Early Alzheimer’s Progressed at the Same Pace on Oral Semaglutide

CATEGORY: BRAIN & MENTAL HEALTH

A daily tablet of semaglutide—the same compound sold by injection for type 2 diabetes and weight loss—did not slow early Alzheimer’s disease in two large randomized trials enrolling 3,808 people, published together in The Lancet.¹ Over two years, participants’ scores on an 18-point dementia rating scale worsened by roughly two points whether they took the drug or a placebo; the difference between the groups was about a tenth of a point or less in both trials, well within the range of chance.

Two Trials, Forty Countries, One Tablet a Day

The trials, named evoke and evoke+, enrolled adults aged 55 to 85 with mild cognitive impairment or mild dementia due to Alzheimer’s disease, confirmed by testing for amyloid, the sticky protein that accumulates in Alzheimer’s brains—1,855 people in one trial and 1,953 in the other, at sites in about 40 countries. Each participant was assigned by chance, with neither participants nor staff knowing who received what, to a once-daily 14 mg semaglutide tablet or an identical placebo for 104 weeks. The main outcome was the Clinical Dementia Rating, sum-of-boxes version: a structured interview with the participant and someone who knows them well, scoring memory, orientation, judgment, community activities, home life, and personal care into a single 18-point total, where higher means worse.

A Tenth of a Point Between Drug and Placebo

In evoke, scores worsened by 2.3 points in both groups—the difference between semaglutide and placebo was −0.08 points (95% confidence interval −0.35 to 0.20). In evoke+, scores worsened by 2.2 points with semaglutide and 2.1 with placebo, a difference of 0.10 points (−0.17 to 0.38). Among participants who began with mild cognitive impairment, the drug also did not meaningfully change how many progressed to dementia. The tablet was not sitting idle: people on semaglutide lost 5.8% of their body weight on average while the placebo group gained slightly, and side effects—mostly nausea, reduced appetite, and other stomach complaints—matched the drug’s familiar profile. The medication was doing its known metabolic work; it simply was not slowing Alzheimer’s.

Why Researchers Expected More

The trials were launched on a promising trail of clues. Studies following large populations had found that people taking GLP-1 drugs—the class that mimics a gut hormone to curb appetite and steady blood sugar—went on to be diagnosed with dementia less often, up to 64% less in some analyses, and a smaller trial of the related drug liraglutide had hinted at preserved brain volume.² But people who take these medications differ from people who do not in weight, health care, and dozens of other ways, and following people’s own choices cannot untangle the drug from the person. The randomized comparison could—and the association did not survive it. A biological substudy sharpened the puzzle: in the spinal fluid of 199 participants, levels of tau, a protein tied to Alzheimer’s damage, ran roughly 10% lower on semaglutide, and an inflammation marker in the blood fell substantially—yet the disease’s visible course did not change.¹ The drug behaved like a cleanup crew sweeping some debris from a building that keeps settling at the same rate.

What the Verdict Closes, and What Stays Open

These trials tested one thing: whether semaglutide could slow Alzheimer’s disease that had already begun, over two years, in people with confirmed disease. On that question the answer was clear enough that a planned third year of the trials was discontinued. The result says nothing about the drug’s established benefits for diabetes, obesity, and heart disease, which stand on their own trials. Whether GLP-1 drugs might matter much earlier—before symptoms appear—was not tested here, and Alzheimer’s researchers have noted that decisive negative trials like these redirect a field’s effort rather than end it.

Key Takeaways

• In two randomized trials of 3,808 adults with early Alzheimer’s disease, a daily oral semaglutide tablet did not slow decline: drug and placebo groups both worsened by roughly two points on an 18-point scale over two years.
• The trials were run because population studies had linked GLP-1 drugs to less dementia; the randomized comparison did not confirm a treatment effect—a reminder that association is not causation.
• Alzheimer’s-related proteins fell modestly in a spinal-fluid substudy, but that biological change did not translate into slower memory or functional decline.
• The trials tested treatment of existing early Alzheimer’s over two years; they did not test prevention in people without symptoms.

References

1. Cummings JL, Atri A, Sano M, et al. Efficacy and safety of oral semaglutide 14 mg (flexible dose) in early-stage symptomatic Alzheimer’s disease (evoke and evoke+): two phase 3, randomised, placebo-controlled trials. The Lancet. 2026;407(10544):2167–2179. link
2. Cummings J, Atri A, Feldman HH, et al. evoke and evoke+: design of two large-scale, double-blind, placebo-controlled, phase 3 studies evaluating efficacy, safety, and tolerability of semaglutide in early-stage symptomatic Alzheimer’s disease. Alzheimer’s Research & Therapy. 2025;17(1):14. link