“Even though we mention this concept a lot – real-time release -- one of the things I’ve heard from the FDA is that they actually don’t like us to emphasize real-time release,” Ruben Lozano, principal scientist at Bristol-Myers Squibb, said at a roundtable discussion on QbD (Quality by Design) and PAT (Process Analytical Technology) in pharmaceutical manufacturing at Interphex on Tuesday.
“They would prefer us to emphasize more the fact that we’re improving our understanding of the products and the process rather than real-time release being our main objective. It’s nice to have real-time release as a natural outcome of your control strategies, but I think the most important aspects of QbD is applying your control strategies and gaining that improved understanding of your product and process,” Lozano added.
The FDA (Food and Drug Administration) defines real-time release testing as the ability to evaluate and ensure the quality of in-process or a final product based on process data. That reliance on in-process testing can often reduce or eliminate end-product testing. And although the FDA acknowledges that QbD isn’t necessarily correlated to real-time release testing, “it would be difficult to justify [real-time release testing] without a science and risk-based approach,” the agency says.
Lozano said he heard the FDA’s stance at a conference in January and that the speaker said she didn’t think real-time release testing should be a company’s “main objective.”
“They don’t want real-time release to become a check box exercise… If I do these five things, I get to do real-time release. That’s so antithetical to the approach of process understanding. If you create this as a hurdle, it defeats the purpose of” real-time release, Carl Anderson, PhD., associate professor at Duquesne University’s Mylan School of Pharmacy, added.
The FDA’s stance on real-time release seems to contrast with the EMA’s (European Medicines Agency), as the European agency last year released guidelines extending the use of real-time release testing to a broader spectrum of biopharma manufacturers.
Calibration and Variability
The topics of calibration and variability were also discussed as points of possible contention for developing compounds, especially in terms of ensuring variability is real and not the result of the calibration.
“One of the things that doesn’t get enough play in the PAT realm in general are the ancillary statistics that help us understand how good the calibration is at any point in its life cycle,” Anderson said. “That part of it is not well enough publicized. I think the regulatory authorities are just now getting on board with that and I think they will be demanding it in the very near future.”
As calibration drifts, there are a number of metrics “to help us understand the quality of a prediction is declining or moving away from what it was in the original calibration,” Anderson added.