Cross-sectional and Longitudinal Associations of Replacing Sedentary Time with Active Behaviors on the Body Mass Index.

Abstract

The magnitude and direction of the association between sedentary time, physical activity (PA) of various intensities and weight status remain unclear. PURPOSE:The purpose of this study was to examine the cross-sectional and longitudinal effects of replacing sedentary time with light PA, moderate PA or vigorous PA on BMI in a large representative sample of adults and older adults. METHODS: We conducted cross-sectional (n=5013) and longitudinal (n=1823) analyses in a prospective cohort study in which PA was measured objectively by accelerometry (ActiGraph GT1M and GT3X+) for 7 days at both baseline (year 2008/09) and after 5 years. Exposure variables were daily time spent sedentary (below 100 cpm), and daily minutes spent performing light (100-2019 cpm), moderate (2020-5999 cpm) and vigorous intensity PA (6000 cpm and up). Outcome variable was BMI (kg/m2) derived from self-reported height and weight. Isotemporal substitution modelling was performed to convey the effects of replacing sedentary time with active behaviours on participant`s BMI. RESULTS: The isotemporal substitution models showed that replacing 10 minutes/day of sedentary time with an equal amount of time of either moderate PA (B=-0.2 95% CI: -0.2 to -0.1) or vigorous intensity PA (B=-1.0; 95% CI: -1.2 to -0.7) were associated with estimated beneficial effects on BMI after adjustments for age and socio-economic status in both genders. For men, replacing sedentary time with light intensity PA also showed small but significant beneficial associations with BMI (B=-0.1; 95% CI: -0.17 to -0.03). There were no longitudinal associations when replacing sedentary time with any of the PA intensities, after adjusting for baseline BMI. CONCLUSIONS: Replacing sedentary time with active behaviours is associated with BMI in cross-sectional but not prospective analyses. This may suggest that substituting sedentary time with PA at one point in time may not predict gain in BMI over time.