GLP-1 Drugs and Eating Disorders: Two Questions, One Studied

Two questions sit under this heading and only one of them has been studied. The researched question is whether these drugs treat binge eating disorder: a 2025 meta-analysis found five studies covering 182 participants, with greater weight loss of 3.81 kg.^[1] The unstudied question is what happens when a powerful appetite suppressant reaches someone with a restrictive eating disorder — a population deliberately excluded from every trial. If you are asking the second question about yourself, the evidence below is not the thing you need.

What has actually been measured

Of all eating disorders, binge eating disorder is the most common, and it is also the one where a drug affecting appetite and reward signaling has an obvious rationale. A 2025 systematic review and meta-analysis searched four databases and found five eligible studies, together enrolling 182 people.^[1] Pooled, GLP-1 treatment produced 3.81 kg more weight loss than comparators.

Note what that outcome is. The headline result is about weight, in a condition defined by episodes of eating and loss of control — not by body size. A treatment can move the scale in someone with binge eating disorder without touching the binges, and weight is the easier thing to measure.

The psychiatric picture is genuinely mixed

A 2026 systematic review looked at semaglutide specifically through a psychiatric lens, screening 342 papers down to 37 covering cohorts, pharmacovigilance, open-label studies and randomized trials.^[2] Its summary is that semaglutide appears effective and well tolerated across several psychiatric populations, with promise in binge eating disorder — acting on cognitive symptoms as well as weight.

And in the same paragraph: these drugs have been linked to depressive symptoms and suicidal ideation, alongside possible antidepressant effects, with the evidence for all of it described as preliminary.^[2] Both directions, in one review, held simultaneously.

That is not a contradiction so much as an honest description of an immature field. Our suicidality article covers the specific question of self-harm, where the larger evidence base points reassuringly and the labels changed accordingly in February 2026.

The question nobody is running trials on

Trials of appetite-suppressing drugs exclude people with active restrictive eating disorders, for obvious ethical reasons. That exclusion means the trial literature is silent on a situation that is now common: someone with anorexia nervosa, bulimia, or subclinical restriction obtaining a drug whose entire purpose is to make eating feel unnecessary.

A drug that removes hunger is not neutral for someone whose illness is built on ignoring it.

Two features of the current market make this worse rather than better. Telehealth prescribing is fast and largely questionnaire-based, and compounded sellers operate with less screening still — our pharmacy verification guide covers what a legitimate intake looks like. A BMI threshold is not an eating disorder screen, and someone with a restrictive disorder at a high body weight will pass one.

• Rapid weight loss can mask a disorder rather than reveal it, because the loss is attributed to the drug and treated as success.
• Losing the physical cue of hunger removes a signal that recovery-oriented treatment works to restore.
• The drugs are effective, which means an eating disorder can progress while every number on the chart looks like a good outcome.

What to do with this

If you have had an eating disorder in the past and are thinking about one of these drugs, that history belongs in the first conversation, not a later one. It is not automatically disqualifying — binge eating disorder and obesity overlap heavily, and the modest evidence points the right way — but it changes who should be involved and what gets monitored.

If someone you know has lost weight rapidly on one of these and their eating has become rigid, secretive or frightening to them, that is worth naming out loud. The drug working is not evidence that nothing is wrong.