By Anna Rose Welch, Editor, Biosimilar Development
In 2017, two experts published an article in BioDrugs entitled, “A Global Reference Comparator For Biosimilar Development.” Though it may have been quietly received at the start, BioApproval’s Christopher Webster and Avalere’s Gillian Woollett put forth an argument that is now earning global attention. Not only has the elimination of bridging studies (thanks to the Global Reference logic) become regularly discussed at biosimilar conferences around the world, but it received attention in the FDA’s Biosimilar Action Plan (BAP) and has since become a core element of biosimilar developers’ strategy for more efficient development. In particular, there has been a great push for regulators to adopt the Global Reference from the leading trade group in this space, The International Generic and Biosimilar Medicines Association (IGBA).
Just like the global reference product argument, the push for tailored biosimilar development (e.g., greater reliance on analytics and pharmacokinetics and pharmacodynamics [PK/PD] data) has taken root over the past few years. As Biosimilar Development editorial board member Francois-Xavier Frapaise argued in his paper, “The End of Phase 3 Clinical Trials In Biosimilar Development,” analytical and technological process evolutions have brought us to an age where quality chemistry, manufacturing, and controls (CMC) data and PK/PD studies could eliminate the need for large comparative trials in the development of certain biosimilars. Following the publication of his article, there was an additional comment from Webster and Woollett agreeing that powered efficacy studies are often redundant and that they can be unethical because their outcomes are not in doubt. Underlying all of these arguments are the scores of real-world evidence from Europe showing that biosimilars are indeed performing as expected — and have been for over a decade.
Now, in recent weeks, Drs. Webster and Woollett, joined by Avalere’s Anny Wong, have penned a bold (perhaps to some “heretical”) call to action to the FDA and other regulators. (Think Martin Luther nailing his 95 Theses to the church door in Wittenberg.) In their newest peer-reviewed article entitled “An Efficient Development Paradigm for Biosimilars,” the authors carefully assembled data from all the biosimilars approved in the U.S., EU, Canada, and Australia to propose a regulatory approval paradigm called “confirmation of sufficient likeness” (CSL). This paradigm substantiates the use of analytical and PK data alone to establish high similarity, in turn eliminating the use of bridging studies, in vivo nonclinical studies, and PD and powered efficacy studies (in most cases). As the authors wrote, CSL promises a more efficient but no less rigorous development process compared to regulators’ current “Totality of the Evidence” (ToE) approach.
In light of the paper’s publication and Webster’s future presentation at the upcoming DIA Biosimilars conference in September, I spoke with him to learn more about the CSL approach and the scientific evidence that has informed this proposed paradigm.
As he stated, “The biosimilar approvals on both sides of the Atlantic have shown over and over again that if you have high analytical and PK similarity, you have clinical equivalence. Moreover, we have over 25 years with comparability, in which over 98 percent of manufacturing changes have been adjudicated on analytical data alone. This amounts to thousands of manufacturing changes for which no clinical data was deemed necessary. So, the question becomes, how many times do you need to conclude biosimilar development with a large, marginally sensitive clinical equivalence study?”
The Paradigm, Unpacked: Why Is It So Important Now?
When writing about new scientific publications, I prefer to let the authors’ own work speak for itself, so I’d encourage you to delve into their article hyperlinked above to get the full gist of their argument. But, in brief, Webster et. al.’s article furthers a steadily intensifying discussion around which and how much data are necessary to approve biosimilars today and how regulators can better integrate lessons learned to date into the biosimilar regulatory requirements.
As biosimilars become more widespread in Europe, and hopefully in the U.S., overall sustainability of the market has become front-of-mind. We’ve heard it said a million times — biosimilars are not cheap or easy products to make. In certain countries, we’re seeing discounts negotiated that raise concerns about long-term feasibility of developing biosimilars using the same resource-intensive regulatory pathway. Hence the ongoing discussion around the need for and worth of large, powered efficacy studies required by regulators for a majority of the biosimilars being developed and approved today. A commonly heard argument is that large clinical studies cannot provide precise insights into analytical composition and, because of their size, are ill-equipped for showing any small clinical differences that may exist between a biosimilar and its reference. In fact, some may argue these trials simply raise more questions than provide answers and have become de facto barriers to competition.
Following the ToE approach today, biosimilar development typically includes analytical characterization, nonclinical development (in vitro and, depending on the country, in vivo studies), PK/PD studies, and finally, a powered (quite large) clinical efficacy study. Of course, regulators have put the opportunity for tailored development on the table via some carefully worded phrases in guidances. In fact, the European Medicines Agency (EMA) has even introduced a pilot program to enable companies to approve a product without comparative trials. (This program has yet to bear fruit, unfortunately.)
Though tailoring opportunities exist as a theoretical opportunity today, approval of biosimilars using this regulatory approach has been a rarity. Two pegfilgrastim biosimilars approved by the FDA and EMA within the last year are currently the most meaningful, recent examples of how an in-depth analytical analysis of a molecule followed by a PK trial can sufficiently demonstrate equivalence.
As Webster et. al., argue in their paper, there are several reasons why it’s time to see the transition from the ToE paradigm to the CSL paradigm. There are a few no-brainers, including the approval and positive performance record of biosimilars for the past 13 years, as well as the analytical technological advancements that improve overall knowledge of the structure and, therefore function, of the biological molecules. Not to mention, regulators have amassed extensive experience with comparability to support reference products’ manufacturing changes — a point Woollett doggedly reiterates at so many conferences. In fact, the FDA’s Janet Woodcock herself stated in Congressional testimony back in 2007 that the FDA requires clinical studies for comparability less than 1 percent of the time. Extrapolation of indications remains the same post-change, and the post-change product is still considered interchangeable with its previous versions.
