Every piece of advice about keeping muscle on a GLP-1 says the same two things. Lift weights, eat protein. No completed trial has tested either one as a way to preserve muscle on these drugs.1
One trial did test exercise alongside a GLP-1. The drug was liraglutide. Randomization came after an eight-week diet, not at the start, and the exercise arm was mostly cycling.2 In it, the placebo group gained more lean mass in the arms and legs than the exercise group did.3
Two drugs have been tested for this exact job. Both preserved lean mass.45
Has anyone tested lifting on semaglutide or tirzepatide?
Not with results, no.
A 2024 review in Diabetes Care came from a group that works on this question. Having surveyed the field, it ended with a proposal, not a finding: "We propose that tailored resistance exercise training may be recommended as an adjunct to incretin therapy to optimize changes in body composition by preserving lean mass while achieving fat loss."1
A proposal is what specialists write when there is no trial to cite, and no stronger statement than that exists in the literature.
Its supporting numbers come from elsewhere. Supervised resistance training over more than ten weeks "can elicit large increases in lean mass (∼3 kg) and strength (∼25%) in men and women," the authors note, in people who were not taking these drugs.1 The inference that this carries over is reasonable. It is still an inference.
What did the one exercise trial actually do?
That exercise trial did something narrower than its reputation suggests.
The trial is usually cited as proof that exercise plus a GLP-1 beats either alone, and on weight it did. Its design matters for muscle.
| What the trial did | |
|---|---|
| Drug | liraglutide 3.0 mg daily, not semaglutide or tirzepatide |
| Starting point | after an eight-week low-calorie diet in which everyone lost about 13 kg |
| Arms | exercise plus placebo, liraglutide plus usual activity, both, or neither |
| Duration | one year after the diet |
| Primary endpoint | change in body weight |
Participants were randomized only after the diet, so every number is measured from a post-diet low point, not from where people started.36
The exercise was not a lifting program. The prescription was two supervised group sessions a week of "30 min of vigorous-intensity indoor cycling and 15 min of circuit training," plus two individual sessions. The target was the World Health Organization's guidance of 150 minutes of moderate activity a week, or the vigorous equivalent.7
The circuit is the part that matters. Later analysis by the same investigators describes it as combining vigorous aerobic work with "muscle-strengthening exercises using body weight or external resistance," typically "3 circuits of 5 exercises performed for 40 seconds interspersed with 20-second breaks."6
So it was not resistance-free. Nor was it progressive resistance training. No loads, sets, reps or progression are described anywhere in the published accounts. The words "resistance training" and "strength training" do not appear in the investigators' own methods paper at all.7
What happened to lean mass in that trial?
Lean mass moved differently depending on which measure you read. One of them points the wrong way.
Arm-by-arm figures for appendicular fat-free mass, meaning the arms and legs, were published in a secondary analysis covering randomization to week 52.3
| Arm | Change in appendicular fat-free mass |
|---|---|
| Placebo | +1.27 kg (95% CI 0.80 to 1.74) |
| Exercise | +0.95 kg (95% CI 0.45 to 1.44) |
| Liraglutide | +0.03 kg (95% CI −0.43 to 0.48) |
| Combination | +0.03 kg (95% CI −0.42 to 0.48) |
Placebo gained the most. The likely reason is unglamorous: the placebo group regained weight after the diet, and regained lean tissue along with it.
Whole-body figures look better for exercise. One narrative review summarizing the trial reports that the exercise arm "increased FFM (+2.1 kg)" while the combination arm held it flat at +0.5 kg and liraglutide alone was unchanged.8 Those are different measures from the appendicular figures and not a direct contradiction, but the review does not give the placebo value, so there is no comparison to make.
What cannot be said from this trial is that exercise protected lean mass and the drug did not. On the one measure reported with all four arms, the group doing nothing came out ahead.
Did anyone get stronger?
Absolute strength did not move.
The same secondary analysis prespecified isometric knee extensor strength as an outcome. Its finding: muscle strength "did not change significantly" in any arm.3 The arm-level changes in peak torque were −4.2 Nm for placebo, +1.2 for exercise, −2.3 for liraglutide and −4.5 for the combination, and no between-group difference reached significance.
