Relationships between body mass index and sleep quality and duration in adults 70 years and older

Objective: Associations between sleep and body mass index (BMI) are age dependent. In older adults, BMI norms representing normal and overweight combine into a single normal-weight category. We aimed to assess the nonlinear associations between age-appropriate BMI categories and sleep duration (SD) and sleep efficiency (SE) in older men and women, controlling for health and functional status. Methods: Secondary data analysis of the Hospitalization Process Effects on Functional Outcomes and Recovery included 719 community-dwelling adults age 70- years hospitalized because of nondisabling diagnoses. Self-report intake data regarding then condition prior to circumstances that led to hospitalization were used to obtain BMI categories (underweight: BMI <= 23, normal weight: 23 BMI < 30, and obese: BMI >= 30), SD, SE, health, and functional status. Analysis of covariance was used for modeling SD and SE separately, additively entering (1) BMI, (2) sex and BMI x sex, and (3) health and functional confounders. Results: For SD and SE, significant BMI group differences in the first model (P<.001) remained significant in the second (P<.001) and third (P<.01) models. High BMI was associated with shorter SD and lower SE compared with normal- and low-BMI groups. Controlling for sex, an inverted J-shaped relationship appeared in women, whereby low- and, more prominently, high-BMI categories were associated with shorter SD and lower SE compared with normal BMI. Although associations remained consistent in the fully adjusted models, effect sizes were small. Conclusions: Findings provide insights into possible mechanisms underlying BMI, sleep, and health and may contribute to informed clinical recommendations, particularly for older women. (C) 2016 National Sleep Foundation Published by Elsevier Inc. All rights reserved.