How Long Should You Stay on Weight Loss Peptides? The Answer Is Not "Forever"

About half of the people who start a GLP-1 medication for weight loss stop it within a year. A Danish nationwide registry of 77,310 patients found 18% had quit by month 3 and 52% by month 12 — and that roughly half of survey respondents had never discussed with their doctor how long they should be on therapy in the first place.

The trial protocols all assumed continuous use. The biology assumes continuous use. Yet many patients are quietly running their own ad-hoc experiment in stopping. The drugs work. The question almost no one is asking out loud is how long you should stay on them — and what changes when you come off.

The duration question patients aren't asking

A November 2025 Kaiser Family Foundation poll found that 1 in 8 U.S. adults have tried a GLP-1 medication. The trials behind those prescriptions ran for 68 or 104 weeks of continuous treatment. That is the entire formal evidence base. The question of when, or whether, to stop is being answered in real time — by patients, insurance plans, and supply shocks rather than by data.

Reimar Thomsen, the Aarhus researcher who presented that 77,310-person Danish registry at the 2024 EASD meeting, was blunt: "Stopping rates are simply very high — too high in my mind." He added what most obesity-medicine colleagues will say privately: "for this to be effective, they need to basically take it the rest of their lives, like antihypertensives and lipid-lowering agents."

Most people don't think of a weight-loss drug the way they think of a statin. They think of it the way they think of antibiotics: take it, fix the problem, stop. The trial data and the biology underneath them say that mental model is wrong — and getting it wrong is what produces the rebound stories.

What the trials actually show when you stop

Four trials are doing most of the heavy lifting on this question, and they tell a remarkably consistent story.

STEP 1 extension. The original STEP 1 put 1,961 adults on semaglutide 2.4 mg or placebo for 68 weeks; investigators then followed 327 of them for another year off-drug. Mean loss during treatment: 17.3% of starting body weight. A year after the last injection, participants had regained 11.6 percentage points — net loss 5.6%. The investigators' summary: participants "regained two-thirds of their prior weight loss." Blood pressure, glucose, and inflammation drifted back toward baseline alongside the scale.

STEP 4. Built specifically to test whether continuing the drug matters. Everyone enrolled spent 20 weeks on semaglutide and lost about 10.6%. They were then randomized to continue the drug or switch to placebo for 48 more weeks. The continuation arm lost another 7.9%; the placebo-switched arm gained 6.9% back. The gap at week 68 was 14.8 percentage points — almost as large as the original treatment effect itself.

STEP 5. The longest published semaglutide trial. Two years on weekly 2.4 mg produced 15.2% mean weight loss vs 2.6% on placebo, with the curve plateauing around week 60 and holding steady through week 104. Of participants with baseline prediabetes, 79.7% reverted to normal blood glucose on semaglutide vs 37% on placebo. Twenty-six percent stopped antihypertensives entirely during the trial. The drug works for as long as you take it.

SURMOUNT-4. The tirzepatide counterpart of STEP 4 — and the most damning result for "lose it and stop." Patients spent 36 weeks on tirzepatide, dropping 21.9% of body weight, then half switched to placebo. A November 2025 JAMA Internal Medicine post hoc analysis showed that 82.5% of the placebo-switched group regained at least 25% of what they'd lost within a year. Nearly 9% gained back more than 100% — they ended above their starting weight. Only 17.5% kept regain below a quarter.

That last figure is worth holding onto. It means one in six people, even after stopping, can hold most of their progress for at least a year. It does not mean five in six can. For the typical patient, the effect of the drug is the drug.

That 2026 University of Cambridge meta-regression pooled six of the largest GLP-1 trials (3,236 patients) and put one-year regain at 60% of the weight lost, with a "half-life" of 23 weeks. The curve extrapolates to a long-term plateau around 75% regained — meaning roughly 25% of the loss persists indefinitely, or about 4-5% of starting body weight. Better than nothing. Not what most patients signed up for.

Why obesity medicine treats this as chronic, not acute

The mismatch between how patients use these drugs and how clinicians think they should be used comes down to a definitional fight most patients have never heard. The WHO, the American Medical Association, the European Association for the Study of Obesity, and (since 2021) the European Commission have all formally classified obesity as a chronic, progressive, relapsing disease. As a 2026 international editorial in Eat Weight Disord put it: nobody waits for a heart attack to treat blood pressure, and nobody should wait for organ damage to treat the biology of obesity.

