By Anna Rose Welch, Director, Cell & Gene Collaborative
We’ve gotten used to hearing the phrase “if, then,” a lot in the biosimilar industry over the past 15 years. This has been especially true on the regulatory front. Beyond the general discussions surrounding extrapolation or interchangeability, there’s been much lively debate on what the biosimilar development paradigm of the future could look like as analytical methodologies and technologies continue to improve. Peppered throughout these discussions has been one phrase: “If there is enough known analytically about the product’s structure and function and there is a biomarker, then we can eliminate the requirement for a large comparative efficacy study.”
To be fair, some progress has been made to establish a more efficient development paradigm. We celebrated the FDA and EMA approval of two different biosimilars for pegfilgrastim that were approved without a comparative efficacy study — though pegfilgrastims do have a PD biomarker. The EMA introduced a pilot program enabling companies to bring forth suitable candidates for tailored development, however this program has yet to bear fruit. Some peer-reviewed publications, such as this one, have even gone so far as to lay out a future development paradigm for biosimilars sans comparative efficacy studies, and some regulators are increasingly on board. Despite these bright spots, however, we’ve had a hard time moving beyond the statement, “If this can be done.”
But a new peer-reviewed paper fresh off the press in early April aims to do just this: lift us out of our “if” prison into concrete, data-supported discussions on how we can do this. The paper published in BioDrugs is entitled “The Path Towards A Tailored Clinical Biosimilar Development” and was penned by Sandoz’s Martin Schiestl and experts from member companies of the International Generic and Biosimilar Medicines Association (IGBA). In an interview with Schiestl, we discussed the paper’s most important findings and what implications these data will have for biosimilar developers and regulators moving forward.
The Paper In Brief: Filling In The Knowledge Gap
Schiestl did an excellent job at giving me the paper’s 30-second (or less) elevator pitch. As he described, “We carried out a retrospective analysis on all development programs for biosimilars approved in the EU and/or the U.S. between April 2006 and November 2019 and found that comparative clinical efficacy was never an issue in the review process. If a biosimilar did not receive regulatory approval, it was not because of data uncovered at the comparative efficacy level. Therefore, we propose that we need to re-think biosimilar development following today’s regulatory science. A more tailored development approach with no routine need for comparative efficacy trials could be adopted while maintaining robust regulatory standards.”
As we all know by this point (hopefully!), the comparative clinical efficacy study is carried out simply to confirm a biosimilar’s similarity, which has already been thoroughly illuminated through extensive analytical characterization and a PK study. However, given the number of biosimilars approved since 2006 — 60-plus in the EU and 23 in the U.S. as of November 2019 — the industry has increasingly questioned whether these large studies provide additional information that has not already been gleaned from the analytical profile. In fact, many are concerned that comparative clinical efficacy trials are only being prescribed because this is what has been required since 2006.
But “Despite the number of conversations and publications on this topic of tailoring, there were no sources containing the answers to the scientific questions we often hear from regulators,” Schiestl explained. “We wanted to unite all the relevant development data in one place so we could ground this discussion in the scientific reality.”
In total, the publication’s authors reviewed 45 biosimilar programs, 42 or which had been approved in the EU and 23 in the U.S. The remaining three did not gain EMA or FDA approval for reasons discussed in detail in the article. While all 42 programs carried out a PK study, 38 of them also had a comparative efficacy trial. Following the completion of those 38 trials, 36 of them were deemed highly similar and subsequently approved. Interestingly, the remaining two biosimilars were deemed equivalent in the comparative efficacy trial; however, each program revealed differences in immunogenicity due to process-related impurities. In turn, both programs returned to the manufacturing process to improve product quality to enable regulatory approval.
It’s also critical to note that these two biosimilars were developed in the early days of biosimilar development, and the study found that no other programs have encountered the same pitfalls since 2010. In fact, given the rate at which technology and scientific knowledge has improved, the authors emphasize it’s highly unlikely we’d see such issues emerging today.
Though the industry and global regulators are currently (and rightfully) distracted by the current Covid-19 response, Schiestl hopes this retrospective analysis serves as the necessary stepping-stone for regulators to arrive at greater consensus about biosimilar development data needs moving forward. “This paper can be the baseline for discussions that move beyond the basic ‘yes, we can,’ or ‘no, we cannot’ debates we’ve been having to this point.”
(Don’t forget to download the supplemental data file following the references in the full-text article for more insights!)
Regulatory Alignment And Learnings: What Potential Does This Data Reveal?
After more than 700 million patient days of effective biosimilar use globally, regulators (and in turn physicians and other stakeholders) have naturally become more comfortable with biosimilars. In the past, there have been concerns about scientific misalignment between the agencies. In a global industry like the biosimilar industry, this can ultimately lead to difficulties efficiently approving the same biosimilar in multiple territories.
This retrospective analysis will go a long way toward continuing the dialogue amongst agencies about where regulatory requirements can continue to align. The agencies know it’s in everyone’s best interest to provide consistent regulations. But what is important to note is that the biosimilar development program proposals that companies bring to the agencies remain individual and complex. “One company may choose to provide certain types of data in an effort to make sure the agency cannot say no to their overarching development plans,” Schiestl clarified. As such, in reviewing data for this publication, the authors weren’t able to make any sweeping assertions about the current level of scientific alignment between agencies for the currently approved biosimilars. “It was impossible to differentiate between data that was requested from the agency or that was part of the company’s individual strategy for that development program,” he explained.
