Selkoe · EMBO molecular medicine 2016 · narrative review · n=?

The amyloid hypothesis of Alzheimer's disease at 25 years.

Cited 6256 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Narrative review synthesizing mechanistic, genetic, biomarker, and clinical trial literature without systematic review methodology

PubMed 27025652 · doi:10.15252/emmm.201606210 · record verified 2026-08-26

What was done

This narrative review synthesizes 25 years of genetic, biochemical, animal model, biomarker, and clinical trial evidence evaluating the amyloid-beta (Aβ) hypothesis of Alzheimer's disease (AD). The authors evaluate genetic links (APP, presenilin, Down syndrome, and ApoE4), mechanistic effects of human-derived soluble Aβ42 oligomers in rodent models and cell cultures, biomarker trajectories in humans, and clinical trial findings for anti-Aβ antibodies (solanezumab, crenezumab, and aducanumab).

What was found

The review reports that all dominant mutations causing early-onset AD affect APP or presenilin, and ApoE4 predisposes to AD in > 40% of cases by impairing Aβ clearance. Soluble human Aβ42 oligomers decrease synapse number, inhibit long-term potentiation, enhance long-term depression, induce tau hyperphosphorylation, and impair memory in rats. In humans, decreased CSF Aβ42 and amyloid-PET positivity precede clinical symptoms by many years. Post hoc analyses of three Aβ antibody trials suggested slowed cognitive decline in mild AD patients. Quantitative effect sizes, sample numbers, and statistical metrics are not reported in the abstract.

Why it matters

The paper consolidates multidisciplinary evidence supporting Aβ accumulation as an initiating driver of AD pathology, arguing that Aβ dyshomeostasis remains the most validated target for disease-modifying therapeutic intervention.

Limits

This is an unsystematic narrative review rather than a systematic review or meta-analysis. Support for clinical efficacy relies on post hoc subgroup analyses of trials rather than prespecified primary endpoints, and the abstract provides no quantitative effect estimates or trial sample sizes.

Cited by