Continuous process validation, expansion of validation activities to cover the entire lifecycle of a product, constant process improvement and innovation instead of fixed workflows and parameters — the concept paper of the EMA addresses many such topics, as does the corresponding FDA guideline. These topics raise many more issues. Proof of this was supplied by the lively discussion on the topic of "process validation" at the GMP Symposium. Please read the following summary of the discussion between the participants and the independent GMP consultant Dr. Christine Oechslein, GMP inspector Klaus Eichmüller and Thomas Peither.
To validate a process one time only then close the files forever – according to the FDA guideline that should no longer be possible. Process validation should cover the entire lifecycle, beginning with the product and process development (process design), continuing through the actual validation on a major industrial scale (process qualification) and proceeding all the way up to monitoring routine production (continued process verification, CPV). This may sound like a novel statement, but if we look at it with open eyes we will see that it is anything but altogether new, at least within Europe. For example, CPV has long since found its way into everyday pharmaceutical routine by way of the product quality review (PQR) that is required here. GMP Inspector Klaus Eichmüller: "The FDA certainly did not invent this principle. And yet American marketing is clearly better. A reasonable, continuously applied PQR with carefully selected data can cover all requirements for continuous process verification."
If you want to implement continuous validation, you have to make a fundamental distinction between whether to convert established processes or introduce new ones. If new developments are involved, a continuous validation approach is usually easier to put into reality (‘quality by design’). When dealing with existing products you typically have extensive data resources from in-process controls, deviations and change workflows. Dr. Christine Oechslein: "The real charm of continuous process verification lies in gaining knowledge from process data that accumulate anyway and using them to control and improve the processes. Colleagues from other sectors such as textiles or the automobile industry are astounded to hear that in the pharmaceutical industry, after only three batches the data are merely collected and there is no further data evaluation to speak of, or if so, it is only carried out at the end of the year. Data cemeteries of this nature should be a thing of the past, considering all the electronic possibilities available in today's world.
The product quality parameters of in-process controls, deviations and changes, etc. must be continuously reviewed and — above all — interpreted. The findings resulting from this should flow back into the process. Thomas Peither: "Learning about the process is not completed with three, four or five initial validation runs. Learning never stops, because much experience is gathered only in the course of the running process. That is why every batch produced is actually a validation batch. I receive data, I can evaluate them statistically and I can take them into consideration for the release. Failing to do so means missing an opportunity at every turn. Each single manufacturing record houses an enormous amount of knowledge and hence capital."
Even though many methodical workflows are already established in this part of the world, letting go of the classic approach and embracing an "enhanced approach" still causes practical problems. It is true that the enhanced approach essentially accelerates a marketing authorisation, since under certain circumstances fewer validation runs will suffice. However, this means that, afterwards, as experience with the process grows, it may become necessary to submit more change requests. For example, specification limits may have to be reset or parameters that were tightly controlled at first could prove to be non-critical after a while since they have no impact at all on the quality of the product. Therefore, many businesses are unable to reap the benefits of the "new" workflow, and even though they may have a choice between the traditional and the "enhanced" approaches, these businesses are inclined to select the conservative variant. The situation in the United States is similar. The industry has been dealing with this topic for a number of years, and by now many companies have developed first concepts. However, no real experience has been gathered in the USA either.
The experts agreed across the board that the question about the (economic) benefit of continuous process validation arises only at the beginning. After all, money spent at the start represents a lasting investment in the future. Consistent and prompt evaluation of all the data that I collect over the course of the entire lifecycle ultimately leads to better understanding of the process. In the future there will be a growing need to call in statisticians (able to think along pharmaceutical lines!) for assistance in evaluating these enormous data quantities. Thomas Peither: “The better I master my process, the sooner will I be able to reduce control steps. I can work effectively long-term and I will clearly benefit economically.”
The true motive for converting to continuous validation throughout the lifecycle should not lie all that heavily in efforts to fulfil requirements of the authorities. The aim should rather be to work towards becoming more familiar with the manufacturing processes. The experts shared the opinion that the FDA has correctly recognised that processes are often not adequately understood. Otherwise, there would not be so many investigation reports.
However, hopes that the new models would make it easier to obtain marketing authorisations have not yet become reality. But this is expected to change, says Thomas Peither: “Marketing authorisation will change when the dialogue between industry and the authorities improves.” Many participants complained that communication with regulatory and supervisory authorities within the FDA is less than ideal ─ to put it mildly. People with visionary concepts often find themselves facing inspectors who still have not necessarily absorbed the underlying philosophy. The experts concurred that this takes time.
Wishing that the authorities would better understand a process explains why the question of how many validation batches are actually needed is willingly volleyed back to the businesses. Klaus Eichmüller: “Considering the degree of heterogeneity of the processes, I cannot specify a fixed number of validation batches. The process know-how is concentrated with the manufacturer, the only one who can judge how many batches are required. Nonetheless, the manufacturer must be able to substantiate this assessment and support it with data.”
Towards the end of the discussion Dr. Christine Oechslein pointed out that the model described by the FDA harbours substantial “potential for misunderstanding and confusion.” The term “process validation” has been expanded to form a generic term and, as already mentioned earlier, it is intended to encompass the preceding developmental stage and the commercial production phase. The phase that was formerly referred to as “process validation” is now hidden, together with qualification of facilities, behind the term “process qualification.” Dr. Christine Oechslein: “New terms have been created and established terms given new meanings. The EMA, however, did not apply this American terminology in their concept paper!” Therefore, the experts warned against “obsequiously accepting” all concepts in Europe. On the other hand, for shipments to the USA or inspections by the FDA, the American terms should be known and familiar to all concerned. At the very least, a glossary should be compiled in which the terminology commonly used in the European linguistic territory is compared to that of the FDA.
Susanne Sailer, Maas & Peither AG, GMP Publishing