The number everyone quotes

If you've started looking into stopping, you've met the statistic already. Two-thirds of the weight comes back within a year. It gets quoted in headlines, in forum replies, and usually by someone who's about to tell you not to bother stopping at all.

It's a real finding from a real trial, and I'm not going to soften it. But it's an average from a specific study design, and the design is the interesting part, because it tells you exactly which lever you have. That's what this piece is about: not whether regain happens, but what the people who avoid the average had in place before their last dose.

A note on names, since the brands differ by country. Tirzepatide is Mounjaro in the UK, Ireland and Australia, and Zepbound in the US. Semaglutide injection is Wegovy in all four. I'll mostly use the generic names.

If you want the full picture of what physically happens when you stop, appetite, the half-life, the cardiometabolic markers, that's covered in quitting GLP-1s: what actually happens. This post picks up where that one leaves off and builds the plan.

What that study actually measured

The two-thirds figure comes from the STEP 1 trial extension. STEP 1 randomised 1,961 adults to semaglutide 2.4 mg or placebo for 68 weeks, alongside a lifestyle programme. When the trial ended, a representative subset of 327 people who'd completed it were followed for another year off treatment.

Here are the actual numbers, as the paper reports them. Over the 68 weeks, the semaglutide group lost an average of 17.3% of their body weight. In the year after treatment stopped, they regained 11.6 percentage points of that, and the group finished 5.6% below its starting weight. (Those three are separately rounded means, so they don't subtract neatly to the decimal.) The cardiometabolic improvements drifted back toward baseline alongside the weight (Wilding et al., 2022).

Read that carefully, because two details usually get lost. First, they didn't end up back where they started. A year on, with nothing at all in place, the group was still 5.6% lighter than at baseline. Second, and this is the detail that matters most for your planning: the lifestyle intervention was withdrawn at the same time as the drug. Everyone stopped everything, at once, by protocol.

That's a clean piece of science and a terrible model of how anyone would actually want to do this. Nobody is asking you to stop your medicine, your food tracking, your support and your exercise programme on the same Monday. The trial had to, to isolate the drug effect. You don't.

Three trials, one consistent finding

Two other withdrawal trials fill in the shape, and they agree with each other.

In STEP 4, 803 people first ran in on semaglutide for 20 weeks, losing an average of 10.6%. They were then randomised either to keep going or to switch to placebo, with the lifestyle programme continuing in both arms. Over the next 48 weeks, the continuing group lost a further 7.9%. The placebo group gained 6.9% (Rubino et al., 2021).

SURMOUNT-4 did the same thing with tirzepatide. After a 36-week lead-in that produced an average 20.9% loss, 670 people were randomised to continue or switch to placebo for 52 weeks. Those who continued lost a further 5.5%; those switched to placebo regained 14.0%. Nearly 90% of the continuing group still held at least 80% of their lead-in weight loss at the end, against 16.6% of the placebo group (Aronne et al., 2024).

The consistent message across all three is that these medicines are treatments for an ongoing condition rather than a course you complete. That's uncomfortable if you were hoping to be finished, and it's worth knowing before you plan around a different assumption.

It's also worth being precise about what these trials did not test. Every one compared the full dose against nothing. None of them tested a reduced maintenance dose, or a structured taper, or stopping the drug while keeping the lifestyle support running. Those are the options most people actually want, and the honest position is that the trial evidence for them is thin. That's a reason to plan carefully with your prescriber, not a reason to assume the worst.

An average is not a verdict

A group mean tells you where the middle of a distribution sits. It tells you nothing about where you'll land, and the spread in these trials was wide.

There's a separate body of evidence about people who lose weight and keep it off long term, and it predates GLP-1s entirely. Roughly 20% of people with overweight succeed at long-term maintenance, defined as losing at least 10% of body weight and holding it for a year or more (Wing & Phelan, 2005). The National Weight Control Registry tracks thousands of them, and the ten-year follow-up of 2,886 members found mean weight loss of 31.3 kg at entry and 23.1 kg still held a decade later, with more than 87% maintaining at least a 10% loss at both five and ten years (Thomas et al., 2014).

What separated the ones who regained from the ones who didn't is the useful part. Regain was associated with dropping leisure-time physical activity, dropping dietary restraint, weighing themselves less often, and a rising share of energy from fat. Note what's on that list: behaviours, all of them trackable, none of them requiring heroics.

Two honest caveats. Registry members are self-selected, so this describes what successful maintainers do, not proof that doing it makes you one. And they lost weight without a GLP-1, so the appetite they're managing may not be the appetite you'll be managing. Still, it's the best long-term evidence we have, and its conclusions line up with common sense.

The playbook: six things to have in place

The through-line here is timing. Every one of these is easier to build while the medicine is still working than after appetite has come back. The window before your last dose is the cheapest time to do the work.

