Table 1

Summary of major studies reporting an association between short sleep duration and obesity

First author, publication year, country	Number studied, age group	Design	Key findings
ALSPAC, Avon Longitudinal Study of Parents And Children; BMI, body mass index (kg/m²); MOR, multivariate odds ratio, TV, television.
Locard,¹⁷ 1992, France	1031, 5-year-olds	Cross-sectional, case–control	Short sleep duration was associated with obesity (OR 4.9); this was independent of watching TV

Kagamimori,¹⁴ 1999, Japan	9668, 3-year-olds	Cross-sectional, case–control Toyoma cohort study	Frequency of sleeping <10 h was greater in obese children (29.3%) v that in non-obese children (13.7%)

Vioque,²⁷ 2000, Spain	1772, ⩾15-year-olds	Cross-sectional	Prevalence OR for obesity 0.43 (CI 0.27 to 0.67) for sleeping ⩾9 h v ⩽6 h; prevalence OR for obesity was 24% lower for each additional sleeping hour/day

von Kries,²⁸ 2002, Germany	6862, 5–6-year-olds	Cross-sectional	Short sleep duration was associated with overweight, obesity and increased body fat percentage; adjusted OR for obesity 0.52 (CI 0.34 to 0.78) for sleeping 10.5–11 h and 0.46 (CI 0.28 to 0.75) for sleeping 11.5 h

Sekine,²¹ 2002, Japan	8274, 6–7-year-olds	Cross-sectional, based on Toyoma cohort study	MOR (adjusted for age, parental obesity, physical activity, watching TV, frequency of eating breakfast, snack frequency) v ⩾10 h was 1.49 (CI 1.08 to 2.14) for 9–10 h, 1.89 (CI 1.34 to 2.73) for 8–9 h and 2.87 (CI 1.61 to 5.05) for <8 h; ORs for boys were greater than for girls

Gupta,¹¹ 2002, USA	383, 11–16-year-olds	Cross-sectional Heartfelt study	Only study to use 1-day actigraphy; short sleep associated with obesity and greater body fat; adjusted OR for obesity decreased by 20% for every hour of increased sleep

Agras,² 2004, USA	150, 9½-year-olds	Prospective study of newborn infants from birth to 9.5-year-olds	Negative correlation between hours of sleep and overweight (−0.21); overweight children sleep less than 30 min on average than children with normal weight, the difference was mainly in daytime sleep; hours of sleep were negatively correlated with high activity levels

Gibson,⁹ 2004, UK	1294, 7–18-year-olds	Cross-sectional, national diet and nutrition survey of young people	Obese children spent less time in bed (10–20 min 1st quintile of age-adjusted BMI v 5th quintile), but this was statistically significant only in boys

Reilly,²⁰ 2005, UK	8234, 7-year-olds	Children of the 90s (ALSPAC) Bristol	25 factors examined at age 30 months that could predispose to obesity at age 7 years; 8 factors found to be significant, including short sleep duration; OR v 12.5 h of sleep was 1.04 (CI 0.76 to 1.42) for 11–11.9 h; 1.35 (CI 1.02 to 1.79) for 10.5–10.9 h; 1.45 (CI 1.10 to 1.89) for <10.5 h

Padez,¹⁸ 2005, Portugal	4511, 7–9-year-olds	Cross-sectional	OR v 8 h sleep; overweight: OR 0.46 (CI 0.40 to 0.51) for 9–10 h, 0.44 (CI 0.38-0.49) for ⩾11 h; obesity: OR 0.44 (CI 0.40 to 0.47) for 9–10 h, 0.39 (CI 0.35-0.42) for ⩾11 h

Knutson,¹⁵ 2005, USA	4486, 15–18-year-olds	Cross-sectional, national longitudinal study of adolescent health	Results significant for men only; linear regression: sleep duration significantly predicted BMI z score (β = −0.08, CI −0.12 to 0.03); logistic regression: sleep duration predicted risk of overweight (OR 0.90, CI 0.82 to 1.00)

Chaput,³ 2006, Canada	422, 5–10-year-olds	Cross-sectional, “Quebec en Forme” project	MOR of obesity v 12–13 h sleep was 1.42 (CI 1.09 to 1.98) for 10.5–11.5 h sleep, 3.45 (CI 2.61 to 4.67) for 8-10 h sleep; waist circumference negatively correlated with sleep duration (r = −0.24), but only significantly in boys

Chen,⁴ 2006, Taiwan	656, 13–18-year-olds	Cross-sectional, case–control	High adequate sleep (defined as 6–8h sleep/night on >4 weekdays/week) was associated with non-obesity, OR 1.74 (CI 1.3 to 2.4)