What Happens When You Stop Semaglutide or Tirzepatide? What the Regain Studies Measured

How fast do people regain weight after stopping semaglutide or tirzepatide?

About 0.8 kg a month.¹ The figure comes from the analysis with the most participants behind it.

A team at the University of Oxford pooled 37 studies covering 9,341 adults and published the result in The BMJ on January 7, 2026.¹ Participants had been on a weight-loss medication for an average of 39 weeks and were followed for an average of 32 weeks after stopping.¹

After stopping	Monthly weight regain	Studies pooled
Semaglutide or tirzepatide	0.8 kg (95% CI 0.7 to 0.9)	10 arms, 1,776 participants
Any incretin mimetic, the paper's own category	0.5 kg (95% CI 0.4 to 0.7)	32 arms, 4,757 participants
Any weight-management medication	0.4 kg (95% CI 0.3 to 0.5)	63 arms, 6,322 participants
A behavioral or lifestyle program	0.1 kg (95% CI 0.08 to 0.13)	a separate review, 249 trials

From the Oxford meta-analysis, read at source. The first row is the one that applies to Ozempic, Wegovy, Mounjaro and Zepbound.¹

Three projections appear in the paper, narrowing as the drug group narrows. For weight-loss medications generally it is 1.7 years (95% CI 1.3 to 2.1). For all incretin mimetics, 1.6 years (1.1 to 2.1). For semaglutide and tirzepatide, the pair the paper calls the newer and more effective incretin mimetics, 1.5 years (1.0 to 1.9).¹ After a behavioral program instead, the same projection runs 3.9 years (2.8 to 4.9).¹

A second endpoint in the same paper is arguably more defensible, and it is shorter. Restricted to the randomized trials, the time until the people who stopped were no longer distinguishable from controls ran 1.4 years for weight-loss drugs in general and 1.3 years (0.8 to 1.8) for semaglutide and tirzepatide.¹

Semaglutide and tirzepatide regain faster than the broader incretin group they belong to, 0.8 kg a month against 0.5, and faster than the whole pool of weight-loss drugs at 0.4, though that pool contains them.¹ Tirzepatide acts on both the GIP and GLP-1 receptors, so the paper's category is incretin mimetics, not GLP-1 agonists alone. They also produce the largest losses, so there is more to come back.

What did the semaglutide withdrawal trial find?

A 10.6% loss became a 6.9% gain.³

STEP 4 was built to answer this question and nothing else.³ Novo Nordisk ran it at 73 sites in 10 countries between June 2018 and March 2020.³ Everyone took semaglutide for 20 weeks first, 16 weeks climbing to 2.4 mg weekly and 4 weeks holding there.³ Of the 902 who started, 803 reached the maintenance dose and were randomized 2 to 1: 535 kept taking semaglutide, 268 switched to placebo.³ Both groups kept the lifestyle program. The trial then ran another 48 weeks.³

The primary endpoint was "percent change in body weight from randomization (week 20) to week 68." The result: "With continued semaglutide, mean body weight change from week 20 to week 68 was −7.9% vs +6.9% with the switch to placebo," a difference of 14.8 percentage points.³

So the people who stopped did not merely stall. They went up, by almost 7% of body weight, in under a year, while still attending the same lifestyle appointments.³ In the arm that switched to placebo, 206 participants gained weight; in the arm that continued, 79 did.³ The trial prints those counts with percentages that do not reconcile against the stated arm sizes, so the counts appear here without denominators.³

What did the tirzepatide withdrawal trial find?

A 14.0% gain, the largest regain figure in either trial.⁴

SURMOUNT-4 used a longer run-in.⁴ Eli Lilly put 783 adults on open-label tirzepatide at the highest dose each person tolerated, 10 or 15 mg weekly, for 36 weeks.⁴ Those 670 had lost a mean of 20.9%.⁴ SURMOUNT-4 randomized those 670 one to one, half continuing and half switching to placebo, for another 52 weeks.⁴

"The mean percent weight change from week 36 to week 88 was −5.5% with tirzepatide vs 14.0% with placebo."⁴

Measured from the very beginning instead, the two groups ended 88 weeks apart in a different way: "The overall mean weight reduction from week 0 to 88 was 25.3% for tirzepatide and 9.9% for placebo."⁴

That second pair is the one to hold onto. Both numbers are still losses. The placebo group finished the study about 10% lighter than they started, having given back roughly half of a 20.9% loss.⁴ A trial can report a 14.0% regain and a 9.9% net loss at the same time, because the two use different starting points. The 14.0% is measured against body weight at week 36, and the 9.9% against body weight at week 0.

Lilly also measured how many people held onto most of what they lost. Of those who continued tirzepatide, 300 of 335 (89.5%) still held at least 80% of their lead-in loss at week 88.⁴ In the placebo group the figure was 16.6%.⁴

One wording difference between the two trials stands out, because the same journal published both. SURMOUNT-4 uses the word regain in its summary and its conclusion: "withdrawing tirzepatide led to substantial regain of lost weight." STEP 4 never applies that word to its placebo group anywhere in its abstract, methods, results or tables.³ STEP 4 reports "+6.9%" and a count of participants who gained weight, and frames its conclusion around what continuing achieved.³⁴ Both report the same finding. The framing differs.

Does everyone regain the weight after stopping a GLP-1?

Two routine-care cohorts say no, and both carry a caveat about who got counted.

Researchers at the Cleveland Clinic pulled electronic health records from their Ohio and Florida system and found 7,938 adults who started semaglutide or tirzepatide and stopped between three and twelve months later. Discontinuation was defined as a gap of more than 90 days after the previous supply ran out. Published in March 2026, the study tracked what happened over the following year.⁶

Average weight change in that year was 0.5% for patients treated for obesity and −1.3% for patients treated for type 2 diabetes. The standard deviation on that first figure was 9.9%, twenty times the mean. In the obesity group, 842 patients (55.4%) gained weight in that year; the rest held steady or kept losing.⁶ Those figures rest on the 3,810 patients, 48.0% of the cohort, who had both a weight at discontinuation and a weight a year later. Gasoyan's team notes that their analysis assumes "the existence of missing weights is independent of a patient's weight outcomes."⁶ The most reassuring dataset on this page is therefore the one most exposed to missing data.

What explains most of the gap with the trials is how the cohort was assembled. Within a year of stopping, 1,557 patients (19.6%) had restarted the same drug, and 2,794 (35.2%) had received another obesity treatment, usually a different medication. These were also short courses: average loss at the point of stopping was 8.4% for the obesity group, against the 10.6% and 20.9% run-in losses in the two trials.³⁴⁶

A second routine-care study found the mirror image of the trial headline. Researchers tracked 4,182 patients, 2,567 on semaglutide and 1,615 on tirzepatide, for six months after their last prescription, and published the result in Biology Methods and Protocols in 2026. Its authors are not independent of the platform: four of the five work for nference, the company whose platform supplied the data, and the fifth reports consulting for it.⁷ Counting regain as a rise of at least 2% in the six months after that last prescription, 33.0% of the semaglutide group and 27.6% of the tirzepatide group regained. Murugadoss and colleagues put the rest this way: "Approximately two-thirds of patients showed stable weight or continued weight loss during this period." Mean change across all patients was 0.29%, and the median was zero.⁷

Three caveats belong with that figure, two of them from the authors. The anchor was the last prescription, and they say it "does not necessarily correspond to medication discontinuation." Follow-up ran six months, not a year. In the 300-patient subset where clinician notes were checked, cessation was documented in only 119 cases, although in those 119 some 72% were in the non-regain group.⁷ Among patients who did regain, the average gain reached 4% within four months.⁷

So two studies put two-thirds on opposite sides of the same question. The STEP 1 extension found two-thirds of the lost weight regained, in trial participants who had lost 17.3% and then lost their lifestyle support.⁸ The nference cohort found roughly two-thirds of patients not regaining at all, in routine care after a smaller loss, with no confirmation that most of them had stopped.⁷ Neither is wrong. They are answers to different questions wearing the same fraction.

A separate follow-up of Lilly's Chinese trial, SURMOUNT-CN, tracked 152 participants for 26 weeks after they stopped, including 44 who had been on placebo all along. Weight rose 9.1% in the group that had taken 10 mg, 12.3% in the 15 mg group and 1.8% in the placebo group. Both were still well below where they started: −8.7% and −10.6% from trial baseline. At that point 42.6% and 53.1% of them were still at least 10% down from baseline.⁹

That study also looked for predictors, and found two. Reported compliance with lifestyle consultations was associated with less regain, and how much weight a person had lost on treatment was associated with more. Sex, age, baseline BMI, hypertension, dyslipidemia, sleep apnea and cardiovascular disease showed no association with weight change.⁹

The question under all of this is how likely you are to keep a useful amount, and the figures that answer it sit in four different studies.

Measure of keeping it	Result	Study
Held at least 80% of the loss, one year after switching to placebo	16.6%	SURMOUNT-44
Still at least 10% below baseline, 26 weeks after stopping	53.1% at 15 mg, 42.6% at 10 mg	SURMOUNT-CN9
Neither regained half the loss nor needed rescue therapy, one year after switching to placebo	30 of 90 (33%)	SURMOUNT-MAINTAIN10
Share of the loss the modeled curve leaves off in the long run	about 25%, roughly 4% to 5% of body weight	meta-regression2

That is four different definitions of success, and no single number answers the question. Read together they put the odds of holding most of a large loss low, and the odds of keeping something real reasonably high.

How many people stop semaglutide or tirzepatide in the first place?

Most of them, within a year.¹¹

Truveta records from 30 US health systems supplied the largest study of this, following 125,474 adults with overweight or obesity.¹¹ Discontinuation was defined as "the first date a patient was 60 days or more without any GLP-1 RA on hand."¹¹ By one year, 53.6% had discontinued, with a confidence interval of 53.4% to 53.9%.¹¹ By two years, 72.2% had.¹¹

Discontinued by	With type 2 diabetes	Without type 2 diabetes
One year	46.5%	64.8%
Two years	64.1%	84.4%

From the Truveta cohort.¹¹

Among the people taking these drugs for weight, not diabetes, almost two-thirds had stopped within a year and more than four in five within two. That reframes every number above this one. Regain is not an edge case here. It is what happens after an event most people who start these drugs go through.

Stopping is also often temporary. Within a year of discontinuing, 47.3% of the diabetes group and 36.3% of the group without diabetes had filled another GLP-1 prescription.¹¹ The Truveta analysis found weight pulling people back: "Weight regain of 1% from discontinuation was significantly associated with increased hazards of reinitiation of 2.3%" in the diabetes group, and 2.8% in the group without it.¹¹

Where does the "two-thirds regained" semaglutide figure come from?

Novo Nordisk published the source of that figure in 2022, and its design is not what most coverage implies.⁸

Novo Nordisk followed a subset of STEP 1 participants for a year after the trial ended and published the result in Diabetes, Obesity and Metabolism in 2022.⁸ The paper calls itself "The present observational study" and refers throughout to "the off-treatment observational extension phase." Participants had taken semaglutide 2.4 mg or placebo for 68 weeks, and the extension tracked them to week 120.⁸

The headline sentence is the paper's own: "One year after withdrawal of once-weekly subcutaneous semaglutide 2.4 mg and lifestyle intervention, participants regained two-thirds of their prior weight loss, with similar changes in cardiometabolic variables."⁸

The underlying figures: a 17.3% mean loss at week 68 became a 5.6% net loss at week 120, with 11.6 percentage points of body weight regained over the off-drug year. In the former placebo arm, a 2.0% loss became a 0.1% loss.⁸

Four things about that study change how much weight the number can carry.

