Adolescents spend up to 6-8 h/day in sedentary screen behaviour and screen time is an independent risk factor for obesity. However, the mechanisms by which screen time confers obesity risk remain unclear. Via community level recruitment this study examined whether the relationship between screen time behaviours and body mass index (BMI: kg/m2) was mediated by total energy intake or macronutrient consumption. In a cross-sectional study of post-pubertal adolescents (N = 283: 86M, 197F) with overweight or obesity at baseline of an intervention for weight control, we examined self-reported total energy intake (mean Calories from 3 day food diary), macronutrient intake (grams/day of carbohydrate, fat, protein) and total screen time (aggregate of hours/day watching TV, playing seated video games, and recreational computer use). BMI was objectively measured and converted to standardized scores (z-BMI). Simple and multiple mediation analyses were conducted using the bootstrapping approach described by Preacher and Hayes. Covariates included age, sex, ethnicity, parental education, Tanner stage, and self-reported physical activity. The relationship between screen time and z-BMI was significantly mediated by energy intake. Higher levels of carbohydrate intake, but not fat or protein intake, significantly mediated the relationship between screen time and z-BMI (95% bias-corrected and accelerated confidence interval [0.0004, 0.0074]). Higher carbohydrate intake mediated the relationship between TV viewing and z-BMI, and video gaming and z-BMI. In conclusion, the relationship between screen time and BMI appears to be mediated by increased energy intake, primarily in the form of higher carbohydrate intake. It is possible that that reducing time spent watching TV and playing video games may reduce food intake and help promote dietary adherence needed for weight management in obese adolescents.
