The general principle of the GMP-compliant documentation (e.g. for manufacture, cleaning, analysis, qualification, maintenance, etc.) is that a plan or protocol containing all the work stages to be carried out and the exact acceptance criteria is first drawn up and then authorised by the responsible person. While the work is being carried out, the actual data is recorded and entered in a record. The actual data is then compared with the requirements, conclusions are drawn and this is all authorised by the responsible person. This may either take place as part of the record or as a separate outline report.

The documentation for validation projects is also structured accordingly. Requirements for the format and structure of the validation protocol and report should be set out in the validation master plan.

1 Elements of the validation protocol

Unless it is already governed by the VMP, a validation protocol must take the aspects listed in figure 1 into consideration (see also PIC/S PI 006 and Annex 1 of the EMEA "Note for Guidance on Process Validation"):

Determination of the validation batches

The production and/or packaging batches which are intended to be included into the process validation must be listed in the validation protocol.

At least 3 consecutive batches must be produced for the prospective and the concurrent validations. The size of these batches must correspond to the recent commercial scale. If flexibility in batch size is desired for routine production, these different batch sizes should be covered in the validation (cf.. EMEA "Note for Guidance on Process Validation", Paragraph 5), for example, at least one batch should be manufactured in both the smallest and the largest intended batch sizes. If the batch size of the validation batches differs greatly from the intended production scale (e.g. by more than a factor of 10), for example, because the active pharmaceutical ingredient is very expensive, this deviation must be justified in the validation protocol and the possible effect of the difference on the validity of the validation must be discussed. If production is later in the routine, the new batch size must be concurrently revalidated (see chapter 7.D.2 Incidences requiring revalidation).

In this context, the concept of a "batch" occasionally causes difficulties, especially when there is continuous production or production in lots (with intermediate cleaning only). For the purpose of process validation it should be tried to stick as close as possible to the official definition ("a defined quantity of starting material, packaging material or product processed in one process or series of processes so that it could be expected to be homogeneous" - EU GMP Guideline) or to written, internal definitions (e.g. batch = lot) of the company.

If the concept of a batch is to be defined differently to usual for validation purposes only (e.g. between two intermediate cleaning procedures), this must be extremely well justified (in writing, in the validation protocol or even in the VMP).

Detailed sequence of operation with in-process controls

The basis of this is an approved master production record (reference to which must be given) with required setting parameters and acceptance limits for the equipment and machines. If the batch size of the validation batches differs from the routine production batch size by more than a factor of 10 , this deviation must be justified in the validation protocol and the possible effect of the difference on the validity of the validation must be discussed. If production is later in the routine, the new batch size must be concurrently revalidated.

If several dosages of a preparation are to be validated together using the bracketing, matrixing or family grouping concept, the rationale and the exact procedure must be detailed in the validation protocol. (See chapter chapter 7.F Validation master plan.)

Quality attributes

These include content, yield and physical characteristics of the product, for example.

However, this involves more than simply fulfilling the final product specifications - these are checked by the batch release in any event (EMEA Note for Guidance on Process Validation, Paragraph 1). The point of validation is to make certain that the process is reproducible, in addition to fulfilling the specifications. The question is therefore: "how can I tell that my process is running well?" or even: "what conditions must be fulfilled for my product to be good?"

Examples of such quality attributes - in addition to fulfilling the final product specifications - are:

• Fulfilment of special requirements for the yield
• Specifications of intermediate products
• Critical control points (CCPs)
• Process or product parameters that are not included in the final product specification
• Sampling procedure suitability
• Suitability of in-process controls
• Process capability investigations

Quality of the raw materials used

The quality of the starting materials which are used in the process validation has a direct influence on the result of the validation. Therefore, the raw materials must be specified and the qualified suppliers must be listed in the validation protocol, unless the specifications and suppliers are determined by internal product numbers. If the validation is carried out by third parties, it is important to decipher internal product codes in the validation protocol so that the raw material quality required is clearly comprehensible to the contractual partner.