The lifestyle program stopped at the same time as the drug. The paper is explicit: "At week 68, treatments (including lifestyle intervention) were discontinued," and the Discussion refers to "the absence of structured lifestyle intervention following semaglutide withdrawal."⁸ What had stopped was counseling every four weeks on a 500 kcal daily deficit and 150 minutes of weekly activity.⁸ So two-thirds describes stopping the drug and the support together. STEP 4 kept the lifestyle program running in both arms and still measured a 6.9% gain in the placebo group.³ That trial's estimate of the drug's own contribution is the between-group figure, 14.8 percentage points.³

That extension covered 17% of the trial. STEP 1 randomized 1,961 people.⁸ The extension screened 336 and enrolled 333, of whom 327 were analyzed and 312 finished.⁸ Week 120 weights measured at a study site were available for 290: 197 on semaglutide and 93 on placebo.⁸

Entry required finishing 68 weeks of treatment, so the sample tilts toward the people it worked for. The authors say so: week 0 to 68 weight loss was "slightly greater in the ExAS than in the FAS, as expected given that the ExAS only included participants who completed 68 weeks of treatment."⁸ They also concede that the way sites were chosen "could introduce an element of selection bias."⁸

The off-treatment year was not drug-free for everyone, and the paper gives two counts for it. Classified by prescription code, 2 of 228 in the semaglutide arm (0.9%) used obesity pharmacotherapy during the extension.⁸ Classified by what investigators reported, 21 (9.2%) did, including 14 (6.1%) on a GLP-1 receptor agonist, for a mean of 172 days. The paper's authors argue the small numbers did not confound the result; the two counts differ by a factor of ten.⁸

One more constraint on how the figure can be used: "All analyses in the extension were exploratory," no formal significance testing was done, and "No power calculations were performed to determine the sample size." That figure therefore carries no confidence interval, because none was calculated.⁸

What the paper found that is usually left out. Not everything reverted. Blood pressure returned to baseline in both arms, but CRP, HDL, VLDL and triglycerides "remained improved relative to the placebo arm" at week 120.⁸ HbA1c ran 5.7% at baseline, 5.2% at week 68 and 5.6% at week 120 on semaglutide.⁸

Does all the weight come back, or does the regain level off?

They do not. Two of the four meta-analyses published on this question in the last year disagree about the one number people most want.

The Oxford team fit a constant monthly rate and projected a return to the starting weight in 1.5 years for these two drugs, with a confidence interval of 1.0 to 1.9.¹ A nonlinear meta-regression in eClinicalMedicine, published March 4, 2026, fit a decelerating curve to six randomized trials covering 3,236 people and found something else. At one year, "60% of the weight lost during treatment was regained."² Past that the curve flattens, with regain "estimated to plateau at 75.3%" and a confidence interval of 68.9 to 81.6, "of the weight lost on treatment." The meta-regression authors add that "the estimated plateau is below the pre-treatment baseline weight" and that "25% of the initial weight loss may be sustained in the long term," which they put at roughly a 4% to 5% net reduction in body weight.² The same model gives regain a timescale: a rate constant of 0.0302 per week, "corresponding to a half-life of 23.0 weeks (95% CI 17.3-34.3)," meaning half of the eventual regain has happened by about six months off drug.²

	Oxford meta-analysis1	Nonlinear meta-regression2
Behind the estimate	63 intervention arms, 9,341 people	6 randomized trials, 3,236 people
Model	a constant monthly rate	a decelerating exponential curve
At one year off drug	0.8 kg a month for these two drugs	60% of the loss regained
Where it ends	back at the starting weight, in 1.5 years	a plateau at 75.3% regained, below the starting weight
Risk of bias in the pool	18 of 35 randomized trials rated high	most studies rated moderate

The weaker claim is the return to baseline, and the model is the reason, not the data. A constant monthly rate applied to a curve that is slowing down will overshoot. The slowing is visible in the primary trials. STEP 1's extension authors saw it in their own data and wrote that it "suggested a slowing of weight regain towards the end of the extension phase."⁸ They then declined to call it a plateau, because they could not distinguish one from a curve flattening as it neared the starting weight.⁸

That caution cuts both ways. No large trial has reported weight data past 52 weeks off drug, and the meta-regression authors say as much: "predictions made outside of this window are extrapolations."² So the first year is measured and the destination is modeled. Both analyses find fast, substantial regain inside that first year; they part company only past the data.

One finding from a third meta-analysis explains part of why short studies look reassuring. Within its obesity subgroup, regain ran 7.31 kg in the studies that followed people beyond 26 weeks against 2.51 kg in those that stopped earlier, a difference significant at p = 0.007.¹² A study that looks for regain for six months finds about a third of what a longer one finds.

Why do GLP-1 regain numbers differ so much between studies?

Because the studies are not measuring the same thing. Design decides the answer before any data arrive.

Design	What it measures	What it reports	What it cannot tell you
Randomized withdrawal (STEP 4, SURMOUNT-4)	A deliberate switch to placebo, blinded, with lifestyle support continuing	6.9% and 14.0% gains over about a year34	What happens to people who stop for ordinary reasons, like cost or side effects
Off-treatment extension (STEP 1 extension)	A year with no drug and no lifestyle program, observed, not assigned	two-thirds of the loss regained, a 17.3% loss becoming 5.6%8	What the drug alone contributed, since the support stopped at the same time
Real-world cohort (Cleveland Clinic)	Prescription records and recorded weights in routine care	0.5% average gain at one year6	Who actually stopped, since many restarted or switched treatments
Pooled meta-analysis (Oxford)	All of the above, converted to a monthly rate	0.8 kg a month for these two drugs1	Individual trajectories, which the average hides

That is four designs and four different-looking answers, each reconcilable with the others once the design is known.

Randomized trials give the cleanest causal answer, since the only thing that changed was the drug. They also describe a situation few people are ever in: a blinded switch to a placebo injection, with scheduled lifestyle visits continuing and no idea which group you are in.

Real-world cohorts describe the situation most people are actually in, and that is exactly why their numbers are smaller. Stopping in routine care often is not stopping.

Does anything besides weight come back after stopping semaglutide?

STEP 4 published a table of seven measures, and six of them moved the wrong way after the switch to placebo.

Measured over 48 weeks	Continued semaglutide	Switched to placebo
Waist circumference	−6.4 cm	+3.3 cm
Systolic blood pressure	+0.5 mm Hg	+4.4 mm Hg
HbA1c	−0.1%	+0.1%
Total cholesterol	+5%	+11%
LDL cholesterol	+1%	+8%
Triglycerides	−6%	+15%
HDL cholesterol	+18%	+18%

Estimated mean change from week 20 to week 68, read at source from the trial's own results table.³ The four lipid rows are percent change in blood concentration, not percent of body weight, and the HbA1c row is percentage points.

Read down the right-hand column and the pattern is plain. Waist, pressure, blood sugar and three of four lipid measures all moved in the wrong direction once the drug stopped. HDL is the exception, rising by the same 18% in both groups, with a between-group difference of zero and a p-value of .83.³

SURMOUNT-4 published only one of these. Waist circumference fell 4.3 cm in the people who continued tirzepatide and rose 7.8 cm in those who switched to placebo, a gap of 12.1 cm.⁴

Two pooled analyses add the multi-drug version, and they are not interchangeable. One compares each discontinued group against its own end-of-treatment values, which gives 5.63 kg of weight (95% CI 3.52 to 7.73), 3.81 cm of waist (2.15 to 5.47), 4.15 mm Hg of systolic pressure (1.87 to 6.44) and 0.25% of HbA1c (0.18 to 0.32) across the obesity trials, with heterogeneity above 96% on every one of those and an abstract that still calls the certainty "moderate."¹² A pre-post change inside one arm has no comparison group, so none of it can be set beside STEP 4's 14.8 percentage point difference. The other analysis does run the comparison, finding across six trials and 8,993 patients that "compared with the continued treatment group, the weight difference in the discontinuation group was" "mean difference (MD) = 17.90%, 95% confidence interval (CI) [14.11-21.69]."¹³ Directions agree across all of it. The sizes should be read as approximate.

The Oxford meta-analysis reaches the same conclusion across all 37 studies and attaches a timeline: cardiometabolic markers were projected to return to baseline within 1.4 years, slightly ahead of the 1.7 years it projected for weight across all weight-loss drugs.¹

Some of it moves faster than that. The meta-regression reviewed systolic pressure across 19 studies and found roughly 70% to 80% of the on-treatment reduction regained within twelve weeks, while HbA1c across 17 studies gave back about half its reduction by eight to twelve weeks and rarely returned fully to baseline even at a year.² Those are narrative summaries and not pooled estimates, because the data were too sparse to model.² Taken at face value, blood pressure reverts first, blood sugar reverts partly, and weight takes longest.

One measure moves the other way. Resting heart rate falls when the drug stops, by a pooled 3.22 beats per minute (95% CI 1.38 to 5.05) off drug compared with on it.¹² Our guide to resting heart rate on a GLP-1 covers why these drugs raise it in the first place.

Does keeping the weight off protect the benefits after stopping tirzepatide?

Yes. The cardiometabolic harm tracks how much weight comes back, not the act of stopping.

Eli Lilly ran a post hoc analysis of SURMOUNT-4 and published it in JAMA Internal Medicine, describing all of its analyses as exploratory.¹⁴ The analysis took the 308 participants who had lost at least 10% during the 36 weeks on tirzepatide and then switched to placebo, sorted them by how much of that loss they gave back, and compared what happened to their cardiometabolic numbers.¹⁴

Share of the lost weight regained	Waist	Systolic BP	HbA1c	Non-HDL cholesterol
Under 25%	+0.8 cm	+6.8 mm Hg	+0.14%	−0.4%
25% to under 50%	+5.4 cm	+7.3 mm Hg	+0.15%	+1.6%
50% to under 75%	+10.1 cm	+9.6 mm Hg	+0.27%	+8.4%
75% or more	+14.7 cm	+10.4 mm Hg	+0.35%	+10.8%

Mean change from week 36 to week 88. Waist circumference, systolic pressure and HbA1c each differed significantly across the four groups, at P < .001, P = .002 and P < .001.¹⁴

All four columns show the same gradient. In the group that gave back less than a quarter of what they had lost, the week 36 to week 88 changes in waist circumference, non-HDL cholesterol and fasting insulin were not significantly different from their week 36 values, meaning those three held at the level the drug had achieved.¹⁴

The same band is the reason the answer above is a qualified yes and not a clean one. Blood pressure and blood sugar got worse even in the best maintainers: systolic pressure rose 6.8 mm Hg with a confidence interval of 3.9 to 9.7, and HbA1c rose 0.14% with an interval of 0.06 to 0.22, neither crossing zero.¹⁴ So holding the weight held three of the five measures and softened the other two. It did not hold everything.

Three measures did not follow that pattern, and omitting them would overstate the result. Fasting insulin peaked in the 50% to 75% group at 46.2% and came in lower, at 26.3%, in the group that regained most. Diastolic pressure was not monotonic across the four groups. HDL cholesterol ran the other way entirely, holding its largest improvement in the group that regained least at 19.7% and declining to 13.4% in the group that regained most.¹⁴

Pulse rate stayed better than its on-drug value in all four groups, between 4.2 and 6.7 beats per minute lower, matching the pooled heart-rate finding above.¹²¹⁴

Lilly funded this analysis of its own trial. Lilly sponsored an analysis of its own trial, and six of the ten authors are Lilly employees who hold Lilly stock.¹⁴ The published statement says the sponsor was "involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication."¹⁴ The conclusion favors continued treatment, and continued treatment is the product. It is also the only analysis found for this page that separates the effect of stopping from the effect of regaining.

Is weight regain a withdrawal effect from stopping semaglutide or tirzepatide?

Two timelines argue against it, and no trial has described one.

Semaglutide clears in about five weeks and tirzepatide faster, on figures the labels state and a later section quotes.¹⁵⁵¹⁶

So the drug is gone within five to seven weeks. The weight takes well over a year, and the two published projections put the endpoint at 1.5 years or at a plateau short of the starting weight.¹² It would track the drug leaving the body if it were one. Regain does not, and the leading explanation for what it does track, appetite returning on a body that is now lighter, is a hypothesis, not a measurement. Nobody has measured it, as the section below sets out.¹⁷¹⁸

The side-effect data point the same way. In STEP 4, gastrointestinal events were reported by 41.9% of people who continued semaglutide and 26.1% of those who switched to placebo, and discontinuation for adverse events ran 2.4% against 2.2%.³ In SURMOUNT-4, the placebo group reported fewer of every gastrointestinal event the trial listed: nausea 2.7% against 8.1%, diarrhea 4.8% against 10.7%, vomiting 1.2% against 5.7%. Serious adverse events were identical at 3.0% in both arms.⁴ One line in that table runs the other way, and it belongs here: three cases of acute cholecystitis occurred in the placebo group and none on tirzepatide.⁴

A fourth meta-analysis reaches the same place from six pooled trials: after withdrawal "the overall adverse reactions of patients decreased, gastrointestinal adverse reactions decreased, and" "the incidence of cardiovascular events was not affected by drug withdrawal."¹³

Stopping made people feel better, not worse. What it did not do was hold the weight.

Three findings on this page line up in a way no trial has tested directly. The STEP 1 extension withdrew the lifestyle program along with the drug and saw two-thirds of the loss return.⁸ STEP 4 kept the program running in both arms and saw 6.9%.³ And the one study that looked for predictors found reported compliance with lifestyle consultations to be the only modifiable factor associated with less regain.⁹ No trial has compared stopping with support against stopping without it, so this is an inference from three different designs and not a measured effect.

