Anders Grøntved

Background Children spend increasing amounts of time on recreational screen media, which may lead to an obesogenic environment. Objectives We investigated the association of trajectories of screen time across ages 3, 5 and 7 years with body composition at age 7 in the Odense Child Cohort. Methods Data were collected in the Municipality of Odense, Denmark, between 2010 and 2019. Group‐based trajectory modelling was applied to group participants into four trajectories of prospective parent‐reported screen time. Body composition was assessed using dual‐energy x‐ray absorptiometry with calculated fat‐mass index (FMI) as the primary outcome. Primary models were linear multivariable regression models adjusted for participants' sex, age, birthweight, maternal origin, maternal education, maternal body‐mass‐index, and maternal age. Further models were adjusted for additional possible confounders. Selection bias was addressed by inverse probability weighting. Results In total, 803 children (48.2% female) were included in the primary analysis. Participants with screen time at all time points were assigned to four trajectory groups [constant low screen time (12.7%), low increase (36.3%), high increase between ages 3 and 5 (33.5%) and high increase in screen time (17.5%)]. Sample characteristics differed across missing data status and trajectories. Mean FMI (kg/m 2 ) and standard deviation (SD) were 3.7 (SD 1.3) and 3.9 (SD 1.6) for the constant low versus high screen time, respectively. No differences in FMI were found between screen time trajectory groups at age 7 (adjusted mean difference 0.1 kg/m 2 , 95% confidence interval −0.3, 0.5 for constant low versus high screen time). No consistent associations between screen time groups and secondary body composition outcomes were found. Conclusions Results from this study do not suggest that recreational screen time from age 3 to 7 years is associated with adiposity or other measures of body composition.