci ., 64 , 335 – 337 . 
20. Khalil , S. A. , Ali , L. M. , and Abdel - Khalek , M. M. ( 1974 ), Effects of aging and relative 
humidity on drug release . Pharmazie , 29 , 36 – 37 . 
21. Georgarakis , M. , Htzipantou , P. , and Kountourelis , J. E. ( 1988 ), Effect of particle size, 
content in lubricant, mixing time, and storage relative humidity on drug release from 
hard gelatin ampicillin capsules , Drug Dev. Ind. Pharm ., 14 , 915 – 923 . 
22. Dahl , T. C. , Sue , I. T. , and Yum , A. ( 1991 ), The effect of pancreatin on the dissolution 
performance of gelatin - coated tablets exposed to high - humidity conditions , Pharm. Res ., 
8 , 412 – 414 . 
23. Kornblum , S. S. , and Sciarrone , B. J. ( 1964 ), Decarboxylation of p - aminosalicylic acid in 
the solid state , J. Pharm. Sci ., 53 , 935 . 
24. Vromans , H. , and van Laarhoven , J. A. H. ( 1992 ), A study on water permeation through 
rubber closures of injection vials , Int. J. Pharm ., 79 ( 1 – 3 ), 301 – 308 . 
25. Lamy - Freud , M. T. , Ferreira , V. F. N. , Faljoni - Alario , A. , and Schreiter , S. ( 1993 ), Effect 
of aggregation on the kinetic of autoxidation of the polyene antibiotic amphotericin B , 
J. Pharm. Sci ., 82 , 162 – 166 . 
26. Smith , G. B. , DiMichele , L. , Colwell , L. F. , et al . ( 1993 ), Autooxidation of simvastin , 
Tetrahedron , 49 , 4447 – 4462 . 
27. Khan , M. M. T. , and Martel , A. E. ( 1967 ), Metal ion and metal chelate catalyzed oxidation 
of ascorbic acid by molecular oxygen. I. Cupric and ferric ion catalyzed oxidation , 
J. Am. Chem. Soc ., 89 , 4176 – 4185 . 
28. Jensen , J. , Cornett , C. , Olsen , C. E. , Tjornelund , J. , and Hansen , S. H. ( 1992 ), Identifi cation 
of major degradation products of 5 - aminosalicylic acid form in aqueous solutions 
and in pharmaceuticals , Int. J. Pharm ., 88 , 177 – 187 . 
29. Yeh , S. , and Lach , J. L. ( 1961 ), Stability of morphine in aqueous solution III. Kinetics of 
morphine degradation in aqueous solution , J. Pharm Sci ., 50 , 35 – 42 . 
REFERENCES 677

678 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
30. Pitman , I. H. , Higuchi , T. , Alton , M. , and Wiley , R. ( 1972 ), Deutrium isotope effects on 
degradation of hydrocortisone in aqueous solution . J. Pharm. Sci ., 61 , 818 – 820 . 
31. Boccardi , G. , Deleuze , C. , Gachon , M. , Palmisano , G. , and Vergnaud , J. P. ( 1992 ), Autoxidation 
of tetrazepam in tablets: Prediction of degradation impurities from oxidative 
behavior in solution , J. Pharm. Sci ., 81 , 183 – 185 . 
32. Greehill , J. V. , and McLelland , M. A. ( 1990 ), Photodecomposition of drugs , Prog. Med. 
Chem ., 27 , 51 – 121 . 
33. International Conference on Harmonization (ICH) ( 1996 ), Harmonized tripartite 
guideline: on stability testing: Photostability testing of new drug substances and products, 
Q1B, 1996 ICH, Geneva. 
34. Matsuda , Y. , Inouye , H. , and Nakanishi , R. ( 1978 ), Stabilization of sulfi somidine tablets 
by use of fi lm coating containing UV absorber: Protection of coloration and photolytic 
degradation from exaggerated light , J. Pharm. Sci ., 67 , 196 – 201 . 
35. Frank , M. J. , Johnson , J. B. , and Rubin , S. H. ( 1976 ), Spectrophotometric determination 
for sodium nitroprusside and its photodegradation products , J. Pharm. Sci ., 65 , 44 . 
36. Asker , A. F. , and Larose , M. ( 1987 ), Infl uence of uric acid on photostability of sulfathiazole 
sodium solutions , Drug Dev. Ind. Pharm ., 13 , 2239 . 
37. Asker , A. F. , Canady , D. , and Cobb , C. ( 1985 ), Infl uence of DL - methionine on the photostability 
of ascorbic acid solutions , Drug Dev. Ind. Pharm ., 11 , 2109 . 
38. Vandenbossche , G. M. , deMuynk , C. , Colardyn , F. , and Remon , J. P. ( 1993 ), Light stability 
of molsidomine in infusion fl uids , J. Pharm. Pharmacol ., 45 , 486 . 
39. Akimoto , K. , Kurosaka , K. , Nakagawa , H. , and Sugimoto , I. ( 1988 ), A new approach to 
evaluating photo - stability of nifedipine and its derivatives in solution by actinometry , 
Chem Pharm. Bull ., 36 , 1483 – 1490 . 
40. Thoma , K. , and Klimek , R. ( 1991 ), Photostabilization of drugs in dosage forms without 
protection from packaging materials , Int. J. Pharm ., 67 , 169 – 175 . 
41. Matsuda , Y. , and Minamida , Y. ( 1976 ), Stability of solid dosage forms: II. Coloration and 
photolytic degradation of sulfi somide tablets by exaggerated ultraviolet irradiation , 
Chem. Pharm. Bull ., 24 , 2229 – 2236 . 
42. Murthy , K. S. , Reisch , R. G. , and Fawzi , M. B. ( 1989 ), Dissolution stability of hard - shell 
capsule products: Part I: The effects of exaggerated storage conditions , Pharm. Technol ., 
13 ( 3 ), 72 – 86 . 
43. Teraoka , R. , Konishi , Y. , and Matsuda , Y. ( 2001 ), Photochemical and oxidative degradation 
of the solid - state tretinoin tocoferil , Chem. Pharm. Bull ., 49 ( 4 ), 368 – 372 . 
44. Matsuda , Y. , Itooka , T. , and Mitsuhashi , Y. ( 1980 ), Photostability of indomethacin 
in model gelatin capsules: effects of fi lm thickness and concentration of titanium 
dioxide on the coloration and photolytic degradation , Chem. Pharm. Bull ., 28 , 
2665 – 2671 . 
45. Desai , D. S. , Abdelnasser , M. A. , Rubitski , B. A. , and Varia , S. A. ( 1994 ), Photostabilization 
of uncoated tablets of sorivudine and nifedipine by incorporation of synthetic iron 
oxides , Int. J. Pharm ., 103 , 69 – 76 . 
46. Carstensen , J. T. , Johnson , J. B. , Spera , D. C. , and Frank , M. J. ( 1968 ), Equilibrium phenomena 
in solid dosage forms , J. Pharm. Sci. , 57 , 23 
47. Garrett , E. R. , and Carper , R. F. ( 1955 ), Prediction of stability in pharmaceutical preparations 
I. Color stability in a liquid multisulfa preparations . J. Am. Pharm. Assoc. Sci. 
Ed ., 44 , 515 – 518 . 
48. Garrett . E. R. ( 1956 ), Prediction of stability in pharmaceutical preparations II. 
Vitamin stability in liquid multivitamin preparations . J. Am. Pharm. Assoc. Sci. Ed ., 
45 , 171 – 178 . 

49. Garrett , E. R. ( 1956 ), Prediction of stability in pharmaceutical preparations III. 
Comparison of vitamin stabilities in different multivitamin preparations . J. Am. Pharm. 
Assoc. Sci. Ed ., 45 , 470 – 473 . 
50. Powell , M. F. ( 1986 ), Enhanced stability of codeine sulfate: Effect of pH, buffer, and 
temperature on the degradation of codeine in aqueous solution , J. Pharm. Sci ., 75 , 901 . 
51. Carney , C. F. ( 1987 ), Solution stability of ciclosidomine , J. Pharm. Sci ., 76 , 393 . 
52. Magalhales , N. S. S. , Cave , G. , Seiler , M. , and Benita , S. ( 1991 ), The Stability and in vitro 
release kinetics of clofi bride emulsion , Int. J. Pharm ., 76 , 225 – 237 . 
53. Shi , L. , DeHaven , P. A. , and Burke , C. J. ( 2005 ), Biopharmaceutical stability studies: 
stable vaccine dosage form development for commerical use , Am. Pharm. Rev ., 8 ( 6 ), 
86 – 92 . 
54. Tydeman , M. S. , and Kirkwood , T. B. L. ( 1984 ), Design and analysis of accelerated degradation 
tests for the stability of biological standards. I. Properties of maximum likelihood 
estimators , J. Biol. Stand ., 12 , 195 – 206 . 
55. Flynn , G. L. ( 1996 ), Cutaneous and transdermal delivery: Processes and systems of 
delivery , in Banker , G. S. , and Rhodes , C. T. , Eds., Modern Pharmaceutics , 3rd ed., Marcel 
Dekker , New York , p. 239 . 
56. Carstensen , J. T. ( 1974 ), Stability of solids and solid dosage forms , J. Pharm. Sci ., 63 ( 1 ), 
1 – 14 . 
57. Horhota , S. T. , Burgio , J. , Lonski , L. , and Rhodes , C. T. ( 1976 ), Effect of storage at speci- 
fi ed temperature and humidity on properties of three directly compressible tablet formulations 
, J. Pharm. Sci ., 65 , 1746 – 1749 . 
58. Alam , A. S. , and Parrott . E. L. ( 1971 ), Effect of aging on some physical properties of 
hydrochlorothiazide tablets , J. Pharm. Sci ., 60 , 263 – 266 . 
59. Carstensen , J. T. , and Kothari , R. ( 1983 ), Solid - state decomposition of alkoxyfuroix acids 
in the presence of microcrystalline cellulose , J. Pharm. Sci ., 72 , 1149 . 
60. Ansel , H. C. , Allen , L. V. , and Popovich , N. G. ( 1999 ), Pharmaceutical Dosage Forms and 
Drug Delivery Systems . 7th ed., Lippincott Williams and Wilkins , Baltimore, MD , p. 98 . 
61. Hecht , G. , Roehs , R. R. , Lang , J. C. , Rodheaver , D. P. , and Chowhan , M. A. ( 1996 ), Design 
and evaluation of ophthalmic pharmaceutical products , in Banker , G. S. , and Rhodes , 
C. T. , Eds., Modern Pharmaceutics , 3rd ed., Marcel Dekker , New York , p. 489 . 
62. Lee , J. Y. ( 1992 ), Sterilization control and validation for topical ointments , Pharm. Tech ., 
16 , 104 – 110 . 
63. Ibrahim , Y. K. E. , and Olurinola , P. R. ( 1991 ), Comparitive microbiological contamination 
levels in wet granulation and direct compression methods of tablet production , 
Pharm. Acta Helv ., 66 , 298 . 
64. Carstensen , J. T. ( 1996 ), Preformulation , in Banker , G. S. , and Rhodes , C. T. , Eds., Modern 
Pharmaceutics , 3rd ed., Marcel Dekker , New York , p. 213 . 
65. Guillory , J. K. , and Higuchi , T. ( 1962 ), Solid state stability of some vitamin a compounds , 
J. Pharm. Sci ., 51 , 100 – 105 . 
66. Pikal , M. J. , Lukes , A. L. , Lang , J. E. , and Gaines , K. ( 1978 ), Quantitative crystalline 
determinations for b – lactam antibiotics by solution calorimetry: Correlation with stability 
, J. Pharm. Sci ., 67 , 767 – 773 . 
67. Krahn , F. U. , and Mieck , J. B. ( 1989 ), Effect of type and extent of crystalline order on 
chemical and physical stability of carbamazepine , Int. J. Pharm ., 53 , 25 – 34 . 
68. Matsuda , Y. , Akazawa , R. , Teraoka , R. , and Totsuka , M. ( 1994 ), Pharmaceutical evaluation 
of carbamazepine modifi cations: Comparative study of photostability of carbamazepine 
polymorphs by using fourier - transformed refl ection - absorption infrared 
spectrocopy and colorimetric measurement , J. Pharm. Pharmacol ., 46 , 162 – 167 . 
REFERENCES 679

680 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
69. DeVilliers , M. M. , van der Watt , J. G. , and Lotter , A. P. ( 1992 ), Kinetic study of the solid - 
state photolytic degradation of two polymorphic forms of furosemide , Int. J. Pharm ., 88 , 
275 – 283 . 
70. Nakagawa , Y. , Itai , S. , Yoshida , T. , and Nagai , T. ( 1992 ), Physicochemical properties and 
stability in the acidic solution of a new macrolide antibiotic, clarithromycin, In comparison 
with erythromycin , Chem. Pharm. Bull ., 40 , 725 – 728 . 
71. Pikal , M. , and Lukes , A. L. ( 1976 ), Kundsen vapor pressure measurements on pure 
materials and solutions dispersed in porous media: Molded nitroglycerin tablets , J. 
Pharm. Sci ., 65 , 1269 . 
72. Fusari , S. A. ( 1973 ), Nitroglycerin sublingual tablets II: preparation and stability of a 
new stabilized sublingual molded nitroglycerin tablet , J. Pharm. Sci ., 62 , 2012 . 
73. Finholt , P. , Jurgensen , G. , and Kristiansen , H. ( 1965 ), Catalytic effect of buffers on 
degradation of penicillin G in aqueous solution , J. Pharm. Sci ., 54 , 387 – 393 . 
74. Tsuji , A. , Nakashima , E. , Deguchi , Y. , Nishide , K. , Shimizu , T. , Horiuchi , S. , Ishikawa , K , 
and Yamana , T. ( 1981 ), Degradation kinetics and mechanism of aminocephalosporins in 
aqueous solution: Cefadroxil , J. Pharm. Sci ., 70 , 1120 – 1128 . 
75. Zia , H. , Teharan , M. , and Zargarbashi . R. ( 1974 ), Kinetics of carbenicillin degradation 
in aqueous solutions , Can. J. Pharm. Sci ., 9 , 112 – 117 . 
76. Pramar , Y. , and Gupta , V. D. ( 1991 ), Preformulation studies of spironolactone: Effect of 
pH, two buffer species, ionic strength, and temperature on stability , J. Pharm. Sci ., 80 , 
551 – 553 . 
77. Allen , A. E. , and Das Gupta , V. ( 1974 ), Stability of hydrocortisone in polyethyleneglycol 
ontiment base , J. Pharm. Sci ., 63 , 107 – 109 . 
78. Das Gupta , V. ( 1978 ), Effect of vehicles and other active ingredients on stability of 
hydrocrtisone , J. Pharm. Sci ., 67 , 299 – 302 . 
79. Jun , H. W. , Whitworth , C. W. , and Luzzi , L. A. ( 1972 ), Decomposition of aspirin in polyethyleneglycols 
, J. Pharm. Sci ., 61 , 1160 – 1162 . 
80. Whitworth , C. W. , Luzzi , L. A. , Thompson , B. B. , and Jun , H. W. ( 1973 ), Stability of aspirin 
in liquid and semi - solid bases. II. Effect of fatty additives on stability in a polyethyleneglycol 
base , J. Pharm. Sci ., 62 , 1372 – 1374 . 
81. Ekman , R. , Liponkoski , L. , and Kahela , P. ( 1982 ), Formation of indomethacin esters in 
polyethylene glycol suppositories , Acta Pharm. Suec ., 19 , 241 – 246 . 
82. Castello , R. A. , and Mattocks , A. M. ( 1962 ), Discoloration of tablets containing amines 
and lactose , J. Pharm. Sci ., 51 , 106 – 108 . 
83. Tonnesen , H. H. ( 2001 ), Formulation and stability testing of photolabile drugs , Int. J. 
Pharm ., 225 , 1 – 14 . 
84. Baugh , R. , Calvert , R. T. , and Fell , J. T. ( 1977 ), Stability of phenylbutazone in the presence 
of pharmaceutical colors , J. Pharm. Sci ., 66 , 733 – 735 . 
85. Kornblum , S. S. , and Sciarrone , B. J. ( 1964 ), Decarboxylation of p - aminosalicylic acid in 
the solid state , J. Pharm. Sci ., 53 , 935 . 
86. Liebe , D. C. ( 1996 ), Packaging of pharmaceutical dosage forms , in Banker , G. S. , 
and Rhodes , C. T. , Eds., Modern Pharmaceutics , 3rd ed., Marcel Dekker , New York , 
p. 681 . 
87. Dean , D. A. ( 2000 ), Packaging, package evaluation, stability and shelf - life , in Carstensen , 
J. T. , and Rhodes , C. T. , Eds., Drug Stability, Principles and Practices , Marcel Dekker , 
New York , p. 483 . 
88. Code of Federal Regulations , Food and drugs, Title 21, Part 211, Current good manufacturing 
practice for fi nished pharmaceuticals, U.S. Government Printing Offi ce, Washington, 
DC, rev. Apr. 1, 2001. 

89. Croce , C. P. , Fischer , A. , and Thomas , R. H. ( 1986 ), Packaging materials science , in, 
Lachman , L. , Lieberman , H. A. , and King , J. L. , Eds., The Theory and Practice of Industrial 
Pharmacy , 3rd ed., Lea & Febiger , Philadelphia , p. 711 . 
90. Sciarra , J. J. ( 1996 ), Pharmaceutical aerosols , in Banker , G. S. , and Rhodes , C. T. Eds., 
Modern Pharmaceutics , 3rd ed., Marcel Dekker , New York , p. 547 . 
91. Lusina , M. , Cindric , T. , Tomaic , J. , Peko , M. , Pozaic , L. , and Musulin , N. ( 2005 ), Stability 
study of losartan/hydrochlorothiazide tablets , Int. J. Pharm ., 291 , 127 – 137 . 
92. Allinson , J. G. , Dansereau , R. J. , and Sakr . A. ( 2001 ), The effects of packaging on the 
stability of a moisture sensitive compound , Int. J. Pharm ., 221 , 49 – 56 . 
93. Mahajan , R. , Templeton , A. , Harman , A. , Reed , R. A. , and Chern , R. T. ( 2005 ), The effect 
of innert atmospheric packaging on oxidative degradation in formulated granules , 
Pharm. Res ., 22 ( 1 ), 128 – 140 . 
94. Al - Zein , H. , Riad , L. E. , and Abd - Elbary , A. ( 1999 ), Effect of packaging and storage on 
the stability of carbamazepine tablets . Drug Dev. Ind. Pharm ., 25 ( 2 ), 223 – 227 . 
95. Veillard , M. , Bentejac , R. , Duchene , D. , and Carstensen , J. T. ( 1979 ), Moisture transfer 
tests in blister package testing , Drug Dev. Ind. Pharm . 5 , 227 – 244 . 
96. Amidon , G. E. , and Middleton , K. R. ( 1988 ), Accelerated physical stability testing and 
long - term predictions of changes in the crushing strength of tablets stored in blister 
packages , Int. J. Pharm ., 45 , 79 – 89 . 
97. Khalil , S. A. H. , Barakat , N. S. , and Boraie , N. A. ( 1991 ), Effects of package type on in 
vitro release and chemical stability of amoxycillin in capsules , Pharm. Ind ., 53 ( 7 ), 
698 – 701 . 
98. Gouda , M. W. , Moustafa , M. A. , and Molokhia , A. M. ( 1980 ), Effect of storage conditions 
on erythromycin tablets marketed in saudi arabia , Int. J. Pharm ., 5 , 345 – 347 . 
99. Pikal , M. J. , Bibler , D. A. , and Rutherford , B. ( 1977 ), Polymer sorption of nitroglycerin 
and stability of molded nitroglycerin tablets in unit - dose packaging , J. Pharm. Sci ., 66 ( 9 ), 
1293 – 1297 . 
100. Barakat , N. S. ( 2006 ), Etodolac - liquid - fi lled dispersion into hard gelatin capsules: An 
approach to improve dissolution and stability of etodolac formulation , Drug Dev. Ind. 
Pharm ., 32 , 865 – 876 . 
101. Santoro , M. I. R. M. , Oliveira , D. A. G. D. C. E. , Kedor - Hackmann , E. R. , and Singh , 
A. K. ( 2005 ), The effect of packaging materials on the stability of sunscreen emulsions , 
Int. J. Pharm ., 297 , 197 – 203 . 
102. Geary , T. G. , Akkod , M. A. , and Jensen , J. B. ( 1983 ), Characteristics of cheloroquine 
binding to glass and plastic , Am. J. Trop. Med. Hyg . 32 , 19 – 23 . 
103. Yahya , M. A. , McElnay , J. C. , and D ’ Arcy , P. F. ( 1985 ), Binding of chloroquine to glass , 
Int. J. Pharm ., 25 , 217 – 223 . 
104. Stanaszek , W. F. , and Pan , I. H. ( 1978 ), Comparison of drug stability in glass versus 
plastic containers: Analysis of prefi lled syringe admixtures , Proc. Okla. Acad. Sci ., 58 , 
102 – 105 . 
105. Arsene , M. , Favetta , P. , Favier , B. , and Bureau , J. ( 2002 ), Comparison of ceftazidime 
degradation in glass bottles and plastic bags , J. Clin. Pharm. Ther ., 27 ( 3 ), 202 – 209 . 
106. Crane , I. M. , Mulhern , M. G. , and Nema , S. ( 2003 ), Stability of reconstituted parecoxib 
for injection with commonly used diluents . J. of Clin. Pharm. Thera ., 28 , 363 – 369 . 
107. Wong , M. , Marion , R. , Reed , K. , and Wang , Y. ( 2006 ), Sorption of unoprostone isopropyl 
to packaging materials , Int J. Pharm . 307 , 163 – 167 . 
108. Driscoll , D. F. , Ling , P. R. , and Bistrian , B. R. ( 2007 ), Physical stability of 20% lipid injectable 
emulsions via simulated syringe infusion: effects of glass versus plastic product 
packaging , J. Parenteral Enteral Nutr ., 31 ( 2 ), 148 – 153 . 
REFERENCES 681

682 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
109. Yuen , P. H. , Denman , S. L. , Sokoloski , T. D. , and Burkman , A. M. ( 1979 ), Loss of nitroglycerin 
from aqueous solution into plastic intravenous delivery systems . J. Pharm. Sci ., 
68 , 1163 – 1166 . 
110. Baaske , D. M. , Amann , A. H. , Wagenknecht , D. M. , Carter , J. E. , Hoyt , H. J. , and Stoll , 
R. G. ( 1980 ), Nitroglycerin compatibility with intravenous fl uid fi lters, containers and 
administration sets , Am. J. Hosp. Pharm ., 37 , 201 – 205 . 
111. Cossum , P. A. , Roberts , M. S. , Galbraith , A. J. , and Boyd , G. W. ( 1978 ), Loss of nitroglycerin 
from intravenous infusion sets , Lancet ., 2 , 34 – 350 . 
112. Crouthamel , W. G. , Dorsch , B. , and Shangraw , R. ( 1978 ), Loss of nitroglycerin from 
plastic intravenous bags , New Engl. J. Med ., 299 , 262 . 
113. Sokoloski , T. D. , Wu , C. C. , and Burkman , A. M. ( 1980 ), Rapid adsorptive loss of nitroglycerin 
from aqueous solution to plastic , Int. J. Pharm ., 6 , 63 – 76 . 
114. Malick , A. W. , Amann , A. H. , Baaske , D. M. , and Stoll , R. G. ( 1981 ), Loss of nitroglycerin 
from solutions to intravenous plastic containers: A theoretical treatment . J. Pharm. Sci ., 
70 , 798 – 800 . 
115. Yuen , P. H. , Denman , S. L. , Sokoloski , T. D. , and Burkman , A. M. ( 1979 ), Loss of nitroglycerin 
from aqueous solution into plastic intravenous delivery systems , J. Pharm. Sci ., 
68 ( 9 ), 1193 – 1166 . 
116. Illum , L. , and Bubdgaard , H. ( 1982 ), Sorption of drugs by plastic infusion bags , Int. J. 
Pharm ., 10 , 339 – 351 . 
117. Pikal , M. J. , and Lang , J. E. ( 1978 ), Rubber closures as a source of haze in freeze - dried 
parenterals: Test methodology for closure evaluation , J. Parent. Drug. Assoc ., 32 , 
162 – 173 . 
118. Jahnke , R. W. O. , Kreuter , J. , and Ross . G. ( 1991 ), Content/container interactions: The 
phenomenon of haze formation on reconstitution of solids for parenteral use , Int. J. 
Pharm ., 77 , 47 – 55 . 
119. Parker , W. A. , and MacCara , M. E. ( 1980 ), Compatibility of diazepam with intravenous 
fl uid containers and administration sets , Am. J. Hosp. Pharm ., 37 , 496 – 500 . 
120. Morris , M. E. ( 1978 ), Comaptibility and stability of diazepam injection following dilution 
with intravenous fl uids , Am. J. Hosp. Pharm ., 35 , 669 – 672 . 
121. Parker , W. A. , Morris , M. E. , and Shearer , C. A. ( 1979 ), Incompatibility of diazepam 
injection in plastic intravenous bags , J. Am. Hosp. Pharm ., 36 , 505 – 507 . 
122. Van Amerongen , G. J. ( 1964 ), Diffusion in elastomers , Rubber Chem. Technol ., 37 , 
1065 – 1152 . 
123. Vromans , H. and van Laarhoven , J. A. H. ( 1992 ), A study on water permeation through 
rubber closures of injection vials , Int. J. Pharm ., 79 , 301 – 308 . 
124. U.S. Food and Drug Administration (FDA) ( 1998, Oct. ), Guidance for industry: Metered 
dose inhaler and dry powder inhaler drug products, FDA Washington, DC. 
125. Castner , J. , Anderson , J. , and Benites , P. ( 2007 ), Strategy for development and characterization 
of HPLC methods to investigate extractables and leachables , Am. Pharm. 
Rev ., 10 ( 3 ), 10 . 
126. Northup , S. J. ( 2005 ), Assesing the biological safety of extractable and leachable chemicals 
in pharmaceutical and medicial products , Am. Pharm. Rev ., 8 ( 4 ), 38 – 43 . 
127. U.S. Food and Drug Administration (FDA) ( 2002, July ), Nasal spray and inhalation 
solution, suspension, and spray drug products.: Chemistry, manufacturing, and controls 
documentation, Guidance for industry. FDA, Washington, DC, 1 – 45. 
128. Norwood . D. L. ( 2007 ), Understanding the challenges of extractables and leachables for 
the pharmaceutical industry — Safety and requlatory enviroment for pharmaceuticals , 
Am. Pharm. Rev ., 10 ( 2 ), 32 . 

129. U.S. Food and Drug Administration (FDA), Extra 6: Safety thresholds and best practices 
for extractables and leachables in orally inhaled and nasal drug products, PQRI 
Leachables and Extratables Working Group, available: http://www.pqri.org/pdfs/ 
LE_Recommendations_to_FDA_09-26-06.pdf . 
130. IPAC - RS (1991, Mar.) Extra 7: Leachables and extractables testing: points to consider, 
available: http://www.ipacrs.com/leachables.html . 
131. Allan , L. , and Wang , Q. ( 2007 ), Impact of package on the stability of pharmaceutical 
products , Am. Pharm. Rev ., 10 ( 4 ), 38 – 44 . 
132. Waterman, K. C. , Adami, R. C. , Alsante , K. M. , Hong , J. , Landis , M. S. , Lombardo , F. , and 
Roberts , C. J. ( 2002 ), Stabilization of pharmaceuticals to oxidative degradation , Pharm. 
Dev. Tech ., 7 ( 1 ), 1 – 32 . 
133. Reed , R. A. , Harmon , P. , Manas , D. , Wasylaschuk , W. , Galli , C. , Biddell , R. , Bergquist , P. 
A. , Hunke , W. , Templeton , A. C. , and Ip , D. ( 2003 ), The role of excipients and package 
components in the photostability of liquid formulations , PDA J. Pharm. Sci. Technol ., 
57 ( 5 ), 351 – 368 . 
134. Hong , J. , Lee , E. , Carter , J. C. , Masse , J. A. , and Oksanen , D. A. ( 2004 ), Antioxidant - 
accelerated oxidative degradation: A case study of transistion metal ion catalyzed oxidation 
in formulation , Pharm. Dev. Tech ., 9 ( 2 ), 171 – 179 . 
135. Quarry , M. A. , Sebastian , D. S. , and Diana , F. ( 2002 ), Investigation of 4,5 - epoxymorphian 
degradation during analysis by HPLC , J. Pharm. Biomed. Anal . 30 , 99 – 104 . 
136. Nakamura , K. , Yokohama , S. , Sawada , M. T. , and Sonobe , T. ( 2003 ), A new stressed test 
to predict the foreign matter formation of minodronic acid in solution , Int. J. Pharm ., 
251 , 99 – 106 . 
137. Bohere , D. , Nascimento , P. C. , Binotto , R. , and Becker , E. ( 2003 ), Infl uence of the glass 
packing on the contamination of pharmaceutical products by aluminium. Part III. Interaction 
of the container - chemicals during heating for sterilization , J. Trace Elem. Med. 
Biol ., 17 ( 2 ), 107 – 115 . 
138. Shen , F. W. , and McKellop , H. A. ( 2002 ), Interaction of oxidation and crosslinking in 
gamma - irradiated ultrahigh molecular - weight polyethylene , J. Biomed. Mater. Res ., 
61 ( 3 ), 430 – 439 . 
139. Welle , F. ( 2005 ), Migration of radiolysis products from radiation - sterilized plastics , 
Pharm. Ind ., 67 ( 8 ), 970 – 972 . 
140. Nassar , M. N. , Nesarker , V. V. , Lozano , R. , Yande , H. , and Palaniswany , V. ( 2005 ), Degradation 
of a lyophilized formulation of BMS - 204352: Identifi cation of degradants and 
the role of elastomeric closure . Pharm. Dev. Tech ., 10 ( 2 ), 227 – 232 . 
141. Glaser , V. ( 2005 ), Addressing stability of biological drugs: forced degradation studies 
predict effects on bioproducts in drug development and manufacture , GEN Genetic eng. 
news , 25 ( 6 ), 1 – 5 . 
142. Code of Federal Regulations , Food and drugs, Title 21, Part 211.65, U.S. Government 
Printing Offi ce, Washington, DC, rev. Apr. 1, 2004. 
143. Federal Food, Drug and Cosmetic Act as amended through Dec. 21, 2000, Chapter V, 
Drugs and devices, Section 501(a)(3), U.S. Government Printing Offi ce, Washington DC, 
2001. 
144. Taborsky , C. J. , and Sheinin , E. B. ( 2006 ), A critical approach to the evaluation of packaging 
components and the regulatory and scientifi c considerations in developing a 
testing strategy , Am. Pharm. Rev ., 9 ( 4 ), 10 – 15 . 
145. Gudat , A. E. , and Firor . R. L. ( 2006, Sept. ), The determination of extractables and leachables 
in pharmaceutical packaging materials using headspace GC/MS, Application Note, 
Agilent Technology. 
REFERENCES 683

684 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
146. Janimak , J. J. , and Marteleur , M. ( 2004 ), On the stability of cerium oxide glass for terminal 
radiation sterilization , Rad. Phys. Chem . 71 , 195 – 198 . 
147. DeGrazio , F. L. ( 2005 ), Parenteral packaging concerns for drugs: Early understanding 
of drug to package compatibility lowers the cost in the long run . GEN: Genetic Eng. 
News (Bioprocess. Channel) , 25 ( 21 ), 2 . 
148. DeGrazio , F. L. ( 2005 ), Parenteral packaging concerns for drugs: Early understanding 
of drug to package compatibility lowers the cost in the long run . GEN: Genetic Eng. 
News (Bioprocess. Channel) , 25 ( 21 ), 2 . 
149. Romacker , M. ( 2005 ), Prefi lled syringes: Why new developments are important in injectable 
delivery today , in Roessling , G. , Ed., Prefi lled Sringes: Innovations That Meet the 
Growing Demand , ONdrugDeliver y, Sussex, United Kingdom . 
150. Wang , W. , Wang , Y. J. , and Kelner , D. N. ( 2003 ), Coagulation factor VIII: Strucuture and 
stability , Int. J. Pharm ., 259 , 1 – 15 . 
151. Schulman , S. , Gitel , S. , and Martinowitz , U. ( 1994 ), Stability of factor VIII concentrates 
after reconstitution , Am. J. Hematol ., 45 , 217 – 223 . 
152. Osterberg , T. , Fatouros , A. , Neidhardt , E. , Warne , N. , and Mikaelsson , M. ( 2001 ), 
B - domain - deleted recombinant factor VIII formulation and stability , Semin. Hematol ., 
38 , 40 – 43 . 
153. Andrade , J. D. , and Hlady , V. ( 1986 ), Protein adsorption and materials biocompatibility: 
A tutorial review and suggested hypothesis , Adv. Polym. Sci ., 79 , 1 – 63 . 
154. Duncan , M. R. , Lee , J. M. , and Warchol , M. P. ( 1995 ), Infl uence of surfactants upon 
protein/peptide adsorption to glass and polypropylene , Int. J. Pharm ., 120 , 179 – 188 . 
155. Wang , P. L. , Udeani , G. O. and Johnston , T. P. ( 1995 ), Inhibition of granulocyte colony 
stimulating factor (c - csf) adsorption to polyvinyl chloride using a nonionic surfactant , 
Int. J. Pharm ., 114 , 177 – 184 . 
156. Burke , C. J. , Steadman , B. L. , Volkin , D. B. , Tsai , P. K. , Bruner , M. W. , and Middaugh , C. 
R. ( 1992 ), The adsorption of proteins to pharmaceutical container surfaces , Int. J. Pharm ., 
86 , 89 – 93 . 
157. Weber , S. S. , Wood , W. A. , and Jackson , E. A. ( 1977 ), Availability of insulin from parenteral 
nutrient solutions , Am. J. Hosp. Pharm ., 34 , 353 – 357 . 
158. Weeren , R. V. , and Gibboni , D. J. , Barrier packaging as an integral part of drug delivery, 
available: http://www.drugdeliverytech.com/cgi-bin/articles=52 (accessed June 1, 2007). 
159. U. S. Patent 5,028,431 : Franz , T. J. , Shah , K. R. , and Kydonieus , A. Article for the Delivery 
to Animal Tissue of a Pharmacological Active. July 2, 1991 ; U. Patent 5,242,433 : Smith , 
J. A. , and Murphy , B. , Packaging System with In - Tandem Applicator Pads for Topical 
Drug Delivery. Decemeber 7, 1992 . 
160. U. S. Patent 5,698,217 : Wilking , S. L. Transdermal Drug Delivery Device Containing a 
Dessicant. December 16, 1997 . 
161. Hoye , W. L. , Mogalian , E. M. , and Myrdal , P. B. ( 2005 ), The effects of extreme temperatures 
on drug delivery of albuterol sulfate hydrofl uoroalkane inhalation products , 
Am. J. Health - Syst. Pharm ., 62 ( 21 ), 2271 – 2277 . 
162. European Pharmacopoeia (EP) ( 2001 ), Preparations for inhalation, EP supplement 
2001: 2.9.18, EP, Strasbourg, France. 
163. U.S. Food and Drug Administration (FDA) ( 1998 ), Guidance for industry: Metered dose 
inhaler (MDI) and dry powder inhaler (DPI) drug products, FDA CDER, Washington, 
DC. 
164. Petersson , G. , and Wiren , J. E. ( 1980 ), The brochodilator response from inhaled terbutaline 
is infl uenced by the mass of small particles: A study on a dry powder inhaler 
(Turbuhaler) , Eur. Respir . 2 , 253 – 256 . 

165. Newman , S.P. , and Busse , W. W. ( 2002 ), Evolution of dry powder inhaler design, formulation 
and performance , Respir. Med ., 96 , 293 – 304 . 
166. Carter , P. A. , Rowley , G. , Fletcher , E. J. , and Sylianopoulos , V. ( 1998 ), Measurement of 
electrostatic charge decay in pharmaceutical powdersand polymer materials used in dry 
powder inhaler devices , Drug Dev. Ind. Pharm ., 24 , 1083 – 1088 . 
167. Shire , S. J. ( 1996 ), Stability, characterization and formulation of recombinant human 
deoxyribonuclease (Pulmozyme ®, Dornase Alpha) , in Pearlman , R. , and Wang , Y. J. Eds., 
Formulation, Characterization and Stability of Protein Drugs , Plenum Press , New York , 
pp. 393 – 426 . 
168. Arnum , P. V. ( 2007 ), 3M drug delivery systems advances inhalation and transdermal 
drug delivery, Pharm. Technol. Online, available http://www.pharmtech.com , Jan. 
31. 
169. Jinks , P. , and Marsden , S. ( 1999 ), The development and performance of a fl uoropolymer 
lined can for suspension metered dose inhaler product, in Fradley , G. , Ed., Proceedings 
of Drug Delivery to the Lungsa X , The Aerosol Society, Portshead, United Kingdom, 
pp. 177 – 180 . 
170. Wilby , M. ( 2005 ), Increasing dose consistency of pMDis , Drug Deliv. Technol ., 5 ( 9 ), 
59 – 65 . 
171. Wilby , M. ( 2006 ), Novel valve designs to eliminate loss of prime , in Dalby , X. R. N. 
et al. , Eds., Proceedings of Respiratory Drug Delviery , Vol. 2, Davis Healthcare, 
Boca Raton, FL, pp. 373 – 376 . 
172. Jinks , P. , and Hunt , K. ( 2006 ), Improving suspension MDI dose consistency in patient 
use by incorporation of a novel semi - permemable system component , in Van Arnum , P. , 
Ed., Proceedings of Drug Delivery to the Lungs XVII , The Aerosol Society, Portshead, 
United Kingdom, pp. 172 – 175 . 
173. Fairweather , W. R. , Lin , T. Y. D., and Roswitha , K. ( 1995 ), Regulatory, design, and analysis 
aspects of complex stability studies , J. Pharm. Sci ., 84 ( 11 ), 1322 – 1326 . 
174. Dyrstad , K. , Veggeland , J. , and Thomassen , C. ( 1999 ), A multivariate method to predict 
the water vapour diffusion rate through polypropylene packaging , Int. J. Pharm ., 188 , 
105 – 109 . 
175. Dyrstad , K. , Thomassen , C. , and Eivindvik , K. ( 1999 ), An opportunistic stability strategy: 
simulation with real data , Int. J. Pharm ., 188 , 97 – 104 . 
176. Ruberg , S. J. , and Hsu , J. C. ( 1991 ), Multiple comparison procedures for pooling batches 
in stability studies , Technometrics , 34 , 465 – 472 . 
177. Su , X. Y. , Po , A. L. W., and Yoshioka , S. ( 1994 ), A bayesian approach to arrhenius prediction 
of shelf life , Pharm. Res ., 11 , 1462 – 1466 . 
178. Yoshika , S. , Aso , T. , and Kojima , K. ( 1996 ), Statistical evaluation of shelf - life of pharmaceutical 
products estimated matrixing , Drug Stab . 1 , 147 – 151 . 
179. Kontny , M. J. , Koppenol , S. , and Graham , E. T. ( 1992 ), Use of the sorption - desorption 
moisture transfer model to assess the utility of a desiccant in a solid product , Int. J. 
Pharm ., 84 , 261 – 271 . 
180. Badway , S. I. F. , Gawronski , A. J. , and Alvarez , F. J. ( 2001 ), Application of sorption - 
desorption moisture transfer modeling to the study of chemical stability of a moisture 
sensitive drug product in different packaging confi gurations , Int. J. Pharm ., 223 ( 1 – 2 ), 
1 – 13 . 
181. Kontny , M. J. , and Mulski , C. A. ( 1998 ), Gelatin capsule brittleness as a function of 
relative humidity at room temperature , Int. J. Pharm ., 54 ( 1 ), 79 – 85 . 
182. Illum , L. , Bundgaard , H. , and Davis , S. S. ( 1983 ), A constant partition model for examining 
the sorption of drugs by plastic infusion bags , Int. J. Pharm ., 17 , 183 – 192 . 
REFERENCES 685

686 EFFECT OF PACKAGING ON STABILITY OF DRUGS AND DRUG PRODUCTS 
183. Nakabayashi , K. , Shimamoto , T. , and Mima , H. ( 1980 ), Stability of packaged solid dosage 
forms. I. shelf - life prediction for packaged tablets liable to moisutre damage , Pharm. 
Chem. Bull ., 28 , 1090 – 1098 . 
184. Nakabayashi , K. , Shimamoto , T. , and Mima , H. ( 1980 ), Stability of packaged solid dosage 
forms. II. Shelf - life prediction for packaged sugar - coated tablets liable to moisture and 
heat damage , Pharm. Chem. Bull ., 28 , 1099 – 1106 . 
185. Jenke , D. R. ( 1993 ), Modeling of solute sorption by polyvinyl chloride plastic infusion 
bags , J. Pharm. Sci ., 82 ( 11 ), 1134 – 1139 . 
186. Roberts , M. S. , Kowaluk , E. A. , and Polack , A. E. ( 1991 ), Prediction of solute sorption 
by polyvinyl chloride plastic infusion bags , J. Pharm. Sci ., 80 ( 5 ), 449 – 455 . 