FDA's own adverse-event database points the same way, with a caveat about how to read an absence. A search of the FDA's FAERS system through its openFDA interface, last updated July 30, 2026, returns 21,023 reports naming Wegovy and 64,947 naming Zepbound. Listing the 300 most-reported reactions for each of Wegovy, Ozempic, Zepbound and Mounjaro turns up no term coded to withdrawal, dependence or rebound. The 300th term on those four lists carries 29, 109, 31 and 74 reports, so the claim is that such reports fall below those thresholds and not that none exist. Run in reverse, 87,542 FAERS reports carry the term DRUG WITHDRAWAL SYNDROME, and none of the eight GLP-1 brand and generic names checked appears among the 300 most-named drugs in them, where that list bottoms out at 100 reports.¹⁹

One peer-reviewed group reads the same database differently. A 2023 disproportionality analysis in Pharmaceuticals applied the standardized MedDRA query for "Drug abuse, dependence and withdrawal" to FAERS reports from January 2018 through December 2022 and calculated a proportional reporting ratio of 4.05 for drug withdrawal syndrome with semaglutide, the same value it found for drug abuse.²⁰

That reading is the weaker of the two, for four reasons, three of them supplied by the paper. A proportional reporting ratio is a ratio, so 4.05 is compatible with a small absolute number of reports, exactly what the thresholds above imply. The per-term counts behind the ratio sit in a table published as an image and cannot be read from the text. The authors themselves state that spontaneous reporting cannot establish the finding and that further empirical research is needed.²⁰

The fourth reason is the line most often quoted from that paper. That paper's description of "rebound and craving phenomena after withdrawing from semaglutide" is not its characterization of the FAERS data. The full sentence calls the ratio "consistent with a range of anecdotal Redditors' observations" that highlight those phenomena.²⁰ What circulates as a pharmacovigilance finding about withdrawal is, in the paper, a comparison with forum anecdote.

What do the Wegovy and Zepbound labels say about stopping?

Not much, and one omission does more work than anything they do say.

Neither weight-management label states how long to treat. Section 2 of each covers dose escalation and maintenance doses and stops there, with no total duration and no criteria for continuing or stopping.¹⁵¹⁶ The only duration-like wording sits inside the indication, where both say the drug is for use "to reduce excess body weight and maintain weight reduction long term."¹⁵¹⁶ That phrase is the strongest textual support in either label for treating these as ongoing therapy, and it sits in the indication, not in a dosing instruction.

Neither label has a section on drug dependence, because the section is not there at all. In the Wegovy, Ozempic, Zepbound and Mounjaro labels, the table of contents runs straight from section 8 to section 10. Section 9, Drug Abuse and Dependence, is absent from all four.¹⁵⁵¹⁶ Each contents table carries the standard footnote: "Sections or subsections omitted from the full prescribing information are not listed."¹⁵

That absence is easy to over-read in either direction. Nothing in them states that there is no risk of dependence. They say nothing about dependence, and saying nothing is what the labeling format does when a section is judged not applicable. So a page claiming the FDA has ruled out dependence for these drugs is describing something no label says, and a page claiming the FDA has flagged a dependence risk is doing the same.

The regain data is on the label, and the label does not call it regain. Zepbound's section 14.1 describes SURMOUNT-4 as "an 88-week randomized withdrawal trial in which all patients received open-label ZEPBOUND during a 36-week lead-in period," "followed by randomization to either continue ZEPBOUND or switch to placebo for 52 weeks."¹⁶ Zepbound's Table 6 prints the placebo arm's week 36 to week 88 change as +14.0 against −5.5.¹⁶ Wegovy's section 14.2 does the same for the STEP 4 design, with Table 7 giving −7.9% against +6.9%.¹⁵ The word regain appears nowhere in the Wegovy label or the Ozempic label, and Zepbound reports the figure without using it.¹⁵⁵¹⁶

Zepbound's label also records what the trial lost before it started measuring. "Of the 783 patients who started ZEPBOUND at Week 0, 14.4% discontinued treatment before randomization at Week 36," with adverse events the most common reason at 6.8%. After randomization, 10.4% of the tirzepatide group and 17.9% of the placebo group discontinued study drug.¹⁶

Every discontinuation instruction in these labels is about safety, not weight. Wegovy's section 5.2 says "If pancreatitis is suspected, discontinue WEGOVY and initiate appropriate management."¹⁵ Section 5.9 says "If patients experience a sustained increase in resting heart rate, discontinue WEGOVY," and section 8.3 instructs clinicians to discontinue "at least 2 months before they plan to become pregnant."¹⁵ Nothing in any of the four tells a prescriber when treatment is finished.

What happened when a trial lowered the tirzepatide dose instead of stopping?

One trial ran that comparison, and the two arms finished 6.7 percentage points apart.

SURMOUNT-MAINTAIN put 441 adults through 60 weeks of tirzepatide at the maximum dose they tolerated, then randomized 378 of them three ways for another 52 weeks: continue, drop to 5 mg, or switch to placebo.¹⁰

At week 112	Weight change from baseline	Regained at least half their loss and received rescue therapy
Continued at 10 or 15 mg	−21.9%	11 of 138 (8%)
Reduced to 5 mg	−16.6%	35 of 142 (25%)
Switched to placebo	−9.9%	60 of 90 (67%)

From the published trial, read via the abstract and registry record.¹⁰

The third column is not a clean regain count, and that matters for the whole comparison. SURMOUNT-MAINTAIN allowed rescue tirzepatide for anyone whose regain passed 50%, and its published sentence ties those three percentages to receipt of that rescue therapy and not to regain alone.¹⁰ So two-thirds of the arm meant to represent stopping went back onto the drug. The −9.9% beside it is therefore not what a year of stopping produces; it is what a year of stopping produces when most of the group is rescued partway through. The full text could not be read for this page, so the exact composite definition stays unresolved.

The 5 mg arm finished 6.7 percentage points of body weight below the arm that stopped.¹⁰ Three limits come with that finding: 5 mg is an ordinary rung of the tirzepatide ladder and not a reduced or improvised dose, semaglutide has no equivalent trial, so none of it transfers to Ozempic or Wegovy, and the rescue therapy makes that placebo arm a soft comparator. Our GLP-1 microdosing guide goes through what that trial does and does not license.

Is regain after a GLP-1 different from regain after dieting?

Faster, and by a wide margin: 0.8 kg a month against 0.1. Only one analysis has run the comparison directly.

After a behavioral or lifestyle program ends, the regain rate was 0.1 kg a month. After a weight-loss medication stops, it was 0.4 kg a month across all drugs and 0.8 kg for semaglutide and tirzepatide. The comparison the Oxford team ran put medications ahead by 0.3 kg a month, with a confidence interval of 0.22 to 0.34.¹

One thing about that comparison belongs with it. That behavioral figure is not pooled from the 37 studies in this review. The figure comes from a separate earlier review of 249 trials and 43,151 participants, reanalyzed here with follow-up capped at two years, which the paper states and does not hide.¹ So the contrast holds between two bodies of literature, not between two arms of one trial.

Weight regain is therefore not a peculiarity of these drugs. It follows weight loss by almost any route. What the drugs change is the slope.

The study's lead author, Sam West, said in the university's announcement that the pattern "reflects the nature of obesity as a chronic, relapsing condition" and is not "a failing of the medicines." His joint senior author, Susan Jebb, put the same point as "Obesity is a chronic, relapsing condition, not a short-term problem with a quick fix."²¹ Both of them, along with a third author, received or are partly funded by a Novo Nordisk Foundation grant, disclosed by the paper and belonging beside the quotation.¹

What happens in the first weeks after your last semaglutide dose?

Both drugs clear on a schedule. What happens to appetite has not been measured.

Start with the part that is known, because it is all on the labels. The Wegovy label states that "semaglutide will be present in the circulation for about 5 to 7 weeks after the last dose of 2.4 mg" and refers to "the long half-life of WEGOVY of approximately 1 week."¹⁵ The Ozempic label instructs clinicians to stop the drug "at least 2 months before a planned pregnancy due to the long washout period for semaglutide."⁵ The Mounjaro label gives tirzepatide "an elimination half-life of approximately 5 days."¹⁶

Now the part that is not known. It has not been measured: no study of appetite hormones, gastric emptying or energy expenditure after one of these drugs was stopped turns up in the reviews that would cite one, or in a search of the indexing databases. A 2026 review in Diabetes, Obesity and Metabolism sets out the leading explanation, "restoration of appetite following withdrawal of pharmacological GLP-1 signalling" alongside "compensatory increases in orexigenic hormones such as ghrelin," then states plainly that "evidence for a direct role in post-cessation weight regain remains limited."¹⁷

REST, the trial built to settle it, has not reported. REST, whose protocol appeared in PLOS ONE in 2026, will compare a gradual dose reduction of 25% every four weeks against immediate cessation, measuring body weight, 24-hour blood pressure and fasting ghrelin 16 weeks after withdrawal. Recruitment opened March 1, 2026, data collection is expected to end in July 2028, and analysis is expected in 2030.¹⁸

Weight moves on a slower clock than the drug, but not as slow as a monthly average suggests. At the Oxford rate, a 100 kg person regains under a kilogram in the first month, below what a bathroom scale distinguishes from normal variation.¹ Across the pooled studies, though, the share of lost weight already regained by about twelve weeks ran anywhere from 10% to 80%.² And among the routine-care patients who did regain, the average gain reached 4% within four months.⁷ The average is slow. The range is not.

What does the regain data mean for semaglutide or tirzepatide bought online?

Less than it looks, because almost every number above describes a planned stop.

Every trial on this page measured people who stopped on a schedule, inside a health system, with their weight recorded. The real-world cohorts did catch unplanned stops, but they caught them as prescription gaps in insured patients with a prescriber, and they could not confirm who had actually stopped.⁶⁷ A gray-market supply produces something neither design covers: an unplanned stop, at an unplanned time, at a dose nobody verified. No study here modeled that, and none can tell you what a particular interruption will do.

The Cleveland Clinic authors list drug shortages during their study window as a possible contributor to the discontinuations they counted, and they note that compounded products were not captured in their records at all.⁶ So even the real-world dataset is partly a record of people who stopped because supply ran out.

A second detail concerns what the regain data say about dose history. Regain scales with how much was lost, and how much was lost depends on the dose reached and the time spent there. STEP 1 ran 68 weeks to a 17.3% loss. SURMOUNT-4 ran 36 weeks to 20.9%. The Cleveland Clinic cohort averaged 8.4% over three to twelve months.⁴⁶⁸ A vial of unverified concentration gives no reliable version of either input, so the published regain curves cannot be mapped onto it.

What would actually settle the regain question?

Seven gaps, and only one of them has a study running.

The study nobody has run	Why the gap matters
A trial that separates the drug from the support	The two-thirds figure comes from a study where structured lifestyle support ended with the drug.8 STEP 4 kept the support but used a blinded placebo injection, not how anyone stops in real life.3 Neither design answers what happens when someone stops the drug and keeps the program.
Follow-up past one year	In the Oxford pool, mean follow-up after stopping was 32 weeks and the longest was 104.1 Both published endpoints, a return to baseline and a plateau below it, sit beyond where the data stop.12
Individual trajectories instead of means	The Cleveland Clinic cohort had a mean one-year change of 0.5% with a standard deviation of 9.9%.6 An average that small on a spread that wide describes very few individuals, and no study has published the full distribution behind it.
A semaglutide dose-reduction trial	Tirzepatide has one.10 Semaglutide has none, so the alternative to stopping that has been tested exists for only one of the two drugs.10
A study of what a restart achieves	Nearly a fifth of the Cleveland Clinic cohort went back on the same drug within a year, and none of the studies cited here reports a separate weight outcome for the people who restarted.6
A head-to-head trial of a wind-down against stopping	SURMOUNT-MAINTAIN tested holding a lower dose, which is not a wind-down to zero.10 REST is running the wind-down comparison, with results expected around 2030.18
Any measurement of the physiology	Nobody has published appetite hormones, gastric emptying or energy expenditure measured after one of these drugs was stopped, so the mechanism is a hypothesis stated as one.1817

Common questions

Will I gain all the weight back if I stop Ozempic or Wegovy?

Do I have to take semaglutide or tirzepatide forever?

Is there a withdrawal syndrome from stopping a GLP-1?

Why do some studies say you regain two-thirds and others say almost nothing?

Does blood pressure, blood sugar or cholesterol go back up after stopping?

Is it better to reduce the dose than to stop?

Should I taper off instead of stopping suddenly?

Is regain worse after tirzepatide than after semaglutide?

Does anything predict who regains?

How long does semaglutide stay in your system after the last dose?