Strength normalized to body weight did improve, and this is where the trial gets quoted misleadingly. Relative strength changed by −7.8% on placebo against −0.4% with exercise, +1.0% with liraglutide and +3.3% with the combination. The authors state the reason in the same sentence: the gains came "because of lower weight and preserved absolute strength."3
That is a division problem, not a training effect. The same muscle moving a lighter body produces a better ratio.
Fitness did improve with exercise. The combination cut stair-climb time by 1.2 seconds against liraglutide alone, about 8.6%, and raised peak oxygen consumption. Liraglutide alone improved no measure of physical fitness.3 Fitness and strength are different things and only one of them moved.
How much of the weight is lean tissue anyway?
Between a quarter and a third, and three randomized trials now agree on that.
| Trial | Drug | Share of weight lost that was lean tissue |
|---|---|---|
| SURMOUNT-1 substudy | tirzepatide | about 25% |
| BELIEVE | semaglutide 2.4 mg | about 29% |
| EMBRAZE | tirzepatide | 30.2% (95% CI 26.4 to 33.9) |
The tirzepatide substudy of SURMOUNT-1 scanned 160 of 2,539 participants and found weight, fat and lean mass down 21.3%, 33.9% and 10.9% at week 72 against 5.3%, 8.2% and 2.6% on placebo. Its own summary: "Of the weight lost, approximately 75% was fat and 25% lean mass for both tirzepatide and placebo."9
Note the last four words. The lean fraction was the same on the drug as on placebo. The EMBRAZE trial found the same thing from the other direction, with its tirzepatide-plus-placebo arm losing 30.2% of weight as lean tissue.5 Losing some lean tissue while losing weight is what weight loss does, not something these drugs do uniquely.
What about the people who say they got stronger?
Some do, and one cohort measured it. The study design limits what it can mean.
Patients on semaglutide 2.4 mg were followed in a prospective cohort that measured grip strength alongside body composition. Lean mass fell about 3 kg by month seven and then held steady. Handgrip strength improved by 4.5 kg at twelve months, and the share of patients meeting criteria for sarcopenic obesity, meaning low muscle alongside excess fat, fell from 49% to 33%.10
There was no control group and no randomization. Everyone got the drug, and everyone also got the diet counselling that comes with it. Before-and-after designs cannot separate the drug from the counselling, from people taking up exercise on their own, or from simply getting better at the test.
The measurement is still real and it points somewhere useful. Lean mass on a scanner fell while grip strength rose. Those two findings are not in conflict, for a reason that comes down to what a scanner measures.
Is lean mass the same as muscle?
No, and the gap is larger than most coverage implies.
Writing in the Journal of the Endocrine Society in 2024, two researchers make the point directly. "It is essential to recall that even corrected FFM is not directly synonymous with skeletal muscle mass." The reason is that "All major categories of molecules, including water, protein, minerals, glycogen, and lipids, are included."8
Its conclusion is that "the entirety of FFM loss is not directly attributable to skeletal muscle."8
The quantitative version comes from a study of diet-induced weight loss, reproduced in that paper. Of 0.9 kg of skeletal muscle lost, "0.7 kg (46% of FFM loss) was due to reductions of water while only 0.15 kg (10% of FFM loss) was due to actual protein loss."8 Those proportions are from dieting, not from these drugs, and they reach this page through a review that reproduced them.
A scanner reporting lean mass down three kilograms is not reporting three kilograms of muscle gone. That is why the trial now underway on this question uses magnetic resonance imaging of the thigh as its primary measure instead of a body composition scan.11
What has been tested, and worked?
Both of the things that worked were drugs, tested in randomized phase 2 trials published in 2026.
Bimagrumab, in a trial called BELIEVE, randomized 507 adults with obesity across nine groups for 48 weeks. On the treatment regimen estimand, total body lean mass changed by +1.0% to +1.1% with bimagrumab, −4.7% to −6.9% with semaglutide, and −0.8% to −2.3% with the combination, against −0.9% on placebo.4 Fat loss went the other way: the combination arms lost 30.6% to 42.2% of fat mass against 18.4% to 24.8% for semaglutide alone.