687 
7.4 
PHARMACEUTICAL PRODUCT 
STABILITY 
Andrew A. Webster 
McWhorter School of Pharmacy, Birmingham, Alabama 
Contents 
7.4.1 Introduction 
7.4.2 Kinetic Equations and Half - Life 
7.4.2.1 Rate Expression 
7.4.2.2 Order Determination 
7.4.2.3 Predicting Shelf Life 
7.4.2.4 Arrhenius Equation and Accelerated Stability Testing 
7.4.3 Degradation Pathways of Pharmaceuticals 
7.4.4 Chemical Degradation 
7.4.4.1 Solvolysis 
7.4.4.2 Oxidation 
7.4.4.3 Photolysis 
7.4.4.4 Dehydration 
7.4.4.5 Racemization 
7.4.5 Physical Degradation 
7.4.5.1 Polymorphism 
7.4.5.2 Vaporization 
7.4.6 Microbial Degradation 
7.4.7 Stability Guidelines and Regulations 
7.4.8 ICH Quality Guidelines 
References 
7.4.1 INTRODUCTION 
Pharmaceutical product stability may be defi ned as the capability of a particular 
formulation to remain within its physical, chemical, microbiological, therapeutic, 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

688 PHARMACEUTICAL PRODUCT STABILITY 
and toxicological specifi cations while in a specifi c container closure system. A collection 
of valid data on the drug in its specifi c container leads to assurances that the 
product will be stable for the assigned shelf life. 
Often, stability of a drug has been referred to as the time from the date of manufacture 
and packaging until its chemical or biological activity is not less than a predetermined 
level of labeled potency without the physical characteristics changing 
appreciably. For most drugs, 90% of labeled potency is generally recognized as the 
minimum acceptable potency. 
There are many factors that can affect the stability of a pharmaceutical product. 
These include the stability of the active drug(s), interactions between active and 
inactive ingredients, the dosage form, manufacturing process, the container system, 
and environment for shipping, handling, and storage. 
The U.S. Pharmacopeia 29/National Formulary 24 (USP29/NF24) defi nes stability 
as the extent to which a product retains within specifi ed limits and throughout its 
period of storage and use (i.e., shelf life) the same properties and characteristics 
that it possessed at the time of its manufacture [1] . The USP29/NF24 further identi- 
fi es fi ve generally recognized types of stability: 
• Chemical degradation — Each active ingredient retains its chemical integrity 
and labeled potency within specifi ed limits. 
• Physical — The original physical properties, including appearance, palatability, 
uniformity, dissolution, and suspendability, may be affected. 
• Microbiological — Sterility or resistance to microbial growth is retained according 
to the specifi ed requirements. Antimicrobial agents that are present retain 
effectiveness within the specifi ed limits. 
• Therapeutic — The therapeutic effect remains unchanged. 
• Toxicological — No signifi cant increase in toxicity occurs. 
Credible pharmaceutical product expiration dates are obtained by rigorous, scientifi 
cally designed studies using reliable, meaningful, and specifi c stability-indicating 
assays, appropriate statistical concepts, and computers to analyze the resulting data 
[2] . A comprehensive review of all aspects of pharmaceutical product stability has 
been published by Lintner [3] and more recently by Connors et al. [4] . 
7.4.2 KINETIC EQUATIONS AND SHELF LIFE 
Consider the reaction 
a b p Q A B P q + > + (1) 
where A and B are the reactants, P and Q are the products, and a, b, p, q are the 
stoichiometric coeffi cients describing the reaction. The rate of change of the concentration 
C of any of the species can be expressed by 
. . dC
dt 
dC
dt 
dC
dt 
dC
dt 
A B P Q , , , 

The reactants decrease in concentration relative to time — hence the negative sign. 
On the other hand, the products increase over time and are preceded by a positive 
sign. The rates of disappearance of A and B and the rates of appearance of P and 
Q are interrelated by equations that take into account the stoichiometry of the 
reaction: 
. =. = = 1 1 1 1 
a 
dC
dt b 
dC
dt p 
dC
dt q 
dC
dt 
A B P Q 
(2) 
7.4.2.1 Rate Expression 
The rate expression is a mathematical description of the rate of the reaction at any 
time t in terms of the concentrations(s) of the molecular species present at that time. 
By simplifying Equation (1) to 
A B products + > (3) 
the rate expression can be written as 
. =. . ( ) ( ) 
dC
dt 
dC
dt 
C Ct 
a 
t 
b A B 
A B 
(4) 
Equation (4) in essence states that the rate of change of the concentration of A at 
time t is equal to that of B and that each of these changes at time t is proportional 
to the product of the concentrations of the reactants raised to the respective powers. 
Note that C A( t ) and C B( t ) are time - dependent variables. As the reaction proceeds, 
both C A( t ) and C B( t ) will decrease. For simplicity, these concentrations can be denoted 
by C A and C B , respectively: 
. =. = dC
dt 
dC
dt 
kC C a b A B 
A B 
(5) 
where k is a proportionality constant, commonly referred to as the reaction rate 
constant or the specifi c rate constant. The format for rate expressions generally 
involves concentration terms of only the reactants and very rarely those of the 
products. The latter occurs only when the products participate in the reaction once 
it has been initiated. 
The order of the reaction, n , can be defi ned as n = a + b . Extended to the 
general case, the order of a reaction is the numerical sum of the exponents of the 
concentration terms in the rate expression. Thus, if a = b = 1, the reaction described 
above is said to be second order overall but fi rst order relative to A and fi rst order 
relative to B. In principle, the numerical value of a or b can be integral or 
fractional. 
At times the rate of reaction is apparently independent of the concentration of 
one of the reactants, even though this reactant is consumed during the reaction. For 
example, in the reaction between an ester and water (hydrolysis) in a predominantly 
aqueous environment, the theoretical rate expression for the ester can be written 
in terms of the concentrations of the ester ( C E ) and water ( C W ): 
KINETIC EQUATIONS AND SHELF LIFE 689

690 PHARMACEUTICAL PRODUCT STABILITY 
. = dC
dt 
kC C E 
E W 
(6) 
If the initial concentration of the ester is less than or equal to 0.5 M , complete 
hydrolysis of the ester will bring about a corresponding decrease in the concentration 
of water of 0.5 M or less. With the initial concentration of water being about 
55 M for an aqueous solution, the relative loss of water through a reaction is insignifi 
cantly small, enabling C W to be considered a constant throughout the entire 
course of the reaction. Therefore, 
. = dC
dt 
k C E 
E . 
(7) 
where k . = kC w . The reaction appears to be fi rst order relative to the ester and zero 
order relative to water. The overall reaction is known as a pseudo - fi rst - order reaction 
with k . the pseudo - fi rst - order constant. 
Pseudo - fi rst - order kinetics are observed whenever the concentration of one of 
the reactants is maintained constant, either by a substantial excess in initial concentration 
or by rapid replenishment of one of the reactants. If one of the reactants is 
the hydrogen ion or the hydroxide ion, its concentration, though probably small 
when compared with that of the drug, can be kept constant throughout the reaction 
by using buffers in the solution. The concentration of an unstable drug in solution 
can be maintained invariant by utilization of a suspension, which provides excess 
solid in equilibrium with the drug in solution. 
7.4.2.2 Order Determination 
Reaction orders can be determined by several methods: 
Substitution Method The accumulated data from a kinetic study can be substituted 
in the integrated form of the various equations that describe reaction 
orders. When the k values of one of the iterations remain constant, the reaction 
is considered to be of that order. 
Graphic Method A plot of the data can be used to ascertain the order. If a plot 
of concentration versus time yields a straight line, the reaction is zero order. 
A straight line from the plot of log( a . x ) versus time is fi rst order and second 
order if the plot of 1/( a . x ) 2 versus time is a straight line (where the initial 
concentrations are equal). 
Half - Life Method For a zero - order reaction the half - life ( t 1/2 ) is proportional to 
the initial concentration. The half - life for a fi rst - order reaction is independent 
of the initial concentration while a second - order reaction is proportional to 
1/initial concentration. 
7.4.2.3 Predicting Shelf Life 
The shelf life of a formulation is currently based on rigorous physical – chemical laws 
and statistical concepts to obtain reliable estimates. McMinn and Lintner have 

developed an information processing system for handling product stability data [5] . 
This system saves the time of formulators in analyzing and interpreting their product 
stability data. For products such as those of vitamins, for example, where large overages 
are required, the statistical portions of this advanced technique aid the manufacturer 
to tailor the formula composition to obtain the desired and most economical 
expiration dating. 
This system stores both physical and chemical data and retrieves the information 
in three different formats (one of which was designed specifi cally for submitting to 
regulatory agencies). It analyzes single - temperature data through analysis of covariance 
and regression. Multiple - temperature data are analyzed either weighted or 
unweighted using the Arrhenius relationship. This method provides estimates of the 
shelf life of the preparation with appropriate confi dence intervals, preprints the 
assay request cards that are used to record the results of the respective assay procedures 
and to enter the data into the system, and produces a 5 - year master stability 
schedule as well as periodic 14 - day schedules of upcoming assays. 
Analysis of stability data obtained at a single temperature is based on a linear 
(zero - order) model 
Y X mn m mn m mn = + + . . . (8) 
where Y mn is the percentage of label of the n th stability assay of the m th lot, X mn is 
the time in months at which Y mn was observed, . m and . m are the slope and intercept, 
respectively, of the regression line of the m th lot, and . mn is a random error associated 
with Y mn . The random errors are assumed to be identically and independently 
distributed normal variables with a zero mean and a common variance . 2 . 
A summary of the regression analysis for each individual lot and for the combination 
of these lots plus a summary of the analyses of covariance and deviation 
from regression is prepared by the computer. 
Since the stability data from individual lots are pooled, these data are examined 
for validity by the F test. The mean square of the regression coeffi cient (slope) is 
divided by the mean square of the deviation within lots, and similarly, the adjusted 
mean ( y intercept) is divided by the common mean square to give the respective F 
ratios. The latter values then are compared with the critical 5% F values. When the 
calculated F values are smaller than the critical F values, the data may be combined 
and the pooled data analyzed. 
7.4.2.4 Arrhenius Equation and Accelerated Stability Testing 
The purpose of stability testing is to assess the effects of temperature, humidity, light, 
and other environmental factors on the quality of a drug substance or product. These 
data sets are used to establish storage conditions, retest periods, and shelf loss and 
to justify overages included in products for stability reasons. The most useful equation 
relating temperature and reaction rate is the Arrhenius equation 
d k 
dT 
E 
RT 
ln a = 2 
(9) 
This can be rewritten as 
KINETIC EQUATIONS AND SHELF LIFE 691

692 PHARMACEUTICAL PRODUCT STABILITY 
k Ae E RT = . a ( ) (10) 
ln 
k
k 
E
R T T 
a 1
2 2 1 
1 1 ( )= . ( ) (11) 
where E a is a constant and the subscripts 1 and 2 denote the two different temperature 
conditions. A plot of ln k as a function of 1/ T , referred to as the Arrhenius plot, 
is linear according to Equation (10) if E a is independent of temperature. This observation 
makes it possible to conduct kinetic experiments at elevated temperatures 
and obtain estimates of rate constants at lower temperatures by extrapolation of 
the Arrhenius plot. This procedure, commonly known as accelerated stability testing, 
is most useful when the reaction at ambient temperature is too slow to be monitored 
conveniently and when E a is relatively high. As an example, in a reaction with an 
E s of 25 kcal/mol with an increase from 25 to 45 ° C, there is about a 14 - fold increase 
in the reaction rate constant. In contrast, a rate increase of just threefold is obtained 
for the same elevation in temperature when E a is 10 kcal/mol. The slope of the 
Arrhenius plot for a reaction yields the magnitude of E a . Hydrolysis reactions typically 
have an E a of 10 – 30 kcal/mol while oxidation and photolysis reactions have 
smaller energies of activation [6] . 
The elevated temperatures most commonly used are 40, 50, and 60 ° C in conjunction 
with ambient humidity. Occasionally, higher temperatures are used. The samples 
stored at the highest temperatures are examined weekly for physical and chemical 
changes. If a substantial change is seen, samples stored at lower temperatures are 
examined. If there is no change after 30 days at 60 ° C, the stability prognosis is excellent. 
Corroborative evidence must be obtained by monitoring the samples stored at 
lower temperatures for longer durations. 
An underlying assumption of the Arrhenius equation is that the reaction mechanism 
does not change as a function of temperature. Since accelerated stability 
testing of pharmaceutical products normally employs a narrow range of temperature, 
it is often diffi cult to detect nonlinearity in the Arrhenius plot from 
experimental data, even though such nonlinearity is expected from the reaction 
mechanism [7] . 
Non - Arrhenius behavior has been observed in pharmaceutical systems [8] . This 
may be attributed to the possible evaporation of solvent, multiple reaction pathways, 
and the change in physical form of the formulation when the temperature of the 
reaction is changed [9] . Nonlinearity in Arrhenius plots frequently is observed in 
following the temperature dependence of protein degradation. Degradation mechanisms 
in proteins often change with temperature [10] . The aggregation of interleukin 
1 . (IL - 1 . ) in aqueous solution follows apparent fi rst - order behavior at 60 ° C to 30% 
of drug remaining. At or below 55 ° C the aggregation deviates from apparent fi rst 
order and becomes biphasic [11] . 
Considerable interest has been generated in the use of accelerated stability 
testing based on a singe condition of elevated temperature and humidity. For abbreviated 
new drug applications (ANDAs) U.S. Food and Drug Administration (FDA) 
stability guidelines suggest that a tentative expiration date of 24 months may be 
granted for a drug product if satisfactory stability results can be documented under 
a stressed condition of 40 ° C and 75% relative humidity [12] . The simplicity of such 

CHEMICAL DEGREDATION 693 
a guideline is attractive because a substantial saving in time can be obtained in 
advancing a drug product to the marketplace [7] . 
7.4.3 DEGRADATION PATHWAYS OF PHARMACEUTICALS 
Many degradation pathways are similar to reactions described for organic compounds 
in organic chemistry texts. The decomposition of a drug is likely to be mediated 
by reaction with water, oxygen, or light. The routes most pharmaceuticals 
degrade include hydrolysis, oxidation, photolysis, and racemization. Chemical degradation 
occurs when a new chemical entity is formed as a process of the degradation 
process. Physical degradation occurs when drug loss does not produce distinctly 
different chemical products [13] . Degradation reactions of a variety of mechanisms 
can often be avoided by decreasing storage temperature and/or buffering pH to a 
stability optimum. 
7.4.4 CHEMICAL DEGREDATION 
7.4.4.1 Solvolysis 
Solvolysis is the decomposition of the active drug with the solvent that is present. 
When water is the solvent, the process is called hydrolysis. The most common solvolysis 
reactions involve liable carbonyl compounds such as esters and . - lactams 
(see Table 1 ). Aspirin is hydrolyzed to acetic acid and salicylic acid in the presence 
of moisture, but in a dry environment this hydrolysis is negligible. The most signifi - 
cant catalysts of hydrolysis are adverse pH and specifi c compounds (e.g., dextrose 
and copper or other metal ions).The rate of hydrolysis is pH and temperature 
dependent. There is a generalized rule of thumb that with each 10 ° C increase in 
temperature, the rate of the reaction increases exponentially. The actual factor of 
rate increase depends on the activation energy of the particular reaction, with the 
activation energy being a function of the specifi c reactive bond and the drug 
formulation. 
Hydrolysis leads to a decrease in active drug and an increase of decomposition 
products. The effect of this change on reaction rate depends on the order of the 
TABLE 1 Liable Functional Groups to Hydrolysis 
Functional Group Examples 
Esters Aspirin, procaine, alkaloids, estrone sulfate, dexmethasone 
sodium phosphate 
Lactams Penicillins, cephalosporins 
Amides Chloramphenicol, thiacinamide 
Lactones Pilocarpine 
Oximes Steroid oximes 
Imides Gluthethimide 
Malonic ureas Barbiturates 
Nitrogen mustards Melphalan 
Azo methines Benzodiazepines 

694 PHARMACEUTICAL PRODUCT STABILITY 
reaction. For zero - order reactions, the decomposition rate is independent of the 
drug concentration. Zero - order hydrolysis shows a lesser percentage of decomposition 
for higher concentration solutions that those of lower concentration. Unlike 
zero - order reactions, hydrolysis following fi rst - order kinetics shows a rate of change 
which is directly proportional to the concentration of the reacting ingredient. Therefore, 
changes in active ingredient concentration do not have an infl uence on the 
percentage concentration. 
Structural modifi cations of the drug may be employed to retard hydrolysis. The 
most frequently encountered hydrolysis reaction involves the ester. Substituents can 
have an effect on these reaction rates. Hansch and Taft [14] and Hammett [15] 
provide excellent reviews of the topic. Additionally, a compound may be stabilized 
by reducing its solubility. This can be accomplished by the addition of lipophilic 
substituents to side chains or aromatic rings. Often less soluble salts or esters have 
been employed to aid in product stability. 
The rate of hydrolysis is impacted by the ionic strength of the total concentration 
of dissolved electrolytes. In general, the rate constant of hydrolysis is directly proportional 
to the ionic strength of ions of like charge (drug cation and excipient 
cation) and inversely proportional to the ionic strength with oppositely charged 
ions. 
7.4.4.2 Oxidation 
Oxidation is a primary cause of instability and usually involves the addition of 
oxygen or the removal of hydrogen. Auto - oxidations are oxidative degradations 
usually mediated through reactions with atmospheric oxygen. Most auto - oxidations 
are free - radical reactions. A drug ’ s susceptibility to auto - oxidation can be determined 
by investigating its stability in a high - oxygen - tension atmosphere, usually 
40% oxygen. Results are compared against those under inert or ambient temperature 
[16] . A listing of several functional groups that are subject to oxidation can be 
found in Table 2 . 
Oxidations are catalyzed by acids, bases, pH values that are higher than the 
optimum, polyvalent metal ions, peroxides, hydroperoxides, and exposure to oxygen 
and ultraviolet (UV) illumination. These reactions may necessitate the use of antioxidant 
chemicals, inert atmospheres, and opaque packaging. Chelating agents 
TABLE 2 Functional Groups Subject to Oxidation 
Functional Group Examples 
Phenols Catecholamines, morphine 
Conjugated dienes Vitamin A 
Thioethers Chlorpromazine 
Nitrites Amyl nitrite 
Aldehydes Paraldehyde, fl avors 
Amines Clozapine 
Carboxylic acids Fatty acids 
Thiols Dimerceprol 
Ethers Diethylether 

CHEMICAL DEGREDATION 695 
added to water sequester heavy metals. Parenteral formulations should not 
come into contact with heavy metal ions during their manufacture, packaging, or 
storage [17] . 
Antioxidants are very effective in stabilizing products undergoing a free - radical 
mediated chain reaction. These products possess lower oxidation potentials than the 
active drug. Ideally, antioxidants are stable over a wide pH range and remain soluble 
in the oxidized form, colorless, and nontoxic. A listing of commonly used antioxidants 
can be found in Table 3 . 
Visual identifi cation of oxidation products is frequently attainable due to the 
introduction of conjugation, but these changes may not be visible in certain 
concentrations. 
7.4.4.3 Photolysis 
When molecules absorb energy, they proceed to a higher energy state where they 
release the energy in a chemical reaction to attain their ground energy state. When 
the energy of activation that is absorbed by the compound comes from a light 
source, the decomposition reaction is called photolytic. The activated species can 
release the energy either as light of a different frequency (fl uorescence) or by 
decomposition (photolysis). Exposure to room light or sunlight can lead to drug 
degradation by causing photo - oxidation and photolysis of covalent bonds. In susceptible 
compounds, usually those with . electrons, photochemical energy creates 
free - radical intermediates, which can perpetuate chain reactions. 
In general, drugs that absorb UV light below 280 nm undergo decomposition in 
sunlight while compounds that absorb above 400 nm have the potential to degrade 
in both sunlight and room light. Photo - induced reactions are common in steroids 
[18] . The photodegradation of sodium nitroprusside in aqueous solution remains a 
classic example of photolytic decomposition [19] . 
7.4.4.4 Dehydration 
Dehydrations are chemical reactions that involve the loss of water. The acid - 
catalyzed dehydration of tetracycline yields the toxic epianhydrotetracycline [20] . 
The physical dehydration of theophylline hydrate and ampicillin trihydrate leads to 
a change of the crystalline structure of the drug [21] . 
TABLE 3 Common Antioxidants 
Aqueous Systems Oil - Based Systems Chelating Agents 
Sodium sulfi te Ascorbyl palmitate Ethylenediamine tetraaceticacid 
(EDTA) 
Sodium metabisulfi te Hydroquinone Dihydroethylglycine 
Sodium bisulfi te Propyl gallate Citric acid 
Sodium thiosulfate Nordihydroguaiaretic acid Tartaric acid 
Ascorbic acid Butylated hydroxytoluene Gluconic acid 
Butylated hydroxyanisole 
. - Tocopherol 
Saccharic acids 

696 PHARMACEUTICAL PRODUCT STABILITY 
7.4.4.5 Racemization 
Racemization is the process of changing from an optically active compound into a 
racemic mixture. Pfeiffer provided one of the earliest discussions on the importance 
of stereospecifi city in drug action [22] . 
The most widely known drugs that undergo racemization are tetracycline, epinephrine 
[23] , pilocarpine [24] , and ergotamine. In tetracycline, the reaction occurs 
rapidly when the dissolved drug is exposed to a pH greater than 3, resulting in a 
steric rearrangement of the dimethylamino group [25] . 
Generally, racemization follows a fi rst - order reaction rate and is dependent on 
temperature, solvent system, catalysts, and the presence or lack of light. Resonance 
stabilization through substituents adjacent to the asymmetric center tends to accelerate 
racemization. 
7.4.5 PHYSICAL DEGRADATION 
7.4.5.1 Polymorphism 
Many pharmaceutical solids can exist in different physical forms. Polymorphism is 
often characterized as the ability of a drug substance to exist as two or more crystalline 
phases that have different arrangements and/or conformations of the molecules 
in the crystal lattice [26] . Polymorphic forms of a drug substance can have different 
chemical and physical properties, including melting point, chemical reactivity, apparent 
solubility, dissolution rate, optical and mechanical properties, vapor pressure, 
and density. These properties can have a direct effect on the ability to process and/or 
manufacture the drug substance and the drug product as well as on drug product 
stability, dissolution, and bioavailability. Thus, polymorphism can affect the quality, 
safety, and effi cacy of the drug product [27] . Amorphous solids consist of disordered 
arrangements of molecules and do not possess a distinguishable crystal lattice. Solvates 
are crystalline solid adducts containing either stoichiometric or nonstoichiometric 
amounts of a solvent incorporated within the crystal structure. If the 
incorporated solvent is water, the solvates are also commonly known as hydrates. 
Drug exposure to changes in temperature, pressure, humidity, and pulverization 
during granulation, milling, and compression may lead to polymorphic phase conversion. 
The extent of conversion generally depends on the relative stability of the 
polymorphs, kinetic barriers for phase conversion, and applied stress [28] . Sulfonamides, 
barbiturates, and steroids are known for their propensity to form polymorphs 
[29] . 
7.4.5.2 Vaporization 
Several drugs and adjuvants possess high vapor pressures at room temperature that 
their vaporization exists as a route of drug loss. Low - molecular - weight alcohols and 
“ aromatics ” used for fl avoring and aroma may be lost through vaporization. Nitroglycerine 
is the most frequently cited example of loss on vaporization. The FDA 
issued a special regulation governing the types of containers that may be used for 

dispensing nitroglycerine tablets [30] . The addition of macromolecules such as 
microcrystalline cellulose allows for the preparation of a stabilized nitroglycerine 
sublingual tablet [31] . 
7.4.6 MICROBIAL DEGRADATION 
Microorganisms could present a risk of infection or degradation of medicinal products. 
These organisms may proliferate during normal storage conditions or during 
patient use, especially in multidose preparations. Preparations containing water bear 
the greatest risk of contamination. These products include solutions, suspensions, 
and emulsions as well as sterile multidose injections and ophthalmic preparations. 
Antimicrobial preservatives are used to prevent or inhibit the growth of microorganisms. 
Factors infl uencing the level of observable effi cacy include the chemical 
structure of the preservative, the physical and chemical characteristics of the pharmaceutical 
product, the concentration of the preservative, and the type and load of 
initial contamination. At no time should preservatives be used as an alternative to 
good manufacturing practice. 
7.4.7 STABILITY GUIDELINES AND REGULATIONS 
FDA guidelines for stability testing are outlined in the Code of Federal Regulations 
(CFR), Part 21, Section 211.166, under Current Good Manufacturing Practice for 
Finished Pharmaceuticals. The guidelines state that there must be a testing program 
designed to assess the stability characteristics of drug products. The results of such 
stability testing are to be used in determining appropriate storage conditions and 
expiration dates. This section of the Federal Register provides guidelines for sample 
size and test intervals, storage conditions for samples, and specifi c test methods 
[32] . 
Section 211.166 guidelines also require an adequate number of batches of each 
drug product tested for expiration date assignment. Accelerated studies, combined 
with basic stability information on the components, drug products, and container 
closure system may be used to support tentative expiration dates provided full shelf 
life studies are not available and are being conducted. Where data from accelerated 
studies are used to project a tentative expiration date that is beyond a date supported 
by actual shelf life studies, there must be stability studies conducted, including 
drug product testing at appropriate intervals, until the tentative expiration date 
is verifi ed or the appropriate expiration date determined. 
Additional guidelines outline the need for reserve samples [33] , expiration dating 
[34] , and laboratory recordkeeping [35] . 
7.4.8 ICH QUALITY GUIDELINES 
The International Conference on Harmonisation of Technical Requirements for 
Registration of Pharmaceuticals for Human Use (ICH) is a project that brings 
together the regulatory authorities of Europe, Japan, and the United States and 
ICH QUALITY GUIDELINE 697

698 PHARMACEUTICAL PRODUCT STABILITY 
experts from the pharmaceutical industry in the three regions to discuss scientifi c 
and technical aspects of product registration. The purpose is to make recommendations 
on ways to achieve agreement in the interpretation and application of technical 
guidelines and requirements for product registration to reduce the need to 
duplicate the testing carried out during the research and development of new medicines. 
The objective of such harmonization is a more economical use of human, 
animal, and material resources and the elimination of unnecessary delay in the 
global development and availability of new medicines while maintaining safeguards 
on quality, safety and effi cacy, and regulatory obligations to protect public 
health [36] . 
The guidelines provide a breath of recommendations from stability testing protocols, 
including temperature, humidity, and trial duration [Q1A(R2)]; basic testing 
protocols required to evaluate the light sensitivity and stability of new drugs and 
products (Q1B); and stability testing for new formulations of already approved 
medicines (Q1C). A description of the various ICH guidelines for stability can be 
found in Table 4 . 
The ICH also provides guidelines in analytical validation [Q2(R1), impurities 
(Q3 series), pharmacopeias (Q4 series), quality of biotechnological products (Q5 
TABLE 4 ICH Document Codes and Guidelines for Stability 
ICH Code Title Description 
Q1A(R2) Stability Testing of New 
Drug Substances and 
Products 
Stability testing protocols including 
temperature, humidity, and trial duration 
Q1B Stability Testing: 
Photostability Testing of 
New Drug Substances 
and Products 
Basic testing protocol required to evaluate 
light sensitivity and stability of new drugs 
and products 
Q1C Stability Testing for New 
Dosage Forms 
Extends main stability guideline for new 
formulations of already approved medicines 
and defi nes circumstances under which 
reduced stability data can be accepted 
Q1D Bracketing and Matrixing 
Designs for Stability 
Testing of New Drug 
Substances and Products 
General principles for reduced stability testing 
and provides examples of bracketing and 
matrixing designs 
Q1E Evaluation of Stability 
Data 
Explains possible situations where 
extrapolation of retest periods / shelf lives 
beyond real - time data may be appropriate; 
provides examples of statistical approaches 
to stability data analysis 
Q1F Stability Data Package for 
Registration Applications 
in Climatic Zones III and 
IV 
Besides proposing acceptable storage 
conditions for long - term and accelerated 
studies, gives guidance on data to cover 
situations of elevated temperature and / or 
extremes of humidity; referenced literature 
provides information on classifi cation of 
countries according to climatic zones 

series), specifi cations (Q6 series), good manufacturing practice (Q7), pharmaceutical 
development (Q8), and risk management (Q9)]. 
REFERENCES 
1. U.S. Pharmacopeia (USP) , Stability considerations in dispensing practice, USP 29/NF 24, 
USP, Rockville, MD, p. 3029 . 
2. Guillory , P. , and Poust , R. ( 2002 ), Chemical kinetics and drug stability , in Banker and 
Rhodes , Eds., Drugs in the Pharmaceutical Sciences , Vol. 121, Marcel Dekker , New York , 
pp. 139 – 166 . 
3. Lintner , C. C. ( 1973 ), Quality Control in the Pharmaceutical Industry , Vol. 2, Academic , 
New York , p. 141 . 
4. Connors , K. A. , Amidon , G. L. , and Stella , J. V. ( 1986 ), Chemical Stability of Pharmaceuticals 
, 2nd ed., Wiley , New York. 
5. McMinn , C. S. , and Lintner , C. J. ( 1973 , May), paper presented at the American Pharmaceutical 
Association Academy of Pharmaceutical Sciences Meeting, Pharmaceutical Technical 
Section, Chicago, IL. 
6. Lachman , L. , DeLuca , P. , and Akers , J. J. ( 1986 ), Kinetic principles and stability testing , 
in Lachman , L. , Lieberman , H. A. , and Kanig , J. L. , Eds., The Theory and Practice of 
Industrial Pharmacy , 3rd ed., Lea & Febiger , Philadelphia , p. 766 . 
7. Fung , H. - L. , and King , S. - Y. ( 1983 ), in Pharmaceutical Technology Conference ’ 83 Proceedings 
, Aster Publishing, Springfi eld, OR. 
8. Pikal , M. J. , Lukes , A. L. , and Lang , J. E. ( 1977 ), J. Pharm. Sci. , 66 , 1312 . 
9. Woolfe , J. , and Worthington , H. E. C. ( 1974 ), Drug Dev. Commun. , 1 , 185 . 
10. Wang , W. ( 1999 ), Instability, stabilization, and dormulation of liquid protein pharmaceuticals 
, International Journal of Pharmaceutics , 185 , 129 – 188 . 
11. Gu, L.C. , Erdos , E. A. , Chang , H.-S. (1991), Pharm. Res . 8 , 485 . 
12. (a) Center for Drugs and Biologics, U.S. Food and Drug Administration (FDA) ( 1987 , 
Feb.), Guideline for submitting documentation for the stability of human drugs and Biologics, 
FDA, Rockville, MD. (b) Center for Drug Evaluation and Research and Center 
for Biologics Evaluation and Research, U.S. Department of Health and Human Services, 
FDA ( 1998 , June), Guidance for industry: Stability testing of drug substances and drug 
products: Draft guidance, FDA, Rockville, MD. 
13. U.S. Pharmacopeia (USP) , Pharmaceutical stability, USP 29/ NF 24, USP, Rockville, MD. 
p. 2994 . 
14. Hansch , A. L. , and Taft , R. E. ( 1991 ), A survey of Hammett substituent constants and 
resonance and fi eld parameters , Chem. Rev. , 91 , 165 . 
15. Hammett , L. P. ( 1970 ), Physical Organic Chemistry , 2nd ed., McGraw - Hill , New York . 
16. Johnson , D. M. , and Gu , L. C. ( 1998 ), Autooxidation and antioxidants , in Swarbrick , J. , 
and Boylan , J. C. Eds., Encyclopedia of Pharmaceutical Technology , Marcel Dekker , New 
York , pp. 415 – 449 . 
17. Vadas , E. B. ( 2000 ), Stability of pharmaceutical products , in Remington: The Science and 
Practice of Pharmacy , 20th ed, Lippincott Williams and Wilkins , Baltimore, MD , p. 989 . 
18. Ogata , M. , Noro , Y. , Yamada , M. , Tahara , T. , and Nishimura , T. ( 2000 ), Photo - degradation 
products of methylprednisolone sulphanate in aqueous solution — Evidence of a 
bicyclo[3.1.0]hex - 3 - en - 2 - one intermediate , J. Pharm. Sci. , 87 ( 1 ), 91 – 95 . 
19. Hauser , U. , Oestreich , V. , and Rohrweck , H. D. ( 1977 ), On optical dispersion in transparent 
molecular systems , Zeits. Phys. A , 280 , pp. 17 – 25 . 
REFERENCES 699

700 PHARMACEUTICAL PRODUCT STABILITY 
20. Yuen , P. H. , and Sokoloski , T. D. , ( 1977 ), Kinetics of concomitant degradation of tetracycline 
to epitetracycline, anhydrotetracycline, and epianhydrotetracycline in acid phosphate 
solution , J. Pharm. Sci. , 66 ( 11 ), 1648 – 1650 . 
21. Shefter , E. , Fung , H. - L. , and Mok , O. ( 1973 ), Dehydration of crystalline theophylline 
monohydrate and ampicillin trihydrate , J. Pharm. Sci. , 62 , 791 . 
22. Pfeiffer , C. C. ( 1956 ), Optical isomerism and pharmacological action, a generalization , 
Sci. , 123 , 3210, 29 – 31 . 
23. Schroeter , L. C. , and Higuchi , T. ( 1958 ), Racemization of epinephrine , J. Am. Pharm. 
Assoc. (Baltimore) , 47 , 6, 426, 30 . 
24. Nunes , M. A. , and Brochmann - Hanssen , E. ( 1974 ), Hydrolysis and epimerization kinetics 
of pilocarpine in aqueous solution , J. Pharm. Sci. , 63 , 716 . 
25. Sheberstova , N. V. , Perel ’ son , M. E. , and Kuzovkov , A. D. ( 1975 ), Study of the epimerization 
of tetracycline by the NMR method , Chem. Nat. Comp. , 10 ( 1 ), 61 – 65 . 
26. Haleblian , J. , and McCrone , W. ( 1969 ), Pharmaceutical applications of polymorphism , J. 
Pharm. Sci. , 58 , 911 . 
27. U.S. Department of Health and Human Services Food and Drug Administration Center 
for Drug Evaluation and Research (CDER), U.S. Food and Drug Administration (FDA) 
( 2004 , Dec.), Guidance for industry ANDAs: Pharmaceutical solid polymorphism chemistry, 
manufacturing, and controls information, FDA, Rockville, MD. 
28. Vippagunta , S. R. , Brittain , H. G. , and Grant , D. J. W. ( 2001 ), Crystalline solids , Adv. Drug 
Deliv. Rev. , 48 , 3 – 26 . 
29. Kuhnert - Brandstatter , M. ( 1971 ), Thermomicroscopy , in The Analysis of Pharmaceuticals , 
Pergamon , Oxford , pp. 37 – 42 . 
30. Federal Register , 37, 15959 ( 1972 ). 
31. Fung , H. - L. , Yap , S. K. , and Rhodes , C. T. ( 1974 ), Development of a stable sublingual 
nitroglycerin tablet. I. Interaction of nitroglycerin with selected macro - molecules , 
J. Pharm. Sci. , 63 , 1810 . 
32. Code of Federal Regulations , Title 21, Food and drugs, Part 211, Current good manufacturing 
practice for fi nished pharmaceuticals, Subpart I, Laboratory controls, Section 211.166, 
Stability testing. 
33. Code of Federal Regulations , Title 21, Food and drugs, Part 211, Current good manufacturing 
practice for fi nished pharmaceuticals, Subpart I, Laboratory controls, Section 211.170, 
Reserve samples. 
34. Code of Federal Regulations , Title 21, Food and drugs, Part 211, Current good manufacturing 
practice for fi nished pharmaceuticals, Subpart G, Packaging and labeling control, 
Section 211.137, Expiration dating. 
35. Code of Federal Regulations , Title 21, Food and drugs, Part 211, Current good manufacturing 
practice for fi nished pharmaceuticals, Subpart J, Records and reports, Section 211.194, 
Laboratory records. 
36. International Conference on Harmonisation of Technical Requirements for Registration 
of Pharmaceuticals for Human Use, available: http://www.ich.org/cache/compo/363 - 272 - 
1.html . 

701 
7.5 
ALTERNATIVE ACCELERATED 
METHODS FOR STUDYING DRUG 
STABILITY: VARIABLE - PARAMETER 
KINETICS 
Giuseppe Alibrandi 
Universit a di Messina, Messina, Italy 
Contents 
7.5.1 Introduction 
7.5.2 Theory 
7.5.3 Experimental 
7.5.3.1 Computer Simulation 
7.5.3.2 Devices to Obtain Variable - Parameter Conditions 
7.5.3.3 Analytical Instruments 
7.5.3.4 Software for Processing Experimental Data 
7.5.4 Examples of Variable - Parameter Kinetic Experiments 
7.5.4.1 Variable - Temperature Kinetic Experiments 
7.5.4.2 Variable - Concentration Kinetic Experiments 
7.5.4.3 Variable - Ionic - Strength Kinetic Experiments 
7.5.5 Conclusions 
References 
7.5.1 INTRODUCTION 
Studies on drug stability have a central role in physicochemical profi ling [1 – 6] . It is 
important to know (i) how long the substrate maintains its chemical identity in 
various environments carrying out its therapeutic action and (ii) the pathway followed 
when it degrades. Both sets of information are vitally important issues in the 
pharmaceutical industry. The fi rst is necessary to decide whether to continue to 
investigate the molecule of interest or not: An unstable drug can reduce its effi ciency 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

702 VARIABLE-PARAMETER KINETICS 
to a little percentage and can generate products with undesired effects; the second 
is necessary to understand whether it could be possible to fi nd a remedy or not, for 
example synthesizing a new drug with a similar structure and reduced reactivity or 
devising a suitable carrier able to stabilize the molecule and deliver it to the desired 
target. 
Temperature, pH, ionic strength, concentration of a metal ion, and other environmental 
parameters infl uence a chemical reaction and varying their values can signify 
drastic changes, depending on the case, on the rate, mechanism, or direction of the 
reaction. For this reason quantitative studies on the effects of physical parameters 
on reactivity often take a very long time in this kind of research. 
The usual way consists of isolating and characterizing the reaction of interest and 
then determining the rate constant under fi rst - or pseudo - fi rst - order conditions for 
different values of the involved parameters. The result is a polydimensional representation 
of the specifi c rate as a function of these parameters that can be of great 
use for practical applications and, particularly, constitutes the experimental structure 
on which to base the investigation directed to clarify the reaction mechanism. 
In fact, the dependence of kobs on parameters is a consequence of their infl uence on 
the rate - determining step and fi tting kinetic data to suitable mathematic models 
gives information that enables one to formulate a reaction scheme, discriminate 
between kinetically equivalent mechanisms, and determine the values of elementary 
kinetic constants [1 – 4, 7, 8] . 
Many examples are present in the scientifi c literature underlining the effort in 
producing kinetic data [9 – 11] . The Edwards historical study that started the investigation 
on the mechanism of the hydrolysis of aspirin required hundreds of kinetic 
experiments [12, 13] . Several examples are reported by Carstensen [1] in his review 
on the subject where, beside the large space dedicated to the determination of the 
pH – rate profi le, the effect of temperature, ionic strength, buffer concentration, and 
dielectic constant on the stability of drugs was treated. 
Physicochemical profi ling concerns the determination of some basic molecular 
properties of pharmaceutical interest that can support the successive clinical 
studies and can help a rapid identifi cation and elimination of compounds with 
unsuitable physicochemical and pharmacokinetic properties. New accelerated 
methods for physicochemical profi ling have assumed particular importance in recent 
years because of the great amount of new chemical entities coming from modern 
synthetic strategies. Thousands of compounds have to be profi led each year 
and traditional technologies cause a bottleneck in drug development. Methods 
have been developed for high - throughput physicochemical profi ling for solubility, 
permeability, p Ka , lipophilicity, stability, and integrity, and a great interest from 
pharmaceutical industries and instrument maker groups is reserved for this subject 
[5, 6] . 
Among physicochemical properties, the stability of drug candidates is receiving 
increasing attention. Unfortunately, its evaluation very often requires a lot of experimental 
time and profi ling of many molecules would be nearly impossible without 
new, computer - aided methods. Any effort useful to facilitate this fi rst part of pharmaceutical 
investigation is appreciated because it can be converted to a considerable 
lowering of the research cost. 
Here a panoramic picture of the effort directed to accelerating the characterization 
of the reactivity in solution of molecular candidates is shown. In particular, 

THEORY 703 
variable - parameter kinetics (VPaK) is considered, a potential new way to save time 
and chemicals while studying drug stability. 
7.5.2 THEORY 
Consider a molecule A degrading in products, 
A P > (1) 
the rate law, operating under fi rst - or pseudo - fi rst - order conditions, is given by Equation 
(2) , where C is the molar concentration of the substrate and k obs (Par 1 , Par 2 , . . . ) 
is the observed rate constant function of various parameters: 
. = dC 
dt 
k C obs 1 2 Par , Par , . . .) ( 
(2) 
In the usual kinetic experiments all these parameters must be rigorously 
constant so that the experimental C – t data, that is, the kinetic profi le obtained 
following the change in concentration with time, can be easily fi tted to the 
differential form of the fi rst - order equation (2) or to its integrated, exponential 
[Equation (3) ] or logarithmic [Equation (4) ], forms to obtain the optimized values 
of k obs : 
C C k t = .( ) 0 exp obs (3) 
ln ln obs C k t C = . + 0 (4) 
The dependence of k obs on one parameter (temperature T , pressure P , pH, ionic 
strength I , etc.), that is, the k obs (Par i ) profi le, can be obtained by carrying out several 
experiments for different values of that parameter. This yields a set of k obs – Par i data 
and, having an analytical model describing this dependence, by a second step fi tting 
treatment, the terms regulating such dependence can be obtained. For example, for 
the parameter temperature ( T ) the dependence function can be the Eyring equation 
(5) and the terms regulating the dependence, the entropy of activation . S ‡ and the 
enthalpy of activation . H ‡ ( k = Boltzmann ’ s constant, h = Planck ’ s constant, R = 
gas constant) [1, 7, 8] : 
k 
kT 
h 
S
R 
H 
RT obs exp exp = .. 
.. 
. .. 
.. 
. . ‡ ‡ 
(5) 
The Eyring equation is usually used in the logarithmic form (6) to have a linear 
plot of kinetic data. The intercept gives the . S ‡ value and the slope gives the . H ‡ 
value: 
ln ln obs k
T 
k
h 
S
R 
H 
RT 
= + . 
. . ‡ ‡ 
(6) 

704 VARIABLE-PARAMETER KINETICS 
Figure 1 shows an example of this kind of study where 25 constant - temperature 
kinetic (CTK) experiments were necessary to fi nd the activation parameters for the 
decomposition of fi ve molecules. When the parameter is a concentration of a species 
present in the reaction environment (H + , OH . , Me n + , a nucleophile, a catalyst, etc.), 
the dependence function can be, for example, of the type shown in the equation 
k k obs OH OH = . [ ] (7) 
and the fi t of k obs – [OH . ] data, obtained in several experiments carried out at constant 
[OH . ], gives the term k OH (Figure 2 ). These dependence functions, where concentrations 
are concerned, can be of great complexity but they are very important 
information for the formulation of the reaction mechanism [1, 3, 7, 8] . 
When the parameter is the ionic strength the dependence function is given by 
the Bronsted–Bjerrum equation (8) [1, 7, 8, 15] , where k 0 is the rate constant at zero 
ionic strength, a is a constant (for water at 25 ° C, a = 0.50925 L 1/2 /mol 1/2 [16] ) and 
Z A Z B is the product of charges of the reacting species taking part in the rate - 
determining step: 
k k 
aZ Z 
I 
I 
obs 
A B = . + 
0 
2 
1 10 (8) 
Even in this case, traditionally, a logarithmic form is used [Equation (9) ] to treat 
kinetic data. The intercept of the linear plot (Figure 3 ) gives the value of k 0 and the 
slope the value of Z A Z B , and this is very useful information for ionic reactions to 
discriminate between kinetically equivalent mechanisms: 
log log obs A B k k aZ Z 
I 
I 
= + 
+ 0 2 
1 
(9) 
FIGURE 1 Eyring plots (solid lines) for reactions of homologous series of fi ve molecules 
obtained with kinetic experiments carried out at constant temperature (plain markers) (simulated 
data from ref. 14 ). 
1/T (K) 
In [(k/T) (s–1) K–1

THEORY 705 
Constant - parameter kinetic (CPaK) experiments are very time consuming. Nowadays, 
the use of personal computers and new - generation analytical instruments 
enables a step forward to be made in collecting kinetic data. 
VPaK [17] can be defi ned as that part of chemical kinetics concerning experiments 
carried out while varying the value of an environmental parameter in a controlled 
way. The aim is to obtain, in a single run, the dependence of the observed 
rate constant on that parameter in a time 10 – 100 times less compared to that usually 
FIGURE 2 k obs ([OH . ]) profi les (solid lines) obtained for homologous series of fi ve molecules 
by 25 kinetic experiments (plain markers) carried out at constant concentration of OH . 
(simulated data). 
[OH–] 
kobs (s–1) 
FIGURE 3 Br o nsted – Bjerrum plots (solid lines) obtained by constant - ionic - strength 
kinetic experiments (plain markers) for homologous series of fi ve compounds having the 
same mechanism (same values of Z A Z B = +1) and different intrinsic reactivity (different 
values of k 0 ) (simulated data). 
log (k) (s–1) 
I /(1+ I )

706 VARIABLE-PARAMETER KINETICS 
spent using traditional methods. This is possible by fi tting the kinetic profi le obtained 
during the reaction to a suitable mathematic model analytically describing it. 
VPaK can be considered a generalization of nonisothermal kinetics [18, 19] , 
largely applied in thermal analysis [20 – 25] for solid - state systems. Several examples 
of variable - temperature experiments concerning solution chemistry were reported 
in the past. Different approaches were used depending on the different scientifi c 
background of the authors (inorganic, organic, organometallic, and pharmaceutical) 
and on the instruments available at the time [14, 26 – 51] . Few examples of variable - 
concentration kinetic (VCK) experiments were reported [31, 35, 52, 53] . Just one 
example of variable - ionic - strength kinetic experiment has so far been reported [54] . 
Here a simple approach is described making a reference to important similar 
studies. 
The mathematical form describing a VPaK experiment is given by the equation, 
. = . 
dC 
dt 
k t C i i { [ ()]} obs Par Par Par 
(10) 
where C is the concentration of the monitored reacting species. In this equation k obs 
is not a constant but depends on a parameter (Par i ) varying with time. Then 
k obs [Par i ( t )] is a function of a function: k obs (Par i ) is the dependence function ( D ) 
describing the dependence of the rate constant on the parameter i and Par i ( t ) is the 
modulating function ( M ) showing the way the parameter changes with time while 
the other parameters are maintained constant. The experimental kinetic profi le 
contains in each point information about the value of the rate constant at that time, 
as can be seen writing Equation (10) in the form of Equation (11) . The ratio of the 
derivative of the concentration on the concentration itself gives k obs . Then, the 
k obs (Par i ) profi le can be obtained in a single kinetic run: 
. = . 
1 
C 
dC 
dt 
k ti i { [ ()]} obs Par Par Par 
(11) 
Furthermore, knowing the mathematical form of the dependence function, a single 
step fi tting to Equation (10) , or its integrated form, gives the terms regulating such 
dependence. For example, in variable - temperature kinetic (VTK) experiments the 
dependence function, as stated above, is the Eyring equation and the terms are the 
activation parameters . S ‡ and . H ‡ . If the temperature changes in a linear way, that 
is, M: T = T 0 + . t , where T 0 is the initial temperature and . is the temperature gradient, 
the model changes to the equation 
. = 
+ .. 
.. 
. 
+ 
.. . 
.. . 
dC 
dt 
k T t 
h 
S
R 
H 
R T t 
C 
( ) 
( ) 
0 
0 
. 
. 
exp exp 
. . ‡ ‡ 
(12) 
or to the relative integral form 
C C 
k T t 
h 
S
R 
H 
R T t 
d 
t 
= . + .. . 
.. . 
... 
. 
+ 
.. . 
.. .
. 0 
0 
0 
0 
exp exp exp 
( ) 
( ) 
. 
. 
. . ‡ ‡ 
t.. . 
(13) 

The fi t of C – t experimental data, that is, the kinetic profi le obtained in a single VTK 
run, to one of these equations gives the optimized values of . S ‡ and . H ‡ . 
For an experiment concerning the investigation of the dependence on the concentration 
of a catalyst Y, [Y], the molar concentration of the catalyst could be 
varied in a linear way, that is, M : [Y] = . t , where . is the concentration gradient. If 
the dependence function is known, for example, D: k obs = k Y [Y], the mathematical 
model will assume the forms 
. = dC 
dt 
k tC Y. 
(14) 
C C k t = .( ) 0 
2 1
2 
exp Y. 
(15) 
A single fi t of the C – t experimental data to one of these equations gives the optimized 
value of the term k Y . 
When the dependence function is not known, Equation (16) has to be used: 
. = 1 
C 
dC 
dt 
k t obs Y]( {[ )} 
(16) 
and the k obs ([Y]) profi le can be obtained by dividing the derivative of the kinetic 
profi le to the profi le itself. This method is particularly important when no kind 
of information is available on the dependence of the rate constant on concentrations 
of species present in the reaction environment and taking part in the rate - 
determining step. The VCK experiment, in this case, gives the empirical relation that 
enables a hypothesis on the dependence function to be formulated. 
7.5.3 EXPERIMENTAL 
Some basic points have to be discussed to conveniently carry out VPaK experiments: 
1. Computer simulation 
2. Devices to obtain variable - parameter conditions 
3. Analytical instruments 
4. Software for processing experimental data 
All these points involve in some way the use of a personal computer. It is important 
to underline that it is the availability of fast and inexpensive computers that 
enables, nowadays, VPaK to be applied, providing a real saving of time and acceptable 
confi dence in the obtained data. 
7.5.3.1 Computer Simulation 
Computer simulation can be very useful in the phase of projecting a VPaK experiment. 
VPaK profi les are often diffi cult to imagine compared to CPaK ones where 
EXPERIMENTAL 707

708 VARIABLE-PARAMETER KINETICS 
the rate always decreases exponentially with time [Equation (3) ]. In VPaK the reaction 
rate is given at each time t by the product of two factors: kobs [Par i ( t )] and C . 
The concentration always decreases with time but kobs [Par i ( t )] can vary in many ways 
depending on the form of both the dependence function and the modulating function. 
The search for good experimental parameters could signify a waste of precious 
time. Using a personal computer it is possible in a few minutes to have an acceptable 
profi le, just obtaining some preliminary data from tests that are always carried 
out before a kinetic study [55] . Any software able to plot a function and, when necessary, 
to evaluate numerically a derivative or an integral can be used. We found 
the MicroMath SCIENTIST program [56] to be very versatile, very easy to use, and 
enabling us to use the differential form of the rate equation (10) without the necessity 
of integrating it. With this tool, giving as input the dependence function and the 
modulating function, the kinetic profi le can be obtained immediately for every value 
of the experimental parameters. Figure 4 shows a typical list of a SCIENTIST 
program where the modulating function, dependence function, and fi rst - order differential 
equation are indicated by arrows for a VTK simulation. 
Figure 5 shows a profi le obtained imposing as experimental data the values T0 = 
298 K and . = 0.0033 K/s. It is also possible to obtain in a single simulation several 
profi les relative to different values of terms conditioning them. Figure 6 , for example, 
shows a multiparametric simulation of VCK profi les for a reaction of basic hydrolysis 
with dependence function kobs = kOH [OH . ] and a modulating function [OH . ] = 
.t , with 10 . 3 . . M s . 1 = 1, 2.5, 4, 5.5, 7. 
7.5.3.2 Devices to Obtain Variable - Parameter Conditions 
A device is necessary to obtain VPaK conditions inside the reaction vessel. The 
parameter must change homogeneously and coherently with the modulating function 
and without altering any other factors important for the reaction or its analytical 
monitoring. 
The parameter can change in a vessel being part of the analytical instrument, for 
example, an ultraviolet – visible (UV – Vis) spectrophotometric cell [39, 41, 45, 14, 47, 
48] , an infrared (IR) cell [42, 46] , or a fl uorometer cell [45, 51] , or a polarimetric 
tube [27, 49] . It can change in a reactor vessel where the analytical signal can be 
read in some way, for example using an optical fi ber cell for spectrophotometry 
[52 – 54] or a conductometric cell [16, 34, 40] . Another possibility is to transport the 
solution from the reaction vessel to the analytical instrument by a peristaltic pump 
[38] . When altenative ways are not practicable, samples can be taken at suitable time 
intervals and analyzed apart [29, 31, 35, 39, 43, 50] . 
Any system satisfying the conditions cited above can be useful as VPaK devices. 
For VTK programmable thermostatic baths (with the possibility of external circulation 
) and temperature programmers using the Peltier effect are very convenient. It 
is always necessary to homogenize the temperature and the composition of the 
solution with good stirring, monitoring the temperature by measuring it inside the 
reaction vessel, possibly memorizing it in a computer. A linear increasing temperature 
is suffi cient and it is enough easy to realize, but other modulating functions, for 
various reasons, have been tested [33, 35, 47] . However, for future particular applications, 
the use of a computer with suitable software can easily generate T ( t ) profi les 
having various shapes. 

