FDA Emphasizes Regulatory Flexibility In Biosimilar Reviews
By Anna Rose Welch, Editorial & Community Director, Advancing RNA
Prior to attending the GPhA Biosimilars Council Leading on Biosimilars conference, I read a study published in the British Journal of Clinical Pharmacology (BJCP). This study, which I urge you to read in full for a clearer glimpse of the data, reviewed the clinical trial submissions for the 21 approved biosimilars in the EU. The researchers examined the European Medicines Agency’s (EMA’s) European public assessment reports (EPARs) for each of the 21 biosimilars and compared the number of patients recruited; the number of trials carried out; trial designs; and the endpoints and equivalence margins for pharmacokinetic (PK) and pharmacodynamic (PD) and Phase 3 trials. This study concluded:
Interestingly, even for biosimilars with the same reference product, the endpoints, statistical model and equivalence margins, and also the sample size, were not in all cases comparable. This shows that there seems to be some flexibility for the sponsor on the decision as to how best to conduct the development plan. ... [Deviations from recommendations in the EMA’s product-specific guidelines also suggest it’s] possible to get approval for a product without following all regulatory guidelines in the assessment, as long as the overall application is convincing.
Though the EMA’s role in reviewing these applications was not explicitly discussed in this study, I found these results indicative of the great challenge facing both the EMA and FDA today. In particular, the varying development models call attention to the great need for regulatory flexibility. The results of the BJCP study go a long way toward providing a clear example of just what regulatory flexibility looks like in the EMA. But where does the FDA, which got a late start with biosimilars, stand in terms of regulatory flexibility today?
How Flexible Is The FDA?
We already know the FDA is currently lagging behind the EU in the number of biosimilar approvals. That said, with upwards of 60 programs currently in development and three biosimilars approved, the FDA has certainly grown in its understanding of biosimilarity. As Steven Kozlowski, director of the FDA’s office of biotechnology products within the office of pharmaceutical quality, described in his presentation, the agency issued several biosimilar guidances after the Biologics Price Competition and Innovation Act (BPCIA) was signed in 2010. These guidances, which discussed the general scientific, clinical, analytical, and quality considerations for developing biosimilars, were finalized three years later. As he described, “Though there was a lot of discussion about biosimilars happening in the agency between 2010 and 2013, the concepts in these guidances remained mostly unchanged. The FDA was figuring out how these concepts actually applied to the biosimilar candidates in its meetings.”
To determine biosimilarity, Kozlowski singled out the need for analytical studies; animal studies (including toxicity); and clinical studies examining PK, PD, and one or more conditions of use. However, “In the small print, there is some flexibility here,” Kozlowski clarified. “Any one of these types of data may not be necessary.” For instance, should a biosimilar company have a more thorough analytical package for its candidate, the additional clinical studies could be less extensive.
As I described in my take-away article from the GPhA show, one of the big topics of discussion throughout the conference was data, in particular how much and what is absolutely necessary to demonstrate biosimilarity. The phrase “totality of the evidence” has become common in a number of FDA documents and presentations. Ultimately, this means, “The agency is going to be flexible when thinking about what it needs scientifically to establish biosimilarity. This will not necessarily fit some sort of checklist,” Kozlowski explained.
The Human Aspect Of Regulation
One of the biggest challenges for the FDA has been how to decide how close is close enough in terms of analytical similarity. A great example of this challenge was the FDA advisory committee meeting for Sandoz’s etanercept biosimilar. During this meeting, both Sandoz and the committee homed in on the misfolded protein present in both the reference product and biosimilar. Though the biosimilar contained a misfolded component, it only had roughly half as much of the misfolded product. Naturally, a situation such as this raises questions about how close the biosimilar has to be to the reference product. For instance, does a biosimilar have to match the impurities found in the reference product? How could these impurities impact the molecule’s potency and the patient?
Unfortunately, there is no magical, mathematical formula capable of proving a drugmaker’s evidence clearly demonstrates biosimilarity. As Kozlowski said, “It’s really a human issue.” In order for the “totality of evidence” concept to work, the FDA emphasizes team review. Because a biosimilar application will contain a variety of evidence from different disciplines (e.g., analytical, clinical), experts from these disciplines deserve a place at the table to share information throughout the review process.
This is not to say statistical analysis won’t play a role in determining biosimilarity. The FDA is currently at work on guidance on this topic. However, a big challenge is ensuring the agency drafts reasonable statistical standards. It’s dangerous to rely on a fixed statistical rule or the expectation the biosimilar will match everything within the reference molecule. “We don’t want standards that are so limiting there can’t be any flexibility in the evaluation of biosimilars,” said Kozlowski. And, as many companies come forward bearing newer analytics and technologies, the FDA needs this flexibility to be able to make the best judgement about each biosimilar. “We want rigor, we want statistics,” said Kozlowski. “But we really do think the science will play out.”