Cross-sectional associations between objectively measured sleep characteristics and body mass index in European children and adolescents

Background: Short sleep duration has been found to be associated with a higher risk for overweight and obesity. However, previous studies have mainly relied on subjective measures of sleep duration and other sleep characteristics (eg quality, timing) have often been neglected. Therefore, we aimed to investigate associations between several, mainly objectively measured sleep characteristics and body mass index (BMI). Further, we aimed to identify distinct sleep subtypes based on these sleep characteristics and to study their association with BMI. Methods: Children aged 9-16 years participating in the European I.Family study (N = 559, 51.2% girls, 32.9% overweight/obese) wore an accelerometer for one week on their wrist and recorded their daily wake-up and lights-off times in a sleep diary. Information on sleep duration, sleep efficiency and sleep latency was derived. To identify sleep subtypes, we conducted a latent class analysis using all five sleep variables. Associations between single sleep variables, sleep subtype and age- and sex-specific BMI z-score were investigated using linear mixed-effects regression models to accommodate clustering among siblings. Results: No statistically significant associations were observed between the single sleep variables (sleep duration, sleep efficiency, sleep latency, wake-up and lights-off times) and BMI z-score. Four sleep subtypes were identified and children were assigned to one of the groups based on their highest probability for latent group membership: "early birds" (17.5% of the sample), "short sleep duration" (14.7%), "optimal sleep" (47.6%) and "poor sleep quality" (20.2%). Sleep subtype was not associated with BMI z-score. Conclusions: Using objective sleep data, we did not find convincing evidence for associations between the sleep variables under investigation and BMI.