In order to cover possible differences in quality between the various approved suppliers or even differences between batches from the same supplier in the process validation, the production of validation batches should be planned such that different batches of raw materials are used. This is particularly important for the active pharmaceutical ingredient, but also for quality relevant excipients, for example excipients that affects the dissolution, or for (synthetically modified) natural products. When validating packaging processes, it is important to purposely use different batches of packaging materials.

Occasionally, it is difficult to obtain different batches of excipients or packaging materials from the manufacturer for the purposes of validation - especially when these are new types of packaging or excipients and only one batch is available from just one manufacturer. The supplier is then often confronted with the requirement for three different batches to be supplied. In these cases, thorough consultation between the supplier and pharmaceutical manufacturer is necessary and in both parties' interests. The question arises of how much more knowledge and explanatory power the process validation would provide in regard to this point if the three required batches of starting materials are generated from a single production lot with more or less arbitrary subdivision, e.g. by briefly switching off the equipment and starting it up again, or by using different packaging processes. In this case, the "three batches" of starting material are indeed largely identical. The amount of work required for the process validation remains the same however - without the desired gain in knowledge.

In this case, it makes more sense to complete the validation using the single batch of starting material in the first place (to be justified in the validation protocol). Then, as soon as new starting material batches are actually available (good communication between the manufacturer and pharmaceutical customer is required), the effect of these new batches can be investigated in an concurrent validation (only for this aspect - complete revalidation is not required).

List of the facilities and equipment used, including measuring and recording instruments and any alternatives

The result of a process validation is only ever valid for the facilities and machines used when the validation was carried out. A transfer to similar or other equipment should be proven, where necessary, with revalidation. It is therefore important that the equipment be precisely defined in the validation protocol. Its calibration status must also be noted. If alternatives are permitted for certain processing steps, e.g. different types of blenders, this must be explained in the validation protocol. In this case, the validation must be planned in such a way that suitable data is collected for all specified types of blenders. More than three validation batches may be required.

Risk analysis

As part of the risk analysis, the effect of any deviations on the quality of the final product is evaluated for each processing step. The risk analysis may be drawn up as a separate document and referred to in the validation protocol (see chapter 7.G Risk analysis).

Definition of the critical processing steps and process parameters

As a result of risk analysis, the processing steps are divided into non-critical and critical. Critical processing steps require validation. The relevant samplings and tests, the results of which justify the process parameters provided in the manufacturing instructions, must therefore be established in the validation protocol.

Occasionally, it is revealed that process parameters which were classified as "critical" in the risk analysis cannot be technically influenced in any way, such as the non-conditioned air supply for a dryer.

If this is the case, the range (e.g. limits for humidity) within which the process (from experience) is reliable and reproducible must be determined and this must be documented in the process validation.

If at a later stage the parameters move outside this validated range (e.g. dryer air supply too moist) and it is not possible to influence this in any way, the process must not be carried out on this day because it is not running under validated conditions.

If the process goes ahead in spite of this, this deviation must be thoroughly investigated and a thorough risk analysis carried out: how great is the risk caused by the product because it was manufactured with a non-validated step in the process?

However, it may also mean that certain parameters may have to be modified, e.g. the dryer air supply must be attached to a cold trap.

Compilation of test plans

Detailed description of the investigations that are carried out for validation purposes in addition to the normal in-process controls specified in the manufacturing instructions. These tests may include:

• More frequent monitoring of process parameters
• More frequent in-process controls
• Additional examinations, e.g. content uniformity or particle size distribution
• Tests after intentionally built in events, e.g. longer holding times, reduced/increased production speeds.
• Function controls (e.g. for automatic detection systems)

The test method must be detailed for every individual test.

When establishing a test plan, it is important to make sure that enough data is collected so that any subsequent statistical interpretation is meaningful.

