← Back
NeuroTrials.ai
Neurology Clinical Trial Database

LUMINESCE

Safety and efficacy of satralizumab in patients with generalised myasthenia gravis (LUMINESCE): a randomised, double-blind, multicentre, placebo-controlled phase 3 trial

Year of Publication: 2025

Authors: Ali A Habib, Chongbo Zhao, Inmaculada Aban, ..., Hiroyuki Murai

Journal: Lancet Neurology

Citation: Lancet Neurol 2025; 24: 117–27

Link: https://clinicaltrials.gov/ct2/show/NCT04963270


Clinical Question

Is satralizumab, an IL-6 receptor inhibitor, safe and effective for treating seropositive generalised myasthenia gravis compared to placebo?

Bottom Line

Satralizumab was well tolerated and resulted in statistically significant but small improvements in MG-ADL and QMG scores at week 24 in AChR-IgG-positive generalised myasthenia gravis patients. However, the effect size was modest, and no difference was seen in quality of life measures. The sponsor halted further development for this indication.

Major Points

  • Primary endpoint met: Satralizumab significantly improved MG-ADL score vs placebo at week 24 (difference −1.02; 95% CI −1.88 to −0.16; p=0.020)
  • QMG score also showed significant improvement (difference −1.63; p=0.0062)
  • Response observed as early as week 4 and sustained through week 24
  • No significant difference in MG-QoL15r total score (p=0.11), breaking the hierarchical testing
  • 48% of satralizumab patients achieved ≥3-point QMG reduction vs 29% placebo
  • Safety profile similar to placebo with no deaths or adverse events of special interest
  • Open-label extension terminated early due to sponsor's decision to halt further development
  • IL-6 receptor occupancy >95% achieved across bodyweight ranges

Design

Study Type: Randomized, double-blind, placebo-controlled, multicentre, phase 3 trial

Randomization: 1

Blinding: Double-blind; sponsor, investigators, and patients masked; prefilled syringes identical in appearance; dual-assessor approach with separate treating (safety) and examining (efficacy) investigators

Enrollment Period: October 19, 2021 to August 15, 2023

Follow-up Duration: 24 weeks double-blind period (plus optional 2-year open-label extension, terminated early)

Centers: 105

Countries: Multiple countries across Europe, North America, South America, Australia, Asia

Sample Size: 188

Analysis: ANCOVA model adjusting for treatment group, visit, stratification factors, and baseline score; mixed model repeated measures for missing data imputation; jackknife technique for variance correction; hierarchical gatekeeping for multiplicity; two-sided α=0.05; 85% power for sample size of 160 AChR-IgG+ patients


Inclusion Criteria

  • Age ≥12 years
  • Diagnosis of seropositive generalised myasthenia gravis (AChR-IgG, MuSK-IgG, or LRP4-IgG positive)
  • MGFA severity class II-IV at screening
  • MG-ADL score ≥5 with >50% attributed to non-ocular symptoms
  • Receiving stable-dose background therapy: acetylcholinesterase inhibitors (stable 2 weeks), oral corticosteroids (stable 4 weeks, ≤30 mg/day), azathioprine (treatment ≥6 months, stable 2 months), or ciclosporine/tacrolimus/mycophenolate (≥3 months treatment, stable 4 weeks)

Exclusion Criteria

  • History of thymectomy within 6 months before screening
  • Ocular myasthenia gravis (MGFA class I)
  • Myasthenic crisis within 3 months before screening (MGFA class V)
  • Any known disease that would interfere with the study (e.g., severe rheumatoid arthritis, symptomatic thyroid disease)
  • IVIG or SCIG within 6 weeks before randomisation
  • Plasma exchange within 8 weeks before randomisation

Arms

FieldSatralizumabControl
InterventionSubcutaneous satralizumab: 120 mg for bodyweight ≤100 kg or 180 mg for bodyweight >100 kg at weeks 0, 2, 4, and every 4 weeks thereafter until week 24, plus stable background therapyMatched placebo subcutaneous injection at weeks 0, 2, 4, and every 4 weeks thereafter until week 24, plus stable background therapy
Duration24 weeks24 weeks

