HEALEY ALS Platform Trial
(2025)Objective
To describe the operational framework and results from the first four regimens of an adaptive platform trial in ALS
Study Summary
• Platform design validated for ALS drug development: master protocol IRB approval in 120 days, regimen amendments in 21 days, 84-90% completion rates; individual results published in JAMA 2025 and JAMA Neurol 2025
Intervention
Four concurrent regimens: Zilucoplan, Verdiperstat, Trehalose, Pridopidine, plus CNM-Au8
Inclusion Criteria
Adults with ALS, enrolled at 54 NEALS consortium sites across the US
Study Design
Arms: Regimen A (Zilucoplan) vs B (Verdiperstat) vs C (Trehalose) vs D (Pridopidine) vs shared placebo
Patients per Arm: ~120 active per regimen, ~40 placebo per regimen, shared placebo pool of ~124-164
Outcome
• CNM-Au8 biomarker signal: plasma NfL decreased 2.3% vs increased 7.9% in placebo (p=0.04) without clinical benefit
• Platform operational success: 84-90% completion rates, concurrent enrollment, shared placebo reduced total placebo exposure
Bottom Line
The HEALEY ALS Platform Trial demonstrated that an adaptive platform design enables concurrent testing of multiple ALS treatments with shared placebo and accelerated regulatory timelines. While none of the first four regimens showed efficacy, the infrastructure proved operationally efficient.
Major Points
- Master protocol achieved single IRB approval in 120 days; fourth regimen amendment approved in just 21 days -- demonstrating regulatory efficiency of platform design
- Enrollment completed in 15 months for Regimens A-C and 11 months for Regimen D; 84-90% completion rates
- 54 NEALS consortium sites participated across the US with concurrent enrollment into multiple regimens
- Shared placebo pool of 122-164 participants increased statistical power while reducing total placebo exposure
- None of five tested regimens met efficacy threshold: Pridopidine DRR 0.99 (95% CrI 0.80-1.21), Verdiperstat DRR 0.98 (0.77-1.24), CNM-Au8 DRR 0.97 (0.78-1.18), Zilucoplan DRR 1.08, Trehalose DRR 0.87
- CNM-Au8 showed significant biomarker effect (plasma NfL decreased 2.3% vs increased 7.9% in placebo, p=0.04) but no clinical benefit
- Individual regimen results published separately: Pridopidine in JAMA 2025, Verdiperstat in JAMA Neurol 2025, CNM-Au8 in JAMA 2025
Study Design
- Study Type
- Adaptive platform trial, phase 2/3, multicenter, randomized, double-blind, placebo-controlled
- Randomization
- Yes
- Blinding
- Double-blind
- Sample Size
- 653
- Follow-up
- 24 weeks per regimen
- Centers
- 54
- Countries
- United States
Primary Outcome
Definition: Disease Rate Ratio (DRR) from Bayesian shared-parameter joint model combining ALSFRS-R slope and survival; DRR <1 indicates slowing of disease progression
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| Shared placebo pool (42 regimen-specific + shared concurrent placebo per regimen) | Pridopidine DRR 0.99 (95% CrI 0.80-1.21, probability <1 = 0.55); Verdiperstat DRR 0.98 (95% CrI 0.77-1.24, probability = 0.57); CNM-Au8 DRR 0.97 (95% CrI 0.783-1.175, probability <1 = 0.65); Zilucoplan DRR 1.08; Trehalose DRR 0.87 | - (See individual regimen CrIs above) | None met prespecified efficacy threshold |
Limitations & Criticisms
- None of the first five tested regimens (Zilucoplan, Verdiperstat, Trehalose, Pridopidine, CNM-Au8) demonstrated clinical efficacy
- 24-week treatment duration may be insufficient for disease-modifying agents in ALS
- Shared placebo design introduces statistical complexity and potential temporal confounding across regimens
- US-only sites (54 centers) may limit generalizability to non-US ALS populations
- Bayesian DRR endpoint not widely used outside this trial, limiting cross-study comparison
- Platform master protocol allows concurrent enrollment which may create selection bias across regimens
- CNM-Au8 biomarker signal (NfL reduction p=0.04) did not translate to clinical benefit, questioning NfL as a surrogate endpoint
Citation
Muscle Nerve. 2025;72(2):294-305