FDA Evaluation of Dystrophin Expression and Functional Endpoints in Duchenne Muscular Dystrophy

How FDA evaluates dystrophin expression and functional endpoints in Duchenne muscular dystrophy

FDA has used two evidentiary routes for Duchenne muscular dystrophy (DMD) products. The first is accelerated approval based on an increase in dystrophin (or an engineered micro-dystrophin) as a surrogate endpoint. The four exon-skipping antisense oligonucleotides and the original Elevidys gene therapy approval all took this route. The second is traditional approval based on a statistically significant effect on a clinical outcome assessment such as muscle strength, time to stand, or 4-stair climb. Deflazacort, vamorolone and givinostat took this route. Across the record, the contested questions have stayed the same: how much dystrophin is enough, whether the assay and controls can be trusted, and whether effort-dependent functional tests can be read against external or historical controls.

Key takeaways

• FDA guidance calls dystrophin a biomarker of particular interest and a potential surrogate endpoint for accelerated approval. It is not an automatically validated one, and its use requires sufficient scientific evidence and acceptable analytical methods ⁵⁶.
• Exon-skipping approvals rested on small increases in truncated dystrophin: about 0.2 to 0.3 percentage points of normal for eteplirsen, 0.92 points for golodirsen, a 0.594-point placebo-adjusted difference for casimersen, and about 5.3 to 5.4 points for viltolarsen ^{111112113257142323324}.
• FDA has repeatedly said that the amount of dystrophin needed for clinical benefit is unknown ^259144328. One review states that the regulations set no numerical threshold for the "reasonably likely" judgment ^114115116.
• For micro-dystrophin, CBER reviewers said that measuring an engineered transgene product shows expression, not necessarily benefit. The randomized data did not show a clear link between expression and North Star Ambulatory Assessment (NSAA) outcomes ^{287288289290291}.
• In several of these files the reviewers and the final decision-maker disagreed. The Center Director overruled the review team on eteplirsen ⁴¹⁰⁴¹¹ and on the original Elevidys approval ⁴⁶³⁴⁶⁴, and the CBER Center Director again reached a different conclusion from reviewers on the 2024 Elevidys supplement ¹⁶⁸¹⁶⁹.
• Traditional approvals each rested on a single prespecified functional primary endpoint, and for givinostat FDA also relied on confirmatory evidence after a missing-data sensitivity analysis lost significance. FDA was explicit when secondary functional endpoints did not survive multiplicity adjustment ^{225226194351354}.

FDA's guidance framework

FDA's DMD guidance treats dystrophin deficiency as the apparent driver of the disease's symptoms and functional consequences. That is the basis for considering dystrophin a potential surrogate ⁵⁶. Baseline dystrophin may be prognostic. An increase may indicate drug activity and help guide dose or route selection ⁵⁶. Interpretation may depend on the size of the change, the tissue sampled, disease stage or age, and patient factors. Sponsors are expected to demonstrate analytical validity where possible, assess assay performance, and document quality control. Trials with clinically meaningful endpoints should also collect relevant biomarkers so that biomarker-outcome correlations can be established ⁷.

The guidance does not prescribe a fixed set of clinical outcome measures. Instead it asks sponsors to:

• agree on endpoints with FDA early;
• assess multiple endpoints when feasible;
• choose measures that capture improvement as well as decline without floor or ceiling effects;
• cover skeletal, respiratory and cardiac function even if only one of them is primary ⁴¹⁹⁴²⁰.

The guidance gives these stage-specific pointers:

Exon-skipping antisense oligonucleotides: dystrophin as the accelerated approval surrogate

Eteplirsen (Exondys 51): the precedent-setting case

Assays and early data. FDA evaluated dystrophin in skeletal-muscle biopsies using two methods ^104105106:

• Immunofluorescence/immunohistochemistry (IHC) to count dystrophin-positive fibers and show where the protein was located.
• Western blot to quantify protein against a normal-muscle reference.

