Abstract
Objective To evaluate the causal effects of sleep traits (i.e., chronotype, insomnia, and sleep duration) on bioavailable testosterone (BT), sex hormone-binding globulin (SHBG), and total testosterone (TT) levels in women and men. Methods We performed Mendelian randomization (MR) using random-effect inverse-variance weighted (IVW) and 7 other MR analyses. Exposure data for sleep traits were obtained from the largest-to-date genome-wide association study (GWAS) from 339,926 to 1,331,010 individuals. Summary data for testosterone levels were obtained from GWAS based on the UK Biobank. Results For women, our study supported that chronotype was associated with decreased BT (IVW: β = − 0.042, 95% CI − 0.060, − 0.023, p = 1.17E–05) and TT (IVW: − 0.053, 95% CI − 0.075, − 0.031, p = 2.30E–06). Besides, insomnia can significantly increase BT (IVW: β = 0.025, 95% CI 0.009, 0.041, p = 0.002). These findings were significant in most sensitivity analyses. For men, statistical significance was found between chronotype and BT (β = − 0.027, 95% CI − 0.048, − 0.005, p = 0.016), and insomnia and TT (β = − 0.028, 95% CI − 0.049, 0.007, p = 0.009) in IVW. However, the effect estimates were not broadly consistent with other sensitivity analyses. Our study did not find support for causal effects of sleep duration on testosterone levels in both women and men. Conclusion Our study reveals the sex differences in the effects of sleep traits on testosterone levels. A healthy sleep habit is vital for the maintenance of testosterone homeostasis in women. Further studies are warranted to investigate the associations between sleep traits and testosterone levels in men.
Funding Information
  • West China Hospital, Sichuan University (2020HXBH016, 2019HXBH087)

This publication has 42 references indexed in Scilit: