Mendelian randomization did not support the causal effect of diabetes on aortic diseases.

ABSTRACT

Observational studies revealed paradoxically inverse associations between diabetes and aortic diseases (aortic aneurysm or aortic dissection), yet the causality remains to be determined. To investigate the causal associations between diabetes and aortic diseases using Mendelian randomization (MR) analyses. Summary-level data for exposures (type 1 diabetes, type 2 diabetes, fasting glucose, fasting insulin, glycated hemoglobin) and outcomes (aortic dissection and aortic aneurysm) were obtained from public genome-wide association study data. The principal analysis was the inverse-variance weighted (IVW) method. Sensitivity analyses were also carried out, including weighted median, MR-Egger, and multivariable MR methods. According to IVW results, type 1 diabetes (odds ratio [OR] 0.99; 95% confidence interval [CI] 0.93-1.07; P = 0.87), type 2 diabetes (OR 0.97; 95% CI 0.77-1.20; P = 0.75), fasting glucose (OR 1.16; 95% CI 0.48-2.84; P = 0.74), fasting insulin (OR 2.75; 95% CI 0.53-14.26; P = 0.23), or glycated hemoglobin (OR 0.33; 95% CI 0.09-1.17; P = 0.09) had no causal effect on aortic dissection. Similarly, type 1 diabetes, type 2 diabetes, fasting glucose, fasting insulin, or glycated hemoglobin had no causal effect on aortic aneurysm. Sensitivity analyses revealed consistent results. MR-Egger method and funnel plot yielded no indication of directional pleiotropy. Diabetes had no causal associations with aortic dissection or aortic aneurysm. The observed inverse associations in previous cohort studies may be explained by confounding factors or reverse causation.