P042 Actigraphy-derived sleep characteristics of healthy children who benefit the most from a sleep extension protocol

Abstract Introduction Although experimental manipulation of sleep can extend children’s sleep via earlier bedtimes, this is not always possible. We aimed to determine true sleep gains made with a one hour earlier bedtime, and which actigraphy variables at baseline predicted the best gains. Methods Secondary analysis within a randomised crossover trial (99 children; 49.5% female; mean age=10.3 years, range 8-12 years). Each child underwent one week of sleep restriction and extension in a random order (1h later or earlier bedtime respectively) separated by a one-week washout. Sleep data comparing baseline and extension weeks were analysed using linear or logistic regression, controlling for randomised order. Results The mean (SD) difference in total sleep time (TST) with sleep extension was 14.3 (44.1) min/night. Thirty-three percent extended their sleep by at least 30 min/night. At baseline, having lower sleep quantity (TST) and quality (efficiency, WASO, waking frequency) significantly predicted more sleep gain. For each 1 h lower TST at baseline, 29.7 min (95% CI 19.4, 40.1) more sleep was gained. At baseline, a TST below 8 h 49 min was the optimal cut-point to predict whether children could extend their sleep by at least 30 min/night (found in 40% of children). Adjusting for relevant demographics and sleep hygiene made little difference. Conclusion Children who slept less than 8 h 49 min/night obtained the most sleep gain from a 1 hour earlier bedtime maintained over a week. Maintenance over the longer term is unknown. Findings have implications for understanding the sleep requirements of children.