Semaglutide Muscle Loss: Why the Published Numbers Disagree
Search semaglutide muscle loss and you get a confident percentage. Forty percent of the weight lost is muscle, says one page. Forty-five, says another. A third quotes 13.9%, which is not answering the same question. All weight loss draws on fat and lean tissue both, so the real question is how much, and measured how. On that, the evidence does not agree with itself.
Four Tier 1 studies exist, across four populations, two doses and three measurement methods, and none states the figure that gets quoted. This page covers what each reports and why they cannot be stacked. For background, start with the semaglutide guide.
Key Takeaways
- None of the four studies below states a "percent of weight lost that was muscle" figure. The circulating numbers are calculations, secondary citations, or a different measurement mislabelled as this one.
- STEP 1's substudy reports change within each tissue. Fat fell 19.3% of starting fat and lean mass 9.7% of starting lean over 68 weeks. Different denominators, not convertible into a share of weight lost.
- Lean mass improved relative to fat. In that substudy the lean-to-fat ratio rose more in people who lost at least 15% of body weight than in those who lost less.
- Volume fell without function measurably worsening. SLIM LIVER found psoas volume down 9.3% on MRI, while chair-rise time and gait speed showed no significant change. The authors concluded function was maintained, not improved.
- Real-world grip strength went up. In SEMALEAN, lean mass fell early then held steady, handgrip rose 4.1 kg by month 12, and sarcopenic obesity fell from 49% to 33%.
- Resistance training and adequate protein are the levers with evidence behind them, and neither is specific to semaglutide.
First three tiles: the semaglutide arm (n=95) of the STEP 1 body composition substudy of 140 adults with overweight or obesity without diabetes, 68 weeks (Wilding et al., J Endocr Soc, 2021). Fourth: SEMALEAN, n=106, real-world obesity cohort (Alissou et al., 2026). Different studies, not a series.
Why You Will See Wildly Different Semaglutide Muscle Loss Numbers
Take the roughly 40% first. No semaglutide trial states it, so we tried to reconstruct it. The STEP 1 abstract gives a mean starting weight of 98.4 kg, 43.4% fat and 53.9% lean. Apply its reported changes, 19.3% off fat, 9.7% off lean, 15.0% off body weight, and lean tissue works out at 34.9% of the kilograms lost, or 38.4% of fat and lean combined. Neither reaches 40%.
That is our calculation, not a trial finding, and we cannot show it is how anyone else reached 40%. Naming an origin we cannot evidence would repeat the error this page is about, so we say only that the semaglutide-specific figure has no stated source.
A published figure for this quantity does exist, though not for semaglutide alone. A 2026 Annals of Internal Medicine systematic review of 35 randomised trials by Batsis and colleagues found that across incretin drugs as a class, the median share of weight loss attributable to muscle-based indices was 28.3%, interquartile range 15.9% to 39.9%. Among the trials using DXA or bioelectrical impedance, the method STEP 1 used, it was about 29%, interquartile range 16.6% to 43.1%. So 40% is not unknowable, and not extreme against that spread either. It is simply a figure nobody in STEP 1 published, and one that does not reproduce from the abstract's own inputs.
The roughly 45% is traceable but unverifiable. It appears in a 2025 Cell Metabolism research letter by Karasawa and colleagues reporting a mouse study. That letter is reported to cite STEP 1 as lean mass down 6.92 kg against 15.3 kg of weight loss, a pairing we relay rather than having read: the full text is paywalled and we could not open it. The 15.3 kg matches the main STEP 1 trial of 1,961 participants. The 6.92 kg cannot, because STEP 1 measured body composition only in the 140-person substudy, so no lean mass figure exists for the full trial.
The 13.9% on consumer drug reference sites is a different quantity again. It pairs with a 6.9 kg lean loss, which reads to us as lean tissue against starting lean mass, the same kind of measure as the substudy's 9.7%, though we cannot confirm that is the denominator intended. It also does not reproduce from STEP 1: 6.9 kg against that substudy's 53.0 kg of baseline lean gives 13.0%, not 13.9%. It is not a share of weight lost, and setting it beside 40% and 45% is the conflation.
