← Back
NeuroTrials.ai
Neurology Clinical Trial Database

AMULET

Safety and efficacy of the anti-α-synuclein monoclonal antibody amlenetug for the treatment of patients with multiple system atrophy (AMULET): a phase 2, randomised, double-blind, multicentre trial

Year of Publication: 2026

Authors: Kjærsgaard L, Wiedemann J, Singer W, ..., Luthman J; AMULET Trial Group

Journal: The Lancet Neurology

Citation: The Lancet Neurology, Volume 25, Issue 6, June 2026, Pages 560-570

Link: https://doi.org/10.1016/S1474-4422(26)00100-6


Clinical Question

Does the anti-α-synuclein monoclonal antibody amlenetug slow clinical disease progression in patients with multiple system atrophy?


Study Overview

Objective

To assess the efficacy and safety of amlenetug, a monoclonal antibody targeting aggregated α-synuclein, versus placebo in slowing clinical disease progression in patients with multiple system atrophy (MSA).

Study Summary

  • Primary endpoint NOT met: Bayesian probability of 89.4% for true slowing of progression did not reach predefined 97.5% threshold
  • Non-significant 19% slowing of clinical progression (effect parameter 0.81, 2.5th-97.5th percentile 0.56-1.13; 2.5th-97.5th percentile of % slowing -13 to 44) with amlenetug vs placebo
  • Generally well tolerated: TEAEs in 100% (amlenetug) vs 95% (placebo); serious TEAEs in 30% vs 33%
  • Two deaths in each group; only one (placebo) considered possibly treatment-related
  • Findings supported further evaluation in a phase 3 trial

Intervention

Intravenous amlenetug 4.2 g (80 mL) every 4 weeks for 48–72 weeks vs matched placebo

Patients per Arm

Amlenetug n=40; Placebo n=21

Bottom Line

In this phase 2 trial, amlenetug did not meet its primary endpoint for slowing disease progression in MSA (Bayesian probability 89.4% vs predefined 97.5% threshold), though a non-significant 19% slowing of clinical progression and acceptable safety profile supported advancement to a phase 3 trial.

Major Points

  • Primary endpoint not met: Bayesian probability of 89.4% for true slowing of disease progression did not reach the predefined 97.5% threshold
  • Effect parameter 0.81 (2.5th-97.5th percentile 0.56-1.13) corresponded to a non-significant 19% slowing of clinical progression (2.5th-97.5th percentile -13% to 44%) for amlenetug vs placebo
  • First randomized controlled trial comparing an anti-α-synuclein antibody vs placebo in MSA
  • Bayesian progression model used to enable feasible sample size in rare, rapidly progressing disease
  • Amlenetug was generally well tolerated with comparable rates of TEAEs (100% vs 95%) and serious TEAEs (30% vs 33%) vs placebo
  • Trial design used 2:1 randomization and a common close design with 48–72 weeks of treatment
  • Despite negative primary endpoint, findings support further evaluation in a phase 3 trial; open-label extension is ongoing

Design

Study Type: Phase 2, randomised, double-blind, placebo-controlled, parallel-group, multicentre trial

Randomization: 1

Blinding: Double-blind: participants, study partners, treating investigators, study site personnel, and sponsor were all masked to treatment allocation; independent site personnel prepared trial medication

Allocation: 2:1 (amlenetug:placebo), stratified by region (USA or Japan) and blood NfL concentration (low ≤20 pg/mL, high >20 pg/mL, or missing), block size of three, via interactive web-based response technology; recruitment capped at 25% from Japan

Enrollment Period: Nov 16, 2021 to Oct 6, 2022

Follow-up Duration: 48–72 weeks (common close design ending when last randomised participant completed week 48 visit)

Centers: 18

Countries: USA, Japan

Sample Size: 64

Analyzed: 61

Analysis: Full analysis set (FAS): all participants with valid baseline and ≥1 valid post-baseline UMSARS total score assessment; safety assessed in all treated participants. Bayesian repeated measures model with proportional treatment effect (θ); threshold for statistical significance = 97.5% probability of active treatment slowing progression vs placebo

Power Calculation: Sample size of 60 (2:1 allocation) estimated to provide 75% power to detect 40% slowing of MSA progression at one-sided α=0.05, assuming 20% dropout, ~1 point UMSARS total worsening per month, 50% contributing 60 weeks and 25% contributing 72 weeks of data

