FDA's Rare Disease Evidence Principles and Evidentiary Flexibility for Ultra-Rare Disease Programs

FDA's Rare Disease Evidence Principles (RDEP) and evidentiary flexibility for small-population and ultra-rare disease programs

FDA has not created a lower approval standard for rare diseases. Its rare disease guidance states that the statutory effectiveness requirement is the same for rare and common diseases, and that the Orphan Drug Act does not create a different approval standard for orphan drugs ¹¹⁸²⁷⁶. What FDA has done is define, with growing specificity, how sponsors can meet the substantial evidence standard when patient numbers make conventional development programs impractical. The Rare Disease Evidence Principles (RDEP), announced in September 2025, is the most targeted of these mechanisms. It sits within a wider set of guidance, pilot programs and review practices covering single-trial approvals, natural history and external controls, surrogate endpoints, innovative trial designs and individualized therapies.

What RDEP is

FDA announced RDEP on September 3, 2025, as a review process intended to give sponsors clearer direction on how to demonstrate substantial evidence of effectiveness for therapies that target rare diseases with very small patient populations, significant unmet need and a known genetic defect ¹⁹. CDER and CBER developed and implement the process jointly ²³²⁴.

FDA's stated rationale is that in very small rare disease populations, running multiple traditional clinical trials can be difficult or impossible while still meeting the statutory evidence standard. RDEP is meant to offer a rigorous, viable alternative and more predictability about the supportive evidence FDA will consider ¹⁹²⁵. FDA also described traditional designs, including placebo-controlled studies, as especially difficult in very small populations ^134135136.

Eligibility criteria

Evidence standard under RDEP

Under RDEP, approval may be based on one adequate and well-controlled study plus robust confirmatory evidence. FDA was explicit that this does not mean approval on a single trial alone ²¹. Examples of confirmatory evidence FDA cited include:

• mechanistic or biomarker evidence, including effects on the disease's direct pathophysiology ²¹¹³⁶
• data from relevant nonclinical models ²¹¹³⁶
• clinically significant pharmacodynamic data ²¹¹³⁶
• case reports and expanded access data ²¹¹³⁶
• natural history studies ²¹

Products approved through this route may carry additional postmarketing requirements to further establish safety and effectiveness ²³²⁴.

Process and timing

A sponsor requests RDEP review as part of a formal meeting request, and the request must be made before the pivotal trial begins ²². Sponsors accepted into the process then meet with the relevant FDA review team to determine which data will substantiate safety and effectiveness, and FDA encourages patient and expert input throughout ^22137138. The announcement did not set a fixed review or approval timeline ¹²³.

What RDEP does not do

• RDEP is separate from orphan drug designation. Participation neither confers orphan designation nor affects eligibility for it ²⁴¹³⁹.
• RDEP does not change the statutory standard. It clarifies how the existing "one adequate and well-controlled investigation plus confirmatory evidence" route applies to a narrowly defined population ²¹¹³⁴.
• Public reporting through late 2026 has not disclosed the number of RDEP requests received, programs accepted, or products approved under the process ¹²³¹²⁷. No CDER MAPP or CBER SOPP describing RDEP procedures has been identified in FDA's published procedure documents; the publicly available procedural detail comes from FDA's announcement ⁶¹.

FDA positioned RDEP as building on and complementing its earlier Accelerating Rare Disease Cures (ARC) program, the START pilot and the Rare Disease Endpoint Advancement (RDEA) pilot ¹⁵².

The legal and policy foundation: one trial plus confirmatory evidence

RDEP applies a principle FDA has articulated for decades. FDA's final guidance Providing Clinical Evidence of Effectiveness for Human Drug and Biological Products (May 1998) explains that substantial evidence ordinarily rests on independent substantiation, but that the statute permits FDA to rely on one adequate and well-controlled investigation plus confirmatory evidence. A single favorable study is not automatically sufficient; its design and results must be persuasive, and conflicting studies matter ^194195196197.