None of the three is a semaglutide muscle loss statistic in the way it gets used. Quoted without their context, all three read as trial findings. They are not, and nothing below uses any of them as a measurement.
These studies were never designed to be compared.
| Study | Population | Dose | Method | Comparator |
|---|---|---|---|---|
| STEP 1 substudy (n=140) | Overweight or obesity, no diabetes | 2.4 mg | DEXA | Placebo |
| SUSTAIN 8 substudy (n=178 scanned, 114 analysed) | Type 2 diabetes on metformin | 1.0 mg | DXA | Canagliflozin, no placebo |
| SLIM LIVER (n=51) | HIV with fatty liver disease | 1.0 mg | MRI, psoas muscle | None, single arm |
| SEMALEAN (n=106) | Obesity, real-world clinic cohort | 2.4 mg | DXA plus handgrip | None, observational |
What DEXA and DXA Actually Measure, and It Is Not Just Muscle
A body composition scan sorts tissue into fat mass, bone mineral, and lean soft tissue. That third bucket, shortened to lean mass, holds skeletal muscle along with body water, glycogen-bound fluid, organs, and connective tissue. Muscle is a large share of it, not all of it. Because glycogen binds water, an early drop in carbohydrate intake shrinks glycogen stores and pulls measured lean mass down without any contractile tissue being lost. That cuts both ways: nobody can claim the muscle loss was smaller than the reading either.
MRI answers a narrower question, the volume of one named muscle group. Handgrip, chair-rise timing, and gait speed answer a third: what the muscle can do. Three layers, three quantities. A decline in one does not imply a decline in another, and the sections below show muscle volume falling measurably while function did not.
Obesity Dose Versus Diabetes Dose
STEP 1 ran at 2.4 mg weekly in people with obesity and no diabetes. SUSTAIN 8 ran at 1.0 mg weekly in type 2 diabetes on metformin. Different dose, different metabolic state, different amount of weight loss, so the two sets of figures should never be averaged. The semaglutide dosing guide covers how the schedules diverge.
What the STEP 1 Body Composition Data Actually Reports
One caveat first. This analysis appeared as a conference abstract in the Journal of the Endocrine Society supplement in 2021 and has never become a full peer-reviewed paper. An abstract carries less methodological detail than a full report.
It scanned 140 participants at nine STEP 1 sites, 95 on semaglutide 2.4 mg and 45 on placebo, capped at a BMI of 40. Three quarters were women, mean starting weight 98.4 kg, mean BMI 34.8. Scans ran at screening and week 68.
| Change from baseline to week 68 | Semaglutide 2.4 mg |
|---|---|
| Body weight | -15.0% (placebo -3.6%) |
| Total fat mass | -19.3% |
| Visceral fat mass | -27.4% |
| Total lean body mass | -9.7% |
| Lean mass as a share of total body mass | +3.0 percentage points |
Source: Wilding et al., STEP 1 body composition exploratory analysis, J Endocr Soc, 2021 (conference abstract).
Read those two percentages carefully. The 19.3% and the 9.7% are each measured against that tissue's own starting amount. Fat shrank by about a fifth of the fat someone began with; lean tissue by about a tenth of the lean. Neither is a share of the kilograms that came off. Converting one framing into the other is the arithmetic that produced the circulating percentages, and doing it silently is how a derived estimate ends up quoted as a trial result.
The finding that rarely travels is the ratio. Lean-to-fat mass started at 1.34 and rose 0.23 by week 68, tilting composition toward lean tissue rather than away. Split by weight lost, the improvement was 0.41 among the 44 who lost at least 15% of body weight against 0.03 among the 39 who lost less. More weight lost went with a better ratio, and any account of semaglutide as muscle-wasting has to explain that direction.
