Supplementary Material for: Long-term change in BMI for children with obesity treated in family-centered lifestyle interventions

Introduction Several evaluations of lifestyle interventions for childhood obesity exist; however, follow-up beyond two years is necessary to validate the effect. The aim of the present study was to investigate long-term weight development following children participating in one of two pragmatic family-centered lifestyle interventions for children with obesity. Methods This real-life observational study included Danish children 4-17 years of age classified as having obesity. Data from 2010-2020, from two community-based family-centered lifestyle interventions (designated hereafter as the Aarhus- and the Randers-intervention) were merged with national registers and routine health check-ups, including height and weight. Adjusted mixed effect models were used to model changes in Body Mass Index (BMI) z-score. We performed exploratory analyses of the development in BMI z-score within stratified subgroups of children treated in the interventions before investigating potential effect modifications induced by sex, age, family structure, socioeconomic or immigration status. Results With a median follow-up of 2.8 years (interquartile range: 1.3;4.8), 703 children participated in an intervention (445 the Aarhus-intervention; 258 the Randers-intervention) and 2,337 children were not invited to participate (no-intervention). Children in both interventions experienced a comparable reduction in BMI z-scores during the first 6 months compared to the no-intervention group (Aarhus-intervention: -0.12 SD/year and Randers-intervention: -0.25 SD/year). Only children in the Randers-intervention reduced their BMI z-score throughout follow-up (Aarhus-intervention vs. no-intervention: 0.01 SD/year; CI:-0.01;0.04; Randers-intervention vs. no-intervention: -0.05 SD/year; CI:-0.08;-0.02). In a subgroup comparisons combining the two interventions, family income below the median (-0.05 SD/year, CI: -0.02;-0.09), immigrant background (0.04 SD/year, CI: 0.00; 0.07), or receiving intervention less than one year (0.04 SD/year, CI: 0.00;0.08) were associated with a yearly change in BMI z-score. In addition, effect modification analyses did not observe any interaction by sex, age, family structure, socioeconomic or immigration. Conclusions Although the more dynamic intervention with longer duration obtained and sustained a minor reduction in BMI z-score, the clinical impact may only be modest. However, this effect may at the best be only modest and still not effective enough to induce a long