GLP-1 Medications and Alzheimer's Disease: What the Research Shows

For nearly a decade, the question of whether GLP-1 receptor agonists could slow Alzheimer's disease has been one of the most closely watched stories in neurology. Two large Phase 3 trials of oral semaglutide — EVOKE and EVOKE+ — were designed to answer it. In March 2026, Novo Nordisk announced the results: neither trial slowed cognitive decline.

The story is more nuanced than that headline suggests. Biomarkers of tau pathology and neurodegeneration improved in patients on semaglutide. A separate Phase 2b trial of liraglutide, published in Nature Medicine in late 2025, missed its primary endpoint but reported significantly less brain volume loss and a slower decline on a cognitive subscale. Real-world cohort studies continue to find that people taking GLP-1 drugs for diabetes develop dementia less often than people on other diabetes medications.

Here is what the science actually shows about GLP-1 medications and Alzheimer's disease in 2026.

Why Researchers Targeted GLP-1 for Alzheimer's

Alzheimer's disease is driven by extracellular amyloid-beta plaques, intracellular tau tangles, chronic neuroinflammation, and progressive synaptic loss. Insulin resistance in the brain — sometimes called "type 3 diabetes" — appears to amplify all four processes. That overlap is what put GLP-1 receptor agonists on the radar in the first place.

GLP-1 receptors are expressed in the hippocampus, cortex, and other regions affected early in Alzheimer's. Preclinical work over the past 15 years identified several mechanisms that made GLP-1 drugs biologically plausible candidates for disease modification:

Reduced amyloid and tau pathology — GLP-1 signaling activates PI3K/Akt and inhibits GSK-3β, the enzyme that hyperphosphorylates tau.
Microglial modulation — GLP-1 agonists shift microglia away from chronic pro-inflammatory states implicated in synapse loss.
Mitochondrial and bioenergetic support — Better brain insulin sensitivity may correct the glucose hypometabolism seen on FDG-PET scans of Alzheimer's patients.
Neurotrophic signaling — Activation of CREB/BDNF pathways supports synaptic maintenance and neuronal survival.

Evidence: "GLP-1RAs reduce pro-inflammatory cytokines, inhibit harmful microglial activation, decrease amyloid-β and tau aggregation, and provide neurotrophic support through CREB/BDNF pathways, providing a multi-target rationale for Alzheimer's disease." — Du H, et al. Front Endocrinol. 2022;13:1033479. DOI: 10.3389/fendo.2022.1033479

The mechanistic case was strong, and it was reinforced by epidemiology. The question was whether any of it would survive a properly powered randomized trial in symptomatic patients. The parallels with the GLP-1 story in Parkinson's disease are striking — and the lessons turned out to be similar.

The Real-World Signal: Cohort Studies on Dementia Risk

Long before EVOKE reported, large medical-record studies had built a consistent — if observational — case that people taking GLP-1 drugs for type 2 diabetes developed dementia less often than matched controls.

A 2025 propensity-matched analysis of 87,229 patient pairs published in BMJ Open Diabetes Research & Care compared GLP-1 receptor agonists to metformin as first-line therapy. After adjustment, GLP-1 use was associated with significantly lower rates of all-cause dementia and Alzheimer's disease specifically.

Evidence: "Among 87,229 matched patients per cohort, GLP-1 receptor agonist use was associated with a significantly lower risk of overall dementia (adjusted HR 0.90; 95% CI 0.85 to 0.95) and Alzheimer's disease (adjusted HR 0.88; 95% CI 0.83 to 0.94) compared with metformin." — Tang H, et al. BMJ Open Diabetes Res Care. 2025;13(4):e004902. DOI: 10.1136/bmjdrc-2025-004902

A separate population-based cohort study from Taiwan, following 109,778 adults with type 2 diabetes between 2013 and 2021, found a similar 10% reduction in dementia risk for GLP-1 users versus non-users.