Two limits. Those subgroups total 83 of the 95 in the semaglutide arm, observed data with no imputation. And the smaller-loss group's confidence interval runs from -0.05 to 0.12, so that 0.03 is indistinguishable from no change.
Model Your Semaglutide Protocol
See how semaglutide accumulates across your dose schedule before you discuss pace with your prescriber.
The Type 2 Diabetes Picture: SUSTAIN 8
The SUSTAIN 8 body composition substudy, published by McCrimmon and colleagues in Diabetologia in 2020, scanned 178 adults with uncontrolled type 2 diabetes on stable metformin, 88 assigned semaglutide 1.0 mg weekly and 90 canagliflozin 300 mg daily. Only 114 of them, 53 and 61, had observed week 52 data, so the figures below rest on 64% of those scanned.
Over 52 weeks, fat mass fell 3.4 kg on semaglutide against 2.6 kg on canagliflozin, and lean mass 2.3 kg against 1.5 kg. Lean mass as a proportion of body mass rose in both arms, by 1.2 and 1.1 percentage points.
Those two columns are not different from each other. The estimated treatment difference was -0.79 kg for fat (95% CI -2.10 to 0.51) and -0.78 kg for lean (-1.61 to 0.04). Both cross zero, and the authors concluded the changes were not significantly different. Reading 2.3 against 1.5 as semaglutide costing more lean tissue reads a gap the statistics do not support.
There was no placebo arm. Canagliflozin is an SGLT2 inhibitor that itself causes weight and fluid loss, so this design compares two active drugs. The authors noted numerical improvements in both arms and called the specific impact of either treatment, absent a placebo arm, speculative. Neither arm's change can be attributed to its drug alone, with no untreated group showing what would have happened anyway.
These are kilograms in type 2 diabetes at 1.0 mg. The STEP 1 figures are percentages in obesity without diabetes at 2.4 mg. The two sets share a molecule and nothing else.
Muscle Volume Versus Muscle Function: the SLIM LIVER Findings
SLIM LIVER, reported by Ditzenberger and colleagues in Clinical Infectious Diseases, is the smallest and most specific study here: 51 people living with HIV who also had metabolic dysfunction-associated steatotic liver disease, all on semaglutide 1.0 mg, no comparator group. Muscle measures covered 46 of them, baseline to week 24, using MRI of the psoas rather than a whole-body scan.
Psoas volume fell 9.3% over 24 weeks (95% CI -13.4 to -5.2), an absolute decrease of 1.49 mL. Fat inside the muscle did not change significantly, down 0.42%, p = 0.16.
The function tests did not follow it down. Ten-time chair-rise time and gait speed both moved favourably, by 1.27 seconds and 0.05 metres per second, but neither reached significance (P = .069 and P = .078), and the study was not powered to detect a change in physical function. The one function measure that did reach significance was slow-gait prevalence, down from 63% to 46% (P = .029).
The authors concluded that despite the decrease in psoas volume there was no significant change in physical function, suggesting function was maintained. Maintained is the supported claim. Improved is not.
What this does and does not establish. Two statistically silent tests in an underpowered single-arm study mean the question went unanswered, not answered in the negative. What it shows is that a volume number alone cannot tell you whether someone got weaker: the volume loss was significant, no function decline detectable. Fifty-one people, a niche HIV and fatty liver population, 1.0 mg, no control group, no power.
Real-World Data: the SEMALEAN Study
SEMALEAN, published by Alissou and colleagues in Diabetes, Obesity and Metabolism in 2026, followed a real-world clinic cohort rather than a trial. Of 115 people who started treatment, 106 completed 12 months on semaglutide 2.4 mg weekly. Nearly 69% were women, mean BMI 46.3. Measurements combined DXA with handgrip strength.
The entry criterion makes this cohort nearly the inverse of the STEP 1 one. SEMALEAN required grade 3 obesity, a BMI of 40 or above; the STEP 1 substudy capped eligibility at 40 or below. The two groups overlap at a single point.
