How much muscle do you lose on a GLP-1?

In the SURMOUNT-1 DXA sub-study (tirzepatide), participants lost 21.3% body weight over 72 weeks — 33.9% of fat mass and 10.9% of lean mass. Roughly three-quarters of the weight lost came from fat and about one-quarter from lean tissue, which is comparable to what's seen with diet-driven weight loss of similar magnitude.

Do you regain the muscle back after stopping the medication?

This is the honest answer: nobody has measured it in a published trial. There is no DXA body composition data from the weight-regain phase of STEP-1 or SURMOUNT-4. Physiologic reasoning suggests regained weight skews disproportionately toward fat, but this hasn't been directly verified.

Why would regained weight be mostly fat instead of muscle?

Muscle takes deliberate stimulus (resistance activity) and adequate protein to rebuild — neither happens automatically as appetite returns and calorie intake rises. Fat, by contrast, is what excess calories are stored as by default. Without a specific muscle-preserving effort, the composition of regain tends to favor fat.

Does this mean stopping GLP-1s leaves you with worse body composition than before you started?

Potentially — this is the specific concern that's driven the current research interest. If someone lost 25% of a large weight loss as lean tissue, then regained most of the weight as fat, their fat-to-lean ratio at their new weight could be worse than at their starting weight. This is a plausibility argument based on physiology, not a measured finding, but it's the reason muscle preservation during and after treatment is emphasized.

What actually preserves muscle during and after GLP-1 use?

The two levers with the strongest support are adequate protein intake (commonly 0.7–1 g per pound of goal body weight) and regular resistance activity (two or more sessions per week). A large trial specifically testing this combination on GLP-1s (LEAN-PREP) is currently underway, so this is still officially in the 'under investigation' category — but the reasoning is sound and the downside is minimal.

Is muscle loss worse with tirzepatide than with semaglutide?

The head-to-head evidence is limited, but the proportion of weight lost as lean mass appears broadly similar across GLP-1s in the DXA sub-studies. What differs is the total weight loss — because tirzepatide typically produces larger total loss than semaglutide, the absolute muscle loss is often larger in tirzepatide trials, but the percentage of loss coming from lean tissue is comparable.

Stopping GLP-1: Why Regain Is Fat, Not Muscle

Written by the GLP-1 Coach editorial team. Last updated August 13, 2026.

If you've spent time in GLP-1 communities, you've almost certainly seen the concern raised: "you regain fat, not muscle." It's compelling, it's memorable, and it's largely correct in the way it's meant — but it's also more precisely characterized as a physiologically well-reasoned inference that hasn't yet been directly measured.

Here's what the actual body composition data shows, what it doesn't yet cover, and why the asymmetry matters enough that muscle preservation has become the durable-outcome question at the center of modern GLP-1 practice.

What does the on-medication body composition data actually show?

The clearest body composition data from a GLP-1 trial comes from the DXA sub-study of SURMOUNT-1, published in Diabetes, Obesity and Metabolism in 2025. Participants on tirzepatide 10 or 15 mg were assessed at week 72:

Measure	Change from baseline
Body weight	−21.3%
Fat mass	−33.9%
Lean mass	−10.9%
Visceral adipose tissue	−33.6%

Doing the math on the composition of total weight lost: roughly three-quarters from fat, one-quarter from lean tissue. That ratio — about 75/25 — is broadly in line with what's seen with diet-driven weight loss of similar magnitude and is not an unusually muscle-hostile pattern.

The STEP-1 DXA sub-study for semaglutide showed a similar directional pattern: total fat mass and visceral fat fell substantially, absolute lean mass decreased in proportion to total weight loss, and the ratio of lean-to-total body mass actually improved because fat loss outpaced lean loss.

The broader GLP-1 literature places lean-tissue loss at roughly 26–40% of total weight lost across trials — meaningfully higher than what deliberate protein-plus-resistance-training programs typically achieve (which can push lean loss down closer to 10–15%), but not the "half your weight loss is muscle" narrative that some social media coverage suggests.

So what's the actual concern, if the on-drug numbers aren't alarming?

The concern isn't primarily what happens on the medication. It's what happens after.

Here's the gap in the evidence base: no published clinical trial has measured DXA body composition during the weight-regain phase after GLP-1 cessation. The STEP-1 extension tracked weight regain over the year following semaglutide cessation — roughly two-thirds of the lost weight returned, as detailed in what actually happens when you stop Wegovy. The SURMOUNT-4 trial showed a ~20 percentage point weight swing when tirzepatide was replaced with placebo, covered in what happens when you stop Mounjaro or Zepbound. Neither study reported the composition of that regained weight.

That leaves a specific and important question unanswered: when the weight comes back, what tissue is it coming back as?

Why the physiologic reasoning points toward asymmetric regain

Muscle and fat regain do not follow the same rules.

Fat regain is largely automatic. Any sustained calorie intake above maintenance is stored, with fat as the default depot. As appetite returns after cessation and daily intake climbs back toward pre-treatment volumes, the caloric surplus needed to regain weight arrives without any behavioral effort required.

Muscle regain is not automatic. Rebuilding lean tissue requires two things simultaneously: adequate protein intake (protein synthesis needs substrate) and adequate mechanical stimulus (resistance activity signals the tissue to build). Neither happens by default when appetite returns. If a person's post-cessation life includes higher calorie intake but no deliberate protein target and no resistance training, their fat stores will rebuild while their lean tissue will not.

The mechanistic prediction is straightforward: without a specific muscle-preserving effort, the composition of regain skews toward fat. Combined with the lean mass lost during active treatment, this means someone's fat-to-lean ratio at their new post-regain weight could plausibly be worse than at their original starting weight — even if the scale reads similar.