Determination of the acceptance criteria

It is important to determine beforehand the results that are expected and the range of variation which is acceptable, or the other way round which results cannot be accepted because they indicate that the process is not running in a controlled manner.

Sampling plan

Almost each test begins with taking a sample - unsuitable sampling can hardly result in a meaningful test result. Therefore, it is important that each sample that is to be taken as part of the validation project is described in detail in a sampling plan, which is part of the validation protocol. Sampling is often paid little attention during the planning stages, but inadequate sampling is quickly blamed when unexpected results are obtained. Besides it is important to plan for samples for validation purposes in addition to the regular in-process control samples specified in the manufacturing instructions, e.g. regularly during a manufacturing operation or from various points in a container, especially for critical production stages.

The following must be described in detail in the sampling plan:

• Place (where? top - middle - bottom)
• Time (when? time, frequency or start - middle - end)
• Type (how? using which instrument?)
• Quantity (how many g, ml or pieces per sample?)
• Number of samples (how many samples at each time?)
• Use (for content uniformity, determination of physical properties, residual moisture, etc.)
• Packaging and storage (in the case samples are not tested immediately at the sampling side, but transferred to a laboratory)
• Labelling (traceability of sample series)

An intetended use must be known for all samples. Samples must not be taken "just in case". In contrast, the testing schedule can allow for specific testing (e.g. uniformity of content), e.g. 40 samples to be taken according to the plan, only 20 of which are to be tested initially and the other samples only to be analysed in a predetermined case (e.g. standard deviation exceeds xyz).

It is also important that the members of staff who carry out the sampling are specially trained.

Examples for sampling can be found in figure 2 and figure 3.

It is useful to keep all samples taken during validation runs until the validation is fully complete (i.e. authorised). Of course, this excludes those cases in which the samples are not stable for this length of time, for example because they separate, are affected by oxidation or deteriorate in other ways. Regardless of the length of time that the samples are stored, the maximum length of time that the prepared analytical samples are stored must naturally have been established in the corresponding method of analysis and verified in the analytical method validation.

Reference documents

For example: validation master plan, SOPs, development reports, reports on pilot batches, IQ/OQ/PQ reports.

Departments involved in the validation and responsible members of staff

Qualification status of the facilities and equipment used

A prerequisite for carrying out validation is that the qualification of the facilities, machines, utilities devices and computer programs is complete (see chapter 7.E.4 Prerequisites for carrying out a validation project). The actual status of these activities must be documented in the validation protocol.

Validation of analytical methods

Valid analytical methods must be available for the testing defined in the validation protocol. The validation status of the analytical methods must be documented in the validation protocol.

Schedule

Manufacturing a validation batch takes considerably more time than routine production. It is important to take this aspect into consideration when scheduling production and to draw up a detailed schedule to prevent misunderstandings. The three validation batches must be scheduled for manufacture on different days, with different staff or during different shifts, if possible, to make this as realistic as possible.

Techniques for interpretation of the test results

If the validation master plan does not detail this sufficiently, the validation protocol must establish which statistical tools may be used to interpret the trial. This will prevent any critics from subsequently implying that the result of the validation was influenced by a choice of questionable statistical techniques.

The PIC/S document PI 006 also names process capability investigation as a tool to be used in the framework of process validation..

Procedure in the event of changes to the validation protocol

Unless it is detailed in the validation master plan, this procedure must be established in case it is realised during production that changes are required to the validation protocol. Changes may be documented, justified and authorised by the responsible person in the form of an appendix to the validation protocol. It is particularly important that all staff involved are informed of the changes

Authorisation of the validation protocol

To prevent misunderstandings, all departments involved must check the validation protocol before it is authorised by the responsible person(s) named in the Validation Master Plan, e.g. Head of Production, Head of Quality Assurance and/or Qualified Person (figure 4).

