The four trials that stopped the drug
Almost everything known about stopping comes from four randomized trials in adults with obesity and without diabetes. In each one, people lost weight on semaglutide (Wegovy) or tirzepatide (Zepbound) and were then switched, without knowing it, to a placebo injection.[1][2][3][4]
| Trial | Drug and time on it | What happened after switching to placebo | Net result |
|---|---|---|---|
| STEP 1 extension (2022)[1] | Semaglutide 2.4 mg, 68 weeks; then drug and lifestyle program both stopped | Regained 11.6 percentage points of the 17.3% lost over the next 52 weeks, about two-thirds of the loss | Still 5.6% below starting weight at week 120; 48% still ≥5% down |
| STEP 4 (2021)[2] | Semaglutide 2.4 mg, 20 weeks run-in; counselling continued | Placebo group regained 6.9% over 48 weeks while the continuing group lost a further 7.9% | Difference of 14.8 percentage points between arms |
| SURMOUNT-4 (2024)[3] | Tirzepatide to max tolerated dose, 36 weeks (lost 20.9%) | Placebo group regained 14.0% over 52 weeks; continuing group lost a further 5.5% | Placebo group still 9.9% below baseline, keeping about half the loss; only 17% kept ≥80% of it (vs 90% who continued) |
| SURMOUNT-MAINTAIN (2026)[4] | Tirzepatide max dose, 60 weeks; then stay, step down to 5 mg, or placebo for 52 weeks | Weight change: −0.2% (stay) / +7.0% (5 mg) / +15.2% (placebo) | Share of loss kept: 96.5% / 67.9% / 42.8% |
Put simply: on average, people who stop regain roughly half to two-thirds of what they lost within a year, and the trials stop watching at that point.[1][3][4] No semaglutide or tirzepatide study has followed people off the drug for longer than about a year.[7] A 2026 meta-analysis in the BMJ pooled the regain rate at about 0.8 kg per month, roughly 9.9 kg in the first year, and projected a return to baseline around 1.5 years after stopping; another model projected a plateau at about 75% regained.[7][8] Both are extrapolations past the data.
Who these trials describe. Everyone in them tolerated the drug and reached a high dose; the Wegovy label itself warns that STEP 4 "may not reflect the experience of patients in the general population who are first starting" the drug.[6] All four were funded by the manufacturers.[1][2][3][4]
The spread: 100 people who stopped tirzepatide
Averages hide the most useful fact. In a post hoc analysis of SURMOUNT-4, researchers looked at the 308 people who had lost at least 10% on tirzepatide and then spent a year on placebo, and sorted them by how much of the lost weight came back.[5]
SURMOUNT-4, placebo year (n=308)
Regained less than a quarter
18of 100 people
The same pattern shows up in the semaglutide data. In the STEP 1 extension, the standard deviation of regain was 7.7 percentage points on a mean of 11.6, which means some people ended above their starting weight and others stayed far below it.[1] People who had lost the most kept the most in absolute terms: those who lost 20% or more were still 12.1% down a year after stopping, while those who had lost under 5% ended slightly above baseline.[1]
"Everyone regains everything" and "most people keep it off" are both wrong. Roughly one in six kept most of it; roughly one in four lost most of it back.[5]
Why the weight comes back
Two things are established, and one widely assumed thing is not measured.
- The drug leaves. Semaglutide's half-life is about a week; the Wegovy label says it stays in the blood for about 5 to 7 weeks after the last 2.4 mg dose.[6] Tirzepatide's half-life is about 5 days, so it is largely gone in about three and a half weeks.[9]
- Weight loss by any method triggers push-back. In a classic study of people who lost 14% on a very-low-calorie diet (no drugs), hunger hormones were still shifted toward "eat more" a full year later, even after regain had begun.[10] The body defends the old weight.
