Eating disorders in the GLP-1 era

Glucagon-like peptide-1 receptor agonists are now prescribed for weight management well beyond the specialist clinics that first used them. That has put a question in front of clinicians which the trials were never designed to answer: what semaglutide and tirzepatide do to people whose relationship with food is already disordered. A systematic review and meta-analysis published in eClinicalMedicine in 2026 found that drugs in this class reduced binge eating and at the same time increased dietary restraint, and its authors wrote plainly that whether that restraint is pathological is unclear^[1]. That sentence is a fair summary of where the whole field currently sits.

What the trial evidence actually shows

The eClinicalMedicine review pooled 25 randomised controlled trials covering 8,069 participants, around two thirds of them women and a majority white, drawn from a literature search running from January 2005 to April 2026^[1]. Against control conditions, participants allocated to a GLP-1 receptor agonist reported moderate reductions in binge eating (Hedges' g of -0.23, 95% CI -0.36 to -0.10), lower loss of control eating, lower eating disinhibition (g of -0.54) and lower emotional eating (g of -0.30). The effects were larger in the trials that specifically enrolled people with diagnosed binge eating disorder^[1].

An earlier meta-analysis in Eating and Weight Disorders pointed the same way from a much smaller base: five included reports covering 182 participants with binge eating disorder, with Binge Eating Scale scores improving by 8.14 points against controls (95% CI -13.13 to -3.15) alongside greater weight loss^[2]. Those authors described the evidence for eating disorders other than binge eating disorder as limited, and asked for long-term trials across more diverse groups before anyone draws conclusions^[2].

The finding that has clinicians uneasy

In the same pooled analysis, cognitive or dietary restraint went up rather than down (g of 0.31, 95% CI 0.17 to 0.44)^[1]. Restraint here is a measured research construct, not a diagnosis. Read from an obesity clinic, a rise in it looks like the treatment doing its job. Read from an eating disorder service, it is one of the things clinicians keep an eye on. The review did not resolve which reading is correct, and it said so rather than papering over the gap.

The second problem is who was allowed into the trials in the first place. A 2026 systematic review of psychiatric outcomes on this drug class, published in Diabetes, Obesity and Metabolism, points to the routine exclusion of people with significant psychiatric comorbidity from GLP-1 trials, and treats that exclusion as a standing evidence gap for precisely the higher risk groups now being prescribed these medicines^[3]. The same review reported modest antidepressant signals, inconsistent findings on suicidality, and no settled answer on either^[3].

What the labels and the guidelines actually require

Less than most readers assume. The United States prescribing information for Wegovy lists warnings and precautions covering acute pancreatitis, gallbladder disease, hypoglycaemia, kidney injury from volume depletion, severe gastrointestinal reactions, hypersensitivity, diabetic retinopathy complications, heart rate increase and pulmonary aspiration under anaesthesia^[4]. Eating disorders appear nowhere on that list. The label's own record of recent major changes also shows that the suicidal behaviour and ideation section was removed in February 2026^[4].

UK guidance is uneven in a way worth noticing. NICE guideline NG246 tells clinicians to take a person's current or previous experience of eating disorders or disordered eating into account before discussing weight at all, and reminds them that someone can be affected by an eating disorder at any weight^[5]. For bariatric surgery the guideline goes considerably further, requiring the multidisciplinary team to provide specialist assessment for eating disorders, with referral or signposting to specialist services where appropriate^[6]. The medicines section that covers tirzepatide, semaglutide, liraglutide and orlistat carries no equivalent requirement^[6]. Our tirzepatide regulation page tracks where each of those medicines is licensed and in which settings it may be prescribed.

What we do not yet know

Most of the things that would matter. Every study in the eClinicalMedicine pool was rated as carrying a high risk of bias or as raising some concerns, and heterogeneity across them was substantial^[1]. The authors' own conclusion was that adequately powered trials in people with diagnosed binge eating disorder, followed for longer, are needed before any of this informs practice^[1]. Both reviews say the evidence for eating disorders other than binge eating disorder is thin, and that long-term data is simply missing^[2].

Nothing in either review speaks to what happens after treatment stops, and the samples skew heavily female and white, which limits what the numbers say about everyone else^[1]. The people obtaining these drugs outside a clinical pathway are not in any of this data at all. We have covered that access question separately in our write-up of a US survey on eating disorder screens and GLP-1 interest, and the online dimension in our report on how a pro-eating-disorder community on X discusses Ozempic.

If you are worried about your own eating

This article reports what clinicians and researchers are flagging. It is not guidance, and it deliberately does not describe how anyone eats, restricts or medicates. If your eating feels out of your control, or if weight loss on one of these medicines has started to crowd out everything else, that is worth raising with a healthcare professional, and worth raising early rather than late. A GP or your prescriber is the right first call. You do not have to wait until you are certain something is wrong.

Prescription status, licensed indications and who may prescribe differ by country, and our semaglutide regulation page sets out the current picture. This article is for educational and journalistic purposes and is not medical advice. Any decision about starting, continuing or stopping a weight management medicine belongs with your prescribing clinician.