RESUMO
OBJECTIVES: To assess changes in sleep parameters and circadian rhythm metrics measured by actigraphy in preschool-aged children. DESIGN: Longitudinal analysis over 1 year. PARTICIPANTS: Ninety-four children living in Tijuana and Ensenada, Mexico. MEASUREMENTS: Children wore accelerometers on the right hip for one continuous week at baseline and 1-year follow-up. Parents recorded child bedtime, waketime, and naps in sleep diaries. We used cosinor and nonparametric approaches to calculate circadian rhythm metrics. RESULTS: At baseline, children had a mean age of 4.2 years, and 51.1% were girls. In multivariable models adjusted for age, gender, BMI category, parental education, household income and city, at follow-up children had significantly earlier waketimes (ß = -7.99 minutes, p < .001) compared to baseline. Children also had lower sleep onset latency (ß = -2.32 minutes, p = .057), and longer nighttime sleep (ß = 9.38 minutes, p = .079), but these changes were not significant at the α < 0.05 level. We found significant increases in log relative amplitude (ß = 0.017, p = .009), and decreases in log midline estimated statistic of rhythm (ß = -0.084, p = .017) and log of the least active 5-hour period (ß = -0.057, p = .010). When we adjusted for co-sleeping, we found significant decreases in the number of nighttime awakenings (ß = -1.29, p = .011) but otherwise similar results. There were no other changes in sleep parameters or circadian rhythm metrics. CONCLUSIONS: Mean increases in nighttime sleep and earlier wake times over one year were concomitant with decreases in overall activity levels and increases in circadian rhythm robustness. Co-sleeping was a predictor of sleep disturbances. This study provides longitudinal evidence regarding changes in sleep and circadian metrics in a sample of children from an under-researched sociodemographic group during an important, early life period.