- Assumed but unmeasured: that returning appetite drives the regain. None of the four withdrawal trials measured hunger, fullness or food intake after stopping.[1][2][3][4] In interviews at an Irish obesity clinic, people described appetite coming back "within a week" or two and weight creeping back after about 6 to 8 weeks, but those are individual accounts, not measured rates.[11]
What this means in plain language: regain after stopping is the expected biology of obesity reasserting itself once the treatment is withdrawn, the same way blood pressure rises when a blood-pressure medicine is stopped. It is not a sign of personal failure, addiction or that the drug "stopped working".[7][12]
Blood pressure and blood sugar after stopping
The health improvements track the weight. In the STEP 1 extension, improvements in blood pressure, blood fats and HbA1c (the three-month blood sugar average) drifted back toward baseline as weight returned.[1] In the SURMOUNT-4 post hoc analysis, the people who regained 75% or more were back to roughly their starting values on most markers, while those who regained under 25% mostly kept their gains. Systolic blood pressure rose in every regain group, by 6.8 to 10.4 mmHg.[5] A rise in HbA1c was visible within 16 weeks of stopping.[5]
In SURMOUNT-MAINTAIN, 92.7% of people with prediabetes who stayed on full-dose tirzepatide had normal blood sugar at the end, versus 51.0% of those switched to placebo.[4] An 18-trial meta-analysis in obesity found an average 5.6 kg regain and a 0.25 percentage-point HbA1c rise after discontinuation.[13]
If you have type 2 diabetes. Blood sugar can rise within weeks of stopping, and other diabetes medicines may need adjusting. This is a reason the decision to stop, and how, belongs in a conversation with your prescriber rather than on your own.[5][15]
What has been tested for keeping it off
Only a few strategies have randomized evidence after a GLP-1-class drug, and most of it points the same way.
- Staying on a medicine. This is the only strategy with strong evidence. Continuing full-dose tirzepatide kept 96.5% of the loss; stepping down to 5 mg kept 67.9%; stopping kept 42.8%.[4] After injectable treatment, switching to the daily oral drug orforglipron maintained about 75 to 79% of the loss over a year, versus 38 to 49% on placebo.[14] The 2026 American Diabetes Association standards describe continued pharmacotherapy as the approach supported by evidence for maintenance.[15]
- A lower dose. In SURMOUNT-MAINTAIN the 5 mg arm regained about 7% over a year and had not plateaued by the end, so a lower dose slowed regain but did not stop it.[4]
- Exercise built during treatment. In the S-LiTE trial (liraglutide, an older GLP-1 drug), people who had done supervised exercise alongside the drug regained 7.1 kg in the year after everything stopped, versus 9.6 kg for drug alone and 3.6 kg for exercise alone; the combination group stayed 5.1 kg below placebo.[16][17] It is the best hint that habits built during treatment carry over, with the caveats that the drug was liraglutide, the exercise was supervised and aerobic, and a third of participants were not followed.
- Protein, lifting, daily weighing. Reasonable, but the evidence comes from diet-only maintenance trials, not from people coming off a GLP-1.[7][18] No trial after GLP-1 withdrawal has tested them. The BMJ meta-analysis found no effect of behavioural support offered after stopping.[7]
Does tapering help?
Short answer: nobody knows. A 2026 meta-analysis noted that every included trial stopped the drug abruptly and that guidelines give no tapering recommendations.[19] The only tapering data is a company-authored conference abstract from a Danish digital clinic: 353 people who tapered semaglutide were on average 1.5% lighter than at cessation six months later, but there was no randomized comparison and the tapering was self-selected.[20] One randomized tapering trial is registered with results pending.[21] Semaglutide's long half-life already produces a built-in 5-to-7-week fade after the last dose, so whether a deliberate step-down adds anything is an open question.[6]
Real-world stopping
Trial participants are not typical. In a US database of 125,474 people on these drugs without diabetes, 64.8% had stopped by one year and 84.4% by two years; 36.3% of those who stopped restarted within a year, and each 1% of weight regained was associated with a 2.8% higher likelihood of restarting.[22] In 7,881 Cleveland Clinic patients, one-year weight loss was 3.6% for those who stopped within three months, 6.8% for those who stopped between three and twelve months, and 11.9% for those who continued.[23] Real-world regain looks milder than in trials, partly because people often took lower doses, stopped earlier, or restarted.[22][23]
Heard vs evidence
The short version
Short on the way. A 35-second captioned version of this page is in production for the Fewer Bites channel. It will be embedded here when it is live.
