Abstract
Sleep duration and sleep efficiency are observed risk factors for all-cause mortality, but the role of sleep regularity is unknown. Irregular sleep patterns are associated with ageing, cancer, depressed mood, and cardio-metabolic diseases, and may therefore increase risk of mortality. We extracted Sleep Regularity Index (SRI) scores, sleep duration, and sleep efficiency from one week of actigraphy in 60,997 UK Biobank participants (62.8±7.8 years, 55.6% female). Using Cox Proportional Hazards models, we assessed whether SRI, sleep duration, and sleep efficiency predicted all-cause mortality up to 7.6 years post-actigraphy. A one standard deviation increase in SRI score (+10.7), indicating more regular sleep patterns, predicted a 20.4±2.0% reduced hazard of all-cause mortality (p < .0001). This relationship remained robust after controlling for age, sex, ethnicity, education, income, and employment status (+1SD in SRI predicted an 18.7±2.4% hazard reduction, p < .0001). SRI was a relatively stronger predictor of mortality hazard than sleep duration (explaining 19.0±2.1% vs. 12.4±2.6% of hazard reduction) or sleep efficiency (19.0±2.1% vs. 13.1±2.6%) in two additional Cox models. Sleep regularity may play a causal role in mortality, via contribution to poor health outcomes, or it may be a marker of poor health leading to mortality. These findings suggest that sleep regularity should be monitored and/or adjusted alongside traditional sleep metrics for optimal health and longevity.