PREVAIL
Efficacy and Safety of Gefurulimab in Generalized Myasthenia Gravis: The PREVAIL Phase 3 Randomized Clinical Trial
Year of Publication: 2026
Authors: Gwathmey KG, Saccà F, Howard JF Jr, et al; PREVAIL Trial Investigators
Journal: JAMA Neurology
Citation: JAMA Neurol. 2026 Jul 27:e262333. Online ahead of print. DOI: 10.1001/jamaneurol.2026.2333
PDF: Clinical Question
Does once-weekly subcutaneous gefurulimab, a dual-binding anti-albumin/anti-C5 nanobody, improve MG-ADL score at 26 weeks vs placebo in adults with AChR-Ab+ generalized myasthenia gravis, with acceptable safety?
Bottom Line
In the phase 3 PREVAIL trial of 260 adults with AChR-Ab+ generalized myasthenia gravis, once-weekly subcutaneous gefurulimab significantly reduced MG-ADL at 26 weeks vs placebo (LS mean change −4.2 vs −2.6; treatment difference −1.6, 95% CI −2.4 to −0.8; P<.001), with concordant benefit on QMG (−4.5 vs −2.4; difference −2.1, P<.001) and MGC (−7.8 vs −4.7; difference −3.1, P<.001). Improvements appeared as early as week 1 and were sustained through 26 weeks, with a safety profile similar to placebo and no meningococcal infections. Gefurulimab is an efficacious, self-administered, weekly subcutaneous option for AChR-Ab+ gMG.
Major Points
- Phase 3, randomized, double-blind, placebo-controlled trial at 113 sites in 20 countries; 260 adults with AChR-Ab+ gMG (MGFA II–IV, MG-ADL ≥5) randomized 1:1 to weight-based weekly subcutaneous gefurulimab (n=131) or matching placebo (n=129) for 26 weeks; 249 (95.8%) completed the RCT.
- Primary end point (LS mean change in MG-ADL from baseline to week 26) achieved: −4.2 (95% CI −4.7 to −3.6) vs −2.6 (−3.1 to −2.0); treatment difference −1.6 (95% CI −2.4 to −0.8; P<.001).
- Key secondary QMG change at week 26: −4.5 (−5.2 to −3.8) vs −2.4 (−3.1 to −1.7); difference −2.1 (−3.1 to −1.1; P<.001). MGC change: −7.8 vs −4.7; difference −3.1 (−4.4 to −1.7; P<.001).
- Responder analyses: ≥5-point QMG reduction 46.3% vs 25.2% (OR 2.56, 1.42–4.62; P=.002); ≥3-point MG-ADL reduction 65.5% vs 52.2% (OR 1.74, 1.03–2.93; P=.04).
- Onset of effect was early: MG-ADL separated by week 1 (difference −1.3; 95% CI −1.9 to −0.8; P<.001) and QMG by week 4 (difference −1.8; 95% CI −2.5 to −1.1; P<.001); benefit sustained through week 26.
- Rescue therapy required by 7 patients (5.3%) on gefurulimab vs 16 (12.4%) on placebo.
- Safety: any treatment-emergent AE 75.6% vs 80.6%; serious AEs 9.2% vs 11.6%; trial-related serious AEs 0 vs 2 (Bacillus bacteremia, herpes zoster in placebo); injection-site reactions 9.9% vs 3.1% (mostly mild, first 3 weeks, none led to discontinuation); no meningococcal infections in either arm.
- One death per arm; neither considered related to trial intervention (dengue fever on gefurulimab; metabolic acidosis from multiorgan failure on placebo).
Design
Study Type: Phase 3, randomized, double-blind, placebo-controlled multicenter trial with open-label extension (this report: 26-week RCT period only)
Randomization: 1
Blinding: Double-blind (patients and investigators); matching placebo prefilled syringes; 1:1 allocation stratified by geographical region and baseline body weight (<80 vs ≥80 kg).
Enrollment Period: November 2022 – November 2024
Follow-up Duration: 26-week randomized controlled treatment period (open-label extension up to 202 weeks ongoing)
Centers: 113
Countries: Global — 20 countries across North America, South America, Europe, Japan, China, and rest of Asia-Pacific
Sample Size: 260
Power Calculation: Based on the CHAMPION MG (ravulizumab) result and 10% attrition, 254 patients provided ~90% power at 2-sided α=0.05 to detect a treatment effect on change from baseline in MG-ADL at week 26.
