Background: Obesity is a heterogeneous chronic disease in which eating behavior phenotypes may influence treatment response. Yet, anti-obesity medication (AOM) selection is still largely guided by anthropometric and metabolic parameters, with limited use of behavioral phenotyping in routine practice. We evaluated whether multidimensional eating behavior changes, measured by the Brazilian Eating Behavior Phenotype Scale (Escala de Fenótipos do Comportamento Alimentar, EFCA), differ across commonly used AOMs in a real-world cohort. Methods: We conducted a retrospective, observational, real-world study in obesity outpatient care settings in São Paulo, Brazil. Adults with obesity (18–65 years) treated with a single principal AOM for 6 months and paired baseline/6-month follow-up EFCA and anthropometric data were included. Analyses focused on early responders (≥5% total body weight loss at 3 months). Five AOM groups available in Brazil were analyzed: semaglutide (oral or subcutaneous), naltrexone/bupropion, sibutramine, topiramate, and tirzepatide. Outcomes included percent weight loss, EFCA total score, and five EFCA subscales (hedonic, emotional, compulsive, hyperphagic, disorganized). Within-medication behavioral changes were assessed using paired tests and standardized effect sizes (Cohen’s dz, 95% CI), summarized in heatmap form. Results: The analytical cohort comprised 66 early responders with paired EFCA assessments at baseline and 6 months. EFCA profiling revealed distinct behavioral response fingerprints across AOMs. Effect size mapping showed predominantly large behavioral effects (many dz ≥ 0.8) in hedonic, emotional, hyperphagic, and compulsive domains. Strongest signals included emotional eating reductions with naltrexone/bupropion (dz 2.04), tirzepatide (dz 1.77), semaglutide (dz 1.52), and topiramate (dz 1.54); hedonic reductions with tirzepatide (dz 2.06), semaglutide (dz 1.55), and naltrexone/bupropion (dz 1.52); hyperphagic reductions with tirzepatide (dz 1.50) and semaglutide (dz 1.34); and compulsive reductions with topiramate (dz 1.41) and consistent effects across tirzepatide, semaglutide, and sibutramine (≈dz 0.95–0.96). Disorganized eating showed heterogeneous/attenuated responsiveness, from near-null with tirzepatide (dz 0.03) to large but imprecise effects in smaller groups (e.g., topiramate dz 1.24, wide CI). Conclusions: In this responder-enriched real-world cohort, AOMs showed distinct and reproducible EFCA behavioral signatures, supporting a clinically actionable phenotype-informed framework to prioritize, sequence, and monitor obesity pharmacotherapy beyond nonspecific weight reduction, while highlighting disorganization as a potential target for adjunctive behavioral strategies.
Pineda-Wieselberg et al. (Thu,) studied this question.