Defining short and long sleep duration for future paediatric research: A systematic literature review

Short and long sleep patterns in children have been associated with a range of poor health outcomes. However, there is no consensus regarding the definitions of these abnormal sleep parameters in childhood for use in paediatric research. Given that there is a clear lack of definitions for sleep duration throughout paediatric sleep literature, this review aimed to establish recommendations for standard cut-offs of short and long sleep for children aged 1-16 years to enable homogeneity in future studies of paediatric sleep duration. Four databases were systematically searched to identify prospective studies that defined short or long sleep patterns in children. Included papers (38) were assessed for methodological quality, and their definitions were extracted to examine the current applied cut-offs in the literature for short or long sleep duration. The definitions were analysed in a regression model to summarize applied cut-offs from subjective data into cut-offs for short and long sleep duration. These models were fitted to reference values of three commonly cited paediatric population studies to establish new definitions of sleep duration for future use in research. Across the age groups there was little consensus in applied cut-offs for short and long sleep duration. This study found the best compromise for short sleep was defined as the 2.5th centile (hours = 0.25*age + 11) and long sleep as the 97.5th centile (hours = 0.017*age²  - 0.68*age + 16) of sleep duration in children. Recommendations for the hourly cut-offs of short and long sleep duration based on these percentiles were described.