Two later drafts expand on this:

• Demonstrating Substantial Evidence of Effectiveness With One Adequate and Well-Controlled Clinical Investigation and Confirmatory Evidence (draft, September 2023) states that FDA assesses the investigation and the confirmatory evidence together, case by case. A highly persuasive investigation may need less confirmatory support; a less persuasive one needs more compelling corroboration. Sponsors should discuss the approach early ¹⁷⁴¹⁷⁰.
• Demonstrating Substantial Evidence of Effectiveness for Human Drug and Biological Products (draft, June 24, 2026) lists the factors FDA weighs for a single-trial approach: trial design, conduct, analysis and persuasiveness; the source and strength of corroboration; disease seriousness, unmet need and prevalence; and whether another trial would be ethical and practicable. It generally expects either a highly persuasive trial or strong confirmatory evidence ¹⁷². For rare diseases, it acknowledges that smaller trials yield less precise estimates and stresses interpretable design (randomization, blinding, appropriate endpoints, covariate adjustment, adequate duration), while allowing scientifically sound flexibility in the source and strength of confirmatory evidence ¹¹¹. It recommends discussing a proposed single-trial approach with FDA before that trial begins, and no later than the end-of-phase 2 meeting ³¹⁵³¹⁶.

Types of confirmatory evidence FDA has described

A RAPS report from April 2026 quoted an FDA official describing a newer default of one well-designed clinical trial when it provides sufficient evidence, noting that single-trial evidence was already used in rare disease development and that the newer approach extends beyond it ¹⁶². FDA's June 2026 draft guidance describes both single-trial and multiple-trial approaches without stating a universal default ¹⁷²³¹⁴.

Flexibility in rare disease development guidance

The final guidance Rare Diseases: Considerations for the Development of Drugs and Biological Products (December 2023) is the core reference for program design. Its key flexibilities:

• Trial design. FDA may consider design flexibility case by case, particularly where natural history is predictable and the treatment has a large effect on an objectively and reliably measured endpoint, while generally recommending a randomized concurrent control ¹⁰². The guidance describes Bayesian methods, n-of-1 trials, delayed-start and crossover designs, and master protocols, and notes that modified designs can reduce placebo exposure ¹⁰⁷.
• Endpoints. Natural history studies can inform disease course and endpoint selection ¹¹⁹. A novel clinical outcome assessment may be needed where existing measures are infeasible, with early FDA discussion recommended ¹⁰⁰.
• Safety database. Exposure expectations developed for chronic drugs in non-life-threatening conditions do not apply to many life-threatening rare diseases, and fewer exposed patients may be acceptable when the intended population is small ¹²².
• Clinical pharmacology. The draft Rare Diseases: Early Drug Development and the Role of Pre-IND Meetings (October 2018) says FDA can be flexible about which clinical pharmacology studies are essential, whether they are integrated into efficacy and safety trials, and which can be deferred until after approval. It also supports modeling and simulation to inform dosing and trial design ¹²¹¹⁰⁵.

FDA's expedited programs guidance also states that FDA applies flexibility to the particular development challenges of rare diseases ²⁷⁷. Disease-specific guidance can make the single-trial expectation explicit: the draft Mucopolysaccharidosis Type III guidance states that, given the disease's rarity, one adequate and well-controlled trial showing a clinically meaningful effect plus confirmatory evidence can support approval ²⁶⁵.

Natural history data and external controls

FDA treats natural history as valuable but not as a substitute for a credible comparison. Randomized concurrent controls remain the preferred approach for separating a drug effect from disease progression ¹⁰².

• The December 2023 rare disease guidance advises sponsors to assess the quality of existing natural history knowledge early and favors prospectively designed, protocol-driven natural history studies ²⁵⁴. External or historical controls are usually reserved for well-characterized natural history, predictable mortality or progressive morbidity, and a large, self-evident drug effect ²⁵⁷²⁵⁸.
• The draft Rare Diseases: Natural History Studies for Drug Development (March 2019) describes natural history studies as useful for planning controlled trials and potentially supplying external control data, and identifies concurrent external data as a stronger design than past patient-level data ²⁵⁵²⁵⁶.
• The final Human Gene Therapy for Rare Diseases (January 2020) allows a single-arm trial with historical controls when randomization is infeasible, but only if treated and control populations are adequately matched on disease state, demographics, concurrent treatment and other factors ²⁵⁹²⁶⁰.
• The draft Considerations for the Design and Conduct of Externally Controlled Trials for Drug and Biological Products (February 2023) confirms that an appropriate externally controlled trial can serve as an adequate and well-controlled investigation ²⁶¹²³². It warns that unmeasured confounding and lack of blinding cannot be eliminated, and recommends prespecifying the external control and analysis before the trial rather than selecting a comparator after seeing single-arm results ^262263233235. Historical comparisons are especially vulnerable to changes in diagnosis and standard of care ^238239264.