The trial's cleanest number is the share of weight loss that came from fat. At week 48 it was "100% (bimagrumab 30 mg kg−1), 71.1% (semaglutide 2.4 mg) and 92.3% (high-dose combination)."4
Apitegromab, in a trial called EMBRAZE, randomized 102 people on tirzepatide to the drug or placebo for 24 weeks. Lean mass loss was 1.6 kg with apitegromab against 3.5 kg with placebo, a difference of 1.9 kg, with total weight loss similar in both arms.5 As a share of weight lost, lean tissue accounted for 14.6% with apitegromab and 30.2% without.
Both are small and neither is definitive. EMBRAZE used 80% confidence intervals and a two-sided significance level of 20%, with no adjustment for multiple comparisons, which is a deliberately permissive design for a phase 2.5 BELIEVE reports its lean mass results under two different estimands that disagree by about a percentage point.4
The point is not that these drugs are the answer. It is that the question has been asked properly of a monoclonal antibody twice, and never of a barbell.
What about protein?
Same position as lifting. No completed randomized trial.
The advice to eat more protein while losing weight does not come from research in people taking these drugs. It comes from the general weight-loss literature, which is the same inferential step the lifting advice makes, and the 2024 review that proposes resistance training treats protein the same way.1
Three separate questions sit inside it and none has been answered in this population. Whether added protein changes lean mass on its own. Whether it adds anything on top of training. Whether either does enough to matter against the quarter-to-a-third of weight that comes off as lean tissue regardless.
Protein has its own arm in the trial now underway, with a target of 1.6 grams per kilogram of body weight per day.11 That figure is reported here as the trial's own protocol parameter. This site does not publish nutrition targets, and a number chosen for a trial population is not a recommendation for anyone. Nothing published tests creatine, branched-chain amino acids or any other supplement for muscle preservation on these drugs either. The absence is the same shape: those compounds have been studied elsewhere and not here.
Does it matter if you were already training?
No published trial answers that, because none reports its results by whether people were already training.
The obvious way to find out is to look at whether people who arrived at a trial already active lost less lean tissue than people who did not. No published body composition substudy of these trials reports lean mass broken out by participants' baseline physical activity or training status.
The tirzepatide substudy of SURMOUNT-1 is the clearest example, because it did run subgroups. The subgroups were "by sex (female or male), age (<50, 50 to <65, or ≥65 years) and total weight reduction tertiles."9 Activity level is not among them.
That absence is about published substudies, not about every table in every supplement. But if a trial had found that prior training changed the outcome, it would be a headline, not a footnote.
What is being tested now?
One trial is aimed directly at the question. It has no results yet.
A protocol published in BMJ Open in April 2026 describes a six-month randomized trial at the Dasman Diabetes Institute in Kuwait. The trial allocates 232 adults with obesity in equal groups to control, resistance exercise, protein supplementation, or both, all while taking semaglutide or tirzepatide.11
The four arms are the point. Training and protein are almost always recommended together, which means no existing evidence separates them. Four groups can show whether either does anything alone and whether the pair adds up to more than its parts.
Its primary outcome is the cross-sectional area of the quadriceps measured by magnetic resonance imaging, with body composition, strength and physical function as secondary measures.11 Choosing imaging of one muscle over a whole-body scan sidesteps the problem in the section above, where fat-free mass and muscle are not the same quantity.
The resistance exercise arm is home-based, three sessions a week, progressing from one set to three.11 As with the protein target, those are design choices and not guidance.
So the claim is not that nobody has tried. It is that nobody has reported.
What would actually settle this?
Settling it takes three studies. The first is already running.
| What is missing | Why it would settle something |
|---|---|
| Results from the registered trial | the only randomized test of resistance training and protein in people on these drugs |
| A strength outcome that is not a ratio | the one exercise trial found no change in absolute strength in any arm |
| A baseline-activity analysis | the question a reader who already trains is actually asking |
Until the first of those reports, the advice to lift and eat protein rests on what resistance training does in people who are not taking these drugs. That is a reasonable place to stand. It is not the same as having been tested, and those are two different claims.
