Yes, and the tissue that comes back is not the tissue that left. A review of body composition during and after incretin therapy cites lean mass losses averaging 6 to 7 kg. Bone mineral density falls alongside it. After stopping, regain averages 5.6 kg within a year, and the review describes that regain as fat-preferential. Body-composition findings also depend on the clinical population. The semaglutide studies in people living with HIV address a distinct safety and fat-distribution question.
So the scale can read the same at both ends of a course while the body underneath has changed. That is the finding, and it is not the one a weight chart shows.
Several trials have tested a second drug to protect muscle. In 102 adults on tirzepatide, randomizing to add apitegromab held 1.9 kg more lean mass at 24 weeks [2]. Total weight loss was similar in both arms, so the lean mass was kept without blunting the loss.
Bimagrumab is the other candidate. It blocks the activin type II receptor, a signal that limits muscle growth. Given alone to 75 adults with type 2 diabetes, it added 1.70 kg of lean mass over 48 weeks [4]. Lean mass on placebo fell 0.4 kg. No GLP-1 was used in that trial.
A later trial paired it with semaglutide in 507 adults with obesity [5]. At 48 weeks, weight fell 17.8 kg on the high-dose pair. It fell 14.2 kg on semaglutide 2.4 mg alone and 3.3 kg on placebo. The abstract reports weight as its endpoint. Lean mass is not among the results it gives.
Two recent cohorts measured how the weight lost splits. In 35 adults on tirzepatide, fat made up 85.7 percent of the loss [6]. Skeletal muscle fell 2.8 kg, or 8.3 percent. That clinic measured body composition by bioelectrical impedance. A second cohort matched 396 GLP-1 users to nonusers from connected-device data [7]. Median weight fell 9.8 percent. Fat mass fell 7.5 percent and muscle mass 2.3 percent. Neither cohort was randomized.
Bone runs the other way from what the review's density figures suggest. Pooled across twenty-five randomized trials, bone mineral density was higher on a GLP-1 at every site measured [3]. Meanwhile whether fractures fell is a question those trials cannot answer.
Weight coming back after stopping is well established and covered elsewhere on this site. This review is about something narrower and less discussed: what the returning weight is made of.
What it gathers
A narrative review searching three databases from 2005 to mid-2026, assembled under a standard framework for this kind of review. [1] Every number in it is drawn from other studies rather than measured. A narrative review chooses what to include. These are figures it found, presented as it read them.
On treatment, lean mass fell by 6 to 7 kg on average, alongside reductions in bone mineral density. After stopping, weight rose by 5.6 kg on average within one year, described as fat-preferential, with observational reports of dose-dependent cardiometabolic worsening.
The persistence figure
The review cites 85% of patients discontinuing within the second year of real-world use. That is the premise of the whole paper. If most people stop, what happens after stopping is most of the story.
That figure comes from the literature the review surveys, not its own data. Persistence estimates vary with how long a gap counts as stopping. The direction is not in dispute. Do not quote the exact number as though it were measured here, any more than a multiplier quoted without its base should.
What it recommends, and what this page will not
Structured exercise. Exercise and pharmacotherapy produce different body composition outcomes. Exercise preserves lean mass where the drug does not.
The review offers prescription guidance. This page does not, because what is appropriate depends on age, joints, cardiac fitness and what somebody is already doing. What is safe to say is that the cessation transition has no established framework. The authors say so. Nobody selling a subscription is planning one, and it is not among the things sellers publish.