The thermal expansion of the reaction solution and vessel, because of their infl uence 
on the concentration of the reacting species, can be a problem for coherence 
to the kinetic model. The best way to solve it is to limit the range of temperature 
to 20 – 25 ° C; otherwise corrections are required that make the procedure less simple. 
Furthermore, a limited thermal excursion assures one about the constancy of the 
activation parameters during the experiments. 
For VCK autoburettes releasing concentrated solution of species having some 
effect on the reaction are usually used. The concentration gradient . (in molars per 
second ) inside the vessel is given by the relation 
FIGURE 4 List of SCIENTIST model used to simulate VTK experiment. The fi rst three 
arrows show, respectively, the modulating function, dependence function, and fi rst - order 
kinetic equation. The last two arrows show experimental parameters T0 and . . 
EXPERIMENTAL 709

710 VARIABLE-PARAMETER KINETICS 
. = gM 
V0 
(17) 
where g (in liters per second) is the releasing rate, M (moles per liter) the molar 
concentration of the added solution, and V 0 (liters) the initial reaction volume. The 
chemical homogeneity is assured by good stirring. A problem can arise when the 
added percentage volume is large and cannot be neglected in the mathematical 
model. Computer simulation can help to fi nd suitable conditions [55] . Some attempts 
have been made to introduce a correction in the kinetic equation [31] . The species 
released into the reaction environment must infl uence the reaction exclusively in 
the way described by the dependence function. 
FIGURE 5 Simulated kinetic profi le for VTK experiment obtained by SCIENTIST model 
in Figure 4 . 
t (s) 
C (mol/L) 
. 
FIGURE 6 Simulated VCK profi les obtained by SCIENTIST multiparametric model. 
D: k obs = k OH [OH . ], M : [OH . ] = . t. 10 . 3 . . M s . 1 = 1, 2.5, 4, 5.5, 7. 
t (s) 
C (mol/L)

Equation (10) is generally valid for any parameter provided that a suitable apparatus 
is available to generate the relative experimental conditions. Some cases are 
very diffi cult to manage. For example, to our knowledge, variable - pressure kinetic 
experiments have so far not been reported. 
7.5.3.3 Analytical Instruments 
Theoretically, any analytical instrument useful to follow a reaction in a traditional 
constant - parameter kinetic experiment can be used for VPaK experiments. Obviously, 
some practical applications can be problematic. For example, while it is very 
easy to change the temperature inside a UV – Vis spectrophotometric cell, changing 
the concentration of a species inside a nuclear magnetic resonance (NMR) tube can 
be very complicated. 
Following a physical property of the reacting species instead of their concentration 
leads to a series of consequences. First, while Equations (2) , (3) , and (4) used 
in CPaK experiments change respectively to the equations 
. 
. 
= . . 
. 
d 
dt 
k 
( ) 
( ) 
. . 
. . obs 1 2 Par , Par , . . .)( 
(18) 
. . . . . = . . . . ( ) ) 0 exp( obs k t (19) 
ln ln obs ( ) ( ) . . . . . =. + . . . k t 0 (20) 
Equation (10) and its integral form change respectively to the equations 
. 
. 
= . . 
. . 
d 
dt 
k ti 
( ) 
{ [ ()]} ( ) 
. . 
. . obs Par Par Par 1 
(21) 
. . . . = . ... 
.. 
+ . . . ( ) [ ()] 0 
0 
exp Par obs 
t 
i k tdt 
(22) 
where . is a physical quantity related to the compounds involved in the reaction at 
time t , . 0 and . . its values, respectively, at the start and at the end of the reaction. 
The parameter . can be the absorbance, the conductance, the optical rotation, the 
area of an NMR peak, and so on. 
The more the precision of the instrument and the more the points for the 
time unit in the acquired profi le, the better the result of the fi tting of experimental 
data. For this reason instruments with a low measure error and connectable to a 
computer for the automatic and continous aquisition of data are very much 
prefered. The UV – Vis spectrophotometer is by far the most used instrument in 
chemical kinetics. It has a good sensitivity and a good control of the temperature. 
It is connected or easily connectable to a computer and is available nearly everywhere. 
The absorbance has a very low dependence on the temperature so that, 
in the used temperature range, its variation can be neglected during the VTK 
experiments. 
For VCK experiments the usual 10 - mm cuvette may be not suitable as a reaction 
vessel so an external reactor can be used where the absorbance can be read by an 
EXPERIMENTAL 711

712 VARIABLE-PARAMETER KINETICS 
UV – Vis sensor or using a fl ux cell. A problem can arise by adding a reagent absorbing 
in the same range where the reacting species or products absorb. When this is 
unavoidable , a correction of the kinetic profi le can be made by a blank scan [53] . 
Unfortunately not all the molecules absorb so that other physical properties and 
other analytical instruments have to be used. 
For reactions involving change in optical rotation, a professional polarimeter is 
a good solution. It has high sensitivity and can be connected to a computer for the 
automatic acquisition of the analytical data. The changing and monitoring of temperature 
can be easily done [49] . The effect of temperture on optical rotation is not 
large. Examples of VCK experiments using this instrument have not been reported, 
but they could be realized, for example, using an external reactor connected to a 
fl ux cell. 
Infrared [42, 46] , fl uorimeters [45, 51] , and conductometers [16, 34, 40] have been 
used with success. Gas – liquid chromatography (GLC) and high - performance liquid 
chromatography (HPLC) are very much used in pharmaceutical studies, for both 
analytical investigations and kinetic studies. Unfortunately they are not the ideal 
for VPaK experiments. The kinetic profi le obtained is made by a few points (each 
one requiring a time - consuming cromatographic analysis) and the fi t to the model 
can lead to a large error in the evaluation of the terms. Nevertheless, examples are 
reported and the results are acceptable [38, 39, 43, 50] . If large numbers of routine 
measurments would require it, suitable automation in the analysis [38] and software 
for the fast processing of the computer acquired chromatograms could be devised. 
7.5.3.4 Software for Processing Experimental Data 
When a VPaK experimental profi le has been produced and stored on a computer, 
suitable software is needed for its quick processing. While in the past the treatment 
of constant - temperature kinetic data was easy to do, even without computers, in 
VPaK the use of a computer is of fundamental importance for both the complexity 
of the mathematical model, which does not require time - consuming calculations, 
and confi dence in the obtained results. Algorithms are necessary to fi t the experimental 
data to Equation (21) or (22) in their particular form depending on the 
investigated parameter and the used analytical instrument. Algorithms are also 
necessary for the evaluation of the derivative and sometimes the integral [e.g., in 
Equation (11) it is not possible to solve the integral in terms of elementay functions 
so it must be evaluated numerically]. Any language can be used for a personalized, 
home - made program, but application programs are commercially available to facilitate 
the work. The MicroMath SCIENTIST, cited for simulation for the search of 
the experimental conditions, is an example. The same models used to simulate VPaK 
profi les are used to fi t the obtained data with the difference that the terms given 
for the simulation are in this case the input values of the terms to be optimized. 
SCIENTIST uses a Powell modifi ed Marquadt for the fi tting and the (default) 
Episode method to solve the differential equation [57] . 
The modulating function is always known because the value of the parameter is 
imposed by a program and/or monitored by a sensor. The dependence function is 
sometimes known and sometimes not. For example, for the parameter temperature 
the dependence function is always known (Arrhenius equation, Eyring equation). 
For the parameter concentration the dependence function can be known in studies 

where a mechanism has already been proposed or guessed and a series of similar 
substrates have to be quantitatively investigated in detail. It can be unknown for 
new studies. In these cases Equation (21) can be written in the form 
. 
. 
= 
. 
. 
1 
. . 
. d
dt 
k ti i { [ ()]} obs Par Par Par 
(23) 
to underline that the main calculation required is for the evaluation of the derivative 
of the kinetic profi le. The ratio of the derivative on . . . . , that is, the profi le minus 
its value at the end of the reaction, gives at each point the value of k obs . 
Some examples concerning pharmaceutical systems will be discussed. 
7.5.4 EXAMPLES OF VARIABLE - PARAMETER 
KINETIC EXPERIMENTS 
7.5.4.1 Variable - Temperature Kinetic Experiments 
Figure 7 shows a kinetic profi le relative to a VTK experiment concerning the racemization 
of ( . ) - adrenaline in acidic aqueous solution [49] . It was obtained polarimetrically 
following the optical rotation . . The modulating function used was T = 
T 0 + . t , where T 0 = 309.6 K and . = 0.001694 K/s, realized by a circulation of water 
coming from a programmed thermostatted bath. The temperature was read inside 
the polarimetric cell by a platinum resistor. Temperature and absorbance were 
automatically acquired by a computer connected to the instruments. The typical 
sigmoidal shape comes from the relation r = k obs [ T ( t )] C , that is, the reaction rate is 
given by the product of two terms: k obs [ T ( t )], which always increases for the increasing 
temperature, and C , which decreases continously during the reaction. For this 
reason the reaction is accelerated in its fi rst part but after the infl ection point the 
low concentration of C is overwhelming and the reaction decelerates until the rate 
FIGURE 7 Change in optical rotation (solid line) during racemization of ( . ) - adrenaline in 
aqueous solution (1 M HCl) carried out under variable - temperature conditions. M: T (K) = 
309.6 + 0.001694 t (dashed line). 
t (s) 
a (deg) 
T (K) 
. . 
EXAMPLES OF VARIABLE-PARAMETER KINETIC EXPERIMENTS 713

714 VARIABLE-PARAMETER KINETICS 
gets to zero at the end of the reaction (Figure 8 ). A direct fi tting of this profi le to 
Equation (24) , with . 0 , . . , . S ‡ , and . H ‡ the parameters to be optimized, gave 
the activation parameters ( . S ‡ = . 27 ± 1 J/K · mol, . H ‡ = 95 ± 1 kJ/mol, R 2 = 
0.99999). Their values were identical to those obtained by fi ve traditional VTK 
experiments: 
. . . . 
. 
= . . ... 
... 
+ { 
. 
+ 
.
. . 
. 
. . ( ) ( ) 
( ) 
0 
0 
0 
0 
exp exp 
exp 
k
h 
S
R 
T t 
H 
R T t 
t . 
. 
‡ 
‡ 
. . 
}+ . dt . 
(24) 
Figure 9 shows the profi le relative to another VTK experiment carried out spectrophotometrically 
[45] . The reaction followed was the hydrolysis of aspirin [12, 13, 
59, 60] at pH 4.50. The temperature was controlled by a cell compartment thermostatted 
by a Peltier temperature programmer and measured by a platinum resistor 
inserted into the spectrophotometric cell. Magnetic stirring was assured by a suitable 
device. Lots of data points were memorized by a computer connected to the 
analytical instrument and easy processings, both differential and integral, were 
carried out. Values of activation parameters were in agreement with each other and 
in agreement with those obtained by comparative constant - temperature kinetic 
experiments carried out in the same conditions ( . S ‡ = . 115 ± 1 J/K · mol, . H ‡ = 69 
± 1 kJ/mol, R 2 = 0.99999). 
The sigmoidal profi le in Figure 10 is relative to the hydrolysis of aspirin carried 
out at pH 7.00 under variable - temperature conditions. In this case a fl uorometer 
was used as an analytical instrument irradiating the salicylic acid at 310 nm and 
recording the fl uorescence signal at 404 nm [45, 61] . A programmable thermostatted 
bath was used to apply a linear increasing temperature ( T 0 = 323.05 K, . = 1.631 . 
10 . 3 K/s) in the thermostatted fl uorometric cell compartment. The temperature was 
FIGURE 8 Change in reaction rate during racemization of ( . ) - adrenaline as obtained using 
Savitzky – Golay method [58] by derivative of VTK profi le reported in Figure 7 . 
. . 
t (s) 
r (deg/s)

read inside the cell and monitored and stored by a computer together with the fl uorescence 
intensity. 
A fl uorometer is more sensitive than a spectrophotometer and enables a very 
low concentration of substrate to be used. This can be convenient in the fi rst stage 
of pharmaceutical studies. Unfortunately the dependence of fl uorescence intensity 
on temperature cannot be avoided. A blank scan at the end of the reaction enables 
a temperature - independent signal [45] useful for VTK processing to be obtained. 
The results were consistent ( . S ‡ = . 109 ± 1 J/K · mol, . H ‡ = 71 ± 1 kJ/mol, R 2 = 
0.99999) with those obtained spectrophotometrically under both CTK and VTK 
conditions. 
Polarimeters, spectrophotometers, and fl uorimeters connected to the computer 
can store, during a VPaK experiment, hundreds or, when necessary, even thousands 
FIGURE 9 VTK profi le (solid line) obtained spectrophotometrically ( . = 298.5 nm) for 
hydrolysis of aspirin in water (pH = 4.50). M: T (K) = 304.36 + 3.647 . 10 . 4 t (dashed line). 
t (s) 
A (au) 
T (K) 
1.104 0 
1.2
1 
0.8 
0.6 
0.4 
0.2
0 
340 
335 
330 
325 
320 
315 
310 
305 
300 
2.104 3.104 4.104 5.104 6.104 7.104 8.104 
FIGURE 10 Change in fl uorescence intensity (solid line; 404 nm) during hydrolysis of 
aspirin (pH 7.00) in variable - temperature kinetic experiment. M: T (K) = 323.05 + 1.631 . 
10 . 3 t (dashed line). 
t (s) 
I (au)
T (K) 
. . . 
EXAMPLES OF VARIABLE-PARAMETER KINETIC EXPERIMENTS 715

716 VARIABLE-PARAMETER KINETICS 
of analytical points. This increases the fi tting of experimental data to the mathematical 
model or the evaluation of the derivative of the kinetic profi le. 
Figure 11 shows a VTK profi le obtained by HPLC analysis concerning the hydrolysis 
of aspirin at pH 7.00 and at a linear increasing temperature, with T 0 = 323.2 K 
and . = 1.631 . 10 . 3 K/s [50] . The experiment was carried out in a reaction vessel 
immersed in a programmable thermostatic bath. Samples of the reaction mixture 
were taken at suitable time intervals and injected into the liquid chromatography 
(LC) column. Few points were collected (27) to build it up because of the time 
required for the chromatographic separation. Nevertheless, the fi t to the relative 
mathematical model gave values of activation parameters similar to those obtained 
spectrophotometrically, although with greater statistical error ( . S ‡ = . 102 ± 8 J/ 
K · mol, . H ‡ = 73 ± 2 kJ/mol, R 2 = 0.9997). 
7.5.4.2 Variable - Concentration Kinetic Experiments 
The dependence on the concentration of species present in the reaction environment 
is certainly the most important in the context of mechanistic studies. A species 
can infl uence the reaction by taking part in it as a reagent or as a catalyst or simply 
by altering in some way the physicochemical character of the reaction environment: 
for example, a nucleophile in a nucleophilic substitution, an ionic metal in a reaction 
catalyzed by it, or a salt altering the ionic strength or molecules of solvent altering 
the dielectric constant. In the fi rst case the dependence function is the heart of the 
rate law. It indicates the species taking part in the rate - determining step and gives 
an idea of the way they interact with each other [1, 3, 7, 8] . For this reason the search 
for its form and for the values of the terms inside it requires a great deal of 
attention. 
A study on the nucleophilic substitution on the square planar complex trans - 
[Pt(PEt 3 ) 2 Cl 2 ], a substrate similar to cisplatinum and other compounds largely used 
as antitumoral agents [53, 62 – 65] , has been carried out spectrophotometrically while 
changing the concentration of the nucleophile with time. The reaction vessel was 
FIGURE 11 VTK profi le obtained by HPLC peak area (PA) of salicylic acid (plain circles) 
during hydrolysis of aspirin at pH 7.00. M: T (K) = 323.2 + 1.631 . 10 . 3 t (dashed line). Solid 
line is relative theoretical curve. 
t (s) 
PA 
PA 
T 
T (K) 
. .

immersed in a thermostatted bath, an autoburette released a suitable concentrated 
solution of nucleophile (thiourea, iodide, bromide, thiocianide), the absorbance 
was read by an optical fi ber cell, and good stirring was assured by an immersion 
stirrer. 
Pseudo - fi rst - order conditions were assured. In classic constant - concentration 
kinetic (CCK) experiments this means the use of a reactant in a high enough concentration, 
compared to that of the substrate, to neglect its consumption during the 
reaction. In this way the simple fi rst - order kinetic equation can be used for the data 
treatment. In VCK experiments these conditions have to be considered in a different 
way: The concentration of the species must vary with time in a way that deviation 
from the modulating function, caused by its consumption, can be neglected. Concentrated 
solutions able to create a suitable excess of reagent in the reaction environment 
can be easily prepared. In particular cases, when this is not possible, the 
consumption of the species can be inserted into the modulating function [53] . 
Figure 12 shows the variable - concentration kinetic profi le obtained for the reaction 
of the platinum complex with SCN . . The modulating function was [SCN . ] = . t , 
with . = 0.048 M/s. Even in this case the reaction is accelerated in the fi rst part of 
the kinetics, but this is caused by the increasing concentration of nucleophile. The 
dependence function for this reaction is given as 
k k k obs S Y Y] = + [ (25) 
where k S is a solvolytic constant and k Y is the direct attack constant for a generic 
nucleophile Y [62, 63] . Fitting the profi le to the model 
. = + .. 
dA 
dt 
k k t A A ( )( ) S Y. 
(26) 
gave the optimized values of k S and k Y (respectively, 0.92 . 10 . 4 s . 1 and 
0.351 M . 1 · s . 1 ) 
FIGURE 12 VCK profi le obtained spectrophotometrically (plain cirlce, selected points) 
and theoretical curve (solid line) (280 nm) for reaction trans - [Pt(PEt 3 ) 2 Cl 2 ] + 2 SCN . > trans - 
[Pt(PEt 3 ) 2 (SCN) 2 ] + 2 Cl . in methanol. M : [SCN . ] = 0.048 t M (dashed line); T = 303.2 K. 
t (s) 
A (au)
[SCN–] 
EXAMPLES OF VARIABLE-PARAMETER KINETIC EXPERIMENTS 717

718 VARIABLE-PARAMETER KINETICS 
Figure 13 shows the SCIENTIST list of the model used for the fi tting procedure 
of the kinetic profi le, where KS, KY, A0, and AI are input values to be optimized. 
The reaction of the same substrate with thiourea in variable - concentration conditions 
was also followed conductometrically with good results [16] . The reaction was 
carried out in a thermostatted reaction vessel. An autoburette was used to add a 
concentrated solution of thiourea. The conductance was read by means of a conductometric 
cell and acquired by a computer. 
Among the studies on the dependence of kobs on the concentration of species 
present in the reaction environment, in the pharmaceutical fi eld a particular space 
is dedicated to pH – rate profi les, that is, the dependence on [H + ] and/or [OH . ]. They 
can give a lot of information on the reaction mechanism and on the way to confront 
the instability [1, 3, 4] . Very often these studies require lots of kinetic experiments 
because the hydrogen ion concentration varies over 14 orders of magnitude. To 
FIGURE 13 SCIENTIST list of model used for processing of VCK profi le in Figure 11 . The 
fi rst three arrows show, respectively, the modulating function, dependence function, and fi rst - 
order kinetic equation. The last three arrows show the experimental parameters g, M , and 
V0 . 

delineate the pH – rate profi le of aspirin, for example, about 50 kinetic experiments 
are necessary. Considering different temperatures, comparative molecules, and other 
various effects, a study on homologous series can be very hard to carry out. 
With VCK experiments, a single run can be enough to obtain the pH – rate profi le 
of a reacting substrate. 
Figure 14 shows the variable - pH kinetic (VpHK) profi le obtained spectrophotometrically 
for the reaction of hydrolysis of aspirin with pH varying in the range 2 – 10 
at T = 342.5 K. The variable - concentration conditions were realized by adding a 
concentrated solution of NaOH (0.6 M ) to the thermostatted reaction vessel containing 
the aqueous solution of acetylsalicylic acid and a buffer composed of acetic 
acid (0.01 M ), fosforic acid (0.01 M ), and boric acid (0.01 M ). In this way an almost 
linear increase of pH was generated. The absorbance was read by an optical fi ber 
cell and stored in a computer. The pH was monitored by a pH sensor connected to 
a computer. 
This profi le is more complex than the others seen before. This is caused by the 
fact that, in the reaction rate equation, given by the product of k obs and the concentration 
of substrate, k obs [pH( t )] varies with the pH in an irregular way because of 
the particular shape of the pH – rate profi le. Figure 15 shows the derivative of the 
pH – rate profi le as obtained by the Savitzky – Golay method [58] . It gives an idea of 
the variation of the reaction rate during the VpHK experiment. There is an acceleration 
in the fi rst part, caused by the increase of k obs with the pH, followed by a 
deceleration for the stabilization of k obs and the continuous decrease of the substrate 
concentration. Then, the rate increases again for the increase of k obs for higher values 
of pH and after decreases again for concentration approaching zero. 
The ratio of the derivative of the profi le to A . A . gave, according to Equation 
(27) , the entire pH – rate profi le in a single scan (Figure 15 ). 
For such large ranges of pH, the dependence function can be very complex. For 
example, in the case of aspirin, changing the pH, four different mechanisms of reaction 
operate and the global rate constant requires several terms [3] . Nevertheless, 
once the pH – rate profi le has been obtained and a reaction mechanism formulated, 
FIGURE 14 Change in absorbance (solid line) during a VpHK experiment concerning 
hydrolysis of aspirin at T = 342.5 K, . = 298.5 nm. The dotted line shows the variation of pH. 
M : not known analytically. 
t (s) 
A (au) 
pH 
. . 
EXAMPLES OF VARIABLE-PARAMETER KINETIC EXPERIMENTS 719

720 VARIABLE-PARAMETER KINETICS 
FIGURE 15 Change in derivative of VpHK profi le reported in Figure 14 (solid line) and 
kobs [pH( t )] profi le as obtained by Equation (27) (dashed line). 
t (s) 
v (au/s)
kobs (s–1) 
. 
. 
. 
a dependence function can be inserted into Equation (27) and a global fi t can be 
done for the evaluation of the values of the elementary constants: 
. 
. 
= . . 
. 
d A A 
dt 
k t A A 
( ) 
{ [[ ]()]}( ) obs 
+ H (27) 
7.5.4.3 Variable - Ionic - Strength Kinetic Experiments 
Figure 16 shows the kinetic profi le for a variable - ionic - strength kinetic (VIK) experiment 
concerning the hydrolysis of indomethacin [54, 66, 67] at NaOH 0.005 M and 
T = 299.0 ± 0.1 K at a varying ionic strength I = I0 + .t with I0 = 0.005 M , due to the 
concentration of NaOH, and . = 1.77 . 10 . 5 M/s, calculated by the equation . = 
gM / V0 . The experimental apparatus was similar to that used for the VpHK experiment. 
In this case a concentrated solution of LiCl (3 M ) was released by the auto- 
FIGURE 16 Change in absorbance (plain circle, selected points) during VIK experiment 
concerning reaction of indomethacine with NaOH 0.005M and T = 299.0 ± 0.1 K. The dashed 
line shows the increase in ionic strength. The solid line is relative to the theoretical model. 
t (s) 
A (au) 
I (mol/L)

burette into the reaction vessel. Hundreds of absorbance data points were acquired. 
A fi t of the experimental data to the mathematical model (28) was performed using 
the MicroMath SCIENTIST program with A 0 , A . , k 0 , and Z A Z B the parameters to 
be optimized. The fi tting was excellent ( R 2 = 0.99999) and the results were in good 
agreement with those obtained in the traditional way [ k 0 = (1.02 ± 0.04) 10 . 4 s . 1 , 
Z A Z B = 1.01 ± 0.04]: 
. = . . + 
. 
dA 
dt 
k AA 
Z Z 
I 
I 
0 
1 04 
1 10 
. 
( ) A B 
(28) 
The differential method was also applied for processing the experimental data. 
The result was in good agreement. 
7.5.5 CONCLUSIONS 
Stability is an essential property of the drug product. A fast screening of new 
molecular entities and low - time - consuming detailed studies on selected candidates 
can avoid delay and consequent high cost in the fi rst part of pharmaceutical 
investigation. 
Potentially, VPaK provides a powerful new method for collecting kinetic data. It 
is a new way to look at kinetic experiments. Instead of obtaining a single value of 
the observed rate constant, with a single run, it is possible to obtain the entire 
dependence of the rate constant on a physical parameter. Usually, in mechanistic 
studies, what one looks for is the intimate way of interaction of the reagents 
in the rate - determining step. The nature of the reagent and products are well 
known because of the very good analytical instuments available. Kinetic experiments 
are just routine operations but, so far, they are of fundamental importance 
because kinetics is the only way to look at this aspect of the reactivity of the 
substrate. 
Without fear of spending too much time it is possible to obtain a full panoramic 
picture of the chemical behavior of a long series of homologous compounds. Plain 
mechanistic studies can be carried out using easily accessible instruments and software. 
Furthermore, kinetic data are obtained using a single sample, avoiding the 
inhomogeity characterizing traditional CPaK experimental data, loss of time in 
preparing more samples, and often, but not of secondary importance, using a lower 
quantity of compound. 
REFERENCES 
1. Carstensen, J. T. (2000), Solution Kinetics; Kinetic pH profi les; Oxidation in solution; Catalysis, 
Complexation, and Photolysis , in Carstensen, J. T. and Rhodes , C. T. , Eds., Drug Stability, 
Principle and Practice , 3rd ed., Marcel Dekker , New York, Chapters 2–5, pp. 19 – 143. 
2. Connors , K. A. , Amidon , G. L. , and Stella , V. J. ( 1986 ), Chemical Stability of Pharmaceuticals: 
A Handbook for Pharmacists , 2nd ed., Wiley-Interscience , New York. 
3. Loudon , G. M. ( 1991 ), Mechanistic interpretation of pH – rate profi les , J. Chem. Ed ., 68 , 
973 – 984 . 
4. Jenks , W. P. ( 1969 ), Catalysis in Chemistry and Enzymology , McGraw Hill , New York . 
REFERENCES 721

722 VARIABLE-PARAMETER KINETICS 
5. Kerns , E. H. ( 2001 ), High throughput physicochemical profi ling for drug discovery , 
J. Pharm. Sci ., 90 , 1838 – 1858 . 
6. Avdeef , A. , and Testa, B. (2002), Physicochemical profi ling in drug reserch: A brief survey 
of the state - of - the - art of experimental techniques , Cell. Mol. Life Sci ., 59 , 1681 – 1689 . 
7. Moore J. W. , and Pearson R. G. (1981), Kinetics and Mechanism , Wiley , New York . 
8. Wilkins R. G. ( 1991 ), Kinetics and Mechanism of Reactions of Transition Metal Complexes , 
VCH , Weinheim . 
9. Kresge , A. J. ( 1987 ), Unusual reactivity of prostacyclin: Rational drug design through 
physical organic chemistry , Acc. Chem. Res ., 20 , 364 – 370 . 
10. Page , M. I. , and Webster , P. ( 1990 ), The hydrolysis of azetidinyl amidinium salts. Part 1. 
The unimportance of strain release in the four - membered ring , J. Chem. Soc. Perkin 
Trans ., 2 , 805 – 811 . 
11. Lajis , N. H. , Noor , H. M. , and Khan , M. N. ( 1995 ), Kinetic and mechanism of the alkaline 
hydrolysis of securinine , J. Pharm. Sci ., 84 , 126 – 130 . 
12. Edwards , L. J. ( 1950 ), The hydrolysis of aspirin. A determination of the thermodynamic 
dissociation constant and a study of the reaction kinetics by ultraviolet spectrophotometry 
, Trans. Faraday Soc ., 46 , 723 – 735 . 
13. Edwards , L. J. (1952), The hydrolysis of aspirin. Part 2, Trans. Faraday Soc ., 48 , 696 – 699 . 
14. Romeo , R. , and Alibrandi , G. ( 1997 ), Structure - reactivity correlations for the dissociative 
uncatalyzed isomerization of monoalkylbis(phosphine)platinum(II) solvento complexes , 
Inorg. Chem ., 36 , 4822 – 4830 . 
15. Carstensen , J. T. ( 1970 ), Kinetic salt effect in pharmaceutical investigation , J. Pharm. Sci ., 
59 , 1140 – 1143 . 
16. Koryta , J. , Dvorak J. , and Kavan , L. ( 1993 ), Principles of Electrochemistry , Wiley , 
Chichester . 
17. Alibrandi, G. (1994), Variable-concentration kinetics , J. Chem. Soc. Chem. Commun ., 23 , 
2709 – 2710 . 
18. Koch , E. ( 1977 ), Non - Isothermal Reaction Analysis , Academic , London . 
19. Brown , M. E. , and Phillpotts , C. A. R. ( 1978 ), Non - isothermal kinetics , J. Chem. Ed ., 55 , 
556 – 560 . 
20. Kissinger , H. E. ( 1957 ), Reaction kinetics in differential thermal analysis , Anal. Chem ., 
29 , 1702 – 1706 . 
21. Flynn , J. H. ( 1969 ), in Schwenker , R. V. , and Garn , P. D. , Eds., Thermal Analysis , Vol. 2, 
Academic , New York , p. 1111 . 
22. Daniels , T. ( 1973 ), Thermal Analysis , Kogan Page , London . 
23. Dollimore , D. (1996), Thermal analysis , Anal. Chem ., 68 , 63R – 71R . 
24. Ozawa, T. (2000), Thermal analysis—Review and prospect, Thermoch. Acta , 335 , 35 – 42 . 
25. Glass , B. D. , Vov a k , Cs. , and Brown M. E. ( 2004 ), The thermal and photostability of solid 
pharmaceuticals. A review , J. Therm. Anal. Calorim ., 77 , 1013 – 1036 . 
26. Rogers , A. R. ( 1963 ), An accelerated stage test with programmed temperature rise , 
J. Pharm. Pharmacol ., 15 , 101T – 105T . 
27. Ahlberg , P. , and Wold , S. ( 1970 ), Evaluation of activation parameters for a fi rst order 
reaction from one kinetic experiment. Theory, numerical methods and computer program , 
Acta Chem. Scand ., 24 , 618 – 632 . 
28. Ahlberg , P. ( 1970 ), Determination of activation parameters in one kinetic experiment . 
Acta Chem. Scand ., 24 , 1883 – 1893 . 
29. Madsen , B. W. , and Anderson , R. A. ( 1974 ), Integral approach to nonisothermal estimation 
of activation energy , J. Pharm. Sci ., 63 , 777 – 781 . 

30. Edel , B. , and Baltzer , M. O. ( 1980 ), Nonisothermal kinetics with programmed temperature 
steps , J. Pharm. Sci ., 69 , 287 – 290 . 
31. Tucker , I. G. , and Owen , W. R. ( 1982 ), High information kinetic studies: Non - isothermal 
programmed acid concentration kinetics , Int. J. Pharm ., 10 , 323 – 337 . 
32. Gonzalez , J. L. , and Salvador , F. ( 1982 ), Kinetics of reactions in solution: Method for the 
treatment of data from non - isothermal chemical kinetic experiments , React. Kinet. Catal. 
Lett ., 21 ( 1 – 2 ), 167 – 171 . 
33. Gonzalez , J. L. , and Salvador , F. ( 1984 ), Comparative studies of non - isothermal methods 
in linear and non - linear temperature variation , React. Kinet. Catal. Lett ., 25 ( 1 – 2 ), 125 – 
130 . 
34. Mason , T. J. , and Lorimer , J. P. ( 1983 ), A method for the determination of the activation 
energy for a reaction from a single kinetic run , Comp. Chem ., 7 ( 4 ), 159 – 163 . 
35. Li Wan Po , A. , Elias , A. N. , and Irwin , W. J. ( 1983 ), Non - isothermal and non - isopH kinetics 
in formulation studies , Acta Pharm. Suec ., 20 , 277 – 286 . 
36. Ortiz Uribe , M. I. , Romero Salvador , A. , and Irabien Gulias , A. ( 1985 ), Kinetic analysis 
for liquid - phase reactions from programmed temperature data. I. Simple analysis of 
potential kinetic laws , Thermochem. Acta , 94 , 323 – 331 . 
37. Kipp , J. E. ( 1985 ), Non - isothermal kinetics — Comparison of two methods of data treatment 
, Int. J. Pharm ., 26 , 339 – 354 . 
38. Kipp , J. E. , Jensen , M. M. , Kronholm , K. , and McHalsky , M. ( 1986 ), Automated liquid 
chromatography for non - isothermal kinetic studies , Int. J. Pharm ., 34 , 1 – 8 . 
39. Bunce , N. J. , Forber , C. L. , and McInnes , C. ( 1988 ), Single - step methods for calculating 
activation parameters from raw kinetic data , J. Chem. Soc. Perkin Trans ., II , 363 – 368 . 
40. Schoenemann , E. , Hahn , H. , and Bracht , A. ( 1991 ), Determination of kinetic parameters 
from non - isothermal conductivity measurements by an integral method , Thermochim. 
Acta , 185 ( 1 ), 171 – 176 . 
41. Alibrandi , G. ( 1994 ), Non - isothermal spectrophotometric kinetics applied to inorganic 
reactions , Inorg. Chim. Acta , 221 , 31 – 34 . 
42. Zhang , S. , and Brown , T. L. ( 1995 ), Application of non - isothermal approach to the kinetics 
of organometallic reactions: The substitution of ( .5 - pentamethylcyclopentadienyl) 
dicarbonylrhodium(I) , Inorg. Chim. Acta , 240 , 427 – 433 . 
43. Junnarkar , G. H. , and Stavchansky , S. ( 1995 ), Isothermal and nonisothermal decomposition 
of famotidine in aqueous - solution , Pharm. Res ., 12 ( 4 ), 599 – 604 . 
44. Lee , M. L. , and Stavchansky , S. ( 1995 ), Isothermal and nonisothermal decomposition of 
thymopentin and its analogs in aqueous - solution , Pharm. Res ., 15 ( 11 ), 599 – 604 . 
45. Alibrandi , G. , Micali , M. , Trusso , S. , and Villari , A. ( 1996 ), Hydrolysis of aspirin studied 
by spectrophotometric and fl uorometric variable - temperature kinetics , J. Pharm. Sci ., 85 , 
1105 – 1108 . 
46. Maeder , M. , Molloy , K. J. , and Schumacher , M. M. ( 1997 ), Analysis of non - isothermal 
kinetic measurements , Anal. Chim. Acta , 337 , 73 – 81 . 
47. Hodgson , S. C. , Ngeh , L. N. , Orbell , J. D. , and Bigger , S. W. ( 1998 ), A student experiment 
in non - isothermal chemical kinetics , J. Chem. Ed ., 75 ( 9 ), 1150 – 1153 . 
48. Ficarra , R. , Villari , A. , Micali , N. , Tommasini , S. , Calabr o , M. L. , Di Bella , M. R. , Melardi , 
S. , Agresta , M. F. , Coppolino , S. , and Stancanelli , R. ( 1999 ), Stability study of piroxicam 
and cinnoxicam in solid pharmaceuticals , J. Pharm. Biomed. Anal ., 20 , 283 – 288 . 
49. Alibrandi , G. , Coppolino , S. , D ’ Aliberti , S. , Ficarra , P. , Micali , N. , and Villari , A. ( 2002 ), 
Temperature - rate profi les by polarimetric variable - temperature kinetic experiments to 
study racemization reactions , J. Pharm. Biomed. Anal ., 29 , 1025 – 1029 . 
REFERENCES 723

724 VARIABLE-PARAMETER KINETICS 
50. Alibrandi , G. , Coppolino , S. , D ’ Aliberti , S. , Ficarra , R. , Micali , N. , and Villari , A. ( 2003 ), 
Fast drug stability determination by LC variable - parameter kinetic experiments , J. Pharm. 
Biomed. Anal ., 32 , 1073 – 1079 . 
51. Mood , A. R. H. , Haghighi , S. , and Gholami , M. R. ( 2004 ), Fluorometric variable - temperature 
kinetic investigations of the transesterifi cation reaction of procaine with aliphatic 
alcohols , J. Pharm. Pharm. Sci ., 7 ( 1 ), 88 – 91 . 
52. Alibrandi , G. , Coppolino , S. , Micali , N. , and Villari , A. ( 2001 ), Variable - pH kinetics: An 
easy determination of pH - rate profi le , J. Pharm. Sci ., 90 , 270 – 274 . 
53. Alibrandi , G. , D ’ Aliberti , S. , and Tresoldi , G. ( 2003 ), Spectrophotometric variable - 
concentration kinetic experiments applied to inorganic reactions , Int. J. Chem. Kinet ., 35 , 
497 – 502 . 
54. Alibrandi , G. , Coppolino , S. , D ’ Aliberti , S. , Ficarra , P. , Micali , N. , and Villari , A. ( 2003 ), 
Variable - ionic strength kinetic experiments to study drug stability , J. Pharm. Sci ., 92 , 
1730 – 1733 . 
55. Alibrandi , G. , D ’ Aliberti , S. , and Pedicini , R. ( 2001 ), Computer simulation of variable - 
parameter kinetic experiments , Chem. Ed ., 6 , 185 – 191 . 
56. MicroMath Scientifi c Software, Salt Lake City, UT. 
57. Press , W. H. , Flannery , B. P. , Teukolsky , S. A. , and Vetterling , W. T. ( 1986 ), Numerical 
Recipes , Cambridge University Press , Cambridge . 
58. Savitzky , A. , and Golay , M. J. E. ( 1964 ), Smoothing and differentiation of data by simpli- 
fi ed least squares procedures , Anal. Chem ., 36 , 1627 – 1639 . 
59. Garrett , E. R. ( 1957 ), The kinetics of solvolysis of acyl esters of salycilic acid , J. Am. Chem. 
Soc ., 79 , 3401 – 3408 . 
60. Fersht , A. R. , and Kirby , A. J. ( 1967 ), The hydrolysis of aspirin. Intramolecular general 
base catalysis of ester hydrolysis , J. Am. Chem. Soc ., 89 , 4857 – 4863 . 
61. Miles , C. I. , and Schenk , G. H. ( 1970 ), Fluorescence of acetylsalicylic acid in solution and 
its measurement in presence of salicylic acid , Anal. Chem ., 42 , 656 – 659 . 
62. Tobe , M. L. ( 1987 ), Substitution reactions , in Wilkinson , G. , Gillard , R. D. , and McCleverty , 
J. A. , Eds., Comprehensive Coordination Chemistry , Vol. 1, Pergamon , Oxford , Chapter 
7.1, pp. 281 – 329 . 
63. Belluco , U. , Cattalini , L. , Basolo , F. , Pearson , R. G. , and Turco , A. ( 1965 ), Nucleophilic 
constants and substrate discrimination factors for substitution reactions of platinum(II) 
complexes , J. Am. Chem. Soc ., 87 , 241 – 246 . 
64. Sherman , S. E. , and Lippard , S. J. ( 1987 ), Structure aspect of platinum anticancer drug 
interactions with DNA , Chem. Rev ., 87 , 1153 – 1181 . 
65. Nicolini , M. , Ed . (1988 ), Platinum and Other Metal Coordination Compounds in Cancer 
Chemiotherapy , Martinus Nijoff Publishing , Boston, MA . 
66. Hajratwala , B. R. , and Dawson , J. E. ( 1977 ), Kinetics of indomethacin degradation I: 
Presence of alkali , J. Pharm. Sci ., 66 , 27 – 29 . 
67. Cipiciani , A. , Ebert , C. , Linda , P. , Rubessa , F. , and Savelli , G. ( 1983 ), Kinetics and mechanism 
of the basic hydrolysis of indomethacin and related compounds: A reevaluation , 
J. Pharm. Sci ., 72 , 1075 – 1076 . 

