By Paul Thomas, Senior Editor
At the beginning of his talk at last week’s Bioprocess Development & Production Week event in San Diego, Amit Banerjee, a research fellow in global biologics with Pfizer, posed the following question: Is the QbD approach aligned with FDA’s new Process Validation guidance?
In other words, Banerjee mused, “Are we trying to force a square peg into a round hole?” It’s something that the industry will find out in time, he suggested.
But during his presentation, Banerjee went on to suggest that, indeed, QbD is aligned with the Process Validation guidance, and vice versa. “QbD provides an opportunity for a systematic process design,” he said. (“Process Design” of course is the first of three stages in FDA’s PV guidance.) In addition, he said, the risk assessment that is a fundamental pillar of QbD is important groundwork for later PV work in that it:
• Helps prioritize experiments;
• Identifies the relationship of quality attributes to process parameters;
• Helps to develop a process control/monitoring strategy.
So in essence Banerjee answered his own question by saying that a Quality by Design approach is aligned with Process Validation and, if executed properly, will reduce the challenge of Process Validation.
Process Validation: You’ve Come a Long Way, Baby
The next speaker, E.J. Brandreth, VP of Quality and Regulatory Affairs for Althea Technologies, suggested the same. Between FDA’s new Process Validation guidelines and the work that organizations such as PDA have been doing, “we’ve come a long way,” he said.
Whereas 1987’s Guideline on General Principles of Process Validation—which encouraged “Three runs and you’re done” said Brandreth—was vague and limited, there are now multiple documents that seek to clarify and specify process validation. These include the ICH documents, two from PDA (Technical Reports <42> and <14>), and even the recent EMA PV concept paper.
“Everything dovetails,” he believes. “You’re going to look at your process, break it down, and validate those things that are important.”
Current process validation, he noted:
• Emphasizes that “Quality, safety, and efficacy are designed or built into the product”;
• Is science-based;
• Obviates the need for worst-case approaches at large-scale;
• Encourages matrix approaches (counter to the 1987 guidance);
• Encourages a team approach: “The guidance strongly recommends input in the validation process from a wide range of disciplines, as well as the full support of senior management”;
• Emphasizes the importance of training: “You need to come up with some reason why the deviation didn’t impact the Process Validation,” he said. “It’s a good practice to have a record of people trained on the protocol.”
Banerjee’s Pfizer colleague Carol Kirchhoff, Senior Principal Scientist for BioTherapeutics Pharmaceutical Sciences, later spelled out her views on the broad overlap between the PV guidance and the key ICH documents. Covered in both guidelines, she noted, are an emphasis on:
• Models to gain process knowledge and understanding;
• Experimentation (especially Design of Experiments);
• Risk assessments;
• Control strategy;
• Facility design and qualification of utilities and equipment;
• Process performance qualification runs ;
• Documentation; and
• Continuous Process Verification (CPV).
“The QbD concepts of ICH Q8, 9, and 10 mesh very well with the FDA Process Validation guidance,” she said.
Validation Runs: The 800-Pound Gorilla?
If there is a “square peg, round hole” issue, it may be the “three runs and you’re done” issue that Brandreth spoke of. Brandreth addressed the contentious topic. PDA, he noted, recommends a minimum of three consecutive validation lots. The draft European guidelines are leaning towards approximately three to five runs, he noted.
In its new guidance, FDA does not specify, leaving it up to industry to decide. The Agency’s key criteria:
• Do I have confidence in my manufacturing process?
• How do I demonstrate that my process works as intended?
• How do I know that my process remains in control?
This is a topic that our contributing editor Ali Afnan has addressed: “An entire industry has grown around the ‘three-batch practice.’ Would any conscientious reader agree that three consecutive batches, even after a large number of engineering batches, satisfy either the spirit or the letter of the law? . . . So, after 24 years, the Agency’s thinking on process validation has changed. What does this mean? The routine of ‘three batches and you are done’ is no longer valid; truly ‘statistically significant’ sampling is now a reality.”
The beauty of the new Process Validation guidance, according to Afnan, is that it’s far less prescriptive than it could have been. But that leaves open the three-batch quandary.
“The biggest challenge facing organizations attempting to bridge the classical paradigm of ‘three batches and we’re done’ is in understanding how the new Process Validation stages work together to build the argument for process predictability,” consultant Bikash Chatterjee has written.
No sooner had Brendreth finished his talk in San Diego than did Peter Calcott, president of Calcott Consulting, rise from the audience and comment on what he called the “800-pound gorilla” still looming in discussions of process validation: “Thou shalt do three runs.”
“Three runs doesn’t really tell you much,” said Calcott. So how many then? “Let’s say 20 . . . that might frighten people but it’s small enough to be manageable,” he said. But with QbD, does the manufacturer “have the data to loosen the licensure requirements, to make just a commitment to evaluate 20 lots and send in the formal data later?”
Calcott turned to the rest of the audience: “Has anyone been successful in saying, ‘We don’t need three runs because we have this wonderful [QbD] dossier?’”
Brandreth responded: “They [FDA] want to see both. They want to see the dossier but also proof that you can run it in large-scale. Three of course is not statistically significant. On the other hand, it’s challenging to do three successive runs in a biological process without significant deviations . . . Three times at least is a trend, goes the old saying.”
Calcott summarized: “Then we know what the desired space (or whatever you want to call it) is, and we should be able to get a point of zero (runs), but we’re not there yet.”
Another questioner from the audience brought the challenge of validation runs back into a realistic perspective: “Agencies have different expectations in different regions,” he said. “In general, the more the merrier in terms of the different number of runs.”
“The best idea is to do at least the basic three runs in your commercial facility,” Brandreth agreed.
Proof in Pudding
Getting back to the square-peg-round-hole question, the one other issue that was raised was the failure of FDA’s Process Validation guidance to mention the words design space. “That, to me, is huge,” said Pfizer’s Kirchhoff.
“Process Validation is referring back to a lot of ICH guidances, though not necessarily just the QbD ones [i.e, Q8, Q9, Q10],” Kirchhoff added. “So Process Validation and QbD do mesh overall, but not 100%.”
“The proof will be in the pudding,” said Pfizer’s Banerjee. “I’m waiting for someone to take their experience with Process Validation and say whether or not it aligned with QbD.”
image courtesy mediaexplored.blogspot.com
- FDA Publishes Long-Awaited Process Validation Guidance
- The Agency's new document provides a much-needed update to its 1987 guidelines....
- Will the New Process Validation Guidance Make a Difference?
- Yes, says expert Jim Agalloco, but the guidance "falls heaviest on smaller firms that are often ill-equipped to to the...
- Final Process Validation Guidance: Any Day Now?
- Jim Agalloco and Jim Akers consider why final Process Validation guidance is taking FDA so long, and how QbD may...
- Process Validation on Your Mind
- A list and look at some noteworthy commentary regarding FDA's recently released final process validation guidance....
- Process Validation: Can We Now Get Back to Basics?
- FDA’s new guidance on process validation gets back to the original language of the GMPs on the subjects of control...