Is Confirmatory Evidence Having a Moment?

As we begin the final quarter of 2022 and the leaves here on the east coast begin to turn and fall, it seems the clock may be running out on FDA and the Center for Drug Evaluation and Research (CDER) to meet its goal of publishing a draft guidance on confirmatory evidence this year.  For the past two years, CDER’s Guidance Agenda has included a spot for this eagerly anticipated guidance, tentatively titled ‘Meeting the Substantial Evidence Standard Based on One Adequate and Well-Controlled Clinical Investigation and Confirmatory Evidence’.  To be fair, the annual guidance agenda consistently presents a lengthy and somewhat aspirational to do list, and it is not surprising when one slips through the cracks.  However, we wonder whether recent and ongoing approval decisions are shaping FDA’s thinking, possibly pushing back issuance of a draft.

The two previous FDA guidance documents on meeting the patient benefit part of the statutory standard for drug approval – substantial evidence of effectiveness – were released in May 1998 (Providing Clinical Evidence of Effectiveness for Human Drug and Biological Products, available here) and December 2019 (Demonstrating Substantial Evidence of Effectiveness for Human Drug and Biological Products, available here), just over two decades apart!  Additional guidance on confirmatory evidence is sorely needed as the December 2019 draft guidance spends just two of its 18 pages telling us what FDA thinks of such a key piece of statutory language.  Moreover, the draft guidance is the first time FDA tried to explicitly describe its position on such a key phrase (“confirmatory evidence”) that was made a part of the plain language of the statutory standard for drug approval in 1997, a 22-year interval of near public silence.

Even in the absence of a guidance document dedicated to explaining FDA’s views on interpreting the statutory construct of “confirmatory evidence,” a number of drug approvals over the past few years give us a great deal of insight into FDA’s current thinking and how FDA is applying this standard.  These illuminating drug approvals include FDA’s September 29th approval of Amylyx Pharmaceutical’s therapy for amyotrophic lateral sclerosis (ALS), Relyvrio (see additional recent coverage from HPM’s FDA Law Blog, here, and the Pink Sheet, here).  While Relyvrio’s approval is just the most recent and perhaps high-profile application of this single study approval standard, we think confirmatory evidence is having a moment all its own through such FDA actions.

We use the term “confirmatory evidence” as short-hand for one of the two ways explicitly defined in the Federal Food Drug & Cosmetic Act (“FD&C Act”) for demonstrating substantial evidence of effectiveness.  Since 1962, the Act has given FDA authority to reject a drug application that fails to provide substantial evidence of effectiveness.  The Act defines substantial evidence as that consisting of “adequate and well-controlled [clinical] investigations” upon which qualified scientific experts could conclude that a drug will have its purported effect.  In 1997, the Food and Drug Administration Modernization Act (“FDAMA”) amended the definition to make clear that substantial evidence could alternatively be demonstrated by “one adequate and well-controlled clinical investigation plus confirmatory evidence” (emphasis added).  However, the Act does not define what may constitute such confirmatory evidence other than to say FDA’s interpretation is to be “based on relevant science.”

In the December 2019 draft guidance, FDA describes the rationale behind confirmatory evidence by drawing an analogy to its interpretation of the original 1962 evidentiary bar.  In prevalent conditions, FDA nearly universally requires that substantial evidence be demonstrated via two adequate and well-controlled clinical investigations, for example, two Phase 2 or 3 trials (as long as they have the characteristics delineated at 21 CFR § 314.126).  Some may confuse this requirement for two studies as a need for replication.  Those who do should be forgiven as we often think of replication as the appropriate means to demonstrate that the results of a scientific study are not due to chance.  But replication implies identical repetition of the first study and thus may be prone to any flaws inherent to the first study.  More importantly, replication is but one means to substantiate the results of a clinical investigation or scientific study.

Conducting a second investigation that tests the same fundamental principles (e.g., that drug A can treat disease X) but with some variation in its design (e.g., different endpoints, populations, or durations) may provide more compelling substantiation that a drug has the effect it purports to have.  Showing that the drug has consistent effects across two slightly different patient populations with the same disease reinforces the evidence that the drug has a beneficial effect.  However, both the FD&C Act and FDA acknowledge that running two adequate and well-controlled investigations may not be necessary to provide assurance that the drug will have the effect it is purported to have.  Rather, confirmatory evidence can be a means to substantiate the effects observed in a single pivotal trial.