RT-PCR confirmed exon 51 skipping. FDA said that result alone did not establish efficacy ^104105106.

In Study 201/202, adequately controlled IHC analyses showed no evidence of increased dystrophin through week 48 ^107108109110. At week 180:

• Western blot showed a mean of 0.93% ± 0.84% of normal. The sponsor reported 0.08% in untreated controls, but FDA cautioned that the controls and assay limits complicated the comparison ^107108109110.
• A revised IHC method showed 17.4% positive fibers versus 1.1% in baseline/external-control samples. FDA found these results hard to interpret because the old and new methods differed sharply and IHC did not agree with Western blot ^107108109110.

The pivotal biomarker data. FDA asked for additional Western blots from Study 301 (PROMOVI), which had paired biopsies taken at baseline and at week 48 ^111112113. Usable results were available for 12 of 13 patients. The mean increase was about 0.2 to 0.3 percentage points of normal and was statistically significant. In one imputation approach the increase was 0.218 points, from 0.260% to 0.478% ^111112113. Reviewers accepted that dystrophin sits on DMD's causal pathway. Whether this small amount was reasonably likely to predict benefit was treated as a judgment about effect size, with no regulatory numerical threshold ^114115116.

Functional endpoints. FDA found no demonstrated functional benefit:

• The 6MWT showed no significant effect versus placebo through week 24, and no advantage in the prespecified week 48 extension analysis ⁸⁰⁸¹.
• The sponsor reported a 141-meter difference at 36 months versus historical controls. FDA said the nominal p-value was not meaningful because the comparison was open-label and not adequate and well-controlled ⁸².
• FDA cited differences in baseline severity and steroid use; expectation, motivation and coaching effects on walking; unreliable "zero" 6MWT distances in some controls; and unequal testing opportunities between treated and natural-history patients ⁸³⁸⁴⁸⁵.
• The review team concluded that the treated patients' course was not clearly distinguishable from natural history ⁸⁶.

Advisory committee. At the April 25, 2016 Peripheral and Central Nervous System Drugs Advisory Committee meeting on NDA 206488 ⁵²⁸, members answered two questions ^486487488489:

Members questioned biopsy controls, differences in sampled muscles, the lack of a dose-response, and the mismatch between dystrophin levels and clinical results ⁴⁹⁰. The review noted a disconnect between supportive patient testimony and formal functional testing, which showed worsening ⁴⁹⁴.

Decision and confirmatory obligation. The review division, supported by the Office of New Drugs Director, favored a complete response. The CDER Director concluded that the surrogate evidence supported accelerated approval and overruled the review team ⁴¹⁰⁴¹¹. The review team appealed through the scientific-dispute process. The Commissioner's office found no procedural basis to overturn the decision ⁴¹²⁴¹³. The postmarketing requirement was a two-year randomized, double-blind trial comparing 30 mg/kg weekly with a substantially higher-exposure regimen (for example, 30 mg/kg daily), with NSAA as the primary endpoint ⁴¹⁴.

Golodirsen (Vyondys 53)

In Study 4053-101 (25 patients), Western blot dystrophin rose from 0.10% to 1.02% of normal after 48 to 59 weeks, an increase of 0.92 percentage points (p<0.001). FDA considered the method adequate ²⁵⁷²⁵⁸. The accelerated approval rationale relied explicitly on the eteplirsen precedent: the increase was similar in magnitude to the one CDER had judged reasonably likely to predict benefit. FDA acknowledged that the dystrophin level needed for meaningful benefit was unknown ²⁵⁹. ESSENCE (Study 4045-301) was designated to confirm benefit ²⁶⁰²⁶¹.

The golodirsen file shows the surrogate's effect size being weighed against risk. FDA issued a Complete Response Letter on August 19, 2019. The letter cited vascular-access infections and potential renal toxicity, and stated that the small, unverified benefit (about a 0.9% increase in dystrophin) did not outweigh those risks ⁴³⁷⁴³⁹. FDA granted Sarepta's appeal on November 22, 2019. The deciding official concluded that clinically meaningful benefit was reasonably likely and that renal function could be monitored, including with a baseline direct GFR measurement. Accelerated approval followed on December 12, 2019 ^440441442.