Do most people even stay on semaglutide or tirzepatide?

Does the FDA say these drugs cause dependence?

Is weight regain unique to these drugs?

Sources

1. West S, Scragg J, Aveyard P, Oke JL, Willis L, Haffner SJP, Knight H, Wang D, Morrow S, Heath L, Jebb SA, Koutoukidis DA. Weight regain after cessation of medication for weight management: systematic review and meta-analysis. BMJ 2026;392:e085304. Published January 7, 2026. doi:10.1136/bmj-2025-085304. PMID 41500720. PROSPERO CRD42024532069. Full text read at source. Objective: to "quantify and compare the rate of weight regain after cessation of weight management medications" in adults with overweight or obesity. Inclusion: randomized controlled trials, non-randomized trials and observational studies, each including a weight management medication for at least 8 weeks with at least 4 weeks of follow-up after stopping, against non-drug comparators or placebo. Included 37 studies, 63 intervention arms, 9,341 participants. Mean treatment duration 39 weeks, range 11 to 176. Mean follow-up after cessation 32 weeks, range 4 to 104. Pooled monthly weight regain, with the paper's own category names: any WMM 0.4 kg (95% CI 0.3 to 0.5) across 63 arms and 6,322 participants; all incretin mimetics 0.5 kg (95% CI 0.4 to 0.7) across 32 arms and 4,757 participants; newer and more effective incretin mimetics, meaning semaglutide and tirzepatide, 0.8 kg (95% CI 0.7 to 0.9) across 10 arms and 1,776 participants. The 9,341 figure is the whole review's population and not the denominator of any one estimate. The behavioral comparator of 0.1 kg a month (95% CI 0.08 to 0.13) is not pooled from these 37 studies: it comes from a previous prospectively registered review of 249 randomized trials and 43,151 participants, whose search "was not updated," reanalyzed here with follow-up limited to two years. Regain after medication was faster than after a behavioral program by 0.3 kg a month (95% CI 0.22 to 0.34). Projected return to baseline weight, by category: any WMM 1.7 years (95% CI 1.3 to 2.1); all incretin mimetics 1.6 years (1.1 to 2.1); newer incretin mimetics 1.5 years (1.0 to 1.9); and for behavioral programs 3.9 years (2.8 to 4.9). A second endpoint, time to no difference from control in the randomized trials, gives 1.4 years (95% CI 0.9 to 1.8) for any WMM, 1.1 years (0.6 to 1.5) for all incretin mimetics and 1.3 years (0.8 to 1.8) for the newer pair. Cardiometabolic markers were projected to return to baseline within 1.4 years. An earlier draft of this page stated that no drug-specific projection existed in the paper and attributed the 1.5-year figure to the university's press release; that was wrong, and the figure is in the results with a confidence interval. Risk of bias: of the 35 randomized trials, 12 low across all domains, 5 with some concerns and 18 high; of the 2 non-randomized studies on ROBINS-I, one serious and one critical. Only 28 of the 35 randomized trials had a control group during both intervention and follow-up, with the other 7 analyzed as single-arm. Funnel plots and Egger's test showed little evidence of publication bias. A sensitivity analysis restricted to low risk-of-bias studies gave 0.65 kg a month (95% CI 0.52 to 0.74) for all medications, higher than the main estimate and not lower. GRADE certainty was moderate for the rate of weight regain. The conclusion is quoted in parts because the sentence exceeds the retrieval tool's quote limit: "cessation of WMM is followed by rapid weight regain" and "reversal of beneficial effects on cardiometabolic markers." One internal inconsistency recorded and not used on this page: the mixed-model randomized estimate appears as 0.3 kg (0.2 to 0.4) in one section and 0.3 kg (0.3 to 0.4) in another. Funded in part by the NIHR Oxford Biomedical Research Centre, grant IS-BRC-1215-20008, with an NIHR advanced fellowship, NIHR senior investigator and applied research collaboration roles, a Wellcome Trust grant 225494/Z/22/Z and a China Scholarship Council visiting scholarship; the funder had no role in the study. Competing interests: four authors were investigators in publicly funded trials where interventions were donated by Nestlé Health Science, Oviva or Second Nature, outside the submitted work, and three authors, including the lead and one senior author, received or are partially funded by a Novo Nordisk Foundation grant.
2. Budini B, Luo S, Tam M, Stead I, Lee A, Akrami A, Vidal-Puig A, Park A. Trajectory of weight regain after cessation of GLP-1 receptor agonists: a systematic review and nonlinear meta-regression. eClinicalMedicine 2026;93:103796. Published online March 4, 2026. doi:10.1016/j.eclinm.2026.103796. PMID 41938838. PROSPERO CRD420250631751. Full text of the open-access version of record read at source through an open-access full-text service; the publisher's own site refused every attempt. Review scope: "A total of 48 relevant studies were included," comprising 36 randomized and 12 non-randomized studies, from 6,574 initial records, with the search running to August 28, 2025. The model was fit on a narrower set: "Six RCTs with 3236 participants were included in the nonlinear meta-regression," those being STEP 1 (n=228), STEP 4 (n=268), STEP 10 (n=138), the three SURMOUNT-1 dose arms (n=247, 262, 253), SURMOUNT-4 (n=335) and SCALE Obesity (liraglutide, n=1505). Entry to the meta-regression required a randomized design, at least 100 participants in the treatment arm, at least 3 kg of on-treatment loss, and either two post-cessation data points or one at or after 12 weeks. SCALE Diabetes and a beinaglutide trial were excluded as outliers. Model: "a mixed-effect exponential recovery model with random effects applied to the rate constant." Results: "At 1 year post-cessation, 60% of the weight lost during treatment was regained"; "Beyond 52 weeks, weight trajectories were extrapolated, with weight regain estimated to plateau at 75.3%" and "(95% CI 68.9-81.6) of the weight lost on treatment"; "The rate constant was 0.0302 per week (95% CI 0.0202-0.0401)," "corresponding to a half-life of 23.0 weeks (95% CI 17.3-34.3)." Across studies, roughly 10% to 80% of treatment-induced loss had been regained by about 12 weeks. On the endpoint: "Notably, the estimated plateau is below the pre-treatment baseline weight," "This residual weight loss corresponds approximately to a 4-5% reduction in body weight," and "it also suggests that 25% of the initial weight loss may be sustained in the long term." In one study post-cessation weight exceeded the pre-treatment baseline, and the authors attribute that to participants switching to insulin. Interpretation: "GLP-1RA cessation is associated with a predictable and decelerating pattern of weight regain, which appears to plateau below pre-treatment levels, suggesting that partial weight-loss benefit may persist long-term but is substantially attenuated." Risk of bias was assessed with "the Cochrane RoB 2 tool for RCTs and the Cochrane ROBINS-I tool for non-randomised studies," and "Most studies were assessed to have a moderate risk of bias." The GRADE rating and the per-category risk-of-bias counts sit in an online appendix that could not be retrieved. Limitations named by the authors: "the trial data used to fit our model only extend to 52 weeks after cessation," "No large-scale trials have reported weight regain data beyond this point," "thus predictions made outside of this window are extrapolations," and "Assessment of publication bias was not feasible as the meta-regression included only six studies." No drug-specific regain analysis was run; the paper states only that "Weight regain trajectories appeared broadly similar across trials of liraglutide, semaglutide and tirzepatide," without figures. No lifestyle-support subgroup analysis was run either: the authors write that they "were unable to run sensitivity analyses to determine the extent to which other variables such as" "concomitant medications, weight maintenance strategies or comorbidities affected the trajectory of weight regain," and that "reporting and implementation of post-cessation weight maintenance strategies was inconsistent." Secondary outcomes were narrative and not pooled, because data were insufficient for modeling: across 17 studies reporting HbA1c, about half of the on-treatment reduction was regained by 8 to 12 weeks and HbA1c rarely returned fully to baseline even at one year; across 19 studies reporting systolic pressure, roughly 70% to 80% of the reduction was regained within 12 weeks. Funding: the abstract states "None." and the methods state "There was no funding source for this study." Declaration of interests: the senior author "has received payments from Eli Lilly and Novo Nordisk for lecturers/presentations and meetings" and "has also served as Principal Investigator for GLP-1RA trials conducted by Eli Lilly and Novo Nordisk," and a co-author "currently sits on the strategic science advisory committee of Hisun Pharmaceutical." A June 2025 preprint of this work exists and carries different numbers, including a 75.6% plateau, a 44-study review and a search ending December 2024. None of the preprint figures is used here; every figure above is from the peer-reviewed version.
3. Rubino D, Abrahamsson N, Davies M, et al. Effect of continued weekly subcutaneous semaglutide vs placebo on weight loss maintenance in adults with overweight or obesity: the STEP 4 randomized clinical trial. JAMA 2021;325(14):1414-1425. doi:10.1001/jama.2021.3224. PMID 33755728. NCT03548987. Key Points, abstract, introduction, methods, results and Tables 1 to 3 read at source. The Discussion and the Article Information block are not present in the served page text and were not read, so no claim here rests on either. Design: "Randomized, double-blind, 68-week phase 3a withdrawal study conducted at 73 sites in 10 countries" from June 2018 to March 2020, in adults with BMI of at least 30, or at least 27 with one weight-related comorbidity, and without diabetes. "A total of 902 participants received once-weekly subcutaneous semaglutide during run-in." After 20 weeks, comprising 16 weeks of dose escalation and 4 weeks at the maintenance dose, 803 (89.0%) who reached 2.4 mg weekly were randomized 2:1 to 48 further weeks of semaglutide (n=535) or placebo (n=268), "plus lifestyle intervention in both groups." Primary endpoint: "The primary end point was percent change in body weight from randomization (week 20) to week 68." Result: "With continued semaglutide, mean body weight change from week 20 to week 68 was −7.9% vs +6.9% with the switch to placebo" and "(difference, −14.8 [95% CI, −16.0 to −13.5] percentage points; P < .001)." Run-in loss 10.6%. Table 2, estimated mean change from week 20 to week 68, continued semaglutide against switched to placebo: waist circumference −6.4 cm (−7.1 to −5.7) against 3.3 cm (2.3 to 4.3); systolic blood pressure 0.5 mm Hg (−0.6 to 1.6) against 4.4 mm Hg (2.9 to 6.0); HbA1c −0.1% against 0.1%; total cholesterol 5% (4 to 6) against 11% (10 to 13); LDL cholesterol 1% (−1 to 3) against 8% (5 to 10); HDL cholesterol 18% (17 to 20) against 18% (15 to 21), between-group difference 0 (−2 to 3), P = .83; triglycerides −6% (−9 to −2) against 15% (7 to 23). Safety: "similar proportions discontinued treatment because of adverse events with continued semaglutide (2.4%) and placebo (2.2%)" and "Gastrointestinal events were reported in 41.9% of participants who continued subcutaneous semaglutide vs 26.1% with placebo." Serious adverse events in the placebo arm 15 (5.6%). Two extraction cautions, both handled: the table row for participants who gained weight prints "206 (82.4)" against a stated placebo n of 268, and 206 of 268 is 76.9%, so that row's denominator appears to be a different analysis set; this page prints the count of 206 and not the percentage. The VLDL cholesterol row returned values identical to the triglycerides row on two independent reads, most likely a text-extraction artifact, so no VLDL figure from this trial appears on this page. One further note on framing: the word regain is not applied to the placebo group anywhere in the sections that were read, and its only appearance is a general statement in the introduction, "with weight regain often following weight loss." Because the Discussion could not be read, that observation is limited to the sections named above.
