← Back
NeuroTrials.ai
Neurology Clinical Trial Database

DMD-ELEVIDYS-MA

Duchenne Muscular Dystrophy and Delandistrogene Moxeparvovec Gene Therapy in Children: A Systematic Review and Meta-Analysis

Year of Publication: 2026

Authors: Antonello BB, Cargnelutti Fontoura F, Braga Albuquerque AL, ..., Zattar Ribeiro PV

Journal: Neurology Genetics

Citation: Neurol Genet. 2026 Jul;12(4):e200408.

Link: https://doi.org/10.1212/NXG.0000000000200408

Bottom Line

Pooled 1-year data from 4 studies (302 boys) show that delandistrogene moxeparvovec produces small but statistically significant improvements in NSAA, time to rise, and dystrophin expression versus standard treatment — with a modest ~0.7-point NSAA benefit in the RCT-only subset — accompanied by frequent gastrointestinal and hepatic adverse events and rare but serious hepatotoxicity, myocarditis, and post-marketing fatal acute liver failure.

Major Points

  • Systematic review/meta-analysis (PROSPERO CRD42025635605) of PubMed/Embase/Cochrane through Jan 2025; included 4 studies (2 RCTs: EMBARK, Study 102; 2 non-randomized: ENDEAVOR, Study 101) enrolling 302 ambulatory boys aged 4 to <8 y with confirmed DMD on ≥12 wk oral corticosteroids.
  • Primary NSAA analysis (all data): MD +2.48 points (95% CI 0.15 to 4.81; I²=98.9%; p=0.04) favoring gene therapy; LSMD-only sensitivity +1.52 (−0.04 to 3.08; p=0.06) tended toward but did not meet significance; RCT-only LSMD +0.70 (0.48 to 0.92; I²=0%; p<0.01) is the most robust estimate.
  • Time to rise from supine improved with therapy: MD −0.85 s (95% CI −1.23 to −0.47; I²=96.4%; p<0.01) overall, and LSMD-only −0.90 s (−1.4 to −0.40; p<0.01).
  • 10-meter walk/run overall not significant (MD −0.28 s; 95% CI −1.23 to 0.66; p=0.56) but LSMD-only leave-one-out sensitivity showed significant reduction of −0.71 s (−1.29 to −0.14; p=0.02).
  • Muscle dystrophin content by Western blot at 12 weeks increased by 28.39% of normal (95% CI 15.39 to 41.39; I²=0%; p<0.01) versus baseline/controls.
  • Non-randomized studies (ENDEAVOR, Study 101) reported larger treatment effects than RCTs (EMBARK, Study 102), likely reflecting external-cohort selection and reduced blinding; heterogeneity was very high (I² ≈96–99%) for most outcomes.
  • Safety was manageable in trials (95–100% AEs, mostly mild GI or transaminase elevations; rare rhabdomyolysis, myocarditis, hepatotoxicity), but post-marketing surveillance has reported at least 3 US deaths from acute liver failure within 2 months of treatment and 1 fatal case in Brazil, prompting ongoing FDA safety review of AAVrh74-platform therapies.
  • Delandistrogene is contraindicated with DMD exon 8/9 deletions (immune-mediated myositis risk); pre-treatment anti-AAVrh74 antibody testing is required, and post-treatment antibody development may preclude future AAV-based gene therapies.

Design

Study Type: Systematic Review and Random-Effects Meta-Analysis

Randomization:

Blinding: N/A (meta-analysis; included studies: 2 double-blind RCTs and 2 open-label non-randomized cohorts)

Enrollment Period: Searches from database inception through January 2025

Follow-up Duration: 48–52 weeks (standardized to 1 year); ENDEAVOR 1 y, EMBARK 2 y, Study 101 5 y, Study 102 2 y

Centers: Multiple (across 4 included studies)

Countries: United States, Europe, Asia (multinational EMBARK)

Sample Size: 302

Analysis: Random-effects inverse-variance meta-analysis using R meta package (metacont), REML between-study variance; primary summary measure mean difference / least-squares mean difference in change from baseline; heterogeneity by I² and Cochran Q (I² >25% or p<0.10 = significant); leave-one-out and LSM-only sensitivity analyses; RCTs-only subgroup analysis; ROBINS-I for non-randomized and RoB 2 for RCTs; funnel-plot for publication bias; R v4.2.3.