Analysis: All randomized patients who received ≥1 dose analyzed regardless of intercurrent events (rescue therapy, discontinuation). Continuous end points analyzed by mixed-effects model for repeated measures (MMRM) with fixed effects for treatment, stratification factors (region, baseline body weight <80 vs ≥80 kg), baseline score, visit and treatment×visit; missing data not imputed (assumed MAR). Responder end points by generalized linear mixed model. Primary and key secondary tested in fixed-sequence hierarchical procedure at 2-sided α=0.05; if both significant, remaining secondaries tested with Holm-Bonferroni. Sensitivity analyses on missing data and per-protocol set.
Inclusion Criteria
- Adults aged ≥18 years
- Documented diagnosis of generalized myasthenia gravis ≥90 days before screening
- Positive serological test for anti-AChR autoantibodies at screening
- MGFA clinical classification II–IV at screening
- MG-ADL total score ≥5 at screening and on day 1 before randomization
- Meningococcal (N meningitidis) vaccination within 3 years before day 1 (or vaccinated before first dose, with prophylactic antibiotics if <2 weeks pre-dose)
- Continued use of previously prescribed MG therapies (corticosteroids, nonsteroidal immunosuppressants, IVIg/SCIg) permitted if doses stable and well tolerated
Exclusion Criteria
- Prior treatment with a complement or FcRn inhibitor within less than 5 half-lives before day 1
- Inadequate meningococcal vaccination without appropriate mitigation
- Additional protocol-specified medical, laboratory, and safety exclusions (full list in Supplement 1)
Baseline Characteristics
| Characteristic | Gefurulimab (n=131) | Placebo (n=129) |
|---|---|---|
| Age at first dose, mean (SD), y | 53.0 (14.42) | 52.7 (17.00) |
| Female n (%) | 78 (59.5) | 79 (61.2) |
| Male n (%) | 53 (40.5) | 50 (38.8) |
| Race — White n (%) | 69 (52.7) | 74 (57.4) |
| Race — Asian n (%) | 43 (32.8) | 38 (29.5) |
| Race — Black or African American n (%) | 3 (2.3) | 3 (2.3) |
| Region — Europe n (%) | 57 (43.5) | 56 (43.4) |
| Region — North America n (%) | 16 (12.2) | 15 (11.6) |
| Region — China n (%) | 16 (12.2) | 15 (11.6) |
| MGFA class II (IIa/IIb) n (%) | 48 (36.6) | 45 (34.9) |
| MGFA class III (IIIa/IIIb) n (%) | 76 (58.0) | 77 (59.7) |
| MGFA class IV (IVa/IVb) n (%) | 7 (5.4) | 7 (5.4) |
| Baseline MG-ADL total score, mean (SD) | 9.0 (2.29) | 9.0 (2.03) |
| Baseline QMG total score, mean (SD) | 14.9 (4.38) | 14.7 (4.39) |
| Baseline MGC total score, mean (SD) | 16.4 (5.32) | 16.1 (5.39) |
| Age at MG diagnosis, mean (SD), y | 43.8 (17.55) | 44.6 (19.37) |
| Time from MG diagnosis to first dose, mean (SD), y | 9.2 (8.45) | 8.2 (8.79) |
| Receiving corticosteroids at baseline n (%) | 84 (64.1) | 92 (71.3) |
| Receiving nonsteroidal immunosuppressant at baseline n (%) | 79 (60.3) | 72 (55.8) |
Arms
| Field | Gefurulimab | Control |
|---|---|---|
| Intervention | Gefurulimab (dual-binding anti-albumin/anti-C5 nanobody, ALXN1720): weight-based loading dose 600 or 900 mg SC on day 1, then maintenance 300 or 600 mg SC weekly (patient/caregiver self-injection with prefilled syringe) for 26 weeks | Matching placebo SC weekly with prefilled syringe for 26 weeks |
| N | 131 | 129 |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| Change from baseline in MG-ADL total score (range 0–24, higher = worse) at week 26; MMRM in the analysis set of all randomized patients receiving ≥1 dose. | Primary | LS mean change −2.6 (95% CI −3.1 to −2.0) | LS mean change −4.2 (95% CI −4.7 to −3.6) | <0.001 | |
| Change from baseline in QMG total score at week 26 (key secondary; range 0–39, higher = worse) | Secondary | LS mean −2.4 (95% CI −3.1 to −1.7) | LS mean −4.5 (95% CI −5.2 to −3.8) | <0.001 | |
| ≥5-point reduction in QMG total score at week 26 (responder) | Secondary | 33/118 (28.0%); adjusted 25.2% | 58/124 (46.8%); adjusted 46.3% | 0.002 | |
| ≥3-point reduction in MG-ADL total score at week 26 (responder) | Secondary | 63/121 (52.1%); adjusted 52.2% | 82/125 (65.6%); adjusted 65.5% | 0.04 | |
