GLP-1 weight regain: what the trials actually show.
STEP-4, SURMOUNT-4, and the off-treatment extension data. How much weight comes back when therapy stops — what the magnitude actually is, why it happens, and what the trials say about preventing it.
Why "weight regain" is the most-asked question of 2026
For most of the cultural conversation around GLP-1 receptor agonists (GLP-1 RAs), the headline number has been the on-treatment weight loss. STEP-1 reported a mean −14.9% at 68 weeks [Wilding 2021]. SURMOUNT-1 reported −20.9% at 72 weeks for tirzepatide 15 mg [Jastreboff 2022]. Those are the numbers people remember.
The number people are now asking about is the off-treatment number. With supply constraints, with reimbursement changes, and with the simple fact that a lot of patients want to know what happens when they eventually stop — the withdrawal data has become the second-most-cited body of trial evidence in the entire class.
The short version: weight regain after discontinuation is real, substantial, and predictable. It is not a sign that the drug "doesn't work." It is a sign that the drug is doing exactly what its mechanism predicts.
Regain is not a failure of willpower. It is the visible signature of a metabolic adaptation that the drug had been suppressing.
STEP-4: the cleanest withdrawal data
STEP-4 is the most informative trial for the regain question because it was specifically designed to test what happens when therapy is stopped. Participants received open-label semaglutide 2.4 mg weekly for a 20-week run-in (titration plus stabilization), then were randomized to continue semaglutide or switch to placebo for an additional 48 weeks [Rubino 2021].
The continuation arm lost an additional ~7.9% from the randomization point (so a cumulative ~17% from baseline). The placebo arm gained ~6.9% from randomization — ending up at roughly −5% from baseline. The between-group difference at week 68 was ~14.8 percentage points. That gap is the cleanest "with vs. without drug" comparison in the published literature.
STEP-1 extension: 1 year off drug
A separate extension study followed STEP-1 participants for one year after the trial's 68-week active-treatment period ended [Wilding 2022 extension]. In that off-treatment year, the semaglutide group regained approximately two-thirds of the weight they had lost. Cardiometabolic improvements (glycemia, blood pressure, lipids) followed the same pattern — improving on drug, partially reversing off drug.
Two-thirds regain over one off-treatment year is the right number to anchor on. It is more than the casual reading of the headline result suggests, and it is consistent with the broader literature on pharmacologic weight loss across drug classes — the magnitude of regain tracks the magnitude of initial loss [Garvey 2020].
SURMOUNT-4: tirzepatide withdrawal
SURMOUNT-4 mirrors the STEP-4 design with tirzepatide. After 36 weeks of open-label tirzepatide at the maximum tolerated dose, participants were randomized to continue tirzepatide or switch to placebo for an additional 52 weeks [Aronne 2024]. The continuation arm lost an additional ~5.5%; the placebo arm regained ~14% — for a between-group difference at week 88 of ~19.4 percentage points.
The regain magnitude is larger in absolute terms than STEP-4, which is consistent with tirzepatide producing larger initial weight loss. The proportional regain pattern is broadly similar.
None of these trials say that the drug "stops working." They say that the drug works for as long as it is being administered, and the biology it had been suppressing reasserts itself when it is not. That is the same pattern observed with antihypertensives, statins, and SSRIs. We do not describe lisinopril as failing when blood pressure rises after discontinuation.
Why regain happens — the biology, not the willpower
Three mechanisms converge to drive post-discontinuation regain:
1. The set-point defended by the hypothalamus
Weight is regulated centrally. The hypothalamic arcuate nucleus integrates leptin, insulin, ghrelin, GLP-1 and other signals to produce a homeostatic "set-point" defended by adjustments to appetite and energy expenditure [Schwartz 2017]. Sustained weight loss shifts these signals — leptin falls, ghrelin rises, resting energy expenditure drops more than body-mass change alone predicts [Sumithran 2011]. These adaptations persist for years after weight loss.
2. The drug was suppressing those adaptive signals
GLP-1 RAs override the post-weight-loss hyperphagia signal at the hypothalamic level. Removing the drug does not reset the set-point — it removes the suppression. The hunger and appetite changes that STEP-4 participants reported on placebo were not a return to baseline appetite. They were a return to the elevated appetite signal that defends the higher set-point, now visible because the drug was no longer masking it.
3. Lean mass loss compounds the math
Roughly 25-40% of weight lost on GLP-1 therapy is fat-free mass, much of which is skeletal muscle [Wilding 2021]. This is comparable to other forms of caloric deficit, but the absolute lean-mass loss is larger because total loss is larger. Lower lean mass means lower resting energy expenditure, which makes regain mathematically easier in any nutritional environment.
