← Back
NeuroTrials.ai
Neurology Clinical Trial Database

DEVOTE

High-dose nusinersen for spinal muscular atrophy: a phase 3 randomized trial

Share Share Copied! Compare

Year of Publication: 2026

Authors: Finkel RS, Crawford TO, Mercuri E, ..., Fradette S

Journal: Nature Medicine

Citation: Finkel RS, et al. High-dose nusinersen for spinal muscular atrophy: a phase 3 randomized trial. Nat Med. 2026;32:1095-1104.

Link: https://doi.org/10.1038/s41591-025-04193-6

Clinical Question

In patients with spinal muscular atrophy, does high-dose nusinersen (50-mg loading, 28-mg maintenance) improve motor outcomes more rapidly than standard-dose nusinersen or sham control?

Bottom Line

High-dose nusinersen (50/28 mg) met its primary endpoint, demonstrating significantly greater improvement in CHOP-INTEND scores compared to matched sham controls (15.1-point improvement vs 11.1-point worsening; P < 0.0001) and achieved substantially greater reductions in plasma neurofilament levels (94% vs 30%; P < 0.0001), indicating more rapid slowing of neurodegeneration. The safety profile was similar to standard-dose nusinersen with no new safety concerns identified.

Major Points

  • Primary endpoint met: CHOP-INTEND total score at day 183 improved 15.1 points with 50/28 mg nusinersen versus -11.1 points with matched ENDEAR sham (difference 26.19, 95% CI 20.7-31.74; P < 0.0001)
  • Plasma neurofilament light chain levels reduced 94% with 50/28 mg versus 30% with matched sham at day 183 (P < 0.0001), indicating substantially slowed neurodegeneration
  • More rapid reduction in NfL levels at day 64: 88% with 50/28 mg versus 77% with 12/12 mg (nominally significant P = 0.0050)
  • HINE-2 responder rate and score improvement significantly higher with 50/28 mg versus matched sham (P < 0.0001)
  • Lower risk of death or permanent ventilation with 50/28 mg versus matched sham (HR 0.322, nominal P = 0.0006) and trend toward lower risk versus 12/12 mg (HR 0.701, P = 0.2775)
  • Safety profile of 50/28 mg similar to 12/12 mg regimen with no new safety concerns
  • Part C participants (previously treated with 12/12 mg) showed mean HFMSE improvement of 1.8 points at day 302; 53% experienced increased scores
  • 50/28 mg regimen achieved higher CSF trough concentrations more quickly (one dose of 50 mg vs three doses of 12 mg)

Design

Study Type: Phase 2/3, three-part trial

Randomization: 1

Blinding: Open-label

Allocation: Part B: 2:1 randomization (50/28 mg vs 12/12 mg); Part C: open-label

Enrollment Period: November 12, 2020 to August 8, 2023

Follow-up Duration: 302 days primary analysis period

Centers: 0

Countries: Multiple countries

Sample Size: 139

Analyzed: 139

Analysis: Primary analysis using joint-rank test with ANCOVA adjusting for disease duration and baseline CHOP-INTEND; multiple imputation for missing data

Power Calculation: Powered to compare treatment-naive infantile-onset participants receiving high-dose nusinersen with matched subset of ENDEAR sham control group; not powered for comparison of 50/28 mg vs 12/12 mg groups

Registration: ClinicalTrials.gov: NCT04089566; EudraCT: 2019-002663-10

Inclusion Criteria

  • Part B: Treatment-naive individuals with infantile-onset or later-onset SMA
  • Part C: Individuals with SMA who had received 12/12 mg nusinersen for more than 1 year
  • Genetic confirmation of SMA (specific criteria not detailed in abstract)

Exclusion Criteria

  • Specific exclusion criteria not detailed in the published article

Arms

FieldPart B Infantile-Onset 50/28 mgControlControlPart B Later-Onset 50/28 mgControlPart C Open-Label 50/28 mg
N50252016840
InterventionNusinersen 50-mg loading doses followed by 28-mg maintenance doses every 4 months, intrathecal administrationNusinersen 12-mg loading doses followed by 12-mg maintenance doses every 4 months, intrathecal administrationSham procedure (matched historical control from ENDEAR trial)Nusinersen 50-mg loading doses followed by 28-mg maintenance doses every 4 months, intrathecal administrationNusinersen 12-mg loading doses followed by 12-mg maintenance doses every 4 months, intrathecal administrationNusinersen 50-mg loading doses followed by 28-mg maintenance doses every 4 months, intrathecal administration (transition from prior 12/12 mg)
Duration302 days primary analysis period302 days primary analysis periodMatched to day 183302 days primary analysis period302 days primary analysis period302 days primary analysis period

