The Journal of Prevention of Alzheimer's Disease March 18, 2022

Two Randomized Phase 3 Studies of Aducanumab in Early Alzheimer's Disease

Samantha Budd Haeberlein, P. S. Aisen, F. Barkhof, S. Chalkias, T. Chen, S. Cohen, et al.

Bottom Line

In two identically designed phase 3 trials halted early for futility, aducanumab robustly reduced brain amyloid but produced discordant clinical results—showing a significant 22% reduction in cognitive decline in EMERGE but no clinical benefit in ENGAGE—igniting intense debate upon its subsequent FDA accelerated approval.

Key Findings

1. In the EMERGE trial, high-dose aducanumab demonstrated a statistically significant 22% reduction in clinical decline compared to placebo on the primary endpoint, the CDR-SB, at week 78 (difference of -0.39; 95% CI, -0.69 to -0.09; P=0.012).
2. In the ENGAGE trial, high-dose aducanumab failed to demonstrate any clinical benefit on the primary endpoint (CDR-SB difference of 0.03; 95% CI, -0.26 to 0.33; P=0.833).
3. Both EMERGE and ENGAGE demonstrated robust, dose- and time-dependent target engagement, with aducanumab successfully reducing amyloid beta plaques on PET imaging.
4. The incidence of Amyloid-Related Imaging Abnormalities with edema (ARIA-E) was notably high, occurring in 35.2% of patients in the 10-mg/kg aducanumab arm compared to 2.7% in the placebo group.
5. Among ApoE ε4 carriers receiving the high-dose (10 mg/kg) treatment, the incidence of ARIA-E was disproportionately higher (43.0%) compared to non-carriers (20.3%).

Study Design

Design
Phase 3 RCT
Double-Blind
Sample
3,285
Patients
Duration
78 wk
Median
Setting
20 countries
Population Patients aged 50 to 85 years with confirmed amyloid pathology who met clinical criteria for mild cognitive impairment due to Alzheimer's disease or mild Alzheimer's disease dementia.
Intervention Aducanumab IV infusion every 4 weeks (low dose 3/6 mg/kg or high dose 10 mg/kg target) over 76 weeks.
Comparator Placebo IV infusion every 4 weeks over 76 weeks.
Outcome Change from baseline to week 78 on the Clinical Dementia Rating Sum of Boxes (CDR-SB) score.

Study Limitations

Discordant clinical efficacy outcomes between two identically designed Phase 3 trials (EMERGE being positive, ENGAGE being negative) left the true clinical benefit of aducanumab highly uncertain.
Both trials were prematurely halted based on a prespecified futility analysis; positive findings in EMERGE emerged only from complex post-hoc analyses on a truncated, later-acquired dataset.
The high incidence of symptomatic and asymptomatic ARIA-E required dose modifications and MRI monitoring, potentially unblinding investigators and patients and introducing bias.
Mid-trial protocol amendments (specifically Protocol Version 4, which allowed ApoE4 carriers to titrate up to the maximum 10-mg/kg dose) complicated the interpretation of the dose-response relationship and the differing outcomes of the two trials.
The study population was overwhelmingly white, severely limiting the generalizability of both safety and efficacy outcomes across racially and ethnically diverse patient groups.

Clinical Significance

Aducanumab's development served as a polarizing yet critical inflection point in Alzheimer's therapeutics. By definitively proving that a monoclonal antibody could clear brain amyloid, it provided major validation for the amyloid cascade hypothesis. However, the discordant clinical data between EMERGE and ENGAGE raised profound questions about whether biomarker improvements (plaque clearance) reliably translate into meaningful cognitive benefits. The intense safety scrutiny also established the modern clinical framework for routine MRI monitoring and ApoE4 genetic screening for ARIA, which paved the pathway for subsequent, more definitively efficacious anti-amyloid therapies like lecanemab and donanemab.

Historical Context

Despite the conflicting trial outcomes and near-unanimous opposition from its advisory committee (which voted 10-0 against approval, leading to three high-profile resignations), the FDA granted aducanumab (Aduhelm) accelerated approval in June 2021 based solely on its plaque-clearing biomarker effect. This controversial regulatory decision sparked fierce international backlash regarding approval standards, drug pricing, and patient safety. Ultimately, the Centers for Medicare & Medicaid Services (CMS) restricted coverage exclusively to patients in clinical trials, stalling the drug's commercial viability. Following the successful development of the clearer-acting agent lecanemab (Leqembi), Biogen officially discontinued the development and commercialization of Aduhelm in early 2024.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

How does aducanumab's mechanism of action relate to the amyloid cascade hypothesis of Alzheimer's disease, and why were these trials specifically designed to target patients in the 'early' stages of the disease?

Key Response

This question tests foundational understanding of Alzheimer's pathophysiology. Aducanumab is a monoclonal antibody that targets aggregated forms of beta-amyloid (oligomers and fibrils) to promote microglial clearance. The amyloid hypothesis postulates that amyloid accumulation is an early, upstream event that drives subsequent tau pathology and neurodegeneration. Targeting patients in early stages (Mild Cognitive Impairment or mild dementia) is critical because once widespread neurodegeneration occurs, removing amyloid is unlikely to reverse cognitive deficits.

Resident
Resident

A patient with mild cognitive impairment due to Alzheimer's disease is starting aducanumab. What are the specific baseline investigations and ongoing monitoring requirements needed to safely manage the risk of Amyloid-Related Imaging Abnormalities (ARIA)?

