Sleep duration and mortality - Influence of age and occupational group in retired individuals.

The importance of sleep duration for health or mortality attracts much public attention. Prior work indicates that both long and short sleep duration predicts mortality, with optimal sleep duration (lowest risk) at 7 h. However, we believe this may differ between subgroups. This may be the case with, for example, age groups (due to aging), or blue-collar and white-collar worker (due to work exposure). It is also likely that retirement, which permits extension of the time in bed, may confound analyses. The purpose of the present study was to investigate how occupational group (blue-collar/white-collar worker) and age influence the pattern of association between sleep duration and mortality in retired individuals. Retired individuals were selected since it was hypothesized that effects of occupation may accumulate over years and since the transition into retirement may be a confounder. We used a sample of 14 000 individuals from the Swedish Twin Registry, which had provided data on sleep duration and a number of covariates. Cox proportional hazards analysis was applied to data. The results show that occupational group did not influence the association, but showed significant hazard ratios (HR) for long (≥9.5 h) and short (<6.5 h) sleep in both groups (HR > 1.35), with optimal sleep duration (lowest HR) with a wide span (6.5-9.5 h). Age groups in tertiles also showed significant U-shapes, with a wide span (6.5-9.5 h) for the younger 2/3 (54-74 years), but a weaker pattern for the oldest third (≥75 years), for which optimal sleep fell in the 6.5-7.5 h interval. It was concluded that occupational group does not influence the association between sleep duration and mortality in retired individuals, but that age does.