Antidiabetic medications and risk of cognitive disorders in type 2 diabetes: A retrospective cohort study

To evaluate the associations between commonly used antidiabetic regimens and incidence of mild cognitive disorder, Alzheimer disease, and vascular dementia in adults with type 2 diabetes.

This retrospective cohort study used the TriNetX US Collaborative Network of electronic health records from 2010 to 2024. Adults aged 40-69 years with type 2 diabetes and at least 1 year of follow-up were included. Propensity score matching was applied to balance covariates.

Patients were classified into 5 treatment groups: metformin only (reference), metformin plus DPP-4 inhibitors, metformin plus GLP-1 receptor agonists, metformin plus SGLT-2 inhibitors, and insulin monotherapy. Exposure was defined by first recorded prescription and continued use during follow-up.

Primary outcomes were incident mild cognitive disorder, Alzheimer disease, and vascular dementia, identified using ICD-10 codes. Hazard ratios with 95% confidence intervals were estimated from Cox proportional hazards models, with landmark analyses for follow-up shorter and longer than 5 years.

Among 1,528,885 adults with type 2 diabetes (mean age, 58 years; 49.2% women), GLP-1 receptor agonists plus metformin were associated with lower incidence of vascular dementia (HR, 0.46; 95% CI, 0.37-0.57), mild cognitive disorder (HR, 0.79; 95% CI, 0.66-0.95), and Alzheimer disease (HR, 0.46; 95% CI, 0.31-0.70). SGLT-2 inhibitors plus metformin reduced vascular dementia risk (HR, 0.68; 95% CI, 0.54-0.87) but not other outcomes. Insulin monotherapy was associated with higher incidence of vascular dementia (HR, 3.27; 95% CI, 3.03-3.52), mild cognitive disorder (HR, 1.71; 95% CI, 1.56-1.87), and Alzheimer disease (HR, 1.56; 95% CI, 1.32-1.84).

GLP-1 receptor agonists and SGLT-2 inhibitors combined with metformin were associated with reduced risk of cognitive decline. Insulin monotherapy was associated with higher incidence of all three outcomes; however, this association is likely influenced by confounding by indication, unmeasured markers of diabetes severity (including diabetes duration, cardiovascular and renal disease severity, and microvascular and macrovascular complications), and shorter follow-up among insulin users, and should not be interpreted as a direct drug effect. Because standard Cox models do not account for death as a competing event, reported hazard ratios reflect cause-specific hazards and may not directly correspond to cumulative incidence, particularly for the insulin group, in which mortality and censoring were substantially higher. Antidiabetic medication choice may influence long-term cognitive outcomes and should be considered in diabetes management.