This is the specific mechanism behind the "sarcopenic obesity" concern that has driven recent obesity medicine research interest in GLP-1 body composition outcomes.

What has been directly measured, and what hasn't

For clarity on where the evidence actually stands:

Question	Evidence status
How much lean mass is lost during GLP-1 treatment?	Measured in SURMOUNT-1 DXA sub-study, STEP-1 DXA sub-study, and others. Roughly 25% of total weight lost on average.
How much weight is regained after cessation?	Measured in STEP-1 extension (~2/3 regained over 1 year) and SURMOUNT-4 (~20 pp swing over 1 year).
What is the composition of that regained weight?	Not directly measured in any published trial. Inferred from physiology and prior weight-cycling literature.
Does resistance training + adequate protein preserve lean mass on GLP-1s?	Actively being tested — the LEAN-PREP RCT is underway but not yet reported. Currently supported by case series and physiological reasoning.
Does resistance training + adequate protein change the composition of regain after cessation?	Not directly measured. No published trial has assessed this.

Any content presenting the asymmetric regain pattern as a proven finding is going beyond what the data actually shows. Any content dismissing the concern because it hasn't been directly measured is ignoring a strong physiological argument. The honest position sits in between.

Why the concern still matters even without the direct measurement

Two reasons the asymmetric-regain concern is worth taking seriously despite the evidence gap:

First, the downside of assuming it's real is minimal. The steps you'd take to protect against it — hitting a protein target, doing resistance activity — are the same steps that have well-established benefits for general health, longevity, and metabolic function. There is no meaningful cost to acting as if the concern is validated even if some future study finds regain composition is less lopsided than physiology predicts.

Second, weight-cycling literature outside GLP-1s already shows composition asymmetry. Studies of weight loss followed by regain in non-GLP-1 contexts have repeatedly shown that regained weight tends to preferentially rebuild fat mass before lean mass, particularly in the absence of resistance training. There's no strong mechanistic reason to expect GLP-1 users to be an exception to that pattern — and the concern is that they might be particularly exposed because of the rapid weight-loss trajectory GLP-1s enable.

What actually helps preserve muscle through the transition

Two levers with strong physiological support and reasonable (though not conclusive) evidence:

Protein intake — commonly 0.7–1 g per pound of goal body weight per day. This is the reference range most obesity medicine and sports nutrition sources use for weight-loss and weight-maintenance phases. The upper end (~1 g/lb) is particularly emphasized during the appetite-return window, when calorie intake is rising and lean tissue is most at risk if protein isn't prioritized within that larger intake. Note that inadequate protein during active treatment — not just after stopping — can produce a plateau that looks like medication failure; the full explanation is in is your weight-loss plateau actually muscle loss in disguise?

Resistance activity — two or more sessions per week. The mechanical stimulus that signals lean tissue to build or maintain doesn't have to come from a gym or heavy weights. Bodyweight-based resistance work, resistance bands, or basic strength training equipment all provide adequate stimulus for most people. Consistency matters more than intensity for the preservation goal.

Neither of these is a heroic intervention. Both are well within reach for most people. The reason they're emphasized so consistently in GLP-1 discussions isn't that they're advanced techniques — it's that they're often deprioritized during the appetite-suppression phase (when eating enough protein feels harder than usual) and then never restarted once appetite returns.

The LEAN-PREP randomized controlled trial (protein 1.6 g/kg/day plus 3x/week resistance training on tirzepatide) is currently underway and will provide the first controlled data on whether this combination materially changes body composition outcomes on GLP-1s. Until it reports, the case rests on physiological reasoning and observational data — reasonable, but not definitive.

The bottom line

Bottom line: About 25% of the weight lost on GLP-1s comes from lean mass in DXA sub-studies — roughly in line with diet-driven loss, not unusually high. The unstudied question is what regains: no published trial has measured post-cessation body composition, but physiology strongly predicts regain skews toward fat, not muscle. That would leave a worse fat-to-lean ratio than at baseline. Protein (0.7–1 g/lb) and resistance activity twice weekly are the standard hedge against that asymmetry, even without conclusive trial evidence.

Related tool: our free post-therapy timeline calculator shows the week-by-week window where protein and resistance activity matter most.

Final Thoughts

The muscle-preservation conversation is where the GLP-1 literature is most obviously ahead of the marketing on both sides. The manufacturer materials tend to celebrate total weight loss without engaging with composition. The critics tend to overstate the muscle loss concern by treating an unmeasured phase as if it had already been measured badly.

The honest position is narrower and more actionable: the lean mass loss during treatment is real but not exceptional, the composition of post-cessation regain is genuinely unknown, and the interventions that most likely mitigate both concerns are the same interventions that have independent health benefits. There is no reasonable case for skipping protein and resistance activity while on treatment or during the transition off — the downside is nonexistent, the physiologic upside is well-supported, and the trial that will formally confirm the effect is coming.

This site's tools focus on the protein and consistency side of the equation because that's the lever most likely to matter over the arc of a GLP-1 journey — during treatment, through the appetite-return window, and into whatever "maintenance" looks like once the drug is no longer doing the work.

Related:

This article is educational and not a substitute for medical advice. Decisions about GLP-1 medication, protein targets, and exercise programs should be made in the context of your personal health situation with your prescriber and, where appropriate, a registered dietitian.

Sources: Look et al., SURMOUNT-1 DXA body composition sub-study, Diabetes, Obesity and Metabolism 2025; Wilding et al., STEP 1 primary trial, NEJM 2021; Wilding et al., STEP 1 extension, Diabetes, Obesity and Metabolism 2022; Aronne et al., SURMOUNT-4, JAMA 2023; LEAN-PREP RCT protocol, PMC 2025; Muscle health in the modern era of incretin therapies, PMC 2025.