What the trials suggest about preventing it
We are explicit here: the trial literature does not yet contain a validated, durable post-GLP-1 maintenance protocol. What follows is what the published evidence suggests is reasonable, not a protocol.
Continuation
The simplest and best-supported strategy for not regaining the weight is to not stop the drug. That is what STEP-4 and SURMOUNT-4 demonstrate directly. The cardiorenal benefits demonstrated in SELECT and FLOW [Lincoff 2023] [Perkovic 2024] reframe the question — continuation is a cardiometabolic decision, not only a weight one.
Dose tapering rather than abrupt cessation
Dose tapering is not specifically tested in the registration trials, but it is a reasonable clinical approach by analogy to other chronically dosed receptor agonists. Whether tapering meaningfully blunts regain or merely smooths the trajectory is a question for forthcoming trials — we do not have a definitive answer in 2026.
Protein floor and resistance training
A protein intake of 1.6-2.2 g/kg of reference body weight is the evidence-based threshold for muscle preservation in caloric deficit [Phillips 2016]. Two to three sessions per week of progressive resistance training preserves lean mass during weight loss in non-pharmacologic trials, and is almost certainly transferable to the GLP-1 context. The dedicated RCT data in GLP-1 cohorts is still emerging.
Intermittent or maintenance dosing
Several investigator-initiated trials are exploring lower maintenance doses or intermittent dosing schedules for weight maintenance rather than active loss. The pharmacology supports it. The outcomes data does not yet exist at registration-trial scale.
Community use of GLP-1s for weight maintenance often involves lower doses or extended dosing intervals. That use predates dedicated trial data and should be treated as informal observation rather than evidence. The trials catching up to clinical practice is one of the stories of 2026.
A tiered framework
We do not write protocols. We write frameworks that you take to a clinician. With that established:
Before starting therapy, establish the protein and resistance-training habits that will be required during weight loss and after it. Baseline labs (HbA1c, lipids, fasting insulin, eGFR). Establish a durable relationship with a clinician before, not after.
Where the indication exists and the response is good, the trial data supports continuation rather than withdrawal. The cardiorenal data reframes the cost-benefit. Reassess with metabolic labs annually.
Lower-dose or extended-interval maintenance is increasingly common in clinical practice and not yet validated in registration-trial form. If you go this route, do it with a clinician monitoring weight, metabolic markers, and the cardiometabolic indications that justified the prescription in the first place.
We will not tell you to source from research-chemical sites. We will not tell you to compound on your own. We will not tell you that stopping the drug abruptly is safe to do without a clinician conversation. Every framework here assumes a physician relationship.
References
- Wilding JPH, et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity (STEP 1). N Engl J Med. 2021;384:989-1002.
- Rubino D, et al. Effect of Continued Weekly Subcutaneous Semaglutide vs Placebo on Weight Loss Maintenance (STEP 4). JAMA. 2021;325(14):1414-1425.
- Wilding JPH, et al. Weight Regain and Cardiometabolic Effects after Withdrawal of Semaglutide: The STEP 1 Trial Extension. Diabetes Obes Metab. 2022;24(8):1553-1564.
- Jastreboff AM, et al. Tirzepatide Once Weekly for the Treatment of Obesity (SURMOUNT-1). N Engl J Med. 2022;387:205-216.
- Aronne LJ, et al. Continued Treatment with Tirzepatide for Maintenance of Weight Reduction in Adults with Obesity (SURMOUNT-4). JAMA. 2024;331(1):38-48.
- Lincoff AM, et al. Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes (SELECT). N Engl J Med. 2023;389:2221-2232.
- Perkovic V, et al. Effects of Semaglutide on Chronic Kidney Disease in Patients with Type 2 Diabetes (FLOW). N Engl J Med. 2024;391:109-121.
- Sumithran P, et al. Long-term persistence of hormonal adaptations to weight loss. N Engl J Med. 2011;365:1597-1604.
- Schwartz MW, et al. Obesity Pathogenesis: An Endocrine Society Scientific Statement. Endocr Rev. 2017;38(4):267-296.
- Phillips SM, Chevalier S, Leidy HJ. Protein requirements beyond the RDA. Appl Physiol Nutr Metab. 2016;41(5):565-572.
- Garvey WT, et al. Two-year sustained weight loss and metabolic benefits with controlled-release phentermine/topiramate in obese and overweight adults. Am J Clin Nutr. 2020;95(2):297-308.