During the past 20 years since the introduction of process validation encompassing all aspects of production, the principles and implementation of the exercises have evolved. However, the basic tenets have remained the same. You begin by developing your Phase III process, which will become the forerunner of your commercial process, albeit at a larger scale in most cases. While this development progresses linearly, you are manufacturing your Phase III clinical material to fuel your pivotal trials. Given sufficient resources, you attempt to optimize the unit operations for economic reasons as well as reliability, efficiency and compliance ones. Often this work is not fully complete or comprehensive, failing to explore all parameters and inputs. Catastrophes occur even with the best laid plan, resulting in one’s having to go back to the drawing board or make major surgery to the process. If I am describing this process pessimistically, it is because I have seen this often since the inception of process validation.

Armed with all the information at hand — and it is often incomplete — we transfer the technology to the commercial operations and pray that our smaller scale operations were optimized in a manner that is representative of commercial scale and that our manufacturing staff is skilled at dealing with the unexpected. After several false starts, we are convinced that the operations are working and we proceed into the process validation phase. Each unit operation is validated in that we perform it three times, hoping that the outcome is as expected against a protocol with preapproved acceptance criteria. Once all the unit operations are validated, we run the complete process to yield product. We perform it again the magic three times, and if we get three consecutive successes we declare victory, write the report and breathe a sigh of relief. The reports are filed, never to be reviewed again.

More often than not, we do not get three consecutive successful runs because things go wrong. We then examine why the first, second and fourth worked but the third failed, trying to convince ourselves that there were factors that did not matter to the demonstration. Sometimes we do convince ourselves and then hope the inspectors will agree. Sometimes they do. These are submitted to the dossier and made available for inspection.

If we have been lucky, the inspection goes well, the filing is approved and we are off to commercialization, making product in our approved plant. Looking back over the years and examining all the work that went into the process validation activity, I am not always convinced that the effort is a good investment, except to say it did lead to our commercial approval of the product. Can we honestly say that, even if the first three batches were successful, we are now confident that batches four through six will also be successful? I think not. As commercial operations move along, there is a mad rush to fix problems that crop up.

For already approved products, where we have to validate the process retrospectively, the payback is less except that we are allowed to continue to sell the product. At least with these older products you do have a track record of how good (or bad) and reliable (or unreliable) the processes were, hopefully.

So, a little more than two years ago, the FDA released its new guidance, which had been awaited with bated breath. (The EMA followed suit a year later.) Would the agencies unveil some forward thinking that might add value to the activity, or would they not? I am happy to announce that the new guidances, while not being perfect, are a major improvement on the old paradigms.

The new guidances describe the rationale and reasoning for process validation in a manner that leads us to believe that the activities may add value commensurate with the effort we put in. They reason that the act of process validation is not a single point in time that can then be forgotten and put aside, but rather describes the activities to be dependent and linked very closely with pre-existing work. They also note that the activity, when completed, becomes the basis for continuously monitoring the process to verify that it is still working the way it is designed to and has not changed unintentionally. The concept of lifecycle is described and it is applicable. Although they describe how a Quality by Design (QbD) development program can add value to the process validation activity, it is not required; there are other ways to get to the desired state. So how do the new paradigms work in real life? There are three components to consider.

Component 1
We must have a robust process development program that considers and expects that the product is clearly defined as to what the patient needs. This is the so-called “Product Profile.” We now link this to Critical Quality Attributes (CQA). These CQAs are the outputs of each unit operation and the final drug product (DP) that lead to the realization of the correct Product Profile. How do we assure that we attain the CQA for the unit operations and DP? We research each unit operation and optimize so that we understand the impact of all the inputs (read: raw materials and components). Some will have limited impact while others will be more significant. We also determine which parameters and controls govern the output. That is, those parameters that must be tightly controlled and those that are less critical, requiring less control (See figure 1). Remember that the output of one unit operation becomes the input for the next. These relationships are worked out by the intelligent use of design of experiments and Quality Risk Management principles to define acceptable ranges for all critical parameters and are described in ICH Q8, Q9 & Q11. In the end, we focus on the important and ignore the unimportant. With this robust development program, we define the Design Space (See figure 2), an area where we can operate and stay in control with the assurance that that the outputs will always lead to product meeting the CQAs and Product Profile. All of this work is performed at small scale with a valid scaled-down model that is representative of the final commercial- scale operation. The premise is that results at this scale are directly related to those that will occur at large scale. Thus the validity of the scaled down model is paramount. This is the ideal QbD approach, but there are alternatives.