4. Aronne LJ, Sattar N, Horn DB, Bays HE, Wharton S, Lin WY, Ahmad NN, Zhang S, Liao R, Bunck MC, Jouravskaya I, Murphy MA; SURMOUNT-4 Investigators. Continued treatment with tirzepatide for maintenance of weight reduction in adults with obesity: the SURMOUNT-4 randomized clinical trial. JAMA 2024;331(1):38-48. Published online December 11, 2023. doi:10.1001/jama.2023.24945. PMID 38078870. NCT04660643. Key Points, abstract, introduction, methods, results and Tables 1 to 3 read at source; the Discussion and Article Information block are absent from the served page text and were not read. This replaces an earlier citation of the same trial that rested on an institutional repository record, and every figure in that record checked out. Design: 783 adults entered a 36-week open-label lead-in and "received once-weekly subcutaneous maximum tolerated dose (10 or 15 mg) of tirzepatide for 36 weeks." Of those, 670 completed and "experienced a mean weight reduction of 20.9%," then "a total of 670 participants were randomized (1:1)," 335 continuing and 335 switching to placebo for a "52-week, double-blind, placebo-controlled period," weeks 36 to 88. Primary endpoint: "The primary end point was the mean percent change in weight from week 36 (randomization) to week 88." Result: "The mean percent weight change from week 36 to week 88 was −5.5% with tirzepatide vs 14.0% with placebo," difference −19.4% (95% CI −21.2% to −17.7%), P < .001. The paper prints the placebo figure without a plus sign even though it is a gain; this page shows it as printed in the table context and as a gain in prose. From the start of the study: "The overall mean weight reduction from week 0 to 88 was 25.3% for tirzepatide and 9.9% for placebo." Key secondary endpoint: 300 participants (89.5%) on tirzepatide maintained at least 80% of the lead-in weight loss at week 88, "compared with 16.6% receiving placebo," P < .001. Waist circumference from week 36 to 88: −4.3 cm (−5.3 to −3.2) against 7.8 cm (6.9 to 8.8), difference −12.1 cm, P < .001. Conclusion: "withdrawing tirzepatide led to substantial regain of lost weight," whereas continued treatment "maintained and augmented initial weight reduction." Safety over the double-blind and follow-up period, tirzepatide against placebo: adverse events leading to discontinuation 6 (1.8%) against 3 (0.9%); at least one adverse event 202 (60.3%) against 187 (55.8%); serious adverse events 10 (3.0%) against 10 (3.0%); nausea 8.1% against 2.7%; diarrhea 10.7% against 4.8%; vomiting 5.7% against 1.2%; acute cholecystitis 0 against 3 (0.9%); three adjudicated major adverse cardiovascular events in the tirzepatide group and none in the placebo group. The paper does not compute a share-of-loss-regained percentage, and none is attributed to it here. Blood pressure, lipid and HbA1c results for week 88 are not in Table 2, enumerated row by row to confirm this; the methods point to an online supplement that could not be reached, so no such figure from this trial appears on this page. Limitations surfaced in the readable text: the lead-in was open-label, 71% of participants were women and 80% were White. Funding read via a university repository record that reproduces the paper's own statement: "Funding/Support: This study was sponsored by Eli Lilly and Company." The conflict-of-interest block could not be read on any route; what is verifiable at source is that the affiliation list places five authors at Eli Lilly and Company in Indianapolis and one at Eli Lilly and Company in Moscow. A metadata registry lists a National Center for Advancing Translational Sciences award against this DOI, an institutional translational-science award and not the trial funder; it is recorded here only so it is not mistaken for one.
5. OZEMPIC (semaglutide) injection prescribing information, NDA 209637, revised 05/2026. US Food and Drug Administration. accessdata.fda.gov. Read at source on the 05/2026 revision. Section 12.3: "With an elimination half-life of approximately 1 week" and "semaglutide will be present in the circulation for about 5 weeks after the last dose." Indication: "OZEMPIC is indicated:" "as an adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes mellitus," plus indications for cardiovascular event reduction and kidney outcomes. This is not a weight-management indication. Section 9, Drug Abuse and Dependence, is absent; the numbering runs from 8 to 10. Section 8.3: "Discontinue OZEMPIC in women at least 2 months before a planned pregnancy due to the long washout period for semaglutide." The word regain does not appear in the label. Cited on this page for the elimination figures, the pregnancy washout statement, the absent section 9 and the indication wording.
6. Gasoyan H, Schulte R, Boyer CB, Casacchia NJ, Butsch WS, Le P, Aminian A, Griebeler ML, Burguera B, Rothberg MB. Obesity treatments and weight changes in clinical practice after discontinuation of semaglutide or tirzepatide. Diabetes, Obesity and Metabolism 2026;28(6):4795-4805. Published online March 12, 2026. doi:10.1111/dom.70660. PMID 41816857. Full text read at source through the open-access deposit; the publisher's own page refused on two attempts. Design: "Data for this retrospective cohort study were obtained from the Cleveland Clinic electronic health record (EHR)," covering sites in Ohio and Florida. Cohort: "This study included 7938 patients who filled an initial prescription for injectable semaglutide" or tirzepatide, 6,184 and 1,754 respectively, who initiated between January 2021 and December 2023 and "discontinued the medication between 3 and 12 months after treatment initiation." Discontinuation was "defined as a gap of greater than 90 days between exhaustion of the previous supply" and the next. Mean percentage weight change from baseline to discontinuation: "when treating obesity was −8.4% (7.6%)," 95% CI −8.7% to −8.1%, and "when treating T2D was −4.4% (6.1%)," 95% CI −4.7% to −4.2%. Mean percentage weight change in the year after discontinuation: "when treating obesity was 0.5% (9.9%)," 95% CI 0.0% to 1.0%, and "when treating T2D was −1.3% (7.8%)," 95% CI −1.6% to −1.0%. In the obesity group, "842 (55.4%) gained weight in the year after discontinuation"; in the type 2 diabetes group, 1,000 (43.6%) did. Treatment after stopping: "Overall, 1557 patients (19.6% of the study cohort) reinitiated their index medication within 1 year" and "2794 patients (35.2% of the study cohort) received an alternative obesity treatment," comprising another medication 27.4%, a lifestyle modification visit 13.7% and metabolic or bariatric surgery 0.6%. The paper states that "the obesity treatment categories were not mutually exclusive, patients could have contributed to each of these events," so the two headline shares cannot be summed and no total appears on this page. The post-discontinuation weight analysis included "n = 3810, 48.0% of the study cohort," those with both a discontinuation weight and a one-year weight; the paper gives no group-specific denominators. Conclusion: "The average weight change 1 year after index medication discontinuation was relatively small" and "there was considerable individual-level variability." Limitations named by the authors: a single integrated health system in two states, limiting generalizability; "Shortages of semaglutide and tirzepatide during the study period could have contributed"; "Some of the observed weight reduction may be associated with other interventions"; compounded medications were not captured; "This study did not capture specific reasons for discontinuation"; and the analysis "assumes that the existence of missing weights is independent of a patient's weight outcomes." The paper's introduction cites STEP 1 and SURMOUNT-4 but its discussion does not set out why its result differs from theirs. Funding came from a Cleveland Clinic Mandel accelerator grant, number MAG0030. Disclosures include one author "currently employed by Novo Nordisk A/S," another reporting research funding from Novo Nordisk A/S, one reporting consulting fees from Janssen Pharmaceuticals, one reporting research grants from Amgen, Ethicon and Medtronic, one reporting National Cancer Institute grant funding and one reporting consulting fees from the Blue Cross Blue Shield Association.
7. Murugadoss K, Varma G, Venkatakrishnan AJ, Gibson CM, Soundararajan V. Weight trajectories after last tirzepatide or semaglutide prescription across a federated health network. Biology Methods and Protocols 2026;11(1):bpag020. doi:10.1093/biomethods/bpag020. PMID 42163990. Read at source. Peer-reviewed article with received, revised and accepted dates printed, not a preprint. Data source: the nference nSights Federated Clinical Analytics Platform, de-identified electronic health records from US tertiary centers tied to academic medical centers. Four of the five authors are identified with nference or nference Labs, the company whose platform supplied the data, and the fifth reports consulting with nference. The page states this where the study is cited. Cohort: 4,182 patients, 2,567 on semaglutide and 1,615 on tirzepatide, plus a 300-patient single-center subset whose clinician notes were examined. Follow-up window ran from 365 days before to 180 days after the index date. The index was set as "the date of their last GLP-1RA prescription," specifically the prescription followed by at least six months without another prescription for the same medication, which means a patient who switched between semaglutide and tirzepatide can be counted as having had a last prescription. The authors state that this anchor "does not necessarily correspond to medication discontinuation," and clinician-documented cessation, defined as "Evidence of treatment cessation documented in notes within ±90 days of the last prescription date," was found by automated extraction in 119 of the 300-patient subset, or 40%. Regain was defined as an increase of at least 2% in the six months following the last prescription, measured from the index weight and not from a pre-treatment baseline. Results: 33.0% of the semaglutide group and 27.6% of the tirzepatide group regained; "Approximately two-thirds of patients showed stable weight or continued weight loss during this period"; the results text puts it as "more than two-thirds of patients fell into the non-regain group" in both cohorts; mean patient-level change 0.29% and median 0. Among patients who did regain, average gain reached 4% within four months. In the 119 with documented cessation, 72% (86 of 119) were in the non-regain group. Pre-index weight loss was modest against the trials, roughly 7% for semaglutide and 11% for tirzepatide.
8. Wilding JPH, Batterham RL, Davies M, Van Gaal LF, Kandler K, Konakli K, Lingvay I, McGowan BM, Kalayci Oral T, Rosenstock J, Wadden TA, Wharton S, Yokote K, Kushner RF; STEP 1 Study Group. Weight regain and cardiometabolic effects after withdrawal of semaglutide: the STEP 1 trial extension. Diabetes, Obesity and Metabolism 2022;24(8):1553-1564. Published online May 19, 2022. doi:10.1111/dom.14725. PMID 35441470. NCT03548935. Full text read at source two ways, through an open-access full-text service and through the CC BY version of record deposited in a university repository; the publisher's own page refused on three attempts and the article's own PMC page served a robot check. Volume, issue and page numbers come from three agreeing metadata registries rather than from the article face, because the version of record available carries early-view pagination. Design, in the paper's words: "The present observational study examined changes in body weight and cardiometabolic risk factors over 52 weeks," referred to throughout as "the off-treatment observational extension phase." Parent trial: "STEP 1 (NCT03548935) was a randomized, double-blind, placebo-controlled trial conducted at 129 sites across 16 countries," with participants "randomized to 68 weeks of treatment with once-weekly s.c. semaglutide 2.4 mg (n = 1306) or placebo (n = 655) (2:1), plus lifestyle intervention." The extension ran weeks 68 to 120, with visits at weeks 68, 75, 80, 104 and 120, and was "offered in five selected countries (Canada, Germany, Japan, the UK and the United States)." Sample: "336 participants from the main phase were screened for the extension," "333 entered the extension, including 232 participants who received semaglutide during the main phase and 101 who received placebo," "327 participants (98.2%) were included in the ExAS," and "The majority of participants completed the extension (93.7%; n = 312/333)." Week 120 weights were "available from site visits for 290 participants," 197 on semaglutide and 93 on placebo; a further 22 completions were self-reported by telephone or email and were not evaluated in the analyses. The 333 who entered are 17.0% of the 1,961 randomized, a figure the paper does not state and which is arithmetic from its own numbers. Results: "observed mean weight loss was 17.3% (standard deviation [SD]: 9.3%) with semaglutide versus 2.0% (SD: 6.1%) with placebo" at week 68; "Participants regained a mean of 11.6 percentage points (SD: 7.7) of body weight in the semaglutide arm versus 1.9 percentage points (SD: 4.8) in the placebo arm"; and "The net mean body weight loss over the full duration of the main treatment phase and off-treatment extension phase (from week 0 to week 120) was 5.6% (SD: 8.9%) in the semaglutide arm versus 0.1% (SD: 5.8%) in the placebo arm." Abstract conclusion: "One year after withdrawal of once-weekly subcutaneous semaglutide 2.4 mg and lifestyle intervention, participants regained two-thirds of their prior weight loss, with similar changes in cardiometabolic variables. Findings confirm the chronicity of obesity and suggest ongoing treatment is required to maintain improvements in weight and health." Discussion conclusion: withdrawal "led to most of the weight loss being regained within 1 year, and a similar change in some cardiometabolic variables back to baseline, reinforcing the need for continued treatment to maintain weight loss and cardiometabolic benefits." Lifestyle support stopped with the drug: "At week 68, treatments (including lifestyle intervention) were discontinued," "participants were withdrawn from treatment (including lifestyle intervention)," and the Discussion refers to "the absence of structured lifestyle intervention following semaglutide withdrawal." The main-phase program had been "counselling every 4 weeks on diet (500 kcal deficit per day relative to total estimated energy expenditure at randomization)" and "physical activity (150 minutes per week)." During the extension, "No treatments were actively administered or prohibited," and "Participation in lifestyle interventions that might impact weight were not recorded." Obesity pharmacotherapy during the extension is reported two ways: by prescription classification, "obesity pharmacotherapies were initiated or ongoing in the extension in two (0.9%) and three (3.0%) participants in the semaglutide and placebo arms, respectively"; by investigator report, "there were 21 (9.2%) obesity pharmacotherapy users in the semaglutide arm and six (6.1%) in the placebo arm during the extension, including GLP-1 receptor agonist use in 14 (6.1%) and four (4.0%)," with a mean duration of 172.1 and 264.7 days. No participant became pregnant or had bariatric surgery during the extension. Cardiometabolic results: "Cardiometabolic improvements seen from week 0 to week 68 with semaglutide reverted towards baseline at week 120 for most variables," with named exceptions, since "CRP and lipids increased from week 68 to week 120 in the semaglutide arm," "but remained improved relative to the placebo arm for CRP, HDL cholesterol, VLDL cholesterol and triglycerides at week 120." On blood pressure: "After treatment withdrawal, mean SBP and DBP increased in both treatment arms, reverting to baseline levels by week 120." Systolic pressure on semaglutide ran 121 at week 68 and 131 at week 120, diastolic 78 and 82. HbA1c on semaglutide ran 5.7% at week 0, 5.2% at week 68 and 5.6% at week 120. On the trajectory: "Mean body weight data suggested a slowing of weight regain towards the end of the extension phase in participants withdrawn from semaglutide," and "Whether the slowing in regain in the off-treatment STEP 1 extension represents a plateauing of mean body weight below baseline levels, or simply a slowing of weight regain as body weight approached baseline, cannot be determined without longer follow-up." Statistical status: "All analyses in the extension were exploratory," "In addition, all analyses were exploratory and thus no formal tests of statistical significance were performed," and "No power calculations were performed to determine the sample size." There is therefore no confidence interval or p-value attached to any regain figure, and none is printed here. Limitations, from the Discussion: "The key limitations of the extension were the relatively small sample size compared with the main STEP 1 trial population and the selection of sites based on those with the highest recruitment in the STEP 1 main phase, which could introduce an element of selection bias"; "Weight losses with semaglutide versus placebo from week 0 to week 68 were slightly greater in the ExAS than in the FAS, as expected given that the ExAS only included participants who completed 68 weeks of treatment"; and "Another limitation was the potential for obesity pharmacotherapy use during the extension," to which the authors answer that "few participants used obesity pharmacotherapy so our findings were not confounded." Two variables that this page does not attribute to this paper, because two separate term searches confirm they appear nowhere in it: waist circumference and fasting plasma glucose. Free fatty acids were measured but deliberately not reported, "because of different fasting conditions in the main and extension phases." Funding: "The STEP 1 trial was funded by Novo Nordisk A/S." Three of the fourteen authors are Novo Nordisk employees, one of whom also holds company stock, and the remaining eleven each declare a Novo Nordisk research, consulting, advisory or speaking relationship. A metadata registry lists a National Institute for Health Research award against this article which does not appear in the funding statement that was read, so this page does not describe the study as NIHR-funded.
9. Chen C, Ying Z, Tang Q, Zhao L, Lu Y, Fan X, Xu L, Si S, Li Y, Xu J, Wang Y, Dong Y, Li X, Chen Y. Weight loss maintenance after tirzepatide cessation in people with overweight/obesity: a real-world follow-up of the phase 3 SURMOUNT-CN trial. Life Metabolism 2025;4(5):loaf024. doi:10.1093/lifemeta/loaf024. PMID 42058132. NCT05024032. Read at source. The paper describes itself as a "real-world, observational 26-week follow-up study" of participants who completed the 52-week phase 3 SURMOUNT-CN trial. Of 160 who completed and consented, 152 formed the eligible analysis population: tirzepatide 10 mg n=57, tirzepatide 15 mg n=51, placebo n=44. Follow-up ran week 52 to week 78, with no study drug. Mean percentage change in body weight from week 52 to week 78 was "9.1% (7.4), 12.3% (9.9), and 1.8% (5.2)" for the 10 mg, 15 mg and placebo groups. Net mean change from trial baseline to week 78 was −8.7% (6.9), −10.6% (10.2) and −2.5% (7.0). At week 78, the proportions still at least 10% below trial baseline were "53.1% and 42.6% versus 15.4%" for 15 mg, 10 mg and placebo, and at least 15% below were "34.7% and 16.7% versus 5.1%." On predictors, multivariate regression found that "Patient-reported partial to complete compliance to lifestyle consultations was negatively associated with weight change," while percentage weight loss during treatment and percentage regain at week 56 were positively associated with it; no association was found with sex, age, baseline BMI, hypertension, dyslipidemia, obstructive sleep apnea or atherosclerotic cardiovascular disease. Conclusion, quoted in two parts because the sentence exceeds the retrieval tool's limit: "Despite weight regain after treatment cessation, study participants exhibited a net weight loss" and "from trial baseline to Week 78, 26 weeks after tirzepatide cessation." Funding: "This study was funded by Eli Lilly and Company," with editorial support from a consultancy. Five of the fourteen authors "are employees of Eli Lilly and Company," and one is an editorial board member of the journal who was blinded from decisions on the manuscript.
10. Horn DB, Aronne LJ, Wharton S, Bays HE, le Roux CW, Srinath R, et al. Tirzepatide for maintenance of bodyweight reduction in people with obesity in the USA (SURMOUNT-MAINTAIN): a multicentre, double-blind, randomised, placebo-controlled trial. The Lancet 2026;407(10545):2305-2318. Published online May 12, 2026. doi:10.1016/S0140-6736(26)00656-2. PMID 42119587. NCT06047548. Phase 3b trial, read via the published abstract and the registry record. Carried over from this site's GLP-1 microdosing guide, where the figures were verified against the journal abstract, the indexing record and a conference summary, all three agreeing. 441 adults entered a 60-week open-label period on tirzepatide at the maximum tolerated dose; 378 were randomized 3:3:2 at week 60 to continue, reduce to 5 mg, or placebo for 52 further weeks. At week 112: −21.9% (95% CI −23.5 to −20.3), −16.6% (−18.0 to −15.1) and −9.9% from baseline. The 8%, 25% and 67% figures are not a plain regain count: the published sentence reads "Among participants who regained at least 50% of lost bodyweight, observed means were 11 (8%) of 138, 35 (25%) of 142," "and 60 (67%) of 90 participants received rescue therapy in the tirzepatide MTD, 5 mg tirzepatide, and placebo, respectively," and the methods state that "participants could receive rescue tirzepatide if their weight regain exceeded 50%." So the percentages are tied to receipt of rescue tirzepatide, two-thirds of the placebo arm received it, and the −9.9% is not a clean estimate of what stopping produces. The full text is behind a publisher this environment cannot reach and the paper is not open access, so the exact composite endpoint definition could not be read and is recorded here as unresolved. An earlier draft of this page, and the site's published microdosing guide, present the three percentages as shares who regained at least half their loss. One discrepancy recorded in the earlier verification: the conference presentation of the same trial reported −22.4%, −17.0% and −10.1%, and the published figures are the ones used.
11. Rodriguez PJ, Zhang V, Gratzl S, Cartwright BMG, Baker C, Gluckman TJ, Stucky NL, Emond JA. Discontinuation and reinitiation of dual-labeled GLP-1 receptor agonists among US adults with overweight or obesity. JAMA Network Open 2025;8(1):e2457349. doi:10.1001/jamanetworkopen.2024.57349. PMID 39888616. Read at source. Retrospective cohort using Truveta Data, electronic health records from 30 US health care systems with linked third-party dispensing and social-determinants data. 125,474 adults, of whom 76,524 had type 2 diabetes and 48,950 did not; the reinitiation analysis covered the 41,792 who discontinued. Follow-up ran up to two years for discontinuation and two further years for reinitiation, with median time to censoring of 618 days (IQR 470 to 730). Discontinuation was defined as "the first date a patient was 60 days or more without any GLP-1 RA on hand" "using fill dates and number of days of medication supplied." Kaplan-Meier estimates: 53.6% discontinued by one year (95% CI 53.4% to 53.9%) and 72.2% by two years; with type 2 diabetes 46.5% and 64.1%; without type 2 diabetes 64.8% by one year and 84.4% by two. Reinitiation, defined as "the first fill of any GLP-1 RA after discontinuation," reached 47.3% at one year in the diabetes group and 36.3% in the group without. On weight: "Weight regain of 1% from discontinuation was significantly associated with increased hazards of reinitiation of 2.3%," given as 2.3% (95% CI 1.9% to 2.8%) with diabetes and 2.8% (95% CI 2.4% to 3.2%) without. The study models regain as a predictor of reinitiation and does not publish a headline post-discontinuation regain magnitude, so none is attributed to it here. The author list was taken from the publisher page; verify the full string before print if it is to be printed in full.
12. Tzang CC, Wu PH, Luo CA, Chen ZT, Lee YT, Huang ES, Kang YF, Lin WC, Tzang BS, Hsu TC. Metabolic rebound after GLP-1 receptor agonist discontinuation: a systematic review and meta-analysis. eClinicalMedicine 2025;90:103680. doi:10.1016/j.eclinm.2025.103680. PMID 41399474. PROSPERO CRD42025646185. Full text read at source through an open-access full-text service and, for the discussion and back matter, through the article's open-access deposit. Scope: "Eighteen RCTs (3771 participants) were included," yielding 20 treatment-arm cohorts. Inclusion required trials that "investigated the effects of GLP-1RA treatment (including dual GLP-1/GIP receptor agonists such as tirzepatide)" in people "with metabolically related baseline conditions such as obesity, type 1 diabetes, or type 2 diabetes," that "reported outcomes at both the end of treatment phase (prior to cessation)" "and at one or more time points following cessation," and that "had ≥12 weeks of follow-up after cessation to ensure clinical relevance." Drugs represented across the pool: liraglutide, semaglutide, tirzepatide, beinaglutide, exenatide and albiglutide. Effect measure, and this was misdescribed on an earlier draft of this page: it is a random-effects mean difference computed within the discontinued arms, from end of treatment to end of post-cessation follow-up, not a comparison against people who continued. The methods state that "mean differences and 95% confidence intervals were calculated by comparing" "post-cessation values to end-of-treatment values," and the extraction section gives the metric as "the absolute change from the end of the GLP-1RA treatment phase" "(i.e., the point of planned cessation) to the end of the post-cessation follow-up period." There is no continuation control group in these estimates. Obesity subgroup: weight 5.63 kg (95% CI 3.52 to 7.73), p<0.0001, I² 99.57%, 11 studies, n=2,779; weight 5.81% (3.49 to 8.13), I² 99.61%; waist circumference 3.81 cm (2.15 to 5.47), I² 98.03%, 7 studies, n=1,978; BMI 2.34 kg/m² (1.21 to 3.46); systolic pressure 4.15 mm Hg (1.87 to 6.44), I² 96.60%, 9 studies, n=2,966; diastolic pressure 1.15 mm Hg (−0.27 to 2.57), p=0.11, not significant; fasting plasma glucose 0.45 mmol/L (0.32 to 0.59); HbA1c 0.25% (0.18 to 0.32), 7 studies, n=1,567; heart rate −3.22 beats per minute (−5.05 to −1.38), p<0.0001, I² 86.52%; HDL cholesterol 0.1 mmol/L (−0.03 to 0.22), p=0.13, not significant; LDL cholesterol 0.12 mmol/L (0.06 to 0.17); total cholesterol 0.19 mmol/L (0.13 to 0.25); triglycerides 0.17 mmol/L (0.1 to 0.25); VLDL 0.23 mmol/L (0.08 to 0.38). Type 2 diabetes subgroup: weight 2.03 kg (1.63 to 2.42), I² 42.28%, 5 studies, n=853; HbA1c 0.65% (0.22 to 1.08), 4 studies, n=219; fasting glucose 0.90 mmol/L (−0.36 to 2.17), not significant; no lipid results reported. Type 1 diabetes, 2 trials and 66 participants: weight 2.04 kg (−1.39 to 5.46), p=0.24, not significant. Drug-stratified analysis was run for liraglutide against semaglutide in the obesity subgroup only: weight regain 4.29 kg (2.43 to 6.15) against 8.21 kg (4.47 to 11.96), between-group p=0.008, and systolic pressure 1.56 mm Hg (−0.09 to 3.2, not significant) against 7.09 mm Hg (4.3 to 9.87), between-group p<0.001. Tirzepatide, exenatide, beinaglutide and albiglutide were not analyzed separately. Follow-up duration meta-regression in the obesity subgroup, longer than 26 weeks against 26 weeks or less: weight 7.31 kg (4.92 to 9.69) against 2.51 kg (0.97 to 4.05), p=0.007; waist 4.78 cm against 1.60 cm, p=0.02. Interpretation: "The magnitude and consistency of these effects underscore the physiological consequences of GLP-1RA discontinuation. As the clinical use of GLP-1RAs continues to grow in the management of obesity and diabetes, it is imperative that treatment guidelines address not only initiation and titration but also discontinuation and long-term maintenance strategies to sustain therapeutic gains." Funding: "This study received no funding." Interests: "We declare no competing interests." Two things about this paper are recorded here because they bear on how much weight its effect sizes can carry. First, "The risk of bias assessment using the ROB 2.0 tool rated all included studies as having a low risk," which is an unusually uniform judgment and which conflicts with the separate meta-regression that judged most studies in a larger overlapping literature to be at moderate risk. Second, the abstract labels the certainty of the 5.63 kg weight estimate "moderate" while reporting an I² of 99.57% alongside it, and the paper does not explain how both statements hold; its GRADE summary says certainty was "rated as moderate for most associations, with some outcomes downgraded to low certainty due to high heterogeneity and lack of trials reporting specific outcomes." Limitations the authors name: "the analysis did not account for lifestyle factors, such as exercise or dietary interventions, despite their potential influence on outcomes," with the example that one included trial "included a strict 8-week low-calorie diet prior to randomization, resulting in substantial initial weight loss (13.1 kg), which may have exaggerated rebound effects and limited generalizability to routine clinical settings"; that "reliance on study-level data precluded exploration of individual-level modifiers such as age, sex, baseline BMI, or comorbidities"; and that "the availability of rebound data was limited." Neither this paper nor the meta-regression reports a dichotomous glycemic-relapse outcome, so no figure for the proportion of people whose blood sugar returned to a diabetic range exists in either, a confirmed absence and not an unread item.
