Supplementary Material for: Risk scores for metabolic dysfunction-associated fatty liver disease in pediatric obesity

Introduction: Early non-invasive detection of incipient liver damage is crucial to prevent long-term adverse health outcomes. A variety of scores to assess liver status have been proposed, mostly for adult populations. Validation of non-invasive hepatic scores to identify children at risk of metabolic dysfunction-associated fatty liver disease (MAFLD) is a gap in research, particularly in youth with severe obesity considering pubertal stage and sex. Methods: In a well-characterized pediatric population aged 9 to 19 years (n=115) 19 published liver scores were analyzed. AUROC for determination of MAFLD as assessed by magnetic resonance imaging was calculated. Results: The pediatric indices PNFI, B-AST and M-APRI and several scores developed in adults significantly differed in children with MAFLD compared to children without, while some established indices did not. Only NAFLD-LFS and the model by Cao et al. showed acceptable predictive accuracy (AUROC > 0.8) independently of pubertal stage and sex. When stratifying for pubertal stage and sex, the GSG-Index was superior in pubertal girls, and NAFLD-LFS performed best in pubertal boys. Conclusion: NAFLD-LFS and the model by Cao et al. were well suited to predict MAFLD in youth with severe obesity. In pubertal children, GSG-Index and NAFLD-LFS performed best in girls and boys, respectively.