Long-Awaited Guidance on FDAMA 115: Confirmatory Evidence Finally Has Its Moment (to be Crossed Off the FDA’s Guidance To-Do List)October 3, 2023
This time last year, we wrote about a long-overlooked FDA statutory authority and wondered if this provision, known colloquially as the “single study plus confirmatory evidence” pathway, was having a moment (previous post here). Two weeks ago, FDA published a draft of its latest drug development guidance explaining how drug and biological product developers can use this pathway to meet the statutory standard for efficacy. The eagerly anticipated guidance, Demonstrating Substantial Evidence of Effectiveness With One Adequate and Well-Controlled Clinical Investigation and Confirmatory Evidence (i.e., the Confirmatory Evidence Guidance, available here) had appeared on CDER’s Guidance Agenda for the past three years.
A brief explainer on confirmatory evidence: how we got here
Before we dive into the specifics of the new guidance, we should explain that the statutory standard we allude to is the requirement that an applicant provide “substantial evidence of effectiveness” to obtain approval of an NDA or BLA. Since 1962, the Federal Food Drug & Cosmetic Act (FDC 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 also be demonstrated by “one adequate and well-controlled clinical investigation and confirmatory evidence.”
However, the Act does not define what may constitute such confirmatory evidence. Instead, FDA’s determination of whether a single study plus confirmatory evidence could provide substantial evidence is to be “based on relevant science.” The original 1962 articulation of the efficacy requirement is similarly infused with elements of deference to scientific judgment. The studies constituting substantial evidence are to be conducted “by experts qualified by scientific training and expertise to evaluate the effectiveness of the drug involved.” Those studies must be able to support a “fair and responsibl[e]” conclusion by similarly qualified scientific experts that the drug has the claimed effect. Just as the original 1962 efficacy standard makes scientific judgment central to a determination of substantial evidence, the 1997 amendment makes scientific judgment central to a determination of what constitutes sufficient confirmatory evidence to substantiate a single adequate and well-controlled trial.
The new guidance is one of three policy documents dedicated to explaining FDA’s interpretation of this statutory authority and their approach to exercising scientific judgment in evaluating drug effectiveness. However, the two previous FDA guidance documents, 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), differ from the new guidance. The May 1998 and December 2019 guidances deal with substantial evidence broadly and include only a small amount of discussion of what constitutes confirmatory evidence. In contrast, the new September 2023 Confirmatory Evidence Guidance is dedicated solely to the single study plus confirmatory evidence pathway.
However, even before publication of this new guidance, we have observed a dramatic uptick in discussions about the confirmatory evidence between FDA and product sponsors. Notably, the discussions have occurred with increasing frequency since publication of the December 2019 Guidance, in both CDER and CBER, and across nearly all review divisions. The timing suggests to us that publication of even that short discussion of confirmatory evidence in 2019 led to a greater willingness on the part of both sponsors and FDA review divisions to engage meaningfully on how to use the single study plus confirmatory evidence pathway. This has coincided with a similar increase in FDA review divisions citing to, analyzing, and applying this authority during their review of and, ultimately, approval of recent NDAs and BLAs. Use of this single study authority has been crucial to breakthroughs in drug development for rare and serious conditions including those with unmet medical needs. Several of these approvals, including examples in which we were directly involved—Relyvrio for ALS and Skyclarys for Friedreich’s Ataxia—expanded upon the examples FDA previously provided to illustrate their interpretation of the single study plus confirmatory evidence statutory standard (see our earlier blog posts here and here).
We are heartened to see that this latest guidance reflects many of the advances we observed in practice since 2019. And, while we were eager, perhaps impatient, to see FDA publish this document, we perceive that more time may have led to a better articulation of FDA policy as more direct experience with the single study plus confirmatory evidence pathway accumulated across review divisions. As we explain in greater detail below, the new guidance elaborates upon certain policy positions that FDA mentions only briefly in their earlier guidance. Although we will also point out where we think the FDA could further improve and clarify their policy, we commend the Agency for this substantial (pun intended) expansion of its current thinking.
The strength the confirmatory evidence is relative to that of the single study
The new Confirmatory Evidence Guidance begins its discussion much like FDA’s previous guidances on substantial evidence with overarching considerations for applying the statutory standard. FDA reiterated its long-held notion that substantial evidence consists of two primary components, the quantity and quality of evidence. While the current guidance does not elaborate on previous guidances’ explanation of how these concepts apply to the single study itself, it instead makes clear that the quantity of confirmatory evidence necessary to provide substantial evidence depends upon both (1) how robust and convincing both the primary source of evidence (i.e., the single trial) is and (2) how robust and convincing the confirmatory evidence is. In a sense, the primary evidence (i.e., the single trial) and secondary evidence (i.e., the confirmatory evidence) exist on something of sliding scale with respect to one another. As articulated by FDA, “[t]he quantity (e.g., number of sources) of confirmatory evidence necessary to support effectiveness may vary across development programs . . . [and] will be impacted by the features of, and results from, the single adequate and well-controlled clinical investigation that the confirmatory evidence is intended to substantiate.” FDA explains further that a more convincing single trial could be supported by a lesser quantity of confirmatory evidence.
