← Back
NeuroTrials.ai
Neurology Clinical Trial Database

A4

Anti-Amyloid Treatment in Asymptomatic Alzheimer's Disease (A4) - Trial of Solanezumab in Preclinical Alzheimer's Disease

Year of Publication: 2023

Authors: Reisa A. Sperling, Michael C. Donohue, Rema Raman, ..., for the A4 Study Team

Journal: New England Journal of Medicine

Citation: N Engl J Med 2023;389:1096-107

Link: https://doi.org/10.1056/NEJMoa2305032


Clinical Question

Does solanezumab, which targets monomeric amyloid, slow cognitive decline in cognitively unimpaired persons with elevated brain amyloid levels over 240 weeks compared to placebo?

Bottom Line

Solanezumab did not slow cognitive decline compared to placebo over 240 weeks in persons with preclinical Alzheimer's disease. Despite evidence of target engagement (slowed amyloid accumulation), there was no clinical benefit, and numerical worsening was observed in the treatment group.

Major Points

  • 1169 cognitively unimpaired persons with elevated brain amyloid randomized 1:1 to solanezumab vs placebo
  • Dose increased mid-trial from 400 mg to 1600 mg IV every 4 weeks due to concerns from prior trial suggesting inadequate dosing
  • Primary endpoint (PACC score change at 240 weeks) showed no significant difference (P=0.26) with numerically greater decline in solanezumab group
  • Solanezumab slowed amyloid accumulation (11.6 vs 19.3 centiloids) but did not reduce amyloid below baseline levels
  • Tau PET showed similar increases in both groups, suggesting no effect on tau pathology
  • ARIA-E was rare (<1% both groups); ARIA-H rates were similar between groups
  • 33.4% of participants had progression of CDR score to >0 at 240 weeks
  • Results suggest slowing amyloid accumulation without reducing fibrillar amyloid levels below baseline does not slow clinical progression
  • Trial complicated by mid-trial dose increase and COVID-19 pandemic disruptions

Design

Study Type: Phase 3 randomized controlled trial

Randomization: 1

Blinding: Double-blind; participants, investigators, site personnel, coordinating center, and sponsor staff blinded to treatment assignment; independent DSMB reviewed masked data quarterly

Enrollment Period: 2014-2022 (double-blind phase)

Follow-up Duration: 240 weeks (4.5 years)

Centers: 67

Countries: United States, Canada, Australia, Japan

Sample Size: 1169

Analysis: Modified intention-to-treat; primary analysis using constrained longitudinal data analysis with natural cubic splines; MMRM sensitivity analysis; time-to-progression analyzed by Kaplan-Meier; graphical testing scheme for multiple comparisons


Inclusion Criteria

  • Age 65-85 years
  • Living independently without diagnosis of MCI or dementia
  • Global CDR score of 0
  • MMSE score 25-30
  • Wechsler Memory Scale LMDR score 6-18
  • Elevated brain amyloid on 18F-florbetapir PET (SUVR ≥1.15 or SUVR 1.10-1.15 with positive visual read)
  • Study partner available to provide information about participant functioning

Exclusion Criteria

  • LMDR score >18 (to enhance likelihood of enrolling persons with elevated amyloid)
  • Unstable medical conditions
  • Cognitive impairment (CDR >0)

Arms

FieldSolanezumabControl
InterventionSolanezumab IV every 4 weeks; initially 400 mg, increased to 1600 mg mid-trial (2017) when ~970 participants already enrolledMatching placebo IV every 4 weeks
Duration240 weeks240 weeks

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Change in Preclinical Alzheimer Cognitive Composite (PACC) score at 240 weeks - sum of 4 z-scores (FCSRT, Logical Memory IIa delayed recall, Digit Symbol Substitution Test, MMSE)Primary−1.13 (95% CI −1.45 to −0.81)−1.43 (95% CI −1.83 to −1.03)0.26
CFI combined score (participant + partner) change at 240 weeksSecondary1.512.09NS (hierarchical testing failed)
ADL partner score change at 240 weeksSecondary−1.63−2.24NS (hierarchical testing failed)
CDR-SB score change at 240 weeksSecondary0.600.71NS (hierarchical testing failed)
Progression of global CDR score (probability at 240 weeks)Secondary0.320.35NS
Amyloid PET change (centiloids)Secondary+19.3+11.6Mean difference 7.7 (95% CI 5.1 to 10.4)
Any adverse eventAdverse97.6%97.9%
Serious adverse eventAdverse26.7%30.1%
DeathAdverse1.2%1.0%
ARIA-EAdverse0.3%0.2%
ARIA-H with microhemorrhageAdverse32.0%27.6%
ARIA-H with superficial siderosisAdverse3.2%3.3%
Adverse event leading to discontinuationAdverse5.8%4.9%

Subgroup Analysis

Prespecified subgroup analyses showed similar results across subgroups. Post hoc analysis by baseline amyloid tertiles showed higher baseline amyloid associated with faster cognitive decline in both groups, but no formal statistical conclusions can be drawn. In tau PET substudy (114 per group), similar increases in neocortical and medial temporal tau in both groups.


Criticisms

  • Participants not representative of general older population at risk for cognitive decline; few Black participants enrolled
  • Mid-trial dose increase from 400 mg to 1600 mg complicated interpretation
  • COVID-19 pandemic caused widespread site disruption, affecting visit timing and assessments
  • Numerical worsening in solanezumab group was unexpected; possible that 4-fold dose increase severely depleted monomeric Aβ causing harm similar to secretase inhibitors
  • Final-visit PACC observations missing for ~30% of participants in both groups
  • Solanezumab does not bind fibrillar deposits, limiting ability to reduce existing plaque burden
  • Conjecture that monomeric Aβ depletion may have caused cognitive worsening requires further investigation

Funding

Public-private-philanthropic partnership: National Institute on Aging (R01AG063689, U19AG010483, U24AG057437), Eli Lilly (also provided study drug and placebo), Alzheimer's Association, Accelerating Medicines Partnership through FNIH, GHR Foundation, Davis Alzheimer Prevention Program, Yugilbar Foundation, anonymous foundation, additional private donors; in-kind support from Avid Radiopharmaceuticals, Cogstate, Albert Einstein College of Medicine, Foundation for Neurologic Diseases

Based on: A4 (New England Journal of Medicine, 2023)

Authors: Reisa A. Sperling, Michael C. Donohue, Rema Raman, ..., for the A4 Study Team

Citation: N Engl J Med 2023;389:1096-107

Content summarized and formatted by NeuroTrials.ai.