Sex differences in sleep regularity and its associations with mortality

Abstract

In characterizing sleep and circadian health, the day-to-day regularity of sleep-wake timing strongly predicts health outcomes, outperforming short sleep duration in prospective associations with mortality and new-onset disease. It is unknown whether biological (e.g., sleep and circadian physiology) and sociocultural (e.g., exposures that affect sleep-wake timing) sex differences lead to differences in day-to-day sleep-wake regularity or modify its prospective associations with health outcomes. Here, we present findings from a UK Biobank study of 506,582 person-days of accelerometer recordings across 73,647 middle-aged adults preceding 549,009 person-years of follow-up. We compared SRI scores between males and females and evaluated whether all-cause, cardiovascular, and cancer mortality differed across SRI groups by sex. Custom contrasts were used to compare estimated marginal means across specific SRI-sex combinations. Females were overrepresented among very high (SRI [≥]90) and underrepresented among very low (SRI <60) groups. After adjustment for demographic, health, and behavioral covariates, males still had higher odds than females of exhibiting SRI <60. Low SRI and male sex were synergistically associated with higher mortality rate. Demographic, health, and behavioral covariate-adjusted models showed stronger and more dose-dependent associations between low SRI and mortality among males than females. Although the omnibus SRI x sex interaction terms were not statistically significant, the interaction contrast for SRI <60 versus [≥]90 differed by sex, suggesting a possible sex difference in mortality rates at very low SRI. Together, our findings suggest that males may be more vulnerable than females to the mortality risk associated with highly irregular sleep-wake schedules.