VALIDATION 
SECTION 8


727 
8.1 
ANALYTICAL METHOD VALIDATION: 
PRINCIPLES AND PRACTICES 
Chung Chow Chan 
Azopharma Contract Pharmaceutical Services, Miramar, Florida 
Contents 
8.1.1 Introduction 
8.1.2 Why Validate Analytical Procedures 
8.1.3 Current Good Manufacturing Practices in Twenty - First Century 
8.1.4 Cycle of Analytical Methods 
8.1.5 Analytical Method Validation Characteristics 
8.1.5.1 Accuracy 
8.1.5.2 Method Precision 
8.1.5.3 Specifi city 
8.1.5.4 Detection Limit 
8.1.5.5 Quantitation Limit 
8.1.5.6 Linearity 
8.1.5.7 Range 
8.1.5.8 Robustness 
8.1.6 Process of Analytical Method Validation 
8.1.7 Information Required in Analytical Procedure 
8.1.8 Phase - Appropriate Method Validation 
8.1.9 Method Verifi cation 
8.1.10 Method Revalidation 
8.1.11 Conclusion 
References 
8.1.1 INTRODUCTION 
Validation of an analytical procedure is the process by which it is established, by 
laboratory studies, that the performance characteristics of the procedure meet the 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

728 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
requirements for its intended use. All analytical methods that are intended to be 
used for analyzing any clinical samples will need to be validated. Validation of analytical 
methods is an essential but time - consuming activity for most analytical development 
laboratories. It is therefore important to understand the requirements of 
method validation in more detail and the options that are available to allow for 
optimal utilization of analytical resources in a development laboratory. 
8.1.2 WHY VALIDATE ANALYTICAL PROCEDURES 
There are many reasons for the need to validate analytical procedures. Among them 
are regulatory requirements, good science, and quality control requirements. The 
Code of Federal Regulations (CFR) 311.165c explicitly states that “ the accuracy, 
sensitivity, specifi city, and reproducibility of test methods employed by the fi rm shall 
be established and documented. ” Of course, as scientists, we would want to apply 
good science to demonstrate that the analytical method used had demonstrated 
accuracy, sensitivity, specifi city, and reproducibility. Finally management of the 
quality control unit would defi nitely want to ensure that the analytical methods that 
the department uses to release its products are properly validated for its intended 
use so the product will be safe for human use. 
8.1.3 CURRENT GOOD MANUFACTURING PRACTICES IN 
TWENTY - FIRST CENTURY 
The overarching philosophy in current good manufacturing practices (cGMPs) of 
the twenty - fi rst century and in robust modern quality systems is quality should be 
built into the product, and testing alone cannot be relied on to ensure product quality. 
From the analytical perspective, this will mean that analytical methods used to test 
these products should have quality attributes built into them. To have quality attributes 
built into the analytical method will require that fundamental quality attributes 
be applied by the bench - level scientist. This is a paradigm shift that requires 
the bench - level scientist to have the scientifi c and technical understanding, product 
knowledge, process knowledge, and/or risk assessment abilities to appropriately 
execute the quality functions of analytical method validation. It will require (1) the 
appropriate training of the bench - level scientist to understand the principles involved 
with method validation and able to validate an analytical method and understand 
the principles involved with the method validation, (2) proper documentation and 
understanding and interpreting data, and (3) cross - functional understanding of the 
effect of their activities on the product and the customer (the patient). It is the 
responsibility of management to verify that skills gained from the training are 
implemented in day - to - day performance. 
8.1.4 CYCLE OF ANALYTICAL METHODS 
The analytical method validation activity is not a one - time study. This is illustrated 
and summarized in the life cycle of an analytical procedure in Figure 1 . An analytical 

ANALYTICAL METHOD VALIDATION CHARACTERISTICS 729 
method will be developed and validated for use to analyze samples during the early 
development of an active pharmaceutical ingredient (API) or drug product. As drug 
development progresses from phase 1 to commercialization, the analytical method 
will follow a similar progression. The fi nal method will be validated for its intended 
use for the market image drug product and transferred to the quality control laboratory 
for the launch of the drug product. However, if there are any changes in the 
manufacturing process that have the potential to change the analytical profi le of the 
drug substance and drug product, this validated method may need to be revalidated 
to ensure that it is still suitable to analyze the API or drug product for its intended 
purpose. 
8.1.5 ANALYTICAL METHOD VALIDATION 
CHARACTERISTICS 
Typical analytical performance characteristics that should be considered in the validation 
of the types of procedures described in this chapter are listed below. Each 
validation characteristic is defi ned to ensure consistency in usage of terminology 
and interpretation: 
Accuracy 
Precision 
Repeatability 
Intermediate precision 
Specifi city 
Detection limit 
Quantitation limit 
Linearity 
Range 
Robustness 
FIGURE 1 Life cycle of analytical method. 
Method development 
QC laboratory 
Method validation/revalidation 

730 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
8.1.5.1 Accuracy 
The International Convention on Harmonization (ICH) defi nes the accuracy of an 
analytical procedure as the closeness of agreement between the values that are 
accepted either as conventional true values or an accepted reference value and the 
value found. For drug substance, accuracy may be defi ned by the application of the 
analytical procedure to an analyte of known purity (e.g., a reference standard). For 
the drug product, accuracy will be determined by application of the analytical procedure 
to synthetic mixtures of the drug product components to which known 
amounts of analyte have been added within the range of the procedure. The ICH 
document also recommends assessing a minimum of nine determinations over a 
minimum of three concentration levels covering the specifi ed range (e.g., three 
concentrations/three replicates). 
Accuracy is usually reported as percent recovery by the assay (using the proposed 
analytical procedure) of known added amount of analyte in the sample or as the 
difference between the mean and the accepted true value together with the confi - 
dence intervals. The range for the accuracy limit should be within the linear range. 
Typical accuracy of the recovery of the drug substance is expected to be about 99 – 
101%. Typical accuracy of the recovery of the drug product is expected to be about 
98 – 102%. Values of accuracy of recovery data beyond this range need to be investigated 
as appropriate. 
8.1.5.2 Method Precision 
The precision of an analytical procedure expresses the closeness of agreement 
(degree of scatter) between a series of measurements obtained from multiple 
samples of the same homogeneous sample under prescribed conditions. Precision is 
usually investigated at three levels: repeatability, intermediate precision, and reproducibility. 
For simple formulation it is important that precision be determined using 
authentic homogeneous samples. A justifi cation will be required if a homogeneous 
sample is not possible and artifi cially prepared samples or sample solutions are 
used. 
Repeatability Repeatability is a measure of the precision under the same operating 
conditions over a short interval of time, that is, under normal operating conditions 
of the analytical method with the same equipment. It is sometimes referred 
to as intra - assay precision. 
The ICH recommends that repeatability be assessed using a minimum of nine 
determinations covering the specifi ed range for the procedure (e.g., three concentrations/
three replicates as in the accuracy experiment) or using a minimum of six 
determinations at 100% of the test concentration. Reporting of the standard deviation, 
relative standard deviation (coeffi cient of variation), and confi dence interval 
is required. The assay values are independent analyses of samples that have been 
carried through the complete analytical procedure from sample preparation to fi nal 
test result. Table 1 provides an example set of repeatability data. 
Intermediate Precision Intermediate precision is defi ned as the variation within 
the same laboratory. The extent to which intermediate precision needs to be estab

ANALYTICAL METHOD VALIDATION CHARACTERISTICS 731 
lished depends on the circumstances under which the procedure is intended to be 
used. Typical parameters that are investigated include day - to - day variation, analyst 
variation, and equipment variation. Depending on the extent of the study, the use 
of experimental design is encouraged. Experimental design will minimize the 
number of experiments that need to be performed. It is important to note that ICH 
allows exemption from doing intermediate precision when reproducibility is proven. 
It is expected that the intermediate precision should show variability that is in the 
same range or less than repeatability variation. ICH recommends the reporting of 
standard deviation, relative standard deviation (coeffi cient of variation), and confi - 
dence interval of the data. 
Reproducibility Reproducibility measures the precision between laboratories. 
This parameter is considered in the standardization of an analytical procedure 
(e.g., inclusion of procedures in pharmacopeias and method transfer between different 
laboratories). 
To validate this characteristic, similar studies need to be performed at different 
laboratories using the same homogeneous sample lot and the same experimental 
design. In the case of method transfer between two laboratories, different approaches 
may be taken to achieve the successful transfer of the procedure. The most common 
approach is the direct - method transfer from the originating laboratory to the receiving 
laboratory. The originating laboratory is defi ned as the laboratory that has 
developed and validated the analytical method or a laboratory that has previously 
been certifi ed to perform the procedure and will participate in the method transfer 
studies. The receiving laboratory is defi ned as the laboratory to which the analytical 
procedure will be transferred and that will participate in the method transfer studies. 
In the direct - method transfer, it is recommended that a protocol be initiated with 
details of the experiments to be performed and acceptance criteria (in terms of the 
difference between the means of the two laboratories) for passing the method 
transfer. Table 2 provides examples of a set of method transfer data between two 
laboratories. 
8.1.5.3 Specifi city 
The ICH defi nes specifi city as the ability to assess unequivocally an analyte in the 
presence of components that may be expected to be present. In many publications, 
TABLE 1 Repeatability Data 
Replicate Percentage of Labeled Claim 
1 100.6 
2 102.1 
3 100.5 
4 99.4 
5 101.4 
6 101.1 
Mean 100.9 
Percentage relative standard deviation (%RSD) 0.90 

732 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
selectivity and specifi city are often used interchangeably. However, there are debates 
over the use of specifi city over selectivity and some authorities, for example, the 
International Union of Pure and Applied Chemistry (IUPAC), have preferred the 
term selectivity , reserving specifi city for those procedures that are completely selective. 
For pharmaceutical application, the above defi nition of ICH will be used. 
For identity test, compounds of closely related structures which are likely to be 
present should be discriminated from each other. This could be confi rmed by obtaining 
positive results (by comparison with a known reference material) from samples 
containing the analyte, coupled with negative results from samples which do not 
contain the analyte. Furthermore, the identifi cation test may be applied to material 
structurally similar or closely related to the analyte to confi rm that a positive 
response is not obtained. The choice of such potentially interfering materials should 
be based on sound scientifi c judgment with a consideration of the interferences that 
could occur. 
The specifi city for an assay and impurity tests should be approached from two 
angles: 
1. When Impurities Are Available The specifi city of an assay method is determined 
by comparing test results from an analysis of sample containing the 
impurities, degradation products, or placebo ingredients with those obtained 
from an analysis of samples without the impurities, degradation products, or 
placebo ingredients. For a stability - indicating assay method, degradation peaks 
need to be resolved from the drug substance. However, these impurities do 
not need to be resolved from each other. 
For the impurity test, the determination should be established by spiking drug 
substance or drug product with the appropriate levels of impurities and demonstrating 
the separation of these impurities individually and/or from other 
components in the sample matrix. Representative chromatograms should be 
used. 
2. If Impurities Are Not Available. Specifi city may be demonstrated by comparing 
the test results of samples containing impurities or degradation products 
to a second well - characterized procedure or other validated analytical procedure 
(orthogonal method). This should include samples stored under relevant 
stress conditions (light, heat, humidity, acid/base hydrolysis and oxidation). For 
the assay method, the two results should be compared; for impurity tests, the 
impurity profi les should be compared. Peak homogeneity tests should be performed 
using PDA or mass spectrometry to show that the analyte chromatographic 
peak is not attributable to more than one component. Figure 2 
illustrates the selectivity of a method to resolve known degradation peaks 
from the parent peak. 
TABLE 2 Results from Method Transfer between Two 
Laboratories 
Runs Average Percent 
Originating laboratory 12 100.7 
Receiving laboratory 4 100.2 

ANALYTICAL METHOD VALIDATION CHARACTERISTICS 733 
8.1.5.4 Detection Limit 
The detection limit (DL) is a characteristic for the limit test only. It is the lowest 
amount of analyte in a sample that can be detected but not necessarily quantitated 
under the stated experimental conditions. The detection is usually expressed as the 
concentration of the analyte in the sample, for example, percentage, parts per million 
(ppm), or parts per billion (ppb). 
There are several approaches to establish the DL. Visual evaluation may be used 
for noninstrumental (e.g., solution color) and instrumental methods. In this case, the 
DL is determined by the analysis of a series of samples with known concentrations 
and establishing the minimum level at which the analyte can be reliably detected. 
Presentation of relevant chromatograms or other relevant data is suffi cient for justifi 
cation of the DL. 
For instrumental procedures that exhibit background noise, it is common to 
compare measured signals from samples with known low concentrations of analyte 
with those of the blank samples. The minimum concentration at which the analyte 
can reliably be detected is established using an acceptable signal - to - noise ratio of 
2 : 1 or 3 : 1. Presentation of relevant chromatograms is suffi cient for justifi cation of 
the DL. 
Another approach estimates the DL from the standard deviation of the response 
and the slope of the calibration curve. The standard deviation can be determined 
either from the standard deviation of multiple blank samples or from the standard 
deviation of the y intercepts of the regression lines done in the range of the DL. 
This estimate will need to be subsequently validated by the independent analysis of 
a suitable number of samples near or at the DL: 
DL = 3. 
S 
FIGURE 2 Overlay chromatogram of impurity solution with sample solution. 

734 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
where . is the standard deviation of the response and S is the slope of the calibration 
curve. 
8.1.5.5 Quantitation Limit 
The quantitation Limit (QL) is a characteristic of quantitative assays for low levels 
of compounds in sample matrices, such as impurities in bulk drug substances and 
degradation products in fi nished pharmaceuticals. QL is defi ned as the concentration 
of related substance in the sample that will give a signal - to - noise ratio of 10 : 1. 
The QL of a method is affected by both the detector sensitivity and the accuracy 
of sample preparation at the low concentration of the impurities. In practice, QL 
should be lower than the corresponding ICH report limit. 
ICH recommends three approaches to the estimation of QL. The fi rst approach 
is to evaluate it by visual evaluation and may be used for noninstrumental methods 
and instrumental methods. QL is determined by the analysis of samples with known 
concentrations of analyte and by establishing the minimum level at which the 
analyte can be quantitated with acceptable accuracy and precision. 
The second approach determines the signal - to - noise ratio by comparing measured 
signals from samples with known low concentrations of anlayte with those of 
blank samples. QL is the minimum concentration at which the analyte can be reliably 
quantifi ed at the signal - to - noise ratio of 10 : 1. 
The third approach estimates QL by the equation 
QL = 10. 
S 
The slope S may be estimated from the calibration curve of the analyte. The value 
of . may be estimated by (1) calculating the standard deviation of the responses 
obtained from the measurement of the analytical background response of an appropriate 
number of blank samples or (2) calculating the residual standard deviation 
of the regression line from the calibration curve using samples containing the 
analyte in the range of the QL. 
Whatever approach is applied, the QL should be subsequently validated by the 
analysis of a suitable number of samples prepared at the QL and determining the 
precision and accuracy at this level. 
8.1.5.6 Linearity 
ICH defi nes linearity of an analytical procedure as the ability (within a given range) 
to obtain test results of variable data (e.g., absorbance and area under the curve) 
which are directly proportional to the concentration (amount of analyte) in the 
sample. The data variables that can be used for quantitation of the analyte are the 
peak areas, peak heights, or the ratio of peak areas (heights) of analyte to the internal 
standard peak. Quantitation of the analyte depends on it obeying Beer ’ s law for 
the spectroscopic method over a concentration range. Therefore, the working sample 
concentration and samples tested for accuracy should be in the linear range. 
There are two general approaches for determining the linearity of the method. 
The fi rst approach is to weigh different amounts of standard directly to prepare 

ANALYTICAL METHOD VALIDATION CHARACTERISTICS 735 
linearity solutions at different concentrations. However, it is not suitable to prepare 
solution at very low concentration, as the weighing error will be relatively high. 
Another approach is to prepare a stock solution of high concentration. Linearity 
is then demonstrated directly by dilution of the standard stock solution. This is more 
popular and the recommended approach. Linearity is best evaluated by visual 
inspection of a plot of the signals as a function of analyte concentration. Subsequently, 
the variable data are generally used to calculate a regression line by the 
least - squares method. At least fi ve concentration levels should be used. Under 
normal circumstances, linearity is acceptable with a coeffi cient of determination ( r 2 ) 
of . 0.997. The slope, residual sum of squares, and y intercept should also be reported 
as required by ICH. 
The slope of the regression line will provide an idea of the sensitivity of the 
regression, and hence the method that is being validated. The y intercept will 
provide an estimate of the variability of the method. For example, the ratios percent 
of the y intercept with the variable data at nominal concentration are sometimes 
used to estimate the method variability. 
For the determination of potency assay of a drug substance or a drug product, 
the usual range of linearity should be ± 20% of the target or nominal concentration. 
For the determination of content uniformity, it should be ± 30% of the 
target or nominal concentration. Figure 3 illustrates the linearity of a sample set 
of data. 
8.1.5.7 Range 
The range of an analytical procedure is the interval between the upper and lower 
concentration of analyte in the sample for which it has been demonstrated that the 
analytical procedure has a suitable level of precision, accuracy, and linearity. The 
range is normally expressed in the same units as test results (e.g., percent, parts per 
million) obtained by the analytical procedure. 
For the assay of drug substance or fi nished drug product, it is normally recommended 
to have a range of 80 – 120% of the nominal concentration. 
For content uniformity, a normal range would cover 70 – 130% of the nominal 
concentration, unless a wider and more appropriate range (e.g., metered - dose inhalers) 
is justifi ed. 
For dissolution testing, a normal range is ± 20% over the specifi ed range. If the 
acceptance criterion for a controlled - release product covers a region from 20% after 
FIGURE 3 Linearity plot of peak - area response versus concentration. 
0 
2,000 
4,000 
6,000 
8,000 
10,000 
12,000 
14,000 
16,000
0 50 100 150 200 250 
Concentration (.g/mL) 
Peak area

736 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
1 h, and up to 90% after 24 h, the validated range would be 0 – 110% of the label 
claim. In this case, the lowest appropriate quantifi able concentration of analyte will 
be used as the lowest limit as 0% is not appropriate. 
8.1.5.8 Robustness 
Robustness of an analytical procedure is a measure of the analytical method to 
remain unaffected by small but deliberate variations in method parameters and 
provides an indication of its reliability during normal usage. 
The evaluation of robustness is normally considered during the development 
phase and depends on the type of procedure under study. Experimental design 
(e.g., fractional factorial design or Plackett – Burman design) is common and useful 
to investigate multiple parameters simultaneously. The result will help to identify 
critical parameters that will affect the performance of the method. Common method 
parameters that can affect the analytical procedure should be considered based on 
the analytical technique and properties of the samples: 
1. Sample preparation 
a. Extraction time 
b. Sample solvent (pH ± 0.05 unit, percent organic ± 2% absolute) 
c. Membrane fi lters 
d. Sample and standard stability 
2. High - performance liquid chromatography (HPLC) conditions 
a. Mobile - phase composition (pH ± 0.05 unit, percent organic ± 2% 
absolute) 
b. Column used (equivalent columns, lots and/or suppliers, age) 
c. Temperature 
d. Flow rate 
3. Gas chromatography (GC) conditions 
a. Column used (lots and/or suppliers, age) 
b. Temperature 
c. Flow rate 
When the results are affected by some critical experimental parameters, a precautionary 
statement should be included in the analytical procedure to ensure that 
this parameter is tightly controlled between experiments. For example, if percent 
ionpairing of mobile phase affects the results signifi cantly, the analytical procedure 
should explicitly be written with a precautionary statement for aqueous component, 
for example, 40% aqueous 20 m M octanesulfonic acid ± 2% absolute. 
Other robustness considerations for ruggedness of the analytical procedure 
during validation include the following: 
(a) Sample Extraction Mechanical shaking is preferred over sonication as the 
latter is affected by a number of factors, for example, water level in bath and 
position of sample. 

(b) Dilution of Sample and Solvent. Minimize the number of dilution steps to 
reduce introduction of error. Dilution solvent should be as similar to mobile 
phase as possible. 
8.1.6 PROCESS OF ANALYTICAL METHOD VALIDATION 
The typical process that is followed in an analytical method validation is chronologically 
listed below: 
1. Planning and deciding on the method validation experiments 
2. Writing and approval of method validation protocol 
3. Execution of the method validation protocol 
4. Analysis of the method validation data 
5. Reporting the analytical method validation 
6. Finalizing the analytical method procedure 
The method validation experiments should be well planned and laid out to ensure 
effi cient use of time and resources during execution of the method validation. The 
best way to ensure a well - planned validation study is to write a method validation 
protocol that will be reviewed and signed by the appropriate person (e.g., laboratory 
management and quality assurance). 
The validation parameters that will be evaluated will depend on the type of 
method to be validated. Analytical methods that are commonly validated can be 
classifi ed into three main categories: identifi cation, testing for impurities, and assay. 
Table 3 lists the ICH recommendations for each of these methods. 
Execution of the method validation protocol should be carefully planned 
to optimize the resources and time required to complete the full validation 
study. For example, in the validation of an assay method, linearity and accuracy may 
be validated at the same time as both experiments can use the same standard solutions. 
A normal validation protocol should contain the following contents at a 
minimum: 
(a) Objective of the protocol 
(b) Validation parameters that will be evaluated 
(c) Acceptance criteria for all the validation parameters evaluated 
(d) Details of the experiments to be performed 
(e) Draft analytical procedure 
The data from the method validation data should be analyzed as the data are 
obtained and processed to ensure a smooth fl ow of information. If an experimental 
error is detected, it should be resolved as soon as possible to reduce any impact it 
may have on later experiments. Analysis of the data includes visual examination of 
the numerical values of the data and chromatograms followed by statistical treatment 
of the data if required. 
PROCESS OF ANALYTICAL METHOD VALIDATION 737

738 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
Upon completion of all the experiments, all the data will be compiled into a 
detailed validation report that will conclude the success or failure of the validation 
exercise. Depending on the company ’ s strategy a summary of the validation data 
may also be generated. Successful execution of the validation will lead to a fi nal 
analytical procedure that can be used by the laboratory to support future analytical 
work for the drug substance or drug product. 
8.1.7 INFORMATION REQUIRED IN ANALYTICAL PROCEDURE 
The minimal information that should be included in a fi nal analytical procedure are 
as follows: 
(a) Rationale of the analytical procedure and description of the capability of the 
method. Revision of analytical procedure should include the advantages 
offered by the new revision. 
(b) Proposed analytical procedure. This section should contain a complete description 
of the analytical procedure in suffi cient detail to enable another 
analytical scientist to replicate it. The write - up should include all important 
operational parameters and specifi c instructions, for example, preparation of 
reagents, system suitability tests, precautions, and explicit formulas for calculation 
of the test results. 
(c) List of permitted impurities and its levels in an impurity assay. 
(d) Validation data. Either a detailed set or summary set of validation data is 
included 
TABLE 3 Validation Parameters 
Type of Analytical Procedure 
Characteristics Identifi cation 
Testing for 
Impurities 
Assay . Dissolution 
(Measurement 
Only) . Content/ 
Potency Quantitation Limit 
Accuracy . + . + 
Precision . 
Repeatability . + . + 
Intermediate precision + a . + a 
Specifi city b + + + + 
Detection limit . . c + . 
Quantitation limit . + . . 
Linearity . + . + 
Range . + . + 
Note : . , characteristic not normally evaluated; +, characteristic normally evaluated. a In cases where 
reproducibility has been performed, intermediate precision is not needed. b Lack of specifi city of one 
analytical procedure could be compensated by other supporting analytical procedure(s). c May be needed 
in some cases. 

(e) Revision history. 
(f) Signature of author, reviewer, management, and quality assurance. 
8.1.8 PHASE - APPROPRIATE METHOD VALIDATION 
The original intent of the cGMPs was to describe standards and activities designed 
to ensure the strength, identity, safety, purity, and quality of pharmaceutical products 
introduced into commerce. However, the GMPs are silent on explicit guidances for 
the development phase of pharmaceuticals in several areas. 
Regulatory bodies recognize that knowledge of the drug product and its analytical 
methods will evolve through the course of development. This is stated explicitly 
in ICH Q7A: Changes are expected during development, and every change in 
product, specifi cations, or test procedures should be recorded adequately. It is therefore 
reasonable to expect that changes in testing, processing, packaging, and so on, 
will occur as more is learned about the molecule. However, even with the changes, 
the need for ensuring the safety of subjects in clinical testing should not be 
compromised. 
According to the ICH guidance, the objective of method validation is to demonstrate 
that analytical procedures “ are suitable for their intended purpose. ” Therefore 
the method ’ s purpose should be linked to the clinical studies and the pharmaceutical 
purpose of the product being studied. 
The purpose in the early phase of drug development is to deliver a known dose 
that is bioavailable for clinical studies. As product development continues, increasing 
emphasis is placed on identifying a stable, robust formulation from which multiple, 
bioequivalent lots can be manufactured and ultimately scaled up, transferred, 
and controlled for commercial manufacture. 
The development and validation of analytical methods should follow a similar 
progression. The purpose of analytical methods in early stages of development is 
to ensure potency, to understand the impurity and degradation product profi le, 
and to help understand key drug characteristics. As development continues, the 
method should be stability indicating and capable of measuring the effect of key 
manufacturing parameters to ensure consistency of the drug substance and drug 
product. 
Analytical methods used to determine purity and potency of an experimental 
API that is very early in development will need a less rigorous method validation 
exercise than would be required for a quality control laboratory method at the 
manufacturing site. An early phase project may have only a limited number of lots 
to be tested and the testing may be performed in only one laboratory by a limited 
number of analysts. The ability of the laboratory to “ control ” the method and its 
use is relatively high, particularly if laboratory leadership is clear in its expectations 
for the performance of the work. 
The environment in which a method is used changes signifi cantly when the 
method is transferred to a quality control laboratory at the manufacturing site. 
The method may be replicated in several laboratories, multiple analysts may use 
it, and the method may be one of many methods used in the laboratory daily. 
The developing laboratory must therefore be aware of the needs of the receiving 
PHASE-APPROPRIATE METHOD VALIDATION 739

740 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
laboratories, for example, quality control laboratory, and regulatory expectations for 
the successful validation of a method to be used in support of a commercial 
product. 
An example of the minimum requirement for potency assay of the drug substance 
and drug product is tabulated in Table 4 . Note that the postponement of intermediate 
precision is aligned with previous discussion that the use of early phase analytical 
method resides mainly in one laboratory and is used only by a very limited number 
of analysts. Each individual company ’ s phased method validation procedures and 
processes will vary, but the overall philosophy is the same. The extent of and expectations 
from early phase method validation are lower than the requirements in the 
later stages of development. The validation exercise becomes larger and more 
detailed and collects a larger body of data to ensure that the method is robust and 
appropriate for use at the commercial site. 
However, certain fundamental concepts of cGMPs must be applied regardless of 
the details of the phased appropriate method validation strategy used. Examples 
are (1) proper documentation, (2) change control, (3) deviations, (4) equipment and 
utilities qualifi cation, and (5) proper training. 
A detailed method validation report may not be necessary until submission of 
the fi nal market application. However, summary reports should be available to 
facilitate effi cient data retrieval and fulfi ll requests from regulatory agencies for the 
information when required. 
8.1.9 METHOD VERIFICATION 
The U.S. Food and Drug Administration (FDA) regulation 21 CFR 211.194(a)(2) 
specifi cally states that users of analytical methods in the U.S. Pharmacopeia/National 
Formulary (USP/NF) are not required to validate the accuracy and reliability of 
these methods but merely verify their suitability under actual conditions of use. USP 
TABLE 4 Assay Method Validation in Early Phase for Drug Substance and Drug 
Product 
Drug Substance Drug Product 
Accuracy Inferred from precision, 
linearity, and specifi city 
Recovery at 100% for each strength 
(bracket for multiple strengths) 
Repeatability Three sample preparation 
at 100% nominal 
Three sample preparations at 100% 
nominal 
Intermediate precision To be completed at later 
stages of development 
To be completed at later stages of 
development 
Specifi city Resolution from most 
likely impurities 
Resolution from impurities and 
excipients 
Quantitation/detection 
limit 
Not required Not required 
Linearity Minimum three levels 
from 80 to 120% 
Minimum three levels from 80 to 
120% 
Range Defi ned in linearity Defi ned in linearity 
Robustness Solution stability Solution stability 

has issued a guidance for verifi cation in general chapter . 1226 . . This proposal provides 
general information to laboratories on the verifi cation of compendial procedures 
that are being performed for the fi rst time to yield acceptable results utilizing 
the laboratories ’ personnel, equipment, and reagents. 
Verifi cation consists of assessing selected analytical validation characteristics 
described earlier to generate appropriate, relevant data rather than repeating the 
validation process for commercial products. The guidance in this general chapter is 
applicable to applications such as titrations, chromatographic procedures (related 
compounds, assay, and limit tests), and spectroscopic tests. However, general 
tests (e.g., water, heavy metals, residue on ignition) do not typically require 
verifi cation. 
Table 5 summarizes the comparison of the validation requirements with the veri- 
fi cation requirements of the HPLC assay of an example fi nal dosage form. ICH 
requires the validation of accuracy, precision, specifi city, linearity, and range. Generally, 
verifi cation will only require a minimal of precision and specifi city validation. 
The accuracy requirements will be dependent on the specifi c situation of the fi nal 
dosage form. 
8.1.10 METHOD REVALIDATION 
There are various circumstances under which a method needs to be revalidated. 
Some of the common situations are described below: 
1. During the optimization of the drug substance synthetic process, signifi cant 
changes were introduced into the process. To ensure that the analytical method 
will still be able to analyze the potentially different profi le of the API, revalidation 
may be necessary. 
2. If a new impurity is found that makes the method defi cient in its specifi city, 
this method will need to be modifi ed or redeveloped and revalidated to ensure 
that it will be able to perform its intended function. 
3. A change in the excipient composition may change the product impurity 
profi le. This change may make the method defi cient in its specifi city for the 
assay or impurity tests and may require redevelopment and revalidation. 
TABLE 5 Validation and Verifi cation Requirements for 
HPLC Assay of Final Dosage Forms 
Performance Characteristics Validation Verifi cation 
Accuracy Yes Maybe 
Precision Yes Yes 
Specifi city Yes Yes 
Limit of detection (LOD) No No 
Limit of quantifi cation 
(LOQ) 
No No 
Linearity Yes No 
Range Yes No 
METHOD REVALIDATION 741

742 ANALYTICAL METHOD VALIDATION: PRINCIPLES AND PRACTICES 
4. Changes in equipment or suppliers of critical supplies of the API or fi nal drug 
product will have the potential to change their degradation profi le and may 
require the method to be redeveloped and revalidated. 
8.1.11 CONCLUSION 
This chapter summarizes the validation parameters that are required according to 
the requirements of ICH Q2R(1). The paradigm shift under cGMP in the twenty - 
fi rst century that requires the bench - level scientist to have the scientifi c and technical 
understanding, product knowledge, process knowledge, and/or risk assessment 
abilities to appropriately execute the quality functions of analytical method validation 
is presented in detail. The method validation process and the minimum requirements 
to be included in a regulatory method are also discussed. An overview of 
phase - appropriate method validation, method verifi cation, and method revalidation 
are presented to stimulate ideas and the thought process to follow when such situations 
are encountered. 
FURTHER READING 
1. International Conference on Harmonization (ICH) ( 2005 , Nov.), Harmonised tripartite 
guideline Q2(R1), Validation of analytical procedures: Text and methodology. 
2. International Conference on Harmonization (ICH) ( 1999 , Oct.), Harmonised tripartite 
guideline Q6A, Specifi cations: Test procedures and acceptance criteria for new drug substances 
and new drug products: Chemical substances. 
3. International Conference on Harmonization (ICH) ( 2000 , Nov.), Harmonised tripartite 
guideline Q7A GMP for active pharmaceutical ingredient. 
4. International Conference on Harmonization (ICH) ( 2006 , Sept.), Guidance for industry: 
Quality systems approach to pharmaceutical cGMP. 
5. Code of Federal Regulations (CFR) , Part 211, Current good manufacturing practice for 
fi nished pharmaceuticals. 
6. Chan , C. C. , et al. , ( 2004 ), Analytical Method Validation and Instrument Performance Veri- 
fi cation , J Wiley , Hoboken, NJ . 
7. U.S. Pharmacopeia (USP) , General Chapter . 1225 . , Validation of compendial procedures, 
USP, Rockville, MD. 
8. U.S. Pharmacopeia (USP) , General Chapter . 1226 . , Verifi cation of compendial procedures, 
USP, Rockville, MD. 

743 
8.2 
ANALYTICAL METHOD VALIDATION 
AND QUALITY ASSURANCE 
Isabel Taverniers , Erik Van Bockstaele , and Marc De Loose 
Institute for Agricultural and Fisheries Research (ILVO), Scientifi c Institute of the Flemish 
Community, Merelbeke, Belgium 
Contents 
8.2.1 Introduction 
8.2.2 Traceability and Measurement Uncertainty 
8.2.2.1 Introduction: Quality of Analytical Results 
8.2.2.2 Role of Method Validation in Traceability and MU 
8.2.2.3 Guidelines on Traceability and Uncertainty of Results 
8.2.2.4 Concept of Traceability 
8.2.2.5 Concept of MU 
8.2.2.6 Different Operational Defi nitions of MU 
8.2.2.7 Approaches to Establish MU 
8.2.2.8 Importance of Traceability and MU 
8.2.2.9 Summary 
8.2.3 Method Validation and Quality Assurance 
8.2.3.1 Role of Method Validation in AQA 
8.2.3.2 Guidelines and Guidance on AQA 
8.2.3.3 Approaches for Evaluating Acceptable Methods of Analysis 
8.2.3.4 Method Performance Characteristics and Criteria Approach 
8.2.3.5 Analytical Quality Assurance 
8.2.4 Summary 
References 
8.2.1 INTRODUCTION 
Credibility of analytical data has never caught the public ’ s eye more than today. 
Rather than on the used techniques and methodologies themselves, attention is 
nowadays paid to the quality and reliability of the fi nal results. This is infl uenced by 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

744 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
a higher demand for regulatory compliance, a higher consciousness of the customer 
— the client wants to know the level of confi dence of the reported result — and 
under impulse of new, more exigent European and international standards such as 
the International Organization for Standardization/International Electrotechnical 
Commission (ISO/IEC) 17025 norm for laboratory accreditation. The underlying 
key principle is comparability of results between laboratories and on a wider, international 
basis. In order for results to be comparable, they must be reported with a 
statement of measurement uncertainty (MU) and they must be traceable to common 
primary references. Methods must be validated to show that they actually measure 
what they are intended to measure — that they are fi t for their specifi c purpose. 
Because validation and quality assurance (QA) apply for a specifi c analytical 
method, it is important to approach each method on a case - by - case basis. An analytical 
method is a complex, multistep process starting with sampling and ending 
with the generation of a result. Although every method has its specifi c scope, application, 
and analytical requirement, the basic principles of validation and QA are 
the same regardless the type of method or sector of application. The information in 
this chapter is mainly taken from analytical chemistry, but it applies to other sectors 
as well. The validation of analytical methods, the establishment of traceability of 
results, and the assessment of MU should be done in a uniform, harmonized way, 
conforming with internationally recognized standards from institutions such as 
Eurachem, the International Union of Pure and Applied Chemistry (IUPAC), or 
ISO. 
It is important to issue a common understanding on the topics of method validation, 
traceability, and uncertainty of measurements. Here, the interrelationships 
between method validation, traceability, and MU of results will be elucidated. 
Throughout the landscape of guidelines and standards, the most relevant information 
is selected, compiled, and summarized. Great importance is attached to the 
different method performance parameters and their defi nitions, ways of expression, 
and approaches for practical assessment. We discuss the role of method validation 
within QA as well as the topics of standardization, internal and external quality 
control (IQC and EQC, respectively), and accreditation and the links between these 
different aspects. 
This chapter provides a general, complete, and up - to - date overview of the topic 
of quality of analytical measurements in the wide sense of the word. It is useful for 
the completely inexperienced scientist as well as those involved in this topic for a 
long time who have lost their way in the labyrinth and are looking for more explanation 
on a particular aspect or deeper insight and knowledge. 
8.2.2 TRACEABILITY AND MEASUREMENT UNCERTAINTY 
8.2.2.1 Introduction: Quality of Analytical Results 
Innumerable types of analytical methods exist in the fi elds of analytical and bioanalytical 
chemistry, biochemistry, biology, clinical biology, and pharmacology and 
related application domains such as forensic, toxicological, environmental, agricultural, 
and food analyses. Regardless of the type of method, scope, and application, 
laboratories must be able to produce reliable data when performing analytical tests 
for a client or for regulatory purposes. Together with the fast development of ana

TRACEABILITY AND MEASUREMENT UNCERTAINTY 745 
lytical methodologies, great importance is nowadays attached to the “ quality ” of the 
measurement data. Quality of analytical measurement data encompasses two essential 
criteria: utility and reliability (Figure 1 ) [1] . Utility means that analytical results 
must allow reliable decision making. A key aspect of reliability or validity of 
results is that they are comparable , whatever their origin. Comparability between 
results in the strict sense is provided by traceability to appropriate standards. Traceability 
to common reference standards underlies the possibility of making a comparison 
(i.e., a distinction) between different results. If results are to be compared 
also in terms of their quantities or levels of analyte, additional information on the 
analytical result is needed: measurement uncertainty . Uncertainty of results arises 
from the combination of all uncertainties of the reference values (to which the 
results are traceable) and all additional uncertainties associated with the measurement 
procedure. Measurement uncertainty and traceability are related concepts 
both defi ning the quality of analytical data (Figure 1 ) [2, 3] . 
Quality of results refl ects adequacy (or inadequacy ) of a method in terms of the 
extent to which the method fulfi lls its requirements or is fi t for its particular analytical 
purpose (see below). Quality is always a relative notion, referring to the requirements 
fi xed beforehand on the basis of national or international regulations or 
customer needs [1, 4] . The need for reliability of analytical data is stressed by the 
fact that measurement results will be used and may form the basis for decision 
making. Unreliable results bring along a high risk for incorrect decisions and may 
lead to higher costs, health risks, illegal practices and so on. Imagine, for instance, 
the consequences if results are false positives or if the uncertainty is much larger 
than reported [1, 5, 6] . 
8.2.2.2 Role of Method Validation in Traceability and MU 
An analysis is a complex multistage investigation of the property values of materials, 
being the identity and the concentration of a specifi c component in a specifi c sample 
FIGURE 1 Relationship between quality, traceability, and MU of results [3] . 
Quality 
Utility 
Reliability 
(validity) 
Comparability 
Traceability MU estimation 
Calibration 
1. Are 2 results 
comparable to, i.e., 
distinguishable, from 
each other? 
Y1 = Y2? 
2. Which of the 2 results 
contains, with a certain level 
of confidence, the highest 
level of analyte? 
Y1>Y2 or Y1 
linearity; if systematic 
trends > 
non - linearity 
• 
Calculate relative signals 
= 
signal/concentration and 
plot them as function of 
concentration 
• 
If horizontal line 
> 
linearity; linearity limits 
correspond to 95% and 
105% relative signal values 
>range 
TABLE 5 Continued 
768

Parameter Defi nition Expression Requirements Practical Assessment 
Ruggedness 
Robustness 
The (intralaboratory tested) 
behavior of analytical 
process when small 
changes in environmental 
and/or operating 
conditions are made 
(generally used term) 
Measure of capacity of 
analytical procedure to 
remain unaffected by 
small but deliberate 
variations in method 
performance parameters, 
which provides an 
indication of its reliability 
during normal usage 
(term used by USP/ICH 
only) 
Ruggedness : 
measure for 
variability (reproducibility 
of results obtained under 
variety of conditions); 
expressed as %RSD 
(interlaboratory ) 
• 
Evaluate effect of different 
small changes in paramters 
(in days, instruments, 
analysts, reagents, material, 
amount of sample material 
used, etc.) individually 
• 
Calculate precision data 
Sensitivity Change in response of 
measuring instrument 
divided by corresponding 
change in stimulus 
Slope of calibration curve 
(arbitrary) 
Source: From refs. 4, 15, 21, 55 – 56, 72, 75, and 76 . 
Note: 
Defi nitions, ways of expression, requirements or acceptance criteria, and guidelines for practical assessment (for more details, see text). 
t p,v is the Student factor corresponding to the confi dence level 1 . 
. 
and 
v degrees of freedom. The symbol p represents the percentile or percentage point of the 
t distribution. For one - sided intervals, p 
= 
1 
. 
. ; for two - sided intervals, p 
= 
1 
. 
. /2. Values of t can be found in the IUPAC nomenclature ( t 
= 2.776 for 
n 
= 5 and 
t 
= 3.182 for 
n 
= 4 at 
p 
= 0.95) [67] . 
X is the mean determined value and n is the number of measurements for which the SD was calculated. If SD data of the certifi ed reference materials are not available, 
95% confi dence limits may be used as an estimate of CRM SD (see second form of formula for z 
score) [21] . 
x 
bl is the mean of the blank measurements, s 
bl is the SD on the blank measurements and S is the sensitivity of the method or the slope of the calibration function. 
The calibration function is the relationship between the measured response x L and the concentration c 
L or amount q 
L [56, 
72, 
95 – 96]. 
769

770 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
Precision relates to the random error of a measurement system (see Figure 8 ) 
and is a component of MU (see also Section 8.2.2 and Figure) [2] . 
Trueness is expressed in terms of bias or percentages of error. Bias is the difference 
between the mean value determined for the analyte of interest and the accepted 
true value or known level actually present [87] . It represents the systematic deviation 
of the measured result from the true result. Method trueness is an indicator 
also for utility and applicability of that method with real samples [88] . 
Different sources of systematic errors contribute to the overall bias (Figure 8 ). 
Thompson and Wood [8] describe persistant bias as the bias affecting all data of the 
analytical system over longer periods of time and being relatively small but continuously 
present. Different components contribute to the persistant bias, such as laboratory 
bias, method bias, and the matrix variation effect. Next to persistant bias, 
the larger run effect is the bias of the analytical system during a particular run 
[4, 8, 15] . 
One or more of these bias components are encountered when analyzing RMs. In 
general, RMs are divided into certifi ed RMs (CRMs, either pure substances/solutions 
or matrix CRMs) and (noncertifi ed) laboratory RMs (LRMs), also called QC 
samples [89] . CRMs can address all aspects of bias (method, laboratory, and run 
bias); they are defi ned with a statement of uncertainty and traceable to international 
standards. Therefore, CRMs are considered useful tools to achieve traceability in 
analytical measurements, to calibrat equipment and methods (in certain cases), to 
monitor laboratory performance, to validate methods, and to allow comparison of 
methods [4, 15, 30] . However, the use of CRMs does not necessarely guarantee 
trueness of the results. The best way to assess bias practically is by replicate analysis 
of samples with known concentrations such as reference materials (see also Section 
8.2.2 ). The ideal reference material is a matrix CRM, as this is very similar to the 
samples of interest (the latter is called matrix matching ). A correct result obtained 
with a matrix CRM, however, does not guarantee that the results of unknown 
samples with other matrix compositions will be correct [4, 89] . 
The usefulness of CRMs for validation (in particular for trueness assessment) 
and traceability purposes has been debated for years. This is illustrated by the enor- 
TABLE 6 Horwitz Function 
Analyte Percent Analyte Ratio Unit Horwitz %RSD AOAC PVM 
100 1 100% 2 1.3 
10 1 . 10 . 1 10% 2.8 2.8 
1 1 . 10 . 2 1% 4 2.7 
0.1 1 . 10 . 3 0.10% 5.7 3.7 
0.01 1 . 10 . 4 100 ppm 8 5.3 
0.001 1 . 10 . 5 10 ppm 11.3 7.3 
0.0001 1 . 10 . 6 1 ppm 16 11 
0.00001 1 . 10 . 7 100 ppb 22.6 15 
0.000001 1 . 10 . 8 10 ppb 32 21 
0.0000001 1 . 10 . 9 1 ppb 45.3 30 
Note: This function if given as an empirical relationship between the precision of an analytical method 
and the concentration of the analyte regardless of the nature of the analyte, matrix, and method used. 
Acceptable RSD R and RSD r values according to [75] and to AOAC International [56, 62] ; % RSD = 
percentage relative standard deviation, PVM = peer - verifi ed methods program. 