Evidence: "Among 109,778 individuals with type 2 diabetes, GLP-1 receptor agonist use was associated with a reduced risk of dementia compared with non-use (adjusted HR 0.90; 95% CI 0.83 to 0.97)." — Liao YT, et al. Diabetes Metab Res Rev. 2025;41(5):e70058. DOI: 10.1002/dmrr.70058

These studies are subject to all the usual caveats of observational research — healthy-user bias, residual confounding, indication for treatment — but the consistency of the signal across countries, comparators, and time windows was difficult to ignore. It is what made the Phase 3 trials so anticipated.

ELAD: Liraglutide in Mild-to-Moderate Alzheimer's

The first major Phase 2 trial in symptomatic Alzheimer's was ELAD, led by Paul Edison at Imperial College London. ELAD randomized 204 patients with mild-to-moderate Alzheimer's disease — without diabetes — to liraglutide 1.8 mg daily or placebo for 12 months across 24 UK clinics. Results were published in Nature Medicine in late 2025.

The primary endpoint was change in cerebral glucose metabolism on FDG-PET. The trial missed it: the difference in glucose metabolism between groups was -0.17 (95% CI: -0.39 to 0.06; P=0.14), favoring placebo numerically but not statistically.

Secondary and exploratory outcomes told a more interesting story:

MRI volumes — Liraglutide-treated patients had nearly 50% less volume loss in frontal, temporal, parietal, and total gray matter over the trial.
ADAS-Exec — Cognitive performance on the Alzheimer's Disease Assessment Scale executive domain favored liraglutide (between-group difference 0.15; 95% CI: 0.03 to 0.28; P=0.01).
Overall cognitive decline — Decline on a composite cognitive measure was approximately 18% slower in the liraglutide group at 12 months.

Evidence: "In participants with mild to moderate Alzheimer's disease, liraglutide did not significantly affect cerebral glucose metabolism, but treatment was associated with reduced regional brain atrophy and slower decline on the executive subscale of ADAS-cog. The drug was safe and well tolerated." — Edison P, et al. Nat Med. 2025. DOI: 10.1038/s41591-025-04106-7

ELAD was a Phase 2b trial with 204 patients. It was not powered to settle the question; it was powered to justify a Phase 3 trial. By the time it was published, two Phase 3 trials in early Alzheimer's were already nearing readout.

EVOKE and EVOKE+: The Phase 3 Verdict

The EVOKE and EVOKE+ trials were the largest and longest GLP-1 trials ever conducted in a neurodegenerative disease. Together, they randomized approximately 3,800 patients with early-stage symptomatic Alzheimer's disease to oral semaglutide 14 mg or placebo for 104 weeks. Topline results were announced in March 2026, with the full paper published in The Lancet.

The primary outcome — change from baseline in the Clinical Dementia Rating Sum of Boxes (CDR-SB) score at week 104 — did not differ significantly between groups in either trial. Patients on semaglutide declined at the same rate as patients on placebo.

Evidence: "Neither trial confirmed superiority of oral semaglutide 14 mg over placebo on the primary outcome of change in CDR-SB at week 104. Oral semaglutide did not slow cognitive or functional decline in early symptomatic Alzheimer's disease." — EVOKE and EVOKE+ Investigators. Lancet. 2026. Article

The biomarker substudy complicated the verdict. In cerebrospinal fluid, semaglutide produced statistically significant reductions of up to 10% in phosphorylated tau (p-tau181, p-tau217), total tau, neurogranin (a synaptic marker), and YKL-40 (a glial inflammation marker). The drug was clearly engaging Alzheimer's-relevant biology — it just was not translating into clinical benefit over two years.

The 12-month open-label extension was discontinued on the basis of these results. Novo Nordisk has not announced plans for further Alzheimer's-specific trials.

What the EVOKE biomarker signal might mean

The disconnect between biomarkers and clinical outcomes is the central puzzle of the EVOKE results. Several interpretations are plausible:

Wrong stage of disease. By the time patients have CDR-SB scores high enough to qualify for an early-Alzheimer's trial, irreversible synaptic and neuronal loss may already dominate.
Wrong endpoint horizon. Two years may be too short to translate a 10% reduction in tau biomarkers into a measurable CDR-SB difference.
Wrong dose or wrong molecule. Oral semaglutide reaches lower CSF concentrations than injectable formulations, and 14 mg may not be the optimal dose for central nervous system penetration.
Real biological effect, insufficient mechanism. GLP-1 signaling may modulate tau and neuroinflammation but leave amyloid pathology and downstream synaptic loss largely untouched.

