Sleep characteristics have been associated with body mass index (BMI) among adolescents, but population-representative evidence on sleep timing and variation across demographic subgroups remains limited. We conducted a cross-sectional analysis of 2017–2018 California Health Interview Survey data from 880 adolescents aged 12–17 years. Weeknight sleep duration was obtained from the CHIS-derived sleep-duration variable, based on reported school-night bedtimes and school-day wake times. Sleep-timing measures included bedtime, wake time, and the derived sleep midpoint. Pooled survey-weighted linear and logistic regression models used 160 jackknife replicate weights and adjusted for age, sex, race/ethnicity, household poverty category, and survey year; sensitivity models additionally adjusted for dietary and behavioral factors. Sleep duration was not associated with BMI overall (p = 0.161) or with the odds of having overweight or obesity (p = 0.36). The association between sleep duration and BMI varied by age group (p for interaction = 0.042). Each additional hour of sleep was associated with a lower BMI among adolescents aged 14–17 years (β = −0.895 kg/m2; 95% CI: −1.464 to −0.326; p = 0.002), but not among those aged 12–13 years. However, there was no evidence of an interaction when age was modeled continuously (p for interaction = 0.36). Bedtime, wake time, and sleep midpoint were not independently associated with BMI. Exploratory evidence that the association between sleep duration and overweight or obesity varied by race/ethnicity was attenuated after adjustment for dietary and behavioral factors (p for interaction changed from 0.04 to 0.05). In this cross-sectional study, neither sleep duration nor sleep timing was associated with BMI in the overall sample, and sleep duration was not associated with BMI-defined weight status. Subgroup findings were not consistent across model specifications and should be examined in longitudinal studies using objective and repeated measures of sleep.
Zamora et al. (Wed,) studied this question.