13. Qi H, Gao Q, Chen H, Du X, Pan B. Weight regain following discontinuation of glucagon-like peptide-1 receptor agonists in adults who are overweight or obese: a systematic review and meta-analysis. PeerJ 2026;14:e21713. doi:10.7717/peerj.21713. PMID 42729958. Abstract read at source through an open-access metadata record holding the publisher's own text; the publisher's article page refused, so the methods, tables and discussion were not read for this page and nothing from them is used. Open access under CC BY. Searches covered "PubMed, Embase, Cochrane Library and Scopus," with a period running "from the establishment of the database to December 2025," for randomized and controlled trials of "tirzepatide, semaglutide, liraglutide, and dulaglutide." Of 699 studies retrieved, "six studies involving 8,993 patients were" "included in the quantitative analysis, comprising 5,553 patients in the discontinuation group and" "3,440 in the continued treatment group." That between-group estimate does exist, and one paper reports it: "compared with the continued treatment group, the weight difference in the discontinuation group was" "mean difference (MD) = 17.90%, 95% confidence interval (CI) [14.11-21.69], P < 0.0001," with "statistical heterogeneity among the studies (P = 0.0082)." It also reports the within-arm version separately: "the percentage difference in body weight between after and before drug withdrawal was" "MD = 9.11%, 95% CI [7.91-10.30], P < 0.0001." On the drugs: "Subgroup analysis further revealed that the weight rebound amplitude after discontinuation of tirzepatide" "was significantly higher than that of semaglutide," which the authors attribute to possible mechanistic differences. On safety after stopping: "after drug withdrawal," "the overall adverse reactions of patients decreased, gastrointestinal adverse reactions decreased, and" "the incidence of cardiovascular events was not affected by drug withdrawal." Conclusion: "There is a significant weight rebound phenomenon after discontinuation of GLP-1RAs, and the rebound" "magnitudes vary among different types of drugs." Risk of bias was assessed with the Cochrane tool version 2.0 and the analysis run in R; the per-study ratings sit in the full text and were not read. Funding and competing interests were not in the record read and are therefore unverified. This paper was missed by the first research pass on this page and was found by an independent fact check; the page previously described the literature as holding two meta-analyses.
14. Horn DB, Linetzky B, Davies MJ, Laffin LJ, Wang H, Murphy MA, Zimner-Rapuch S, Lau E, Arad AD, Lee CJ. Cardiometabolic parameter change by weight regain on tirzepatide withdrawal in adults with obesity: a post hoc analysis of the SURMOUNT-4 trial. JAMA Internal Medicine 2026;186(2):157-167. Published online November 24, 2025. doi:10.1001/jamainternmed.2025.6112. PMID 41284285. NCT04660643. Abstract, methods and results read at source on the publisher's page; funding, sponsor role and conflict-of-interest statements read at source through the article's open-access deposit, because the publisher's served page truncates before that block. Note the dual date: online first in 2025, issue-dated 2026. Design, in the paper's words: "This post hoc analysis of the SURMOUNT-4 trial included tirzepatide-treated participants" "with 10% or greater weight reduction at week 36 initially randomized to placebo," n=308, with the analyses described as exploratory. Participants were grouped by the share of their lead-in weight loss that they regained, in four bands, and the figures are mean change from week 36 to week 88. Weight regain as a percentage of the week 0 to 36 loss ran from 3.2% to 22.9% across the bands, 15.1% overall. Waist circumference by band: "<25% weight regain, 0.8 cm; 95% CI, −1.0 to 2.6; 25% to <50%, 5.4 cm; 95% CI, 4.0-6.8;" "50% to <75%, 10.1 cm; 95% CI, 8.9-11.3; ≥75%, 14.7 cm; 95% CI, 12.7-16.7; P < .001." Systolic pressure: "6.8 mm Hg [95% CI, 3.9-9.7], 7.3 mm Hg [95% CI, 4.8-9.8], 9.6 mm Hg [95% CI, 7.1-12.1]," "and 10.4 mm Hg [95% CI, 8.0-12.8], respectively; P = .002." Non-HDL cholesterol: "−0.4% [95% CI, −7.3 to 6.5], 1.6% [95% CI, −2.3 to 5.5]," "8.4% [95% CI, 3.9-12.9], and 10.8% [95% CI, 5.3-16.3], respectively." HbA1c: "0.14% [95% CI, 0.06-0.22], 0.15% [95% CI, 0.09-0.21], 0.27% [95% CI, 0.21-0.33]," "and 0.35% [95% CI, 0.29-0.41], respectively; P < .001." Fasting insulin: "−4.0% [95% CI, −20.7 to 12.7], 15.4% [95% CI, 2.3-28.5], 46.2% [95% CI, 29.5-62.9]," "and 26.3% [95% CI, 9.6-43.0], respectively." Point estimates read from the article's table but without individually quoted confidence intervals, in the same four-band order, and therefore reported here without intervals: diastolic pressure 2.8, 1.6, 3.9 and 4.3 mm Hg; fasting glucose 3.8, 6.9, 9.2 and 9.0 mg/dL; triglycerides 5.5%, 5.3%, 29.6% and 18.9%; HDL cholesterol 19.7%, 16.7%, 13.6% and 13.4%; pulse rate −5.4, −6.7, −5.7 and −4.2 beats per minute. In the band that regained less than 25%, the week 36 to week 88 changes in waist circumference, non-HDL cholesterol and fasting insulin were not significantly different from week 36. The relationship is graded but not uniform: waist circumference, systolic pressure, HbA1c, non-HDL cholesterol and fasting glucose rose monotonically with regain, while fasting insulin peaked in the third band, diastolic pressure was non-monotonic, HDL cholesterol ran in reverse, and pulse rate showed no trend and stayed below its week 36 value in every band. The figure legend notes that no statistical comparison among bands was available for fasting insulin. Conclusion: "In this post hoc analysis of the SURMOUNT-4 trial, among participants with obesity who achieved weight reduction with 36-week tirzepatide treatment, withdrawing tirzepatide led to 25% or greater weight regain in most participants within 1 year and was associated with a greater reversal of their initial cardiometabolic parameter improvements compared with those who maintained weight reduction. These findings underscore the importance of continued obesity treatment." Funding: "This study was sponsored by Eli Lilly and Company." Role of the funder, quoted because it is unusually explicit: "Eli Lilly and Company was involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The sponsor did not have the right to veto publication or to control the decision regarding to which journal the manuscript was submitted. Final decisions resided with the authors, which included employees of the sponsor." Disclosures: six of the ten authors "reported employment and stock ownership from Eli Lilly and Company during the conduct of the study and outside the submitted work." The first author reported grants from Eli Lilly during the study, plus grants from Eli Lilly and Novo Nordisk and personal fees from Eli Lilly, Novo Nordisk, AstraZeneca, Boehringer Ingelheim, Zealand Pharmaceuticals and Amgen outside the submitted work. Two further authors reported extensive industry grants, advisory board roles and personal fees across Novo Nordisk, AstraZeneca, Boehringer Ingelheim, Eli Lilly and others.
15. WEGOVY (semaglutide) injection prescribing information, NDA 215256, revised 06/2026. US Food and Drug Administration. accessdata.fda.gov. Read at source. Indication, quoted in segments because the sentence spans several lines: "WEGOVY injection is indicated in combination with a reduced calorie diet and increased physical activity:" then "To reduce excess body weight and maintain weight reduction long term in:" then "Adults and pediatric patients aged 12 years and older with obesity." and "Adults with overweight in the presence of at least one weight-related comorbid condition." Section 1 gives no numeric BMI thresholds, using the terms obesity and overweight. The label also carries a cardiovascular-risk-reduction indication and an accelerated-approval MASH indication. Section 9, Drug Abuse and Dependence, is absent: the FULL PRESCRIBING INFORMATION contents run from "8 USE IN SPECIFIC POPULATIONS" to "10 OVERDOSAGE," and section 15 is likewise absent. The contents table carries the standard footnote "Sections or subsections omitted from the full prescribing information are not listed," which does not name what was omitted. That footnote text was read on this label; the body's reference to it is scoped to Wegovy for that reason, while the absence of section 9 was confirmed on all four labels. Section 2 contains no treatment duration and no criteria for continuing or stopping; its only discontinuation language is a product-switch instruction, "One week after discontinuing WEGOVY 2.4 mg injection, initiate 25 mg of WEGOVY tablets orally once daily." The word "chronic" appears only in unrelated senses such as chronic kidney disease and is never used of the therapy. Section 12.3 Elimination: "semaglutide will be present in the circulation for about 5 to 7 weeks after the last dose of 2.4 mg." Section 10 Overdosage: "taking into account the long half-life of WEGOVY of approximately 1 week." Section 14.2, Study 5, is the STEP 4 design: "Study 5 was a 68-week trial that enrolled 902 patients with obesity," "All patients received WEGOVY during the run-in period of 20 weeks that included 16 weeks of dose escalation," "803 patients reached WEGOVY 2.4 mg and were then randomized in a 2:1 ratio to either continue on WEGOVY or receive placebo." Table 7, change from randomization at week 20 to week 68: WEGOVY continued, n=535, −7.9%; placebo, n=268, +6.9%; difference from placebo −14.8 (−16; −13.5), p<0.001. The label's only prose on the result is neutral: "From randomization (week 20) to week 68, treatment with WEGOVY resulted in a statistically significant reduction" "in body weight compared with placebo (Table 7)." The word regain does not appear anywhere in the label. Section 6.1 references "a 7-week off-drug follow-up period" after the 68 treatment weeks but gives no weight figure for it. Safety-driven discontinuation instructions include 5.2 "If pancreatitis is suspected, discontinue WEGOVY and initiate appropriate management," 5.9 "If patients experience a sustained increase in resting heart rate, discontinue WEGOVY," 5.10 on "temporarily discontinuing WEGOVY" to reduce retained gastric contents, and 8.3 instructing clinicians to "discontinue WEGOVY in patients at least 2 months before they plan to become pregnant." Section 6.1: "Permanent discontinuation of treatment as a result of a gastrointestinal adverse reaction occurred in 4.3%." Two details read on the 10/2025 revision of the same label, 215256s026lbl.pdf, because the 06/2026 file could not be read that far: the section 14.2 table and the section 12.3 elimination sentence. The 06/2026 file supplied the indication, the missing section 9 and the section 2 and 5 material.
16. ZEPBOUND (tirzepatide) injection prescribing information, NDA 217806, revised 08/2026, and MOUNJARO (tirzepatide) injection prescribing information, NDA 215866, revised 08/2026. US Food and Drug Administration. Zepbound and Mounjaro. Both read at source. Zepbound indication, quoted in segments: "ZEPBOUND® is indicated in combination with a reduced-calorie diet and increased physical activity:" "to reduce excess body weight and maintain weight reduction long term in adults with obesity or adults with overweight" "in the presence of at least one weight-related comorbid condition," plus a second indication for moderate to severe obstructive sleep apnea in adults with obesity. Mounjaro indication: "MOUNJARO® is indicated as an adjunct to diet and exercise to improve glycemic control in adults and pediatric patients" "10 years of age and older with type 2 diabetes mellitus." Section 9, Drug Abuse and Dependence, is absent from both; confirmed on the 08/2026 files and independently on the 04/2026 revisions 217806Orig1s050lbl.pdf and 215866Orig1s052lbl.pdf. Zepbound section 2 gives dose escalation and maintenance doses only, with no total treatment duration and no stopping criteria. Mounjaro section 12.3: tirzepatide is absorbed "with an elimination half-life of approximately 5 days," and "Steadystate plasma tirzepatide concentrations were achieved following 4 weeks of once weekly administration," with the source printing "Steadystate" unhyphenated. Mounjaro section 10 repeats "the half-life of tirzepatide of approximately 5 days." Section 8.3 of the Mounjaro label carries no pregnancy washout statement, unlike the semaglutide labels; its only timing guidance concerns oral contraceptive absorption. Zepbound section 14.1, Study 4, is SURMOUNT-4, NCT04660643: "an 88-week randomized withdrawal trial in which all patients received open-label ZEPBOUND during a 36-week lead-in period," "followed by randomization to either continue ZEPBOUND or switch to placebo for 52 weeks." Table 6, least-squares mean percent change in body weight from week 36 to week 88: switched to placebo +14.0; continued Zepbound at maximum tolerated dose of 10 or 15 mg −5.5; difference −19.4 (95% CI −21.2, −17.7). The label reports that figure without using the word regain. Also from section 14.1: "Of the 783 patients who started ZEPBOUND at Week 0, 14.4% discontinued treatment before randomization at Week 36," "and adverse events were the most common reason for discontinuation (6.8%)," and "The proportions of patients who discontinued study drug after randomization at Week 36 were 10.4%" "for the ZEPBOUND-treated group and 17.9% for the placebo-treated group." Section 6.1: "Across both trials, 4.8%, 6.3%, and 6.7% of patients treated with 5 mg, 10 mg, and 15 mg of ZEPBOUND, respectively," "permanently discontinued treatment as a result of adverse reactions compared to 3.4% of patients treated with placebo," and "The majority of patients who discontinued ZEPBOUND due to adverse reactions did so during the first few months" "of treatment due to gastrointestinal adverse reactions." This footnote replaces a figure previously carried on this site from the 01/2026 Mounjaro revision; the five-day half-life is now confirmed on the 08/2026 label at source.