1. Settle the actual question with your prescriber first: stop, or step down? People arrive at "coming off" for very different reasons: cost, side effects, a supply problem, pregnancy plans, or just wanting to know who they are without it. Those reasons point to different answers, and a lower maintenance dose is a real thing to ask about rather than an all-or-nothing choice. Go in with your reason stated plainly and let them weigh it up.

2. Set the protein floor before appetite returns. Protein is the most filling macronutrient, and it's the lever that protects lean mass while weight is changing (Cava et al., 2017). Muscle is also what keeps your energy expenditure up, so losing it during the weight-loss phase makes the maintenance phase harder. Work out your daily range with the protein target calculator and get used to hitting it while eating is still easy to control.

3. Start resistance training now, not when the weight moves. Two or three short sessions a week is the dose the preservation research is built on. The full protocol, and why it matters more on a GLP-1 than on an ordinary diet, is in protecting muscle on a GLP-1. If you want a routine that needs no equipment, there's one in bodyweight training on a GLP-1.

4. Rebuild the food environment while willpower is cheap. On a suppressed appetite, changing what's in your kitchen costs you almost nothing emotionally. Six months later it will cost a great deal. Sort the shopping list, the default breakfast, the thing you eat when you get in from work, and the snacks within arm's reach, now. Your surroundings do more of this work than your resolve does, which is the whole argument of the obesogenic environment piece.

5. Keep weighing, and judge the trend. Frequent self-weighing is one of the behaviours most consistently associated with holding a loss, and dropping it is associated with regain (Thomas et al., 2014). The point isn't the daily number, which is mostly water. It's that a trend line tells you about a drift of two kilos while it's still two kilos.

6. Expect the hunger, so it doesn't feel like a personal failure. After weight loss, appetite hormones stay shifted toward eating more, and they stayed shifted a full year later in the study that tracked them (Sumithran et al., 2011). On top of that, you're removing a drug that was actively suppressing appetite. Hunger returning is the expected outcome of both, on schedule, and it is not evidence that you've lost your discipline.

The coming-off planner puts these into a timeline you can take to an appointment, so the conversation with your prescriber starts from a plan rather than from "I think I want to stop."

What the first twelve weeks tend to look like

A rough shape rather than a promise, because individual experiences vary a lot.

Weeks 1 to 3. Semaglutide and tirzepatide both have half-lives around five to seven days, so the drug is still meaningfully present. Most people notice little. This is the window to make sure the routines are habits rather than intentions.

Weeks 4 to 8. Appetite generally comes back over this stretch, often described as food becoming interesting again rather than as raging hunger. Portion sizes creep first, snacking second. Weight may be flat or drifting slightly up.

Weeks 8 to 12. Whatever the new normal is going to be usually becomes visible here. This is the point at which a trend line earns its keep, because it distinguishes a settling of a couple of kilos of water and food volume from an actual upward drift.

Any gastrointestinal side effects you had should ease as the drug clears. Anything new or severe is a reason to contact your prescriber rather than to wait and see.

Choosing your restart line in advance

This is the single most useful thing in the whole piece, and it takes about a minute.

Before you stop, decide with your prescriber what would make restarting the right call. A weight, or a percentage regained, or a number of weeks of an upward trend. Write it down.

The reason is straightforward. If you don't choose a line while you're calm, you'll choose one while you're upset, and the decision made at 6am on a bad scale morning is rarely the one you'd have made with a clear head. A pre-agreed line turns restarting from an admission of failure into a step in a plan you wrote yourself.

Restarting, if it happens, usually means going back to a low dose and titrating up again, which is a prescriber decision. Stop-start cycles are common in real life and are not a mark against you.

The questions people ask

Will I definitely regain two-thirds?

No. That was the average in one extension study of 327 people who stopped their medication and their lifestyle programme simultaneously, by design. The individual spread was wide, and the group was still 5.6% below its starting weight a year later.

Does tapering the dose down prevent regain?

We don't have good trial evidence either way, because the withdrawal trials compared a full dose against placebo rather than testing a taper. It's a reasonable thing to discuss with your prescriber, and it shouldn't be presented to you, or by you, as a proven strategy.

Can I just eat less to hold the weight?

You can, and some people do. It's harder than it sounds because you're doing it against shifted appetite hormones and a lower energy requirement at the same time. That's why the playbook leans on protein, resistance training and environment rather than on restriction, which tends to cost muscle and make the arithmetic worse.

How long do I have to keep this up?

Honestly, indefinitely, in the same sense that any chronic condition needs ongoing management. One encouraging finding: maintenance appears to get easier with time, and people who have held a loss for two to five years have a substantially better chance of holding it longer (Wing & Phelan, 2005).

This is information, not medical advice. Stopping, tapering, switching to a maintenance dose and restarting are all decisions for the prescriber who knows your history. Don't change or stop a prescribed medicine on the strength of an article.

If you're thinking about coming off, the coming-off planner builds the timeline and the checklist, and the protein target calculator gives you the daily number to hit. Both are free and neither needs an account.