mous number of papers published on this topic. We mention here only some interesting 
references [89 – 91] . Examples of the use of pure substance RMs, matrix CRMs, 
or LRMs can be found in the special issues of Accreditation and Quality Assurance 
(volume 9, 2004) and Analytical and Bioanalytical Chemistry (volume 278, 2004) on 
biological and environmental reference materials and the special issue of Trends in 
Analytical Chemistry (volume 23, 2004) on challenges for achieving traceability of 
environmental measurements. 
If no such (certifi ed) reference materials are available, a blank sample matrix of 
interest can be “ spiked ” with a known amount of a pure and stable in - house material, 
called the spike or surrogate . Recovery is then calculated as the percentage of 
FIGURE 8 Composition of error of analytical result related to accuracy of analytical 
method [4, 8] . 
Trueness Precision Indicator for difference between 
expected value and true value 
Indicator for difference between 
result and expeced value 
Analysis of CRMs + 
statistical control 
Duplicate analysis 
Random 
bias 
Minimally needed in method validation 
Single-laboratory validation 
Matrix variation 
effect 
Method 
bias 
Laboratory 
bias 
Run 
bias 
Persistent bias Run effect 
Variations within whole analytical 
system over longer periods 
Variations during 
particular run 
True value 
Analytical result 
Error 
Inaccuracy 
Difference between 
analytical result 
and true value 
Bias 
Systematic error 
Difference between 
expected value and 
true value 
Expected value 
(limiting mean) 
Imprecision 
Random error 
Difference between 
analytical result and 
expected mean value 
Intermediate 
precision 
Reproducibility Repeatability 
Interassay precision= 
variability over a short 
time interval under 
same conditions 
Within-lab variation due 
to random effects= 
variability over longer 
period of time, under 
different conditions 
Interlaboratory 
variation, tested by 
collaborative studies 
METHOD VALIDATION AND QUALITY ASSURANCE 771

772 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
the measured spike of the matrix sample relative to the measured spike of the blank 
control or the amount of spike added to the sample. The smaller the recovery 
percent, the larger the bias which is affecting the method and thus the lower the 
trueness [4, 56, 72, 92, 93] . An indication of trueness can also be obtained by comparing 
the method with a second, well - characterized reference method under the 
condition that the precision of the established reference method is known. Results 
from the two methods, performed on the same sample or set of samples, are compared. 
The samples may be CRMs, in - house standards, or just typical samples [15] . 
For a comparison between the three ways of establishing bias, see also Section 8.2.2 
on analytical quality. It should be clear here that the use of recovery estimates and 
comparing methods are alternative ways which encompass serious limitations. They 
can give an idea about data comparability, but trueness cannot be assured [89] . 
Trueness or exactness of an analytical method can be documented in a control 
chart. Either the difference between the mean and true value of an analyzed (C)RM 
together with confi dence limits or the percentage recovery of the known, added 
amount can be plotted [56, 62] . Here, again, special caution should be taken concerning 
the used reference. Control charts may be useful to achieve trueness only if a 
CRM, which is in principle traceable to SI units, is used. All other types of references 
only allow traceability to a “ consensus ” value, which however is assumed not to be 
necessarely equal to the “ true ” value [89] . The expected trueness or recovery percent 
values depend on the analyte concentration. Therefore, trueness should be estimated 
for at least three different concentrations. If recovery is measured, values should be 
compared to acceptable recovery rates as outlined by the AOAC Peer Verifi ed 
Methods Program (Table 7 ) [56, 62] . Besides bias and percent recovery, another 
measure for the trueness is the z score (Table 5 ). It is important to note that a considerable 
component of the overall MU will be attributed to MU on the bias of a 
system, including uncertainties on reference materials (Figures 5 and 8 ) [2] . 
Recovery is often treated as a separate validation parameter (Table 5 ). Analytical 
methods intend to estimate the true value of the analyte concentration with an 
uncertainty that is fi t for the intended purpose. However, in such analytical methods, 
the analyte is transferred from the complex matrix to a simpler solution whereby 
there is a loss of analyte. As a consequence, the measured value will be lower than 
the true concentration present in the original matrix. Therefore, assessing the effi - 
TABLE 7 Acceptable recovery percentages as a function of the analyte concentration 
[56] 
Analyte Percent Analyte Ratio Unit Mean Recovery (%) 
100 1 100% 98 – 102 
10 1 . 10 . 1 10% 98 – 102 
1 1 . 10 . 2 1% 97 – 103 
0.1 1 . 10 . 3 0.10% 95 – 105 
0.01 1 . 10 . 4 100 ppm 90 – 107 
0.001 1 . 10 . 5 10 ppm 80 – 110 
0.0001 1 . 10 . 6 1 ppm 80 – 110 
0.00001 1 . 10 . 7 100 ppb 80 – 110 
0.000001 1 . 10 . 8 10 ppb 60 – 115 
0.0000001 1 . 10 . 9 1 ppb 40 – 120 

ciency of the method in detecting all of the analyte present is a part of the validation 
process. Eurachem, IUPAC, ISO, and AOAC International state that recovery values 
should always be established as a part of method validation. Recovery or spiking 
studies should be performed for different types of matrices, several examples of 
each matrix type, and each matrix type at different levels of analyte concentration 
[2, 4, 15] . 
Specifi city and Selectivity Specifi city and selectivity both give an idea of the reliability 
of the analytical method (for defi nitions see Table 5 ). Some authors give different 
defi nitions for both terms while for others they are identical. The term specifi c 
generally refers to a method that produces a response for a single analyte only, while 
the term selective is used for a method producing responses for different chemical 
entities or analytes which can be distinguished from each other. A method is called 
selective if the response is distinguished from all other responses. In this case, the 
method is perfectly able to accurately measure an analyte in the presence of interferences 
[56, 94] . According to Eurachem, specifi city and selectivity principally 
refl ect the same characteristic and are related very closely to each other in such a 
way that specifi city means 100% selectivity. In other words, a method can only be 
specifi c if it is 100% selective. Another related term is confi rmation of identity , which 
is the proof that “ the measurement signal, which has been attributed to the analyte, 
is only due to the analyte and not to the presence of something chemically or physically 
similar or arising as coincidence ” [15] . A method must fi rst show high specifi city 
before true quantifi cation can be performed [87] . 
There is no single expression form for specifi city. It is rather something which 
must be demonstrated. The way in which this is done depends on the objective and 
the type of analytical method (see also below). For identifi cation tests, the goal is 
to ensure the identity of an analyte. Specifi city is here the ability to discriminate 
between compounds of closely related structures which can be present. 
For impurity tests (limit impurity test, quantitative impurity test) and assay tests, 
the accent lies on the ability to determine or discriminate for the analyte in the 
presence of other interferants. Selectivity can be assessed by spiking samples with 
possible interferants (e.g., degradation products) [55, 56, 72] . 
Limit of Detection There is no analytical term or parameter for which there exists 
a larger variety in terminology and formulations than for the limits of detection and 
quantifi cation. Limit of detection or detection limit is the terminology most widely 
used as accepted by Eurachem. ISO uses minimum detectable net concentration 
while IUPAC prefers minimum detectable (true) value [15] . All offi cial organizations 
however refer to the same defi nition: the lowest amount of an analyte in a sample 
which can be detected but not necessarily quantifi ed as an exact value. In general, 
the LOD is expressed as a concentration cL or a quantity qL , derived from the smallest 
signal xL which can be detected with reasonable certainty for a given analytical 
procedure. The lowest signal xL is the signal which lies k times SD blank above the 
mean blank value, whereby k is a numerical factor chosen according to the level of 
confi dence required [56, 72, 95 – 97] . 
The larger the value of k , the larger the confi dence level. Eurachem and IUPAC 
recommend a value of 3 for k , meaning that the chance that a signal more than 3s 
above the sample blank value is originating from the blank is less than 1%. The 
METHOD VALIDATION AND QUALITY ASSURANCE 773

774 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
detection limit is thus the concentration or amount corresponding to a measurement 
level (response, signal) three sbl units above the value for zero analyte (Table 5 ). At 
the concentration or amount level of 3 times the sbl , the RSD or coeffi cient of variation 
(CV) on the measured signal is 33% (measure for uncertainty) [2, 4, 15, 75, 95, 
98] . According to USP/ICH, the detection limit corresponds to that signal where the 
signal - to - noise ratio is 2 : 1 or 3 : 1 [72, 85] . 
It is not true — as is often thought — that detection or quantifi cation is impossible 
below the determination limit, but at these lower levels, the uncertainty of the detection/
quantifi cation measurement is higher than the actual value itself [21] . In this 
context, Huber [56] defi nes the limit of detection also as the point at which a measured 
value is larger than the uncertainty associated with it. According to Krull and 
Swartz [88] , the LOD is a concentration point where only the qualitative identifi cation 
is possible but no accurate and precise quantifi cation. 
For qualitative methods, the LOD is defi ned as the threshold concentration at 
which the test becomes unreliable. A series of blank samples, spiked with different 
concentrations of the analyte, are each analyzed at least 10 times. The threshold, or 
cutoff , concentration is determined visually based on a response curve, plotting the 
percentage of positive results versus the concentration. In this respect, the LOD is 
also defi ned as the concentration at which 95% of the experiments give a clearly 
positive signal [15] . 
The LOD may not be confused with the sensitivity of the method. The latter is 
the capability of the method to discriminate small differences in concentration or 
mass of the test analyte and is equal to the slope of the calibration curve (see below) 
[56] . 
Limit of Quantifi cation For the limit of quantifi cation , or limit of determination , 
defi nitions and formulas are very similar to those of LOD, except that for LOQ, k 
is taken to be 5, 6, or even 10 [2, 4, 15, 56, 72, 96] . A value of 10 for k means that 
the %RSD at the limit of quantifi cation is 10%. The LOQ thus corresponds to that 
concentration or amount of analyte quantifi able with a variation coeffi cient not 
higher than 10% [98] . The LOQ is always higher than the LOD and is often taken 
as a fi xed multiple (typically 2) of the detection limit [4] . Also, the determination 
limit is referred to as the signal 10 times above the noise or background signal, corresponding 
to a signal - to - noise ratio of 10 : 1 [72, 85] . 
In practice, the LOQ can be calculated in an analogous way as for the LOD, as 
indicated in Table 5 . An alternative way of practically assessing the LOD and LOQ 
is the following. In a fi rst step, 10 independent sample blanks are measured each 
once, the blank standard deviation sbl is calculated, and the lowest signals corresponding 
to both the LOD and the LOQ are calculated as xLOD = xbl + 3sbl and xLOQ 
= xbl + 10 sbl , respectively. In a second step, sample blanks are spiked with various 
analyte concentrations (e.g., 6) close to the LOD. Per concentration, 10 independent 
replicates are measured and the standard deviation of the measured signals calculated. 
These standard deviations s (or the %RSD) are then plotted against the concentration. 
LOD and LOQ values are those concentrations of analyte corresponding 
to %RSD values of 33 and 10%, respectively [15, 21] . 
As is said for the LOD, neither is it true that at and below the LOQ quantifi cation 
becomes impossible. Quantifi cation is possible; however it becomes unreliable 
as the uncertainty associated with it at these lower levels is higher than the measure

ment value itself. Quantifi cation becomes reliable as soon as the MU is lower than 
the value measured [21] . 
Decision Limit and Detection Capability: For Specifi c Sectors Only In the 
context of analytical method validation, the terms decision limit (CC . ) and detection 
capability (CC . ) as well as minimum required performance limits (MRPLs) are often 
used and need some clarifi cation. These terms are applicable for the measurement 
of organic residues, contaminants, and chemical elements in live animals and animal 
products, as regulated within the European Union (EU) by directives 96/23/EC [99] , 
2002/657/EC [82] , and 2003/181/EC [100] . The Commission distinguishes group A 
substances for which no permitted limit ( maximum residue level , MRL) has been 
established and group B substances having a fi xed permitted limit. 
The decision limit CC . is the limit at and above which it can be concluded with 
an error probability of . that a sample is noncompliant . If a permitted limit (PL) 
has been established for a substance (group B or the regulated compounds), the 
result of a sample is noncompliant if the decision limit is exceeded (CC . = xPL + 
1.64 sMRL ). If no permitted limit has been established (group A), the decision limit 
is the lowest concentration level at which the method can discriminate with a statistical 
certainty of 1 - . that the particular analyte is present (CC . = xbl + 2.33 ssample ). 
The detection capability CC . is the smallest content of the substance that may be 
detected, identifi ed, and/or quantifi ed in a sample with an error probability of . 
(CC. = CC . + 1.65 ssample ). 
Minimum required performance limits have been established for substances for 
which no permitted limit has been fi xed and in particular for those substances whose 
use is not authorized or even prohibited within the EU (group A). A MRPL is the 
minimum content of an analyte in a sample which at least has to be detected and 
confi rmed. A few MRPLs for residues of certain veterinary drugs have been published 
so far in directive 2003/181/EC. 
For group A substances (no PL established), CC . and CC . are comparable with 
LOD and LOQ, respectively, as their concentrations correspond to measured signals 
laying y times above the blank signal. For substances having a PL (group B), CC . 
and CC . are not related to LOD and LOQ but are expressed in relation to this PL. 
It is important to note that these terms apply specifi cally for inspection of animals 
and fresh meat for the presence of residues of veterinary drugs and specifi c contaminants 
and are therefore different from LOD and LOQ [82, 99 – 102] . 
Linearity and Range For assessment of the linearity of an analytical method, 
linear regression calculations do not suffi ce. In addition, residual values should be 
calculated (Table 5 ). The latter represent the differences between the actual y value 
and the y value predicted from the regression curve for each x value. If residual 
values, calculated by simple linear regression, are randomly distributed about the 
regression line, linearity is confi rmed, while systematic trends indicate nonlinearity. 
If such a trend or pattern is observed, this suggests that the data are best treated by 
weighted regression. For either simple or weighted linear regression, linearity supposes 
that the intercept is not signifi cantly different from zero [4, 15, 21, 75] . 
An alternative approach to establish linearity is to divide the response by the 
respective concentrations and to plot these “ relative responses ” as a function of the 
concentration on a log scale. The obtained line should be horizontal over the full 
METHOD VALIDATION AND QUALITY ASSURANCE 775

776 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
linear range, with a positive deviation at low concentrations and a negative deviation 
at higher concentrations. By drawing parallel horizontal lines, corresponding to, for 
example, 95 and 105% of the horizontal relative response line, the intersection 
points can be derived at which the method becomes nonlinear [56] . 
It is important that a linear curve is repeatable from day to day. However, linear 
ranges may be different for different matrices. The reason for this is a possible effect 
of interferences inherent to the matrix. A test for general matrix effects can be 
performed by means of “ standard additions ” or the method of analyte additions. 
For a set of samples, obtained by adding different concentrations of analyte to a 
certain matrix, the slope of the calibration curve is compared with the slope of the 
usual calibration function. A lack of signifi cance (curves are parallel) means that 
there is no matrix effect [21, 75] . 
Ruggedness and Robustness Although the terms ruggedness and robustness are 
often treated as the same and used interchangeably, separate defi nitions exist for 
both terms, as indicated in Table 5 . 
To have an idea about the ruggedness, Eurachem recommends to introduce 
deliberate variations to the method, such as different days, analysts, instruments, 
and reagents and variations in sample preparation or sample material used. Changes 
should be made separately and the effect evaluated of each set of experimental 
conditions on the precision and trueness [4, 15, 85] . To examine the effects of different 
factors, a “ factorial design ” methodology can be applied, as described by von 
Holst et al. [103] . By combining changes in conditions and performing a set of 
experiments, one can determine which factors have a signifi cant or even critical 
infl uence on the analytical results. In ICH/USP guidelines, ruggedness is not defi ned 
separately but treated under the same denominator as reproducibility precision: It 
is “ the degree of reproducibility of the results obtained under a variety of conditions, 
expressed as %RSD ” [56] . 
Robustness is a term introduced by USP/ICH [88] . Although Eurachem has 
included the term robustness in its offi cial list of defi nitions, the term is not used by 
offi cial organizations other than USP/ICH. According to Eurachem, both parameters 
do present the same and are thus synonyms [15, 72, 85] . 
Sensitivity The sensitivity of a method is the gradient of the response curve. In 
practical terms, sensitivity refers to the slope of the calibration curve. Sensitivity is 
often used together with detection and quantifi cation limits. Indeed, the slope of the 
calibration curve is used for the calculation of limits of detection and quantifi cation. 
A method is called sensitive if a small change in concentration or amount of analyte 
causes a large change in the measured signal [4, 15, 21] . Sensitivity is not always 
mentioned as a validation parameter in offi cial guidelines. According to Thompson 
et al. [4] , it is not useful in validation because it is usually arbitrary, depending on 
instrument settings. USP/ICH does not mention sensitivity at all. 
8.2.3.5 Analytical Quality Assurance 
Quality assurance is the complete organizational infrastructure that forms the basis 
for all reliable analytical measurements [8] . It stands for all the planned and systematic 
activities and measures implemented within the quality system [2, 104] . A 
quality system has a quality plan, which emphasizes the implementation of good 

laboratory practices (GLPs). GLPs are comparable to the good manufacturing 
practices (GMPs) and the larger HACCPs (hazard analysis critical control points) 
quality systems of food production factories. Attention goes to all aspects of quality 
management in the laboratory organization, including staff training, the maintenance 
and calibration of all equipment used, the laboratory environment, safety 
measures, the system of sample identifi cation, recordkeeping, and storage — the 
latter may be simplifi ed by the use of a laboratory information management system 
(LIMS), the use of validated and standardized methods, and the documentation of 
these methods and of all information concerning the followed procedures (standard 
operating procedures, or SOPs). 
Quality assurance embraces both QC and “ quality assessment. ” QC is defi ned as 
the mechanism or the practical activities undertaken to control errors, while quality 
assessment is the mechanism to verify that the system is operating within acceptable 
limits. Quality assessment and control measures are in place to ensure that the 
measurement process is stable and under control [2, 8] . 
Within QC, internal and external quality control are distinguished. In general, 
QA comprises the following topics as also schematized in Figure 6 . 
Use of Validated Methods: In -House Versus Interlaboratory Validation Wherever 
possible or practically achievable, a laboratory should use methods which have been 
“ fully validated ” through a collaborative trial, also called interlaboratory study or 
method performance study. Validation in collaborative studies is required for any 
new analytical method before it can be published as a standard method (see below). 
However, single - laboratory validation is a valuable source of data usable to demonstrate 
the fi tness for purpose of an analytical method. In - house validation is of 
particular interest in cases where it is inconvenient or impossible for a laboratory 
to enter into or to organize itself a collaborative study [4, 5] . 
On the one hand, even if an in - house validated method shows good performance 
and reliable accuracy, such a method cannot be adopted as a standard method. In - 
house validated methods need to be compared between at least eight laboratories 
in a collaborative trial. On the other hand, a collaborative study should not be conducted 
with an unoptimized method [58] . Interlaboratory studies are restricted to 
precision and trueness while other important performance characteristics such as 
specifi city and LOD are not addressed [105] . For these reasons, single - laboratory 
validation and interlaboratory validation studies do not exclude each other but must 
be seen as two necessary and complementary stages in a process, presented in Figure 
9 . The added value of single - laboratory validation is that it simplifi es the next 
step — interlaboratory validation — and thereby minimizes the gap between internally 
(validated or not) developed methods and the status of interlaboratory validation. 
By optimizing the method fi rst within the laboratory, as a kind of preliminary 
work, an enormous amount of collaborators ’ time and money is saved [58] . 
The importance of conducting such a single - laboratory preliminary validation 
step is increasingly highlighted by international standardization agencies. The 
IUPAC and AOAC International include a preliminary work paragraph in their 
guidelines for collaborative studies, stating that within - laboratory testing data are 
required on precision, bias, recovery, and applicability. Additionally, a clear description 
of the method, including statements on the purpose of the method, the type of 
method, and the probable use of the method, is required within this preliminary 
work [62, 63] . It is however not only in the harmonized guidelines for collaborative 
METHOD VALIDATION AND QUALITY ASSURANCE 777

778 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
trials (see below) that the link between a single - laboratory prevalidation step and 
the collaborative trial is emphasized. Separate guidelines for single - laboratory validation 
of methods of analysis have recently been published by the IUPAC, ISO, and 
AOAC International [4] . The IUPAC guidelines have also been considered and 
accepted by the CCMAS [77] . In addition, specifi c, individual working groups or 
scientists are presenting their own framework for pre - , single - laboratory validation 
of methods of analysis. The latter do not concern offi cial, published guidelines but 
can be found on the Internet [see, e.g., 58, 85 ] or are distributed through national 
or international specifi c working groups. The objectives of a single - laboratory or 
in - house validation process are depicted in Figure 9 . Depending on the type of 
method (Table 4 ), data can be obtained for all criteria except for the reproducibility 
(interlaboratory) precision. However, it is this among - laboratories variability or 
reproducibility which is the dominating error component in analytical measurement 
and which underlies the need for interlaboratory validation [106] . 
Interlaboratory or collaborative validation studies can be organized by any laboratory, 
institute, or organization but should preferably be conducted according to 
one of the following recognized protocols: (1) ISO 5725 on Accuracy (Trueness and 
Precision) of Measurement Methods and Results [68] or (2) IUPAC Protocol for 
the Design, Conduct and Interpretation of Method - Performance Studies [63] . The 
latter revised, harmonized guidelines have been adopted also by AOAC International 
as the guidelines for the AOAC Offi cial Methods Program [62] . The main 
requirements for collaborative studies outlined in these guidelines are shown in 
Figure 9 . 
Precision plays a central role in collaborative studies. Wood [84] defi nes a collaborative 
trial as “ a procedure whereby the precision of a method of analysis may 
be assessed and quantifi ed. ” Precision is the objective of interlaboratory validation 
studies, and not trueness or whichever other method performance parameter. Evalu- 
FIGURE 9 Hierarchy of, relationship between, and objectives and requirements for prevalidation 
[106] , validation [62, 63, 68] , and standardization of analytical methods [62, 63, 67, 
68, 75, 84] : RSD = relative standard deviation, CV = coeffi cient of variation, SOP = standard 
operating procedure. 
Prevalidation 
Single-laboratory optimization 
Description of analytical system: 
- Purpose of method? 
- Type of analyte? 
- Type of method? 
(Full) validation Standardization 
Interlaboratory or 
collaborative trial: 
Adoption by internationally 
Recognized standardization body 
Applicability/ intended use of method: 
- Type(s) of material/ matrix(matrices) 
- Concentration rate of analyte 
Writing a SOP 
Fixing the analytical requirement 
Evaluation of method performance 
characteristics 
1. ISO 5725 [68] 
2. IUPAC [63] 
Precision data must be given 
in terms of RSD or CV (%) 
Minimum of 5 materials 
Minimum of 8 laboratories 
Both repeatability and reproducibility 
precision data must be given 
Precision data must be documented 
both without and with outliers 
(Cochran test, Grubbs test) 
Precision: calculated values of RSD must be 
in compliance with Horwitz (Horrat) values 
Method has been validated collaboratively 
[63, 68] 
Evaluation of precision and other statistical data 
by an accepted method of statistical analysis 
(Cochran, Grubbs) 
Precision: not more than 1 of the 5 sets of data 
give more than 20% statistically outying results 
Mandatory standard format for text and 
presentation of results

ation of the acceptability of precision data is important for the standardization of 
methods (see below). 
Use of Standardized Methods The fi rst level of AQA is the use of validated 
or standardized methods. The terms validated and standardized here refer to the 
fact that the method performance characteristics have been evaluated and have 
proven to meet certain requirements. At least, precision data are documented, 
giving an idea of the uncertainty and thus of the error of the analytical result. 
In both validated and standardized methods, the performance of the method is 
known. 
Validated methods can be developed by the laboratory itself or by a standardization 
organization after interlaboratory studies. Standardized methods are developed 
by organizations such as the AOAC International, ISO, USP (see Table 3 ), U.S. 
Environmental Protection Agency (EPA), American Society for Testing and Materials 
(ASTM), or Food Standards Association (FSA) [56] . Here exactly lies the difference 
between a validated and a standardized method: An analytical method can 
only be standardized after it has been validated through interlaboratory comparisons. 
The main prerequisite for a standards organization is that a method has been 
adequately studied and its precision shown to meet a required standard, as summarized 
in Figure 9 . The format of a standard method as outlined in ISO Guide 78/2 
[58] is shown in Table 8 [2] . A specifi c IUPAC protocol [67] describes in detail how 
to present AQA data such as the performance characteristics. 
TABLE 8 Mandatory Text Format for Standardized Methods 
1. Scope States briefl y what method determines 
2. Defi nitions Precise defi nition of analyte or parameter determination by 
method 
3. Fields of application Type of materal(s)/matrix(ces) to which method is applicable 
4. Principle Basic steps involved in procedure 
5. Apparatus Specifi c apparatuses required for determination are listed 
6. Reagents Analytical reagent - grade reagents needed for determination 
are listed 
7. Sampling 
8. Procedure Divided into numbered paragraphs or subclauses, includes a 
preparation - of - test - sample step and reference to QA 
procedures 
9. Calculation and 
expression of results 
Indication of how fi nal results are calculated and units in which 
results are to be expressed 
10. Notes Additional information as to procedure; may be in form of 
notes, placed here or in body of text 
Annex Includes all information on analytical quality control, such a 
precision clauses (repeatability and reproducibility data), 
table of statistical data outlining accuracy (trueness and 
precision) of method 
References Reference to published report on collaborative study which was 
carried out prior to standardization of method 
Source: According to ISO Guide 78/2 [2, 58] . 
METHOD VALIDATION AND QUALITY ASSURANCE 779

780 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
Effective Internal Quality Control ( IQC) In the IUPAC harmonized guidelines 
for IQC, Thompson and Wood [8] defi ne IQC as a “ set of procedures undertaken 
by the laboratory staff for the continuous monitoring of operation and the results 
of measurements in order to decide whether results are reliable enough to be 
released ” . IQC guarantees that methods of analysis are fi t for their intended purpose, 
meaning the continuous achievement of analytical results with the required standard 
of accuracy. The objective of IQC is in fact the elongation of method validation: 
to continuously check the accuracy of analytical data obtained from day to day in 
the laboratory. In this respect, both systematic errors, leading to bias, as well as 
random errors, leading to imprecision, are monitored. In order to be able to monitor 
these errors, they should remain constant. Within the laboratory, such constant 
conditions are typically achieved in one analytical run. The word internal in IQC 
implicates that repeatability conditions are achieved. Thus monitoring the precision 
as the objective of IQC concerns not reproducibility or interlaboratory precision 
but only repeatability or intralaboratory precision. In fact, the monitoring of accuracy 
of an analytical method in IQC can be translated into the monitoring of the 
analytical system [8] . 
Two aspects are important for IQC: (1) the analysis of control materials such as 
reference materials or spiked samples to monitor trueness and (2) replication of 
analysis to monitor precision. Of high value in IQC are also blank samples and blind 
samples. Both IQC aspects form a part of statistical control, a tool for monitoring 
the accuracy of an analytical system. In a control chart, such as a Shewhart control 
chart , measured values of repeated analyses of a reference material are plotted 
against the run number. Based on the data in a control chart, a method is defi ned 
either as an analytical system under control or as an analytical system out of control . 
This interpretation is possible by drawing horizontal lines on the chart: x. (mean 
value), x. + s (SD) and x. . s , x. + 2 s (upper warning limit) and x. . 2 s (lower warning 
limit), and x. + 3 s (upper action or control limit) and x. . 3 s (lower action or control 
limit). An analytical system is under control if no more than 5% of the measured 
values exceed the warning limits [2, 6, 85] . 
Participance in Profi ciency Testing Schemes Profi ciency testing (PT) is the periodic 
assessment of the competency or the analytical performance of individual 
participating laboratories [23] . An independent coordinator distributes individual 
test portions of a typical uniform test material. The participating laboratories analyze 
the materials by their method of choice and return the results to the coordinator. 
Test results obtained by different laboratories are subsequently compared with each 
other and the performance of each participant evaluated based on a single competency 
score [64, 107] . International harmonized protocols exist for the organization 
of PT schemes [59, 60, 64, 69, 79] . 
Participation in profi ciency tests is not a prerequisite or an absolute substitute 
for IQC measures or vice versa. However, participance in profi ciency tests is meaningless 
without a well - developed IQC system. IQC underlies participance in PT 
schemes, while IQC and participance in PT schemes are both important substitutes 
of AQA (Figure 6 ). It is shown that laboratories with the strongest QC procedures 
score signifi cantly better in PT schemes [8, 50] . Participance in PT can to a certain 
extent improve the laboratory ’ s performance; however unsatisfactory performance 
in schemes (up to 30% of all participants) has been reported. This means that there 

is no correlation between good analytical performance and participation in PT [108] . 
However, PT has a signifi cant educational function as it helps a laboratory to demonstrate 
competency to an accreditation body or another third party [60] . 
The terms PT schemes and collaborative trials are often confused with each other, 
as in both QA measures, a number of different laboratories are involved. However, 
there is a clear distinction between both. The mean differences with respect to the 
objective and application, the results, used method, test materials, and participating 
laboratories are summarized in Table 9 . It is important to note also that the results 
obtained from PT schemes, as well as those from collaborative performance studies, 
can be used for assessing the MU (see Section 8.2.2 ). 
External Quality Control and Accreditation Participation in PT schemes is an 
objective means of evaluating the reliability of the data produced by a laboratory. 
Another form of external assessment of the laboratory performance is the physical 
inspection of the laboratory to ensure that it complies with externally imposed 
standards. Accreditation of the laboratory indicates that it is applying the required 
TABLE 9 Differences between Method Performance Studies and Profi ciency - Testing 
Schemes 
Characteristic 
Collaborative/Interlaboratory 
(Method Performance) studies Profi ciency - Testing Schemes 
Main Objective Validation of new methods Competency check of analytical 
laboratories 
Application New methods, required for full 
validation and 
standardization, fi rst 
prerequisite for IQC and QA 
Routinely used (validation and/ 
or standardization methods), 
recommended within IQC and 
QA system 
Results aimed at Precision : multiple results, both 
repeatability and 
reproducibility; > %RSD is 
compared to theoretical 
Horwitz Horrat values 
Trueness : Single result per test 
material, > calculation of Z 
score as measure for bias 
Method and protocol 
used 
Single prescribed method for 
which SOP must strictly be 
followed 
Multiplicity of methods; 
participants have free choice 
of (validated and/or 
standardized) method 
Test materials Minimum of fi ve different 
materials, no stipulations for 
homogeneity and stability of 
test samples 
No minimum, per round often 
less than fi ve test samples; 
homogeneity and stability of 
materials must be assured 
Participating 
laboratories 
Minimum of eight, assumed to 
be equally competent 
No minimum, variety in 
participants is possible 
throughout one scheme 
(different rounds); not 
assumed to have equal 
competency (will be tested) 
Source: From refs. 8, 78, and 107 . 
METHOD VALIDATION AND QUALITY ASSURANCE 781

782 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
QA principles. The “ golden standard ” ISO/IEC 17025 [17] , which is the revised 
version of ISO Guide 25 [70] , describes the general requirements for the competence 
of calibration and testing laboratories. In Europe, the accreditation criteria 
have been formalized in European standard EN45001 [109] . Participation in PT 
schemes forms the basis for accreditation, because PT is a powerful tool for a laboratory 
to demonstrate its competency. Accreditation guides use the information 
obtained by PT schemes [6, 17, 60, 64] . 
Accreditation is a formal recognition that a laboratory is competent to carry out 
specifi c (types of) calibrations or tests [2] . After the use of validated and standardized 
methods, the introduction and use of appropriate IQC procedures and the 
participation in PT schemes, accreditation to ISO/IEC 17025 is the fourth basic 
principle related to laboratory QA in general [4] . Guidelines on the implementation 
of ISO/IEC 17025, including the estimation of MU (see also Section 8.2.2 ), are 
published in the literature and by offi cial accreditation bodies such as Eurachem, 
CITAC, EA, Eurolab, and ILAC (see Table 1 ) [2, 60, 80, 81, 110] . It is worthwile to 
mention that accreditation, just like participance in PT schemes, does not necessarily 
indicate good performance of the laboratory [108] . 
8.2.4 SUMMARY 
Together with the fast development of analytical methodologies, great importance 
is nowadays attached to the quality of the measurement data. Besides the necessary 
reporting of any result with its MU and traceability of the results to stated standards 
or references (Section 8.2.2 ), a third crucial aspect of analytical methods of whichever 
type is their status of validation. It is internationally recognized that validation 
is necessary in analytical laboratories. However, less is known about what is validation 
and what should be validated, why validation is important, when and by whom 
validation is performed, and fi nally, how it is carried out practically. This Chapter 
has tried to answer these questions. 
Method validation is defi ned in detail and different approaches to evaluate the 
acceptability of analytical methods are described. Great importance is attached to 
the different method performance parameters, their defi nitions, ways of expression, 
and approaches for practical assessment. Validation of analytical methodologies is 
placed in the broader context of QA. The topics of standardization, internal and 
external quality control, and accreditation are discussed as well as the links between 
these different aspects. Because validation and QA apply for a specifi c analytical 
method, it is important to approach each method on a case - by - case basis. An analytical 
method is a complex, multistep process, starting with sampling and ending 
with the generation of a result. Although every method has its specifi c scope, application, 
and analytical requirement, the basic principles of validation and QA are 
the same, regardless of the type of method or sector of application. The information 
in this work is mainly taken from analytical chemistry, but it applies to other sectors 
as well. 
Section 8.2.3 on quality in the analytical laboratory provides a good, complete, 
and up - to - date collation of relevant information in the fi elds of analytical method 
validation and QA. It is useful for the completely inexperienced scientist as well as 
for those involved in this topic for a long time but somewhere having lost the way 

in the labyrinth, looking for more explanation on a particular aspect, or longing for 
deeper insight and knowledge. 
ACKNOWLEDGMENTS 
The authors wish to thank Simon Kay for preliminary discussions on the topic, Janna 
Puumalainen for reading and commenting on early versions of the chapter, Andrew 
Damant for giving suggestions, Arne Holst - Jensen for refreshing ideas, and Friedle 
Vanhee for many helpful discussions, reading, and assistance throughout the writing 
process. 
REFERENCES 
1. van Zoonen , P. , Hoogerbrugge , R. , Gort , S. M. , Van de Wiel , H. J. , and Van ’ t Klooster , 
H. A. ( 1999 ), Some practical examples of method validation in the analytical laboratory , 
Trends Anal. Chem. , 18 , 584 – 593 . 
2. CITAC/Eurachem Guide ( 2002 ), Guide to Quality in Analytical Chemistry — An Aid to 
Accreditation , available: http://www.eurachem.org . 
3. Eurachem/CITAC Guide ( 2003 ), Traceability in Chemical Measurement. A Guide to 
Achieving Comparable Results in Chemical Measurement , Joint Eurachem/CITAC 
Working Group on Measurement Uncertainty and Traceability, available: http://www. 
eurachem.org . 
4. Thompson , M. , Ellison , S. , and Wood , R. ( 2002 ), Harmonised guidelines for single - 
laboratory validation of methods of analysis, IUPAC Technical Report , Pure Appl. 
Chem. , 74 , 835 – 855 . 
5. Battaglia , R. ( 1996 ), Quality assurance in a food analytical laboratory — The introduction 
of EN 45001 in the food analytical laboratories of a retail company , Accred. Qual. Assur. , 
1 , 256 – 261 . 
6. Mesley , R. J. , Pocklington , W. D. , and Walker , R. F. ( 1991 ), Analytical quality assurance: 
A review , Analyst , 116 , 975 – 990 . 
7. Eurachem Guide EEE/RM/062rev3 ( 2002 ), The Selection and Use of Reference Materials. 
A Basic Guide for Laboratories and Accreditation Bodies, available: http://www. 
eurachem.org . 
8. Thompson, M. , and Wood, R. (1995), Harmonised guidelines for internal quality control 
in analytical chemistry laboratories , Pure Appl. Chem. , 67 , 649 – 656 . 
9. Maroto , A. , Boqu e , R. , Riu , J. , and Rius , F. X. ( 1999 ), Evaluating uncertainty in routine 
analysis , Trends Anal. Chem. , 18 , 577 – 584 . 
10. Hund , E. , Massar , D. L. , and Smeyers - Verbeke , J. ( 2001 ), Operational defi nitions of 
uncertainty , Trends Anal. Chem. , 20 , 394 – 406 . 
11. Fleming , J. , Albus , H. , Neidhart , B. , and Wegschieder , W. ( 1996 ), Glossary of analytical 
terms (II) , Accred. Qual. Assur. , 1 , 87 – 88 . 
12. Quevauviller , Ph. ( 2004 ), Traceability of environmental chemical measurements , Trends 
Anal. Chem. , 23 , 171 – 176 . 
13. Eurachem/CITAC Guide ( 1995 ), in Ellison, S. L. R., Rosslein, M., and Williams, A., Eds., 
Quantifying Uncertainty in Analytical Measurement , 1st ed., available: http://www. 
eurachem.org . 
REFERENCES 783

784 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
14. Eurachem/CITAC Guide ( 2000 ), in Ellison, S. L. R., Rosslein, M., and Williams, A., Eds., 
Quantifying Uncertainty in Analytical Measurement , 2nd ed., available: http://www. 
eurachem.org . 
15. Eurachem Guide ( 1998 ), The Fitness for Purpose of Analytical Methods. A Laboratory 
Guide to Method Validation and Related Topics , LGC, Teddington, available: http://www. 
eurachem.org . 
16. International Organization for Standardization (ISO) GUM ( 1995 ), Guide to the 
Expression of Uncertainty in Measurement , ISO , Geneva . 
17. International Organization for Standardization (ISO)/IEC 17025 ( 1999 ), General 
Requirements for the Competence of Calibration and Testing Laboratories , ISO , 
Geneva . 
18. CX/MAS 01/8 ( 2001 ), Codex Alimentarius Commission, Codex Committee on 
Methods of Analysis and Sampling (FAO/WHO), Measurement uncertainty. Relationship 
between the analytical result, the measurement uncertainty and the specifi cation 
in Codex standards, agenda item 4a of the 23rd session, Budapest, Hungary, Feb. 26 – Mar. 
2, 2001. 
19. CX/MAS 02/6 ( 2002 ), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling (FAO/WHO), Proposed draft guidelines on measurement 
uncertainty, agenda item 5 of the 24th session, Budapest, Hungary, Nov. 18 – 22, 2002. 
20. CX/MAS 02/13 ( 2002 ), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling (FAO/WHO), The use of analytical results: Sampling, relationship 
between the analytical results, the measurement uncertainty, recovery factors and 
the provisions in Codex standards, agenda item 9 of the 24th session, Budapest, Hungary, 
Nov. 18 – 22, 2002. 
21. CX/MAS 02/4 ( 2002 ), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling (FAO/WHO), Proposed draft guidelines for evaluating acceptable 
methods of analysis, agenda item 4a of the 24th session, Budapest, Hungary, Nov. 
18 – 22, 2002. 
22. European Cooperation for Accreditation of Laboratories (EAL) ( 1996 ), The Expression 
of Uncertainty in Quantitative Testing , EAL - G23 , The Netherlands . 
23. International Laboratory Accreditation Cooperation (ILAC) ( 2002 ), Introducing the 
Concept of Uncertainty of Measurement in Testing in Association with the Application of 
the Standard ISO/IEC 17025 , ILAC - G17, ILAC Technical Accreditation Issues Committee, 
available: www.ilac.org . 
24. Valcarc e l , M. , and Rios , A. ( 1999 ), Traceability in chemical measurements for the end 
users , Trends Anal. Chem. , 18 , 570 – 576 . 
25. Walsh , M. C. ( 1999 ), Moving from offi cial to traceable methods , Trends Anal. Chem. , 18 , 
616 – 623 . 
26. Valcarc e l , M. , and Rios , A. ( 1997 ), Is traceability an exclusive property of analytical 
results? An extended approach to traceability in chemical measurement , Fresen. J. Anal. 
Chem. , 359 , 473 – 475 . 
27. Quevauviller , Ph. , and Donard , O. F. X. ( 2001 ), Stated references for ensuring traceability 
of chemical measurements for long - term environmental monitoring , Trends Anal. 
Chem. , 20 , 600 – 613 . 
28. Pan , X. R. ( 1996 ), The traceability scheme in chemical measurement , Accred. Qual. 
Assur. , 1 , 181 – 185 . 
29. Charlet , P. , and Marschal , A. ( 2004 ), Improvement in the traceability of environmental 
analysis by the relevant use of certifi ed pure solutions and a matrix certifi ed reference 
material , Trends Anal. Chem. , 23 , 178 – 184 . 

