For quite a while, the new Process Validation Guideline by FDA had been expected, now it has been published as a draft. On a total of 20 pages and subdivided into 7 chapters, the FDA now describes their current thinking with regard to process validation. The chapters are subdivided into:
The introduction points out expressly that process validation is connected to a product life cycle. Thus, process validation includes development towards routine production and routine production itself. Furthermore, the guidance document intends to promote modern manufacturing principles, process improvement, innovations and sound science. It also refers to ICH Q8, 9 and 10 in connection with the life cycle concept.
The introduction then lists the products subject to the guideline: "human drugs", "veterinary drugs", "biological and biotechnology products", "finished products", "combination products" (drug and medical device), but also "Active Pharmaceutical Ingredients" (with reference to ICH Q7a). As a countermove, exclusions are listed as well, e. g. medical devices and dietary supplements. It is also pointed out that the guideline does not give any information on the documentation necessary to apply for a marketing authorisation. The guideline does not apply to the validation of "automated process control systems".
The history briefly deals with the FDA Guideline on Process Validation of 1987, the basic principles of which have been taken up again in the new draft. Interestingly, the text expressly mentions the GHTF Guideline on Process Validation relevant to medical devices as being likewise useful for drug manufacturing. The current guidance document is based on experience values gathered since 1987 and also on the FDA Initiative cGMPs for the 21st century - a risk-based approach. Here, the text contains another reference to modern manufacturing techniques and to "risk management" and "quality management tools" and "concepts" - however, without going into detail.
What is new is the definition of process validation as "the collection and evaluation of data, from the process design stage through production, which establishes scientific evidence that a process is capable of consistently delivering quality products". Thus, process validation is now split up into 3 stages:
The text states expressly that in practice these 3 stages might overlap. With emphasis, it urges manufacturers to prove with a high degree of assurance that the product can be manufactured according to the quality attributes before a batch is placed on the market. For this purpose, data from laboratory-, scale-up and industrial scale are meant to be used. The data are explicitly meant to cover conditions involving a great risk of process variation.
For this reason, the manufacturer should
Again, the text points out expressly that qualification activities lacking the basis of a sound process understanding will not lead to an accordingly qualitatively safe product. The chapter closes by pointing out that the process must be maintained during routine operation. This includes materials, equipment, the environment, personnel and changes in the manufacturing procedures.
III Statutory and Regulatory Requirements for Process Validation
In this chapter, the FDA refers to paragraphs in 21 Code of Federal Regulation 210/211 that also have to be applied with regard to validation (and qualification of equipment): 211.100(a), 211.110(a)(b), 211.160(b)(3), 211.165 (a)(c)(d), 211.180(e), 211.42, 211.63, 211,68.
This is the central chapter of the guidance. At the beginning, Good Project Management and good archiving are pointed out as effective and efficient means for the product life cycle. A team approach to process validation is mentioned as well, with a statistician listed as possible team member. Furthermore, the text reminds of the fact that the full support of senior management is necessary. All studies conducted within the framework of process validation should be documented accordingly and conducted on the basis of sound scientific principles.
Then the recommended activities in the 3 stages are dealt with emphatically.
Stage 1 - Process Design
a) Building and Capturing Process Knowledge and Understanding: In this stage 1, the manufacturing process is meant to be defined, which will then be reflected in the manufacturing and testing documentation. Also in view of Q10, it is expressly pointed out that earlier development stages do not have to be conducted under cGMP. However, here, too, the basis should be sound scientific methods and principles, including Good Documentation Practice. It is considered to be no regulatory expectation that the process be developed and tested until it fails, but the combination of conditions involving a high process risk should be known. In order to achieve this level of process understanding, among other things the implementation of design of experiments in connection with risk analysis tools is recommended. However, other methods, like classical laboratory tests, are also considered as acceptable. What is considered to be essential is the adequate documentation of the process understanding based on rationales, above all in view of the life cycle.
b) Establishing a Strategy for Process Control : Process knowledge and understanding are considered to be the basis for process control. Apart from in-process controls, the text mentions the possible use of Process Analytical Technologies (PAT) and quotes the corresponding PAT guideline.