Accelerated approval and surrogate endpoints

Accelerated approval allows reliance on a surrogate or intermediate endpoint reasonably likely to predict clinical benefit for serious conditions where a drug offers a meaningful advantage over available therapy ^363364365. For rare diseases, sponsors may consider surrogate endpoints when a clinical endpoint changes too slowly to assess in a trial of reasonable duration, and natural history data, animal models, patient data and early pharmacodynamic data can help support a proposed biomarker ³⁶⁶³⁵¹. The rare disease gene therapy guidance also states that sponsors may consider accelerated approval based on a surrogate endpoint ¹²⁰³⁶⁷.

The draft Accelerated Approval: Expedited Program for Serious Conditions (December 2024) adds rare disease-specific nuance. Where trial evidence linking surrogate changes to clinical outcomes is limited, FDA may weigh converging evidence from animal models, epidemiology and clinical data, and in certain circumstances compelling nonclinical evidence may support a rare disease surrogate without clinical data establishing that relationship ^347368348369. Pharmacologic activity alone is insufficient ³⁴⁷. The draft identifies the RDEA pilot as an option for collaborating on novel surrogate or intermediate endpoints ³⁷⁰. Confirmatory postapproval studies remain required, and FDA can specify protocol, enrollment and completion milestones by the time of approval ^371372373.

Cell and gene therapy in small populations

The draft Innovative Designs for Clinical Trials of Cellular and Gene Therapy Products in Small Populations (September 24, 2025) describes several design options ³²¹¹:

• Disease progression modeling of natural history using biomarker trajectories, clinical endpoints and patient heterogeneity, with caveats about variable progression, evolving standard of care and model validation in underrepresented subgroups ³²¹.
• Externally controlled studies using historical or real-world data, with suitability judged case by case ³²¹.
• Single-arm trials with participants as their own control, comparing outcomes with a reliably established baseline, while noting that fluctuating disease and regression to the mean can confound interpretation ³.
• Adaptive designs, including early stopping, sample-size reassessment, enrichment and dose selection followed by confirmation in one study, all requiring rigorous prespecification ³²³.
• Bayesian designs that can incorporate external clinical data, including existing control data to augment a concurrent control group ³⁴².

On the CMC side, the final Human Gene Therapy for Neurodegenerative Diseases (October 2022) advises sponsors whose small early-phase studies may supply effectiveness evidence to establish critical quality attributes and controls early, and to consider multiple small lots to build process experience ²²². The draft Leveraging Prior Knowledge in the Development of Human Gene Therapy Products Incorporating Genome Editing (June 2026) describes scientifically justified reuse of prior CMC, nonclinical and clinical knowledge to reduce new data requirements ^223224225.

Ultra-rare and individualized therapies

For the smallest populations, FDA has issued guidance aimed at therapies developed for one or a few patients.

Individualized antisense oligonucleotides (ASOs). Two 2021 drafts address ASOs for severely debilitating or life-threatening genetic diseases where the targeted variant is generally reported in only one or two patients ²⁶⁵⁷. The nonclinical draft (April 2021) recommends a less extensive first-in-human safety package than for broader-use ASOs, offset by convincing in vitro or in vivo proof of concept, with off-target binding assessment and route-appropriate safety pharmacology ³²⁶⁰. FDA cautions that abbreviated testing may miss toxicity and that larger populations or commercialization typically warrant additional studies ⁴¹. The clinical draft (December 2021) covers IND content, dosing and monitoring, and asks sponsors to document the variant's rarity ^26583659.

