When you lose weight on a GLP-1 drug, part of what you lose is not fat. In the tirzepatide trial SURMOUNT-1, body scans showed that about 75% of the weight lost was fat and about 25% was lean mass. People on placebo, who lost much less, lost fat and lean mass in the same proportion. For semaglutide, the STEP 1 scans showed roughly 40% lean mass. Here is what those numbers mean, why lean mass is not the same as muscle, and what scans cannot tell you.
Both drug labels sum it up in one line. Wegovy: "Semaglutide lowers body weight with greater fat mass loss than lean mass loss" (Wegovy label, 12.2). Zepbound says the same for tirzepatide (Zepbound label, 12.2). That is true, but some lean mass goes with it.
Most of these numbers come from DXA (dual-energy X-ray absorptiometry), a low-dose whole-body scan that splits the body into fat, bone mineral, and everything else. "Everything else" is called lean mass, or lean soft tissue.
The SURMOUNT-1 body composition paper spells out the terms (Look et al., 2025):
So lean mass includes skeletal muscle, but it also includes organs, connective tissue and body water. A drop in lean mass is not all muscle loss. The nutrition advisory discussed below says muscle makes up about half of lean body mass (joint nutrition advisory, 2025).
Some lean loss is expected. The SURMOUNT-1 authors note that people gain some lean mass as they gain weight, since muscles carry a heavier body, and lose some of it again with large weight loss (Look et al., 2025).
SURMOUNT-1 enrolled 2,539 adults with obesity or overweight, without diabetes, and gave them weekly tirzepatide 5 mg, 10 mg or 15 mg, or placebo, for 72 weeks (Jastreboff et al., NEJM 2022). The main paper reports weight, not body composition. Body composition comes from a substudy of 160 of those participants (124 on tirzepatide, 36 on placebo) who had a DXA scan at the start and at week 72 (Look et al., 2025). They were 73% female, with an average weight of 102.5 kg.
| Week 72 change (DXA substudy) | Tirzepatide (pooled 5, 10, 15 mg) | Placebo |
|---|---|---|
| Body weight | -21.3% | -5.3% |
| Fat mass | -33.9% (-15.9 kg) | -8.2% (-3.6 kg) |
| Lean mass | -10.9% (-5.6 kg) | -2.6% (-1.2 kg) |
| Visceral fat (estimated) | -40.1% | -7.3% |
| Share of loss that was fat / lean | 74% / 26% | 75% / 25% |
On tirzepatide, 74% of the weight lost was fat and 26% was lean mass. On placebo it was 75% and 25%. The drug made people lose far more weight without changing the mix. The split was similar at each dose (75%, 72% and 75% fat at 5, 10 and 15 mg), in women and men, and across age groups. In people aged 65 and older on tirzepatide, 76% of the loss was fat.
In absolute terms, the average person on tirzepatide lost 5.6 kg of lean mass from a starting lean mass of about 50.8 kg. This study did not test strength or function directly.
Caveats: the subgroups are small; the study was funded by Eli Lilly, which makes tirzepatide; and the trial gave diet and activity advice (a 500 kcal a day deficit, at least 150 minutes of activity a week) but no strength training program.
STEP 1 enrolled 1,961 adults and compared weekly semaglutide 2.4 mg with placebo for 68 weeks. Weight fell 14.9% with semaglutide and 2.4% with placebo (Wilding et al., NEJM 2021). A subgroup with a BMI of 40 or less had DXA scans, and fat mass fell more with semaglutide than with placebo (O'Neil and Rubino, 2022).
We could not open the full STEP 1 paper, so here is how two later open papers describe its scan results. The SURMOUNT-1 body composition paper says STEP 1 reported 10.4 kg of fat and 6.9 kg of lean mass lost, about 60% fat and 40% lean (Look et al., 2025). The joint nutrition advisory gives 8.3 kg of fat (62%) and 5.3 kg of lean mass (38%) out of a 13.6 kg loss (joint nutrition advisory, 2025). The kilogram figures differ between the two summaries, but both put lean mass near 40% of the loss.
Do not read that as "semaglutide costs more muscle than tirzepatide." These were different trials with different scanners and people, and no trial has scanned both drugs head to head.
A meta-analysis of 20 randomized trials with 15,782 participants found lean mass made up 25% to 39% of weight lost on these drugs (Eisa and Barood, 2026). Lifestyle programs that lost weight without drugs showed about the same share. Lifestyle programs that added resistance training did best:
| Source | Share of weight lost that was lean mass |
|---|---|
| SURMOUNT-1 DXA substudy, tirzepatide | 26% |
| STEP 1 DXA substudy, semaglutide (as summarized by later papers) | about 38% to 40% |
| Meta-analysis: semaglutide trials | 35.2% |
| Meta-analysis: tirzepatide trials | 25.4% |
| Meta-analysis: liraglutide trials | 26.8% |
| Meta-analysis: lifestyle programs | 26.2% |
| Meta-analysis: lifestyle plus resistance training | 17.5% |
A systematic review in the Annals of Internal Medicine looked at 35 trials using DXA, bioimpedance, CT or MRI (Batsis et al., 2026). Across drug groups, the median share of weight lost from muscle-related measures was 28.3%, and about two thirds of drug groups were above the review's preset benchmarks (about 25% of weight lost for DXA or bioimpedance studies). No study in it measured physical function objectively, such as with strength or walking tests.
The honest summary: GLP-1 drugs do not appear to cost lean mass out of proportion to other ways of losing a lot of weight, but the share is not small, and little is known about long-term effects on strength.
DXA is the standard tool in these trials, but it has blind spots worth knowing before you book one:
The joint advisory from four nutrition and obesity groups lists people at higher risk of muscle loss when appetite and protein intake fall: older adults, people who are perimenopausal or menopausal, people with lower testosterone, and people who are sedentary or do no strength training (joint nutrition advisory, 2025). The same advisory recommends that the starting exam include muscle strength, function and body composition. If you are in one of those groups, that is a fair thing to ask for.
We found one detailed published guidance document on this: the 2025 joint advisory from the American College of Lifestyle Medicine, the American Society for Nutrition, the Obesity Medicine Association and The Obesity Society (joint nutrition advisory, 2025). What it says:
These are expert recommendations, not tested GLP-1 trial results. One trial testing this directly, LEAN-PREP, plans to randomize 232 adults starting semaglutide or tirzepatide to resistance exercise, extra protein, both, or neither for 6 months. Only its protocol has been published so far (Alawadhi et al., 2026). Your needs depend on your kidneys, your other conditions and your current diet, so talk to your prescriber or a registered dietitian before you change your protein intake or start a new training plan.
A bathroom scale cannot show lean mass, but a steady record helps. Log doses and dose changes, weigh in at the same time of day, and note side effects like nausea that cut how much you eat. Dose Track keeps GLP-1 doses, weigh-ins and side effects in one place, along with any other compounds you take. See how weight tracking works and the GLP-1 tracking guide.
See also the tirzepatide and semaglutide pages, the semaglutide vs tirzepatide comparison, and the glossary entry for GLP-1 receptor agonist. If your weight has stalled, read why GLP-1 weight loss levels off.
This article is general information, not medical advice. Talk to your prescriber before you start, change or stop any medicine, diet or exercise program.