A comprehensive analysis of 17 trials involving over 20,000 participants concludes that current anti-amyloid monoclonal antibodies produce minimal cognitive benefit while raising safety and cost concerns.
The debate over drugs that target brain amyloid has taken a new turn after a large systematic review concluded that the group of agents known as anti‑Aβ monoclonal antibodies produce, on average, only modest or no clinically meaningful improvements for people with mild cognitive impairment (MCI) or early Alzheimer’s disease.
The analysis pooled results from multiple trials and found that, despite clear effects on removing amyloid deposits on scans, the translation into better memory, function or independence was very small in common clinical measures.
That conclusion has prompted a heated response from academic researchers who argue that the review mixes drugs with different mechanisms and histories.
Supporters of the review point to its rigorous methodology and the scale of evidence it summarizes, while critics warn against averaging together older failed programmes and newer agents that have shown measurable benefits in individual studies.
Key results from the pooled analysis
The review extracted data from 17 informative clinical trials comprising roughly 20,342 participants, comparing seven antibody programmes — aducanumab, bapineuzumab, crenezumab, donanemab, gantenerumab, lecanemab, and solanezumab — against placebo. At about 18 months of treatment the authors report little or no difference in standard cognitive endpoints. For example, change on the ADAS‑Cog (a commonly used cognitive scale) translated to a standardized mean difference smaller than one point, whereas clinical meaningfulness is typically judged at changes of approximately two to four points. The authors therefore judged that small test score differences were unlikely to reflect real improvements in daily living.
Safety signals, costs and population limits
The review also emphasised safety and practical issues. Treated participants experienced higher rates of amyloid‑related imaging abnormalities (ARIA), a radiological finding that can be associated with swelling or microbleeds; carriers of the APOE ε4 genetic variant were particularly prone to these events. In economic terms the therapies carry steep price tags — roughly €25,000 per year in many settings — which raises questions about value if clinical benefits remain marginal. The authors note limitations in the evidence base, including relatively short follow‑up, younger trial populations than the typical real‑world patient, and frequent industry sponsorship of trials.
Study quality and methodological context
The review follows the standard approach used by the Cochrane collaboration: systematic searches, pre‑specified inclusion criteria, and pooled statistical analysis to estimate average effects. The strength of this approach is its ability to summarise a broad literature and assess bias across trials. Its weakness, critics say, is that pooling distinct molecules with differing targets and binding properties can blur signals from effective agents. The reviewers acknowledged heterogeneity across drugs and trials but maintained that an aggregated assessment is necessary for policy and clinical decision‑making.
Responses from research leaders
Scientists at major dementia centres have disputed the review’s implications. Their principal argument is conceptual: not all anti‑amyloid antibodies are the same. Prominent researchers have pointed to recent trials of lecanemab and donanemab that reported measurable slowing of cognitive decline in large, well‑conducted studies, and they warn that averaging these positive results with older negative trials reduces the apparent benefit. They also highlight unanswered questions about optimal patient selection, dosing, monitoring for ARIA, and long‑term outcomes.
What comes next for patients and policymakers
Both sides agree on priorities: better stratification of participants (including routine APOE genotyping where appropriate), longer follow‑up, more diverse trial populations, and exploration of combination approaches that address non‑amyloid drivers of neurodegeneration. There is also a call for honest conversations about cost, access and how much small statistical effects mean for patients and families. The field is likely to remain contested as regulators, health technology assessors and clinicians weigh limited benefits, safety trade‑offs and affordability against the urgent need for disease‑modifying options.