Plausible Mechanism Framework. On February 23, 2026, CDER and CBER issued the draft Considerations for the use of the Plausible Mechanism Framework to Develop Individualized Therapies that Target Specific Genetic Conditions with Known Biological Cause ²⁹³³³. It addresses evidence of effectiveness and safety for genome-editing and RNA-based therapies when randomized trials are infeasible and data will come from a limited number of patients ^29329433. Key elements:

• FDA anticipates that one adequate and well-controlled investigation with confirmatory evidence could establish effectiveness ²⁰⁴.
• Where untreated natural history can be reasonably characterized, comparing a patient's post-treatment change with baseline or an estimate of the untreated course may be acceptable, particularly when the effect is rapid, dramatic and unlikely to occur spontaneously ⁴⁴²²⁷.
• Sponsors should confirm the diagnosis and show that the targeted variant is causal and rare ⁴⁰.
• FDA described possible single applications and master protocols covering related genome-editing variants ^295159160.
• The choice between traditional and accelerated approval, and specific endpoints, is left to discussion with the review division ²²⁸.

FDA later clarified that it had proposed a framework operating within existing regulatory pathways, not a new pathway ¹⁶¹.

Engagement programs that support evidence planning

Flexibility in recent approvals

Recent FDA approval announcements show how these principles are being applied in specific products:

• Waskyra (Wiskott-Aldrich syndrome gene therapy, December 2025): FDA cited appropriate use of expanded access data and flexibility in trial design, mechanism of action and manufacturing review ²⁹⁶²⁹⁷.
• Avlayah (neurologic manifestations of Hunter syndrome, March 2026): accelerated approval based on a surrogate endpoint reasonably likely to predict benefit, with a clinical benefit trial underway ²⁹⁸²⁹⁹.
• Kresladi (severe leukocyte adhesion deficiency type I gene therapy, March 2026): evidence from one open-label, single-arm study, with accelerated approval on a surrogate endpoint subject to confirmatory studies. FDA said it considers small trial populations and all available sources of evidence in rare diseases ^300301302.
• Otarmeni (genetic hearing loss gene therapy, April 2026): evidence from a single ongoing single-arm trial supported by confirmatory mechanistic nonclinical data and sustained protein expression, approved under the National Priority Voucher program ³⁴⁵³⁴⁶.

These announcements are product-specific examples and do not identify any of these products as RDEP approvals.

Open questions and stakeholder concerns

Stakeholders have welcomed RDEP but questioned consistency in rare disease reviews. At a Senate hearing reported on February 26, 2026, an EveryLife Foundation representative said FDA decisions on rare disease applications appeared inconsistent with its stated commitment to flexibility. She cited hesitation over surrogate endpoints, natural history studies, external controls and real-world evidence, and reported at least 23 complete response letters for rare disease therapies submitted under accelerated approval since the start of 2025 ¹²⁷¹⁴⁰. These criticisms concerned FDA's broader rare disease reviews, not RDEP decisions ¹⁴⁰.

Several practical questions remain open in the public record: how many programs have entered RDEP, how the process will interact with the Plausible Mechanism Framework and the Rare Disease Framework Initiative, and whether the October 2026 initiative will produce changes to evidentiary standards, endpoints, statistics, trial design or postapproval requirements ^123163164165.

Practical implications for sponsors

• Assess RDEP eligibility before the pivotal trial. The request window closes once the pivotal study starts, and it is made through a formal meeting request ²².
• Plan confirmatory evidence deliberately. Under both RDEP and FDA's general single-trial guidance, the persuasiveness of the trial and the strength of corroborating mechanistic, pharmacodynamic, nonclinical, natural history or expanded access data are weighed together ^21172174.
• Invest early in natural history. Prospective, protocol-driven natural history data support endpoint selection, surrogate justification and any external control ^254119366.
• Prespecify any external control and analysis. FDA expects these choices to be fixed before results are known ²³⁵.
• Use the engagement programs. START, RDEA and the Rare Disease Innovation Hub provide additional channels for design and endpoint discussions ^391370291.
• Treat flexibility as case-specific. FDA's guidance consistently frames these approaches as case-by-case, with randomized concurrent controls still preferred where feasible ¹⁰²¹¹¹.