Questions to bring to your prescriber
Print these or screenshot them. They are conversation starters, not a checklist of things to do.
- If I ever needed to stop (cost, supply, side effects, pregnancy plans), what would you want the plan to be?
- Would a lower maintenance dose or a different medicine be an option for me rather than stopping altogether?
- If I have type 2 diabetes or prediabetes, how would we watch my blood sugar in the weeks after stopping, and would other medicines need adjusting?
- What should I expect my blood pressure and cholesterol to do if weight returns?
- Is there a structured exercise or nutrition program you would want me in while I am still on the medicine, so the habits are built before any stop?
Education, not advice. This page explains what trials, labels and guidelines report. It does not tell anyone to start, change or stop a medicine. Your prescriber knows your history; this page does not.
Sources
- Wilding JPH et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: the STEP 1 trial extension. Diabetes Obes Metab 2022. primary read · industry-funded
- Rubino D et al. Effect of continued weekly subcutaneous semaglutide vs placebo on weight loss maintenance (STEP 4). JAMA 2021. primary read · industry-funded
- Aronne LJ et al. Continued treatment with tirzepatide for maintenance of weight reduction (SURMOUNT-4). JAMA 2024. primary read · industry-funded
- SURMOUNT-MAINTAIN: tirzepatide maintenance dose vs step-down vs placebo. Lancet 2026;407:2305-18. primary read (mirror PDF) · industry-funded
- Horn DB et al. Weight regain and cardiometabolic changes after tirzepatide withdrawal: SURMOUNT-4 post hoc analysis. JAMA Intern Med 2025. primary read · post hoc · industry-funded
- FDA. Wegovy (semaglutide) prescribing information, rev. 02/2026, sections 12.3 and 14.2. label
- West et al. Weight regain after stopping weight-management drugs: systematic review and meta-analysis. BMJ 2026. primary read · NIHR-funded
- Budini B et al. Trajectory of weight regain after cessation of GLP-1 receptor agonists. eClinicalMedicine 2026;93:103796. abstract only
- FDA. Zepbound (tirzepatide) prescribing information, 2026, section 12.3. label
- Sumithran P et al. Long-term persistence of hormonal adaptations to weight loss. NEJM 2011. primary read
- Qualitative interviews on stopping GLP-1 therapy, Irish specialist obesity service. Clinical Obesity 2026. primary read · qualitative
- Rajjo T et al. Long-term use of obesity management medications (editorial). Am Fam Physician 2026;113(2):118-119. secondary · editorial
- Tzang CC et al. Metabolic rebound after GLP-1 receptor agonist discontinuation: 18-RCT meta-analysis. eClinicalMedicine 2025;90:103680. abstract only
- ATTAIN-MAINTAIN: orforglipron for maintenance of body weight reduction after injectable therapy. Nature Medicine 2026. primary read · industry-funded
- American Diabetes Association. Standards of Care in Overweight and Obesity 2026. primary read · guideline
- Lundgren JR et al. Healthy weight loss maintenance with exercise, liraglutide, or both combined (S-LiTE). NEJM 2021. primary read
- Jensen SBK et al. Healthy weight loss maintenance with exercise, GLP-1 receptor agonist, or both, followed by one year without treatment (S-LiTE follow-up). eClinicalMedicine 2024. primary read · exploratory
- Wing RR et al. A self-regulation program for maintenance of weight loss. NEJM 2006. primary read · diet-only
- Qi H et al. Weight regain following discontinuation of GLP-1 receptor agonists: meta-analysis. PeerJ 2026;14:e21713. abstract only
- Embla digital clinic. Semaglutide tapering: observational cohort, European Congress on Obesity 2024 (abstract). conference abstract · company-authored · not peer-reviewed
- Randomized trial of GLP-1 receptor agonist tapering, NCT07294950 (protocol published in PLOS ONE). protocol · results pending
- Rodriguez PJ et al. Discontinuation and reinitiation of dual-labeled GLP-1 receptor agonists among US adults with overweight or obesity (Truveta, n=125,474). JAMA Netw Open 2025. primary read · observational
- Gasoyan H et al. Early discontinuation, dose and one-year weight outcomes in 7,881 patients. Obesity 2025. primary read · observational