The discussion of the quantity and quality of evidence is, however, lacking in significant ways that are particularly important for rare disease drug development. In the December 2019 Guidance, FDA writes that the clinical context for a disease (i.e., its rarity, its seriousness or life-threatening nature, and its unmet medical need) inform how much uncertainty the FDA will accept in terms of what constitutes an adequate and well-controlled trial (see Section V of that guidance available here). We found it to be a significant gap that FDA did not apply the same concepts to considering the quantity and quality of confirmatory evidence. Dedicated discussion of this topic is especially important in the context of a rare disease where there may be greater uncertainty about the single study’s findings due to the common challenges that persist to developing orphan drugs. In the context of a disease that is all three of those things, that is, rare, serious and with an unmet medical need, FDA has authority to and should apply the maximum degree of flexibility when making determinations about substantial evidence, including in considering confirmatory evidence.
A constellation of confirmatory evidence can be provided
In addition, FDA made several other noteworthy overarching points in the Confirmatory Evidence Guidance. First, FDA clarifies that confirmatory evidence can consist of data from multiple sources. While the December 2019 Guidance provided several examples of potential sources of confirmatory evidence, it lacked any language to indicate whether FDA viewed the different types of confirmatory evidence listed in the previous guidance as being mutually exclusive. This new guidance confirms something we have seen in practice: that confirmatory evidence may consist of “one or more sources (e.g., clinical data, mechanistic data, animal data).” In our words, a constellation of data from multiple sources may constitute the confirmatory evidence that substantiates the results of one adequate and well-controlled clinical investigation.
Potential pitfalls when pursuing this pathway
The Confirmatory Evidence Guidance also includes two notable areas of caution for drug developers interested in pursuing single trial plus confirmatory evidence approvals. The new guidance reiterates statements from the December 2019 Guidance that the statutory standard for approval requires demonstrations of both safety and efficacy in order to inform the FDA’s benefit-risk assessment. However, the new guidance goes further to warn those that a single pivotal trial may not, on its own, provide sufficient safety information to support an indication.
The guidance’s second warning reminds drug developers that FDA regulations (i.e., 21 C.F.R. 314.50(d)(5)(iv)) require that “marketing submissions [include] a description and analysis of all data or information relevant to an evaluation of the safety and effectiveness of the drug product, from any source, foreign or domestic, to avoid selecting only those sources that favor a conclusion of effectiveness.” FDA grimly states that “results of a clinical investigation or confirmatory evidence can be called into question by conflicting evidence unless there is a sufficient scientific justification that may explain the disparate findings” (emphasis added).
Sources of confirmatory evidence expanded
Ultimately, the portion of the Confirmatory Evidence Guidance that represents some of the most informative changes from earlier guidance is in its discussion of the potential types of confirmatory evidence. In the December 2019 Guidance, FDA provides four examples – (1) evidence from existing adequate and well controlled clinical investigation(s) in a closely related approved indication, (2) data that provide strong mechanistic support, (3) compelling results from the single AWC 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. The new guidance not only expands upon these broad categories individually (more on this below) but adds whole new categories, listing them as follows: (1) clinical evidence from a related indication, (2) mechanistic or pharmacodynamic evidence, (3) evidence from a relevant animal model, (4) evidence from other members of the same pharmacological class, (5) natural history evidence, (6) real-world data/evidence, (7) evidence from expanded access use of an investigational drug.
Before we discuss these categories individually, we are compelled to point out that the new guidance states that this list of examples is not intended to be an “exhaustive list.” In other words, the new Confirmatory Evidence Guidance makes clear that other types or sources of confirmatory evidence exist beyond those which are listed in the guidance. Although the December 2019 Guidance described its four types of confirmatory evidence as “examples,” it has been our experience that some FDA review divisions viewed the list in the December 2019 Guidance as being an exhaustive list and the only acceptable sources of confirmatory evidence. The new guidance both explicitly and implicitly confirms that the guidance’s examples are meant to be illustrations of possible sources of confirmatory evidence and not the exclusive list of acceptable options.
Clinical evidence from a related indication
FDA elaborated significantly upon their earlier discussion of when and how confirmatory evidence could come from clinical evidence in a related indication. FDA has dropped the notion that the closely related indication must be an approved indication. The new guidance adds that the clinical evidence from two related, unapproved indications could provide confirmatory evidence for one another in the setting of concurrent approvals for both indications.
However, we believe that there are also circumstances in which confirmatory evidence could come from an adequate and well-controlled trial in a related indication even if a sponsor is not seeking approval for that indication. The same underlying factors that make an adequate and well-controlled trial a robust and reliable source of evidence in the setting of an approved indication or indication submitted for concurrent approval would also be present even if the indication is not currently under review. Fore example, the new guidance notes that there are at least three factors will be considered when deciding whether evidence from a related indication could be confirmatory evidence; the degree of similarity between (1) the indications, (2) the drug’s mechanism of action [“MOA”] in each indication, and (3) the efficacy endpoints.