Registration: ClinicalTrials.gov NCT05104476


Inclusion Criteria

  • Male or female aged 40–75 years
  • Possible or probable MSA of parkinsonian (MSA-P) or cerebellar (MSA-C) subtype
  • Onset of motor MSA symptoms within 5 years before screening
  • UMSARS part I score ≤16 (omitting item 11 on sexual function)
  • Montreal Cognitive Assessment (MoCA) score ≥22
  • Anticipated survival ≥3 years
  • Care partner with ≥3 h weekly contact able to accompany participant to trial visits

Exclusion Criteria

  • Evidence (clinical or MRI) or history of any other serious neurological disorder, including movement disorders that could mimic MSA
  • Other intracranial or systemic conditions leading to a diagnosis other than MSA
  • Family history of MSA (≥2 blood relatives)
  • Attempted suicide within past 6 months or significant suicide risk
  • Receipt of any investigational products within 30 days before screening
  • Previous treatment with antibodies targeting α-synuclein
  • Contraindications to MRI

Arms

FieldAmlenetugControl
N4021
InterventionIntravenous amlenetug 4.2 g (80 mL) infusion over 30 min every 4 weeks (±3 days)Intravenous matched placebo infusion over 30 min every 4 weeks (±3 days)
Duration48–72 weeks (common close design)48–72 weeks (common close design)

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Disease progression assessed using a Bayesian progression model of longitudinal change from baseline in UMSARS total (parts I and II) score during the double-blind treatment phasePrimaryReference (placebo); MMRM 48-week progression 13.93 (SE 3.42) UMSARS total pointsMMRM 48-week progression 13.32 (SE 2.86) UMSARS total points; Bayesian effect parameter 0.81 = 19% slowing vs placeboBayesian: 19% slowing of clinical progression (2.5th-97.5th percentile -13% to 44%); effect parameter 0.81 (2.5th-97.5th percentile 0.56-1.13); Bayesian probability of true slowing 89.4% (predefined significance threshold 97.5% — NOT met)
Change from baseline in modified UMSARS (mUMSARS, part I scores 0/1 collapsed to 1) — key secondary endpointSecondary27% slower progression (Bayesian probability 97.0%); MMRM change 6.48 (SE 1.33) amlenetug vs 7.07 (SE 1.62) placebo; treatment effect -0.59 (95% CI -3.56 to 2.38)
Change from baseline in UMSARS part ISecondary22% slower progression (Bayesian probability 91.6%); MMRM change 6.98 (SE 1.59) vs 7.64 (SE 1.92); treatment effect -0.66 (95% CI -4.09 to 2.77)
Change from baseline in UMSARS part IISecondary17% slower progression (Bayesian probability 84.4%); MMRM change 6.76 (SE 1.83) vs 7.16 (SE 2.15); treatment effect -0.40 (95% CI -4.04 to 3.24)
MRI-derived percentage volume change from baseline to week 48 (MMRM treatment effect, amlenetug minus placebo)SecondaryPrimary ROIs: pons 1.95 (95% CI -0.42 to 4.32), cerebellar white matter 2.61 (-1.57 to 6.79), cerebellar grey matter -1.41 (-3.27 to 0.46). Secondary ROIs: caudate 0.71 (-1.10 to 2.52), putamen -0.25 (-2.31 to 1.80), brainstem 1.82 (0.09 to 3.56), total grey matter 0.60 (-0.34 to 1.55). Non-significant signals of less volume loss in pons, cerebellar white matter, and brainstem
Change in plasma neurofilament light chain (NfL) concentrationSecondaryNo relevant changes from baseline in either group
Change from baseline to week 48 in UMSARS part IV (global disability) and Schwab and England Activities of Daily Living (SE-ADL) score (MMRM)SecondaryUMSARS IV: 1.13 (SE 0.30) vs 1.35 (SE 0.35); treatment effect -0.22 (95% CI -0.76 to 0.32). SE-ADL: -20.82 (SE 6.52) vs -23.56 (SE 7.40); treatment effect 2.74 (95% CI -7.90 to 13.38)
Clinician, Patient, and Observer Global Impression of Severity (CGI-S, PGI-S, OGI-S, 5-point scales) MMRM treatment effectsSecondaryCGI-S -0.24 (95% CI -0.52 to 0.04) favouring amlenetug; PGI-S 0.25 (-0.11 to 0.61) favouring placebo; OGI-S -0.16 (-0.68 to 0.37) favouring amlenetug
Quality of life (EQ-5D-5L VAS) and prespecified disease milestones (speech, swallowing, walking, falls) from UMSARS part I (MMRM treatment effects)SecondaryEQ-5D-5L VAS -0.64 (95% CI -10.87 to 9.58); speech -0.42 (-0.86 to 0.02); swallowing 0.34 (-0.18 to 0.85, favouring placebo); walking -0.17 (-0.60 to 0.26); falls -0.14 (-0.83 to 0.55)
Amlenetug plasma and CSF concentrations at week 48; CSF-to-plasma ratio; free CSF α-synuclein change (target engagement)SecondaryPlasma amlenetug: mean 390,751 ng/mL, median 356,000 ng/mL. CSF amlenetug: mean 1531 ng/mL, median 1350 ng/mL. CSF-to-plasma ratio 0.4%. Free (unbound) CSF α-synuclein reduced 33.0% from baseline (25th-75th percentile -43.8 to -23.5), indicating central target engagement
Any treatment-emergent adverse event (TEAE)SafetyAmlenetug: 40 (100%; 95% CI 91–100) · Placebo: 20 (95%; 95% CI 77–99)
Serious treatment-emergent adverse eventSafetyAmlenetug: 12 (30%; 95% CI 18–45) · Placebo: 7 (33%; 95% CI 17–55)
Drug-related TEAEsSafetyAmlenetug: 25 (63%; 95% CI 47–76) · Placebo: 8 (38%; 95% CI 21–59)
TEAEs potentially related to hypersensitivitySafetyAmlenetug: 4 (10%; 95% CI 4–23) · Placebo: 1 (5%; 95% CI 1–23)
TEAEs leading to discontinuationSafetyAmlenetug: 3 (8%; 95% CI 3–20) · Placebo: 2 (10%; 95% CI 3–29)
Deaths (adverse events leading to death)SafetyAmlenetug: 2 (5%; 95% CI 1–17) — cardiopulmonary failure, asphyxia due to foreign body · Placebo: 2 (10%; 95% CI 3–29) — cardiorespiratory arrest (only death assessed as possibly drug-related), sepsis
Completion of double-blind treatmentSafety48/61 (79%) completed; mean 56 weeks (SD 13) treatment
COVID-19 infectionAdverseAmlenetug 11 (28%) vs Placebo 5 (24%)
Back painAdverseAmlenetug 6 (15%) vs Placebo 2 (10%)
HeadacheAdverseAmlenetug 5 (13%) vs Placebo 1 (5%)
Urinary tract infectionAdverseAmlenetug 4 (10%) vs Placebo 3 (14%)
Peripheral oedemaAdverseAmlenetug 4 (10%) vs Placebo 1 (5%)
FlushingAdverseAmlenetug 4 (10%) vs Placebo 0
HypertensionAdverseAmlenetug 4 (10%) vs Placebo 0