30. Segura , M. , Camara , C. , Madrid , C. , Rebollo , C. , Azcarate , J. , Kramer , G. N. , Gawlik , 
B. M. , Lamberty , A. , and Quevauviller , Ph. ( 2004 ), Certifi ed reference materials (CRMs) 
for quality control of trace - element determinations in wastewaters , Trends Anal. Chem. , 
23 , 194 – 202 . 
31. F o rstner , U. ( 2004 ), Traceability of sediment analysis , Trends Anal. Chem. , 23 , 217 – 
236 . 
32. Theocharopoulos , S. P. , Mitsios , I. K. , and Arvanitoyannis , J. ( 2004 ), Traceability of environmental 
soil measurements , Trends Anal. Chem. , 23 , 237 – 251 . 
33. Sab e , R. , and Rauret , G. ( 2004 ), Challenges for achieving traceability of analytical measurements 
of heavy metals in environmental samples by isotopic dilution mass spectrometry 
, Trends Anal. Chem. , 23 , 273 – 280 . 
34. Drolc , A. , Ros , M. , and Cotman , M. ( 2004 ), Establishment of traceability of ammonium 
nitrogen determination in wastewater , Anal. Bioanal. Chem. , 378 , 1243 – 1250 . 
35. Thompson , M. ( 1998 ), Uncertainty of sampling in chemical analysis , Accred. Qual. Assur. , 
3 , 117 – 121 . 
36. Roy , S. , and Fouillac , A. - M. ( 2004 ), Uncertainties related to sampling and their impact 
on the chemical analysis of groundwater , Trends Anal. Chem. , 23 , 185 – 193 . 
37. Analytical Methods Committee ( 1995 ), Uncertainty of measurement: Implications of its 
use in analytical science , Analyst , 120 , 2303 – 2308 . 
38. Fleming , J. , Albus , H. , Neidhart , B. , and Wegschieder , W. ( 1997 ), Glossary of analytical 
terms (VIII) , Accred. Qual. Assur. , 2 , 160 – 161 . 
39. Hund , E. , Massart , D. L. , and Smeyers - Verbeke , J. ( 2003 ), Comparison of different 
approaches to estimate the uncertainty of a liquid chromatographic assay , Anal. Chim. 
Acta. , 480 , 39 – 52 . 
40. Eurachem/EA Guide 04/10 ( 2002 ), Accreditation for Microbiological Laboratories , 
available: http://www.eurachem.org . 
41. K u ppers , S. ( 1998 ), Is the estimation of measurement uncertainty a viable alternative to 
validation? Accred. Qual. Assur. , 3 , 412 – 415 
42. Ellison , S. L. R. , and Barwick , V. J. ( 1998 ), Using validation data for ISO measurement 
uncertainty estimation. Part 1. Principles of an approach using cause and effect analysis , 
Analyst , 123 , 1387 – 1392 . 
43. Moser , J. , Wegscheider , W. , and Sperka - Gottlieb , C. ( 2001 ), Quantifying the measurement 
uncertainty of results from environmental analytical methods , Fresen. J. Anal. 
Chem. , 370 , 679 – 689 . 
44. Dehouck , P. , Vander Heyden , Y. , Smeyers - Verbeke , J. , Massart , D. L. , Crommen , P. H. , 
Marini , R. D. , Smeets , O. S. , Decristoforo , G. , Van de Wauw , W. , De Beer , J. , Quaglia , 
M. G. , Stella , C. , Veuthey , J. - L. , Estevenon , O. , Van Schepdael , A. , Roets , E. , and Hoogmartens 
, J. ( 2003 ), Determination of uncertainty in analytical measurements from collaborative 
study results on the analysis of a phenoxymethylpenicillin sample , Anal. 
Chim. Acta , 481 , 261 – 272 . 
45. Maroto , A. , Riu , J. , Boqu e , R. , and Rius , F. X. ( 1999 ). Estimating uncertainties of analytical 
results using information from the validation process , Anal. Chim. Acta , 391 , 
173 – 185 . 
46. Maroto , A. , Boqu e , R. , Riu , J. , and Rius , F. X. ( 2001 ), Estimation of measurement uncertainty 
by using regression techniques and spiked samples , Anal. Chim. Acta , 446 , 
133 – 145 . 
47. Barwick , V. J. , and Ellison , S. L. R. ( 2000 , Jan.), VAM Project 3.2.1 : Development and 
Harmonisation of Measurement Uncertainty Principles. Part d. Protocol for Uncertainty 
Evaluation from Validation Data , LGC, UK, available: http://www.vam.org.uk . 
REFERENCES 785

786 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
48. Armishaw , P. (2003), Estimating measurement uncertainty in an afternoon. A case study 
in the practical application of measurement uncertainty , Accred. Qual. Assur. , 8 , 
218 – 242 . 
49. Horwitz , W. ( 1982 ), Evaluation of analytical methods used for regulation of food and 
drugs , Anal. Chem. , 54 , 67A – 76A . 
50. Thompson , M. , and Lowthian , P. J. ( 1997 ), The Horwitz function revisited , J. AOAC Int. , 
80 , 676 – 679 . 
51. King , B. ( 2001 ), Meeting the measurement uncertainty and traceability requirements of 
ISO/IEC standard 17025 in chemical analysis , Fresen. J. Anal. Chem. , 371 , 714 . 
52. Mueller - Harvey , I. ( 2003 ), Do we need quality assurance and quality control of analytical 
measurements in R & D laboratories? Food Agric. Environ. , 1 , 9 – 11 . 
53. R o sslein , M. ( 2000 ), Evaluation of uncertainty utilising the component by component 
approach , Accred. Qual. Assur. , 5 , 88 – 94 . 
54. Mueller , N. ( 2002 ), Introducing the concept of uncertainty of measurement in testing in 
association with the application of the standard ISO/IEC 17025 , Accred. Qual. Assur. , 7 , 
79 – 80 . 
55. International Organization on Harmonization ( 1995 ), Guideline for industry: Text on 
validation of analytical procedures, ICH - Q2A, available: http://www.fda.gov/cder/guidance/
index.htm . 
56. Huber , L. ( 1998 ), Validation of analytical methods. Part 9, in Huber, L., Ed., Validation 
and Qualifi cation in Analytical Laboratories, Interpharm Press, Agilent Technologies, 
available: http://www.labcompliance.com . 
57. Wells , R. J. ( 1998 ), Validation requirements for chemical methods in quantitative analysis 
— Horses for courses? Accred. Qual. Assur. , 3 , 189 – 193 . 
58. Green , M. ( 1996 ), A practical guide to analytical method validation , Anal. Chem. , 68 , 
305A – 309A . 
59. Eurachem Guide ( 2000 ), Selection, Use and Interpretation of Profi ciency Testing (PT) 
Schemes by Laboratories , available: http://www.eurachem.org . 
60. Eurachem/Eurolab/EA Guide EA - 03/04 ( 2001 , Aug.), Use of Profi ciency Testing as a 
Tool for Accreditation in Testing , available: http://www.eurachem.org . 
61. European Committee for Normalisation (CEN) , ( 2006 ), Standardization and related 
activities — General Vocabulary (from ISO/IEC Guide 2:2004), EN45020:2006, www. 
cenorm.be. 
62. Association of Offi cial Analytical Chemists (AOAC) International ( 2000 ), Method validation 
programs (OMA/PVM Department), including Appendix D: Guidelines for collaborative 
study procedures to validate characteristics of a method of analysis, available: 
http://www.aoac.org/vmeth/devmethno.htm . 
63. Horwitz , W. ( 1995 ), Protocol for the design, conduct and interpretation of method performance 
studies , Pure Appl. Chem. , 67 , 331 – 343 . 
64. Thompson , M. , and Wood , R. ( 1993 ), The international harmonized protocol for the 
profi ciency testing of (chemical) analytical laboratories , Pure Appl. Chem. , 65 , 2123 – 
2144 : also J. AOAC Int ., 76 , 26 – 940 ( 1993 ). 
65. Currie , L. A. ( 1995 ), Nomenclature in evaluation of analytical methods including detection 
and quantifi cation capabilities , Pure Appl. Chem. , 67 , 1699 – 1723 . 
66. Currie , L. A. , and Svehla , G. ( 1994 ), Nomenclature for the presentation of results of 
chemical analysis, International Union of Pure and Applied Chemistry (IUPAC) , Pure 
Appl. Chem. , 66 , 595 – 608 . 
67. Pocklington , W. D. ( 1990 ), Harmonized protocols for the adoption of standardized 
analytical methods and for the presentation of their performance characteristics, 

International Union of Pure and Applied Chemistry (IUPAC) , Pure Appl. Chem. , 
62 , 149 – 162 . 
68. International Organization for Standardization (ISO) Guide 5725 ( 1994 ), Accuracy 
(Trueness and Precision) of Measurement Methods and Results , ISO , Geneva . 
69. International Organization for Standardization (ISO) Guide 43 ( 1984 ), Development 
and Operation of Laboratory Profi ciency Testing , ISO , Geneva . 
70. International Organization for Standardization (ISO) Guide 25 ( 1990 ), General Requirements 
for the Competence of Calibration and Testing Laboratories , ISO , Geneva . 
71. International Organization for Standardization (ISO)/IEC 17025 ( 1999 ), General 
Requirements for the Competence of Calibration and Testing Laboratories , ISO , Geneva . 
72. International Conference on Harmonization ( 1996 ), Guidance for Industry: Validation 
of Analytical Procedures: Methodology , ICH - Q2B, available: http://www.fda.gov/cder/ 
guidance/index.htm . 
73. International Conference on Harmonization ( 1999 ), Harmonised Tripartite Guideline: 
Specifi cations: Test Procedures and Acceptance Criteria for Biotechnological/Biological 
Products , ICH - Q6B. 
74. U.S. Food and Drug Administration (FDA)/CDER/CVM ( 2001 ), Guidance for industry 
— Bioanalytical method validation, available: http://www.fda.gov/cder/guidance/ 
index.htm . 
75. CX/MAS 01/4 (2001), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling, Criteria for evaluating acceptable methods of analysis for Codex 
purposes, agenda item 4a of the 23rd session, Budapest, Hungary, Feb. 26 – Mar. 2, 2001. 
76. CX/MAS 02/5 ( 2002 ), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling (FAO/WHO), Criteria for evaluating acceptable methods of 
analysis for Codex purposes, agenda item 4b of the 24th session, Budapest, Hungary, 
Nov. 18 – 22, 2002. 
77. CX/MAS 02/11 ( 2002 ), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling (FAO/WHO), Requirements for single - laboratory validation 
for Codex purposes, agenda item 8b of the 24th session, Budapest, Hungary, Nov. 18 – 22, 
2002. 
78. CX/MAS 02/12 ( 2002 ), Codex Alimentarius Commission, Codex Committee on Methods 
of Analysis and Sampling (FAO/WHO), Validation of methods through the use of results 
from profi ciency testing schemes, agenda item 8c of the 24th session, Budapest, Hungary, 
Nov. 18 – 22, 2002. 
79. International Laboratory Accredidation Cooperation (ILAC) ( 2000 ), Guidelines for the 
requirements for the competence of providers of profi ciency testing schemes, ILAC - G13 
ILAC Technical Accreditation Issues Committee, available: www.ilac.org . 
80. International Laboratory Accredidation Cooperation (ILAC) ( 2001 ), Guidance for 
accreditation to ISO/IEC 17025, ILAC - G15, ILAC Technical Accreditation Issues Committee, 
available: www.ilac.org . 
81. International Laboratory Accredidation Cooperation (ILAC) ( 2002 ), The scope of 
accreditation and consideration of methods and criteria for the assessment of the scope 
in testing, ILAC - G18, ILAC Technical Accreditation Issues Committee, available: www. 
ilac.org . 
82. European Commission (EC) ( 2002 ), Commission decision 2002/657/EC implementing 
council directive 96/23/EC concerning the performance of analytical methods and the 
interpretation of results , Off. J. Eur. Commun. , L 221/8, 17.8.2002. 
83. Holst - Jensen , A. , and Berdal , K. G. ( 2004 ), The modular analytical procedure and validation 
approach and the units of measurement for genetically modifi ed materials in foods 
and feeds , J. AOAC Int. , 87 , 1 – 9 . 
REFERENCES 787

788 ANALYTICAL METHOD VALIDATION AND QUALITY ASSURANCE 
84. Wood , R. ( 1999 ), How to validate analytical methods , Trends Anal. Chem. , 18 , 
624 – 632 . 
85. Waters Corporation ( 2004 ), Validation Guidelines: Terminology and Defi nitions, available: 
http://www.waters.com/WatersDivision . 
86. Hibbert , D. B. ( 1999 ), Method validation of modern analytical techniques , Accred. Qual. 
Assur. , 4 , 352 – 356 . 
87. Fleming , J. , Neidhart , B. , Albus , H. , and Wegscheider , W. ( 1996 ), Glossary of analytical 
terms (III) , Accred. Qual. Assur. , 1 , 135 . 
88. Krull , I. S. , and Swartz , M. ( 1999 ), Analytical method development and validation for 
the academic researcher , Anal. Lett. , 32 , 1067 – 1080 . 
89. Quevauviller , Ph. ( 2004 ), Traceability of environmental chemical measurements , Trends 
Anal. Chem. , 23 , 171 – 176 . 
90. Holcombe , G. , Lawn , R. , and Sargent , M. ( 2004 ), Improvements in effi ciency of production 
and traceability for certifi cation of reference materials , Accred. Qual. Assur. , 9 , 
198 – 204 . 
91. Lauwaars , M. , and Anklam , E. ( 2004 ), Method validation and reference materials , 
Accred. Qual. Assur. , 9 , 253 – 258 . 
92. Fleming , J. , Albus , H. , Neidhart , B. , and Wegschieder , W. ( 1996 ), Glossary of analytical 
terms (IV) , Accred. Qual. Assur. , 1 , 191 . 
93. Thompson , M. , Ellison , S. L. R. , Fajgelj , A. , Willets , P. , and Wood , R. ( 1999 ), Harmonised 
guidelines for the use of recovery information in analytical measurement , Pure Appl. 
Chem. , 71 , 337 – 348 . 
94. Vessman , J. ( 1996 ), Selectivity or specifi city? Validation of analytical methods from the 
perspective of an analytical chemist in the pharmaceutical industry , J. Pharm. Biomed. 
Anal. , 14 , 867 – 869 . 
95. Analytical Methods Committee ( 1987 ), Recommendations for the defi nition, estimation 
and use of the detection limit , Analyst , 112 , 199 – 204 . 
96. Fleming , J. , Albus , H. , Neidhart , B. , and Wegschieder , W. ( 1997 ) Glossary of analytical 
terms (VII) , Accred. Qual. Assur. , 2 , 51 – 52 . 
97. Huber , W. ( 2003 ), Basic calculations about the limit of detection and its optimal determination 
, Accred. Qual. Assur. , 8 , 213 – 217 . 
98. Kuselman , I. , and Sherman , F. ( 1999 ), Assessment of limits of detection and quantifi cation 
using calculation of uncertainty in a new method for water determination , Accred. 
Qual. Assur. , 4 , 124 – 128 . 
99. European Commission (EC) , Council directive 96/23/EC of April, 29 1996, on measures 
to monitor certain substances and residues thereof in live animals and animal products 
and repealing directives 85/258/EEC and 86/469/EEC and decisions 89/187/EEC and 
91/664/EEC, Off. J. Eur. Commun. , L125, 0010 – 0032, 23.05.1996. 
100. European Commission (EC) , Commission decision 2003/181/EC amending decision 
2002/657/EC as regards the setting of minimum required performance limits (MRPLs) 
for certain residues in food of animal origin, Off. J. Eur. Commun. , L 71/17, 0017 – 0018, 
15.3.2003. 
101. Antignac , J. - P. , Le Bizec , B. , Monteau , F. , and Andre , F. ( 2003 ), Validation of analytical 
methods based on mass spectrometric detection according to the “ 2002/657/EC ” European 
decision: Guideline and application , Anal. Chim. Acta , 483 , 325 – 334 . 
102. De Wasch , K. , De Brabander , H. F. , Courtheyn , D. , Van Hoof , N. , Poelmans , S. , and 
Noppe , H. ( 2003 ), The commission decision 2002/657/EC: Disscussion on some new 
analytical aspects. EURO FOOD CHEM XII, in Strategies on Safe Food, Proceedings , 
Vol. 1, Brugge, Belgium, Sept. 24 – 26, 2003, pp. 45 – 48. 

103. von Holst , C. , M u ller , A. , Bj o rklund , E. , and Anklam , E. ( 2001 ), In - house validation of 
a simplifi ed method for the determination of PCB ’ s in food and feedingstuffs , Eur. Food 
Res. Technol. , 213 , 154 – 160 . 
104. Prichard , E. , Albus , H. , Neidhart , B. , and Wegscheider , W. ( 1997 ), Glossary of analytical 
terms (IX) , Accred. Qual. Assur. , 2 , 348 – 349 . 
105. van der Voet , H. , Van Rhijn , J. A. , Van de Wiel , H. J. ( 1999 ), Inter - laboratory, time, and 
fi tness - for - purpose aspects of effective validation , Anal. Chim. Acta , 391 , 159 – 171 . 
106. Horwitz , W. , and Albert , R. ( 1996 ), Reliability of the determinations of polychlorinated 
contaminants (bifenyls, dioxins, furans) , J. AOAC Int. , 79 , 589 – 621 . 
107. Thompson , M. , and Lowthian , P. J. ( 1993 ), Effectiveness of analytical quality control is 
related to the subsequent performance of laboratories in profi ciency tests , Analyst , 118 , 
1495 – 1500 . 
108. King , B. , Boley , N. , Kannan , G. ( 1999 ), The correlation of laboratory performance in 
profi ciency testing with other QA characteristics , Accred. Qual. Assur. , 4 , 280 – 291 . 
109. The Joint European Standard Institution ( 1989 ), General Criteria for the Operation of 
Testing Laboratories , EN45001, CEN/CENELEC , Brussels . 
110. Pritzkow , J. ( 2003 ), Practical experience of the laboratories in implementing the ISO/ 
IEC 17025 , Accred. Qual. Assur. , 8 , 25 – 26 . 
REFERENCES 789


791 
8.3 
VALIDATION OF LABORATORY 
INSTRUMENTS 
Herman Lam 
Wild Crane Horizon, Inc., Scarborough, Ontario, Canada 
Contents 
8.3.1 Introduction 
8.3.2 Scope 
8.3.3 Laboratory Instrument Classifi cations 
8.3.4 Validation Phases 
8.3.4.1 Planning and Requirements Phase 
8.3.4.2 Qualifi cation (Testing) Phase 
8.3.4.3 Operational Phase 
8.3.4.4 End of Life 
8.3.5 Summary 
References 
8.3.1 INTRODUCTION 
The reliability of chemical and physical measurements is critically dependent on the 
suitability and performance of the instruments from which the measurements are 
obtained. It is a challenge for any laboratory to develop a pragmatically structured 
validation program for laboratory instruments of varying complexity. However, 
implementation of such a program is highly valuable as it provides assurances that 
instruments meet performance requirements and are suitable for their intended use. 
These assurances are required to comply with the good manufacturing practices 
(GMPs) and good laboratory practices (GLPs). 
In recent years, tremendous efforts have been put forth by many organizations 
to address the validation of a wide variety of laboratory instruments of varying 
complexity. Instrument validation – related topics have been discussed at great length 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

792 VALIDATION OF LABORATORY INSTRUMENTS 
in meetings and conferences. Many guidance documents and reference literatures 
on the topic of the validation of laboratory instruments are available [1 – 11] . Several 
key concepts and a common framework for the validation of laboratory instruments 
have been consolidated into the form presented herein: 
1. The fundamental purpose of laboratory instrument validation is to 
provide assurances that the instrument is suitable for its intended use. The 
assurance is supported by documented evidence that the system consistently 
performs according to predetermined specifi cations for its intended 
applications. 
2. The validation of laboratory instruments is the responsibility of the user ’ s 
organization and not the suppliers of the instruments. The users can partner 
with the suppliers to facilitate installation and validation testing. 
3. Most of the instruments used in the laboratory are commercial off - the - shelf 
(COTS) instruments, and consequently the users have little or no input into 
their design. A full system development life - cycle (SDLC) approach [8] , which 
is used to develop complex computerized systems such as Laboratory Information 
Management System (LIMS) or Chromatographic Data System (CDS) 
or custom design laboratory equipment, is not appropriate for COTS instruments. 
Some laboratory instruments such as a pH meter or centrifuge are fairly 
simple and therefore do not warrant the SDLC approach. 
4. Scalable validation for laboratory instruments is being developed to provide 
appropriate levels of coverage for laboratory instruments of various complexities 
for their intended use. 
5. Laboratory instruments are classifi ed into different categories according 
to their complexity to enable a scalable validation approach. The expected 
validation deliverables and documentation requirements for each class of 
laboratory instrument can be defi ned in the validation guide for the 
laboratory. 
6. Validation efforts should be commensurate with the complexity of the different 
classes of instruments as defi ned in the validation guide developed for the 
laboratory. 
8.3.2 SCOPE 
This chapter provides an overview of laboratory instrument validation for GMPs 
and GLPs. The term GxP is used as a general referral to GMP and GLP. The focus 
of this chapter is on the validation of COTS instruments using a scalable 
approach. 
The validation of large - scale computerized systems such as LIMS, chromatographic 
data systems, and custom - designed systems (bespoke systems) are 
beyond the scope of this chapter. Readers are encouraged to consult GAMP 
(Good Automated Manufacturing Practice) Guide for Validation of Automated 
System in Pharmaceutical Manufacturing , Version 4 [8] for the validation for those 
systems. 

8.3.3 LABORATORY INSTRUMENT CLASSIFICATIONS 
A practical way to provide an appropriate level of validation for a wide range of 
laboratory instruments is to develop a classifi cation system to categorize the laboratory 
instruments according to their functional complexity, usage, and risk to data 
integrity. For each category or class of instruments, a set of predetermined scalable 
deliverables and testing is defi ned. These defi nitions will in turn be used to determine 
the extent and type of testing required to validate various instruments. This 
instrument classifi cation system and its associated deliverables provide simple and 
structured guidance on deciding the validation effort required for a whole spectrum 
of laboratory instruments from a simple pH meter to a fully automated dissolution 
workstation. 
An instrument classifi cation system which is based on the complexity of the 
instrument is given in Table 1 . There are similar laboratory instrument classifi cations 
outlined in GAMP: Good Practice Guide: Validation of Laboratory Computerized 
Systems [10] , and the U.S. Pharmacopeia (USP) general chapter on analytical instrument 
qualifi cation [11] . As there can be overlap between categories of instrument 
classifi cations, individual organizations should make modifi cations to the defi nitions 
and the types of instruments in the classifi cation system as required, thereby customizing 
the classifi cation system to their respective operation. After the laboratory 
instrument classifi cation has been established, the deliverables required for the 
validation for each class of instruments will be defi ned to provide the framework 
for the validation process. Table 2 lists the deliverables commonly encountered in 
the laboratories. 
TABLE 1 Instrument Classifi cation 
Category a Defi nition Examples b 
A Equipment/tools used for sample 
preparations 
Hotplates, stirrers, shakers 
B Basic fi rmware instruments used to 
measure fundamental physical 
parameters such as weight, 
dimension, temperature, and pH 
Balances, pH meters, digital 
thermometers, centrifuges, sonicators 
C Advanced fi rmware instruments 
that utilize spectrometry, 
chromatography, dissolution, etc. 
Liquid chromatographs, gas 
chromatographs, dissolution baths, 
Karl Fischer titrators 
D Commercial COTS controlled by 
external computer 
Hybrid systems such as automated 
dissolution workstation with 
high - performance liquid 
chromatography (HPLC) or 
ultraviolet – visible (UV – Vis) 
interface 
Liquid chromatographs, gas 
chromatographs, UV/Vis 
spectrophotometers, Fourier 
transform infrared (FTIR) 
spectrophotometers, near - infrared 
(NIR) spectrophotometers, mass 
spectrometers, atomic absorption 
spectrometers, thermal gravimetric 
analyzers, COTS automation 
workstations 
a Categories may vary. Consult the user ’ s organization laboratory instrument classifi cation guide or references 
in the GAMP guide and USP general chapter . 1058 . [11, 12] . 
b Examples are not all inclusive. 
LABORATORY INSTRUMENT CLASSIFICATIONS 793

794 VALIDATION OF LABORATORY INSTRUMENTS 
8.3.4 VALIDATION PHASES 
Conceptually, a life - cycle approach to laboratory instrument validation can be 
divided into four phases, as shown in Figure 1 . 
8.3.4.1 Planning and Requirements Phase 
The activities in the planning phase are commenced after a business need to implement 
an instrument in the laboratory has been identifi ed. Planning - phase activities 
include concisely documenting the user requirements and functional requirements 
of the instrument. Once established, the requirements are then used to evaluate the 
instrument candidates available from suppliers. A validation plan with deliverables 
that are commensurate with the class of the instrument being implemented should 
be developed during the planning phase. 
Validation Plan A validation plan is prepared to highlight the activities and 
deliverables required in the implementation of the laboratory instrument. The 
details of the validation plan should be scaled according to the complexity of the 
system, its intended use, and the impact on the business. A generic validation plan 
can be used for similar types of laboratory instruments with similar applications. 
For example, one validation plan can be prepared for the HPLC. A validation plan 
typically includes the following sections: 
• Objectives and Scope of Project Rationale for implementing the instrument 
with its key applications and to state the assumptions, exclusions, and limitations 
of the project. 
TABLE 2 Key Instrument Validation Deliverables 
Deliverable 
By Category 
A B C D 
Business requirements a . b . b,c . . 
Compliance assessment (GxP criticality assessment) . b . b,c . . 
Validation plan . b . b . c . 
Electronic records/electronic signatures (ER/ES) 
assessment 
. . . . 
User and functional requirements (design 
qualifi cation) 
. b . a . c . 
Installation qualifi cation . b . c . c . 
Operational qualifi cation . b . c . c . 
Performance qualifi cation . b . c . c . 
Validation report . . c . c . 
Traceability matrix and confi guration documents . . . . 
Standard operating procedures . b . . . 
Performance verifi cation . . . . 
Periodic review . . b . . 
Decommissioning . b . . . 
Note : . : required; . : not required. a Supplier audit may be required for some category D instruments. 
b Optional. 
c Simplifi ed standard forms can be used. 

VALIDATION PHASES 795 
• System Description Hardware, software, and system confi guration (if the 
system is comprised of different modules or components). 
• Compliance Assessment State whether the use of the analytical instrument is 
subject to GxP and regulatory requirements. Instruments subjected to GxP 
will need validation. A rationale should be provided to support the decision of 
the assessment. 
• Validation Approach and Deliverables Outline the required documents, quali- 
fi cation testing, and reports to be included in the validation project. The deliverables 
should be based upon the complexity and the risk associated with the 
intended use of the system. The deliverables will highlight the acceptance criteria 
to demonstrate that the system has met the requirements for its intended 
use. 
• Roles and Responsibilities List the group(s) or personnel that are responsible 
for the validation of the instrument. 
• Project Timeline and Milestones List the projected completion date of major 
tasks in the validation. This information is useful in monitoring the progress of 
the project and making sure proper resources are available at various stages of 
the validation. 
• Record Management Describe the format, numbering sequence, revision, 
version control, and storage of the validation documentation. 
FIGURE 1 Validation phases. URS: user requirement specifi cation; FRS: functional requirement 
specifi cation; DQ, IQ, OQ, PQ: design, installation, operational, and performance quali- 
fi cations; SOP: standard operating procedure. 
Phases 
Planning 
Qualigication 
Operational Operation SOP 
Site 
preparation 
Identify needs 
Calibration and 
performance 
verification 
Qualifcation 
protocol 
Validaton plan 
Maintenance 
and usage logs 
Qualifcation 
IQ, OQ, PQ 
Requirements 
URS and FRS 
Periodic 
review 
Review and 
approval 
DQ and system 
evaluation 
Change 
control 
Summary 
report 
Risk and 
compliance 
assessments 
Activities 
End of life Decommission

796 VALIDATION OF LABORATORY INSTRUMENTS 
• References of Relevant Information List the internal and external instrument 
validation guidance documents, compendial reference methods, qualifi cation 
testing, and related SOPs for validation. 
User Requirements The user requirements outline the applications that the user 
wants the analytical instrument system to run and the major tasks that the instrument 
is required to execute in order to deliver the applications. The requirements 
should be clear, unambiguous, and verifi able or testable. The user requirements 
dictate the amount of validation that will be required. A successful validation should 
ultimately demonstrate that the user requirements can be met by the instrument 
and that the instrument can be used reliably for its intended applications. It is very 
important to state clearly what the instrument is supposed to do. Only the mandatory 
requirements should be validated. Including the “ nice to have ” or optional 
features in the user requirements will unnecessarily add more work to the 
validation. 
For example, a simple HPLC system with an isocratic mobile - phase delivery 
system and a multiple - wavelength UV detector may be suffi cient for simple routine 
product release testing. However, an HPLC system for running impurity assays or 
stability - indicating methods is likely to require a gradient mobile - phase delivery 
system and a diode array UV detector. If only the isocratic applications will be 
required for an HPLC system, there is no need to validate the gradient feature of 
the HPLC system even though the system is capable of running gradient applications. 
Another example is that of a dissolution bath that is solely utilized for running 
paddle dissolution methods. Although the bath may have the capability to run both 
basket and paddle methods, it is only the validation of the paddle method that 
should be considered necessary. 
Functional Requirements The high - level user requirements establish the framework 
for the functional specifi cations which identify the system functions, mode of 
operation, and operation environment necessary to meet the user requirements. 
Again, the requirements should be verifi able or testable to demonstrate that the 
requirements can be met during the qualifi cation testing. 
Functional requirements usually include the following considerations: 
• Functions of each system hardware component or modules: The expected 
operational range of each function and its performance attributes such as the 
accuracy, precision, and linearity required for its intended use should be 
defi ned. 
• Site requirements to support the operations of the instrument: Power supply 
(electrical voltage and current), ventilation, gas supply, water supply, and 
drainage. 
• Health and safety requirements: Mechanical safety for robotic movement, 
radiation safety for systems that involve the use of radioactive sources, and 
laser safety for systems that use high - power lasers. 
• Computer operation system and network requirements. 
• Data type (analog and/or digital output/input) and capacity (size of the data 
fi les). 

VALIDATION PHASES 797 
• Requirements for electronic records and electronic signatures [13 – 15] : System 
security, data integrity, traceability, and data archive/retrieval. 
• Communication and control of other systems. 
• System recovery from a major failure or disaster. 
Typical hardware functional requirements for a gradient HPLC system with 
UV – Vis detection are listed in Table 3 as an example. 
Design Qualifi cation For COTS analytical instruments, the user has little or no 
input into the design. The design information is usually not available to the users. 
In this case, the design qualifi cation demonstrates the user requirements and the 
functional requirements can be fulfi lled by the analytical instrument being considered 
by comparing the requirements against the technical specifi cations from the 
suppliers. For analytical systems which comprise of multiple COTS components 
(e.g., a dissolution system with online HPLC analysis capability), the confi guration 
and compatible components from different suppliers should be demonstrated in the 
DQ. 
TABLE 3 Typical Functional Requirements for HPLC with UV – Vis Detector 
Modules Functional Requirements 
Pump • The pump should be capable of a fl ow rate between 0.50 
and 5.00 mL/min. 
• The pump should be a quaternary gradient pump and have a 
compositional accuracy of ± 1.5% of the theoretical values for 
the four channels. 
• The pump should have relative standard deviation (RSD) 
of . 2.0% for six successive readings from a calibrated 
fl owmeter. 
Autosampler • The autosampler should have an injector capable of injecting 
sample volumes from 1 to 100 . L. 
• The injector should have a precision of . 1.5% RSD. 
• The injector carryover should be . 0.5% of peak area. 
• The autosampler should pick the correct vial. 
• The autosampler should use relays and/or contacts to 
communicate with the laboratory CDS. 
• Temperature - controlled sample racks should be capable of 
maintaining samples in the temperature range of 4 – 15 ° C 
(± 3 ° C). 
UV – Vis detector • The detectors should operate with a wavelength range from 
200 to 800 nm. 
• The detector should have a wavelength accuracy of ± 2 nm. 
• The detector response should be linear with a correlation 
coeffi cient r2 not less than 0.999 over the full dynamic range. 
• The linear range should be up to 2.0 Absorbance Unit Full 
Scale (AUFS) . 
Column compartment • The column oven should be capable of maintaining a 
temperature range of 5 ° C above ambient to 60 ° C. 
• The oven should be within ± 3 ° C of the set temperature and 
the temperature precision should be . 2.0%. 

798 VALIDATION OF LABORATORY INSTRUMENTS 
System Evaluation and Supplier Assessment Usually there are many choices and 
suppliers for common COTS laboratory instruments. The user requirements and the 
operational requirements will provide the basic criteria for the selection. Obviously 
the chosen instrument must be able to fulfi ll the key requirements for its intended 
use. Other factors concerning the instrument such as its ease of use, maintenance, 
and reputation of the suppliers in terms of quality, reliability, and support should be 
considered. From a practical point of view, a supplier audit may not be viable or 
necessary for commonly used COTS instruments. A supplier assessment is sometimes 
used to evaluate whether the supplier has a good - quality system in place to 
support the development and manufacturing of the instrument of interest. The need 
for a supplier assessment depends on the criticality and complexity of the system 
to be obtained. 
8.3.4.2 Qualifi cation (Testing) Phase 
The terms instrument qualifi cation and instrument validation are sometimes used 
indiscriminately. In this chapter, the term qualifi cation refers to the site preparation 
and the testing employed to demonstrate that the instrument is properly installed 
in a suitable environment and the performance meets the predetermined specifi cations 
for its intended use. Qualifi cation is a part of the whole validation life cycle. 
Validation refers to the process to provide assurance that the instrument is suitable 
for the intended application throughout the lifetime of the instrument. Installation 
qualifi cation (IQ), operation qualifi cation (OQ), and performance qualifi cation 
(PQ) are performed to provide evidence that the user requirement specifi cations 
(URSs), functional requirement specifi cations (FRSs), and design qualifi cation 
(DQs) have been met. The sequence of requirements setting and qualifi cation 
events as well as the relationships between IQ, OQ, PQ and URS, FRS, and DQ are 
generally illustrated by the “ V ” diagram shown in Figure 2 . Installation qualifi cation 
demonstrates the fulfi llment of the DQ. Similarly, OQ demonstrates the fulfi llment 
of the functional requirements and PQ demonstrates the fulfi llment of the user 
requirements. 
Site Preparation Prior to the installation of the instrument in the laboratory, all 
the preparations that support the operation of the instrument must be ready. The 
FIGURE 2 Qualifi cation “ V ” diagram. 
User requirements Performance qualification 
Users and supplier 
Functional specifications 
Design specifications 
System 
Installation qualification 
Operation qualification 
Supplier 
User define

VALIDATION PHASES 799 
supplier usually provides a site preparation document which outlines the information 
of the necessary facilities required to support the operations of the instrument. 
The site preparation document usually includes the following information: 
• Physical dimensions and weight of the instrument 
• Environmental conditions for proper operations: temperature, humidity, and 
vibration control 
• Utilities: power supply, water, gas, drainage, ventilation, network connection 
• Health and safety requirements 
It is a common mistake to underestimate the effort and time required for site 
preparation. The users should carefully study the site preparation guide to ensure 
that the necessary preparations to house the new instrument in the laboratory are 
completed prior to installation. Inadequate site preparation can cause major inconveniences 
and long delays in the installation process. It is a waste of time 
and resources to have the service engineer show up in the laboratory but not able 
to do anything due to incomplete site preparation. 
Qualifi cation Approaches If an analytical instrument is comprised of different 
functionally discrete modules, a modular approach to qualifi cation testing that 
focuses on the specifi c operations of the individual module can be suitable for 
certain aspects of some operational testing (such as the fl ow rate precision and 
accuracy testing of a HPLC pump and the temperature accuracy column 
compartment). 
When it comes to performance qualifi cation, a holistic approach must be taken 
to test the analytical system with all the necessary modules working together to 
deliver the intended applications as specifi ed in the user requirement (Figure 3 ). 
The proper functioning of each individual module of the analytical system does not 
FIGURE 3 Modular versus holistic approach for qualifi cation. 
Installation and operation 
qualification 
Verify the components 
modular testing 
System suitability test 
Performance 
qualification

800 VALIDATION OF LABORATORY INSTRUMENTS 
infer proper functioning of the whole system with all the modules working in 
concert. 
Qualifi cation Protocols For COTS instruments, the supplier often provides test 
protocols for installation and operation qualifi cation. It is a common practice to 
purchase the qualifi cation services from the supplier to execute the testing. The 
supplier should have a good knowledge of the instrument being installed and quali- 
fi ed, and therefore utilizing their services to perform these steps can provide signifi - 
cant time and resource savings. However, it is the responsibility of the user to review 
the test protocol for meaningful testing, that proper procedures will be used, and 
that reasonable acceptance criteria for the testing to ensure the objective of the 
testing are met. The supplier procedures and/or testing should be compatible with 
the procedures and practices of the user company. Acceptance criteria which are 
too loose do not provide enough assurance to demonstrate instrument performances. 
In contrast, acceptance criteria which are too tight can lead to unnecessary 
failures that cause a lot of effort and time to investigate and justify. The acceptance 
criteria should refl ect both the operation requirements and the user requirements. 
The protocols must be approved before execution. 
The test protocol typically has a general description of the system, the confi gurations, 
and the intended use. The test scripts in the test protocol provide detailed 
information of the testing procedures. In each test script, the following information 
should be provided: 
• Description of the module to be tested 
• Purpose and objective of the tests 
• Scope and limitations 
• Test procedure 
• Acceptance criteria for the testing 
• Sections to capture the test results 
• A section to document deviation and exceptions encountered during the 
testing 
Installation Qualifi cation Installation qualifi cation provides documented evidence 
that the instrument was received and successfully installed in accordance with 
the approved design requirements and properly installed in an environment suitable 
for its operation. Proper installation is the fi rst step to ensure that the instrument 
will function properly. An improperly installed instrument is likely to cause problems 
during the operational qualifi cation and performance qualifi cation. The following 
are some checks for the IQ process: 
IQ checks before instrument is installed: 
• Verify hardware and software against shipping list. 
• Check for visible damage. 
• Complete site preparation check list. 
• Health and safety precautions. 
• Add the instrument to the instrument inventory list. 

VALIDATION PHASES 801 
IQ checks during installation: 
• Document the location of the instrument. 
• Document all the hardware components, including computer, printer, analytical 
instrument fi rmware, interfaces, and network connection.. 
• Document of all software applications, the operating system, and the storage 
location of actual software package. 
• Document the system confi guration. 
IQ checks after installation: 
• System powers up properly. 
• Proper initialization and homing position. 
• Proper communication between modules. 
• Calibration of modules if necessary. 
• Software version verifi cation. 
• Proper launching of the software and compatibility with hardware. 
• Back up critical fi les for the system settings. 
• Set up log book. 
Availability of reference documents: 
• Site preparation guide. 
• System operation manuals. 
• Sale and shipping documents. 
• Factory testing records if available. 
• Certifi cate of quality compliance from the vendor. 
During system installation and qualifi cation testing, use screen capture to print 
the information for evidence whenever possible as it is a more effi cient and complete 
way to document the results and observations than writing down the information 
on paper. 
Operational Qualifi cation Operational qualifi cation provides documented evidence 
that the instrument will perform in accordance with its functional requirements 
throughout the representative operation range in a suitable environment. The 
OQ for simple instruments such as pH meters, balances, stirrers, water baths, and 
thermometers can simply achieved by execution of a calibration for the instrument. 
For moderately complex systems, OQ verifi es the correct operation of the hardware 
and software for the instrument. 
The testing of the hardware should cover the functionality of the instrument 
expected during normal operation. For example, the testing for an HPLC system 
would include the operation of the pump, the injector, and the detector [15] . 
Typical OQ tests for HPLC modules include: 
• Pump: fl ow rate accuracy and gradient accuracy 
• Detector: linearity of response, noise, drift, and wavelength accuracy 
• Injector: precision, linearity, and carryover 
• Column heater: temperature accuracy 

802 VALIDATION OF LABORATORY INSTRUMENTS 
Typical OQ tests for the UV – Vis spectrophotometers [16] include: 
• Wavelength accuracy and reproducibility 
• Stray light 
• Resolution 
• Photometric accuracy, reproducibility, and linearity 
• Noise 
• Baseline fl atness and stability 
In addition to testing the system components, a test of software functionality 
would be performed to test the system software operation and electronic records 
and electronic signatures (ERES) compliance (security, data integrity, data backup, 
and archive). In order to test the software functionality, a predetermined set of 
instructions can be entered step by step into the system. The system responses are 
then compared to the expected outcomes of the instruction and any problems with 
the execution are determined. Some vendors will provide a standard set of data 
which can be processed by the system to verify the data - handling capability of the 
system. 
Electronic Records and Electronic Signatures Considerations ERES compliance 
testing for computerized (personal - computer - controlled) instruments is required 
to demonstrate the functional requirements in the following three key areas 
[12 – 14] : 
1. Security To prevent unauthorized access. 
• Control unauthorized access to the instrument through physical, logical, and 
procedural controls. Physical control is normally achieved by controlling 
access to the site, building, laboratory, and instrument. Logical security is 
provided by setting up different levels of access accounts such as user, supervisor, 
and system administrator and distinct passwords for each account user. 
Dual logical access controls are usually available from the applications software 
or the operating system. A procedural control using SOPs can be used 
to assign users access. 
2. Data Integrity To demonstrate that records are valid and trustworthy. A 
record is a combination of raw data and the metadata (processing parameters 
plus other related information necessary to reconstruct the records): 
• Traceability: Helps to reconstruct the events. 
• Use of audit trails to track the activities for the creation, modifi cation, or 
deletion of a record. It provides a record of who did what, wrote what, when, 
and why. 
• Use of secure date and time stamp for the creation, modifi cation, and deletion 
of data. 
3. Data Preservation To preserve a complete and accurate set of records by 
copying electronic data to a removable or remote storage medium and to 
manage the data retention. The data integrity must not be compromised during 
the preservation process: 

VALIDATION PHASES 803 
• Data backup and restore: Active data are periodically copied from the hard 
drive or the personal computer controlling the instrument to a suitable 
medium such as CD - ROM or DVD or to a separate location such as a controlled 
fi le server. 
• Data archive and retrieval: Data that are no longer required in the day - to - 
day operation of the instrument are archived in a retrievable manner. 
Performance Qualifi cation Performance qualifi cation is the process that provides 
documented evidence to demonstrate that the instrument has fulfi lled the user 
requirements. Holistic testing which involves all the functional components in the 
system is required for the PQ testing. Performance qualifi cation can be demonstrated 
by running a typical application that requires all the modules to function 
together as a whole system to deliver the intended application and the expected 
results. 
If applicable, it is advantageous to execute a highly reliable test that is frequently 
performed by users on the particular type of instrument for the PQ. For example, 
the PQ test for a gradient HPLC system can be a gradient HPLC method with which 
the user has in - depth experience. In this case, the test results will mainly refl ect the 
performance of the instrument and will not be affected by the uncertainty of the 
method used. 
Test Exceptions and Data Review In case the test results fail to meet the acceptance 
criteria, an investigation is required to determine the cause of the failure. The 
failures may be caused by execution errors of the test procedure or by instrument - 
related problems. Based on the outcome of the investigation, corrective actions can 
be taken to rectify the problem. Once the problem has been resolved, retesting can 
be executed to confi rm that the instrument operations meet the requirements. The 
failure investigation, impact assessment, corrective actions, and retesting must be 
documented in the exception log of the qualifi cation testing. After the qualifi cation 
testing has been completed, the testing process, data generated, and results must be 
reviewed for accuracy, completeness, and justifi able deviations. All major issues with 
regard to compliance have to be addressed prior to releasing the instrument for 
production use. 
Summary Report After completion of the qualifi cation activities, a summary 
report should be prepared to summarize all the qualifi cation testing and results and 
deviations from the validation plan as well as include a conclusion to state whether 
the instrument is ready for its intended uses. Any deviations from planned activities 
including testing failures and requirements not met by the system should be 
addressed in the report. The summary report typically includes the following 
sections: 
• Introduction 
• System description 
• Reference documents 
• Validation activities, testing, acceptance criteria, and results 
• Summary of deviations, problems, and mitigations 

804 VALIDATION OF LABORATORY INSTRUMENTS 
• Restrictions on the system (if any) 
• Deliverables 
• Conclusion 
For simple analytical instruments, a simple table to summarize the qualifi cation 
testing, acceptance criteria, results, and pass/fail decision of the tests will be suffi cient 
since there are fewer tests that are required and the tests are usually relatively 
simple. For complex analytical systems, a more complex table often referred to as 
a traceability matrix which traces the requirements, testing, acceptance criteria, test 
results, and storage locations of the validation documents, test data, and other supporting 
documents is usually included in the summary report for easy reviewing and 
quick references. 
After the actual qualifi cation testing and regular performance verifi cation testing, 
the documents and the related test data are the only proof that the instrument has 
gone through such testing and has been properly installed and maintained to support 
its intended applications. The document should be stored systematically in a centralized 
location and maintained with care in order to prevent any losses. A good 
recordkeeping system can be extremely useful in audit preparation and help to 
speed up the turnaround time for documents during an inspection. 
8.3.4.3 Operational Phase 
After an instrument has been qualifi ed, it is ready for production use. The activities 
in the operational phase support the day - to - day use and maintain the instrument in 
a validated state. 
Standard Operating Procedure A SOP has to be written to provide instructions 
for the operation, maintenance, and calibration of the new instrument. A typical 
SOP should include: 
• A general system description 
• Operation instructions 
• Responsibilities of the system users and system administrators 
• Calibration or performance verifi cation requirements, acceptance criteria, frequency 
of testing, and the actions required if the instrument does not meet the 
performance verifi cation requirements. 
• Maintenance requirements 
• Service, major and minor repairs, and parts replacement that will necessitate a 
requalifi cation of the instrument. For example, the replacement of a UV lamp 
in a UV detector does not require a full requalifi cation, whereas a replacement 
of circuitry board will warrant full requalifi cation. 
Maintenance Normal wear and tear as well as aging of various components may 
compromise the performance of the instrument or lead to operation failure. The 
instrument needs to be maintained in order to function consistently and reliably. A 
preventive maintenance program which identifi es and replaces the consumable 
parts will likely save time and money in the long run. The usage and service records 

VALIDATION PHASES 805 
are kept with the instrument to provide a performance history of the system. These 
records can provide clues to simplify troubleshooting in case of instrument failure 
and in addition are used to help develop a meaningful preventive maintenance 
schedule. 
Performance Verifi cation and Calibration In order to maintain the instrument in 
a validated state, regular performance verifi cation and calibration are required to 
demonstrate the instrument is functioning reliably according to a predetermined set 
of criteria to support the applications required in the user requirements. The current 
GMP (cGMP) requirements also dictate the calibration of instruments at suitable 
intervals in accordance with an established written program. The terms calibration 
and performance verifi cation are very often used interchangeably. Calibration 
involves measuring and adjusting the instrument response using known standards. 
Performance verifi cation verifi es the operation and performance characteristics of 
an instrument against a predetermined set of requirements. Calibration can be 
considered as a part of performance verifi cation. 
The frequency of performance verifi cation testing should be based on the 
knowledge of the operational reliability for the type of instrument and the type 
of operations that the instrument will be supporting. Initial frequency can be 
based on the recommendation from the supplier. A good performance record can 
justify less frequent verifi cation. One potential drawback for overextending 
the period between performance verifi cation can be the increased amount of impact 
assessment on the data generated since the last performance verifi cation in case 
of system failure. Running system suitability testing before the analysis cannot 
replace the need for regular instrument calibration. System suitability testing 
is method specifi c whereas system calibration verifi es the general performance 
of the instrument. The system suitability test only demonstrates that the instrument 
is suitable for a particular method at the time of analysis. It cannot reveal marginal 
performance of the system. For example, the system suitability test for an 
HPLC assay using UV detection is not likely to detect a wavelength accuracy 
problem since both the standards and the samples are analyzed at the same 
wavelength. 
Change Control Change control provides a structured mechanism for requesting, 
authorizing, evaluating, testing, implementing, and releasing changes to validated 
systems. It should be performed in accordance with approved and documented 
procedures. These procedures should include the following elements: 
Prechange: 
• Description of the proposed change 
• Assessment of impact of a proposed change 
• Reviewing and approval prior to execution 
• Communicating changes to system users 
• Management approval to implement the change 
Postchange: 
• Implementing the change 
• Testing after the change is implemented 

806 VALIDATION OF LABORATORY INSTRUMENTS 
• Training and/or retraining system users 
• Management approval to complete the change process 
Typical change control forms to are shown in Figures 4 and 5 . 
The version of fi rmware may be inadvertently updated by the service engineer 
during service or routine maintenance without making the changes known to the 
system owners. The change may not have any impact on the operation of the instrument 
and may not be detected. However, the new fi rmware version is different from 
FIGURE 4 Prechange control form. 
Equipment make and type Instrument identifier 
Current software New software 
Operating system Validation plan reference 
Description of change 
Reason for change 
Impact of change 
Validation document references 
Documentation that requires updating 
Requalification necessary (attach protocols) 
Date 
Date 
Date 
Request initiator 
Mangement approval 
QA approval 
FIGURE 5 Postchange control form. 
Equipment (make and type) Instrument identifier 
Test results summary 
e t a D e c n e r e f e R e c n e r e f e r n o i t a c i f i l a u q e R 
e t a D e c n e r e f e R d e w e i v e r n o i t a t n e m u c o D 
e t a D e c n e r e f e R d e t a d p u s d r o c e r t n e m p i u q E 
e t a D e c n e r e f e R d e w e i v e r P O S 
e t a D e c n e r e f e R d e t a d p u r e t s i g e r m e t s y S
Program validation data 
archived 
e t a D e c n e r e f e R 
e t a D e s u r o f y d a e r m e t s y S 
e t a D r o t a i t i n i t s e u q e R 
e t a D l a v o r p p a t n e m e g a n a M 
e t a D l a v o r p p a A Q

VALIDATION PHASES 807 
the version documented in the validation documents. The system owners should 
work with the service engineers to prevent this potential violation. 
Business Continuity Planning (Disaster Recovery) A disaster recovery plan 
should be in place to ensure the continued operation of the laboratory in case of 
an adverse event that renders the instrument out of commission and hence 
causes interruption to the business processes which the system supports. Adverse 
events like the failure of the critical hardware components of the instrument and 
the failure of the application software do happen in the day - to - day operation of 
a laboratory. The disaster recovery plan should provide the necessary steps to 
restore the systems back to a functional state. The steps typically include instructions 
to reinstall the application software to the personal computer controlling the instrument, 
to reconfi gure the instrument, and to restore the backup data to the 
instrument. 
Periodic Review The performance of the instrument should be reviewed on a 
regular basis, typically once every two to three years, to ensure the instrument is 
still reliable and that it continues to comply with the user requirements. The review 
should include: 
• Operation Environment Changes such as temperature, humidity, and vibration 
that may impact the operation of the instrument. 
• Change Control System confi guration changes; hardware, fi rmware, and software 
change and its impacts. 
• Records Usage, maintenance, services, performance verifi cation testing, and 
location of the records. 
• Documentation Validation documents are current and the requirements are 
being met by the instrument, SOPs, operation manuals, business continuity 
plan, and location of the documents. 
• User Training Adequate training for a new version of software operation. 
The outcome of the review will determine whether the instrument is maintained in 
a validated state. In the case that the records indicated the instrument is more prone 
to certain types of failure, a preventive maintenance program may be desirable to 
avoid system failure during operation. 
8.3.4.4 End of Life 
The decommissioning of an instrument is the last step in the validation life cycle. 
When an instrument is no longer required in the laboratory and is ready for retirement, 
the following activities and related recordkeeping are required: 
• Document in the instrument binder or logbook the reason for decommissioning 
and the effective date. 
• Archive all related records such as the instrument binder or logbook, software, 
and manuals if no longer required in the laboratory. 
• Ensure that all electronic records are archived following established 
procedures. 