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Mean change from baseline in total MG-ADL score at week 24 in AChR-IgG-positive patients (modified intention-to-treat population)Primary−2.57 (SE 0.35; 95% CI −3.25 to −1.88)−3.59 (SE 0.29; 95% CI −4.15 to −3.02)0.020
Mean change from baseline in QMG score at week 24 (AChR-IgG+) | Difference: −1.63 (SE 0.60; 95% CI −2.80 to −0.46)Secondary−1.78 (SE 0.46; 95% CI −2.67 to −0.88)−3.41 (SE 0.41; 95% CI −4.21 to −2.61)0.0062
Mean change from baseline in MG-QoL15r total score at week 24 (AChR-IgG+) | Difference: −1.51 (SE 0.94; 95% CI −3.36 to 0.34)Secondary−4.69 (SE 0.71; 95% CI −6.09 to −3.30)−6.20 (SE 0.69; 95% CI −7.56 to −4.85)0.11 (not significant - hierarchy broken)
Mean change from baseline in MGC total score at week 24 (AChR-IgG+) | Difference: −2.99 (SE 0.81; 95% CI −4.57 to −1.41)Secondary−4.14 (SE 0.62; 95% CI −5.35 to −2.92)−7.13 (SE 0.58; 95% CI −8.27 to −5.99)Not inferential (hierarchy broken)
Proportion receiving rescue therapy | Difference: −6.77% (95% CI −17.25 to 3.70)Secondary13.8% (11 patients; 95% CI 5.58-21.92)7.0% (6 patients; 95% CI 1.01-12.94)Not inferential (hierarchy broken)
Mean change from baseline in Neuro-QoL Fatigue subscale at week 24 | Difference: −2.20 (SE 1.10; 95% CI −4.36 to −0.04)Secondary−3.29 (SE 0.90; 95% CI −5.05 to −1.53)−5.50 (SE 0.75; 95% CI −6.97 to −4.02)Not inferential (hierarchy broken)
≥2-point reduction in MG-ADL at week 24 (responder analysis) | Difference: 12.7% (95% CI −27.3 to 1.9)Secondary59% (47 patients)71% (61 patients)
≥3-point reduction in QMG at week 24 (responder analysis) | Difference: 18.7% (95% CI −33.5 to 3.9)Secondary29% (23 patients)48% (41 patients)
Minimal symptom expression (MG-ADL 0-1) at week 24 | Difference: 1.7% (95% CI −19.4 to 15.9)Secondary13% (10 patients)14% (12 patients)
Any adverse eventAdverse73% (67/92 patients; 261 events)90% (86/96 patients; 286 events)
AEs related to study treatmentAdverse22% (20/92 patients; 47 events)47% (45/96 patients; 102 events)
AEs leading to discontinuationAdverse1% (1/92 patients)2% (2/96 patients)
Serious adverse eventsAdverse7% (6/92 patients; 9 events)3% (3/96 patients; 3 events)
Serious AEs related to study treatmentAdverse0%2% (2/96 patients)
InfectionsAdverse42% (39/92 patients; 59 events)45% (43/96 patients; 56 events)
Serious infectionsAdverse4% (4/92 patients; 4 events)2% (2/96 patients; 2 events)
Injection-related reactionsAdverse7% (6/92 patients; 18 events)15% (14/96 patients; 39 events)
HeadacheAdverse10% (9/92)9% (9/96)
COVID-19Adverse10% (9/92)13% (12/96)
Upper respiratory tract infectionAdverse9% (8/92)6% (6/96)
Urinary tract infectionAdverse7% (6/92)5% (5/96)
DeathsAdverse0%0%
Adverse events of special interestAdverse0%0%

Subgroup Analysis

Subgroup analyses for the primary endpoint showed consistent but small improvements in MG-ADL score with satralizumab compared with placebo across baseline and demographic characteristics including sex, age at diagnosis, race, region, background therapy, baseline disease severity, and dose level. No significant difference was observed in MuSK-IgG-positive patients (n=16), though the study was not powered for this subgroup. Only one LRP4-IgG-positive patient was randomised to satralizumab, precluding independent evaluation.


Criticisms

  • Effect size was small (MG-ADL difference of −1.02 points) with uncertain clinical meaningfulness
  • MG-QoL15r endpoint not significant (p=0.11), breaking the hierarchical testing and limiting inferences for downstream endpoints
  • No further improvement observed during open-label extension beyond week 24 effect
  • Imbalance in refractory MG at baseline (52% placebo vs 38% satralizumab)
  • Underpowered for MuSK-IgG-positive and LRP4-IgG-positive subgroups
  • Only 3 adolescent patients enrolled, precluding efficacy/safety assessment in this age group
  • Limited North American representation (10 patients total) with only 2 Black patients
  • Sponsor terminated open-label extension and halted further development for this indication
  • Cross-trial comparisons with C5 inhibitors and FcRn antagonists difficult due to design differences
  • 24-week duration may not capture long-term efficacy and durability of response

Funding

F Hoffmann La Roche

Based on: LUMINESCE (Lancet Neurology, 2025)

Authors: Ali A Habib, Chongbo Zhao, Inmaculada Aban, ..., Hiroyuki Murai

Citation: Lancet Neurol 2025; 24: 117–27

Content summarized and formatted by NeuroTrials.ai.