Ozempic Effects for the Brain? Semaglutide and Tirzepatide Linked to Lower Risk of Dementia and Ischemic Stroke

A new cohort study in JAMA Network Open adds another layer of discussion to the “unexpected benefits” of GLP-1RA drugs (semaglutide and tirzepatide). An analysis of U.S. electronic health records found that in people with type 2 diabetes and obesity, those who started semaglutide/tirzepatide had a lower risk of dementia, ischemic stroke, and all-cause death than comparable patients on other antidiabetic drugs. This is an association, not proof of causation, but the signal is compelling and consistent with biological mechanisms of neurovascular protection.

Background of the study

Type 2 diabetes and obesity increase the risk of vascular dementia and ischemic stroke: chronic inflammation, insulin resistance of the brain, dyslipidemia, hypertension and microcirculation damage contribute to this. Even with good sugar control, some patients still have a high neurovascular risk, so attention is shifting to therapies that simultaneously improve metabolism and body weight and can affect the vascular and neuroinflammatory link.

GLP-1 receptor agonists (GLP-1RAs) - and especially semaglutide, but also tirzepatide (a dual GIP/GLP-1 agonist) - have shown large effects on weight loss, glycemia, and a range of cardiovascular risk factors over the past few years. There is biological logic to their possible "neurovascular benefit": in animal models, activation of GLP-1 signaling reduces neuroinflammation, improves endothelial function, affects amyloid-tau pathology, and in the vascular bed, affects reactivity and thrombosusceptibility. Clinically, this may manifest as fewer strokes and cognitive decline - but such findings require human data.

Until recently, human data were patchy: small cohort samples, short follow-up, heterogeneous endpoints. Observational comparisons are complicated by indication confounding and the “healthy user effect”: who receives modern drugs often differs in access to care, motivation, and concomitant therapy. Modern EHR platforms and propensity score matching reduce these confounds but do not eliminate them completely, so any signals from cohorts should be interpreted as associations rather than proven causation.

Against this background, a large EHR-based study is a step toward answering a practical question: is initiation of semaglutide/tirzepatide therapy in people with T2D and obesity associated with a lower risk of dementia, ischemic stroke, and death compared with alternative antidiabetic regimens? Even with positive associations, the next necessary step is randomized or at least pragmatic trials with cognitive and vascular outcomes to understand how much of the effect is due to weight loss and improvement in risk factors, and how much is due to direct effects of the class on the brain and blood vessels.

The most important thing in two lines

• Sample: 60,860 adults (after 1:1 matching; mean age ~58, 50% female) from the TriNetX network; followed for up to 7 years.
• Results: lower risk of dementia (HR 0.63; 95% CI 0.50-0.81), ischemic stroke (HR 0.81; 0.70-0.93) and death from any cause (HR 0.70; 0.63-0.78). The effect is more pronounced in people ≥60 years old, in women and with a BMI of 30-40.

Who and how was studied

The study is retrospective, using EHR data (TriNetX, December 2017 - June 2024). Two groups were compared:

• GLP-1RA: semaglutide or tirzepatide;
• Other antidiabetics: metformin, sulfonylurea, DPP-4 inhibitors, SGLT2, thiazolidinediones, α-glucosidase.

We used propensity score matching to adjust for baseline characteristics and Cox proportional hazards models to calculate HRs for the outcomes: dementia, Parkinson's disease, ICI, ischemic stroke, intracerebral hemorrhage, and total mortality.

What has changed - by the numbers

• Dementia: HR 0.63 (−37% relative risk).
• Ischemic stroke: HR 0.81 (−19%).
• Death from any cause: HR 0.70 (−30%).
• Subgroups: greatest benefit in ≥60 years, women, BMI 30-40.

What was not found

• Parkinson's disease - no significant differences.
• Hemorrhagic stroke - also without differences.
  These nuances are also emphasized by the JAMA editorial digest.

How it might work (mechanical clues)

• Anti-inflammatory effect in the central nervous system and blood vessels, reduction of neuroinflammation.
• Improved endothelial function and metabolic control (glucose, weight, blood pressure) → less vascular risk.
• Possible direct effects of GLP-1 signaling in the brain.
  Hypotheses are consistent with the findings of the press release and accompanying materials.

What does this mean for the patient and the doctor?

• In people with type 2 diabetes + obesity, semaglutide/tirzepatide therapy may not only help with sugar and weight, but also be associated with better neurovascular outcomes.
• This is not proof of causality: unobserved factors (lifestyle, access to care, comorbidities) may have influenced treatment choices. Treatment decisions remain individual and based on clinical indications.

Limitations to keep in mind

• Observational design → residual confounding is inevitable, even with matching.
• Based on EHR codes: possible event classification errors.
• There is no dose/duration randomization and no strict control of concomitant medications.
• RCTs and prospective confirmation of neurovascular benefit are needed.

Summary

In obese T2D patients, semaglutide/tirzepatide is associated with a lower risk of dementia, ischemic stroke, and death at up to 7 years. This is an encouraging but observational signal: it reinforces the idea that GLP-1RAs may provide neurovascular benefits beyond glycemic control - and sets the agenda for future randomized trials.

Source: Lin HT et al. Neurodegeneration and Stroke After Semaglutide and Tirzepatide in Patients With Diabetes and Obesity. JAMA Network Open 2025;8(7):e2521016. DOI: 10.1001/jamanetworkopen.2025.21016. Press release published July 15, 2025.