Key Response

Residents must understand the practical management of novel therapeutics and their side effects. ARIA (edema/effusion [ARIA-E] or hemorrhage/hemosiderosis [ARIA-H]) is the most significant adverse effect of anti-amyloid antibodies. Safe prescribing requires a baseline MRI within 1 year before initiation and surveillance MRIs prior to the 5th, 7th, 9th, and 12th infusions. Knowing when to suspend or discontinue dosing based on the radiographic severity of ARIA and the presence of clinical symptoms is essential for patient safety.

Fellow
Fellow

How did the sponsor biologically and pharmacologically explain the discordant clinical outcomes between the EMERGE and ENGAGE trials, despite both cohorts demonstrating robust, dose-dependent amyloid clearance on PET imaging?

Key Response

Fellows need to grapple with complex, contradictory trial data. The sponsor's post-hoc explanation for the discordance was that a mid-trial protocol amendment allowed ApoE4 carriers to titrate up to the highest dose (10 mg/kg). Because EMERGE started slightly later than ENGAGE, a larger proportion of patients in EMERGE received the highest dose for a longer duration before the data cutoff. This highlights the nuanced relationship between pharmacokinetics (cumulative drug exposure), target engagement (amyloid clearance), and downstream clinical efficacy.

Attending
Attending

In the context of the FDA's controversial accelerated approval of aducanumab based on a surrogate biomarker (amyloid reduction), how should clinicians structure shared decision-making conversations regarding the absolute clinical benefit (a 0.39-point difference on the 18-point CDR-SB scale in one trial) versus the known risks and burdens?

Key Response

Attendings must translate complex, controversial evidence into patient care. The 22% relative reduction in cognitive decline in EMERGE translates to a very small absolute difference (0.39 points on the Clinical Dementia Rating-Sum of Boxes). Clinicians must help patients weigh this arguably marginal and uncertain clinical benefit against the high incidence of ARIA (up to 40%), the burden of monthly IV infusions, frequent MRI monitoring, and substantial financial costs.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

From a statistical and trial design perspective, what are the methodological pitfalls of halting a trial early for futility based on an interim pooled analysis, and subsequently utilizing a post-hoc analysis with additional accumulated data to claim drug efficacy?

Key Response

PhDs must critically evaluate research methodology and statistical integrity. The EMERGE and ENGAGE trials were stopped based on a pre-planned interim futility analysis. The subsequent claim of efficacy relied on a larger, later dataset where the high-dose group in EMERGE crossed the significance threshold. Re-analyzing data after a futility stop fundamentally inflates the Type I error rate, introduces operational bias, and renders the calculated p-values statistically uninterpretable without severe penalty adjustments.

Journal Editor
Journal Editor

How does the high incidence of ARIA in the high-dose aducanumab arms threaten the blinding integrity of both trials, and what methodological safeguards or sensitivity analyses would you require as a reviewer to assess the impact of this potential unblinding on a subjective endpoint like the CDR-SB?

Key Response

Editors and peer reviewers must identify fatal flaws such as unblinding. Since ARIA occurred in a large proportion of treated patients—often necessitating extra MRIs, dose interruptions, or causing mild symptoms—patients, caregivers, and potentially raters could deduce treatment assignment. Because the primary endpoint (CDR-SB) relies on subjective caregiver and patient reports, this unblinding introduces severe expectancy bias. A tough reviewer would demand sensitivity analyses comparing the primary outcome in patients with and without ARIA to evaluate this bias.

Guideline Committee
Guideline Committee

Given that aducanumab received FDA accelerated approval based on the surrogate endpoint of amyloid plaque reduction rather than definitive clinical benefit, how should clinical practice guidelines synthesize the discordant EMERGE and ENGAGE data to establish 'Appropriate Use Criteria' and grade the strength of recommendation for routine clinical practice?

Key Response

Guideline committees must decide how to translate controversial regulatory decisions into standard of care. Despite FDA approval, many professional societies (e.g., American Academy of Neurology) and payers (e.g., CMS) restrict its use strictly to patients matching the clinical trial criteria or those enrolled in subsequent randomized trials. The committee must weigh the low certainty of clinical benefit and conflicting Phase 3 evidence against the confirmed biomarker efficacy, ultimately determining if the strength of recommendation warrants broad clinical adoption or highly restricted, monitored use.

Clinical Landscape

Noteworthy Related Trials

2018

EXPEDITION3

n = 2,129 · NEJM

Tested

Solanezumab 400 mg IV every 4 weeks

Population

Patients with mild Alzheimer's disease

Comparator

Placebo

Endpoint

Change in ADAS-cog14 at 80 weeks

Key result: Solanezumab did not significantly slow cognitive decline compared to placebo in mild Alzheimer's disease.
2023

CLARITY AD

n = 1,795 · NEJM

Tested

Lecanemab 10 mg/kg biweekly

Population

Patients with early Alzheimer's disease and confirmed amyloid

Comparator

Placebo

Endpoint

Change in CDR-SB at 18 months

Key result: Lecanemab reduced cognitive decline by 27 percent on the CDR-SB scale compared to placebo at 18 months.
2023

TRAILBLAZER-ALZ 2

n = 1,736 · JAMA

Tested

Donanemab IV every 4 weeks

Population

Patients with early symptomatic Alzheimer disease with amyloid and tau pathology

Comparator

Placebo

Endpoint

Change in iADRS at 76 weeks

Key result: Donanemab significantly slowed clinical progression by 35 percent on the iADRS compared to placebo over 76 weeks.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis