The interaction effect of sleep and physical activity on energy expenditure.
To address this gap, we propose to use data from iDATA to assess the independent, modifying, and synergistic effects of physical activity with sleep duration on components of energy expenditure.
For this proposed analysis, the primary independent variables during waking periods are counts and vector magnitude as measured using the ActiGraph GT3X accelerometer. The primary independent variable during non-waking periods is sleep duration identified using temporally-matched ActivPAL and ACT24 data. Physical activity and sleep duration estimates will be independently evaluated as (1) continuous estimates and (2) clinically meaningful categories based on recommended levels of physical activity and sleep duration.
The primary dependent variable is total energy expenditure (TEE) obtained via doubly labeled water assessment. Sub-analyses will be conducted on the components of total energy expenditure (resting metabolic rate [RMR], physical activity energy expenditure [PAEE). RMR will be estimated based on sex, age, height, and weight. PAEE will be estimated using the standardized method of subtracting the thermic effect of food (standard 10% of TEE) and RMR from TEE. Energy expenditure estimates will be treated as continuous variables.
(1) To quantify the independent association of sleep and physical activity with energy expenditure components (total energy expenditure, resting metabolic rate, and physical activity energy expenditure) using linear regression modeling. Hypothesis: There will be independent associations of sleep duration and physical activity (weekly estimates averaged over days worn) with total energy expenditure when accounting for potential confounding. Tests of hypotheses will be implemented using a flexible permutation approach in which the distributional form of the test statistic is left unspecified.
(2) To assess the inter-day associations between sleep duration and physical activity using linear mixed modeling. Two linear mixed models will evaluate whether change in one variable (i.e., physical activity) significantly impacts concurrent changes in another (i.e., sleep duration). The linear mixed models will include a random intercept to account for within-subject correlations. Hypothesis: There will be bidirectional associations between changes in sleep duration and changes in physical activity. Specifically, changes in sleep duration/physical activity will influence concurrent changes in physical activity/sleep duration, while controlling for total energy expenditure. All outcome and predictor variables in the linear mixed models will be standardized to aid interpretation, and tests of hypothesis will be performed using permutation approach.
(3) To examine the joint and modifying effects of sleep duration on physical activity and physical activity on sleep duration in regards to energy expenditure components using log-linear regression modeling (multiplicative scale). Hypotheses: (1) There will be a statistical (joint) interaction effect with sleep duration and physical activity on energy expenditure when accounting for potential confounding. (2) The association between sleep duration and total energy expenditure will depend upon the proportion of individuals accumulating at least 150 minutes per week of moderate- to vigorous- intensity aerobic activity (active). (3) The association between physical activity and total energy expenditure will depend upon the proportion of individuals sleeping for 7-9 hours per night (optimal sleep duration).
(4) To assess the mechanistic interaction of sleep and physical activity on the energy expenditure components using log-linear regression modeling. Hypothesis: A sufficient cause interaction between physical activity and sleep duration on energy expenditure is present.
1) Jaejoon Song, Department of Biostatistics, The University of Texas MD Anderson Cancer Center
2) Casey P Durand, Department of Health Promotion and Behavioral Sciences, The University of Texas Health Science Center
3) Joseph Cheung, Stanford Center for Sleep Sciences and Medicine, Stanford University
4) Kelley Pettee Gabriel, Department of Epidemiology, Human Genetics, and Environmental Sciences, The University of Texas Health Science Center