Causal Impact of Type 2 Diabetes Mellitus on Cerebral Small Vessel Disease: A Mendelian Randomization Analysis.

Junfeng Liu,Loes Rutten-Jacobs,Hugh S Markus,Matthew Traylor,Ming Liu

doi:10.1161/strokeaha.117.020536

Abstract

The relationship between type 2 diabetes mellitus (T2D) and cerebral small vessel disease (CSVD) is unclear. We aimed to examine the causal effect of T2D, fasting glucose levels, and higher insulin resistance on CSVD using Mendelian randomization. Five CSVD phenotypes were studied; 2 were clinical outcomes associated with CSVD (lacunar stroke: n=2191/27 297 and intracerebral hemorrhage [ICH]: n=2254/8195 [deep and lobar ICH]), whereas 3 were radiological markers of CSVD (white matter hyperintensities: n=8429; fractional anisotropy [FA]: n=8357; and mean diffusivity: n=8357). We applied 2 complementary analyses to evaluate the association of T2D with CSVD. First, we used summarized data from genome-wide association study to calculate the effects of T2D-related variants on CSVD with inverse-variance weighted and weighted median approaches. Second, we performed a genetic risk score approach to test the effects of T2D-associated variants on white matter hyperintensities, FA, and mean diffusivity using individual-level data in UK Biobank. T2D was associated with higher risk of lacunar stroke (odds ratio [OR], 1.15; 95% confidence interval [CI], 1.04-1.28; P=0.007) and lower mean FA (OR, 0.78; 95% CI, 0.66-0.92; P=0.004) but not white matter hyperintensities volume (OR, 1.01; 95% CI, 0.97-1.04; P=0.626), higher mean diffusivity (OR, 1.04; 95% CI, 0.89-1.23; P=0.612), ICH (OR, 1.07; 95% CI, 0.95-1.20; P=0.269), lobar ICH (OR, 1.07; 95% CI, 0.89-1.28; P=0.466), or deep ICH (OR, 1.16; 95% CI, 0.99-1.36; P=0.074). Weighted median and penalized median weighted analysis showed similar effect estimates of T2D on lacunar stroke and FA, but with wider CIs, meaning they were not significant. The genetic score on individual-level data was significantly associated with FA (OR, 0.63; 95% CI, 0.45-0.89; P=0.008) after adjusting for potential confounders. Our Mendelian randomization study provides evidence to suggest that T2D may be causally associated with CSVD, in particular with lacunar stroke and FA.

Highlights

Background and PurposeThe relationship between type 2 diabetes mellitus (T2D) and cerebral small vessel disease (CSVD) is unclear
Five CSVD phenotypes were studied; 2 were clinical outcomes associated with CSVD, whereas 3 were radiological markers of CSVD
T2D was associated with higher risk of lacunar stroke and lower mean fractional anisotropy (FA) (OR, 0.78; 95% confidence interval (CI), 0.66–0.92; P=0.004) but not white matter hyperintensities volume (OR, 1.01; 95% CI, 0.97–1.04; P=0.626), higher mean diffusivity (OR, 1.04; 95% CI, 0.89–1.23; P=0.612), Intracerebral hemorrhages (ICH) (OR, 1.07; 95% CI, 0.95–1.20; P=0.269), lobar ICH (OR, 1.07; 95% CI, 0.89–1.28; P=0.466), or deep ICH (OR, 1.16; 95% CI, 0.99–1.36; P=0.074)