Viltolarsen (Viltepso)

In Study NS-065/NCNP-01-201 (16 boys, 8 per dose), FDA used Western blot normalized to myosin heavy chain as the primary assay. FDA considered this normalization more reliable than α-actinin because it was less variable. Validated mass spectrometry served as a supportive assay ^320321322.

After about 24 weeks, dystrophin rose from 0.3% to 5.7% of normal at 40 mg/kg/week and from 0.6% to 5.9% at 80 mg/kg/week. There was no significant difference between doses ^{323324325326327}. Mass spectrometry showed smaller mean changes of 1.5 and 3.7 points ^{323324325326327}. That gap between orthogonal assays is a useful reminder that "percent of normal" depends on the method.

Accelerated approval relied partly on the prior eteplirsen and golodirsen conclusions. The designated confirmatory study used Time to Stand as its primary endpoint ³²⁸³²⁹.

Casimersen (Amondys 45)

Casimersen is the only exon-skipping approval in this set whose dystrophin evidence came from a randomized, placebo-controlled comparison. In an interim analysis of Study 4045-301, mean Western blot dystrophin increased by 0.811% of normal with casimersen versus 0.217% with placebo, a difference of 0.594 points (p=0.004) ^140141142143. FDA treated the increase as reasonably likely to predict benefit while acknowledging uncertainty about the amount required. Clinical benefit was to be confirmed through ESSENCE outcomes, including the week 96 6MWT ¹⁴⁴¹⁴⁵.

Comparison of exon-skipping dystrophin evidence

Gene therapy: micro-dystrophin and Elevidys

Framing at the advisory committee

At the May 12, 2023 advisory committee meeting, FDA presented week 12 micro-dystrophin expression as Sarepta's proposed surrogate. The question for the committee was whether it was reasonably likely to predict benefit. FDA said that judgment needed three things: biological plausibility, empirical evidence, and clinical data linking expression to an outcome such as NSAA ³⁷³⁸. FDA stressed that the engineered protein differs from naturally occurring shortened dystrophins ³⁷³⁸.

On function, FDA made three points:

• Study 102 Part 1, the only randomized, double-blind, placebo-controlled comparison, did not show a significant NSAA effect ³⁹.
• Exploratory analyses suggested a possible benefit in ages 4 to 5 but none in ages 6 to 7 ³⁹.
• Open-label and external-control NSAA comparisons were unreliable, because NSAA is effort- and administration-dependent, disease progression varies between patients, and the matching methods and covariates raised concerns ⁴⁰⁴¹.

The committee voted 8 to 6 that overall benefit-risk supported accelerated approval in ambulatory patients using week 12 micro-dystrophin as the surrogate ⁴².

The review record and the June 2023 decision

Expression data. In Study 102 Part 1, Western blot showed significantly greater expression with Elevidys than with placebo (p<0.0001) ^287288289.

Link to function. Reviewers gave greatest weight to the randomized data because NSAA is effort-dependent ²⁹⁰²⁹¹. The findings were:

• After adjustment for age and baseline, there was no clear relationship between expression and NSAA change ²⁹⁰²⁹¹.
• In pooled analyses, expression explained about 11% of the variation in NSAA change ²⁹⁰²⁹¹.
• The week 48 NSAA treatment difference was 0.82 points (95% CI −1.03 to 2.67) ²⁹².

