Background and Purpose FXI (factor XI) is involved in thrombus propagation and stabilization. It is unknown whether lower FXI levels have a protective effect on risk of ischemic stroke (IS) or myocardial infarction. This study investigated the effect of genetically determined FXI levels on risk of IS, myocardial infarction, and intracerebral hemorrhage. Methods Two-sample Mendelian randomization analysis was performed. Instruments and genetic association estimates for FXI levels were obtained from a genome-wide association study of 16 169 individuals. Genetic association estimates for IS and its etiological subtypes were obtained from a study of 16 851 cases and 32 473 controls. For myocardial infarction, estimates were obtained from a study of 43 676 cases and 123 504 controls and for intracerebral hemorrhage from a study of 1545 cases and 1481 controls. Results After applying a Bonferroni correction for multiple testing, the Mendelian randomization analysis supported a causal effect of higher, genetically determined FXI levels on risk of any IS (odds ratio [OR] per 1-unit increase in natural logarithm-transformed FXI levels, 2.54; 95% CI, 1.68-3.84; P=1x10(-5)) but not myocardial infarction (OR, 1.01; 95% CI, 0.76-1.34; P=0.94) or intracerebral hemorrhage (OR, 1.81; 95% CI, 0.44-7.38; P=0.41). Examining IS subtypes, the main results supported an effect of higher, genetically determined FXI levels on risk of cardioembolism (OR, 4.23; 95% CI, 1.94-9.19; P=3x10(-4)) and IS of undetermined cause (OR, 3.44; 95% CI, 1.79-6.60; P=2x10(-4)) but not large artery atherosclerosis (OR, 2.73; 95% CI, 1.15-6.45; P=0.02) or small artery occlusion (OR, 1.19; 95% CI, 0.50-2.82; P=0.69). However, the statistically significant result for IS of undetermined cause was not replicated in all sensitivity analyses. Conclusions We find Mendelian randomization evidence supporting FXI as a possible target to reduce risk of the cardioembolic subtype of IS.