Association Between Metformin and Mortality in T2DM Patients With Different Subtypes of Ischemic Stroke in MIMIC‐IV: A Retrospective Cohort Study

Background and Purpose While Type 2 diabetes mellitus (T2DM) is a recognized risk factor for ischemic stroke (IS), metformin’s effect on mortality across distinct IS subtypes remains unclear. This study aimed to investigate the association between metformin and all‐cause mortality in patients with concurrent T2DM and IS, exploring TOAST subtype‐specific efficacy. Methods This retrospective MIMIC‐IV cohort study included T2DM and IS patients receiving glucose‐lowering agents. Primary and secondary endpoints were 90‐ and 365‐day all‐cause mortality. Propensity score matching (PSM) and multivariate logistic regression evaluated associations. Results Among 1946 patients, 375 received metformin. Post‐PSM (nonmetformin n = 750; metformin n = 375), 90‐day mortality was significantly lower in the metformin cohort (9.6% vs. 15.1%, p = 0.014). Multivariate regression confirmed metformin’s independent association with decreased 90‐day (odds ratio [OR] = 0.63, 95% confidence interval [CI]: 0.42–0.94, p = 0.025) and 365‐day (OR = 0.59, 95% CI: 0.42–0.83, p = 0.002) mortality. Subgroup analyses exploring TOAST classification heterogeneity revealed that in the unmatched cohort, metformin was associated with reduced 90‐day mortality in cardioembolism (CE) and stroke of other determined etiology (SOE), as well as reduced 365‐day mortality in large‐artery atherosclerosis (LAA), CE, SOE, and stroke of undetermined etiology (SUE). However, following PSM adjustment, significant 365‐day mortality reduction persisted solely in the SUE subtype (OR = 0.38, 95% CI: 0.16–0.90, p = 0.027). Conclusion Metformin is significantly associated with reduced short‐ and long‐term mortality in patients with T2DM and IS. Prognostic benefits exhibit heterogeneity across stroke etiologies, highlighting the necessity for tailored glucose‐lowering strategies based on TOAST classifications.