808 VALIDATION OF LABORATORY INSTRUMENTS 
• Disconnect all building services from the instrument. 
• Make arrangement for the removal of the instrument. 
• Withdraw or amend any affected SOPs. 
• Update the calibration program and the inventory list. 
8.3.5 SUMMARY 
The fundamental purpose of validating laboratory instruments is to provide assurance 
that the instrument is suitable for its intended use. The validation effort associated 
with a laboratory instrument should be commensurate with the complexity of 
the instrument, its intended use, and the impact of the data generated. A scalable 
validation approach to manage the whole life cycle of the instrumentation in the 
laboratory, from the planning stage to decommissioning, is a good business practice 
for smooth laboratory operation and will avoid preventable instrument failures. A 
systematic approach to instrument validation based on sound scientifi c rationales 
balancing the potential risks and business affordability can convey confi dence to 
the auditors during laboratory inspections. 
REFERENCES 
1. Freeman , M. , Leng , M. , Morrison , D. , and Munden , R. ( 1995 ), Position paper on the 
qualifi cation of analytical equipment , Pharm. Technol. Eur. , November 1995, 40 . 
2. Huber , L. ( 1995 ), Validation of Computerized Analytical Systems , Interpharm Press . 
3. Huber , L. ( 1996 ), Quality assurance and instrumentation , Accred. Qual. Assur. , 1 , 24 . 
4. Huber , L. ( 1999 ), Validation and Qualifi cation in Analytical Laboratories , Interpharm 
Press . 
5. International Organization for Standardization (ISO)/IEC 17025 ( 1999 ), General requirements 
for the compliance of testing and calibration laboratories, ISO, Geneva. 
6. Miller , J. M. , and Crowther , J. B. (2000), Analytical Chemistry in a GMP Environment — A 
Practical Guide , Wiley Interscience , Hoboken, NJ . 
7. International Conference on Harmonization (ICH) ( 2000 ), Quality guideline, Q7A, ICH, 
Geneva. 
8. International Society for Pharmaceutical Engineering (ISPE) ( 2002 ), GAMP (Good 
Automated Manufacturing Practice) Guide for Validation of Automated System in Pharmaceutical 
Manufacturing , Version 4, ISPE. 
9. Lam , H. ( 2004 ), Procurement qualifi cation and calibration of laboratory instrument: An 
overview , in Chan et al. , Eds., Analytical Method Validation and Instrument Performance 
Verifi cation , Wiley Interscience , Hoboken, NJ , Chapter 9. 
10. International Society for Pharmaceutical Engineering (ISPE) ( 2005 ), GAMP (Good 
Automated Manufacturing Practice) Good Practice Guide: Validation of Laboratory Computerized 
Systems , ISPE. 
11. U.S. Pharmacopoeia (USP) ( 2006 ), general chapter . 1058 . , Analytical instrument qualifi - 
cation, USP, Rockville, MD. 
12. U.S. Food and Drug Administration (FDA) ( 2005 ), Guidance for industry, 21 CFR Part 
11: Electronic records and electronic signatures, FDA, Rockville, MD. 

13. International Society for Pharmaceutical Engineering (ISPE) ( 2002 ), Good Practice and 
Compliance for Electronic Records and Signatures, Part 1, Good Electronic Records Management 
, ISPE and PDA. 
14. International Society for Pharmaceutical Engineering (ISPE) ( 2005 ), GAMP (Good 
Automated Manufacturing Practice) Good Practice Guide: A Risk - Based Approach to 
Compliant Electronic Records and Signatures , ISPE. 
15. Lam , H. ( 2004 ), Performance verifi cation of HPLC , in Chan et al. , Eds., Analytical Method 
Validation and Instrument Performance Verifi cation , Wiley Interscience , Hoboken, NJ , 
Chapter 11. 
16. Lam , H. ( 2004 ), Performance verifi cation of UV — vis spectrophotometers , in Chan et al. , 
Eds., Analytical Method Validation and Instrument Performance Verifi cation , Wiley Interscience 
, Hoboken, NJ , Chapter 10. 
REFERENCES 809


811 
8.4 
PHARMACEUTICAL 
MANUFACTURING VALIDATION 
PRINCIPLES 
E. B. Souto, 1,3 T. Vasconcelos, 2,3 D. C. Ferreira, 3 and 
B. Sarmento 3 
1 Free University of Berlin, Berlin, Germany 
2 Laboratory of Pharmaceutical Development, BIAL, S. Mamede do Coronado, Portugal 
3 Faculty of Pharmacy, University of Porto, Porto, Portugal 
Contents 
8.4.1 Introduction 
8.4.2 Scope of Validation Processes 
8.4.3 Validation Master Plan 
8.4.4 Validation Protocols and Reports 
8.4.4.1 Validation Protocols 
8.4.4.2 Validation Reports 
8.4.5 Facilities Validation 
8.4.5.1 Generalities 
8.4.5.2 Design of Facilities 
8.4.6 Manufacturing Process Validation 
8.4.7 Analytical Methods 
8.4.8 Equipment and Computer Systems 
8.4.8.1 Equipment Systems 
8.4.8.2 Computer Systems 
8.4.9 Cleaning Validation 
8.4.10 Conclusions 
References 
8.4.1 INTRODUCTION 
The pharmaceutical industry has been a pioneer in the development of quality and 
safety procedures assuring that the risk of its work is reduced to a minimum. The 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

812 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
establishment in the last century of good manufacturing practices (GMPs) was a 
small step in the pharmaceutical work but a giant step in the reduction of risk to 
patients and operators and fi nancial losses in pharmaceutical industry. Furthermore, 
to assure a GMP, one of the most important means is the validation procedure. 
Validation aims to confi rm that quality is assured step by step in the process and 
not just at the end of the pharmaceutical procedures. Generally, an entire process 
is validated and every step is verifi ed. Validation procedures are intended to ensure 
that all the different parts, steps, and components of a process are well defi ned and 
controlled in order to guarantee that the fi nal product will not change over time. In 
general, validation is the process of checking if something satisfi es a certain criterion 
and providing accurate and documented evidence that a process or system, when 
operated within established parameters, can effectively and reproducibly be performed 
while meeting its predetermined specifi cations and quality attributes. Therefore, 
validation is an integral part of quality assurance, involving the methodical 
study of systems, facilities, and processes aimed at determining whether they perform 
their intended functions adequately and consistently as specifi ed. When validated 
the process will provide a high degree of uniform batch assurance, meeting the 
required specifi cations and therefore being formally approved. Validation does not 
improve processes but confi rms that these have been properly developed and are 
under control. It is not only a requirement by the regulatory authorities but also an 
improvement to the pharmaceutical industry. The pharmaceutical industry benefi ts 
with the validation procedures since decreases the risk of problems and thus assure 
the smooth running of the processes. Also, validation procedures contribute to 
decreasing the risk of defect costs, the risk of regulatory noncompliance, and in 
process controls and end - product testing. 
Validation can also be defi ned as documenting that any procedure, process, 
and activity consistently lead to the expected results, including the qualifi cation 
of systems and equipment. The processes to be validated must be defi ned in a 
validation master plan (VMP) that consists of an approved written plan of objectives 
and actions stating how and when a company will achieve compliance with the 
GMP requirements regarding validation. The validation procedure must be well 
defi ned in a validation protocol, which is a written plan of actions defi ning how the 
process validation will be conducted, specifying as well who will conduct the various 
tasks and describing the testing parameters. The validation protocol must also 
contain sampling plans, testing methods, specifi cations of product characteristics, 
and equipment to be used. Moreover, it must specify the acceptance criteria and 
who will sign/approve/disapprove the conclusions derived from such a scientifi c 
study. The validation report is the fi nal product of a validation procedure and must 
contain the results and interpretation of the tasks defi ned in the validation 
protocol. 
Validation procedures must be applicable to computer systems; cleaning processes; 
manufacturing processes heating, ventilation, and air - conditioning systems; 
water systems; and analytical methods and equipment. 
The process validation provides a high degree of assurance that a specifi c process 
will consistently result in a product that congregates its predetermined specifi cations 
and quality characteristics. Process validation provides documented evidence that 
a process is capable of reliably and repeatedly rendering a product of the required 
quality. Water and air systems validations establish that the system is under control 

over a long period of time. Analytical validation proves that the analytical procedure 
is suitable for its intended purpose. Equipment validation is applicable to critical 
equipment whose performance may have an impact on the quality of the product 
guaranteeing that it works in a suitable manner. Computer validation is intended 
to provide a high degree of assurance that a computerized system analyzes, controls, 
and records data correctly and that data processing complies with predetermined 
specifi cations. Cleaning validation establishes that cleaning procedures are removing 
residues to predetermined levels of acceptability, taking into consideration 
factors such as batch size, dosing, toxicology, and equipment size. 
To summarize, validation procedures are a tool to develop a product with good 
and reproducible characteristics, assuring quality throughout the lifetime and 
improving security for patients and the pharmaceutical industry. 
8.4.2 SCOPE OF VALIDATION PROCESSES 
The scope of validation has been widely stated and attempts to clearly defi ne the 
competences of validation has been described [1] . It is now perfectly understood in 
both regulatory and compliance points that all laboratory processes, including facilities, 
equipment, analytical methods, and computer programs used in the analytical 
testing of pharmaceutical products, must be validated [2] . In the implementation of 
validation, it is fundamental to outline from the beginning the objectives throughout 
the validation process, which means that there should be proper preparation and 
planning before validation is performed as well as a specifi c program for validation 
activities. 
There should be an appropriate and suffi cient structured system, including organizational 
structure and documentation infrastructure, suffi cient suffi cient personnel 
and fi nancial resources to perform validation tasks in a timely manner. 
Management and the personnel responsible for quality assurance should be involved 
in this discussion. Validation performance must be under the responsibility of appropriate 
and experience personnel that should represent different departments 
depending on the validation work to be performed. 
Validation should be performed for new premises or equipment, utilities, or 
even systems, processes, and procedures at periodic intervals and when major 
changes have been made in accordance to written protocols. Validation can be 
prospective, concurrent, or retrospective, depending on when the validation is 
performed. 
There should be a clear distinction between in - process controls and validation. 
In - process tests are performed during the manufacture of each batch using specifi cations 
and methods devised during the development phase. The aim is to monitor 
the process continuously and not exactly validate it. When a new manufacturing 
formula or method is adopted, steps should be taken to demonstrate its suitability 
for routine processing prior to the validation. The defi ned process using the materials 
and equipment specifi ed should be shown to yield a product consistently of the 
required quality. 
Essential validation work must therefore be identifi ed to prove control of the 
critical aspects of the operations. A risk assessment approach should be used to 
determine the scope and extent of validation. 
SCOPE OF VALIDATION PROCESSES 813

814 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
8.4.3 VALIDATION MASTER PLAN 
Validation is intended to establish documented evidence with a high degree of probability 
that a process previously established will constantly perform according to 
the intended specifi ed products. 
According to the World Health Organization (WHO), the VMP is a high - level 
document that establishes an umbrella validation plan for the entire project and 
summarizes the manufacturer ’ s overall philosophy and approach to be used for 
establishing performance adequacy. It provides information on the manufacturer ’ s 
validation work program, including the manufacturer ’ s intentions and methods used 
to establish the adequacy of performance of equipment, systems, and controls and 
the processes to be validated. The VMP is an approved document that provides the 
details and time scales for the validation work to be performed, including the 
responsibilities relating to the plan. VMPs may be defi ned as structured, detailed 
plans of work providing information about the overall manufacturer ’ s philosophy, 
intentions, and approaches to be used to establish performance adequacy and how 
all of the validation work on a project will be controlled. 
Although only a recommendation, the VMP must be, undeniably, the most signifi 
cant document in any validation program. Throughout this document must 
clearly postulate the validation policy of each manufacturer, referring the organizational 
structure of all validation activities. The VMP supplies the main guidelines 
for the validation program, identifi es the responsibilities of the personnel involved 
in the validation activities, identifi es all aspects subject to validation, and expresses 
the nature and extent of testing on each point [3] . Other issues that must be mentioned 
in the VMP are a summary of existent facilities, systems, and equipment; a 
list of processes already validated and those intended to be; the planning and scheduling; 
the changing control; and references of existing documents prior to elaboration 
of the VMP. In this way, it is intended to be a retrospective, current, and 
prospective validation plan. All validation activities relating to critical technical 
operations relevant to product and process controls within a fi rm should be included 
in a VMP as well. This includes qualifi cation of critical manufacturing and control 
equipment. 
The VMP should be a summary document and should therefore be brief, concise, 
and clear. It should not repeat information documented elsewhere but should refer 
to existing documents such as policy documents, standard operating procedures 
(SOPs), and validation protocols/reports. The documentation format is illustrated 
in the VMP. 
A VMP should be divided into chapters covering different subjects. First, an 
introduction should state the manufacture ’ s validation policy, general description of 
the scope of those validation activities covered by the VMP, and their objectives, 
derivation, location, and schedule. Then, it must declare all validation activities and 
their organizational structure in terms of personnel responsibility for the VMP, validation 
protocols, validation work, report and document preparation projects, 
approval of the same validation protocols, reports in all stages of validation processes, 
and the training needs in support of validation. Other requirements of the 
VMP are cross references to other documents and to specifi c characteristics of the 
processes that are critical for yielding a quality product. Next, all validation activities 
comprised in the VMP should be summarized and compiled in a matrix format. Such 
a matrix should provide an overview and contain all items covered by the VMP that 

VALIDATION PROTOCOLS AND REPORTS 815 
are subject to validation describing the extent of validation required. It should 
include validation of analytical techniques which are to be used in determining the 
validation status of other processes or systems, validation approaches, revalidation 
activities, actual status, and future planning. Finally, the VMP must possess a general 
statement on key acceptance criteria for all validation activities. A planning section 
with detailed planning of subprojects must be included with an estimate of staffi ng, 
equipment, validation schedule of activities, and other specifi c requirements to 
complete the validation effort described. This time plan could be included in the 
above - mentioned matrix. A VMP requires regular updating. It must end with a 
statement of the enterprise ’ s commitment to controlling critical changes to materials, 
facilities, equipment, or processes (including analytical techniques) as well as 
references and a glossary. 
The VMP should contain references to the SOP’s and information relevant to all 
aspects of the validation process. These actions are carried out in the GMP manufacturing 
process of a product and executed in accordance with written instructions 
if the manufacture of the product is intended to be GMP compliant. The list of relevant 
SOPs must be included in the VMP and should defi ne how they must be validated. 
By planning and scheduling the validation procedure, the VMP defi nes the 
periodicity required to assure the validation statements. Validation of the different 
activities, facilities, and systems will occur in predetermined areas as defi ned by the 
VMP. The VMP should state who is responsible for preparing the VMP, preparing 
the protocols and SOPs, the validation work, the report and documentation preparation 
and control, approval of validation protocols and reports, tracking systems, and 
defi ning training needs. 
The VMP helps not only the manufacturer and team members but also the 
inspection. It helps all members of the validation team to know their tasks, responsibilities 
of various groups during the validation of the equipment, and utility and 
validation program with respect to time, people, and money. It also helps auditors 
to understand the enterprise approach to validation and the setup and organization 
of all validation activities. Furthermore, the VMP is a living document updated and 
adjusted during the course of the project. Therefore, there are special changes that 
require revalidation, namely software changes, site and operational changes, and 
changes in the source of materials, the process, the equipment, production areas, and 
support systems. 
8.4.4 VALIDATION PROTOCOLS AND REPORTS 
8.4.4.1 Validation Protocols 
A protocol is a written set of instructions broader in scope than a SOP. SOPs are 
the detailed written instructions for procedures routinely performed in the course 
of any of the activities associated with pharmaceutical manufacturing. A protocol 
describes the details of a comprehensive planned study to investigate the consistent 
operation of new system/equipment, a new procedure, or the acceptability of a new 
process before it is implemented. 
Protocols include signifi cant background information, explain the rationale 
behind and the aim of the study, and give a full description of the procedures to be 
followed. This means as well description of the site of the study; the responsible 

816 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
personnel; the equipment to be used; the standards and criteria for the relevant 
products and processes; the type of validation, sampling, testing, and monitoring 
requirements; a description of how the results will be analyzed; and predetermined 
acceptance criteria for conclusive purposes. Validation, stability, and clinical studies 
are examples of written protocols for pharmaceutical manufacturers. 
A validation protocol is a document that describes the item to be qualifi ed, the 
tests and checks to be performed, as well as the results that are expected to be 
obtained. It is a fi le in which the records, results, and evaluation of a completed 
validation program are assembled. It may also contain proposals for the improvement 
of processes and/or equipment [1] . Validation protocols are important in 
ensuring that documented evidence is taken which demonstrates that an equipment 
item, a system, a process, or a method consistently performs at a specifi ed level 
[4] . 
Validation protocols are required to describe the objective, methodology, and 
acceptance criteria for installation, operational, and performance qualifi cations. 
They are written to ensure test methods, and acceptance criteria are reviewed and 
approved before qualifi cation of protocols. In practical terms, there are several 
stages for the production of protocols. First, an acceptable format needs to be 
agreed. No universal format exists for protocols, but to some extent, the type of 
equipment, the size of the project, and the personal preferences will dictate the 
protocol style. However, some norms have been established. Like other controlled 
documents, protocols are assigned unique reference numbers and revision numbers. 
They are titled and numbered on every page and have a particular place for approval 
signatures. Other common elements in protocols tend to be brief descriptions of the 
item being qualifi ed and a clear statement of responsibilities. 
Often the protocol will incorporate test sheets or sections for recording data. In 
this way, once the protocol has been executed, the document constitutes a record 
of the results and conclusion [3] . It must describe the activities to be performed in 
a validation, including the acceptance criteria for the approval of a manufacturing 
process or a part thereof for routine use [1] . 
A validation protocol is necessary to defi ne the specifi c items and activities that 
will constitute any validation study. It is advisable for companies to draw up a VMP 
indicating the overall validation strategy for either the product range or equipment 
type or the entire site. The protocol must be prepared before initiation of the study 
and must either include or refer to the documentation required to provide information 
about a specifi c process, the parameters involved in that specifi c process, the 
personnel responsibilities, and the acceptance criteria. 
The list of items to be qualifi ed in the validation protocol must be produced and 
the approach agreed upon. There are several procedures to perform the validation 
protocols. Nevertheless, it is important that the approach agreed upon is executed 
to ensure internal consistency and prevent items being inadvertently forgotten. 
Thus, for instance, it is possible to group identical items together under one protocol 
or to have a protocol per item. A single or multiple protocols can be written covering 
installation qualifi cation (IQ), operational qualifi cation (OQ), and design quali- 
fi cation (DQ). For complex items such as large utilities, is it acceptable to address 
the generating system and the distribution system in separate protocols. Computer 
systems validation or control systems for process equipment can be documented 
within the mechanical validation protocol or in a separate protocol. 

VALIDATION PROTOCOLS AND REPORTS 817 
The protocol should be approved prior to its use. Furthermore, any changes 
to a protocol should be approved prior to implementation of the change. Once 
the approach and format are agreed upon, the protocol preparation can start. 
Preparing protocols requires individuals who are trained to extract information 
from a variety of sources and synthesize it into a coherent whole. Typically, sources 
of information include material requisition, specifi cations, data sheets, piping 
and instrument diagrams, manufacturer ’ s literature, and equipment operating 
manuals. Signifi cant protocols cannot be written until suffi cient sources of information 
are available. Once the fi rst draft is written, protocols require review. 
Usually, individuals from manufacturing, engineering, and quality assurance will 
perform the review. It is fairly normal for fi rst - draft protocols to require considerable 
modifi cation as this will be the fi rst time that other interested parties have seen 
them. Rather than check everything, it is preferable that the review disciplines 
concentrate on their specialist areas. For example, manufacturing should evaluate 
if the proposed acceptance criteria are consistent with the process requirements; 
engineering should ensure that the list of instruments, major components, and 
utilities requirements are accurate; and quality assurance should ensure that compendial 
requirements have been met and that the protocol meets normal quality 
expectations. 
In a large project, the coordination of protocol review, keeping track of the revision 
status, and the storage and retrieval of protocols are tasks that require careful 
planning. At some points it is probable that a number of protocols at different draft 
stages are circulating for review, some are being modifi ed, new protocols are being 
prepared, and approved protocols are stored prior to execution. A tracking system 
should be implemented that can very quickly identify where a protocol is, providing 
a picture of the current protocol production status. The method and time allowed 
for review must also be agreed upon. Review methods range from traditional circulation 
of a single document to each individual for comment to assembling a team 
for a joint review to electrically accessing the document and revising or commenting 
on it. The method chosen must refl ect the available time, resource, and technology. 
In some cases, the resources for protocol review are a limiting factor that can critically 
affect the schedule [3] . 
8.4.4.2 Validation Reports 
A validation report is a written document that cross - references the validation protocol, 
summarizes the results obtained, describes any deviations observed, and draws 
the necessary conclusions, including recommending changes required to correct 
defi ciencies for the qualifi cation and validation performed [5] . In this report it is 
required to present both the results and conclusions and the secure approval of the 
study. The report should include a summary of the procedures used to clean, sample, 
and test as well as the physical and analytical test results or references for the same. 
The conclusions regarding the acceptability of the results should also be included. 
Other information would be the status of the procedures being validated, any recommendations 
based on the results, or any relevant information obtained during 
the study. These include, revalidation practices (if applicable), the approved conclusions, 
and any deviations of the protocol that might have occurred. In cases where 
it is unlikely that further batches of the product will be manufactured for a period 

818 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
of time, it is advisable to generate temporary reports on a batch - by - batch basis until 
such time as the cleaning validation study has been completed [1] . 
Finally, the departments responsible for the qualifi cation and validation work 
should approve the completed report and the conclusion of the report should state 
if the outcome of the qualifi cation and/or validation was considered successful. The 
fi nal review is performed by the quality assurance department, which gives the 
approval of the report according to the company ’ s quality assurance system [1] . 
8.4.5 FACILITIES VALIDATION 
8.4.5.1 Generalities 
Facilities validation is related to the location, design, and construction of the plant 
to facilitate cleaning, maintenance, and operations in order to be appropriate for 
the type and stage of manufacture [6] . 
All facilities in a pharmaceutical industry must be designed and validated in 
order to assure the minimum risk of cross - contamination. Facilities validation must 
include the products and personnel fl ow, rooms design and cleaning, air and humidity 
systems, and water pipelines. 
8.4.5.2 Design of Facilities 
With regard to the design, the overall facilities should always be developed according 
to the most simplistic route of material fl ow and control of cross - contamination. 
Several layouts have therefore been described [7] , with the aim of separating 
released materials from quarantined or rejected ones. 
One of the most popular layouts consists of a center, or core, that has been conceived 
as a storage or warehouse area for raw materials, packaging components, and 
bulk stocks. Following this warehouse area immediately in the outer perimeter the 
manufacturing and packaging operations should be located to allow the fl ow of raw 
materials and components from the receiving and quarantine areas to approved 
storage. After materials are weighed into batch quantities, they are moved into the 
manufacturing area. When the manufacturing process is fi nished, the obtained products 
are placed in quarantine and then moved to bulk stock upon release. The 
packaging run follows when scheduled, and then the product and packaging components 
are delivered from the bulk stock and approved storage areas. An advantage 
of this layout is space conservation by virtue of having the supply areas close 
to the areas being supplied. Nonetheless, a signifi cant disadvantage is the crossover 
of materials with the potential risk of contamination or mix - up. 
An alternative layout design could be having the receiving, approved raw materials, 
components storage, and dispensing on one side and the manufacturing, quarantine, 
bulk stock, and packaging areas across a central corridor. Material fl ow from 
one area to another occurs in the same way as in the previous layout. However, in 
this case, fl ow is circular, eliminating much of the crossover described above. 
To minimize contamination or mix - up, a third layout could be basic straight - line 
fl ow moving the materials along a critical path. The main advantage over the above - 
mentioned layouts is minimal crossover of materials, thus minimizing the potential 

for contamination or mix - up. The additional space required to accommodate this 
confi guration can be pointed to as the main disadvantage. 
Cross-Contamination Control Two parameters particularly useful in controlling 
the cross - contamination are the air - handling systems and dust collection. 
Regarding the GMPs, air - handling systems are designed with the supply of outside 
air, moving on to the fi ltration systems that will be used, determining where positive 
and negative air pressures are required and whether to recirculate or exhaust spent 
100% air, and fi nally to dust collection and exhaust systems. 
Air fi ltration systems, including prefi lters and particulate matter air fi lters, should 
be used when appropriate on air supplies to production areas. If air is recirculated 
of dust from production in areas where air contamination occurs during production, 
appropriate exhaust systems or other systems adequate to control contaminants 
should be requested. 
A typical design involves one or more bag or cartridge fi lters located close to the 
area of dust generation. These coarse fi ltration devices should remove 95% of the 
dust generated from normal pharmaceutical manufacturing operations [8] . The pre- 
fi ltered air is then mixed with 10 – 15% outside make - up air and passed through a 
high - effi ciency particulate air (HEPA) fi lter and reenters the rooms through the 
supply plenum diffuser. 
Regarding dust collection, the fi rst area that must be addressed is the raw material 
sampling rooms and dispensing area. This area should be an enclosed facility 
with separate booths or hoods where the individual weighing or sampling can take 
place. These areas may be designed using horizontal laminar fl ow or appropriate 
hoods and other dust pickup devices. The supply air to these stations will therefore 
be HEPA fi lters either at the pickup stations or after the dust collector prior to 
returning to the general area or supply air. 
The last area of dust collection must be the packaging. Some machines are 
designed with a self - contained vacuum system that returns the air, fi ltered through 
an absolute fi lter, back to the packaging area. There should also be some provisions 
made at the cottoning stations. 
Humidity and Temperature Control Humidity and temperature control systems 
are also important considerations as these impair both product protection and 
working environment comfort. Unless otherwise stated, 45% room humidity and 
21 ° C are usually appropriate for critical manufacturing areas. Comfortable conditions 
should be provided for all operations. However, temperature controls should 
be such that little or no variation will be caused by external ambient temperatures. 
Thus, comfortable working conditions are achieved and there should be no impact 
on the characteristics of in - process materials. Warehousing operations should have 
adequate ventilation, particularly in areas of high storage, either pallet racks or 
pallet - to - pallet storage. The ventilation could be provided by large roof fans to circulate 
air. In addition, some form of supplemental air heating, such as hot - air 
blowers, should be provided for cold areas, such as shipping or receiving docks 
[7] . 
Water Systems Control The supply of potable water in a plumbing system must 
be free of defects that could contribute contamination to pharmaceutical products. 
FACILITIES VALIDATION 819

820 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
Therefore, an effective water system is required. Nowadays, several techniques can 
be used to obtain water of high pharmaceutical quality. These include ionexchange 
treatment, reverse osmosis, distillation, electrodialysis, and ultrafi ltration. However, 
there is no single optimum system for producing high - purity water, and selection of 
the fi nal system is dependent on factors such as the quality of raw water, intent of 
its use, fl ow rate, and costs. In the pharmaceutical industry, the different water classes 
normally encountered are well water, potable water, purifi ed water, and specially 
purifi ed grades of water, such as water for injection (e.g., MilliQ water). 
Water drawn directly from a well is called well water. The water may not be either 
chemically or microbiologically pure because it is untreated [7] . Therefore, the use 
of well water should be restricted to nonmanufacturing operations, such as lawn 
sprinklers, fi re protection systems, and utilities. 
Potable water is city water or private well water that has usually been subjected 
to some form of microbiological treatment, such as chlorine addition [7] . Potable 
water can be used in processing operations for cleaning and sanitation purposes. 
Periodic monitoring of use points should be conducted to ensure adequate residual 
chlorine levels and the absence of microbial contamination. Purifi ed water is treated 
to attain specifi ed levels of chemical purity and it is the type of water used in most 
pharmaceutical processing operations and fi nal equipment cleaning. Purifi ed water 
generally is produced by deionization or distillation, although reverse osmosis or 
ultrafi ltration systems might be utilized if the required chemical purity could be 
achieved. Water softening or activated carbon fi ltration is frequently employed as 
a pretreatment process to remove calcium and magnesium ions or chlorine and 
organic materials. Ion exchange and demineralization through deionization is a very 
common method to obtain the purifi ed water used in the pharmaceutical industry. 
Deionization equipment should have proper size to allow frequent regeneration. 
A recirculation system should also be installed in the unit that approaches the rated 
fl ow of the deionization unit. Procedures should be written to ensure that all water 
treatment equipment is properly operated, monitored, maintained, and sanitized on 
a regular basis. 
Regarding water fi ltration procedures fi rst, prefi lters should be provided to 
prevent large particulates from entering the system and microfi ltration to remove 
bacteria. Prefi lters are usually the replaceable cartridge type with porosities ranging 
as high as 25 . m. Microfi ltration follows and is generally accomplished with 0.2 - . m 
absolute fi lters, which will remove most bacteria. 
After defi nition of the water type required, the water pipeline must be validated 
in order to ensure an adequate fl ow and purity (chemical and microbiological) of 
water. After the validation process, periodic verifi cation of the pipeline and water 
collection points is required. This verifi cation must be based on well - defi ned 
SOPs. 
Sanitizing is best accomplished through several methods. After periods of low 
water usage, the system should be fl ushed with a supply of water that has residual 
chlorine. Periodic hyperchlorination and microbial control are also recommended. 
Microbial control can be achieved by storing the water at 80 ° C. Alternatively, ultraviolet 
radiation can be applied. 
Pest Control Each manufacture, processing, packing, or holding area of a pharmaceutical 
product should be maintained clean and sanitary. This means that these 

areas must be free of infestation by rodents, birds, insects, and others (other than 
laboratory animals). Trash and organic waste matter should be held and disposed 
of in a timely and sanitary manner. 
A pest control program should be developed in order to assure the integrity and 
quality of products produced and comply with existing guidelines. The program 
should be written to include a general statement of the purpose and the company 
position. Effectiveness of the program should be assured by defi ning the plant individual 
with overall responsibility for the program and how the responsibility will be 
carried out. In addition, the extermination staff, whether they be in - house or subcontracted 
personnel, should have their training and experience requirements well 
defi ned. Assistance in supporting the program may be gained from other plant personnel 
by their pointing out problem areas. 
This program should be treated to line management personnel. Furthermore, its 
content regarding the list of approved pesticides to be used in the plant should also 
be verifi ed periodically. Basic information should be controlled, such as the trade 
name of the pesticide; classifi cation; type of action; chemical name and concentration 
of active ingredient; effectiveness, usefulness, area of usage, mode, and frequency 
of application; toxicities and any specifi c toxic symptoms; status of government 
approval; and specifi c restrictions and cautions [7] . The development of sheets 
depicting such information will serve a twofold purpose. First, these sheets are 
subject to approval by the plant safety organization to determine if the materials 
comply with the Occupational Safety and Health Administration (OSHA) requirements 
and the requirements of other state or local agencies. Additionally, these 
sheets would also facilitate compliance with the GMP regulations. Written procedures 
for use of suitable rodenticides, insecticides, fungicides, fumigating agents, and 
cleaning and sanitizing agents should be provided in order to prevent the contamination 
of equipment, components, drug product containers, closures, packaging, 
labeling materials, or drug. 
All manufacturing areas should be constructed using nonporous materials on the 
walls and fl oors. Any protrusions, such as pipes and electrical boxes, should be minimized. 
Space should be allocated carefully to provide suffi cient rooms for all operations. 
There should be adequate lighting and the areas should be remote from any 
openings to the outside. Adequate training in understanding GMPs should be given 
to all personnel, including outside contractors. 
Scheduled inspections and preventive treatments should be documented specifying 
the problem encountered as well as the special service that has shown effectiveness 
to the treatment. 
Manufacturing Rooms Manufacturing rooms must be well designed in order to 
ensure adequate cleaning and reduction of cross - contamination. Points of dust 
accumulation like 90 ° angles areas must be avoided in the room, dust collectors or 
air lines must be presented restrictly at the wall surface and only the minimum 
equipments must be presented in a room. 
Air pressure inside the manufacturing rooms must be positive or negative 
depending on whether the product is a liquid or a powder, respectively, in 
order to avoid cross - contamination. It must be monitored by validated pressure 
systems. 
FACILITIES VALIDATION 821

822 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
Packaging and Labeling Control Type of construction in the packaging area is 
very important because it is particularly sensitive to mix - ups in either product or 
labeling. Individual packaging lines should have minimum separation between each 
line. Depending on the equipment used, more space might need to be used. Consideration 
should be given to the separation of lines by partitions. The partitions 
should be closed at fl oor level to prevent migration of product and be high enough 
to prevent any crossover of product. 
One possibility might be to start packaging operations with a prestaging area 
large enough to contain all required components for the packaging operation. This 
area is separate but adjacent to or in front of the fi lling area. The fi lling area should 
be equipped, if possible, with air fi ltered through absolute fi lters. Dust collector from 
the central dust collector system should be located in the fi lling area to remove dust 
from the fi lling area. The air circulation in the room must enter in the fi lling area 
and then be moved to the area where cottoning, capping, or labeling takes place. 
Thus, maximum product protection is exercised in the fi lling area while the product 
is exposed, with lesser degrees of control being required in the other areas because 
the product is in containers. 
SOPs should be developed for all packaging operations. These include specifi c 
procedures for line setup, approval of line before start of operations, periodic line 
check during operation, close out of line, line clearance at end of operation, and 
reconciliation of product and components. 
Labeling operations must also receive special attention. The use of cut labels in 
the pharmaceutical industry has already disappeared. Electronic label counting and 
verifi cation are the norm, with bar codes, universal product codes (UPCs) or health 
industry bar codes (HIBCs) being used more frequently as label identifi ers. The 
storage of labels is very important for both security and preservation reasons. The 
verifi cation of labeling at the fi nal label application point is also becoming more 
popular, since the the U.S. Food and Drug Administration (FDA) has identifi ed 
errors. 
SOPs must be developed for label accountability with specifi c tolerances spelled 
out. Provision must be made for label security, such as locked cabinets on operating 
lines. Accountability sheets for other components should be provided so that reconciliation 
between used and fi nished packages can be developed. If possible, a 
cleaning area and shop should be set up for the cleaning of fi lters and parts and the 
disassembly and assembly of fi lling machines. 
8.4.6 MANUFACTURING PROCESS VALIDATION 
The purpose of manufacturing process validation is to ensure that the manufacturing 
process for the drug product is fully controlled and capable of providing a 
product that complies with established quality specifi cations consistently and repetitively. 
This validation process covers manufacturing operations from formulation 
through packaging. Process validation is being conducted in support of launch of 
the drug product. 
Process validation also establishes documented evidence that the manufacturing 
process, when executed according to the VMP and pertinent process SOPs, consis

tently manufactures a drug product that meets all of the critical specifi cations 
required for its purpose. It is designed to ensure that all of the prerequisites are in 
place as well as to allow technical analysis of all applicable performance tests. 
Nowadays the manufacturing validation tendencies tend to focus on pharmaceutical 
development. Therefore, the primary responsibility of the pilot plant is to 
ensure that the developed product is effi ciently, economically, and consistently 
reproducible on a large scale. Since low production costs are of higher competitive 
advantage, attention must be paid to the production costs. Each operation unit 
should therefore be optimized. To avoid excessive production times, the manufacturing 
instructions transferred to the production department should be clearly written, 
readily understood, and unambiguous. Unless the economic advantage of purchasing 
new equipment has been proven to be necessary, in - house equipment should be 
preferentially used. Nonetheless, if international companies are intended to manufacture 
products at several sites, alternate manufacturing equipment and procedures 
might also be required. 
The physical properties and specifi cations for the manufactured formulations 
should match those established earlier by the product formulator, the pilot plant 
staff, and the quality control department. Therefore, the product manufactured on 
a large scale should possess the appropriate quality properties. 
Additional responsibilities for pilot plant staff are those related to the evaluation 
of new processing equipment, which aims to fi nd causes and solutions for problems 
that occasionally might arise during the production. Since the department of pharmaceutical 
research and its development division are responsible for developing the 
formulations and the manufacturing processes for the dosage forms, the experience 
gained in the development of the manufacturing process will be of major importance 
during the validation process. Such validation conducted at a pilot scale should 
simplify large - scale validation. 
It is clear that the aim of pharmaceutical research is to achieve zero defects and 
zero batch rejections, and this can be verifi ed by process validation. One must bear 
in mind that exhaustive fi nished testing of product is not a substitute for in - process 
controls and process validation. 
The validation process will consistently deliver product that has a direct relationship 
to a dosage form on which clinical effi ciency and safety were determined. Thus, 
comparisons between the manufacturing process, the raw material used, in - process 
control, and fi nished - process test results are in order. During scale up, consideration 
of these factors at the development stage must be carefully performed to produce 
quality batches. Matching equipment from laboratory to production helps to eliminate 
possible validation problems afterward. The physicochemical characteristics of 
the active drug substances should be controlled. Special attention should be paid 
to products in which the drug substance comprises a very small percentage of the 
pharmaceutical dosage forms. 
The physicochemical characteristics of raw materials play an important role in 
content uniformity and bioavailability. Therefore, bioavailability of the drug over 
time must be thoroughly investigated before any signifi cant changes are made. Once 
the physicochemical properties of drug substances (e.g., particle size of raw materials) 
can infl uence the availability and clinical effect of a product, the main characteristics 
of raw materials should be considered in a validation program. The 
MANUFACTURING PROCESS VALIDATION 823

824 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
characteristics of raw materials can vary among manufacturers, but also from lot to 
lot of the same manufacturer. Control of the physicochemical characteristics of 
excipients is also important and must be stated in the specifi cations. 
In order to develop a reproducible manufacturing process, attention must be 
given to particular instructions and screening procedures. For instance, excipients 
should be free of lumps and proper screening will aid raw material dispersion. 
Additionally, one should specify the size and design of containers and all equipment 
to be used. 
Changes in process decisions regarding the suitability of new manufacturing 
equipment that is intended to be employed should be performed by management 
and by production personnel. Equipment changes for some products might guarantee 
additional dissolution and/or content uniformity studies. The degree and depth 
of a study are largely dependent upon the specifi c product. For some products with 
very good historical data coming from a reproducible and controlled process, particular 
tests can be obviated. However, if potent dosage forms are to be developed, 
those are usually performed (e.g., sampling the mix in tablets technology) [9] . If 
products with lack of historical data are to be handled, additional fi nished product 
testing should be performed. For instance, with respect to tablet production dissolution 
testing and content uniformity testing are usually implied in the fi nished product 
tests. 
The batches utilized in this type of approach for manufacturing process validation 
should be presented on different scales as the manufacturing process is developed. 
Laboratory - scale batches are of very small size (e.g., 100 – 1000 times less than production 
scale) and are produced at the research and early development laboratory 
stages, which are used to support the formulation and packaging development and 
clinical and/or preclinical studies. In the development pharmaceutics, data validation 
of laboratory - scale batches can contribute to the choice of the appropriate manufacturing 
process (evolution and defi nition of critical product performance characteristics). 
Pilot batches correspond to at least 10% or 100,000 units for tablets (the 
biggest) of the production batches and are used in the development or optimization 
stage to support stability studies. Pilot batches are used to provide data predictive 
of the production batches and therefore provide the link between process development 
and industrial manufacture of the pharmaceutical product. Finally, production - 
scale batches will be produced during routine marketing of the drug product. 
Traditionally a proper process validation performance must be established based 
on the following considerations: 
• All raw materials used for the validation batches must be approved by quality 
control (QC). 
• Drug products are validated concurrently. Samples from three consecutive 
batches should meet all of the predetermined specifi cations without unexplained 
failures. If any of the three batches does not comply, the validation will 
be repeated up to three times to try to obtain three consecutive acceptable 
results. If not obtained, the tests will be suspended until the manufacturing 
process has been reviewed. 
• The validation batches must use the regular production equipment and 
personnel. 

