Causal Association of Sleep Traits with All-Cause and Cause-Specific Mortality: A Prospective Cohort and Mendelian Randomization Study

Abstract

The study aimed to explore the association between different sleep traits and all-cause mortality as well as to validate causality in the association through mendelian randomization (MR). We analyzed 451,420 European ancestry participants from the UK Biobank. Multivariable-adjusted Cox proportional hazards model was conducted to evaluate the association between sleep traits and all-cause mortality. In MR analysis, the inverse variance weighting (IVW) method was applied as the primary analysis to investigate the causal association between sleep traits and mortality. During a median follow-up period of 12.68 years, 34,397 individuals died. Observational analyses showed the multivariate-adjusted hazard ratio (HR) and 95% confidence intervals (CIs) for short sleep, long sleep, early chronotype, daytime sleepiness, daytime napping, and insomnia with mortality, 1.246 (1.195, 1.298), 1.735 (1.643, 1.831), 0.931 (0.909, 0.953), 1.276 (1.212, 1.344), 1.299 (1.254, 1.346), and 1.117 (1.091, 1.142) (All p < 0.0001). Based on UK Biobank, MR analysis indicated the association between daytime napping and an increased risk of all-cause mortality (odd ratio [OR]: 1.219, 95% CI: 1.071–1.387, p = 0.003), which may be largely attributable to cancer disease mortality (OR: 1.188, 95% CI: 1.009–1.399, p = 0.039). We found no causal association between sleep duration, short sleep, long sleep, chronotype, daytime sleepiness, insomnia, and mortality risk. The causal associations between sleep traits and all-cause mortality risk were directionally replicated in FinnGen. Our findings suggest a potential causal association between daytime napping and increased risk of all-cause mortality in middle-aged and older persons. The finding could have important implications for evaluating daytime napping habits to decrease the risk of mortality.