The impact of weight loss interventions on disordered eating symptoms in people with overweight and obesity: a systematic review & meta-analysis
Tsompanaki E., Koutoukidis DA., Wren G., Tong H., Theodoulou A., Wang D., Park RJ., Jebb SA., Aveyard P.
Background: It is unclear whether weight loss interventions worsen disordered eating in people living with overweight/obesity. We aimed to systematically evaluate the association between weight loss interventions and disordered eating. Methods: Six databases were searched from inception until September 2024. Trials of weight loss interventions in people with overweight/obesity were included if they reported a validated score for disordered eating on either the Eating Disorder Examination Interview or the Eating Disorder Examination Questionnaire pre- and post-intervention. Interventions included behavioural weight loss programmes (BWL) and pharmacotherapy licenced for weight loss, with or without concurrent psychological support, provided for at least 4 weeks. Pooled standardised mean differences (SMD) in scores of disordered eating were calculated using random effects meta-analyses. Risk of bias (RoB) was assessed using the Cochrane RoB 2 tool and the Newcastle–Ottawa scale for randomised and single-arm trials, respectively (PROSPERO ID: CRD42023404792). Findings: Thirty-eight studies with 66 eligible arms (61 interventions: 29 BWL, 11 BWL + pharmacotherapy, 20 BWL + psychological intervention, 1 pharmacotherapy + psychological intervention) and 3364 participants in total were included. The mean weight change was −4.7 kg (95% CI: −5.7, −3.7). Compared with baseline, disordered eating scores improved by −1.47 SMD units (95% CI: −1.67, −1.27, p < 0.001, I2 = 94%) at intervention completion (median of 4 months). Seven randomised trials that directly compared a weight loss intervention to no/minimal intervention reported an improvement of −0.49 SMD units (95% CI, −0.93, −0.04, p = 0.0035, I2 = 73%). Sub-group analyses showed: (a) disordered eating scores improved more in people with an eating disorder at baseline compared with people without high scores, (b) no clear evidence that the association depended upon intervention type, and (c) disordered eating scores improved more in trials rated at low overall RoB. Interpretation: Despite heterogeneity in effect size, weight loss interventions consistently improved disordered eating scores. These findings provide reassurance that weight loss interventions might not worsen disordered eating and may improve it. Funding: Novo Nordisk UK Research Foundation Doctoral Fellowship in Clinical Diabetes.