Sleep, Dietary Intake, and Physical Activity: Examining Bidirectionality Using Repeated-Measures Analyses

PURPOSE: Studies examining how sleep impacts diet and activity and vice versa have been limited by the use of aggregate data in cross-sectional designs. We examined how sleep might influence diet and physical activity the next day, and daytime diet and activity relate to sleep that night, using temporal repeated-measures analyses. METHODS: Twenty-four-hour time use (sleep onset, wake after sleep onset, total sleep time, sedentary time, light physical activity [LPA], and moderate-vigorous physical activity [MVPA]) was measured (AX3 accelerometers) over 7 days in adolescents (11-14.9 years) in Dunedin, New Zealand (March-December 2021). Bedtime was reported in daily diaries. Multiple-pass 24-hour dietary recalls were collected on days 3, 5, and 7. Mixed-effects regression models were used with participants as a random effect, adjusted for weekend days. RESULTS: In 75 youth (mean age 12.9 y, 36% indigenous Māori, 37.3% overweight or obese), going to bed 1 hour later was associated with less LPA (mean difference; 95% confidence interval: -3.8 minutes; -6.1 to -1.6) and MVPA (-14.0 minutes; -22.6 to -5.3) the next day. By contrast, LPA and MVPA during the day were not related to sleep that night. On days when bedtime was later, youth were more sedentary (25 minutes; 11-39, for each hour later) and consumed significantly more energy from ultraprocessed (618 kJ; 248-988) and noncore (732 kJ; 350-1113) foods and less from core foods (-419 kJ; -781 to -57). DISCUSSION: Sleep timing appears more strongly related to diet and physical activity than total sleep time in youth. Encouraging earlier bedtimes might be a practical target to reduce obesity risk.