Tirzepatide Linked to Lower Dementia Risk Than Semaglutide or SGLT2 Inhibitors in Diabetes Patients

What the researchers found

Three target trial emulations using propensity score-matched real-world data compared dementia outcomes across diabetes medications over two years:

- Tirzepatide vs. SGLT2 inhibitors (n=14,462): HR 0.66 (95% CI 0.47–0.93, p=0.02) for dementia

- Semaglutide vs. SGLT2 inhibitors (n=57,959): results not specified for dementia HR

- Tirzepatide vs. semaglutide (n=12,246): HR 0.69 (95% CI 0.48–0.99, p=0.04) for dementia

Tirzepatide also showed lower all-cause mortality vs. both semaglutide (HR 0.72, 95% CI 0.58–0.90) and SGLT2 inhibitors (HR 0.29, 95% CI 0.23–0.37). Both tirzepatide and semaglutide reduced major adverse cardiovascular events (MACE) compared to SGLT2 inhibitors.

Why it matters

Dementia is a growing global health crisis with few effective treatments. If certain diabetes medications can reduce dementia risk as a side benefit, this could influence prescribing decisions for the millions of people with type 2 diabetes who are already at elevated dementia risk.

How the study worked

Researchers conducted three target trial emulations using the TriNetX global federated research network, a large real-world database. Adults with type 2 diabetes and no baseline dementia were included. Propensity score matching was used to balance groups, and survival analysis tracked first dementia diagnosis, MACE, and all-cause mortality over two years.

What this study cannot tell us

This is an observational study using real-world data, not a randomized controlled trial, so it cannot prove causation. Unmeasured confounders may exist despite propensity score matching. The two-year follow-up is relatively short for dementia outcomes. Tirzepatide has been available for less time than the comparators, so follow-up duration may differ. The authors explicitly note these findings are hypothesis-generating.

How to read the evidence

This is a large observational study using target trial emulation methodology with propensity score matching — a rigorous approach for real-world evidence. However, it remains observational and cannot establish causation. The authors explicitly describe the findings as hypothesis-generating.

When this study was published

Published in 2026 in Diabetes Research and Clinical Practice, this is a very recent study using up-to-date real-world data including tirzepatide, which only became widely available in recent years.

The bigger picture

There's intense interest in whether GLP-1 receptor agonists can protect against neurodegeneration. This study suggests that tirzepatide — which targets both GLP-1 and GIP receptors — may offer even greater neuroprotective benefits than semaglutide alone, potentially due to its dual-receptor mechanism. This adds to a rapidly growing body of evidence linking these diabetes drugs to brain health.

Questions still open

• Does tirzepatide's dual GLP-1/GIP receptor mechanism explain its greater dementia risk reduction compared to semaglutide's GLP-1-only approach?
• Would a randomized controlled trial confirm these observational findings, and what duration would be needed?
• Is the dramatically lower mortality with tirzepatide vs. SGLT2 inhibitors (HR 0.29) influenced by prescribing bias or channeling effects?

Common questions

Read the original research

Target trial emulations for tirzepatide, semaglutide and SGLT2-inhibitors for dementia in patients with type 2 diabetes: Real world evidence from a retrospective cohort study.

Diabetes research and clinical practice, 113083

Citation

Younis, Alhena; Henney, Alex E; Riley, David R; Anson, Matthew; Zhao, Sizheng S; Ibarburu, Gema H; Malik, Rayaz A; Su, Li; Lip, Gregory Y H; Cuthbertson, Daniel J; Alam, Uazman. (2026). Target trial emulations for tirzepatide, semaglutide and SGLT2-inhibitors for dementia in patients with type 2 diabetes: Real world evidence from a retrospective cohort study.. Diabetes research and clinical practice, 113083. https://doi.org/10.1016/j.diabres.2026.113083