Webster uses the creative analogy of looking at two keys to describe the concept of structure as function, which is an important part of the biosimilar development discussion. When you copy a key, you know from a careful visual comparison of the operative ends of the copy and the original whether the copy will unlock the door. More importantly, you know this even before you use the copy. In the same way, relying on strong analytics and PK data will tell you everything you need to know about the biosimilar’s structure and function and the regulators can be trusted to make this determination.
This is the point Webster et. al., returns to throughout the paper: Of the 70 biosimilar approvals we’ve seen between the U.S. and EU, a candidate with a strong analytical match and PK profile has always been approved as a biosimilar following the clinical efficacy study. Regardless of this fact, “The assumption today is that once we have the analytics and PK, there is still more to know,” Webster said. “We have to reframe this thinking process. If you get to the point where you have good analytics and a solid PK in which the PK curves are literally overlying each other, you know how a clinical equivalence study will conclude. You know it as surely as your own name.” (This is where the argument these large efficacy studies are unethical comes in: You’re entering that equivalence trial without any scientific doubt that you already have a matching molecule.)
“This is no longer in the realm of ‘opinion,’” said continued Webster. “Just look at the historical data. And in situations where the analytical or PK similarity is not quite as close, once again, look at the historical record; it turns out that experienced regulators are extremely good at knowing when additional data are or are not necessary.”
Practicalities: How Could This Paradigm Roll Out?
When I asked Webster how this process could be implemented compared to the current ToE approach, he shared one possible example which, he argued, could be the recipe for even higher quality, “cleaner” biosimilars moving forward.
For example, a manufacturer could carry out its analytical characterization and functional assays, including PK, and amass representative data at scale. These data could be submitted to the regulator for scientific advice. “During this process, the regulator would examine the data and provide responses to the sponsor’s questions, including a request for the concurrence that the data, with any planned and stated complements, appear to be sufficient to support a marketing application and may be submitted in a marketing application at risk,” outlined Webster. This process would give a sponsor a good indication of whether its marketing application would be successful as is, or whether any additional work would be necessary.
Following any additional process revisions, the regulator would then review the data again, and should the additional work be sufficient, the manufacturer would then conduct the PK study. The submission containing the PK data to the regulatory agency would ultimately be the application for market approval (i.e. the marketing authorization application [EU] or the biologics licensing application [U.S.]). The agency would then review the data to determine whether the analytics and PK study confirms the biosimilar’s “sufficient likeness,” in turn eliminating the need for a large clinical trial.
However, it is important to note that Webster and the other authors are not arguing there won’t ever be residual uncertainty following analytical and PK work, nor will every biosimilar development scenario be a fit for the CSL approach. “We have not written that there will never again be a need for a clinical study in biosimilar development,” Webster clarified. “What we have said is that a clinical efficacy study shouldn’t be a routine requirement for every single biosimilar. Rather, a regulator can look at analytical and PK data, and if it’s decided at that point based on the submitted data you don’t need to do a clinical study, then that’s your approval.”
Overall, at the heart of this call-to-action is the desire to see biosimilar regulation evolve with new scientific learnings. It also seeks to ensure that biosimilar sponsors do not have unnecessary and ultimately wasteful regulatory burdens imposed upon them. As has been said many times, regulatory consistency is a must.
Throughout the article, the authors continue to revert back to the use of comparability with reference biologics following a manufacturing change. Many of the leading biologics with biosimilars have undergone multiple manufacturing changes in series, which can lead to shifts in the product’s critical quality attributes. This can occasionally lead to bumps in the road for biosimilar developers. A great example of this occurred with the trastuzumab biologic, which experienced a shift in one of its critical quality attributes. This shift and the impact it had clinically wasn’t discovered until the biosimilar development process. (Information about manufacturing changes is not publicly available in the U.S. However, when pressed at a recent conference to share what data the manufacturer submitted to approve that manufacturing change, EU regulators were unable to comment.)
I’ve even heard of examples in which originator products that have seen changes to the molecule’s primary structure (which would mean it’s no longer even biosimilar to itself) has been approved using nothing more than analytical data — thanks to the strength of that data.
It’s examples such as this that emphasize just how important it is to have a level playing field between biosimilars and biologics. In fact, one possible movement towards equalizing this playing field would be to see reference product manufacturers required to test PK in their products following a manufacturing change.
Though Webster acknowledges the CSL approach outlined in the paper may be deemed “heretical” by some stakeholders in the industry, he points to the fact that he and his fellow authors are hardly alone in their questioning of the current development paradigm. Similar conversations have been occurring more frequently on the international conference front among clinical pharmaceutical experts. But even more importantly, outside of the papers and discussions on this topic, there is scientific data to back up their argument. “This is a movement which is growing, and I believe the status quo is no longer tenable,” Webster said.
Woollett concluded, “To follow the science cannot be heresy, surely. We are open to anyone telling us we have got the science wrong. However, we ask that, if our science is right, it not be overlooked, for patients’ sake.”
In addition to reading Webster, et. al.’s paper, I’d also urge you to check out the upcoming DIA Biosimilars conference September 23-24th in Bethesda, MD. Webster will be discussing this paper and other regulatory challenges and opportunities along with a panel of other industry experts. Woollett will be heading up a panel on the U.S. biosimilar insulin opportunity. Hope to see you there!