Recommendations and decision. The clinical and clinical-pharmacology reviewers recommended against approval, even for a restricted 4- to 5-year-old group ²⁹³²⁹⁴. An integrated assessment found the 4- to 5-year-old results sufficient for accelerated approval given the totality of evidence and unmet need ²⁹⁵²⁶⁷. CBER Center Director Peter Marks overrode the review committee's recommendation ^463464465. On June 22, 2023, FDA granted accelerated approval for ambulatory children aged 4 through 5 years with a confirmed DMD mutation ⁴⁶²⁴⁴⁸. The required confirmatory study was Part 1 of Study SRP-9001-301 (EMBARK) ⁴⁶⁶⁴⁶⁷.

EMBARK and the 2024 supplement

EMBARK missed its primary endpoint. The week 52 NSAA difference was 0.65 points (p=0.244) ^164165166167. Time to rise from the floor and the 10-meter walk numerically favored Elevidys. Because the primary endpoint failed, the statistical reviewer made no inferential claims on those secondary endpoints ^164165166167. Reviewers also cited the lack of prespecified multiplicity control and inconsistent 10-meter walk results across studies, and they recommended a Complete Response ^164165166167.

The CBER Center Director acknowledged the failed primary endpoint. He judged the secondary and exploratory functional findings, together with other studies, to verify clinical benefit in ambulatory patients ¹⁶⁸¹⁶⁹. The June 2024 action expanded the indication to patients aged 4 years and older with a confirmed DMD mutation, with a different approval pathway for each population ^170171172:

• Ambulatory patients: traditional approval.
• Non-ambulatory patients: accelerated approval, based on micro-dystrophin expression, with a confirmatory trial underway.

The label also contraindicates use in patients with deletions in exon 8 and/or exon 9 ^170171172.

This analysis covers the review record through the June 2024 Elevidys action. Later events, including confirmatory-trial readouts for the exon-skipping products and FDA's 2025 safety actions on Elevidys after reports of fatal acute liver failure, are not analyzed here. Check current labeling and FDA communications before relying on the regulatory status of any product described above.

Traditional approvals on functional endpoints

The givinostat review shows how FDA handles a single positive study. A missing-data sensitivity analysis lost statistical significance ^355356354. FDA still found effectiveness on the 4-stair climb result plus confirmatory evidence: less increase in vastus lateralis fat fraction on MRI and longer-term comparisons with natural history ^355356354.

Cross-cutting themes for development programs

1. Dystrophin magnitude remains a judgment call. Each exon-skipping approval acknowledged that the dystrophin level needed for benefit is unknown ^259144328. Later approvals relied on the eteplirsen precedent rather than on a validated threshold ²⁵⁹³²⁸.
2. Assay rigor and controls drive credibility. FDA discounted non-concordant IHC and Western blot data and data from changing methods ^107108109110. It preferred Western blot with less variable normalization ^320321322. It valued paired-biopsy or placebo-controlled designs ^{111112113142143}.
3. Engineered proteins raise a separate surrogate question. For micro-dystrophin, FDA asked for clinical evidence that expression tracks function, not just evidence that expression occurred ^{3738287288289}.
4. External and historical controls for effort-dependent endpoints face heavy scrutiny. FDA's critiques of the 6MWT and NSAA in the eteplirsen and Elevidys files centered on motivation and coaching effects, baseline imbalances, matching methods and testing asymmetries ^8384854041.
5. Multiplicity discipline is applied consistently in reviews. Secondary functional endpoints that fail adjustment are labeled as such, even when they numerically favor the product ^{354164165166167}.
6. Final decisions have diverged from review-team recommendations. This happened on eteplirsen ^410411412413, the original Elevidys approval ⁴⁶³⁴⁶⁴ and the Elevidys expansion ¹⁶⁸¹⁶⁹. Sponsors should expect the record to document dissent in detail.

Questions to explore further

Topics worth following up include:

• the ESSENCE and viltolarsen confirmatory-trial results and their regulatory consequences;
• the status of the eteplirsen dose-comparison confirmatory trial;
• FDA's post-2024 safety and labeling actions on Elevidys;
• the confirmatory trial for non-ambulatory Elevidys patients;
• how FDA treated MRI fat fraction and other imaging biomarkers as confirmatory evidence across DMD programs.