What People Actually Do After Stopping

Everyone knows that stopping a GLP-1 means the weight comes back. A large study of what actually happened to people who stopped semaglutide or tirzepatide in ordinary care found the average weight change over the following year was 0.5%.^[1] Not 50% of the loss. Half of one percent. The explanation is not that the trials were wrong.

What the study found

The diabetes group kept losing. The obesity group gained back a fraction of a percent. Neither looks anything like the rebound described in trials of stopping.

Why this does not contradict the trials

It looks like a contradiction and it is not. In a withdrawal trial, people randomized to placebo receive nothing else — that is the design, and it is what makes it a clean test of the drug. In SURMOUNT-4, people switched to placebo regained about 14% within a year, which we covered in the tirzepatide timeline.

Real patients do not receive nothing. In this study, 19.6% restarted the same medication and 35.2% received an alternative obesity treatment — 27.4% another drug, 13.7% a lifestyle modification visit, and 0.6% metabolic or bariatric surgery.^[1]

More than half of the people who stopped did something else. The trial arm that regains everything is the arm where nothing else happens.

So the two findings answer different questions. The trials tell you what the drug was doing. This tells you what happens to people, in a system where stopping is usually a transition rather than an ending.

A second study, and the flaw it shares

An independent 2026 analysis followed 4,182 patients for six months after their last recorded semaglutide or tirzepatide prescription and found much the same thing: roughly two-thirds had stable weight or were still losing.^[2] Two datasets reaching that answer is more than one.

That matters because someone who quietly kept taking the drug will of course not regain weight, and will be counted as a success of stopping. It is the single largest reason to hold any “people do fine after stopping” finding loosely.

To the authors’ credit, the sub-analysis answers its own objection as far as it can. Among the 119 whose records confirm they stopped, the pattern held: 72% did not regain weight. That is a much smaller number of people, and it is the number that actually addresses the question.

One further observation from the same work, which fits everything above: exercise counseling appeared in the notes of 26.2% of those with durable weight loss against 14.7% of those who regained (P = .04). Read that carefully — it measures what a clinician wrote down, not what a patient did, and documented counseling is a marker of engaged care as much as of exercise. It is consistent with the theme that what happens after stopping matters more than the stopping.

The sentence that limits all of it

The authors write that there was considerable individual-level variability, and that phrase carries more weight than the averages above.^[1] A mean change of +0.5% is entirely compatible with a substantial number of people regaining a great deal while others continued to lose.

Averages of this kind describe a population, not a trajectory. If you are deciding whether you can afford to stop, the honest answer is that the average is reassuring and tells you very little about your own case — and that what you do next appears to matter more than the stopping itself.

What to take from it

• Stopping is not a cliff for most people, at least not within a year, and the catastrophic framing is overstated — though a separate argument holds that the cardiovascular cost may not appear in a weight number at all, which we cover in the cardiovascular cost of stopping.
• What you do next is the variable. Restarting or switching was common and appears to be doing much of the work.
• Diabetes and obesity are different situations. Those treating diabetes kept losing, usually because they moved to another agent.
• Your own result may be nothing like the mean. The authors say so explicitly, and it is the most important sentence in the paper.