FDA Approvals Based on a Single Adequate and Well-Controlled Trial Plus Confirmatory Evidence

FDA approvals based on one adequate and well-controlled trial plus confirmatory evidence

FDA has repeatedly found substantial evidence of effectiveness on the basis of a single adequate and well-controlled clinical investigation plus confirmatory evidence. This is not limited to rare disease. Reviews in Drugs@FDA apply the standard to enzyme replacement therapies and antisense oligonucleotides, and also to treatments for overactive bladder, major depressive disorder, systemic lupus erythematosus, NASH and ROS1-positive non-small cell lung cancer. The confirmatory evidence reviewers accepted falls into a small number of recurring categories:

• additional clinical trials that were not independently persuasive;
• mechanistic and pharmacodynamic data;
• evidence from a related indication or from other drugs in the same class;
• animal model data;
• natural history data.

Reviewers also rejected or discounted proposed confirmatory evidence in several files. Those decisions show where the bar sits.

Legal and policy basis

Section 115(a) of the Food and Drug Administration Modernization Act of 1997 (FDAMA) amended section 505(d) of the Federal Food, Drug, and Cosmetic Act. The amendment clarified that FDA may consider "data from one adequate and well-controlled clinical investigation and confirmatory evidence" to constitute substantial evidence ²⁰⁴. FDA guidance presents this as a case-by-case route, not a default. FDA must determine that the combined evidence is sufficient to establish effectiveness ²⁰⁵²⁰⁶.

Three FDA guidances address the topic:

• Providing Clinical Evidence of Effectiveness for Human Drug and Biological Products (May 1998) ¹⁹⁵
• Demonstrating Substantial Evidence of Effectiveness With One Adequate and Well-Controlled Clinical Investigation and Confirmatory Evidence (draft, September 2023) ²⁰¹²⁰²
• Demonstrating Substantial Evidence of Effectiveness for Human Drug and Biological Products (draft, June 2026) ²⁰³

When FDA considers the approach appropriate

FDA weighs several factors ¹⁸¹:

• the design, conduct, analysis and persuasiveness of the single trial;
• the source and strength of the confirmatory evidence;
• the seriousness of the disease, unmet need and prevalence;
• whether a second adequate and well-controlled trial would be ethical and practicable.

The stronger the single trial, the less confirmatory evidence may be needed, and conflicting evidence must also be considered ¹⁸⁶. FDA expects sponsors to discuss the plan early. Under the June 2026 draft, that discussion should happen before the single trial starts and should explain the proposed trial and the type and quantity of confirmation ^213177214. Establishing effectiveness this way does not settle safety. A further or longer trial may still be needed for an adequate safety and benefit-risk assessment ¹⁸⁸.

Categories of confirmatory evidence in FDA guidance

The June 2026 draft describes related adequate and well-controlled trial data as the generally expected source of strong confirmatory evidence. That includes trials in related diseases or of related products ²¹¹. The draft also describes a separate approach: a "Highly Persuasive Adequate and Well-Controlled Trial Plus Early-Phase Confirmatory Evidence" ²¹².

Approval examples by type of confirmatory evidence

In the table below, "period" is the decision-date window of the FDA review document in which the finding appears. Several examples are supplemental approvals for new indications.

How reviewers applied each category

Additional clinical trials that were not independently persuasive

The most common form is a second clinical study. Reviewers accepted it even when it could not stand on its own as an adequate and well-controlled trial. Examples:

• Acthar Gel: reviewers accepted Studies 04 and 05 although neither was independently interpretable as positive ⁶²⁶³.
• Saphnelo: reviewers accepted phase 2 Study 1013 and phase 3 Study 05 considered together. Neither was a separate successful adequate and well-controlled trial ¹⁰⁰.
• Livdelzi: ENHANCE ended early, which made it insufficient as the pivotal investigation. Its Week 12 data were still accepted as confirmatory ^419420421.
• Different doses: in Gemtesa the supporting studies tested other doses ²⁶. In Daybue the confirmatory evidence came from a dose-ranging study ¹⁴⁹.
• Phase 2 trials: reviewers accepted phase 2 trials for Arikayce ³⁵ and Rezdiffra ³⁶⁰.
• Single-arm or open-label studies: reviewers accepted these for Recorlev ¹¹⁵, Wayrilz ²⁵¹ and Onpattro ³⁰.
• Active-comparator data: in Auvelity, Study MDD-201 showed superiority to bupropion and served as confirmatory evidence ¹⁰².

Mechanistic and pharmacodynamic evidence

Rare monogenic diseases account for most of these examples. Reviewers usually combined three elements:

1. a well-established disease etiology;
2. a mechanism that directly addresses it;
3. a pharmacodynamic biomarker response.

Xenpozyme ⁴⁵⁹⁴⁶⁰, Lamzede ²⁸⁷²⁸⁹, Joenja ¹⁷² and Veopoz ³³¹³³² follow this pattern. Exposure-response relationships added weight in Nulibry ³⁴⁶, Lamzede ²⁸⁹ and Avlayah ²⁴³. Ekterly shows the approach applied outside enzyme replacement. Reviewers accepted HAE pathophysiology plus in vitro and in vivo plasma kallikrein inhibition data as adequate confirmatory evidence ³⁹⁶³⁹⁷.

