Thresholds of physical activity associated with obesity by level of sedentary behaviour in children

Background: It is unknown whether moderate-to-vigorous physical activity (MVPA) thresholds for obesity should be adapted depending on level of sedentary behaviour in children.

Objective: The objective of the study is to determine the MVPA thresholds that best discriminate between obese and non-obese children, by level of screen time and total sedentary time in 12 countries.

Methods: This multinational, cross-sectional study included 6522 children 9-11 years of age. MVPA and sedentary time were assessed using waist-worn accelerometry, while screen time was self-reported. Obesity was defined according to the World Health Organization reference data.

Results: Receiver operating characteristic curve analyses showed that the best thresholds of MVPA to predict obesity ranged from 53.8 to 73.9 min d^-1 in boys and from 41.7 to 58.7 min d^-1 in girls, depending on the level of screen time. The MVPA cut-offs to predict obesity ranged from 37.9 to 75.9 min d^-1 in boys and from 32.5 to 62.7 min d^-1 in girls, depending on the level of sedentary behaviour. The areas under the curve ranged from 0.57 to 0.73 ('fail' to 'fair' accuracy), and most sensitivity and specificity values were below 85%, similar to MVPA alone. Country-specific analyses provided similar findings.

Conclusions: The addition of sedentary behaviour levels to MVPA did not result in a better predictive ability to classify children as obese/non-obese compared with MVPA alone.