Causal associations of sleep apnea and snoring with type 2 diabetes and glycemic traits and the role of BMI

Abstract Objective Sleep apnea and snoring have been associated with type 2 diabetes, with BMI playing a role in the pathway, but the directions of causality are unclear. This study examined the causal associations of sleep apnea and snoring with type 2 diabetes while assessing the role of BMI using multiple genetic methods. Methods Five genetic methods were used: two‐sample; bidirectional univariable Mendelian randomization (MR) inverse variance‐weighted (MR‐IVW); multivariable MR‐IVW; network MR; and latent causal variable method. Results Compared with univariable MR‐IVW, the odds ratio (95% CI) of type 2 diabetes for genetically predicted sleep apnea and snoring using the largest genome‐wide association study decreased dramatically, from 1.61 (95% CI: 1.16‐2.23) to 1.08 (95% CI: 0.59‐1.97) and from 1.98 (95% CI: 1.25‐3.13) to 1.09 (95% CI: 0.64‐1.86) after adjustment for BMI. Network MR showed that BMI accounts for 67% and 62% of the total effect of sleep apnea and snoring on type 2 diabetes, respectively. The latent causal variable suggested that sleep apnea and snoring have no direct causal effect on type 2 diabetes. Conclusions These results first suggest that the associations of sleep apnea and snoring with type 2 diabetes were mainly driven by BMI. The possible indirect effects of sleep apnea and snoring on type 2 diabetes through BMI cannot be ruled out.