Even so, FDA does not make the decision to rely upon a single adequate and well-controlled trial plus confirmatory evidence lightly.  Several factors may influence FDA’s determination, such as the persuasiveness of the single trial, the robustness of the confirmatory evidence, and the public health need due to the seriousness of the disease, an unmet medical need, or a combination of both, and the feasibility of enrolling two clinical trials.  The infeasibility of a second clinical trial more commonly manifests in orphan conditions because they offer limited pools for recruiting patients.  Therefore, we have more commonly encountered use of the confirmatory evidence pathway in rare disease drug approvals.  However, this statutory standard is not restricted to use in rare disease settings.  The confirmatory evidence standard is applicable in any condition, especially in those where the clinical context warrants expedited development as a result of having a serious or life-threatening unmet medical need.

Notably, the single study approval pathway using the “confirmatory evidence” standard is just one approach that allows FDA to use its expert judgment to determine that an application meets the substantial evidence of effectiveness standard.  FDA’s Subpart E regulations and the December 2019 draft guidance also provide for broader application of scientific judgement in review of new drugs, mainly through the acceptance of greater uncertainty based on what can constitute an adequate and well-controlled trial (e.g., a study designed to be externally-controlled).  Such expert judgment should be considered on top of use of the “confirmatory evidence” standard in rare, serious diseases with an unmet medical need.

All of this brings us back to where we started – FDA’s interpretation of “confirmatory evidence.”  In the December 2019 draft guidance, FDA provides four examples of what may constitute such evidence – (1) supportive evidence from existing adequate and well-controlled clinical investigation(s) in a closely related approved indication, (2) evidence from data that provide strong mechanistic support, (3) compelling results from the single adequate and well-controlled clinical investigation supported by additional data from the natural history of the disease, and (4) support from scientific knowledge about the effectiveness of other drugs in the same pharmacological class.  While the guidance clearly describes these as just four illustrations of confirmatory evidence, it has been our experience that individual review divisions within FDA have not always shared this view.  We have at times encountered not only a reluctance to apply the “confirmatory evidence” standard but even experienced refusal to recognize other sources of evidence not explicitly mentioned in the guidance (but equally capable of substantiating a single pivotal trial’s results).

Despite these experiences, we have been involved with several “confirmatory evidence” approvals, especially for rare serious conditions, in the past few years that suggest that the FDA may be taking confirmatory evidence more seriously.  Many of these approvals have been variations of the guidance’s mechanistic confirmatory evidence example.  These examples have often taken a similar form whereby the positive findings on a clinical endpoint in the single adequate and well-controlled trial are supported by an effect on a biomarker, another pharmacodynamic endpoint, or even a drug mechanism of action that fits squarely within the disease’s pathogenic pathway(s).  In these examples, we have seen everything from data from the same clinical trial to data collected during in vitro and in vivo experiments cited as confirmatory evidence.  These variations on the mechanistic confirmatory evidence example provide some indication that the December 2019 draft guidance’s language is intended to illustrate just a portion of what may be possible sources of confirmatory evidence rather than provide a rigid framework to be followed.

In other precedents, we have seen FDA cite to consistent clinical findings in earlier phase (i.e., phase 1 and 2) studies, randomized trials of higher or lower doses of the same drug, and data from the open-label cross-over portion of a single pivotal trial.  Notably, none of these latter examples are mentioned in the December 2019 draft guidance.  Further, FDA notes in that draft guidance that the strength of evidence from the clinical trial and the robustness of the confirmatory evidence are two additional factors to consider when deciding whether a single adequate and well-controlled trial plus confirmatory evidence demonstrates substantial evidence of effectiveness.

All of this is why we find FDA’s application of the “confirmatory evidence” standard in the case of Relyvrio so interesting.  In this example, FDA explained that Relyvrio’s primary finding, a slowing of disease progression as measured by the ALSFRS-R clinical rating scale, was supported by long-term survival benefit observed in the open-label extension and by vital status census of all those originally randomized in the clinical trial.  Here, FDA cited to survival data from the same clinical trial population as confirmatory evidence.  The FDA noted that the ALS functional rating scale and survival were distinct aspects of the disease.  In other words, these two types of data represented phenomena that were sufficiently distinct that a survival effect could substantiate, or confirm, a functional one.

Relyvrio may be just the latest and perhaps most high-profile confirmatory evidence approval precedent to date, but it reinforces the interpretation that FDA’s discussion of confirmatory evidence in the December 2019 draft guidance was meant to illustrate what is possible and was not intended to set rigid bounds or limitations on the nature or scope of types of confirmatory evidence.  We would be remiss not to mention that our extensive analyses of FDA’s rare disease approvals lead us to believe that a substantial number of FDA’s orphan drug approvals have been based on this “confirmatory evidence” standard since its enactment in 1997 even when FDA has not explicitly cited to this part of the law.

However, we also see Relyvrio and other confirmatory evidence approvals as indicia of FDA’s emerging, explicit interest in this 25-year old portion of the FD&C Act.  Even in the absence of a stand-alone guidance, we sense that confirmatory evidence may be having a moment all its own.