S1 Data -

Background
Many studies examining weight trajectories have used adiposity measures shown to be problematic for trajectory analysis in children with obesity, and remission of obesity remains poorly understood.
Objectives
To describe weight trajectories for school-aged children, the rate of obesity remission and factors associated.
Methods
Children between 6 and 11 years of age with ≥3 valid height and weight measurements from an Oregon hospital-system over a minimum six-month period were included. Percent distance from the median body mass index (BMI) was used for modeling. Latent class analysis and linear mixed models were used to classify children based on their weight trajectory.
Results
We included 11,247 subjects with a median of 2.1 years of follow-up, with 1,614 (14.4%) classified as overweight and 1,794 (16.0%) classified as obese. Of subjects with obesity, 1% experienced remission during follow-up, whereas 23% of those with overweight moved to within a healthy weight range. Latent class analysis identified three classes within each weight-based stratum over time. The majority of children with overweight or obesity had a flat trajectory over time. Lower socioeconomic status was associated with a worsening trajectory. Latent class models using alternate measures (BMI, BMI z-scores, tri-ponderal mass index (TMI)) differed substantially from each other.
Conclusions
Obesity remission was uncommon using the adiposity metric of distance from the median though transition from overweight to healthy weight was more common. Children with low socioeconomic status have worse trajectories overall. The choice of adiposity metric may have a substantial effect on the outcomes.