Our guide to what happens after significant weight loss covers the lean mass findings in more detail, including what is sold for this and what it is worth.
Common questions
Does lifting weights protect muscle on a GLP-1?
No completed randomized trial has tested it on semaglutide or tirzepatide. The recommendation comes from resistance training research in people not taking these drugs, where supervised programs over ten weeks or more raise lean mass by around 3 kg. Whether that carries over has not been measured.
Has any trial combined exercise with a GLP-1?
One trial has, using liraglutide. Participants were randomized after an eight-week diet to exercise, liraglutide, both, or neither, for a year. Its primary endpoint was body weight, not muscle, and its exercise program was mostly interval cycling with a 15-minute circuit.
Did exercise preserve lean mass in that trial?
Not on the measure reported for all four arms. Appendicular fat-free mass rose 1.27 kg on placebo against 0.95 kg with exercise, most likely because the placebo group regained weight. Whole-body figures look better for exercise but were not reported with a placebo comparison.
Did people in that trial get stronger?
No. Isometric knee extensor strength showed no change in any arm of the liraglutide trial. Strength relative to body weight improved, which the authors attribute to lower body weight with preserved absolute strength.
How much muscle do you lose on semaglutide or tirzepatide?
Across three randomized trials, lean tissue was roughly a quarter to a third of total weight lost. About 25% in the tirzepatide substudy of SURMOUNT-1, about 29% for semaglutide in BELIEVE, and 30.2% in the tirzepatide arm of EMBRAZE.
Is that worse than losing weight any other way?
The tirzepatide substudy found approximately 75% fat and 25% lean "for both tirzepatide and placebo." The lean fraction was the same without the drug, which suggests this is a feature of weight loss, not of the drug.
Is lean mass the same thing as muscle?
No. Fat-free mass includes water, protein, minerals, glycogen and lipids, and a 2024 review states that "the entirety of FFM loss is not directly attributable to skeletal muscle." In one diet study, 46% of fat-free mass lost was water and only 10% was protein.
Does a scan showing 3 kg of lean mass lost mean 3 kg of muscle gone?
No. Fat-free mass includes water, protein, minerals, glycogen and lipids, so it is not a muscle measurement. Partly for that reason, the trial now testing resistance training on these drugs uses magnetic resonance imaging of the thigh as its primary measure instead of a body composition scan.
Has anything been shown to preserve lean mass on these drugs?
Two drugs, in phase 2 trials. Bimagrumab added to semaglutide held lean mass near flat while increasing fat loss. Apitegromab added to tirzepatide produced 1.9 kg less lean mass loss than placebo at similar total weight loss. Both trials are small, and EMBRAZE used 80% confidence intervals and a 20% significance level.
Is bimagrumab or apitegromab available?
Neither is approved. Both were tested in phase 2 trials published in 2026 and neither has completed the development required for approval.
What about protein intake?
No completed randomized trial has tested a protein intervention in people taking these drugs. One is underway, with a protocol target of 1.6 grams per kilogram per day. That is a trial parameter, not guidance.
Does it help if I was already lifting before I started?
No published trial answers that. No body composition substudy of these trials reports lean mass by participants' baseline activity or training status. The tirzepatide substudy ran subgroups by sex, age and weight-loss tertile, and activity was not one of them.
Is tirzepatide worse than semaglutide for muscle?
A retrospective analysis of routine care records found greater lean mass loss on tirzepatide. That analysis is a preprint, it has not completed peer review, it is observational, and only 1.2% of its cohort had paired body composition measurements. The randomized trials give similar lean fractions for both.
Is there a trial testing lifting right now?
Yes. One published in April 2026 describes a six-month randomized trial of 232 adults taking semaglutide or tirzepatide, with arms for resistance exercise, protein, both, and neither. No results have been reported.
Should I stop worrying about muscle then?
