The argument for the elimination of the need for phase 3 studies in biosimilars is pretty simple: They cost a great deal but what do they add to our knowledge about the safety and efficacy of biosimilars? One of the primary tasks of the Food and Drug Administration (FDA), in educating health care professionals, media, the public, academia, and manufacturers, was to deemphasize the importance of the clinical trial in the totality of evidence approach they use to evaluate biosimilars.
The health care professional community, academia, and patient advocates may take another view: They are complex biochemical medicines and we cannot be sure of their safety and efficacy without carefully controlled studies in large populations. We have been ingrained for 40 years with the need for randomized, phase 3 clinical investigations that it may be very uncomfortable indeed to approve a drug without them.
Phase 3 Studies in Biosimilars: Statistically Speaking
At least one pharmaceutical company (Adello) is seeking FDA approval without phase 3 trials that study large groups of patients with the disease indication. In biosimilars, FDA is willing to extrapolate approvals without any clinical studies in other indications, and indeed, payers and providers are willing to accept this.
Furthermore, the FDA has taken many steps to speed access of biosimilars to approval. Other than altering the intellectual property and exclusivity timelines, what action can save more time in the process (much less money for the biosimilar developer) than the elimination of phase 3 studies in biosimilars?
In a phase 3 clinical trial of the originator biologic vs. a biosimilar, what do we expect to see? Since the expectation is that the physiochemical characteristics of the two molecules are exceedingly similar, and phase 1 trials should have proven equivalent pharmacodynamics, we don’t expect big differences in outcomes by phase 3. If phase 2 studies have been performed successfully, we believe this more emphatically.
At worst, we expect to see clinical effects that are on the edges of anticipated norms for the originator drug but within the range expected. One French investigator wrote this month in BioDrugs that typical phase 3 studies with 600 to 1000 patients are not statistically powered to detect more than major differences in safety. What is the real implications of 2 versus 5 drug withdrawals in patients taking medications that are much more alike than they are different? This author believes that well-designed phase 1 trials in volunteers can sufficiently detect the formation of antidrug antibodies and other immunogenicity differences between biosimilars and their originator drugs. This may be particularly true in patients with autoimmune disorders. When patients are routinely given methotrexate (another immunosuppressant) concomitantly with the biologic therapy, reliable evaluations of immunogenicity of the study medications are very difficult. Finding that hidden safety signal may not be possible.
More Pressure on Postmarketing Surveillance
In other words, it is easier to determine whether a biosimilar drug is “noninferior” to a reference product in clinical testing. The range of expected values is small (and there is little or no expectation that a biosimilar will demonstrate superiority). I’m no statistician, but I’d expect that to detect clinically significant differences among outcomes in this type of comparison, one would need study populations far exceeding that of the typical phase 3 study in biosimilars. Unlike in a clinical trial of a study drug versus a placebo or other standard therapy, large differences may be seen, and population sizes may be less important (hence, phase 2 trials of 100 patients may reveal red flags or lack of effectiveness).
Without the use of phase 3 trials in biosimilars to attain comfort and security, the post-marketing surveillance machinery becomes that much more important. The observation of safety issues based on real-world prescribing and utilization will be a front-line defense, not a backstop, to identify unintended pharmaceutical outcomes. This means that more of the onus will fall on the conduct of registry trials, FDA’s Sentinel program, and notably the Biologics and Biosimilars Collective Intelligence Consortium (BBCIC), which is in the process of preparing for its first comparative-effectiveness studies in long-acting insulins (Q4 2018) and granulocyte colony-stimulating factors (i.e., filgrastim, pegfilgrastim).
This would still be a significant leap of faith, based on the approvals and limited use of biosimilars today, but I can envision other companies gambling, with FDA’s consultation, on skipping this traditional step to drug approval. I wouldn’t bet against it.