The Association between Sleep Timing with Diet and Physical Activity Levels in School-Aged Children

Background: Several studies have reported associations between shortened sleep durations and obesity in children, through mechanisms such as increased food intake and decreased physical activity levels. Recently, research has begun to look at sleep timing (the combination of sleep onset and sleep offset) as an important predictor of behaviour, independent of total sleep duration. However, no research has used objective assessment methods to examine sleep timing behaviour in relation to both diet and activity in school-aged children. Objectives: The primary objectives were to examine the association between sleep timing with diet and physical activity levels in school-aged children from New Zealand. Design: The Physical activity, Exercise, Diet, And Lifestyle Study was a cross-sectional observational study of children aged 8-11 years old and their parents, conducted in the Otago region of New Zealand between May and October 2015. Methods: A total of 468 child participants from primary schools in Dunedin, New Zealand took part in the study. Children were asked to complete lifestyle questionnaires and have anthropometric measurements taken during one school day, and to wear an accelerometer (Actigraph GT3X+) on their non-dominant wrist for eight consecutive days and seven consecutive nights. The participant’s self-reported frequency of breakfast, fruit and vegetables, and ‘extra’ foods consumption was used to assess diet. Accelerometer-measured sleep and moderate and vigorous physical activity intensity was used to assess sleep behaviour and activity patterns, respectively, while body composition was assessed using the participant’s calculated body mass index. Participants were classified into one of four sleep timing behaviour groups, using the median split for sleep onset (9:37pm) and sleep offset (6:50am) times, rounded to the nearest quarter hour: early sleep-early wake; early sleep-late wake; late sleep-early wake; late sleep-late wake. Late to sleep was defined as after 9:30pm, otherwise classified as early to sleep. Late to wake was defined as after 6:45am, otherwise classified as early to wake. Multivariate mixed regression analysis was used to examine associations, accounting for schools as clustering units. Analyses were undertaken using data from 370 participants who had a complete dataset for the outcome variables and covariates relevant to this thesis available prior to the 7th of November 2015. Results: The study population had a mean age of 10.2 (SD 0.6) years, were predominantly of ‘New Zealand European and Other’ ethnicity (89%), had a higher socio-economic status (49%), indicated by a New Zealand Deprivation Index 2013 score between one and three, and had a ‘normal’ body mass index (89%). Almost all participants (98%) provided at least five valid days of accelerometer data and 80% provided six valid days. Across the whole week, participants slept for an average of 8.6 (SD 0.7) hours per night, went to sleep at a mean time of 9:37pm (SD 0:36) and woke up at a mean time of 6:50am (SD 0:30). Girls slept longer and woke up later than boys, particularly on weekend days, where girls accumulated on average 15 minutes more sleep (8.7 vs 8.4 hours, p=0.011) and woke up 15.4 minutes later (7.04am vs 6.47am, p=0.001). In adjusted multivariate regression models, with the sleep timing behaviour group as the independent variable, no association was found with the consumption frequency of breakfast, fruit and vegetables, or ‘extra’ foods. Participants in the early sleep-early wake sleep timing behaviour group accumulated on average 13 more minutes of moderate and vigorous physical activity per day than those in the late sleep-late wake group (83 mins vs 70 mins, p=0.004), independent of total sleep duration. Participants in the group with sleep timing consistent with the shortest sleep duration (late sleep-early wake) had a significantly higher body mass index z-score than those in the group with sleep timing consistent with the longest sleep duration (early sleep-late wake), independent of total sleep duration (0.40 vs -0.04, p=0.017). Conclusion: This study concludes that sleep timing, independent of total sleep duration, was associated with physical activity and body composition, but not diet in this group of school-aged children. These findings suggest that sleep timing is an important factor to consider in addition to total sleep duration for good health. Further investigation of the causal mechanisms underlying the relationship that sleep timing has with health outcomes is necessary to provide direction for the development of effective strategies for both the treatment and prevention of childhood obesity. Key Words: sleep, diet, physical activity, children, timing, obesit