Component 2
Armed with this development knowledge, we design our process validation demonstration. The purpose is to confirm that the data generated at small scale is robust at large scale. We recognize that for complex processes — and biopharmaceuticals all fall into this category as well as some small molecular weight drugs — three runs is just not enough to assure this activity. How many are needed? I routinely use about 20 runs to demonstrate process variability. That is not based on statistical analysis but rather experience; after 20 runs you perform an analysis to determine process variability and compare back to the small-scale work to either confirm or adjust your conclusions on the process. This could result in tightening of the design space, if needed. Do we wait for the 20 runs and the analysis to submit the dossier? Obviously not. Rather, we submit with a commitment to perform the runs and analyze the situation, perhaps submitting a report as a post-approval commitment.

Component 3
Armed with the results of the validation, we proceed in manufacturing and periodically review the trends of the operations, focusing on the critical inputs and control points and parameters, and compare back to the development work and the initial demonstration. We are focused on assuring that the process still stays in control and performs comparably to how we predicted in development and the demonstration (validation) phase. Certain adjustments may be needed to bring back into control, but these will be within the design space. Thus, they can be performed via internal change control rather than submissions to the regulatory bodies. As we gain experience, we may choose to alter the routine operating ranges (but still remaining within the design space) for a variety of reasons, including economics, ease of operation, efficiency or reliability. Likewise, as new technology is made available or altered control strategies are contemplated, we use this body of data to cautiously introduce these changes via change control. After the change, we monitor to assure our assumptions are still valid. Thus, we embark on a program of continuous improvement as defined in ICH Q10. The linkage of these various elements is described in figure 3.

So how are companies managing the change to the strategy for process validation? Different companies approach it differently. Larger pharma and biopharma firms have embraced the concept and created departments with the sole focus of implementing these new guidances. Since they have the resources and the deep and broad pipelines, they have products at all stages, so it makes sense. These departments act as catalysts, linking the process development and commercial operations to assure smooth flow of information and knowledge.

Smaller companies are not so endowed with resources, with some literally operating hand to mouth to keep programs going. I have been approached by a large number of potential clients asking what do they need to do and when to implement, if at all. I advise them based on the following major factors.

The first is the business model they espouse. Some companies plan to be the next big player in the field — the next big pharma or biopharma. These are clearly going to implement all aspects of the new paradigm, but not necessarily all at once. Other companies have no desire to become a big player in the industry. Their business model is to create business leads, demonstrate the molecule and either sell the product or the company to the highest bidder. This package for sale could include animal safety and efficacy data only, as well as Phase I or even early Phase II data. In all these cases, they will not need to worry about process validation and the changes.

Some companies have a business model whereby they plan to develop the molecule through Phase II and as they move to Phase III either go it alone or partner with a big player. In this situation, the new paradigm becomes important. But since they are partnering with one of the big players, they should utilize the partner’s expertise in implementing the new paradigm. That does not mean they do nothing. They should establish the key elements of development either alone or in partnership with the co-developer. Having worked both for big pharma and with big pharma while at a small development company, I know that the big players tend to want to do it their way and value less what other companies do. So one should not spend money on what will be repeated or viewed by the partner as “not our way” and less valuable. It may not have a good return on investment for you. Save it for the next molecule coming down the pipeline, when you go it alone.