Stage 2 - Process Qualification
It is meant to prove that the process design is suitable for reproducibly manufacturing commercial batches. This stage has 2 elements: On the one hand the qualification activities regarding premises and equipment, on the other hand performance qualification (PQ). Strictly speaking, this stage encompasses those activities that are currently summarised under process validation: On the basis of qualified equipment, it is then demonstrated that the process can create a product in conformity with the specifications.
The text deals with the constituents of the qualification activities that are the prerequisites for PQ. Without mentioning the terms DQ, IQ, OQ, these activities are described as constituents of the qualification. The necessary documentation on qualification, too, is dealt with. The description of qualification activities are possible as individual plans or as part of a project plan. The possibility to integrate risk management in order to determine priorities and scope of performance and documentation is only mentioned as a "can" option. The contents of the plan should be:
1. Description of the tests
2. Acceptance criteria
3. a schedule
5. Information on the documentation and release of the qualification results
6. Data on the change control procedure
The results are meant to be summarised in a report making reference to the acceptance criteria in the plan. Both the plan and the report are meant to be reviewed and released by the quality control unit.
A specific subchapter is dedicated to the performance qualification approach, the second element of stage 2 in process validation. The PQ combines the qualified premises and equipment as well as the trained personnel with the commercial manufacturing process. What could in principle be regarded as a definition of PQ is the document's statement: "A successful PQ will confirm the process design and demonstrate that the commercial manufacturing process performs as expected". The PQ is considered to be an important milestone within the product life cycle. Its completion is a prerequisite for marketing, and the decision for marketing the product should be based on data gained from the commercial manufacturing scale, if necessary supported by laboratory and scale-up studies.
The guidance points out expressly that a sound science should serve as approach to PQ. This goes as far as the FDA requiring in the guidance: "We strongly recommend firms employ objective measures (e.g. , statistical metrics), wherever feasible and meaningful to achieve adequate assurance". Insofar, according to the guidance, the scope of PQ regarding sampling, additional tests is more comprehensive than in normal production. Finally, the text points to very specific aspects of biotechnological manufacture and to PAT implementations within the framework of PQ, which take account of a different PQ approach.
The PQ plan (here called "protocol") is meant to discuss the following points:
1. The manufacturing conditions including the parameters to be set, the process limits and the use of raw material
2. Data that are collected and how they are evaluated
3. Tests to be performed (IPC, release, characterisation) and acceptance criteria for each significant process step
4. A detailed sampling plan (where, how much, how often) based on a statistical basis (within a batch and between batches) equalling a risk analysis
5. Criteria providing a rationale for the question whether the process constantly produces a qualitative product. Among them are a description of the statistical methods used for data analysis (referring to variabilities within a batch, but also between batches) and the description of the handling of deviations
6. If need be, qualification aspects of the premises and equipment, training certificates and a verification of the materials used (raw materials and primary packaging materials)
7. Validation status of analytical methods used for measurements in process, for in-process tests and tests on the finished product
8. Review and release by the corresponding departments and the quality unit
An individual item regulates the performance of the PQ tests and reporting: Not new, but still mentioned in the text is the requirement not to carry out the activities until the protocols have been released. Changes to the protocols should be evaluated correspondingly and released by the departments concerned and by the quality unit. The PQ batches are meant to be performed by production staff under normal conditions. This puts and end to discussions about "worst case" conditions within the framework of the validation runs.
