GLP-1 and emotional eating research suggests these drugs can quiet cravings and reduce binge-type eating, though the effect is modest and the trials are small. A 2026 meta-analysis of 25 randomized trials (8,069 people) found GLP-1 drugs lowered emotional eating, binge eating and loss-of-control eating compared with control groups, while increasing dietary restraint.[1] No GLP-1 drug is approved to treat emotional eating or binge eating disorder.
For the basics of how these medicines work, start at the GLP-1 basics hub. For mood effects, see our page on GLP-1 drugs, depression and suicidal thoughts.
Bottom line
- Cravings and appetite: well shown in short, controlled feeding studies of semaglutide and tirzepatide.[2][3]
- Emotional and binge eating: reduced in pooled trials, but most trials were small, varied and at risk of bias.[1]
- Binge eating disorder: promising early data, no approval, and a head-to-head trial is underway.[4][7]
- After stopping: appetite and weight tend to return.[6]
What emotional eating is, and why a GLP-1 might touch it
Emotional eating means eating in response to feelings such as stress, boredom or sadness rather than hunger. It overlaps with binge eating (eating a large amount in a short time with a sense of losing control) and with what people online call “food noise,” constant thoughts about food.
GLP-1 drugs mimic a gut hormone that signals fullness. They slow stomach emptying and act on brain areas tied to appetite and reward. That second effect is why researchers ask whether they change the drive to eat, not just hunger. The same reward-pathway idea is behind research on GLP-1 drugs and addiction.
The key studies
| Study | Design and size | Main result |
|---|---|---|
| Semaglutide appetite trial (2017)[2] | Crossover RCT, 30 adults with obesity, 12 weeks, up to 1 mg | 24% fewer calories at free-choice meals; less hunger and craving; better control of eating; lower preference for high-fat foods |
| Tirzepatide eating-behavior trial (2025)[3] | RCT, 114 adults, 6 weeks, with a liraglutide comparison arm | About 525 fewer calories at a buffet lunch at week 3; less craving, overeating and food reactivity |
| Meta-analysis (2026)[1] | 25 RCTs, 8,069 people | Small-to-moderate drops in binge eating, loss-of-control eating, disinhibition and emotional eating; more restraint |
| Binge eating systematic review (2026)[4] | 12 studies of liraglutide, semaglutide and dulaglutide, mostly under 75 people | Fewer binges and lower binge eating scores; no new psychiatric safety signals |
Semaglutide: less craving, better control
In the 2017 trial, people with obesity took semaglutide or placebo for 12 weeks each, in random order. On semaglutide they ate 24% fewer calories across lunch, dinner and snacks. They also reported less hunger and fewer food cravings, felt more in control of their eating, and showed less preference for fatty, energy-dense foods. Nausea ratings were similar between the two periods, so the drop was not simply from feeling sick.[2]
Tirzepatide: changes in food reactivity
In a 6-week Lilly-funded trial, tirzepatide reduced how much people ate at a test lunch by about 525 calories compared with placebo. It also lowered self-reported cravings, the tendency to overeat and reactivity to tempting foods nearby. Brain scans showed less activation to photos of high-fat, high-sugar foods in reward-related regions. Tirzepatide did not change “volitional restraint,” meaning deliberate dieting.[3]
Pooled trials on emotional and binge eating
The 2026 meta-analysis is the best summary so far. Across trials, GLP-1 drugs reduced emotional eating (4 trials, 429 people), binge eating (9 trials, 2,106 people) and eating disinhibition. Effects were larger in trials of people diagnosed with binge eating disorder. The authors rated every study as having some concerns or high risk of bias, and results varied widely between trials.[1]
Food noise
A 2026 nationally representative US survey found that about 18% of adults frequently experience food noise. It was more common in people who reported loss of control over eating, and also more common among current and former GLP-1 users, which may reflect who seeks out these drugs.[5] No large trial has yet used food noise as a main outcome.
What’s established vs preliminary
Established: semaglutide and tirzepatide lower hunger, cravings and calorie intake in controlled settings.[2][3]
Preliminary:
- whether they treat binge eating disorder as well as existing options such as therapy or lisdexamfetamine[4]
- whether changes in emotional eating last
- whether higher dietary restraint is healthy or a risk for some people[1]
Limitations and cautions
- Short trials. Most lasted weeks to months.
- Self-report. Emotional eating is measured with questionnaires, which can shift if people know they are losing weight.
- Restraint is a double-edged sword. The meta-analysis found more dietary restraint on GLP-1 drugs, and the authors say it is unclear whether that is harmful.[1] People with anorexia, bulimia or a history of restrictive eating were generally excluded from trials.
- The underlying feelings remain. A drug that quiets appetite does not address stress, low mood or grief. Many clinicians pair medication with therapy for this reason.
- Stopping. In the STEP 1 extension, people regained about two-thirds of their lost weight within a year of stopping semaglutide, suggesting appetite effects fade off the drug.[6] See our guide on stopping a GLP-1.
What’s next
A phase 2 trial at Johns Hopkins is comparing tirzepatide with placebo and with lisdexamfetamine in 105 adults with binge eating disorder, with primary completion planned for late 2027.[7] Other trials are testing GLP-1 drugs alongside behavioral therapy.
Questions to discuss with a clinician
- Is my eating mostly hunger, emotional, or binge-type, and would therapy help alongside or instead of medication?
- Do I meet the criteria for a GLP-1 drug for weight or diabetes?
- How will we check that I’m eating enough protein and nutrients? Our GLP-1 diet guide covers this.
- What’s the plan if cravings come back when the dose changes or I stop?