Related indication, related product or same class

• Related indication: Epidiolex's tuberous sclerosis complex approval relied on its established efficacy in Dravet syndrome and Lennox-Gastaut syndrome ³¹.
• Same drug in related transplant settings: the Prograf lung transplant review cited RCTs in liver, kidney and heart transplantation under its "Confirmatory evidence" heading ⁹⁰.
• Same approach in an approved product: Wainua drew on existing evidence that the same antisense approach worked in approved inotersen ¹⁶⁷.
• Same pharmacological class: Jideytro relied on the effectiveness of other approved ROS1 inhibitors in ROS1-positive NSCLC ⁴⁸⁹.
• Different disease stage: in Tzield, meta-analyses in Stage 3 disease supported the Stage 2 indication, although the review noted limitations in their strength ¹⁰³¹⁰⁶.

Animal model data

Reviewers accepted animal efficacy data in several reviews:

• Nulibry: MoCD Type A mouse data showing reduced SSC and improved survival ³⁴⁷³⁴⁸.
• Iwilfin: two neuroblastoma mouse models in which eflornithine prevented or delayed tumor formation ³⁷⁹³⁸⁰.
• Nexviazyme: GAA-knockout mouse studies of Pompe disease, cited under "Confirmatory Evidence" ³⁸¹³⁸².
• Miplyffa: NPC1-null mouse and cell data ²⁴⁷.

Natural history and external data

Reviewers used natural history data in two ways. Sometimes it formed the control arm of the single investigation, as in Nulibry ³⁴⁵ and Prograf ⁸⁹. In other files it was the confirmatory evidence itself:

• Besremi: natural history data used as a historical control group ⁴⁸¹.
• Lytenava (bevacizumab-vikg): disease natural history substantiated the NORSE 2 treatment effect. The reviewer cautioned that it is not an appropriate external control for new therapies ⁴⁸².
• Duvyzat (givinostat): reviewers agreed that natural history control analyses could contribute confirmatory evidence for Study 48 ⁴⁸³.
• Aurlumyn (iloprost): two observational studies against standard-of-care and natural history comparators ⁴⁸⁵.

Skyclarys (omaveloxolone) submitted propensity-matched comparisons with the FA-COMS natural history study as confirmatory evidence. The passage reviewed does not explicitly state final acceptance of those comparisons ⁴⁸⁴. Likewise, a registry comparison submitted for Yartemlea is described as confirmatory, but no final acceptance conclusion was found ^463470474.

Evidence reviewers rejected or discounted

These negative findings are as useful for development planning as the accepted evidence.

• Xenpozyme: reviewers concluded that the nonclinical findings did not rise to confirmatory evidence. They discounted an underpowered mouse toxicology study ⁴⁶⁰⁴⁶¹.
• Iwilfin: reviewers did not accept supportive clinical studies in related populations. The reasons were small numbers, difficulty isolating eflornithine's effect and no attributable objective responses ⁴⁴⁰⁴⁴¹.
• Veopoz: the CD55-knockout mouse was unsuitable as a disease model because it does not show the CHAPLE phenotype ³³³.
• Lamzede: serum oligosaccharides counted as pharmacodynamic confirmation but were not treated as a validated surrogate endpoint ²⁹¹²⁹².
• Livdelzi: FDA did not accept future, post-approval AFFIRM results as the confirmatory evidence needed for approval. AFFIRM was instead intended to verify clinical benefit after accelerated approval ⁴²²⁴²³.
• Lynavoy: reviewers rejected a higher-dose crossover study and evidence from other IBAT inhibitors. They gave an open-label study limited utility and found that the mechanistic and pharmacodynamic evidence did not constitute confirmatory evidence ³¹⁰³¹¹.

Practical takeaways for development programs

• The single investigation need not be a randomized placebo-controlled trial. FDA accepted externally controlled, historically controlled and single-arm designs as the one adequate and well-controlled investigation:

  • Prograf ⁸⁹
  • Nulibry ³⁴⁵
  • Iwilfin ¹⁵⁸
  • Wainua ¹⁶⁴
  • Veopoz ³³⁰
  • Avlayah ²⁴³

  In these files, the confirmatory evidence usually had to be mechanistically strong.

• Mechanistic confirmation works best with a well-defined etiology. It was most persuasive where the disease etiology is well established and a pharmacodynamic biomarker responds in an exposure-dependent way ^287289346.

• Class and related-product evidence depends on the class. Reviewers accepted it for ROS1 inhibitors ⁴⁸⁹ and for inotersen ¹⁶⁷, but rejected it for other IBAT inhibitors in Lynavoy ³¹⁰. The difference turns on shared mechanism and consistency of effect ¹⁹¹.

• Confirmatory evidence must exist at the time of approval. Planned post-approval studies did not qualify ⁴²²⁴²³.

• Plan early with FDA. Guidance expects sponsors to agree on the type and quantity of confirmatory evidence early, ideally before the single trial begins ²¹³²¹⁴.