| Change from baseline in MGC total score at week 26 (range 0–50, higher = worse) | Secondary | LS mean −4.7 (95% CI −5.7 to −3.8) | LS mean −7.8 (95% CI −8.8 to −6.8) | <0.001 | |
| Change from baseline in QMG total score at week 4 (early onset) | Secondary | LS mean −1.5 (95% CI −2.0 to −1.0) | LS mean −3.3 (95% CI −3.9 to −2.7) | <0.001 | |
| Change from baseline in MG-ADL total score at week 1 (early onset) | Secondary | — | Reported significantly better than placebo | <0.001 | |
| Receipt of rescue therapy during 26-week RCT | Secondary | 16/129 (12.4%) | 7/131 (5.3%) | ||
| Any treatment-emergent AE | Safety | 104/129 (80.6%) | 99/131 (75.6%) | ||
| Any AE related to trial intervention | Safety | 36/129 (27.9%) | 36/131 (27.5%) | ||
| Any serious AE | Safety | 15/129 (11.6%) | 12/131 (9.2%) | ||
| Serious AE related to trial intervention | Safety | 2/129 (1.6%) — Bacillus bacteremia, herpes zoster | 0/131 (0%) | ||
| AE leading to discontinuation of trial intervention | Safety | 1/129 (0.8%) — vitamin D deficiency | 1/131 (0.8%) — suicide attempt | ||
| AE of special interest (meningococcal infection) | Safety | 0/129 (0%) | 0/131 (0%) | ||
| Injection site reactions (combined preferred terms) | Safety | 4/129 (3.1%) | 13/131 (9.9%) | ||
| Grade 3 AE | Safety | 16/129 (12.4%) | 13/131 (9.9%) | ||
| Grade 4 AE | Safety | 5/129 (3.9%) | 3/131 (2.3%) | ||
| Death (Grade 5) | Safety | 1/129 (0.8%) — metabolic acidosis/multiorgan failure, unrelated | 1/131 (0.8%) — dengue fever, unrelated | ||
| Headache | Safety | 16/129 (12.4%) | 13/131 (9.9%) | ||
| Back pain | Safety | 2/129 (1.6%) | 10/131 (7.6%) | ||
| Nasopharyngitis | Safety | 9/129 (7.0%) | 9/131 (6.9%) | ||
| Upper respiratory tract infection | Safety | 10/129 (7.8%) | 8/131 (6.1%) | ||
| Diarrhea | Safety | 11/129 (8.5%) | 8/131 (6.1%) | ||
| Urinary tract infection | Safety | 4/129 (3.1%) | 7/131 (5.3%) | ||
| COVID-19 | Safety | 9/129 (7.0%) | 5/131 (3.8%) | ||
| Dizziness | Safety | 8/129 (6.2%) | 2/131 (1.5%) |
Subgroup Analysis
The paper emphasizes that primary and secondary end points were met with statistical significance across the population; formal prespecified subgroup analyses (by region, weight, MGFA class) are shown in Supplement 2 and are described as consistent with the overall result. No effect-modification interactions were highlighted in the main text.
Criticisms
- Short controlled duration (26 weeks) — long-term efficacy, durability, and rare adverse events (including meningococcal infection risk on any C5 inhibitor) require the ongoing open-label extension.
- No active comparator: benefit is quantified vs placebo only. Cross-trial comparisons to eculizumab, ravulizumab, zilucoplan, efgartigimod, rozanolixizumab, and nipocalimab must be interpreted with caution given differences in populations, endpoints, and mechanisms.
- Underrepresentation of Black or African American patients (2.3%) limits generalizability to this population despite the trial's overall geographic diversity.
- Sponsor (Alexion, AstraZeneca Rare Disease) had a role in trial design, conduct, data management, analysis, and manuscript preparation.
- Small imbalances in baseline features (ocular MG as first presentation, prior thymectomy, immunoglobulin use) between groups, though considered unlikely to affect overall results.
- Rescue therapy allowed but analyzed regardless of intercurrent events; higher rescue use in placebo (12.4% vs 5.3%) could bias against showing an even larger gefurulimab effect.
- Meningococcal infection risk of C5 blockade remains a class concern despite mandatory vaccination and none observed in this 26-week window.
Funding
Sponsored by Alexion, AstraZeneca Rare Disease, which had a role in trial design and conduct; data collection, management, analysis, and interpretation; manuscript preparation, review, and approval; and decision to submit for publication.
Based on: PREVAIL (JAMA Neurology, 2026)
Authors: Gwathmey KG, Saccà F, Howard JF Jr, et al; PREVAIL Trial Investigators
Citation: JAMA Neurol. 2026 Jul 27:e262333. Online ahead of print. DOI: 10.1001/jamaneurol.2026.2333
Content summarized and formatted by NeuroTrials.ai.