17. Damhof MA, van Bruggen FH, van Roon EN. Early weight regain after GLP-1 receptor agonist discontinuation: mechanisms and implications for treatment de-escalation strategies. Diabetes, Obesity and Metabolism 2026;28(10):8769-8776. Published online July 3, 2026. doi:10.1111/dom.71075. PMID 42396734. Abstract read via a reproduction of the indexed record; the full text was not opened, so nothing from the body of the review informs this page. A narrative review, not original data. Quoted from the abstract: "restoration of appetite following withdrawal of pharmacological GLP-1 signalling," "compensatory increases in orexigenic hormones such as ghrelin," "alterations in incretin balance including glucose-dependent insulinotropic polypeptide (GIP)," "potential adaptive changes in GLP-1 receptor signalling pathways during prolonged pharmacological exposure," and then, decisively for how this page uses it, "evidence for a direct role in post-cessation weight regain remains limited." The abstract also describes "a transient mismatch between increased biological drive for weight regain" with "a large proportion occurring during the early post-cessation period." On tapering it states that whether a structured taper can facilitate discontinuation "remains unknown and requires prospective evaluation." The abstract does not use the term withdrawal syndrome and does not mention gastric emptying or energy expenditure.
18. Yevusiak T, Weisman A, Retnakaran R, Wharton S, Drucker DJ, Kramer CK. Impact of semaglutide withdrawal on cardiometabolic profile and physiology of energy balance: recovery effects after semaglutide termination, the REST trial study protocol. PLOS ONE 2026;21(7):e0354237. doi:10.1371/journal.pone.0354237. Read at source. A published study protocol with no results. Cited on this page as the strongest available source for the claim that the physiology after stopping has not been measured and that tapering has not been tested against stopping. The protocol states "There is little clinical data on the best strategy to achieve weight maintenance," and describes the mechanism in hedged terms: "This rebound effect is likely related to physiologic mechanisms triggered by weight reduction that induce compensatory changes in energy balance homeostasis" "including increased appetite and reduced energy expenditure." Its future-perspective section says "gradual dose reduction may attenuate this effect by allowing physiological adaptation in energy balance regulation," which is a hypothesis the trial will test. Design: primary outcome is percentage body-weight change from baseline to 16 weeks after complete withdrawal, comparing gradual reduction of 25% every four weeks against immediate cessation; secondary outcomes include 24-hour systolic blood pressure and fasting ghrelin at 16 weeks post-discontinuation, plus a standardized meal test with post-meal hormone sampling, appetite visual analogue scales, eating-behavior questionnaires and dietary intake. Ghrelin appears only as a secondary outcome. Recruitment began March 1, 2026, data collection is expected to end July 31, 2028, and analysis is expected in 2030.
19. FDA Adverse Event Reporting System, queried through the openFDA drug event API. US Food and Drug Administration. api.fda.gov. Queried at source. Every query returned a meta.last_updated value of 2026-07-30. Report totals, each from a single-clause exact search on the field patient.drug.medicinalproduct.exact with limit=1, which is the field that reproduces these numbers; the same query against patient.drug.openfda.brand_name.exact returns different totals on the same data vintage, so the field has to be recorded with any FAERS count: WEGOVY 21,023; ZEPBOUND 64,947; semaglutide as generic name 73,001; and 87,542 reports across all of FAERS carrying the reaction term DRUG WITHDRAWAL SYNDROME. Reaction-term bucket lists were then pulled for each brand with count=patient.reaction.reactionmeddrapt.exact&limit=300. No term containing WITHDRAW, DEPENDEN or REBOUND appears in the 300 most-reported reactions for any of WEGOVY, OZEMPIC, ZEPBOUND or MOUNJARO. The lowest-count bucket returned sets the detection floor for each list: 29 reports for Wegovy, 109 for Ozempic, 31 for Zepbound and 74 for Mounjaro. Run in the other direction, counting the drugs named in reports coded DRUG WITHDRAWAL SYNDROME, 300 buckets returned none containing WEGOVY, OZEMPIC, ZEPBOUND, MOUNJARO, SEMAGLUTIDE, TIRZEPATIDE, SAXENDA or VICTOZA, with that list's lowest bucket at 100 reports. For scale on terms that do appear: WEIGHT INCREASED runs 1,015 for Wegovy, 1,777 for Zepbound, 2,454 for Ozempic and 1,464 for Mounjaro, and FOOD CRAVING runs 301 for Wegovy and 372 for Zepbound. Two constraints on this query, both stated and not worked around. Boolean AND failed through this environment's retrieval path, returning the whole-database total of 20,692,690 for every two-clause query including a known-good control, so no co-occurrence count was obtained and none is estimated anywhere on this page. Because the bucket lists are capped at 300, the correct statement is that withdrawal-coded reports fall below the floors given above, not that there are none. openFDA's own disclaimer is returned with every query: "Do not rely on openFDA to make decisions regarding medical care." and "While we make every effort to ensure that data is accurate, you should assume all results are unvalidated." FAERS reports are voluntary, unverified and do not establish causation.
20. Chiappini S, Vickers-Smith R, Harris D, Papanti Pelletier GD, Corkery JM, Guirguis A, Martinotti G, Sensi SL, Schifano F. Is there a risk for semaglutide misuse? Focus on the Food and Drug Administration's FDA Adverse Events Reporting System (FAERS) pharmacovigilance dataset. Pharmaceuticals 2023;16(7):994. doi:10.3390/ph16070994. Read at source. A disproportionality analysis of FAERS reports from January 2018 to December 2022, data accessed July 4, 2023, using the standardized MedDRA query "Drug abuse, dependence and withdrawal" with preferred terms including drug abuse, substance abuse, intentional product misuse, dependence, drug withdrawal syndrome, withdrawal and withdrawal syndrome. Semaglutide accounted for 8,249 adverse event reports, 26.1% of the 31,542 across the molecules studied; comparators were dulaglutide 11,858, liraglutide 7,883, exenatide 2,585, lixisenatide 390, tirzepatide 290, phentermine-topiramate 183 and albiglutide 104. Proportional reporting ratios for semaglutide against the remaining molecules: drug abuse 4.05, drug withdrawal syndrome 4.05, prescription drug used without a prescription 3.60, intentional product use issue 1.80. No significant difference was found against phentermine-topiramate. On the line most often quoted from this paper: its reference to "rebound and craving phenomena after withdrawing from semaglutide" is the tail of a sentence describing the ratio as "consistent with a range of anecdotal Redditors' observations" highlighting those phenomena. It is the authors' comparison with online anecdote and not their characterization of the FAERS data, and an earlier draft of this page quoted it as the latter. Other quoted fragments: "The current findings, which indicate a possible semaglutide misuse/abuse issue," "the withdrawal symptoms associated here with semaglutide may be conceivably related," and "highlighting both rebound and craving phenomena after withdrawing from semaglutide." The authors note the limits of spontaneous-report disproportionality analysis and state that further empirical research is needed. One limit on this citation: the per-term report counts behind the ratios sit in Table 4, served by the publisher as an image, so the numerator for the 4.05 figure could not be read. The ratio is reported on this page with that gap stated.
21. University of Oxford. New study finds that stopping weight-loss drugs is linked to faster regain than ending diet programmes. January 8, 2026. ox.ac.uk. Read at source. The university's own announcement of the study above. Used on this page for two things the paper's results do not state: the projection of approximately 1.5 years to return to starting weight for semaglutide and tirzepatide specifically, and two researcher quotations. Sam West, lead author: "This isn't a failing of the medicines - it reflects the nature of obesity as a chronic, relapsing condition." Susan Jebb, joint senior author: "Obesity is a chronic, relapsing condition, not a short-term problem with a quick fix." The announcement also gives the monthly figures in pounds, 0.9 and about 1.8, and states that regain after drugs was faster than after behavioral programs by approximately 0.3 kg a month, matching the paper.
22. Marchetti M, Campisano G, Paragliola RM, Francomano D, Cipolloni V, Conforti A. Weight regain after discontinuation of GLP-1 and GIP/GLP-1 receptor agonists: a systematic review. Metabolites 2026;16(10):757. doi:10.3390/metabo16100757. Read at source. A systematic review with narrative synthesis and no new pooled estimate. 18 reports representing 16 study families, from 363 records screened and 59 full texts assessed, searching PubMed and MEDLINE, Europe PMC and ClinicalTrials.gov plus citation searching, from inception to September 2026. That end date is four months after SURMOUNT-MAINTAIN was published, and the review does not include or mention it, naming SURMOUNT-4 as the only tirzepatide withdrawal trial. An earlier draft of this page explained the omission by saying the search closed before that trial appeared, which is not correct. Inclusion covered adults with overweight or obesity with or without type 2 diabetes, quantifying weight or cardiometabolic outcomes at least 12 weeks after planned or unplanned cessation, or testing a post-cessation maintenance strategy. It reproduces the three canonical figures this page cites to their own papers: STEP 4 regain of 6.9% from weeks 20 to 68 against −7.9% continued; SURMOUNT-4 regain of 14.0% from weeks 36 to 88 against −5.5% continued; and the STEP 1 extension's 11.6 percentage points in one year, about two-thirds of prior loss, leaving participants 5.6% below baseline. Cited on this page for two statements of its own. On real-world data: "Estimates were smaller and more variable, partly because discontinuation was often inferred and many patients restarted or switched therapy." On maintenance strategies: "evidence for specific dietary, behavioral, tapering or transition regimens is insufficient," and no included trial directly compared dose reduction with full discontinuation. One bibliographic inconsistency is recorded here: the publisher page's citation metadata lists the final author as the sole author while the article itself carries the full list above.

Sources that could not be read for this page

• The full text of SURMOUNT-MAINTAIN, the dose-reduction trial, which is not open access and whose publisher refused every route. Its figures here come from the published abstract and the registry record, and the exact definition of its rescue-therapy endpoint could not be read.
• The publisher's own site for three of the four meta-analyses. Each was read in full or in abstract through the open-access version of record deposited elsewhere, and each footnote says which.
• The Discussion sections of STEP 4 and SURMOUNT-4, absent from the text their publisher serves. Nothing here rests on either, and the observation that STEP 4 avoids the word regain is scoped to the sections that could be read.
• SURMOUNT-4's conflict-of-interest statement and its week 88 blood pressure, lipid and HbA1c results. The first was unreachable, and the second sits in an online supplement that could not be opened, so no such figure from that trial appears above.
• Table 4 of the 2023 FAERS analysis, published as an image, so the report counts behind its reporting ratio could not be read.
• ClinicalTrials.gov registry records, which returned empty templates, so no posted trial results informed this page.
• Boolean queries against openFDA, which returned the whole-database total through this environment, so no co-occurrence count was obtained and none is estimated above.