Additional Evidence for the Relationship between Type 2 Diabetes Mellitus and Stroke through Observational and Genetic Analyses
While type 2 diabetes mellitus (T2DM) is commonly considered a putative causal risk factor for stroke, the effect of stroke on T2DM remains unclear. The intrinsic link underlying T2DM and stroke has neither been thoroughly examined. We aimed to evaluate the phenotypic and genetic relationships underlying T2DM and stroke. We evaluated phenotypic associations using data from the UK Biobank (N=472,050). We then investigated genetic relationships leveraging genomic data in European ancestry for T2DM, with and without adjusting for body mass index (BMI) (T2DM: Ncase/Ncontrol=74,124/824,006; T2DMadjBMI: Ncase/Ncontrol=50,409/523,897), and for stroke (Ncase/Ncontrol=73,652/1,234,808). We performed additional analyses using genomic data in East Asian ancestry for T2DM (Ncase/Ncontrol=77,418/356,122) and for stroke (Ncase/Ncontrol=27,413/237,242). Observational analyses suggested a significantly increased hazard of stroke among individuals with T2DM (hazard ratio 2.28 [95%CI 1.97, 2.64]), but a slightly increased hazard of T2DM among individuals with stroke (1.22 [1.03, 1.45]) which attenuated to 1.14 [0.96, 1.36] in sensitivity analysis. A positive global T2DM-stroke genetic correlation was observed (rg=0.35; P=1.46×10-27), largely independent of BMI (T2DMadjBMI-stroke: rg=0.27; P=3.59×10-13). This was further corroborated by 38 shared independent loci and 161 shared expression-trait associations. Mendelian randomization analyses suggested a putative causal effect of T2DM on stroke in Europeans (odds ratio 1.07 [95%CI 1.06, 1.09]), which remained significant in East Asians (1.03 [1.01, 1.06]). Conversely, despite a putative causal effect of stroke on T2DM also observed in Europeans (1.21 [1.07, 1.37]), it attenuated to 1.04 [0.91, 1.19] in East Asians. Our study provides additional evidence to underscore the significant relationship between T2DM and stroke.