If obesity is a chronic disease, a drug that controls it is a chronic-disease drug. You don't stop a statin when your LDL hits target. You don't stop an ACE inhibitor when your blood pressure normalizes. The November 2025 JAMA Internal Medicine editorial alongside SURMOUNT-4 argued GLP-1s should be reframed as "weight management" rather than "weight loss" drugs. Matthew Hayes, a psychiatry professor at the University of Pennsylvania's Perelman School of Medicine, said it more bluntly: "Going on a GLP-1 is a lot like going on a statin. Patients must accept that they'll be on the drug, or an alternative, for life."

The cardiometabolic numbers underneath that recommendation are sobering. In SURMOUNT-4, when patients regained at least 75% of their weight after stopping tirzepatide, waist circumference rose 14.7 cm on average, systolic BP climbed 10.4 mmHg, and HbA1c rose 0.35 points — roughly back to baseline. The Cambridge meta-regression found that about half of any HbA1c improvement is gone within 8-12 weeks off-drug, and 70-80% of the systolic BP improvement reverses within 12 weeks. If a patient went on a GLP-1 partly to avoid progressing from prediabetes to diabetes, those are wins for as long as the drug is on board — and not after.

Five real reasons therapy ends

"Lifelong" is not the same as "no exceptions." There are five legitimate reasons a clinician will pull the plug.

Cost and access. Even in Denmark's subsidized system, lower-income patients quit at higher rates. In the U.S., the cost story is sharper: brand-name semaglutide or tirzepatide runs into five figures a year without insurance, and supply shortages have made even paid-for prescriptions intermittent.

Side effects that don't resolve. Most nausea, diarrhea, and constipation peak during dose escalation and fade. But not always. In a 76-patient community Italian obesity clinic, 35% had moderate-to-severe GI symptoms during titration, and a meaningful subset never tolerated the maintenance dose. In trials, GI events drove only 3-4% of discontinuations; in real-world community settings — faster titration, sparser follow-up — the number is dramatically higher.

A genuine plateau. A plateau on therapy is typically defined as eight to twelve weeks of no further loss. It is not failure — the body has adapted to a lower set point. If a patient has reached and held a target weight on a stable dose, that is a defensible point to discuss reduction.

Pregnancy. GLP-1s are contraindicated in pregnancy and during attempts to conceive — both because of inadequate fetal-safety data and because rapid weight loss carries early-pregnancy risk. The Aarhus case-control analysis flagged pregnancy as one of the events most strongly associated with stopping; patients are typically advised to come off the drug at least two months before trying to conceive.

New diagnoses that change the calculus. Pancreatitis, biliary or pancreatic cancer, and severe gastroparesis are reasons to stop entirely. Thyroid disease occasionally is. Bariatric surgery — which paradoxically reduced discontinuation risk by 15% in one Aarhus subgroup — is sometimes a reason to add the drug, not remove it.

Note what is not on this list. "I hit my goal weight" is not by itself a medical reason to stop, even though it is the most common reason patients ask.

Maintenance dosing, tapering, and the off-cycle debate

If "stop entirely" usually fails and "lifelong full dose" is unaffordable for many people, the middle ground is where the actual clinical action is. Three approaches are being tested in real practice.

Lowest effective maintenance dose. The most common recommendation from obesity-medicine physicians is to find "the lowest dose that still helps regulate appetite and stabilize weight." For semaglutide, that often means coming down from 2.4 mg to 1.7 or 1.0 mg weekly; for tirzepatide, dropping a step or two from the 15 mg ceiling. There is no head-to-head trial proof of non-inferiority versus full-dose maintenance, but the clinical experience is favorable enough that most obesity programs are already doing it.

Stepwise taper toward stop. Brufani and Morviducci, writing in Obesity and Endocrinology in 2025, proposed a structured protocol they openly call "anecdotal and hypothesis-generating": 2.4 mg → 1.7 mg for 4-6 weeks → 1.0 mg → 0.5 mg, watching for less than 2% regain at each step before moving down. A separate February 2026 piece in The Conversation suggested a similar three-to-six-month gradient. The logic: abrupt withdrawal triggers a sharper appetite rebound than a gradient does, and tapering gives lifestyle habits time to harden.

Extended-interval dosing. Mitch Biermann at Scripps presented a case series at Obesity Week 2025 showing that older adults who switched from weekly to every-other-week semaglutide kept most of their loss and shifted body composition favorably toward fat (rather than muscle) loss. A Diabetes, Obesity and Metabolism study found every-other-week dosing maintained roughly 75% of the weight loss. None of these approaches are FDA-labeled, but they are how working clinics handle the affordability and supply problem.

A policy note worth flagging: the UK's NICE caps semaglutide use for weight loss at two years, with no equivalent limit on tirzepatide. The Cambridge meta-regression authors openly criticized that policy — there is no biological basis for a 24-month cutoff. Treat it the way regulators treat cholesterol: assess, adjust, don't yank because the calendar says so.