Mechanistic or pharmacodynamic evidence
FDA also provides significant new explanation of mechanistic confirmatory evidence. At first, this discussion sets a rather high bar stating that “strong mechanistic evidence . . . may be appropriate to use as confirmatory evidence” if (1) the pathophysiology of the disease is well understood, (2) the drug’s MOA is clearly understood, and (3) that MOA directly targets the major driver(s) of the disease. Where these connections between the drug’s MOA and the disease pathophysiology are either less direct or less clear (e.g., when the disease pathophysiology is not well understood) “mechanistic data may not provide sufficient confirmatory evidence . . . and additional evidence from other sources may be needed.” While FDA continues to indicate a preference for “relevant and well-understood” pharmacodynamic markers as the source of mechanistic confirmatory evidence, the new guidance acknowledges that in vitro data could provide a source of such evidence.
FDA does something else interesting here with respect to pharmacodynamic markers. In a footnote, FDA states that demonstration of an exposure-response relationship between the drug and a pharmacodynamic biomarker could be “particularly persuasive” confirmatory evidence even where disease pathophysiology is not well understood and the biomarker’s relationship to the drug’s MOA is not clear. In effect, FDA has expanded the notion of mechanistic confirmatory evidence to include both true mechanistic evidence as well as implied mechanistic evidence based on pharmacodynamic data.
Evidence from a relevant animal model
Previously, FDA only discussed a possible role for data from animal models within the context of mechanistic information. The new guidance, however, provides an entirely separate discussion of the role of this data. In doing so, FDA is clearly drawing much narrower boundary around the potential for animal models to contribute to confirmatory evidence than before. “Typically, results of studies conducted in an animal model of disease are intended to support progressing a drug candidate forward from preclinical to clinical development, rather than to support a finding of substantial evidence.” FDA then describes three potential factors to weigh when considering the contribution of animal model data: “ similarity of pathophysiology and manifestations of the disease in the animal model and in humans,  elucidation of the drug’s mechanism of action with evidence of similar pharmacology and pharmacodynamics in the animal model and humans with disease, and  evidence that the results of efficacy studies conducted in the animal model reasonably support clinical benefits and outcomes in humans with disease.”
Ultimately, FDA sets a very high bar for accepting animal model data as confirmatory evidence stating: “Only models that have proved to be translational (i.e., prior drugs with the same intended clinical effect have been shown to have this effect observed in the animal model, with similar exposure-response) are likely to be considered as confirmatory evidence.” For rare diseases and, obviously, those with unmet medical needs this may prove to be an impossible bar to surmount. But as we noted above, FDA did not include a discussion of how their authority to exercise regulatory flexibility in the context of rare, serious conditions and those with unmet medical needs will be applied to evaluating confirmatory evidence. In the context of many rare and ultra-rare conditions, an animal model may be able to meet one or more of FDA’s factors above and, yet, not be able to pass this translational test because there may no prior drugs developed to demonstrate the animal model’s translational ability. Such circumstances appear to be a reasonable place for FDA to exercise a degree of flexibility.
Of the types of confirmatory evidence that were discussed in earlier guidance, natural history evidence also saw a significant expansion in this new guidance. Previously, FDA’s example only raised natural history when there were historical controls to reinforce a finding on an objective endpoint (e.g., mortality). The new guidance removes this narrow interpretation and describes natural history more broadly as an opportunity to leverage external controls, such as: (1) the use of patient-as-own-control study designs supported by natural history data to show the absence of significant period effects or spontaneous resolution of the condition and (2) progressive diseases where using natural history can be used to confirm that the amount of deterioration in the control group is an accurate reflection of disease progression.
While FDA does not make a distinction between using natural history to substantiate the primary efficacy endpoint from the double-blind phase of a single trial from secondary efficacy endpoints and other supportive analyses, our experiences in ALS and Friedreich’s Ataxia have demonstrated that natural history can be used to substantiate findings from key secondary efficacy endpoints and the open-label extension phase of the single study. In these recent experiences (i.e., the approvals of Relyvrio and Skyclarys), the combination of open-label extension data and natural history data proved to be an invaluable source of confirmatory evidence to support approval.
Same pharmacological class, real-world data, & expanded access
FDA’s new Confirmatory Evidence Guidance includes confirmatory evidence from drugs in the same pharmacological class and, for the first time, evidence from real-world data (“RWD”) and expanded access programs (“EAPs”). The discussions of confirmatory evidence from drugs in the same pharmacological class provide little additional thinking compared to what FDA has already laid out in its December 2019 Substantial Evidence Guidance. Similarly, the discussion of RWD is consistent with previous guidance documents on the topic, including limiting consideration of this type of data to support the approval of a new indication for a drug already approved.
In contrast, the addition of EAPs as a potential source of confirmatory evidence is quite noteworthy. While FDA is clear that many EAPs are not be designed to gather data of the quality and quantity to support their use as confirmatory evidence, FDA nonetheless acknowledges that an EAP could gather data that is sufficiently reliable and persuasive (e.g., objective clinical outcomes in the context of detailed medical record collection). In the context of rare diseases, particularly ultra-rare diseases, where each patient represents relatively larger proportions of the known U.S. patient populations, it seems that EAPs should carry additional weight.