Pooled trial data put weight regain after stopping semaglutide or tirzepatide near 1 kg a month

Weight comes back after people stop taking semaglutide or tirzepatide. A re-analysis published in Endocrinology, Diabetes and Metabolism has put a figure on how fast it happens across the trials that measured it: about 1.04 kilograms a month, which is roughly 2.3 pounds.

The four authors, working at IMU University and Monash University in Malaysia, the University of Huddersfield in the UK, and Shah Alam Hospital, did not run a new trial. They collected the published follow-up numbers from studies that had already tracked people after the drug was stopped, then rebuilt the average path using a Bayesian model. That is a statistical approach that reports a range of plausible values, called a credible interval, rather than a single point.

Six trials, 1,776 people, and no more than a year of real follow-up

The pooled set was six studies covering 10 treatment arms and 1,776 participants. Three studies with 634 participants involved semaglutide: STEP 4, STEP 10, and the STEP 1 extension. Three studies with 1,142 participants involved tirzepatide: SURMOUNT-4, the SURMOUNT-1 extension, and the SURPASS-1 off-treatment follow-up.

Observed follow-up after the drug stopped ranged from 4 weeks in one trial to 52 weeks in others, giving 53 measurement points in total. The people in these trials were mostly in their late forties, started in the range of a BMI near 32 to 40, and most did not have diabetes, though the SURPASS-1 group did.

About a kilogram a month came back

At the moment treatment stopped, participants had lost an estimated 15.35 kilograms, or about 34 pounds, with a credible interval of 11.18 to 19.60 kilograms. After that, the model estimated regain of 1.04 kilograms a month, credible interval 0.80 to 1.29.

At that rate, half of the weight that had been lost was projected to return by 7.50 months, with a credible interval of 5.05 to 10.57 months. The authors' own conclusion states the range more loosely, at roughly 7 to 9 months.

Tirzepatide showed numerically faster regain than semaglutide when the numbers were compared without adjustment, but once other factors were accounted for, no clear drug-specific difference held up. People who had lost the most weight showed the strongest signal toward faster regain in absolute terms, though the credible interval for that association included zero, meaning the data cannot rule out no effect at all. Continued behavioral or lifestyle support pointed toward slower regain, but that estimate was imprecise too.

The return-to-baseline figure goes past where the data stops

The headline that will travel furthest is that participants were projected to return to their starting weight by 15.00 months, credible interval 10.10 to 21.13. That number deserves care. No trial in the set followed anyone past 52 weeks after stopping. The 15-month figure comes from assuming the 1.04 kilograms a month continues at a steady rate beyond the last measurement, and the authors say plainly that it should be read as extrapolation and not as an observed outcome.

Real weight trajectories usually flatten rather than continue in a straight line, so a constant-rate assumption may overstate how far regain runs. The paper does not claim otherwise. It says more flexible models will become useful once longer follow-up exists.

What else limits the estimate

The analysis worked from group averages published in papers, not from individual patient records, which restricts what can be tested. Six studies and 10 arms is a small evidence base for this kind of modelling, and the comparison between drugs was underpowered and described by the authors as exploratory.

The trials also stopped the drug on a protocol. In some, participants moved to placebo and kept receiving lifestyle counselling. The amount of support after stopping differed between studies, which the authors list as a limitation, and it is not the same as a person stopping because of cost, side effects, or a lapse in insurance coverage.

The six source trials were run by the companies that make these drugs, Novo Nordisk for the STEP studies and Eli Lilly for the SURMOUNT and SURPASS studies. The re-analysis itself reports no funding and its authors declare no conflicts of interest.

What the work supports is narrow and practical: regain after stopping is fast enough to matter within months, not years, and it is worth planning for before treatment ends rather than after. Decisions about starting, continuing, or stopping any of these medicines belong with a person's own clinician.