The EVOKE and EVOKE+ trials delivered a sobering verdict: oral semaglutide failed to slow clinical progression in early symptomatic Alzheimer's disease, despite producing measurable positive effects on AD-related biomarkers. This disconnection — biological signal without functional benefit — frames the central tension this analysis explores across the GLP-1 receptor agonist (GLP-1RA) literature. Observational data consistently show lower dementia incidence among GLP-1RA users versus other antidiabetic treatments, and preclinical work suggests reduced amyloid and tau pathology, improved mitochondrial function, and dampened neuroinflammation. Yet limited blood-brain barrier penetrance of most GLP-1RAs and the EVOKE null results force a recalibration of mechanism versus clinical hope.
The EVOKE failure likely reflects a fundamental staging problem rather than a drug class failure. Intervening in symptomatic Alzheimer's — when neurodegeneration is already entrenched — may simply be too late for metabolic reprogramming to reverse cognitive trajectories. The real opportunity may lie in the preclinical window, particularly in metabolically vulnerable populations with insulin resistance, vascular dysfunction, and chronic inflammation — the exact phenotype where GLP-1RAs demonstrably excel systemically. This shifts the therapeutic hypothesis from direct neuroprotection to upstream metabolic risk reduction, a meaningfully different and more tractable target. For clinicians and researchers, the implication is clear: GLP-1RA trials enriched for metabolic syndrome at preclinical AD stages are now the scientifically justified next step. Biomarker-positive, cognitively normal adults with metabolic syndrome represent the highest-yield population. This is not a failed drug class — it is a mismatch of timing and target.
