Home / Guides by goal / GLP-1 Muscle Loss: What the Data Show
GLP-1 Muscle Loss: What the Data Show
How much muscle do GLP-1 drugs cost? DXA and MRI data, two 2026 reviews that disagree, what is still unknown, and drugs in trials to protect muscle.
On this page
Many people ask whether GLP-1 drugs cost muscle. The data say something narrower than the headlines: weight loss of any kind includes some lean mass, how much differs across studies, and very few studies measured strength.
The short answer
People who lose weight on semaglutide or tirzepatide also lose lean mass, and typical estimates in the studies and reviews below are roughly a quarter to a third of the weight lost. In a body-composition substudy of the SURMOUNT-1 trial, about 75% of the weight lost on tirzepatide was fat and about 25% lean mass, similar to the placebo group. A 2026 meta-analysis of randomized trials found that the share of weight lost as lean mass with these drugs was broadly similar to weight loss through diet and exercise.
Lean mass measured by DXA is not the same as muscle, and neither is the same as strength.
Two caveats matter. A second 2026 review reads the evidence differently, and a 2026 systematic review of 35 studies found that none reported objective physical-function outcomes. So “GLP-1 drugs destroy muscle” and “there is no muscle loss” are both claims the data do not support.
These drugs are FDA-approved for specific uses; retatrutide and the muscle-preserving drugs mentioned here are not approved.
How this guide is built
This is an explainer, not a ranking of drugs. We use the studies’ own measures and say when a figure comes from an abstract only. We read the studies in this guide through their PubMed abstracts, not the full papers, and the labels through section 12.2 (pharmacodynamics). For muscle-building claims about peptides sold online, see our separate guide to peptides for muscle growth. For the drugs themselves, see our guides to semaglutide, tirzepatide, Wegovy vs Ozempic and Zepbound vs Mounjaro.
What do the labels say?
Both labels describe the same pattern in their pharmacology sections. Wegovy: “Semaglutide lowers body weight with greater fat mass loss than lean mass loss.” Zepbound: “Tirzepatide lowers body weight with greater fat mass loss than lean mass loss” (Wegovy, Zepbound, section 12.2). Neither label gives lean-mass figures or discusses strength.
What do DXA and MRI measure?
- DXA reports lean mass, a category that includes more than skeletal muscle.
- MRI can measure muscle volume directly, but volume is not strength.
- A fall in muscle volume does not by itself prove a loss of strength, and no fall does not prove strength was kept.
We write “lean mass” when a study says lean mass and “muscle volume” when it used MRI.
What did the body-composition studies find?
| Study | Design | Who | Method | Result | Funding |
|---|---|---|---|---|---|
| SURMOUNT-1 substudy, 2025 | Substudy of a phase 3 randomized trial, 72 weeks | Adults with obesity, no type 2 diabetes; 160 with DXA (124 tirzepatide, 36 placebo) | DXA | Weight −21.3% vs −5.3%; fat mass −33.9% vs −8.2%; lean mass −10.9% vs −2.6%. About 75% of weight lost was fat and 25% lean mass in both groups | Lilly |
| SURPASS-3 MRI, 2025 | Post hoc analysis of an open-label phase 3 trial, 52 weeks | Type 2 diabetes; 246 with MRI, vs insulin degludec | Thigh MRI | Tirzepatide: muscle volume −0.64 L; fat infiltration of muscle −0.36 points. The authors say the loss is “broadly in accordance” with the weight lost | Lilly |
| STEP UP, 2026 | Prespecified secondary analysis of a phase 3b trial, 72 weeks | Adults with obesity; only 55 with valid MRI (49 semaglutide, 6 placebo) | MRI | Vs placebo: adipose tissue −11.1 L; thigh muscle volume −1.1 L (p=0.07, not significant); muscle fat infiltration −0.92 points | A Novo Nordisk trial; funding not in the abstract |
| SUSTAIN 8 DXA substudy, 2020 | Substudy of a randomized trial, 52 weeks | Type 2 diabetes; 178 (114 with final data), semaglutide vs canagliflozin | DXA | Fat −3.4 vs −2.6 kg; lean mass −2.3 vs −1.5 kg (difference not significant). No placebo arm; the authors call its impact “speculative” | Not stated in our source |
| Retatrutide phase 2 substudy | Substudy of a phase 2 trial, 36 weeks | Type 2 diabetes; 103 with DXA at baseline and end | DXA | The abstract gives no lean-mass figures; the authors say the share was similar to other treatments | Lilly |
We do not give lean-mass figures for the STEP 1 DXA substudy or for retatrutide, because the abstracts we read have none. In STEP UP, the placebo group had six people, so its comparison is weak.