No. Losing a quarter to a third of weight as lean tissue is real and measured in randomized trials. What has not been tested is the specific claim that lifting and protein change that outcome in people taking these drugs, while two antibodies have been tested and did.
Why is strength measured so inconsistently?
Because the trials use different tools. The exercise trial measured isometric knee extensor torque. An observational cohort of people on semaglutide measured handgrip, and reported it improving by 4.5 kg at twelve months while lean mass fell about 3 kg and then held. Another randomized analysis used a sit-to-stand count. None of those converts into the others, which is part of why the picture looks muddled.
Sources
- Locatelli JC, Costa JG, Haynes A, Naylor LH, Fegan PG, Yeap BB, Green DJ. "Incretin-Based Weight Loss Pharmacotherapy: Can Resistance Exercise Optimize Changes in Body Composition?" Diabetes Care 2024;47(10):1718-1730. Digital object identifier 10.2337/dci23-0100. Read via a reproduction: the publisher returned an error, so the quoted text comes from the abstract as deposited and retrieved through a bibliographic service, where word order is reconstructed, not read. Source for "We propose that tailored resistance exercise training may be recommended as an adjunct to incretin therapy to optimize changes in body composition by preserving lean mass while achieving fat loss," and for the statement that supervised resistance training over more than ten weeks "can elicit large increases in lean mass (∼3 kg) and strength (∼25%) in men and women," which are figures from trials in people not taking these drugs. The absence of a completed randomized trial of resistance training or protein in people on semaglutide or tirzepatide was established by six separate searches of a bibliographic index covering the published literature, each one inspected by hand. That search covers published papers only: a trial registry could not be searched, so this claim covers only what has been published. ↩1 ↩2 ↩3 ↩4
- Lundgren JR, et al. "Healthy Weight Loss Maintenance with Exercise, Liraglutide, or Both Combined." New England Journal of Medicine 2021;384(18):1719-1730. Digital object identifier 10.1056/NEJMoa2028198, confirmed at the identifier registry. Read via a reproduction: the publisher was unreachable, so all quoted text comes from the deposited abstract and no results table was seen. Source for the four-arm design "After an 8-week low-calorie diet, participants were randomly assigned for 1 year to one of four strategies," for the arms as exercise plus placebo, liraglutide 3.0 mg plus usual activity, both, or placebo plus usual activity, for the primary endpoint as change in body weight with body fat percentage secondary, for the diet producing a mean 13.1 kg loss in 195 participants before randomization, and for the one-year weight differences against placebo of −4.1 kg for exercise, −6.8 kg for liraglutide and −9.5 kg for the combination. Lean mass does not appear in the abstract. Registrations printed in the secondary analyses are EudraCT 2015-005585-32 and NCT04122716. ↩1
- Jensen SBK, et al. "Physical Fitness with Exercise and GLP-1 Receptor Agonist Treatment Alone or Combined After Diet-Induced Weight Loss: A Secondary Analysis of a Randomized Controlled Trial in Adults with Obesity." Sports Medicine 2026;56(7):1785-1800. Digital object identifier 10.1007/s40279-025-02386-0. Read at source, open access. Source for the appendicular fat-free mass table, for "Appendicular FFM increased with placebo and with exercise alone," for the peak torque figures and the statement about strength quoted in the body, for the relative strength percentages and the authors' explanation that they arose "because of lower weight and preserved absolute strength," and for the stair-climb and oxygen consumption results. This paper reports 193 participants randomized and treated; the parent trial reports 195 randomized. Both figures are in print and the difference is most likely those who received at least one dose, which this page did not confirm. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6