One of the most interesting case studies we’ve seen that exemplifies the complexity of establishing and reviewing biosimilar programs is that of trastuzumab. The development programs for two different trastuzumab biosimilars — ABP980 and SB3 — uncovered a shift in the reference product’s ADCC critical quality attribute (CQA) that impacted the clinical primary endpoint. Though this shift was due to a presumed manufacturing change and affected only a certain number of the reference product’s lots, the biosimilar development programs were, of course, impacted — but to a different extent depending on the agency. Under the EMA, the primary endpoint was not achieved in either biosimilars’ comparative efficacy study. However, under the FDA’s purview, only the ABP980 program did not meet its primary endpoint. As the authors point out, this was “presumably because the FDA and EMA had different expectations for the analysis plan of the same study.” This goes a long way to showcasing the complexity that can arise in each company’s statistical plans.
However, Schiestl urged looking at the big picture here: in each case, the EMA and FDA came to the same conclusion about these products and ultimately approved them. “This shows the degree of alignment between agencies,” he qualified. “But there will always be some level of differences; it’s inevitable, especially given that different individuals are assessing these products and there are different strategies companies may choose when developing the same molecule.”
Moving forward, the industry must put its best, highest quality proposals forward demonstrating the necessary functional characterization and bioassays have been carried out. In addition, each company will show it has analyzed all process-related impurities and that they are well-controlled.
It’s also important that the regulatory agencies arrive at a consensus on how they plan to evolve the biosimilar pathway together. After all, making biosimilar development more efficient and sustainable is not just for the benefit of the industry; it will also improve regulatory efficiency. This paper comes at the perfect time for the EMA, which released its “Regulatory Science to 2025” strategy in January 2020. These overarching regulatory science goals have prioritized (among many other things) the reevaluation of the biosimilar development framework. To achieve the biosimilar-related goals of this regulatory strategy, it is imperative that the agency take into account and integrate the latest scientific knowledge about the comparability assessment into its biosimilar clinical requirements.
How These Lessons From The Past Can Shape The Road Forward
Now that this retrospective analysis has been completed, the next step is to gain consensus with regulators about how to move forward. It is highly unlikely regulations will be made that eliminate comparative efficacy trials for every single product. Rather, this shift in mindset will happen first with the molecules that regulators are more familiar with — for instance the IgG1 monoclonal antibodies. It will all come down to how much is known about the functional properties of the molecule, and, so far, we do have a large knowledge base for all the antibodies, Schiestl qualified. As an industry, we are accustomed to attaining as much knowledge as we can about molecules’ CQAs. In fact, this practice was established in a systematic manner for all biologics more than a decade ago during the quality by design initiatives in Europe and the U.S. which ultimately led to the International Council for Harmonization (ICH) guidelines Q8, Q9, Q10, and Q11.
Of course, that’s not to say that there won’t be products whose functional profiles aren’t still somewhat of a mystery. Schiestl pointed specifically to erythropoietin. Though this is a simple cytokine, more than 40 percent of the molecule’s mass is attributed to glycosylation, which does impact the functional profiles and is difficult to characterize in its entirety. Erythropoietin’s immunogenicity profile is not currently linked to glycosylation in the scientific literature; regardless, immunogenicity needs to be prevented by all means as it may have severe consequences for patients. Molecules such as this could serve as a special case in which the development package requires additional scrutiny.
But Schiestl also urged we move beyond the notion that we’ve been operating under today: that we can only employ tailored development for products that are equipped with biomarkers. This paper proposes expanding beyond this for a few key reasons. For one, most antibodies do not have biomarkers, and finding them is a risky exercise that may not result in success. Similarly, Schiestl argues that, “Because a biomarker study measures the clinical physiological response, which is required for efficacy, it actually serves the same purpose as a primary endpoint of a comparative efficacy trial. Therefore, we need to think beyond how we can ensure robust regulatory conclusions without the PD biomarker.”
To move the conversation forward, this article proposes an outline of what a streamlined biosimilar development paradigm could look like. Primarily, it would place greater emphasis on confirming that the physiochemical and functional CQAs are highly similar to that of the reference product. Should any of the analytical testing reveal problematic differences, the company would need to go back to the manufacturing process to reoptimize the product’s quality. Depending on the CQA affected, this could mean further purification to eliminate host-cell proteins, media optimization, or, on an even larger scale, going back to the drawing board with a new cell line to ensure the highest quality protein is being produced.
Once high similarity is established, the biosimilar would move into the PK equivalence trial to determine PK and further confirm safety and immunogenicity. Though there will be greater emphasis placed on the PK study in a development paradigm without comparative efficacy studies, Schiestl does not foresee this posing additional challenges for the industry. PK studies are far from novel, and there has been plenty of guidance over the years helping companies structure these trials.
Rather, the emphasis is going to fall much more heavily on the analytical physiochemical and functional characterization. “What our paper shows is that we have the technology and the agency guidance to do this thoroughly today,” he offered. For example, he singled out the FDA’s recent guidance on comparative analytics. This guidance laid out the quality-range approach to help companies determine that their moderate- to high-risk CQAs are in line with those of the reference product’s. To Schiestl and many others in the industry, this method is the “perfect approach” and the “way of the future” to ensure that biosimilar CQAs are well-controlled. (Of course, it will depend upon the reference manufacturer maintaining the same control of their product. That said, it’s not anticipated we’ll see similar CQA drifts in reference products beyond the trastuzumab example.)
As such discussions evolve and we strive to keep stakeholders informed of the data behind biosimilars’ approvals, it will be critical to approach these proposed regulatory changes with the right mindset and language. Revising the regulatory pathway to eliminate comparative efficacy studies must not be considered or labeled a “short-cut” by industry players. The goal is not to do less; rather, these conversations are to ensure that regulators only require the best, most scientifically precise data.
“The regulatory science is a continuum,” Schiestl concluded. “If new knowledge about these products comes up — for example, due to manufacturing changes or biosimilar development — it must and will be included. Integrating this knowledge and acting upon it will not be a radical action; it will be taken step by step.”