Antiamyloid monoclonal antibodies provide no clinically meaningful benefit for patients with mild cognitive impairment (MCI) or early Alzheimer’s disease (AD), results of a large systematic review showed. However, some experts argue the analysis is skewed because it included failed drug trials in its pooled analysis.
Across 17 randomized controlled phase 3 trials involving more than 20,000 participants and covering seven antiamyloid therapies, treatment effects on cognition, dementia severity, and functional ability were consistently small and fell below established thresholds for clinical importance.
The drugs were also associated with a substantially increased risk for amyloid-related imaging abnormalities (ARIA), particularly ARIA-E (edema).
“Unfortunately, the evidence suggests that these drugs make no meaningful difference to patients,” investigator Francesco Nonino, MD, neurologist and epidemiologist at the IRCCS Institute of Neurological Sciences of Bologna in Bologna, Italy, said in a news release.
“There is now a convincing body of evidence converging on the conclusion that there is no clinically meaningful effect,” he added.
The review’s broad class-level conclusion quickly drew pushback from outside experts who noted that only two of the drugs in the review — lecanemab and donanemab — had positive phase 3 data and current clinical use. The remaining five drugs failed to meet their respective study outcomes.
“This review does not clarify the evidence, it blurs it,” Bart De Strooper, MD, PhD, group leader at the UK Dementia Research Institute at University College London in London, England, said in a statement from the UK nonprofit Science Media Centre. Pooling failed first-generation antibodies with lecanemab and donanemab turns therapeutic progress into “statistical noise,” he added.
David Knopman, MD, professor of neurology at the Mayo Clinic in Rochester, Minnesota, raised a similar concern, adding that the analysis does not resolve the still-open question of whether the newer agents deliver clinically meaningful benefit in practice.
The study was published online on April 16 in the Cochrane Database of Systematic Reviews.
Assessing Clinical Benefit
The amyloid cascade hypothesis has long guided AD drug development, leading to the development of agents designed to clear amyloid-beta early in the disease course in hopes of slowing clinical decline.
The FDA has approved three such drugs: aducanumab in 2021, lecanemab in 2023, and donanemab a year later. But manufacturer Biogen pulled aducanumab from the market in early 2024, a decision the company attributed to efforts to “reprioritize its resources” toward other AD agents.
That marked the end of a rocky road for aducanumab, whose approval was met with criticism from an FDA advisory panel that had recommended the agency reject the drug due to insufficient evidence of clinical efficacy. Three members of the panel — including Knopman — resigned in protest following the decision.
Lecanemab and donanemab had a smoother path to market, with both receiving unanimous recommendations for approval from the FDA expert panel.
However, the question of whether amyloid removal translates into meaningful clinical benefits persists.
To examine this further, investigators conducted a systematic review and meta-analysis of 17 placebo-controlled phase 3 trials covering seven antiamyloid monoclonal antibodies.
The trials were conducted between 2014 and 2024 and included 20,342 participants (mean age, 69.5-73.9 years) with MCI or mild dementia due to AD. About two thirds of patients were women, and most were White. Disease duration ranged from 17 to 52 months and follow-up from 18 to 27 months.
Drugs included donanemab, lecanemab, aducanumab, bapineuzumab, crenezumab, gantenerumab, and solanezumab. Pooled data combined results on the two approved drugs and older agents. All studies compared active treatment with placebo.
Primary outcomes included cognitive function, dementia severity, and functional ability, which were assessed using validated scales such as the Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog; range 0-70), Clinical Dementia Rating Scale Sum of Boxes (CDR-SB; range 0-18), and Alzheimer’s Disease Cooperative Study-Activities of Daily Living (ADCS-ADL and variants).
Limited Clinical Effects
At 18 months, pooled effects on cognitive outcomes were consistently small across studies. The absolute difference favoring treatment was approximately 0.85 points on the ADAS-Cog compared with minimal clinically important differences of 2-4 points depending on disease stage.
Dementia severity showed similarly small effects, with a pooled difference of 0.29 points on the CDR-SB, well below the threshold considered clinically meaningful.
Across ADCS-ADL and related scales, standardized mean differences were trivial, the researchers noted, with absolute differences ranging from approximately 0.07 to 1.90.
At 24 months and beyond, results remained consistent, with no evidence of clinically meaningful improvement. No clinically meaningful differences were observed in behavioral symptoms.
While efficacy findings were modest, safety signals were more pronounced. Risk for any ARIA-E was approximately 10-fold higher in the treatment group at 18 months (relative risk [RR], 10.02; 95% CI, 7.49-13.41) and at 24 months (RR, 7.49). Symptomatic ARIA-E was less frequent but still increased compared with placebo, though most events were asymptomatic.
ARIA-H showed variability across trials, with a small increase in risk at later timepoints (RR, 1.86 beyond 24 months). Reporting of symptomatic ARIA was inconsistent, which limited certainty about clinical impact.
Rates of serious adverse events and mortality were similar between groups at all timepoints.
Findings Spark Controversy
The analysis has sparked debate among AD experts, particularly around whether the trials of older failed agents should be pooled with newer, more successful therapies.
The researchers argued that pooling results across therapies tests whether amyloid removal is an effective strategy. However, some experts say this approach may obscure meaningful differences.
Knopman, who was not part of the study, criticized the decision to combine older failed therapies with newer agents that have shown positive results and De Strooper agreed.
“In the review itself, the authors acknowledge that while several first-generation antibodies failed, newer antibodies have produced positive clinical effects, yet they still pool them together and treat the resulting average as if it were an informative judgment on the entire field,” he said.
“The public is left with the impression that antiamyloid therapy has broadly failed and that the field should move on. A serious review should help readers understand that complexity. This one risks obscuring it,” he added.
Knopman also noted that the review does not address regulatory decision-making because several included therapies were never submitted for approval.
More broadly, he said that ARIA risks appear manageable and that the true clinical value of lecanemab and donanemab are still being evaluated in clinical practice.
“I do not dispute the fact that the effect sizes for lecanemab and donanemab are small, but there are no established guidelines as to what constitutes a meaningful change in the Alzheimer field that has no external standard of benefit size to compare to,” Knopman said.
In the end, he said, the review does not address the question of clinical meaningfulness of the two agents that are currently being used.
“Whether lecanemab and donanemab produce clinically meaningful benefits is a matter of debate that will take several years to clarify,” he said.
Disclosure information for study authors is available in the original study publication. The study was funded by the Drug and Medical Devices Governance Area in Emilia-Romagna in Bologna, IRCCS Institute of Neurological Sciences of Bologna, and the Italian Ministry of Health. Knopman reported being a DSMB for Roche for an investigational agent, trontinemab.
https://www.medscape.com/viewarticle/no-clinical-benefit-antiamyloids-alzheimers-review-concludes-2026a1000c32
