Introduction

A number of studies have shown a U-shaped association between sleep duration and mortality. Since sleep duration is partly genetically determined, it seems likely that its association with mortality is also genetically influenced. The purpose of the present study was to investigate the influence on heredity on the association between sleep duration and mortality.

Methods

We used a cohort of 14 267 twins from the Swedish Twin Registry.

Results

A Cox proportional hazards regression analysis, adjusted for a number of covariates, confirmed a clear U shape with a hazard ratio (HR) = 1.34 and 95% confidence interval (CI) = 1.15-1.57 for a sleep duration of = 6.5 hours and HR = 1.18 (CI = 1.07-1.30) for sleep of = 9.5 hours. Reference value was 7.0 hours. A co-twin analysis of 1942 twins discordant on mortality showed a HR = 2.66 (CI = 1.17-6.04) for long (= 9.5 hours) sleep in monzygotic twins and an HR = 0.66 (CI = 0.20-2.14) for short (< 6.5 hours) sleep. In dizygotic twins, no association was significant. The heritability for mortality was 28% for the whole group, while it was 86% for short sleepers and 42% for long sleepers. Thus, the link with mortality for long sleep appears to be more due to environmental factors than to heredity, while heritability dominates among short sleepers.

Conclusions

We found that both long and short sleep were associated with higher total mortality, that the difference in mortality within twin pairs is associated with long sleep, and that short sleep has a higher heritability for mortality, while long sleep is associated with more environmental influences on mortality.