Subgroup Analysis

Prespecified post-hoc analysis in participants with baseline UMSARS total score <40 (n=42, 69% of FAS): mean 48-week UMSARS total progression 12.21 (SE 3.08) amlenetug vs 18.61 (SE 4.16) placebo. Bayesian effect parameter 0.63 (2.5th-97.5th percentile 0.41-0.94) equated to 37% slower clinical progression (2.5th-97.5th percentile 6 to 59%) with Bayesian probability 98.9% (nominal significance). Domain-specific slowing: mUMSARS 42% (13 to 64), probability 99.6%; UMSARS part I 42% (10 to 65), probability 99.3%; UMSARS part II 30% (-6 to 56), probability 95.5%


Criticisms

  • Small sample size (n=61 treated) limited statistical power, particularly for placebo group (n=21)
  • Primary endpoint not met by predefined Bayesian threshold (89.4% vs 97.5% required)
  • Wide credible interval (2.5th-97.5th percentile -13% to 44% slowing) reflects substantial uncertainty about true effect
  • Protocol amendment removed exclusion criteria for advanced disease features after trial initiation
  • Trial limited to USA and Japan, limiting generalizability
  • Sponsor-funded trial (H Lundbeck) with multiple sponsor-affiliated authors

Funding

H Lundbeck

Based on: AMULET (The Lancet Neurology, 2026)

Authors: Kjærsgaard L, Wiedemann J, Singer W, ..., Luthman J; AMULET Trial Group

Citation: The Lancet Neurology, Volume 25, Issue 6, June 2026, Pages 560-570

Content summarized and formatted by NeuroTrials.ai.