2 Content of a validation report

The trial data results and experiences are summarised, analysed and evaluated in the validation report. Problems and deviations during production and testing are also adressed in the validation report. For validation batches that are produced at large intervals from each other and for the concurrent validation, an individual report is drawn up for each validation batch.

A validation report contains the following details:

• Batch production records with the actual settings of the facilities and machines and the actual process parameters (hand-written entries in the batch production records or machine printouts). The raw data is checked for completeness and plausibility and evaluated using the techniques specified in the validation protocol, if necessary, before it is compared with the requirements in the validation protocol.
• Results of the in-process controls, the final product quality controls and the additional validation experiments, including the results of unsuccessful trials. These results are also checked for completeness and plausibility and interpreted using the techniques specified in the validation protocol, if necessary, before it is compared with the requirements in the validation protocol.
• Raw data from the analytical investigations or reference to corresponding documents:
  If the raw data from the analytical investigations is not included in (or appended to) the validation report, the document must state where this raw data can be found.
• Deviations from the validation protocol, additional investigations (with rationale), special observations:
  When the production and test protocol are interpreted, deviations from the requirements in the validation protocol must be justified and explained. The effect on the validation result must be discussed.
• Assessment of the results in comparison to the requirements:
  Once the data and information has been interpreted, it is compared with the requirements in the validation protocol and the results are assessed. There should be a discussion and justification of how the result of the validation can be transferred to the routine conditions.
• Determination of follow-up measures
  The results of the validation may indicate the need to modify the manufacturing instruction for the routine process so that, for example, additional control measures (IPC) are defined or tighter requirements for machine settings or narrower limits for the process parameters are set. However, the investigations often show that the observed process cannot yet be designated as validated. Then, further validation measures must be completed, e.g. more precise investigations of individual processing steps. In the worst cases, the process even has to be optimised before the validation study is repeated. Such follow-up measures must be suggested in the validation report.
• Authorisation of the validation report
  If the results of the validation batches prove that the investigated process is reproducible and results in a product with the specified quality, the validation report must be authorised by the responsible person(s) named in the Validation Master Plan, e.g. Head of Production, Head of Quality Assurance and/or Qualified Person.

2.1 How to deal with deviations from the requirements
in the validation protocol

In practice, problems often occur when validation batches are being produced and the requirements cannot be fulfilled. These problems and deviations must be documented in the validation report, together with the measures implemented to trace and rectify the problems/deviations.

Possible causes may be:

• A process that has been inadequately developed/not optimised (this includes correlations that are not known, for example between granulation time and homogeneity, particle size distribution or bulk weight, or between particle size distribution and homogeneity or disintegration time)
• Insufficient data material on which the machine settings and process parameters were specified
• Unqualified facilities or non-calibrated measuring points

Unexpectedly large variability in the process parameters is often observed during production and this may lead to thinking about extension of the specified limits. However, this sort of extension must be very well justified and the data must prove that it has no effect on the quality of the final product. This can result in a crisis in the line of argumentation, especially if the validation protocol had stated that the corresponding process parameter was "critical" and that a small modification would therefore have a great effect on the product quality.

Deviations from the validation plan which call the validation study into question require follow-up measures (see chapter 2 Content of a validation report).

2.2 Archiving of the validation documents

The documentation drawn up as part of the process validation (plans, raw data, reports) should be available at the manufacturing site (for packaging validation, at the packaging site) so that it is accessible for any inspections by authorities (cf. EMEA "Note for Guidance on Process Validation", Paragraph 4).

The length of time that the validation documentation is to be archived depends on how long the corresponding product is in circulation. The validation documents must be kept for at least as long as the batch documentation for the last ever batch produced. In Europe, this is at least 1 year after the expiration date or 5 years after batch certification by the Qualified Person, whichever is the longer period, in accordance with directive 2001/ 94/ EC. If clinical studies are still being carried out with the product at that time, the documentation must be archived for at least 5 years after the completion or formal discontinuation of the last clinical trial in which the batch was used.