None of these are answered by EVOKE itself. They are the questions the next generation of trials will need to address.

Where the Science Stands in 2026

Three sets of findings now coexist in the literature, and the honest summary is that they do not yet point in the same direction.

Evidence	Population	Key finding	Interpretation
Real-world cohorts (BMJ, Diabetes Metab Res Rev, JAMDA)	Type 2 diabetes, no dementia at baseline	10-40% lower dementia risk on GLP-1 vs comparators	Possible prevention signal; observational
ELAD Phase 2b (Nat Med, 2025)	Mild-moderate AD, no diabetes	Primary endpoint missed; less brain atrophy and slower ADAS-Exec decline	Suggestive but underpowered
EVOKE/EVOKE+ Phase 3 (Lancet, 2026)	Early symptomatic AD	No CDR-SB benefit; biomarker reductions	Negative for disease modification in symptomatic disease

The reasonable read is that GLP-1 receptor agonists do something measurable to Alzheimer's biology — tau, neuroinflammation, synaptic markers — but that something does not translate into slower clinical decline once symptoms are established. Whether earlier intervention, longer duration, or combination therapy with amyloid- or tau-targeted drugs could change that remains untested.

For people currently taking GLP-1 medications for weight loss or diabetes, the practical implications are limited. There is no evidence to start a GLP-1 drug for the purpose of preventing or treating Alzheimer's disease. There is also no evidence that being on one for diabetes or obesity puts cognitive function at risk — safety data across the EVOKE program were reassuring, consistent with the broader GLP-1 safety profile discussed in our overview of GLP-1 medications and brain health.

Conclusion

The EVOKE Phase 3 trials closed the book on the simplest version of the GLP-1-for-Alzheimer's hypothesis: oral semaglutide 14 mg does not slow cognitive or functional decline in patients with early symptomatic Alzheimer's disease over two years. The biomarker signal in the same trials, the secondary endpoints from ELAD, and a decade of consistent observational data on dementia risk reduction suggest the underlying biology is real — but converting it into clinical benefit will require different patients, different timing, or different drug combinations.

For now, GLP-1 receptor agonists remain a treatment for obesity and type 2 diabetes, not for Alzheimer's disease. The next chapter of this research will likely focus on preclinical and prodromal populations, combination regimens with anti-amyloid therapies, and second-generation GLP-1/GIP and triagonist molecules that achieve higher central nervous system exposure.

References

EVOKE and EVOKE+ Investigators. 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. Lancet. 2026. Article
Edison P, et al. Liraglutide in mild to moderate Alzheimer's disease: a phase 2b clinical trial. Nat Med. 2025. DOI: 10.1038/s41591-025-04106-7
Tang H, et al. Evaluating GLP-1 receptor agonists versus metformin as first-line therapy for reducing dementia risk in type 2 diabetes. BMJ Open Diabetes Res Care. 2025;13(4):e004902. DOI: 10.1136/bmjdrc-2025-004902
Liao YT, et al. Impact of glucagon-like peptide-1 receptor agonists on the dementia incidence in patients with type 2 diabetes mellitus: a population-based longitudinal cohort study. Diabetes Metab Res Rev. 2025;41(5):e70058. DOI: 10.1002/dmrr.70058
Du H, et al. The mechanism and efficacy of GLP-1 receptor agonists in the treatment of Alzheimer's disease. Front Endocrinol. 2022;13:1033479. DOI: 10.3389/fendo.2022.1033479
Cummings J, et al. EVOKE and EVOKE+: design of two large-scale, double-blind, randomised, placebo-controlled, phase 3 studies evaluating efficacy, safety, and tolerability of oral semaglutide in people with early Alzheimer's disease. Alzheimers Dement. 2025. PubMed

Last updated: 2026-05-15 Medical review: Dr. James Chen, MD, PhD, FACE