If you are heavily virtual, outsourcing process and product development, clinical and commercial manufacturing operations, does that change the landscape?  Not really. The same factors above apply but you now have a second or third player involved. You must utilize their expertise as appropriate, keeping in mind that all this might be as new to the CMO as it is to you. It will require more coordination but it is not impossible.

I have indicated that companies in early phase development need not worry about the new process validation guidance because commercial processes are not locked down until the Phase III. While that is true processes are evolving during earlier phases, as are the analytical methods used to assess them, and as such it is not appropriate to validate an evolving process. However, a word of caution is pertinent. In early phase development, our knowledge of the molecule is still fragmentary — including its safety — and we have to assure safety of patients to the best of our ability. So it is appropriate to build in layers of validation for these elements. For oral dosage forms, one must build in assurance of purity and protection against cross contamination. For sterile products, the safety goes a further step, assuring the sterilization step and aseptic processing (as appropriate) as well as cell line safety, which must be incorporated into the strategy for biotech products. This often involves elements of process validation. This is unaffected by the new guidances, which are silent in this area.

So the take-home message is this: Understand your business model and be cautious of investing your money and resources unless the return on investment is clear.

The second factor is your portfolio and where your molecules are in development. If you intend to commercialize the product (assuming it works), then position where the activities need to occur and implement that part of the paradigm moving forward as the molecule progresses. The next molecule coming along will pick up the earlier phases you missed.

If you are ready to embark on the development of your Phase III process, then maybe the time to upgrade your process development to a QbD or QbD-like strategy is appropriate. This builds the foundation to move to the next phase.

If you are already in Phase III production and most of your process development is done, you will have little time to really make a paradigm shift in this area. In this case, take what you have and implement the 20-batch demonstration strategy for your submission. That can work if you have a robust development program. But before launching that I advise a review with your regulatory bodies to assure they agree with you. You may be forced into the old paradigm. With whatever strategy you do undertake as you head toward commercial approval, implement a continuous verification/continuous improvement strategy as you gain more experience. But also plan ahead for the next molecule to back-fill those elements not yet implemented.

If you have a commercial product recently approved, there is nothing stopping you from implementing continuous verification and continuous improvement, although you will probably not have a broad design space to operate within. That will have to wait for the next molecule or a post-approval supplement covering aspects of the design space for key unit operations.

In all cases, approach it cautiously and recognize any work you do must fit into the timeline you have been working on and communicated to the market. If you cannot shoehorn it into the present timeline because of resources or other factors, consider not implementing on this molecule. For small companies, a delay in timeline (just to get a slightly better, more progressive submission) may damage shareholder perception of how things are going. Your goal should be to stabilize and add value to your company, not to add unnecessary risk to your success.

Table 1 illustrates three case studies of anonymous companies that are different and the strategies they have implemented.
Understand your pipeline, where the molecules are and your perception of their probability of success. Cautiously choose to implement when it is right for a particular molecule, depending on your business model. It does not all have to happen at once and can be implemented piecemeal when and if the time arises. 




























This article is based on a presentation made by the author at a recent IBC Conference at Huntington Beach, CA: Biopharmaceutical Development and Production 2013: Methods & Molecules.

References

1. Guideline on Process Validation (2012) – EMA/CHMP/CVMP/QWP/70278/2012-Rev1
2. FDA Guidance for Industry (2011) Process Validation: General Principles and Practices Jan 2011
3. ICH Q8 2009 Pharmaceutical Development Q8 (R2)
4. ICH Q9 2005 Quality Risk Management Q9
5. ICH Q10 2008 Pharmaceutical Quality Systems Q10
6. ICH Q11 2012 Development and Manufacture of Drug Substances (Chemical Entities and Biotechnological/Biological Entities) Q11

---

Peter H. Calcott, Ph.D. is president of Calcott Consulting LLC. He can be reached at peterc@calcott-consulting.com.