- "Bimagrumab plus semaglutide alone or in combination for the treatment of obesity: a randomized phase 2 trial." Nature Medicine 2026;32:869-882. Digital object identifier 10.1038/s41591-026-04204-0. Read at source. The trial is named BELIEVE. Source for "507 adults with obesity" randomized to nine groups, for the 48-week blinded period, for the primary endpoint as "absolute change from baseline in body weight at week 48," for the lean mass percentages on the treatment regimen estimand, for the fat mass percentages, and for the fat loss index values of "100% (bimagrumab 30 mg kg−1), 71.1% (semaglutide 2.4 mg) and 92.3% (high-dose combination)." The paper reports lean mass under two estimands which differ: the efficacy estimand gives +2.3% to +2.7% for bimagrumab and −5.3% to −7.9% for semaglutide. No confidence intervals are given for lean mass. The 29% figure used in this page's table is the arithmetic complement of the 71.1% fat loss index for semaglutide 2.4 mg, not a figure the paper states. ↩1 ↩2 ↩3 ↩4
- "Apitegromab for lean mass preservation during tirzepatide-induced weight loss: a randomized, double-blind, placebo-controlled phase 2 trial." Nature Medicine 2026;32(7):2673-2678. Digital object identifier 10.1038/s41591-026-04440-4. Read at source. The trial is named EMBRAZE. Source for 102 participants randomized 1:1 across seven US sites, for dosing every four weeks for 24 weeks, for the primary endpoint as "change from baseline in lean body mass at 24 weeks as assessed by whole-body DEXA," and for "Lean mass loss was 1.6 kg (0.8−2.3) with apitegromab versus 3.5 kg (2.8−4.1) with placebo, with lean mass loss representing 14.6% (10.5−18.7%) of weight loss with apitegromab and 30.2% (26.4−33.9%) of weight loss with placebo." Total weight loss was 11.2 kg against 12.5 kg. The trial used 80% confidence intervals and a two-sided significance level of 20%, and its p values were nominal with no adjustment for multiplicity. The primary p value appears as 0.001 in the abstract and 0.0014 in the results. ↩1 ↩2 ↩3 ↩4
- "Bone Health After Exercise Alone, GLP-1 Receptor Agonist Treatment, or Combination Treatment: A Secondary Analysis of a Randomized Clinical Trial." JAMA Network Open 2024;7(6):e2416775. Digital object identifier 10.1001/jamanetworkopen.2024.16775. Read at source. Source for the description of the circuit as combining vigorous aerobic exercise with the muscle-strengthening movements quoted in the body, for "Typically, 3 circuits of 5 exercises were performed for 40 seconds interspersed with 20-second breaks," and for the statements that "The exercise group had increased lean mass during the 1-year intervention period" and "The combination group preserved lean mass despite the largest fat mass reduction," neither of which is accompanied by a placebo value. This paper describes a six-week ramp-up where the methods paper at footnote 6 describes seven weeks, and that paper refers to six weeks elsewhere in its own text. ↩1 ↩2
- Jensen SBK, et al., methods report of the same parent trial. Nature Communications 2022. Digital object identifier 10.1038/s41467-022-32307-y. Read at source. Source for the exercise prescription as "30 min of vigorous-intensity indoor cycling and 15 min of circuit training" twice weekly plus two individual sessions, for the target of "at least 150 min of moderate-intensity aerobic physical activity" or the vigorous equivalent, for the achieved dose of "156 ± 54 min/week" at "78 ± 4% of maximum heart rate," and for the randomization of 195 participants after the diet. The phrases "resistance training," "strength training" and "progressive overload" do not appear in this paper, and no loads, sets or repetition schemes are specified anywhere in it. A separate secondary analysis reports a different achieved dose of about 116 minutes per week, and this page could not establish which arms or analyses the two figures describe. ↩1 ↩2
- Tinsley GM, Heymsfield SB. "Fundamental Body Composition Principles Provide Context for Fat-Free and Skeletal Muscle Loss With GLP-1 RA Treatments." Journal of the Endocrine Society 2024;8(11):bvae164. Digital object identifier 10.1210/jendso/bvae164. Read at source. Source for the two statements quoted in the body, that corrected fat-free mass is not synonymous with skeletal muscle mass and that all major molecular categories are included in it, for the conclusion about how much of fat-free mass loss is attributable to skeletal muscle, quoted in the body, and for the reproduced water and protein breakdown of 0.9 kg of skeletal muscle lost, also quoted in the body. Those proportions come from a study of diet-induced weight loss which this page did not read in the original, and the paper does not address rapid weight loss or body composition scanning specifically. This review is also the source of the whole-body fat-free mass figures quoted for the exercise trial, +2.1 kg for exercise alone and +0.5 kg for the combination, which are therefore reported here at one remove, not from the trial paper. ↩1 ↩2 ↩3 ↩4
