Sleep Duration, Nap Habits, and Mortality in Older Persons

Objective: To examine the effect of nighttime sleep duration on mortality and the effect modification of daytime napping on the relationship between nighttime sleep duration and mortality in older persons. Design: Prospective survey with 20-yr mortality follow-up. Setting: The Cross-Sectional and Longitudinal Aging Study, a multidimensional assessment of a stratified random sample of the older Jewish population in Israel conducted between 1989-1992. Participants: There were 1,166 self-respondent, community-dwelling participants age 75-94 yr (mean, 83.40, standard deviation, 5.30). Measurements: Nighttime sleep duration, napping, functioning (activities of daily living, instrumental activities of daily living, Orientation Memory Concentration Test), health, and mortality. Results: Duration of nighttime sleep of more than 9 hr was significantly related to increased mortality in comparison with sleeping 7-9 hr (hazard ratio [HR] = 1.31, P < 0.01) after adjusting for demographic, health, and function variables, whereas for short nighttime sleep of fewer than 7 hr mortality did not differ from that of 7-9 hr of sleep. For those who nap, sleeping more than 9 hr per night significantly increased mortality risk (HR = 1.385, P < 0.05) and shorter nighttime sleep reduced mortality significantly in the unadjusted model (HR = 0.71, P < 0.001) but only approached significance in the fully adjusted model (HR = 0.82, P = 0.054). For those who do not or sometimes nap, a short amount of sleep appears to be harmful up to age 84 yr and may be protective thereafter (HR = 1.51, confidence interval [CI] = 1.13-2.02, P < 0.01; HR = 0.76, CI = 0.49-1.17, in the fully adjusted model, respectively). Conclusions: The findings are novel in demonstrating the protective effect of short nighttime sleep duration in individuals who take daily naps and suggest that the examination of the effect of sleep needs to take into account sleep duration per 24 hr, rather than daytime napping or nighttime sleep per se