Do the 2026 reviews agree?
They do not fully agree, and we do not choose between them.
| Review | What it pooled | Finding | Limit |
|---|---|---|---|
| Batsis 2026, systematic review | 35 studies, median 26 weeks and 78 participants; DXA, bioimpedance, CT or MRI | A median of 28.3% of weight lost was in muscle indices (interquartile range 15.9–39.9%); 65% of estimates exceeded a benchmark of about 25%. “No study reported objective physical function outcomes” | Methods and durations varied widely; 42.9% had low risk of bias. Search ran to February 2026, before STEP UP |
| Eisa 2026, meta-analysis | 20 randomized trials, 15,782 participants; DXA or MRI | Lean mass as a share of weight lost: semaglutide 35.2%, tirzepatide 25.4%, liraglutide 26.8%, diet and exercise 26.2% (no difference, p=0.42), diet plus strength training 17.5% | Moderate heterogeneity (I² 68%). Its conclusion about strength training is the authors’ recommendation, not a trial in people taking GLP-1 drugs |
| Network meta-analysis, 2025 | 22 randomized trials, 2,258 participants | GLP-1 drugs: lean mass −0.86 kg (95% CI −1.30 to −0.42), about 25% of weight lost | We read the abstract only |
How to read them side by side: Eisa finds the share similar to losing weight with diet, while Batsis finds more estimates above a threshold with incretin-based drugs. They pooled different sets of studies with different designs, so neither settles the question.
The figures for semaglutide and tirzepatide in Eisa also differ (35.2% vs 25.4%). We found no trial comparing the two directly on body composition, so that gap comes from comparing different trials.
Does losing lean mass mean losing strength?
Few trials measured it. What we found, in specific groups:
- STEP UP (n=55 with MRI): sit-to-stand repetitions in 30 seconds went from a mean of 13 at baseline to 15 at week 72, with the semaglutide and placebo groups pooled; the authors describe the number as “preserved.” It is small and gives no comparison with placebo.
- STEP-HFpEF (529 people with heart failure with preserved ejection fraction and obesity): six-minute walk distance rose 21.5 m vs 1.2 m on placebo, a difference of 20.3 m. It measures function, not muscle.
- STEP 9 (407 people with knee osteoarthritis and obesity): the SF-36 physical function score rose 12.0 vs 6.5 points. See our joint pain guide.
In STEP-HFpEF and STEP 9, function improved more than with placebo. Those two trials measured function, not muscle, and none of this can be generalized to healthy older adults or to people with sarcopenia.
What is still not known?
- Older adults and sarcopenia. Mean participant age in the studies of the Batsis review ranged from 20 to 63.7 years. The SURMOUNT-1 substudy says the proportions of fat and lean mass lost were consistent across most post hoc subgroups, which included people 65 and older; its abstract gives no figures for that group. A trial in older adults is registered, and we found no results.
- Long term and weight regain. Data are limited. A 2026 systematic review of weight regain after stopping these drugs says structured exercise “may help reduce regain and preserve lean mass,” and that evidence for other strategies remains limited. We read its abstract only.
- Measured strength. Few trials measure it, and the Batsis review found none with objective physical function.