Fat mass fell 14.3% by month 7 and 18.9% by month 12. Lean mass fell 3.0 kg by month 7 then stopped falling, holding steady through month 12 while fat loss continued. Handgrip rose 3.7 kg at month 7 and 4.1 kg over the year. Sarcopenic obesity, low muscle mass combined with obesity, fell from 49% at baseline to 33% at 12 months.
Note on the grip figure: the paper's results section reports +4.1 kg at 12 months while its own abstract reports +4.5 kg. We use the results value.
Observational, no comparator arm, and filtered by who stayed on the drug. The protocol excluded anyone stopping before month 7 or with incomplete follow-up; nine were removed, seven for gastrointestinal side effects. So both findings leaned on here, the plateau and the grip gain, are measured only in people who tolerated a year of treatment. Everyone was also under clinical care and may have changed diet and activity, so the grip gain cannot be attributed to the drug.
What Actually Protects Muscle During Semaglutide Treatment
Resistance Training
Progressive resistance training has the most evidence behind it, none semaglutide-specific. A 2024 Diabetes Care review by Locatelli and colleagues examined resistance exercise against body composition on incretin-based drugs as a class. On duration it is specific: supervised programmes running longer than 10 weeks produced large gains in lean mass and strength. It prescribes no weekly session count.
Two to three sessions a week covering the major muscle groups is the usual starting point, from general public activity guidance rather than that review or any semaglutide trial. Turning it into a programme is a conversation for a prescriber or qualified trainer.
Protein Intake
Adequate protein supports lean tissue retention in a calorie deficit, with roughly 1.2 to 1.6 grams per kilogram of body weight daily the range generally cited. Semaglutide suppresses appetite, which is the point of the drug and also what makes any protein target harder to hit. If nausea is eating into intake, the semaglutide side effects guide covers the pattern.
Titration Pace
This one is inference, not a finding. Less weight lost means fewer kilograms of lean tissue lost, which is the argument for slower escalation. No trial has tested titration speed against body composition, so treat it as a discussion point, not a protocol. See the semaglutide weight loss data for how total loss accumulates.
How Semaglutide Muscle Loss Compares to Tirzepatide and Retatrutide
No head-to-head body composition trial has compared semaglutide against tirzepatide or retatrutide. That is an absence, not a tie. Every number comes from a separate trial with its own population, comparator, dose and scanning protocol, so lining them up implies a comparability the designs do not support.
What exists is class-level context. The Locatelli review characterises incretin-based weight loss drugs as a group, semaglutide among them, as causing rapid and significant lean mass loss of roughly 10% or about 6 kg, comparable in its description to a decade or more of aging. That covers the class, not its members.
For the same question asked of the other two drugs, see tirzepatide muscle loss and retatrutide and lean mass preservation.
The same caution applies across species. The Cell Metabolism letter above measured skeletal muscle mass and force-generating capacity in mice, which can suggest a mechanism worth testing but does not describe people. For how the drug acts in the body, see how semaglutide works.
Video: CBS News on GLP-1 muscle loss concerns.
Frequently Asked Questions
Does semaglutide cause muscle loss?
Lean tissue does decline, as with any large weight loss from any cause. In the STEP 1 substudy, lean mass fell 9.7% of its own baseline while fat fell 19.3%, and the lean-to-fat ratio improved. Whether that counts as muscle loss depends on which layer you measure.
What percentage of weight lost on semaglutide is muscle?
None of the four studies covered here reports that statistic, and the STEP 1 substudy gives change within each tissue against its own baseline. Across incretin drugs as a class, a 2026 Annals of Internal Medicine review of 35 trials puts the median at 28.3%. The widely quoted 40% is not one of those findings, and no one in STEP 1 published it.
How do you not lose muscle on semaglutide?
