Новости
PHARMACEUTICAL MANUFACTURING HANDBOOK
Regulations and Quality
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
