Objective: To investigate the associations between sleep duration and mortality in the elderly by controlling for sleep quality.
Method: Data were collected from participants in a cohort study in Shizuoka, Japan. A total of 14,001 elderly residents (aged 65-85 years), randomly chosen from all 74 municipalities in the prefecture, completed questionnaires that evaluated sleep duration, sleep complaints, and the use of hypnotics. Participants were followed from 1999 to 2006. We analyzed 11,395 subjects to estimate the hazard ratios (HR) for mortality from all causes and cardiovascular disease (CVD).
Results: With 60,252 person-years, 1004 deaths were identified. While short sleep duration and mortality were not associated, longer sleep duration was associated with higher risk of mortality in both sexes. Compared with those who slept 7 h, the multivariate HR and 95% confidence interval of CVD mortality for those who slept > or =10 h was 1.95 (1.18-3.21) and, for those who slept < or =5 h, it was 1.10 (0.62-1.93). Although no clear association was found between sleep quality and mortality, long sleep duration was associated with higher risk of CVD mortality among those with poor sleep quality.
Conclusion: Long sleep duration is associated with higher risk of CVD mortality among the elderly with poor sleep quality.