Resistance training and adequate protein, commonly cited as 1.2 to 1.6 grams per kilogram of body weight daily, have the most evidence behind them. The reviewed evidence is about training duration, programmes longer than 10 weeks; the familiar two to three sessions a week comes from general activity guidance, not from a semaglutide trial. Appetite suppression makes protein targets harder to reach.
Does semaglutide make your muscles weak?
The available function data does not show that. In SLIM LIVER, psoas volume fell 9.3%, but chair-rise time and gait speed showed no significant change and the authors concluded function was maintained. Slow-gait prevalence fell from 63% to 46%. In SEMALEAN, handgrip rose 4.1 kg over 12 months. Both were small studies without control groups.
Does Ozempic cause muscle wasting?
Muscle wasting means disease-level loss such as sarcopenia or cachexia, and no semaglutide study reports that pattern. SEMALEAN found sarcopenic obesity fell from 49% to 33% over 12 months. Ozempic is semaglutide dosed for type 2 diabetes, so the 1.0 mg studies above sit closer to it than the 2.4 mg obesity data.
Can you regain muscle lost on semaglutide?
Lean tissue and strength can be rebuilt through resistance training and adequate protein, as after any period of weight loss. No trial has tested muscle regain after stopping semaglutide, so this is general exercise physiology, not a semaglutide-specific finding.
Is semaglutide worse for muscle than tirzepatide or retatrutide?
No head-to-head body composition trial has compared them, so a ranking is not supported. The Locatelli review in Diabetes Care describes incretin-based drugs as a class causing lean mass loss of around 10%, or roughly 6 kg, which it compares to a decade or more of aging. That covers the class rather than separating its members.
Plan Your Semaglutide Dosing
Map your schedule and see when levels plateau
References
- Wilding JPH, Batterham RL, Calanna S, et al. Impact of Semaglutide on Body Composition in Adults With Overweight or Obesity: Exploratory Analysis of the STEP 1 Study. J Endocr Soc. 2021;5(Suppl 1):A16-A17. (Conference abstract.) [DOI]
- McCrimmon RJ, Catarig AM, Frias JP, et al. Effects of once-weekly semaglutide vs once-daily canagliflozin on body composition in type 2 diabetes: a substudy of the SUSTAIN 8 randomised controlled clinical trial. Diabetologia. 2020;63(3):473-485. [PubMed]
- Ditzenberger GL, Lake JE, Kitch DW, et al. Effects of Semaglutide on Muscle Structure and Function in the SLIM LIVER Study. Clin Infect Dis. 2025;80(2):389-396. [PubMed]
- Alissou M, Demangeat T, Folope V, et al. Impact of Semaglutide on fat mass, lean mass and muscle function in patients with obesity: The SEMALEAN study. Diabetes Obes Metab. 2026;28(1):112-121. [PubMed]
- Locatelli JC, Costa JG, Haynes A, et al. Incretin-Based Weight Loss Pharmacotherapy: Can Resistance Exercise Optimize Changes in Body Composition? Diabetes Care. 2024;47(10):1718-1730. [PubMed]
- Batsis JA, Gavras A, Gross DC, et al. Effect of Incretin-Based and Nonpharmacologic Weight Loss on Body Composition: A Systematic Review. Ann Intern Med. 2026;179(7):996-1013. (35 randomised trials; incretin class, not semaglutide alone. A correction notice was published for this review, Ann Intern Med. 2026;179(7):1072; we could not access it.) [PubMed]
- Karasawa T, Choi RH, Meza CA, et al. Unexpected effects of semaglutide on skeletal muscle mass and force-generating capacity in mice. Cell Metab. 2025;37(8):1619-1620. (Animal study.) [DOI]
Medical Disclaimer
This article is for informational and educational purposes only and does not constitute medical advice. Nothing here is a personalised nutrition, exercise, or dosing recommendation. Do not start, stop, or modify any medication regimen without consulting a qualified healthcare provider. Individual results vary. Always discuss treatment options, risks, benefits, and alternatives with your doctor.