Oral semaglutide, the GLP-1 receptor agonist widely used for diabetes and weight management, did not meaningfully slow clinical progression in people with early-stage symptomatic Alzheimer's disease. That is the conclusion of the evoke and evoke+ trials, two phase 3, randomised, placebo-controlled studies of oral semaglutide 14 mg at a flexible dose, published in The Lancet. Over two years of treatment, the drug failed to separate from placebo on clinical decline.
The result is a disappointment because the rationale looked reasonable. Earlier observational research had hinted that people taking GLP-1 drugs might fare better cognitively, and shifts in biomarkers suggested the class could have neuroprotective effects beyond its metabolic and cardiovascular benefits. Early Alzheimer's has become the focus of drug development on the logic that intervening before substantial damage accumulates offers the best chance of delaying disability - which made a well-tolerated oral pill an attractive candidate to test.
On safety, the trials turned up no surprises. Adverse events were broadly in line with what is already known about semaglutide, including more gastrointestinal complaints and weight-related effects than placebo. In other words, the trials did not fail because the drug proved dangerous in this population; they failed on efficacy.
For patients and families, the practical takeaway is that a repurposed metabolic drug will not be joining the short list of treatments that alter the course of Alzheimer's. The findings also serve as a reminder that promising signals from observational data and biomarker changes do not reliably translate into clinical benefit, and that large randomised trials remain the only way to settle the question.