• Critical in - process control parameters must be specifi ed by ranges, and full - 
range monitoring is performed. The limit determined to be the “ worst case ” 
will be challenge at least twice and the other limit challenged at least once to 
complete the full - range monitoring. 
• For critical process control parameters specifi ed by fi xed points, the 
value ought to be challenged within an acceptable tolerance, typically ± 1 
unit. 
• All tests must be conducted by training and experienced technical personnel 
and must be documented in a scientifi c manner using the established format of 
the protocol. 
• Any test function that does not have results which support the parameters 
defi ned in the approved protocol must be conclusively rationalized for their 
nonconformance and approved; otherwise, the qualifi cation will be considered 
invalid. 
Samples are to be taken during and/or after each critical manufacturing step. All 
control parameters for the manufacturing process have to be monitored and 
recorded. Each sample analysis will be performed in duplicate using validated or 
accepted pharmacopeia methods. The sample results will be used to confi rm in - 
process and fi nal product quality attributes as defi ned by the preestablished speci- 
fi cations. Conformance with specifi cations will justify the appropriateness of the 
critical parameters used during the process validation. 
Validation data should be generated for all products to demonstrate the adequacy 
of the manufacturing process. The process validation data may not always be available, 
however, where the manufacturing process uses a nonstandard method of 
manufacture. Data demonstrating the validity of that method should be submitted 
in the marketing authorization fi le. 
8.4.7 ANALYTICAL METHODS 
The aim of validation of an analytical procedure is to demonstrate that the method 
employed in any product testing, such as the identifi cation, control of impurities, 
assay, dissolution, particle size, water content, or residual solvents, is validated in the 
most important characteristics. Identifi cation tests, quantitative tests for impurities 
content, limit tests for control of impurities, and quantitative tests of the active 
moiety in samples of pharmaceutical product are the most common types of analytical 
procedures that validation addresses [1] . 
However, other analytical procedures, such as dissolution testing for dosage form 
or particle size determination for drug substance, are required for validation of 
analytical procedures. The revalidation of an analytical procedure is possible when, 
in particular circumstances, it could show changes in the synthesis of the drug substance, 
the composition of the fi nished product, or the analytical procedure. However, 
certain other changes may require validation as well. 
Method validation should confi rm that the analytical procedure employed for a 
specifi c test is suitable for its intended use. The validation of an analytical method 
ANALYTICAL METHODS 825

826 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
is the process by which it is established by laboratory studies that the performance 
characteristics of the method meet the requirement for the intended application. 
This implies that validity of a method can be demonstrated only through laboratory 
studies. Methods should be validated or revalidated before their introduction and 
routine use whenever the conditions change for which the method has been validated 
(e.g., an instrument with different characteristics) and the method change is 
outside the original scope of the method. 
Depending on the use of the assay, different parameters will have to be measured 
during the assay validation. Validation of analytical assays is the process of establishing 
one or more of the following as appropriate to the type of assay: accuracy precision 
(repeatability, intermediate precision), linearity, range, limit of detection, limit 
of quantifi cation, specifi city, and robustness [1] . For physicochemical methods there 
are accepted defi ned limits for these test parameters: 
1. Specifi city is the ability to assess unequivocally the analyte in the presence of 
components, which may be expected to be present and include impurities, 
degradants, matrices, and so on. 
2. Accuracy is the replica of agreement between the value which is accepted as 
either a conventional true value or an accepted reference value and the value 
found. 
3. Precision is the degree of agreement between a series of measurements 
obtained from multiple sampling of the same homogeneous sample under 
prescribed conditions. Precision may be considered at three levels: repeatability 
(precision under the same operating conditions over a short interval of 
time), intermediate precision (precision within - laboratory variations: different 
days, different analysis, different equipment, etc.), and reproducibility (precision 
between laboratories, collaborative studies, usually applied to standardization 
of methodology). 
4. Detection limit is the lowest amount of analyte in a sample which can be 
detected but not necessarily quantifi ed as an exact value. 
5. Quantitation limit is the lowest amount of analyte in a sample which can be 
quantitatively determined with suitable precision and accuracy. The quantifi cation 
limit is a parameter of quantitative assays for low levels of compounds in 
sample matrices and is used particularly for the determination of impurities 
and/or degradation products. 
6. Linearity is the ability, within a given range, to obtain test results 
which are directly proportional to the concentration of analyte in the 
sample. 
7. Range is the interval between the upper and lower concentration of analyte 
in the sample for which it has been demonstrated that the analytical procedure 
has a suitable level of precision, accuracy, and linearity. 
8. Robustness is a measure of the capacity to remain unaffected by small but 
deliberate variations in the method parameters and provides an indication of 
its reliability during normal usage. 

8.4.8 EQUIPMENT AND COMPUTER SYSTEMS 
8.4.8.1 Equipment Systems 
Generalities With the advent of the industrial revolution in the eighteenth century, 
machinery became essential for our lifestyle. Machinery is one of most important 
key factors of our society and the same occurs in the pharmaceutical industry. In 
the middle of the twentieth century, many procedures in the pharmaceutical industry 
were hand manufactured, requiring a lot of time and manpower. Packaging was 
one of the most rudimentary tasks performed; an image of hundreds of persons 
seated at a table packaging pharmaceutical products was a normal view in the 
middle of the last century. In the second half of the twentieth century manpower 
was replaced by machinery. Blister machines, automatic capsule - fi lling machines, 
and bigger compression and drying equipment became common in industry, dramatically 
increasing output and profi ts. Since then, the equipment used in the manufacture 
of medications has shown great improvement. According to GMPs all 
equipment must be located, designed, constructed, adapted, and maintained to suit 
the operations to be carried out. Their layout and design must minimize the risk of 
errors, allowing effective cleaning and maintenance in order to avoid cross - contamination, 
buildup of dust or dirt, and, in general, any adverse effects on the quality of 
products [10] . Equipment and premises must improve the product quality and safety 
and should never create any risk to the product. Adequate design must be taken 
into account when new equipment is created in order to improve its effi ciency and 
cleanness and reduce errors or breakdowns. 
Equipment should be placed under adequate environmental conditions in order 
to function accurately and in such a way as to prevent any risk of error or contamination. 
The environment must show minimal risk of causing contamination of materials 
or products when considered together with measures to protect the manufacture 
[10] . 
Manufacturing equipment should be designed, located, and maintained to suit its 
intended purpose. Repair and maintenance operations should not present any 
hazard to the quality of the products [10] . It should be designed so that it can be 
easily and thoroughly cleaned. It should be cleaned according to detailed and 
written procedures and stored only in a clean and dry condition. Washing and cleaning 
equipment should be chosen and used in order not to be a source of 
contamination. 
Production equipment should not present any hazard to the products. The parts 
of the production equipment that come in contact with the product must not be 
reactive, additive, or absorptive to such an extent that it will affect the quality of 
the product and thus present any hazard [10] . Equipment should not be made of 
materials that contaminate the fi nal product; therefore manufacturing equipment is 
usually composed of stainless steel and polymeric materials that are easily cleaned. 
According to this, natural materials must be avoided. 
Balances and measuring equipment of an appropriate range and precision should 
be available for production and control operations. Measuring, weighing, recording, 
and control equipment should be calibrated and checked at predetermined intervals 
by appropriate methods. Adequate records of such tests should be maintained [10] . 
EQUIPMENT AND COMPUTER SYSTEMS 827

828 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
All equipment for which calibration is applicable must be periodically calibrated. 
Periodicity of calibration must take into account the type of the equipment, risk 
assessment, and previous results. An extremely short calibration periodicity becomes 
extremely expensive, whereas an extremely long calibration periodicity could result 
in poor verifi cation results. Poor verifi cation results mean that the results generated 
are wrong and therefore all results obtained since the previous calibration must 
be reviewed. In conclusion, adequate evaluation of the calibration periodicity is 
essential. 
All equipment must be well identifi ed with name and code and must have a single 
operation instruction. A good operation instruction must contain at least the following 
points: (i) equipment description, containing the function of the equipment, 
name, code, serial number, model, manufacturer, accessories, dimensions, power 
source, connections, and component descriptions; (ii) function, containing a description 
of a procedure of how to handle the equipment, and all the steps required and 
options; (iii) calibration procedure, described in detail and providing the calibration 
periodicity; (iv) maintenance, a simple and easily maintenance procedures should 
be available; (v) cleaning described in detail; and (vi) security procedures. 
Equipment System Validation The GMP has special features regarding the equipment, 
setting several points for the validation, such as design qualifi cation, installation 
qualifi cation, operational qualifi cation, and performance qualifi cation. Each 
point must be given in a different document, so four documents must be generated 
for each equipment. Without any of these documents, the equipment cannot be 
considered adequate for GMP purposes. GMPs do not refer equipment validations; 
nonetheless, they refer these four items, and therefore, equipment can only be considered 
validated after the approval of these four documents. 
Equipment validation is not reliable, and several special features are required 
instead. GMP is concerned with several procedures before the equipment is placed 
into the service. After entrance into service, maintenance and calibration must 
be made periodically. The following focuses on the procedures required for new 
equipment to enter into service. 
Design Qualifi cation When a pharmaceutical industry intends to purchase new 
equipment, it must fi rst know and defi ne what type of equipment and which speci- 
fi cations it must have. Specifi cations must be set taking into account specifi c requirements 
of the company or facility. All the specifi cations of the equipment must be 
reviewed. This is called design qualifi cation. The requirements of the equipment 
must be defi ned in a specifi c document previous to the purchase by any pharmaceutical 
company. The document generated is used to justify the equipment selection 
from the various proposals. 
Installation Qualifi cation After equipment selection, it is necessary to assure that 
the equipment is installed well. The IQ document describes and validates the procedure 
of the equipment installation. It establishes confi dence that the process 
equipment and ancillary systems are capable of consistently operating within established 
limits and tolerances [10] . The equipment manufacturer and pharmaceutical 
company must agree and check the IQ, which must be approved by the pharmaceutical 
company at the end. This document certifi es that equipment was installed as 
specifi ed by the manufacturer and the purchaser. 

Operational Qualifi cation The OQ document certifi es that the equipment works 
as desired and defi ned by the manufacturer and the purchaser. An example is the 
acquisition of a new high - shear mixer/granulator where the paddle is put on rotation 
with a real calibrated rotation speed tester and, if the value obtained meets the 
specifi cations, the mixer paddle rotation passes the OQ test. If not, additional 
requalifi cation must be performed. All the test results must be introduced and veri- 
fi ed in the OQ report that is approved by the company at the end. The OQ document 
must describe several tests and related specifi cations to perform on the equipment 
in order to evaluate if it is working well, and the test to be performed must be 
described and approved by the manufacturer and the purchaser. Therefore, tests 
must be performed on the equipment, and for each one a description and signature 
of who performed and verifi ed the test are required. Usually the tests are performed 
by the manufacturer and verifi ed by the purchaser. These tests usually consist of 
evaluating if the mechanical and electric components of the equipment are working 
as desired. 
Performance Qualifi cation After evaluation of the way the equipment is working, 
it is required to perform some tests applying real situations to evaluate if the results 
are in agreement with those specifi ed. It is called performance qualifi cation (PQ) 
and establishes confi dence through appropriate testing that the fi nished product 
obtained by a specifi ed process meets all release requirements for functionality and 
safety [11] . 
Commonly when a pharmaceutical industry purchases a new compression 
machine, a PQ is conducted. First, the company must select some of its well - known 
products and prepare them until the compression phase as usual. The product must 
be compressed by the new equipment using the same compression conditions, such 
as compression force and tablet output. The obtained tablets must meet all the 
specifi cations of that product. The parameters specifi ed could be aspect, hardness, 
thickness, diameter, average mass and uniformity of mass, friability, and tablet 
output. If the tablets obtained comply with specifi cations, the compression machine 
is considered reliable to obtain a product with good quality and in a reproducible 
manner and the equipment is considered performance qualifi ed. 
Also in the PQ a document must be created with the description of the tests to 
be performed and the related specifi cations. This document must be verifi ed and 
approved before the test performance. After the tests, the results must be introduced 
in the document and the pharmaceutical company must approve the fi nal version. 
All the equipment qualifi cation documents must defi ne the equipment designation, 
tests to be performed, test specifi cations, materials, operators, reviewers, and 
responsibility for approval. 
After approval of design qualifi cation, installation qualifi cation, operational qualifi 
cation, and performance qualifi cation, the equipment is considered adequate for 
GMP proposes and can be placed in service. 
8.4.8.2 Computer Systems 
Generalities Technological advances are increasing extraordinarily, with the 
advent of computer system processes that were unthinkable a few years ago but 
nowadays are easy and simple. These advances are available every day in almost 
EQUIPMENT AND COMPUTER SYSTEMS 829

830 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
everything. Particularly in the pharmaceutical industry the last decade has seen the 
development of new analytical equipment, such as near Infrared, Fourier transform 
infrared (FTIR), or Raman spectroscopy using complex software to perform modulation 
or high - performance liquid chromatography (HPLC) and mass spectrophotometry, 
which have easier and simpler software. All this evolution in analytical 
equipment allows the industry to produce better and have higher profi ts. The computational 
systems revolution has reached not only analytical equipment used in 
pharmaceutical industry but also manufacturing equipment. Nowadays in the pharmaceutical 
industry almost every operation uses a computer system. Therefore, all 
computer systems in the pharmaceutical industry are supposed to be validated in 
order to assure that the results they generate are accurate and precise. In addition, 
wherever a computerized system replaces a manual operation, there should be no 
resultant decrease in product quality or quality assurance. Consideration should be 
given to the risk of losing aspects of the previous system which could result from 
reducing the involvement of operators. 
A computer system is composed of software and hardware, equipment, a processor, 
and a user, and it is used to execute a specifi c procedure. Regardless of whether 
the computer system is developed in - house or by a contractor or purchased off the 
shelf, establishing documented end - user requirements is extremely important for 
computer systems validation. Without fi rst establishing end - user needs and intended 
use, it is virtually impossible to confi rm that the system can consistently meet them. 
Once established, it should obtain evidence that the computer system implements 
those needs correctly and that they are traceable to system design requirements and 
specifi cations. It is important that the end - user requirements specifi cations take into 
account predicate rules [12] . 
Computer System Validation All computer systems presently used in the pharmaceutical 
industry need to be evaluated for risk assessment. The fi rst task in the 
validation of a computer system must be to determine if a potential failure in the 
computer system can cause a risk in product security, effi cacy, and quality. This must 
be applicable in all pharmaceutical industry areas, related with good clinical practices 
(GCPs), good distribution practices (GDPs), good laboratory practices (GLPs), 
or GMPs. Validation of a computer system also provides evidence that all of its 
components are working perfectly and generating adequate results. Thus, all of its 
components must be validated, such as applications, processes, users, and facilities. 
Risk assessment is the fi rst critical step in the validation of a computer system. 
After risk assessment, the validation protocol must be created, including all the 
points referred in the VMP. The procedures to be executed in a computer system 
should be defi ned, in addition to the specifi cations and tests to be performed. These 
tests could be trials for IQ, OQ, and PQ. After the test procedure and data evaluation, 
a validation report must be available. Therefore, a very well written and periodically 
reviewed (e.g., every year) VMP, validation protocol, and validation report 
are always necessary. 
The extent of validation necessary will depend on a number of factors, including 
the use to which the system is to be put, whether the validation is to be prospective 
or retrospective, and whether novel elements are incorporated. Validation should 
be considered as part of the complete life cycle of a computer system. This cycle 

includes the stages of planning, specifi cation, programming, testing, commissioning, 
documentation, operation, monitoring, and modifying [13] . 
A retrospective validation is applicable to computer systems in use when the 
VMP is elaborated. In this case, the real use of the system must be pointed out in 
order to allow a correct test plan. 
To produce an adequate validation close cooperation between key personnel and 
those involved with the computer systems is essential. People in responsible positions 
should have the appropriate training for the management and use of systems 
within their fi eld of responsibility which utilizes computers. This should include 
ensuring that appropriate expertise is available to provide advice on aspects of 
design, validation, installation, and operation of computerized systems [12] . 
After the validation procedure and personal characterization, it is necessary to 
describe how to handle the computer systems to validate them. As previously 
referred, a computer system is used to execute a specifi c procedure to give a defi ned 
result. Therefore, attention should be paid to siting equipment in suitable conditions 
where inappropriate factors cannot interfere with the system [13] . Therefore, the 
physical place of the computer system must be described in the validation protocol, 
and every time its place is changed a revalidation plan is required. 
A written detailed description of the system should be provided (including diagrams) 
and kept up to date. It should describe the principles, objectives, security 
measures and scope of the system, the main features of the way in which the computer 
is used, and how it interacts with other systems and procedures [13] . A computer 
system is a particular type of equipment, and, as for any equipment, it is 
required to have a SOP available and, if possible, near the equipment. 
The software — a program enabling computer to perform a specifi c task — is a 
critical component of a computerized system. Even a simple calculus sheet could 
be considered software, as it allows a computer to perform a specifi c task. The user 
of such software should take all reasonable steps to ensure that it has been produced 
in accordance with a system of quality assurance [13] . 
The system should include built - in checks of the correct entry and processing of 
data. In order to verify the validation data, some computer systems may periodically 
be submitted to a defi ned group of inputs for which the result is known and the 
result must be kept and fi led. If the results are acceptable, the computer system is 
operating well, but if the results do not match the expected ones, the computer 
system is not working properly and maintenance is required. In this situation, all 
the results obtained from the referred computer system since the last validation 
verifi cation are consided questionable and must be reevaluated. 
Before a system using a computer is brought into use, it should be thoroughly 
tested and confi rmed as being capable of achieving the desired results. If a manual 
system is being replaced, the two should be run in parallel for a time as part of this 
testing and validation [13] . This means that every time maintenance has made an 
intervention in the equipment and any piece has been adjusted or replaced, a revalidation 
is required. This maintenance intervention could be in both hardware and 
software. 
Data should only be entered or amended by authorized people. Suitable methods 
of deterring unauthorized entry of data include the use of keys, pass cards, personal 
codes, and restricted access to computer terminals. There should be a defi ned 
EQUIPMENT AND COMPUTER SYSTEMS 831

832 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
procedure for the issue, cancellation, and alteration of authorization to enter and 
amend data, including changing personal passwords. Consideration should be given 
to systems allowing for recording of attempts to access by unauthorized people [13] . 
A computer system must have different user levels and the access to each application 
should be protected by a user name and password. Critical operations, such as 
method modifi cations, calibration procedures, and adjustments, must only be possible 
by specifi c personnel. Operators must only have access to procedures defi ned 
and without any possibility of changes. All the operations must be automatically 
recorded in the equipment, including operator user name, time, and date. These 
records may be available for a quality assurance audit. 
When critical data are being entered manually (e.g., the weight and batch number 
of an ingredient during dispensing), there should be an additional check on the 
accuracy of the record which is made. This check may be done by a second operator 
or by validated electronic means [13] . Frequently, computer systems require external 
data to realize the specifi ed task. The introduction of an external data is a critical 
step in the computer system procedure, because an error in data introduction may 
produce an erroneous result. To avoid errors one can proceed by double verifi cation 
followed by the automatic registry prove or, alternatively, double verifi cation consisting 
of the introduction of data by one operator while another verifi es the data 
introduced and both sign the register. An automatic registry proves that it is available 
when the values introduced in a computer system came from another computer 
system registry. For example, in the dissolution test for tablets, tablet weights introduced 
in the computer system could be available from a balance associated with an 
automatic registry; that registry is the proof that the data introduced are correct. 
The system should record the identity of operators entering or confi rming critical 
data. Authority to amend entered data should be restricted to nominated people. 
Any change to an entry of critical data should be authorized and recorded with the 
reason for that change. Consideration should be given regarding the creation of a 
complete record of all entries and amendments (an “ audit trail ” ) [13] . 
As previously mentioned, changes to a system or to a computer program should 
only be made in accordance with a defi ned procedure and should include provision 
for validating, checking, approving, and implementing the change. Such a change 
should only be implemented and recorded with the agreement of the person responsible 
for the part of the system concerned. Every signifi cant modifi cation should be 
validated. 
For quality auditing purposes, it should be possible to obtain clear printed copies 
of electronically stored data. Data should be secured by physical or electronic means 
against wilful or accidental damage. Stored data should be checked for accessibility, 
durability, and accuracy. If changes are proposed to the computer equipment or its 
programs, the above - mentioned checks should be performed at a frequency appropriate 
to the storage medium being used. Data should be protected by backing up 
at regular intervals. Back - up data should be stored as long as necessary at a separate 
and secure location [13] . 
All data generated in a computer system must be available, secure, and safely 
stored. Unauthorized people cannot have access to data fi les. It must be impossible 
to overwrite any data, but all recalculation required must generate a new data and 
not substitute the previous one. Safety procedures must be available. Data should 
be backed up periodically following specifi c SOPs, and backup copies must be iden

tifi ed and stored in a defi ned and safe place. All backup data must be evaluated for 
the backup process, ensuring that the data backup has been well performed. Periodically 
all data must be checked for integrity in order to ensure that they are well 
preserved and no data have been lost. 
Adequate alternative arrangements should be available for systems which need 
to be operated in the event of a breakdown. The time required to bring the alternative 
arrangements into use should be related to the possible urgency of the need. If 
the system fails or breaks down, the procedures to be followed should be defi ned 
and validated. Any failures and corrective actions taken should be recorded [13] . 
Special cases of breakdowns are power breakdowns. It is required that some computer 
systems work for 24 h and should not be interrupted after starting data acquisition. 
Consequently, an extra uninterruptible power supply (UPS) is required to 
allow the operation of these computerized systems during power breakdowns. 
A procedure should be established to record and analyze errors and to enable 
corrective action to be taken. Every computer system can originate errors that must 
be documented. Every time an error occurs, it must be analyzed and, if applicable, 
some adjustments in the computer system may be required. Adjustments have to 
be well defi ned and documented; in some cases a revalidation of the system may be 
required. 
When outside agencies are used to provide a computer service, there should be 
a formal agreement including a clear statement of the responsibilities of that outside 
agency [13] . When the release of batches for sale or supply is carried out using a 
computerized system, the system should allow only a qualifi ed person (QP) to 
release the batches and it should clearly identify and record the person releasing 
the batches. This is possible by giving the QP an operational level in the computer 
system that allows it to release batches. No other person should have the same 
operational level authorization as the QP. 
Software Validation Authorities are demanding rules in order to outline the software 
validation principles used in medical device software or the validation of 
software used to design, develop, or manufacture medical devices. 
Guidances recommend an integration of software life - cycle management and risk 
management activities. Software validation and verifi cation activities must be conducted 
throughout the software life cycle [12, 14] . Software verifi cation and validation 
are terms frequently confused. Software verifi cation is defi ned as the process 
that provides objective evidence that the design outputs of a particular phase of the 
software development life cycle meet all of the specifi ed requirements for that phase 
[14] . Software verifi cation consists of tasks performed to evaluate if the software is 
performing desired tasks and providing desirable results. Software verifi cation is 
part of software validation. Software testing is one of the many verifi cation activities 
intended to confi rm that software development output meets its input requirements. 
Other verifi cation activities include various static and dynamic analyses, code and 
document inspections, walkthroughs, and other techniques [14] . 
Although actual guidelines are only applicable in software used as a component 
part or accessory of a medical device (e.g., blood establishment software, programmable 
logic controllers in manufacturing equipment, software that records and 
maintains the device history record), the validation principles presented in these 
guidelines could be applicable to any software validation. 
EQUIPMENT AND COMPUTER SYSTEMS 833

834 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
Software is not a physical entity and, unlike some hardware failures, software 
failures occur without advanced warning. One of the most common software failures 
is branching, that is, the ability to execute alternative series of commands based on 
differing inputs. The software branching capacity makes the commands extremely 
complex and diffi cult to validate once errors occur as an answer of a specifi c input, 
and until the introduction of that specifi c input error has not been detected. Software 
input can be almost any data and, and since it is impossible to introduce all 
data into a software, validation of data is extremely diffi cult. Thus, results are considered 
to be of high confi dence level. The majority of software problems occur as 
a consequence of errors in the software design and development and are not directly 
related to the software manufacture. It is simple to manufacture several software 
copies that work perfectly and as the original one. 
Software validation is not separately defi ned in the quality system regulation. The 
FDA considers software validation to be the “ confi rmation by examination and 
provision of objective evidence that software specifi cations conform to user needs 
and intended uses, and that the particular requirements implemented through software 
can be consistently fulfi lled ” [14] . Software validations have special concerns 
on software installation, implementation, and utilization. A software validation consists 
in several tests, inspections, and verifi cations performed to assure the adequate 
installation and use of software and that the tasks performed meet all the specifi cations 
defi ned. Software validations must be performed under the environmental 
conditions to which software will be submitted. This is particularly important in 
medical devices that are used under special conditions, such as close to or inside the 
human body. 
A software validation plan must take into account the risk analyses for the software; 
therefore a critical software must have a high level of confi dence and must be 
submitted to deep validation processes, whereas noncritical software may be submitted 
to less extensive validation processes. 
Seemingly insignifi cant changes in software code can create unexpected and very 
signifi cant problems elsewhere in the software program. The software development 
process should be suffi ciently well planned, controlled, and documented to detect 
and correct unexpected results from software changes [14] . Maintenance of a software 
must be carefully performed because even a few small changes could develop 
dramatic software results. Accuracy and thorough documentation are essential in 
order to assure the software validation. 
Software validation is not a simple and easy task; therefore an adequate schedule 
and task plan for software validation and verifi cation are required to avoid unnecessary 
time and money expenses. One day of planning with no experiments performed 
is better than one day of experiments without planning. 
Electronic Documents Several special cases of electronic documents have been 
targeted by specifi c offi cial regulations. These are electronic records and electronic 
signatures. For instance, the FDA issued regulations that provide criteria for acceptance, 
under certain circumstances, of electronic records, electronic signatures, and 
handwritten signatures executed to electronic records as equivalent to paper records 
and handwritten signatures executed on paper. These regulations, which apply to all 
FDA program areas, were intended to allow the widest possible use of electronic 
technology compatible with the FDA ’ s responsibility to protect public health [12] . 

Some interpretations for this guideline included almost every record generated 
in a computer as the fi nal target, but another was very simple, considering only a 
few documents under the scope of this guideline. The records required to be maintained 
under predicate rules or submitted to the FDA would apply whenever records 
in electronic format replace paper hardcopies. On the other hand, when the computer 
is used to generate paper printouts of electronic records and those paper 
records meet all the requirements of the applicable predicate rules, the FDA would 
generally not consider people to be “ using electronic records in lieu of paper 
records. ” The same is considered when people rely on paper records to perform 
their regulated activities. In these cases, the use of computer systems in the generation 
of paper records would not trigger the offi cial requirements [12] . Computer 
systems that generate a document that is printed out and signed by hand are 
excluded in this guideline. The guideline only regulates documents generated in a 
computer system that are fi led, signed, and modifi ed electronically. Since the documents 
are not printed and signed, special considerations must be taken in order to 
assure that an adequate fi le is performed, undesirable modifi cations are avoided, 
and the signature is safe and personalized. 
This guideline is only applicable to documents for which it is required to be 
maintained. All documents created and signed electronically that are not required 
to be maintained are not under the scope of this guideline. The decision of which 
documents are under the scope of this guideline is taken by the pharmaceutical 
industry and must be well documented and justifi ed. 
Electronic signatures are intended to be the equivalent of handwritten signatures, 
initials, and other general signings required by predicate rules. They include electronic 
signatures that are used, for example, to document the fact that certain events 
or actions occurred in accordance with the predicate rule (e.g., approved, reviewed, 
and verifi ed). 
Validation Team A well-defi ned validation team with a well - written description of 
responsibilities is required and assures the adequate realization of the validation 
tasks. A validation team should be composed by different responsibilities: responsible - 
of - validation team, team leader, archive manager, test coordinator, quality assurance 
member, tester, and witness. The responsible - for - validation team elaborates and 
approves the VMP, protocols, and reports. The team leader should be responsible for 
the computer system validation and utilization. An archive manager is responsible 
for the management of all computer system validation documents. The test coordinator 
is responsible for the computer system test and coordinates the elaboration and 
operation of tests for evaluating the performance of the computer system. A quality 
assurance member is required to periodically inspect and train the personnel and 
review all the validation documents. The tester is responsible for the execution of 
the tests required to perform the validation protocol. The witness is responsible for 
observing and reviewing the operations of the tester. 
8.4.9 CLEANING VALIDATION 
Cleaning validation is defi ned as the “ the process of providing documented evidence 
that the cleaning methods employed within a facility consistently controls potential 
CLEANING VALIDATION 835

836 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
carryover of product (including intermediates and impurities), cleaning agents and 
extraneous material into subsequent product to a level, which is below predetermined 
levels ” [15] . 
The reasons behind the validation of cleaning procedures are the assurance of 
the safety and purity of the product (customer requirement), it is a regulatory 
requirement in active pharmaceutical ingredient product manufacture, and it assures 
the quality of the process from an internal control and compliance point of view 
[15] . 
Cleaning should be carried out with relative ease and with the use of standard 
cleaning materials [7] . Vacuum facilities should be available for cleaning and contact 
parts fouls be wiped down and sanitized utilizing a sanitizing agent. The equipment 
should be washed, dried, covered, and stored in an equipment storage area. 
As in manufacturing, the packaging operation starts with the generation of the 
packaging order. This consists of the approved packaging components, the batch 
number of the product to be packaged, and quantities of each. A supervisor verifi es 
the accuracy and completeness of the packaging order, including expiration date, 
line being used, and any other special equipment being used for that operation. 
These steps are accomplished prior to bringing components to the line. The complete 
line area, including all equipment, is verifi ed as being properly disassembled 
and cleaned of all product and components from the previous packaging operation. 
After the QA department has verifi ed that the area has been cleared and cleaned, 
the components must be brought to the line for mechanical setup. Once the setup 
mechanic has completed all the adjustments required, a supervisor oversees the 
prestart procedure. This consists of cleaning all products used during the setup; 
counting labels used in the label setup; rechecking all components and lot numbers; 
verifying the bottle count, all stamps, and lot numbers; and signing the packaging 
order that all is in readiness to start packaging operations. 
Nonappropriate cleaning procedures will develop batches of poor quality due to 
the risk of presence of a number of contaminants, such as precursors of the drug, 
degradation products, solvents and other materials employed during the manufacturing 
process, microorganisms, cleaning agents, and lubricants [15] . 
8.4.10 CONCLUSIONS 
Since the middle of the twentieth century the pharmaceutical industry has been a 
leader in terms of quality and security of manufacturing, associating higher production 
effi ciency with higher profi ts. Therefore, validation became essential, that is, the 
confi rmation and evidence that all facilities, equipment and processes work as 
desired, generating quality products. 
In the present chapter, the pharmaceutical industry validation system has 
been reviewed. To have an appropriate validation system it is fi rst required to 
defi ne which equipment, facilities, and processes will be validated, when they 
will be validated, and by whom this must be performed. This defi nition is based 
on a risk assessment priority and is written in a specifi c document, the 
so - called MVP. In order to generate an adequate validation report, all the validation 
activities should be described in the validation protocols, SOPs, and specifi c 
procedures. 

Facilities validation is a critical process in a pharmaceutical industry and the types 
of pharmaceutical forms produced must be considered. Facilities that produce different 
pharmaceutical forms have different specifi c requirements and different critical 
parameters based on risk assessment. All facilities must have an adequate fl ow 
of people, raw materials, bulk products, and fi nished products. These fl ows must be 
created in order to avoid cross - contamination. Additionally, pressurized rooms and 
adequate SOPs should be supplied to minimize the risk of cross - contamination. 
Controlled air temperature and humidity are also required and should be validated 
to ensure adequate stability of the product. 
MVP is a crucial procedure to confi rm that the product manufacture is adequate 
and is generated in a consistent manner with the same quality. Defi nitions of equipment, 
rooms, quality of raw materials, and process parameters are necessary to 
validate the manufacturing process. Maintaining these defi nitions ensures that 
the product has the same quality after the process validation. Changes in equipment, 
raw materials, or process parameters require a manufacturing process 
revalidation. 
Analytical methods validation is one of the most regulated validation processes 
in the pharmaceutical industry. Analytical validations are required to demonstrate 
that the methods employed are the most indicated for each product and that the 
results obtained are reliably correct. All methods employed in raw and fi nished 
product materials analysis are required to be validated. 
Equipment validation is comprised of four critical operations: design, installation, 
operational, and performance qualifi cations. These operations will confi rm that the 
equipment has adequate specifi cations, installation, and functions, manufacturing a 
product with adequate properties. After these procedures, whenever equipment is 
installed, it must be periodically verifi ed and calibrated in order to ensure adequate 
performance. 
Computer systems are a specifi c type of equipment that must be validated as well. 
These systems have specifi c requirements if they are used to collect and process 
data. Associated with computer hardware is the software, carefully validated in 
order to prove that the data generated by them are correct. 
In order to avoid cross - contamination, another concern with respect to equipment 
is the cleaning process, which must comprise cleaning SOPs to ensure adequate 
cleanliness. Cleaning validation must be performed based on risk assessment and 
worst - case scenarios. 
REFERENCES 
1. World Health Organization (WHO) ( 2006 ), WHO expert committee on specifi cations for 
pharmaceutical preparations (fortieth report) — Supplementary guidelines on good manufacturing 
practices (GMP): Validation, WHO, Geneva. 
2. Thomas , E. , Grochulski , A. , Patel , R. , George , S. M. , and Zhang L. ( 2006 , Sept.), The laboratory 
control system: Fulfi lling cGMP requirements , BioPharm Int. , 26 – 32 . 
3. Slater , S. ( 1999 ), Biopharmaceutical Validation: An Overview in Biopharmaceuticals, an 
Industrial Perspective , Springer - Verlag . 
4. Chaloner - Larsson , G. , Anderson , R. , and Egan , A. A WHO Guide to Good Manufacturing 
Practice (GMP) Requirements , 1997, World Health Organization, Geneva. 
REFERENCES 837

838 PHARMACEUTICAL MANUFACTURING VALIDATION PRINCIPLES 
5. International Conference on Harmonization (ICH) ( 2000 ), Good manufacturing practice 
guide for active pharmaceutical ingredients, ICH, Geneva. 
6. European Medicines Evaluation Agency (EMEA) ( 2003 ), Good Manufacturing Practices 
, Chapter 4: Building and facilities, EMEA. 
7. Nash , R. A. ( 1990 ), The essential of process validation , in Pharmaceutical Dosage Forms, 
Tablets , Marcel Dekker Inc , New York . 
8. Hanna , S. A. ( 1990 ), Quality assurance , in Pharmaceutical Dosage Forms, Tablets , Marcel 
Dekker Inc , New York . 
9. Connolly , R. J. , Berstler , F. A. , and Coffi n - Beach , D. ( 1990 ), Tablet production , in Pharmaceutical 
Dosage Forms, Tablets , Marcel Dekker Inc , New York . 
10. European Medicines Evaluation Agency (EMEA) ( 2003 ), Good Manufacturing Practices 
, Chapter 3: Premise and equipment, EMEA. 
11. U.S. Food and Drug Administration (FDA) ( 1987 ), Guideline on general principles of 
process validation, FDA, Rockville, MD. 
12. U.S. Food and Drug Administration (FDA) ( 2003 ), Guidance for industry, CFR Part 11, 
Electronic records; electronic signatures, good manufacturing practices, FDA, Rockville, 
MD. 
13. European Medicines Evaluation Agency (EMEA) ( 2003 ), Good Manufacturing Practices 
, Annex 11: Computerised systems, EMEA. 
14. U.S. Food and Drug Administration (FDA) ( 2002 ), general principles of software validation, 
good manufacturing practices, Rockville, MD. 
15. Active Pharmaceutical Ingredients Committee (APIC) ( 1999 ), Cleaning validation in 
active pharmaceutical ingredient manufacturing plants, APIC. 

839 
INDEX 
Accelerated stability testing, 691–693 
Accuracy, 730 
Active pharmaceutical ingredients, 
546–547, 563 
Additives, 468–473, 467 
Aluminum, 486, 489–491 
Analytical method validation 
characteristics, 729–737 
Anharmonicity, 372–373 
Arrhenius equation, 691–693 
Audit checklist, 225–237 
Audit trail, 30 
Biotechnological products, 668–671 
Biowavers, 84–85 
Box plot, 291 
Bracketing, 592–594 
Bulk pharmaceutical excipients, 547–548 
Cause-and-effect diagram, 288 
cGMP, 202–204 
Change control, 89–91 
Changes in manufacturing sites, 73–74 
Chemical reaction kinetics, 627–628 
Chemometrics, 386–409, 416–419 
Class I, 166 
Class II, 166 
Class III, 166 
Cleaning validation, 835–836 
Climatic zones and recommended storage 
conditions, 578 
Clogenicity, 106–107 
Common Technical Document (CTD), 
333–334 
Competency-based training, 438–439 
Components of validation activity, 93 
Consent decrees, 59 
Container extractables and leachables, 
665–668 
Containers, 481–483 
Contaminants found in water for injection, 
465 
Control chart, 292–293 
Control limits, 305–306 
Corrective and Preventative Action 
(CAPA), 222, 25 
Criminal proceedings, 61–66 
Data mining, 358, 360–361 
Data warehousing, 359 
DEHP, 494, 509 
Delivery systems, 506–516 
Design of stability studies, 589–598 
Detection limit, 733–734 
Disgorgement, 59 
Drug product salvaging, 24 
Pharmaceutical Manufacturing Handbook: Regulations and Quality, edited by Shayne Cox Gad 
Copyright © 2008 John Wiley & Sons, Inc.

840 INDEX 
Drug shelf life for multiple batches, 
604–617 
Elastomeric closures, 499–506 
Electronic documents, 834–835 
Endogenous contaminants, 459–479 
Endotoxin and pyrogenicity testing, 106 
Equipment system validation, 828–829 
European pharmacopeia, 550–551 
Excipients, 36–38, 562–563 
Extractability tests, 506 
Extractables, 667 
Facilities validation, 818–822 
Factor VII, 669–670 
Factors infl uencing stability of drugs and 
drug products, 644–654 
False Claims Act, 59–61 
FDA inspections, 48–49 
FDA’s Offi ce of Criminal Investigations 
(OCI), 47 
Finished products, 562 
Form, 49–51, 483–484, 486–487 
FT spectrometers, 414 
Full stability study design, 591–592 
Functional groups subject to oxidation, 
694–695 
Genetic stability, 109–112 
Glass classifi cation, 483 
Glass containers, 656 
GXP trainer(s), 444 
Harmonization, 87–89 
Hematopoiesis regulation, 108–109 
Histogram, 289 
Human platelet lysate, 101–102 
Hygroscopicity, 652 
Hyperspectral imaging, 412, 428–429 
Immune modulatory effects, 108 
Impurities in offi cial articles, 458 
Injunctions, 58–59 
Instrument classifi cation, 793 
Internal audit, 217–220 
International Conference on 
Harmonization, 133–135 
Japanese pharmacopeia, 551 
Leachables, 667 
Lean manufacturing, 318–320 
Legacy systems, 3–31 
Limit of Detection (LOD), 767, 773–774 
Limit of Quantifi cation (LOQ), 774–775, 
767 
Limulus Amoeobocyte Lysate (LAL), 
534 
Linearity, 734–735, 775–776 
Long-term stability analysis 598–626 
Management trainer(s), 444 
Manufacturing process validation, 822–825 
Master production and control records, 21 
Matrixing, 594–595 
Measurement Uncertainty (MU), 751–752 
Metals, 657 
Microbial degredation, 697 
Microbiological testing, 106 
MIR/NIR chemical imaging, 381 
Multiple Scatter Correction (MSC), 392 
Multivariate image analysis, 418–419 
National GMP regulations, 120–131 
Natural tolerance limits, 305–306 
New drug delivery systems, 671–673 
Nonsterile liquid dosage forms, 661–662 
Normal probability plot, 290–291 
Object-oriented process modeling method 
(CIMp), 179 
Operational classifi cation, 801–802 
Ordinary impurities, 458 
Organoperoxide radicals, 667 
Parental Drug Association (PDA), 504 
Pareto Chart, 288–289 
Particulate matter, 516–525 
Patter recognition, 397–398 
Pharmaceutical excipients, 653–654 
Phenotypic identity of MSC, 106 
Photolysis, 695 
Photostability studies, 573–575 
Plastic containers, 489–490, 494, 498–499, 
501, 656–657 
Polymerase Chain Reaction (PAR), 534 
Polymeric materials, 495 
Polymorphism, 652–653, 696 
Potential tumorigenicity, 109–112 
Precision, 730–731 
Predicting shelf life, 690–691 
Principle-component analysis, 393–395 
Process capability indices, 306–307 
Process controls, 355, 11–13

INDEX 841 
Process Failure Risk Analysis (PFRA), 
184 
Process ownership, 264 
Process validation, 91–92 
Pyrogens, 11 
Quality management, 241–252, 316–318 
Quality risk management, 220–222, 
333–335 
Quantitation limit, 734 
Quantum mechanical model, 369–372 
Racemization, 696 
Radio-Frequency Identifi cation (RFID), 
186–187 
Raman spectroscopy, 376–379 
Range, 735–736, 775–776 
Real-Time Release (RTR), 341–342 
Repeatability, 730 
Reproducibility, 731 
Residual solvents, 480–481 
Returned drug products, 23–24 
Risk-based orientation, 327–328 
Robustness, 736–737, 769, 776 
Root Mean Square Error of Prediction 
(RMSEP), 403 
Root-cause analysis, 355 
RSD (relative standard deviation), 34 
Rubber, 657–658 
Ruggedness, 769, 776 
Safety trainer(s), 444 
Scatter diagram, 289–290 
Section 305 proceedings, 62 
Seizures, 56–58 
Sensitivity, 769, 776 
Shelf life determination, 575–578 
Short-term stability analysis, 626–633 
SIMCA classifi cation, 398–399 
Site changes, 38–40 
Software validation, 833–834 
Solvents, 480–483 
Solvolysis, 693–694 
Sorption-desorption moisture transfer, 
674–675 
Special control charts, 302–306 
Specifi cation limits, 305–306 
Specifi city, 766 
Stability and shelf life, 579–580, 585 
Standard error of calibration (SEC), 404 
Sterile liquid dosage forms, 662–665 
Subject matter experts, 446 
SUPAC, 68–72, 72–85, 85–86 
Tandem mass spectrometry, 537 
Traceability, 747–748 
Training development, 449–451 
Trend analysis, 217 
U. S. Pharmacopeia (USP), 560, 548–550, 
793, 688, 458 
Validation master plan, 814–815 
Validation plan, 794–796 
Validation protocols, 815–818 
Validation reports, 817–818 
Validation, 30 
Vaporization, 696–697 
Variable-concentration kinetic 
experiments, 716–720 
Variable-ionic-strength kinetic 
experiments, 720–721 
Variable-temperature kinetic experiments, 
713–716 
Viability, 106 
Vibrational spectroscopy, 365–386 
Warning letter, 53–54 
Water, 460 
World Health Organization, 131