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Change in CHOP-INTEND total score at day 183 in Part B infantile-onset participantsPrimaryENDEAR matched sham: -11.1 points (95% CI -15.9 to -6.2)50/28 mg: +15.1 points (95% CI 12.4 to 17.8)<0.0001
HINE-2 responder rate at day 183 (Part B infantile-onset)SecondaryENDEAR matched sham: lower proportion50/28 mg: significantly higher proportion<0.0001
HINE-2 score improvement at day 183 (Part B infantile-onset)SecondaryENDEAR matched sham: lower improvement50/28 mg: greater improvement<0.0001
Plasma NfL reduction from baseline at day 183 (Part B infantile-onset)SecondaryENDEAR matched sham: 30% reduction50/28 mg: 94% reduction<0.0001
Plasma NfL reduction at day 64 (Part B infantile-onset)Secondary12/12 mg: 77% reduction50/28 mg: 88% reduction0.0050 (nominally significant)
Event-free survival (Part B infantile-onset)SecondaryENDEAR matched sham50/28 mg: lower risk0.0006 (nominally significant)
Event-free survival: 50/28 mg vs 12/12 mg (Part B infantile-onset)Secondary12/12 mg50/28 mg: trend toward lower risk0.2775
Overall survival (Part B infantile-onset)SecondaryENDEAR matched sham: higher mortality50/28 mg: lower mortality, similar pattern to EFSNot specified
CHOP-INTEND at day 302 (Part B infantile-onset)Secondary12/12 mg: numerically higher LSM actual score50/28 mg: numerically higher LSM ranked scoreNot significant
HFMSE improvement (Part B later-onset)Secondary12/12 mg: numerically lower improvement50/28 mg: numerically greater improvementNot specified
RULM improvement (Part B later-onset)Secondary12/12 mg: numerically lower improvement50/28 mg: numerically greater improvementNot specified
Plasma NfL reduction at day 64 (Part B later-onset)Secondary12/12 mg: lower reduction50/28 mg: greater reduction0.0495 (nominally significant)
HFMSE improvement (Part C)SecondaryBaselineMean improvement 1.8 points (SD 3.99); adult subgroup 2.3 points (SD 3.95)
HFMSE responders (Part C)SecondaryBaseline53% (20/38) experienced increase in score
RULM improvement (Part C)SecondaryBaselineMean improvement 1.2 points (SD 2.14); adult subgroup 0.9 points (SD 1.89)
RULM responders (Part C)SecondaryBaseline62% (16/26) with opportunity to improve experienced increase in score
Any adverse event (Part B infantile-onset 50/28 mg)SafetyMajority were mild to moderate
Treatment-related adverse events (Part B infantile-onset 50/28 mg)Safety<7%
Serious adverse events (Part B infantile-onset)SafetyENDEAR matched sham: 95%; 12/12 mg: 72%50/28 mg: 60%
Treatment-related serious adverse events (Part B infantile-onset 50/28 mg)SafetyNone
Severe adverse events (Part B infantile-onset)SafetyENDEAR matched sham: 90%; 12/12 mg: 56%50/28 mg: 44%
Fatal adverse events (Part B infantile-onset)SafetyENDEAR matched sham: 55%; 12/12 mg: 24%50/28 mg: 20%
Most common AEs (≥15%) in Part B infantile-onset 50/28 mgSafetyPneumonia (20%), respiratory failure (20%), pyrexia (18%), COVID-19 (16%), upper respiratory tract infection (16%)
AEs leading to treatment discontinuationSafetyNone except fatal AEs (assessed as unrelated to treatment)
Meningitis, hydrocephalus, thrombocytopenia, renal/liver failureSafetyNone reported across Parts B and C
Pneumonia 20%, respiratory failure 20%, pyrexia 18%, COVID-19 16%, upper respiratory tract infection 16%Adverse
Pneumonia 14% vs 5%, pneumonia aspiration 14% vs 5%Adverse
<7% in Part B overall; 20% in Part CAdverse
One participant in Part B infantile-onset 12/12 mg group (respiratory failure); none in 50/28 mg groupAdverse
Part B infantile-onset: 50/28 mg 20%, 12/12 mg 24%, ENDEAR matched sham 55%; all assessed as unrelated to treatmentAdverse

Subgroup Analysis

Prespecified matching was performed to create comparator groups from ENDEAR (sham) and CHERISH (sham and 12/12 mg) trials. Adult subgroup in Part C showed HFMSE improvement of 2.3 points and RULM improvement of 0.9 points at day 302.

Criticisms

  • Study not powered to detect statistically significant differences between 50/28 mg and 12/12 mg active treatment groups
  • Use of matched historical controls from ENDEAR rather than concurrent placebo group introduces complexity and potential for unmeasured confounding
  • Broad geographical scope and inclusion of new sites introduced variability in standard of care (e.g., availability of cough assist machines, ventilators)
  • In one country, 10/13 participants met definition for permanent ventilation or death, highlighting impact of differences in standard of care and supportive equipment
  • Small sample size in Part B later-onset cohort (n=24, randomized 2:1) led to highly variable results over time
  • Prespecified matching algorithm could not account for all differences in baseline characteristics; DEVOTE participants were younger at symptom onset with lower baseline CHOP-INTEND scores
  • Part C enrolled heterogeneous population by design with limited opportunity for improvement in some participants already at upper end of scoring
  • Supportive equipment was not provisioned to sites with identified gaps until later in the study, potentially affecting outcomes
  • Higher incidence of pneumonia aspiration in 50/28 mg group (14% vs 5% in matched sham), though assessed as unrelated to treatment

Funding

Biogen (sponsor of the study)

Based on: DEVOTE (Nature Medicine, 2026)

Authors: Finkel RS, Crawford TO, Mercuri E, ..., Fradette S

Citation: Finkel RS, et al. High-dose nusinersen for spinal muscular atrophy: a phase 3 randomized trial. Nat Med. 2026;32:1095-1104.

Content summarized and formatted by NeuroTrials.ai.