The lifestyle handoff before you stop

The most important question before any reduction is body composition — not just the scale. Between 15% and 50% of the weight lost on a GLP-1 is lean tissue, mostly muscle. The Cambridge group cited a 40-60% range for newer agents. After 28 weeks on tirzepatide 15 mg, a typical patient loses 11.2 kg total, including 1.6 kg of fat-free mass. In SURMOUNT-1, tirzepatide produced 26% greater fat loss than placebo, but also 8% greater lean-mass loss. That muscle does not automatically come back when you stop. The fat does.

The clearest evidence on what to do about it comes from Signe Torekov's Copenhagen group. Jensen and colleagues randomized 195 adults who had already lost 12% of body weight to placebo, supervised exercise, liraglutide, or both. At one year off all treatment, only the exercise-plus-drug group maintained significantly more loss than placebo (5.1 kg more). Drug alone was not statistically better. Exercise was the variable that protected the result.

Two protocols are emerging. The protein target is 1.6 g/kg/day, with 1.2 g/kg/day as a floor, spread evenly at roughly 25 g per meal. The exercise prescription is resistance training three times a week — squat, press-up, row, lunge, plank. Holly Wyatt at the University of Colorado Anschutz recommends a first-meal-of-the-day target for patients tapering off: at least 25 g protein, 15 g fiber, under 45 g carbs, minimal added sugar — a structural mimic of the appetite suppression the drug used to provide.

The people who keep their weight off are not the ones with the strongest willpower. They're the ones who used the drug-induced window to add muscle, lock in sleep, and rewire grocery habits — so that when the appetite signal returns, the structure of their day already does most of the work.

A practical framework for your own timeline

The framework the evidence supports:

Don't think in cycles. Think in phases. The meta-regression shows weight regain follows a predictable exponential curve no matter how clean the lifestyle looks afterward. "On for six months, off for three" is a recipe for the red bars in the chart above.

If cost is the problem, drop the dose, not the drug. Lowest-effective-maintenance or every-other-week dosing both have preliminary support. Full dose to nothing is the worst available option for regain.

If you must stop, taper. Three to six months of structured dose reduction gives the appetite system time to adjust and gives you time to put structural defenses in place. An abrupt stop is the worst case, not the best.

Build the lifestyle stack first, then taper. Hit 1.6 g/kg/day of protein, three resistance sessions a week, and seven to nine hours of sleep before reducing the medication. Reverse the order and you're trying to build new habits during an appetite rebound — the worst possible time.

Decide the long-term posture in advance. If your reason for taking the drug is "I want to be 30 pounds lighter for a while," the trial data say that's an expensive way to lose 4-5% permanently. If your reason is "I want to manage a chronic metabolic condition," think about it the way you'd think about lipid or blood pressure management — and accept that the answer may be lifelong.

There isn't one number. The right duration depends on whether you treat the drug as a procedure or a treatment — and whether you've used the time it bought you to build something underneath it. The wrong answer, and the most common one, is "until I hit my goal."

Frequently Asked Questions

How long does it take to gain the weight back after stopping?

The Cambridge meta-regression of six major trials found regain follows an exponential curve with a half-life of about 23 weeks: roughly 60% of the lost weight is back at one year, plateauing around 75% long-term. About one in six people in SURMOUNT-4 maintained most of their loss for at least a year — but the average pattern is steady, predictable regain.

Is it safer to stop cold turkey or to taper?

No head-to-head trial answers this directly, but the clinical consensus is that a structured three-to-six-month taper beats an abrupt stop. Abrupt withdrawal produces a sharper appetite rebound, and a gradient gives time to build the lifestyle infrastructure that has to take over.

Can I do "off cycles" — three months on, three months off?

The evidence does not support it. Every published trial of GLP-1 discontinuation shows the same regain curve. Weight cycling is also independently associated with worse body composition — muscle on the way down, fat on the way back up. If cost is the problem, lower-dose or extended-interval maintenance has more support than periodic stops.

I've reached my goal weight — isn't that a reason to stop?

It is a reason to discuss reducing the dose, not necessarily to stop. The November 2025 JAMA Internal Medicine editorial argued that GLP-1s are "weight management" drugs, not "weight loss" drugs — the goal is sustained metabolic improvement, not a number on a scale. Most clinicians move patients to a maintenance dose and reassess every six to twelve months.

Will I lose muscle if I stay on the drug long-term?

Some lean tissue loss is unavoidable with any weight loss — 15% to 50% of total loss is typically lean mass. The Copenhagen Jensen et al. study showed that supervised resistance exercise alongside a GLP-1 was the only combination that protected lean mass and kept weight off afterward. Drug alone did not. The point of the lifestyle handoff is to keep that lean mass intact during therapy.