Meta-regression maps the shape of weight regain after stopping GLP-1 drugs

A systematic review and nonlinear meta-regression published March 4, 2026 in eClinicalMedicine has produced what is now the clearest single estimate of what happens to body weight after people stop taking GLP-1 receptor agonists. Pooling 48 studies and fitting a regain curve using data from six randomized trials covering 3,236 participants, the authors found that one year after stopping treatment, roughly 60% of the weight lost during treatment had come back ^[1].

The regain did not follow a straight line. It followed an exponential curve with a half-life of about 23 weeks, meaning the fastest regain happens in the first several months after the last dose and then slows ^[1]. The model projected that the curve levels off at 75.3% of the weight originally lost — in other words, on average, about a quarter of the treatment-related weight loss was still retained when the curve plateaued ^[1].

That shape matters as much as the headline number. A half-life of roughly 23 weeks implies that about half of the eventual regain occurs in the first five to six months off the drug, with the remainder accumulating more slowly over the following year and beyond ^[1]. Averages, though, are not individual outcomes. A meta-regression describes the group trajectory across trials; it does not predict what any one person's weight will do, and the source material available does not report how widely individual results varied around that curve.

It is also worth being precise about what the analysis measured. The 60% figure is a share of the weight lost during treatment, not a share of total body weight ^[1]. Someone who lost 40 pounds on treatment and followed the average curve would be expected to regain roughly 24 pounds in the first year off the drug, and to end up around 10 pounds below their starting point once the curve flattened, based on the modeled plateau of 75.3% ^[1].

Why it matters for patients

GLP-1 medicines — semaglutide (Ozempic, Wegovy, Rybelsus) and tirzepatide (Mounjaro, Zepbound) among them — are increasingly framed by clinicians as long-term treatments rather than short courses, and this analysis puts numbers behind that framing ^[1]. Stopping is common in the real world for reasons that have nothing to do with the medicine working: insurance coverage ends, supply runs short, side effects become intolerable, cost becomes unmanageable, or a person reaches a goal weight and wants to stop.

For anyone in that position, the practical value of this paper is that it converts a vague warning — "you may regain the weight" — into a measurable expectation with a timeline. The regain is steep early and slows later, and on average it is substantial but not complete ^[1]. That has implications for how people and their clinicians plan around a pause or a stop, how employers and insurers model the value of covering these drugs, and how anyone interprets before-and-after results seen online.

It also reframes what "maintenance" means. If the group-level curve bottoms out near 75% regain, then the portion of benefit that persists without ongoing treatment is real but modest on average ^[1]. Decisions about whether and how to continue treatment belong between a patient and their prescriber.

What is not yet known

The available source material does not break down whether regain differed by drug, by dose at the time of stopping, by how much weight a person lost, or by whether people continued structured diet and exercise support after stopping ^[1]. It also does not report what happened to blood pressure, blood sugar, or cholesterol during regain, or whether restarting a GLP-1 after a gap restores the earlier level of weight loss. Those questions would require either the full text of the paper or additional studies. Until then, the 60%-at-one-year and 75.3% plateau figures stand as the best available summary estimate of what stopping does ^[1].