Objective: The epidemiological evidence for the association between sleep duration and obesity in the elderly is inconsistent and has not been investigated with objective measures. Furthermore, the role of sleep fragmentation in this relationship is unknown. Our aim was to investigate the association of sleep measures with body mass index (BMI) and obesity in a normal elderly population.
Design: Cross-sectional study.
Subjects: A total of 983 community-dwelling elderly (mean age 68.4+/-6.9 years, range, 57-97).
Measurements: Weight and height were measured, and sleep duration and fragmentation were assessed with on average six nights of actigraphy.
Results: A quadratic model adequately described the association between continuous measures of sleep duration and BMI. Actigraphic sleep duration had a significant U-shaped relationship with BMI (beta of quadratic term=0.30, 95% confidence interval (CI): 0.08, 0.52). Both short sleepers (<5 h: OR, 2.76 (95% CI: 1.38, 5.49), 5 to <6 h: OR, 1.97 (95% CI: 1.26, 3.08)) and long sleepers (>or=8 h: OR, 2.93 (95% CI: 1.39, 6.16)) were more likely to be obese, compared to participants who slept 7 to <8 h. BMI increased with 0.59 kg m(-2) per standard deviation of sleep fragmentation (95% CI: 0.34, 0.84). After adjustment for sleep fragmentation, the association between short sleep and obesity was no longer significant. Exclusion of participants with probable sleep apnea only marginally changed these associations. Self-reported habitual sleep duration was not associated with BMI or obesity.
Conclusions: Sleep duration, as measured with actigraphy, had a U-shaped relationship with BMI and obesity in an elderly population. A highly fragmented sleep is associated with a higher BMI and a higher risk of obesity, and may explain why short sleep is related to obesity. To preclude bias that can be introduced by self-report measures of sleep duration, using multiple measures of sleep parameters is recommended in future research.