Inclusion Criteria

  • Randomized controlled trials or non-randomized cohort studies
  • Compared delandistrogene moxeparvovec with placebo or propensity-matched external natural-history cohorts
  • Male at birth aged ≥4 to <8 years with molecularly confirmed DMD (variant between exons 18 and 79)
  • On stable dose of oral corticosteroids for ≥12 weeks
  • Follow-up of at least 48 weeks
  • Reported at least one clinical outcome of interest (NSAA, 10MWR, TTR, or dystrophin expression)

Exclusion Criteria

  • Reviews, case reports, case series, case-control studies
  • Preclinical research
  • Ongoing studies or abstracts without published results at time of search
  • Non-ambulatory patients or patients outside 4 to <8 y age window
  • Studies not reporting NSAA, TTR, 10MWR, or dystrophin expression
  • Duplicate/overlapping populations (e.g., Study 102 part 2 was excluded to avoid overlap with part 1 RCT comparison)

Baseline Characteristics

CharacteristicDelandistrogene moxeparvovec (Intervention)Standard treatment (Control)Included studies (per Table 1)
N107195 (placebo in EMBARK & Study 102; propensity-matched external natural-history in ENDEAVOR & Study 101)
Sex100% male at birth100% male at birth
Age range (y)4 to <84 to <8
Mean age (y) — ENDEAVOR5.8 (SD 1.1)
Mean age (y) — EMBARK6.03 (1.05)
Mean age (y) — Study 1015.1 (0.9)
Mean age (y) — Study 1026.26 (1.16)
Corticosteroid useAll on stable oral corticosteroids ≥12 wk pre-treatmentAll on oral corticosteroids
DMD variant locationBetween exons 18 and 79 (exon 8/9 deletions excluded)
Dose1.33 × 10¹⁴ vg/kg (ENDEAVOR, EMBARK) or 2.0 × 10¹⁴ vg/kg (Studies 101, 102) IV × 1
Mean age (y) — ENDEAVOR ENHC6.2 (0.4)
Mean age (y) — EMBARK placebo6.03 (1.05)
Mean age (y) — Study 101 external cohort6.4 (0.3)
Mean age (y) — Study 102 placebo6.26 (1.16)
ENDEAVOR (C.M. Zaidman 2023) [NCT04626674]20 intervention vs 91 external natural-history control; non-randomized; US; 1-y follow-up
EMBARK (J.R. Mendell 2024) [NCT05096221]63 intervention vs 62 placebo; RCT; US/Europe/Asia; 2-y follow-up
Study 101 / SRP-9001-101 (J.R. Mendell 2023a) [NCT03375164]4 intervention vs 21 external cohort; non-randomized open-label; Nationwide Children's Hospital, Ohio; 5-y follow-up
Study 102 / SRP-9001-102 (J.R. Mendell 2023b) [NCT03769116]20 intervention vs 21 placebo (part 1 only); RCT; Nationwide Children's Hospital & UCLA; 2-y follow-up

Arms

FieldDelandistrogene moxeparvovecControl
InterventionSingle IV infusion of AAVrh74-based SRP-9001 microdystrophin gene therapy (1.33 × 10¹⁴ vg/kg in EMBARK/ENDEAVOR; 2.0 × 10¹⁴ vg/kg in Studies 101/102) plus stable oral corticosteroidsPlacebo (EMBARK, Study 102 part 1) or propensity-matched external natural-history cohort (ENDEAVOR, Study 101), all on stable oral corticosteroids
DurationSingle dose; outcomes at 48–52 wkMatched follow-up (48–52 wk for analysis)

