Amyloid hopes face harder safety test

Patients offered a new generation of Alzheimer’s drugs face a sharper debate over whether modest biological gains justify the risks of brain swelling, bleeding and intensive monitoring, after a large review found no clinically meaningful benefit from medicines designed to clear amyloid beta from the brain.

The review assessed 17 randomised trials involving 20,342 participants with Alzheimer’s disease or early symptoms linked to the condition. It found that anti-amyloid monoclonal antibodies reduced amyloid deposits, but did not produce a meaningful improvement in cognition, daily functioning or dementia severity when compared with placebo. The findings challenge one of the most heavily backed theories in Alzheimer’s research: that removing amyloid plaques can materially slow the disease for patients.

Safety concerns were more pronounced. Treated patients were more likely to develop amyloid-related imaging abnormalities, known as ARIA, which include brain swelling and small bleeds visible on scans. Many cases did not produce obvious symptoms, but serious and life-threatening events have been recorded, particularly among patients carrying two copies of the APOE4 gene, a known risk factor for Alzheimer’s and ARIA.

The findings arrive at a sensitive point for drugmakers and regulators. Lecanemab, sold as Leqembi by Eisai and Biogen, and donanemab, sold as Kisunla by Eli Lilly, have become symbols of a new disease-modifying era after decades of failed Alzheimer’s trials. Both are approved in the United States for early Alzheimer’s disease, while UK regulators licensed lecanemab in August 2024 and donanemab in October 2024 for selected adults at early stages of the disease.

Access remains constrained. Health technology assessors in England and Wales have judged the benefits of lecanemab and donanemab too small to justify the additional cost and service burden for the NHS, which would need capacity for amyloid testing, genetic risk assessment, repeated infusions and MRI monitoring. Final draft guidance in June 2025 maintained that position, though the assessment process has continued after manufacturer appeals.

The dispute is not simply scientific. It also concerns what counts as a meaningful outcome for patients and families. Earlier pivotal trials showed statistically significant slowing of decline for lecanemab and donanemab in early-stage disease, but the absolute differences on clinical rating scales were small. Supporters argue that slowing deterioration, even modestly, can matter when treatment begins before major functional loss. Critics counter that the gains are too limited, uncertain in everyday practice and accompanied by risks that may be unacceptable for a frail population.

A further complication is the design of the review. Some dementia specialists have cautioned that pooling older anti-amyloid antibodies that failed in trials with newer approved drugs may understate the effect of the latest treatments. Others say the broad analysis is precisely what is needed because the amyloid strategy has consumed vast research investment while leaving many patients with little practical improvement.

For clinicians, the immediate issue is patient selection. Anti-amyloid drugs are intended for people with mild cognitive impairment or mild dementia caused by Alzheimer’s disease, not for advanced dementia or memory loss from other causes. Treatment usually requires confirmation of amyloid pathology through PET scans or cerebrospinal fluid tests, exclusion of patients with certain bleeding risks, and continuing surveillance for ARIA.

Families are also being asked to weigh burdens that extend beyond drug safety. Intravenous treatment schedules can mean repeated hospital visits over many months. MRI monitoring adds cost and logistical pressure. Patients on anticoagulants or those with significant vascular changes in the brain may face higher danger. These practical barriers have raised questions about whether private access could widen inequality, leaving well-resourced patients able to pursue treatment while public systems hesitate.

Alzheimer’s researchers are not abandoning amyloid, but the field is moving towards a broader model. Tau pathology, neuroinflammation, vascular health, metabolic dysfunction and synaptic damage are gaining attention as possible targets. Blood-based biomarkers are also advancing, with the potential to identify disease earlier and reduce reliance on expensive scans, though wider clinical use will require rigorous validation.