| Domain | Key Finding | Evidence Level |
| The question | Do GLP-1RAs reduce binge eating, and does that make them a treatment for binge eating disorder (BED)? | Systematic review, 25 RCTs, n=8,069 |
| The headline number | Binge eating severity: g = −0.23 (95% CI −0.36 to −0.10), 9 trials, 2,106 participants. Small by convention (the authors call it moderate). Loss-of-control eating, disinhibition, and emotional eating moved the same way. | Meta-analysis; every trial high risk or some concerns of bias |
| The downside signal | Cognitive and dietary restraint rose: g = +0.31 (0.17 to 0.44), 13 trials. Whether that is healthy control or rigid restriction, the data cannot say. | Meta-analysis |
| The caveat | Only 3 of 25 trials selected people for binge eating: about 130 patients, 17 weeks at most. The one blinded, placebo-controlled trial in diagnosed BED (n=27) was null on binge episodes, −4.0 vs. −2.5 per week, p = 0.37. | Pilot RCTs, high risk of bias |
| The comparison | Lisdexamfetamine phase III effect sizes were 0.83 and 0.97. Psychological treatment produces binge abstinence in about half of completers. Not head-to-head, but not close. | RCTs; meta-analysis of 39 RCTs |
| Where it fits | A GLP-1RA is a defensible adjunct in obesity with binge eating, not monotherapy for diagnosed BED. Screen for eating pathology first: only 55% of direct-to-consumer sellers asked. | Extrapolation |