- Look M, Dunn JP, Kushner RF, Cao D, Harris C, Hunter Gibble T, Stefański A, Griffin R. "Body composition changes during weight reduction with tirzepatide in the SURMOUNT-1 study of adults with obesity or overweight." Diabetes Obesity and Metabolism 2025;27(5):2720-2729. Digital object identifier 10.1111/dom.16275. Read via a reproduction: the publisher is unreachable, so quoted text comes from the deposited abstract. Source for the substudy of "160 of the 2539 participants," for scanning at baseline and week 72, for "The change in body weight, fat mass and lean mass from baseline to Week 72 was -21.3%, -33.9% and -10.9% with tirzepatide and -5.3%, -8.2% and -2.6% with placebo," for "Of the weight lost, approximately 75% was fat and 25% lean mass for both tirzepatide and placebo," and for the three subgroups it ran, quoted in the body. ↩1 ↩2
- Alissou M, Demangeat T, Folope V, Coëffier M, Achamrah N, et al. "Impact of Semaglutide on fat mass, lean mass and muscle function in patients with obesity: The SEMALEAN study." Diabetes Obesity and Metabolism 2026;28(1):112-121, published online 9 October 2025. Digital object identifier 10.1111/dom.70141. Read via a reproduction: the publisher is unreachable, so quoted text comes from the deposited abstract. This is a prospective observational cohort with no control group and no randomization, which is why this page treats its handgrip result as a cohort observation rather than a training or drug effect. It reports lean mass declining about 3 kg by month seven then stabilizing, handgrip strength improving 4.5 kg at month twelve, and sarcopenic obesity prevalence falling from 49% at baseline to 33% at month twelve. The enrolled figure in the deposited abstract is 115 adults on semaglutide 2.4 mg; the number completing could not be established. ↩1
- Protocol for a randomized controlled trial of resistance exercise and protein intake during semaglutide or tirzepatide therapy, known as LEAN-PREP. BMJ Open 2026;16(4):e116911, published online 22 April 2026. Bibliographic record read at the identifier registry; the article itself could not be opened, and the protocol details here come from the abstract as deposited and retrieved through a bibliographic service, where word order is reconstructed. Two different titles are registered for this paper by two services, so no title is quoted here. Reported design: a single-center six-month trial at the Dasman Diabetes Institute in Kuwait, 232 adults with obesity randomized in equal groups to control, resistance exercise, protein supplementation or both while taking semaglutide or tirzepatide; resistance exercise home-based three times weekly progressing from one to three sets; protein target 1.6 g/kg/day; primary outcome the cross-sectional area of the quadriceps by magnetic resonance imaging; secondary outcomes body composition, strength, physical function, liver fat and intramuscular fat. No results have been published. ↩1 ↩2 ↩3 ↩4 ↩5
- Murugadoss K, Venkatakrishnan AJ, Soundararajan V. "Greater lean-body-mass decline with tirzepatide than semaglutide in routine care, revealed by body-composition digital phenotyping." Preprint posted 13 April 2026, digital object identifier 10.21203/rs.3.rs-9407186/v1, with a second preprint version on another server. Read at source. Not peer reviewed. A retrospective cohort using deidentified electronic health records covering "670,422 first-episode GLP-1 patients," of whom 7,965 had paired lean body mass measurements, which is 1.2% of the cohort. It reports a pattern the authors call the depletive metabotype, defined in the text as more than 20% weight loss with more than 5% lean body mass loss, in "10.3% (n = 142) of tirzepatide-treated patients" against "6.7% (n = 252) of semaglutide-treated patients." The figure caption uses "≥ 5%" where the text uses "> 5%." All three authors are employed by the company that owns the records network.