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Change from baseline in North Star Ambulatory Assessment (NSAA) total score at 1 year — RCT-only LSMD analysis (most robust)PrimaryStandard treatment (placebo in RCTs)Delandistrogene moxeparvovec<0.01
Time to rise from floor (TTR) — primary pooledSecondary−0.85 s<0.01
Time to rise from floor (TTR) — LSMD-only sensitivitySecondary<0.01
10-meter walk/run (10MWR) — primary pooledSecondary−0.28 s0.56 (NS)
10-meter walk/run (10MWR) — LSMD-only sensitivitySecondary0.02
Muscle dystrophin expression by Western blot (% of normal) at 12 weeksSecondary+28.39%<0.01
Patients with ≥1 AEAdverse95%–100% across studies (ENDEAVOR 95%, EMBARK 98.4%, Study 101 100%, Study 102 100%)
VomitingAdverse54–64% (ENDEAVOR 55%, EMBARK 54%, Study 101 64.2%, Study 102 60%)
Decreased appetiteAdverse14–45% (ENDEAVOR 45%, EMBARK 27%, Study 101 14.3%, Study 102 30%)
NauseaAdverse7–40% (ENDEAVOR 40%, EMBARK 31.7%, Study 101 7.1%, Study 102 30%)
GLDH increasedAdverse23.8–40% (ENDEAVOR 40%, EMBARK 23.8%)
Elevated liver enzymes (transaminases/GGT)Adverse20–40% overall; SAE elevations 1–5%
ConstipationAdverse25% (ENDEAVOR 25%)
PyrexiaAdverse~16% (EMBARK 15.9%)
Abdominal pain upperAdverse13–15% (EMBARK 12.7%, Study 102 15%)
ThrombocytopeniaAdverse15% (ENDEAVOR)
Serious treatment-related AEsAdverseEMBARK: transient liver enzyme elevations 3 (4.75%), hepatotoxicity 1 (1.58%), liver injury 1 (1.58%), myocarditis+pyrexia+nausea+vomiting 1 (1.58%), rhabdomyolysis 1 (1.58%); Study 102: rhabdomyolysis 2 (10%), increased transaminases 1 (5%), liver injury 1 (5%); ENDEAVOR: increased transaminases 1 (5%), vomiting 1 (5%); Study 101: 0
On-trial deathsAdverse0 across all 4 included studies
Post-marketing / real-world fatal eventsAdverse≥3 US deaths from acute liver failure within 2 months post-treatment; 3 Brazilian AE reports in 2025 (1 fatal, attributed by Anvisa to severe influenza A complicated by immunosuppression, unlikely causal link); FDA continues to investigate acute liver failure risk with AAVrh74-platform gene therapies

Subgroup Analysis

Three prespecified NSAA analyses: (1) primary pool of all data (MD +2.48, p=0.04); (2) LSMD-only sensitivity (LSMD +1.52, p=0.06); (3) RCTs-only LSMD (LSMD +0.70, p<0.01, I²=0%) — the most robust. Non-randomized studies (ENDEAVOR, Study 101) generally reported larger effects than RCTs (EMBARK, Study 102). Dose subgroup: 1.33 × 10¹⁴ vg/kg (EMBARK, ENDEAVOR) vs 2.0 × 10¹⁴ vg/kg (Studies 101, 102) — later determined equivalent when re-measured by linear standard qPCR.


Criticisms

  • Very small evidence base — only 4 studies and 302 boys, with just 107 treated; limits precision and generalizability.
  • Extreme statistical heterogeneity (I² ≈96–99%) for NSAA, TTR, and 10MWR pooled analyses, reflecting differences in design, controls, and dosing.
  • Two of 4 studies used external natural-history or historical controls rather than concurrent placebo, introducing confounding and inflating apparent benefit; nonrandomized cohorts (ENDEAVOR, Study 101) drove much of the positive signal.
  • Individual RCTs (EMBARK, Study 102) were NOT independently statistically significant for NSAA at 52 weeks; pooled significance depends on inclusion of nonrandomized data.
  • Mixed use of least-squares mean difference vs mean difference across studies required multiple sensitivity analyses; results moved in and out of significance depending on which studies were included.
  • 1-year follow-up is short for a disease with a decades-long trajectory; long-term durability, respiratory/cardiac/QoL outcomes, and effects on loss of ambulation were not assessed.
  • Safety could not be quantitatively meta-analyzed (only qualitative synthesis); post-marketing fatal acute liver failure signal is not captured in the trials pooled.
  • Excluded ongoing pivotal trials (SRP-9001-302 ENVOL, SRP-9001-303 ENVISION) because published results were not available at search date.
  • Population overlap possible across external cohorts drawn from Nationwide Children's Hospital.
  • Restricted to ambulatory boys aged 4 to <8 y; findings do not apply to non-ambulatory, older, or younger patients or to those with exon 8/9 deletions (contraindicated).

Funding

Not specified in the article (no industry funding declared for the meta-analysis itself; included trials were sponsored by Sarepta Therapeutics / F. Hoffmann-La Roche as developers of delandistrogene moxeparvovec).

Based on: DMD-ELEVIDYS-MA (Neurology Genetics, 2026)

Authors: Antonello BB, Cargnelutti Fontoura F, Braga Albuquerque AL, ..., Zattar Ribeiro PV

Citation: Neurol Genet. 2026 Jul;12(4):e200408.

Content summarized and formatted by NeuroTrials.ai.