- Strength training and protein with GLP-1 drugs. Protocols are underway (LEAN-PREP); we read no results.
What do medical societies say?
- ACLM, ASN, OMA and The Obesity Society (joint advisory, 2025) list “muscle and bone loss” among the challenges. They recommend assessing strength, function and body composition before starting, and, during treatment, preserving muscle and bone with strength training and an adequate diet. A correction to the advisory has been published; we did not read it.
- EASO, EFAD and ECPO (2026 consensus) discuss adequate protein intake and progressive resistance exercise as practical approaches that could help preserve fat-free mass and physical function.
These are expert recommendations, not trial results. We do not give protein amounts or exercise plans. For what fits you, ask your clinician or a dietitian.
Are any drugs being tested to protect muscle?
Yes, all in trials, and none approved.
| Drug | Trial | Status on the registry | What is known |
|---|---|---|---|
| Bimagrumab (Lilly) with semaglutide | NCT05616013, phase 2, 507 people | Completed, with results | Published 2026: at 48 weeks, weight −17.8 kg with the high combination vs −14.2 kg with semaglutide alone and −3.3 kg with placebo. Side effects included muscle spasms, diarrhea and acne. We did not read the body-composition data |
| Bimagrumab with tirzepatide | NCT06643728, phase 2, 252 people | Active, not recruiting | No results |
| Bimagrumab with tirzepatide (Massachusetts General Hospital) | NCT05933499, phase 2, 63 people | Recruiting | No results |
| Enobosarm (Veru) | QUALITY, NCT06282458, phase 2, 168 people; PLATEAU, NCT07446998 | Completed; active | We read no publications and do not cite results |
| Apitegromab (Scholar Rock) with tirzepatide | EMBRAZE, NCT06445075, phase 2, 102 people | Completed | We read no publications and do not cite results |
| Trevogrumab, with or without garetosmab (Regeneron), with semaglutide | COURAGE, NCT06299098, phase 2, 1,005 people | Active, not recruiting | We read no results |
Retatrutide is also not approved: a search of FDA’s Drugs@FDA data on Nov. 4, 2026 found no product listed under that name.
We found no evidence that any peptide sold online preserves muscle during GLP-1 treatment. A registered trial is not a result, and none of these drugs should be described as preserving muscle yet.
Which claims are not supported?
- “GLP-1 drugs destroy muscle.” Lean mass falls, in a share similar to other ways of losing weight in one meta-analysis, though the reviews differ.
- “You only lose fat on these drugs.” Lean mass was part of the weight lost in every study above that reported it: about 25% in the SURMOUNT-1 substudy, and a median of 28.3% in muscle-based measures across the studies in the Batsis review.
- “Losing lean mass means losing strength.” Strength has rarely been measured; where function was measured, in specific groups, it improved or held steady.
- “Tirzepatide protects muscle better than semaglutide.” We found no trial comparing them directly on body composition.
- “A research peptide will protect your muscle on a GLP-1.” We found no evidence for that.
What should you ask your doctor?
- Could someone assess my strength, physical function and body composition before and during treatment?
- Is a referral to a dietitian or an exercise professional appropriate for me?
- Does my age, or a condition such as sarcopenia, change how closely I should be monitored?
- If I am offered a product to “protect muscle,” what evidence supports it for me, and is it approved?
We do not give doses, protein amounts or exercise plans, and we do not recommend products or sellers.
Frequently asked questions
Do GLP-1 drugs cause muscle loss?
People lose lean mass on semaglutide and tirzepatide, as they do with other weight loss. In one 2026 meta-analysis the share of weight lost as lean mass was broadly similar to that with diet and exercise, while another 2026 review reads the evidence differently. Neither shows that strength falls.
How much of the weight lost is muscle?
Lean mass measured by DXA is not the same as muscle. In a 160-person tirzepatide substudy, about 25% of the weight lost was lean mass.
