GLP-1 RAs vs Metformin Significantly Reduce Risk of Dementia in Adults with Type 2 Diabetes

A large population-based cohort study has found that treatment with GLP-1 receptor agonists (GLP-1 RA) was associated with a significantly lower risk of incident dementia, particularly Alzheimer disease and nonvascular dementia, compared to treatment with metformin when used as a first-line therapy in adults with type 2 diabetes (T2D).1

The findings, published in JAMA Neurology, suggest that initiating treatment with a GLP-1 RA may offer more robust neuroprotective advantages in addition to glycemic control. Among 558,170 matched individuals, GLP-1 RA users had a 19% lower risk of all-cause dementia over a median 6.8 years of follow-up (hazard ratio [HR] 0.81; 95% CI, 0.75–0.88) compared to metformin initiators.1

Individuals with T2D have a 1.7-fold higher risk of developing dementia compared to the general public,2 first author Mingyang Sun, from the department of anesthesiology and perioperative medicine at the People’s Hospital of Zhengzhou University, Henan, China, and colleagues wrote. Moreover, both metformin and GLP-1 RAs have been explored for their potential to reduce the risk of dementia, with the former linked to reductions in oxidative stress and improved insulin sensitivity and the latter showing strong neuroprotective effects and the ability to reduce amyloid plaques in the brain.3,4 To date, however, there have been no direct comparative studies evaluating the differential efficacy of the 2 drugs in preventing T2D-related dementia.

With this research gap as motivation, Sun and fellow researchers conducted a retrospective cohort study using data from Danish national health registers spanning 2008 to 2021. Eligible individuals were adults without prior dementia who initiated either a GLP-1 RA or metformin as first-line treatment and had no history of insulin or other glucose-lowering drug use. Propensity score matching was applied to balance demographic and clinical variables between the treatment groups, yielding 279,085 matched pairs. The mean age was 56.9 years, and 52.5% were women. Comorbidity profiles also were similar, eg, authors reported similar rates of coronary artery disease and cerebrovascular disease.1

Over the follow-up period, Sun et al identified 1,215 incident dementia cases among GLP-1 RA users and 1,520 among metformin users, corresponding to incidence rates of 0.79 and 1.01 per 1,000 person-years, respectively, with the HR for all cause dementia in favor of GLP-1 RAs. Reductions were also seen in cause-specific analyses:

• Alzheimer disease (HR 0.82; 95% CI, 0.72–0.94)
• Vascular dementia (HR 0.84; 95% CI, 0.74–0.96)
• Other dementia types (HR 0.75; 95% CI, 0.64–0.89)

In contrast, the investigators found that the risk of Parkinson disease, included as a negative control outcome, did not differ significantly between groups (HR 0.91; 95% CI, 0.76–1.10), a reinforcement, they stated, of the specificity of the dementia findings.1

The authors also conducted a series of sensitivity analyses, including adjusting for the year of treatment initiation and using time-conditional propensity scores, which did not materially change the results. In further subgroup analyses the researchers found consistent results across age, sex, and baseline comorbidity levels. Notably, the relative risk reduction in dementia appeared within the first few years of treatment and remained stable over time.1

Among the study's limitations Sun and colleagues noted the observational design and potential for residual confounding, despite rigorous matching and adjustment. They also highlighted that they could not directly assess medication adherence and that dementia diagnoses were based on registry coding rather than clinical adjudication. Furthermore, because GLP-1 RAs were more commonly used in later years of the study period, temporal trends in dementia diagnosis or treatment practices may have influenced results.

The team called for randomized clinical trials to confirm their findings and to clarify mechanisms, including whether GLP-1 RAs exert direct neuroprotective effects beyond glucose lowering.

"Given the significant clinical and societal burden of T2DM-related dementia, determining which therapy— GLP-1 RAs or metformin—provides superior protection against cognitive decline is crucial," they wrote in the study's discussion. They suggested that despite metformin's well-established status as first-line T2D therapy, superior efficacy of GLP-1 RAs in dementia protection, if substantiated, "could redefine treatment guidelines for T2DM.

"Clarifying this therapeutic distinction will offer crucial guidance for clinicians, policymakers, and researchers, potentially reshaping dementia prevention strategies in T2DM management," they concluded.

1. Sun M, Wang X, Lu Z, et al. Evaluating GLP-1 receptor agonists versus metformin as first-line therapy for reducing dementia risk in type 2 diabetes. BMJ Open Diab Res Care. 2025;13:e004902. doi:10.1136/ bmjdrc-2025-004902 https://drc.bmj.com/content/bmjdrc/13/4/e004902.full.pdf
2. Tang B, Sjölander A, Wastesson JW, et al. Comparative effectiveness of glucagon-like peptide-1 agonists, dipeptidyl peptidase-4 inhibitors, and sulfonylureas on the risk of dementia in older individuals with type 2 diabetes in Sweden: an emulated trial study. EClinicalMedicine 2024;73:102689.
3. Tang H, Shao H, Shaaban CE, et al. Newer glucose-lowering drugs and risk of dementia: A systematic review and meta-analysis of observational studies. J Am Geriatr Soc 2023;71:2096–106.
4. Liang Y, Doré V, Rowe CC, et al. Clinical Evidence for GLP-1 Receptor Agonists in Alzheimer’s Disease: A Systematic Review. J Alzheimers Dis Rep 2024;8:777–89.