Required components of the PQ report:
1. Discussion of the results with cross references to the PQ protocol
2. A summary and evaluation of the data according to the specifications in the protocol
3. Evaluation of all unexpected observations and additional data compared to the protocol
4. Summary and discussion of all "manufacturing nonconformances", like e.g. deviations
5. Detailed description of corrective actions or changes with regard to procedures and controls that have already been laid down
6. Clear indication of a result if the data show that the process is in compliance with the specifications in the protocol and that the process is in a sufficient state of control
7. All necessary reviews and releases by the departments and the quality unit
Stage 3 - Continued Process Verification
During this stage 3, the objective is to keep up the validated state of the process also in routine production. For this the manufacturer is required to establish a system detecting unplanned process variations. Shifts are meant to be evaluated accordingly so that the process does not get out of control. There is a direct reference to 21 CFR 211.180(e) in order to support this ongoing programme. The data must be statistically trended, and the analysis be done by a trained person. The text recommends explicitly that a statistician or at least an employee trained in statistical techniques works out the sampling plans and carries out the evaluation of the data with regard to process stability and process capability. The evaluation and the trending should be done according to SOPs. These evaluations are meant to be reviewed by the quality unit in order to detect changes in the process (alert limits) at an early stage and to be able to implement process improvements. Also with regard to unexpected process changes that can also occur in a well developed process, the guidance recommends "that the manufacturer use quantitative, statistical methods whenever feasible" in order to identify and characterise them and investigate the root cause. Here, too, variations within a batch and between batches are addressed explicitly. At the beginning of routine production, the guidance recommends the same scope of monitoring activities and samples as in the process qualification stage until enough data have been collected to allow a - statistically secured - adjustment of this scope.
Data from complaints, OOS results, deviations etc. can also give hints regarding process variability. Employees in the production line and in quality assurance should be encouraged to give feedback on the process performance. Operator errors should also be tracked in order to check if training measures are appropriate. The text explicitly recommends regular meetings between quality assurance and production in order to evaluate the above data and to discuss possible trends and drifts with the corresponding correction and follow-up measures.
The above results can then contribute to process improvements. However, the guidance points out that changes may only be implemented in a structured way and with the final approval by the quality assurance. Additional measures regarding process design (stage 1) and process qualification activities (stage 2) might become necessary. Another topic that is rated very important by the guidance is that of maintenance (including calibration). The corresponding maintenance and calibration cycles should then be performed based on the gathered data.
V Concurrent Release of Performance Qualification Batches
In rare cases concurrent release is also possible, i.e. a release of the product before the entire PQ protocol has been completed. Possible cases include orphan drugs (limited demand for the product) and radiopharmaceuticals. Yet, the guidance recommends the manufacturer to contact the FDA before implementing concurrent release. Batches marketed within the framework of concurrent release should be traced back very closely in order to be able to take immediate action in case of complaints (root cause). Furthermore, each batch manufactured under concurrent release should immediately be included in a stability programme.
Documentation is considered very important during each stage of the process validation life cycle. On the basis of the process design (stage 1), the guidance recommends to draw up process flow charts for the full-scale process. Apart from that, the text refers to 21 CFR 211.22 and 211.100.
VII Analytical Methodology
While analytical methods do not have to be validated in the (early) development stages, but have to be scientifically sound, analytical methods of the clinical phases 2 and 3 are subject to cGMP.
Conclusion: It does not really surprise that the new guidance does not mention a fixed number of validation runs proving process validity. This became already more or less clear in the Compliance Guide on Validation published in 2004. The new guidance relies on a 3-stage life cycle model. The new catchword is "process understanding". New definitions for process validation and performance qualification show the strong connection to "scientific sound". Strong emphasis is also placed on the use of statistical methods. "Modern" methods from the world of six sigma, like DoE, process capability indexes (Cpk) and statistical process control are mentioned directly or indirectly. What does surprise is the non-mention of the qualification stages DQ, IQ, OQ, even though the activities that are usually subsumed under these terms are expressly addressed in the guidance as a basis for PQ. The PQ is now the key element of the process validation life cycle and is meant to be carried out under normal conditions. Thus, worst-case considerations with industrial scale batch sizes are excluded. Within the framework of continued process verification, apart from the topic of trending, maintenance is rated highly. Both revalidation and retrospective validation are not mentioned any longer. It will be interesting to see how industry reacts to this draft. Especially the discontinuation of the "magic 3" - even though long anticipated - will probably (have to?) lead to new rationales in order to prove a validation. The new definition of PQ might lead to irritations. Up till now, PQ was often seen as being primarily related to equipment. It remains to be seen in how far statistics will find their way into process validation.
Here you can find the draft for the new Guidance for Industry Process Validation: General Principles and Practices.
On behalf of the European Compliance Academy (ECA)