Abstract
OBJECTIVE: This study aimed to provide a BMI-adjusted meta-analytical calculation of blood leptin levels across different eating disorders (EDs) including anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder (BED), recovered EDs, and healthy controls (HCs). The goal was to understand BMI-independent leptin alterations and their potential as biomarkers. METHOD: PubMed and ClinicalTrials.gov were searched for studies reporting serum leptin in AN, BN, BED, or recovered EDs. A multilevel network meta-analysis using a linear mixed-effects meta-regression model, adjusting for BMI, sex, and assay type, was performed on 146 studies (5048 patients, 3525 controls). RESULTS: Significant differences in leptin levels were found across EDs. AN patients exhibited the lowest leptin levels, while BED patients had the highest. BN and recovered AN patients had leptin levels similar to AN, significantly lower than HCs. BMI, sex, and assay type were significant covariates. The model accounted for heterogeneity due to diagnostic criteria, assay types, and study-level differences. CONCLUSIONS: Leptin levels in EDs are significantly altered beyond BMI effects, suggesting disease-specific factors. These findings support leptin's potential as a biomarker for ED staging and prognosis. Further research is needed to explore leptin's role in ED pathogenesis and trajectory, to identify subpopulations and improve clinical interventions.