Abstract Body (Do not enter title and authors here): INTRODUCTION
Cerebrovascular disease, including stroke and cerebral small vessel disease, remains second leading cause of death and disability worldwide. Sodium-glucose cotransporter 2 (SGLT2) inhibitors have demonstrated improved cardiovascular and renal outcomes. However, their potential neuroprotective effects remain underexplored. Though preliminary observational studies suggest protective association between SGLT2 inhibition and cerebrovascular outcomes, residual confounding and reverse causation may influence these findings. We aimed to assess causal role of SGLT2 inhibition in cerebrovascular disease through Systematic review and Meta-analysis of Mendelian randomization (MR) studies.

METHODS
We reviewed two-sample MR studies evaluating causal association between genetically predicted SGLT2 inhibition and risk of stroke, including ischemic and any stroke subtypes, through May 2025. Databases included PubMed and Google Scholar. Eligible studies utilized genetic instruments derived from genome-wide association studies (GWAS) of SGLT2-related traits, including atrial fibrillation, blood lipids, circulating plasma proteins, and urinary sodium. Binary random-effects model was used to calculate pooled ORs, I² statistic to assess heterogenicity, leave-one-out sensitivity analysis, and subgroup analysis (ischemic stroke vs. any stroke) to evaluate robustness and consistency.

RESULTS
A total of five MR studies from multiple GWAS datasets were included. Pooled OR [inverse variance weighted] for the association between SGLT2 inhibition and risk of stroke was 0.30 (95% CI: 0.15-0.60; p<0.01), indicating statistically significant protective effect. Substantial heterogeneity was observed (I²=93.5%). However, leave-one-out sensitivity analysis demonstrated that no individual study unduly influenced the overall estimate. Subgroup analysis revealed stronger effect for ischemic stroke (pooled OR=0.29; 95% CI: 0.21-0.39, p<0.01; I²=0%) whereas the association with any stroke was not significant (pooled OR=0.54; 95% CI: 0.16-1.85, p=0.32; I²=97.31%).

CONCLUSION:
Our study provides strong evidence supporting causal, protective effect of SGLT2 inhibition against stroke, particularly ischemic stroke, based on MR. The consistent findings across diverse GWAS exposures and robust sensitivity analyses underscore therapeutic potential of SGLT2 inhibition in stroke prevention. However, high heterogeneity in the overall model and broader stroke subgroup warrants further studies.