GLP-1 and Alzheimer

In 2024, a large retrospective cohort analysis published in NEJM Evidence examined over one million electronic health records from patients with type 2 diabetes. The study compared rates of new Alzheimer's disease and related dementia diagnoses in semaglutide users versus matched controls receiving seven other diabetes medications (including insulin, metformin, SGLT2 inhibitors, and DPP-4 inhibitors).

The results were striking across every comparison group: semaglutide users had 40-70% lower incidence of Alzheimer's and related dementias compared to all other comparator medications. The effect was consistent across age groups, sex, duration of diabetes, and baseline BMI. The magnitude of risk reduction exceeded that seen with any currently available Alzheimer's prevention strategy, including blood pressure control, exercise, or cognitive training.

This is observational data — it cannot prove that semaglutide caused the risk reduction. Patients prescribed semaglutide may differ from those on other medications in ways not fully captured by the analysis (the "healthy user bias"). However, the consistency of the signal across comparators and the biological plausibility given what we know about GLP-1 receptors in the brain have made this one of the most discussed findings in neuroscience and metabolic medicine.

The first successful GLP-1 brain trial was the liraglutide ELAD (Evaluating Liraglutide in Alzheimer's Disease) Phase 2 trial, which demonstrated that 12 months of liraglutide treatment in early Alzheimer's patients reduced progression of brain pathology markers compared to placebo — a landmark result that validated the approach and opened the door to larger trials.

The EVOKE and EVOKE+ trials became the centerpiece of GLP-1 Alzheimer's research because they tested oral semaglutide in large populations with mild cognitive impairment and early Alzheimer's disease. In November 2025, Novo Nordisk reported topline outcomes showing semaglutide did not achieve statistically significant slowing on the primary CDR-SB endpoint versus placebo. This result lowered near-term expectations for semaglutide as a disease-modifying Alzheimer's therapy, though secondary analyses and subgroup interpretation still matter for the scientific story.

Separately, Phase 2 trials of exenatide in Parkinson's disease have shown neuroprotective effects on motor function — the same GLP-1R-mediated brain protection mechanism at work in a different neurodegenerative context. Tirzepatide (Mounjaro/Zepbound), the dual GIP/GLP-1 receptor agonist, has also initiated brain health trials, and given its superior efficacy on metabolic parameters, researchers are optimistic about its neuroprotective potential.

Observational data in type 2 diabetes patients consistently shows that GLP-1 users have 35-50% lower dementia risk compared to patients on other diabetes treatments — a signal that has now appeared in multiple independent datasets across different countries and healthcare systems, making it increasingly difficult to dismiss as confounding alone.