Abstract
Background: Childhood obesity is one of the most serious challenges facing global health, with far-reaching negative and long-term physical, psychosocial and academic consequences. While insufficient sleep has emerged as a strong, independent risk factor for obesity in children, the mechanisms responsible remain uncertain. Observational research suggests that sleep deprivation is more strongly related to dietary intake than movement behaviour in children, but confounding is likely to be high. To date, little experimental research exists to determine causality.
Objectives: The primary aim of the Daily, Rest, Eating, and Activity, Monitoring (DREAM) randomised crossover trial and this thesis was to objectively determine whether mild sleep deprivation increased energy intake in the absence of hunger (EAH) in children aged 8-12 years (feeding experiment). Secondary aims of this thesis were to determine whether mild sleep deprivation influences daily energy intake, quality of dietary food choices, timing of eating, and the reallocation of movement behaviours when awake.
Methods: 105 healthy children (8-12 years) with normal reported sleep (~8-11 hours/night) were randomised to go to bed one hour earlier (sleep extension) or one hour later (sleep restriction) every night for one week each, while keeping morning wake-time constant. The two weeks were separated by a one-week washout. EAH was assessed with one clinic-based feeding experiment at the end of both intervention weeks, and dietary intake by two 24-hour diet recalls during both intervention weeks. Diet quality was classified by level of processing (i.e. ultra-processed versus less processed foods), and as core and non-core foods. Energy intake was examined for the total day, evenings only (after 5 pm) and for weekend and weekdays between interventions. Twenty-four-hour movement behaviours (sleep, physical activity (PA), sedentary behaviour (SB)) were measured via 24-hour actigraphy. Data were analysed according to ‘intention to treat’ and an a priori ‘per protocol’, set at a 30 minute difference in sleep duration between intervention conditions. Mixed effects regression models were used to determine mean differences (95% CI) between intervention conditions. Time gained from any sleep lost that was necessarily reallocated to other movement behaviour components in the 24-hour day (i.e. light PA (LPA), moderate to vigorous PA (MVPA), SB) was modelled using regression and presented graphically.
Results: There was no evidence of a difference in energy intake from EAH between sleep conditions when analysed as a crossover design. However, a learning effect was found, with children eating significantly less during the first course (-239 kJ; 95% CI: -437, -41) and significantly more in the second course (181 kJ; 38, 322) in the second week. A parallel analysis of the first week only indicated a small, non-significant increase in EAH with sleep deprivation (108 kJ; -204,419). The intention to treat analysis for daily dietary intake revealed that children consumed significantly more carbohydrates (10.7 g; 1.6,19.7), total sugar (6 g; 1.0, 11.0), and energy from non-core foods (416 kJ; 6.5, 826) during sleep restriction compared to the extension week. Differences were strengthened when restricted to children who met the a priori per protocol, with significantly higher total energy intake (361 kJ; 20, 702) and energy from ultra-processed foods (UPF) (523 kJ; 93, 952), in those children that had lost at least 30 minutes of sleep each night. There was no evidence of a difference in eating after 5 pm or eating on weekends between interventions, however energy intake was significantly higher on weekdays (520 kJ; 58, 982) during the sleep restriction compared to the extension week. During sleep restriction, children gained ~49 minutes of additional time awake, which was mostly reallocated to SB (28 minutes; 18,37) and LPA (19 minutes; 13,26), and a small amount to MVPA (3 minutes; 0,6), with the overall composition of the day remaining similar in proportional terms.
Conclusion: These results indicated that mild sleep loss in healthy children aged 8-12 years may play a role in paediatric obesity by increasing energy intake, particularly from non-core and ultra-processed foods. By contrast, changes to appetite regulation were not apparent, although the EAH experiment proved unsuitable for an experimental crossover design in children of this age. Furthermore, children were not less active when tired with the activity composition of the waking day remaining similar during both experimental conditions.