Does tirzepatide or semaglutide cost less muscle?
We found no trial comparing them directly. Differences between the two in meta-analyses come from comparing different trials.
Can I prevent muscle loss on a GLP-1 drug?
Medical societies recommend strength training and adequate protein during treatment. These are expert recommendations, not trial results in people taking GLP-1 drugs. Ask your clinician about your own plan.
Is there a drug that protects muscle during weight loss?
Several are in phase 2 trials, including bimagrumab, enobosarm, apitegromab and trevogrumab. None is approved.
From Best Peptides Today
Sources we checked
- Wegovy prescribing information (DailyMed, rev. 06/2026), section 12.2FDA. Accessed .
- Zepbound prescribing information (DailyMed, rev. 08/2026), section 12.2FDA. Accessed .
- SURMOUNT-1 body composition substudy, Look 2025 (PubMed)PubMed. Accessed .
- SURPASS-3 MRI analysis, Sattar 2025 (PubMed)PubMed. Accessed .
- STEP UP MRI analysis, Hjelmesaeth 2026 (PubMed)PubMed. Accessed .
- SUSTAIN 8 DXA substudy, 2020 (PubMed)PubMed. Accessed .
- Retatrutide phase 2 body composition substudy (PubMed)PubMed. Accessed .
- Batsis 2026: systematic review of muscle and incretin-based weight loss, Ann Intern Med (PubMed)PubMed. Accessed .
- Eisa 2026: meta-analysis of lean mass with GLP-1 drugs, Diabetes Obes Metab (PubMed)PubMed. Accessed .
- Network meta-analysis of GLP-1 drugs and lean mass, 2025 (PubMed)PubMed. Accessed .
- STEP-HFpEF trial, Kosiborod 2023, NEJM (PubMed)PubMed. Accessed .
- STEP 9 trial, Bliddal 2024, NEJM (PubMed)PubMed. Accessed .
- Trial protocol in older adults, 2024 (PubMed)PubMed. Accessed .
- Systematic review of weight regain after stopping GLP-1 and GIP/GLP-1 drugs, 2026 (PubMed)PubMed. Accessed .
- LEAN-PREP trial protocol (PubMed)PubMed. Accessed .
- ACLM, ASN, OMA and TOS joint advisory on nutrition and GLP-1 therapy, 2025 (PubMed)PubMed. Accessed .
- Erratum to the ACLM, ASN, OMA and TOS advisory (PubMed)PubMed. Accessed .
- EASO, EFAD and ECPO consensus, Lancet Diabetes Endocrinol 2026 (PubMed)PubMed. Accessed .
- Heymsfield 2026: bimagrumab with semaglutide, Nature Medicine (PubMed)PubMed. Accessed .
- ClinicalTrials.gov NCT05616013 (bimagrumab with semaglutide, phase 2)ClinicalTrials.gov. Accessed .
- ClinicalTrials.gov NCT06643728 (bimagrumab with tirzepatide, phase 2)ClinicalTrials.gov. Accessed .
- ClinicalTrials.gov NCT05933499 (bimagrumab with tirzepatide, MGH, phase 2)ClinicalTrials.gov. Accessed .
- ClinicalTrials.gov NCT06282458 (enobosarm, QUALITY, phase 2)ClinicalTrials.gov. Accessed .
- ClinicalTrials.gov NCT07446998 (enobosarm, PLATEAU)ClinicalTrials.gov. Accessed .
- ClinicalTrials.gov NCT06445075 (apitegromab with tirzepatide, EMBRAZE, phase 2)ClinicalTrials.gov. Accessed .
- ClinicalTrials.gov NCT06299098 (trevogrumab with or without garetosmab, with semaglutide, COURAGE, phase 2)ClinicalTrials.gov. Accessed .
- Drugs@FDA (searched Nov. 4, 2026; no product listed as retatrutide)FDA. Accessed .
Written and maintained by the Best Peptides For Editorial Team. How we rate evidence.