Body Mass Index
The distribution of median BMI z-score is presented in Figure 1. We did not observe any significant difference in BMI z-score over the first six months of treatment (Fig. 1A) or significant changes in weight classification across the study period (Supplementary Table #S2). To evaluate differences in BMI by clinical and demographic variables associated with increased BMI, Wilcoxon rank sum test was performed and evaluated BMI by risk group, sex, and ethnicity at each timepoint (Fig. 1B-D). BMI z-scores were not significantly different between low risk, intermediate risk, and high risk groups at study entry, end of induction, or end of 6-months (Fig. 1B, P = 0.89, P = 0.48, and P = 0.19, respectively). BMI z-scores were also not significantly different between females and males at each timepoint (Fig. 1C, P = 0.69, P = 0.24, and P = 0.94, respectively) or Hispanics and Non-Hispanics at each timepoint (Fig. 1D,P = 0.71, P = 0.88, and P = 0.13, respectively).
We performed median regression analysis to evaluate the relationship between changes in dietary variables and BMI z-score (Supplementary Table #S3). We did not find any additional significant relationships between dietary intake of specific nutrients or food groups and BMI z-score after adjustment for age at study entry, sex, ethnicity, and baseline BMI z-score.