GLP-1s and lean mass: 2026 trials narrow the question without closing it

The question of what incretin drugs do to muscle has been argued mostly with two numbers pulled out of trial appendices. In March 2026 a meta-analysis put a proper denominator under them.

Researchers pooled 20 randomised controlled trials comprising 15,782 participants that reported body composition by DXA or MRI, searching to 20 January 2026, and asked what proportion of total weight lost was lean mass [s1]. With incretin agonists, lean mass made up 25% to 39% of total weight lost: 35.2% with semaglutide (95% CI 31.5–38.9), 25.4% with tirzepatide (22.8–28.0) and 26.8% with liraglutide (23.1–30.5) [s1]. Intensive lifestyle intervention produced a comparable proportion at 26.2% (24.1–28.3), with p = 0.42 for the comparison [s1]. Lifestyle intervention combined with resistance training showed the most favourable profile at 17.5% (14.2–20.8) [s1]. Heterogeneity was moderate (I² = 68%) and Egger's test found no publication bias (p = 0.23) [s1].

That result reframes the debate. The share of weight coming off as lean tissue is substantial with these drugs — but not obviously more substantial than when the same amount of weight comes off by other means.

The supporting and complicating evidence

A paper in Cell Reports Medicine combined four preclinical studies with a proof-of-concept clinical trial registered as NCT05606471 [s2]. In obese mice, GLP-1 medicines predominantly reduced fat with a small but statistically significant decrease in lean body mass; among lean tissues, liver mass loss exceeded the change in muscle mass [s2]. Absolute muscle mass and strength fell while relative muscle mass and strength improved, and running performance improved with them [s2]. The authors also report that although atrophy during immobilisation was similar, GLP-1 medicines produced a distinct muscle proteome compared with calorie restriction [s2]. In the clinical component, patients with obesity improved body composition without a negative effect on strength [s2]. The paper's own framing is middle-aged mice and men: absolute muscle values decrease slightly while body composition and mobility improve [s2].

A pharmacovigilance analysis points the other way, and it is worth being explicit about what that kind of analysis can and cannot do. Using the FDA Adverse Event Reporting System from October 2003 to March 2024, researchers identified 142 reports of muscle atrophy in association with GLP-1 receptor agonist therapy, mostly in adults aged 18 to 64 [s4]. Disproportionality analysis produced signals for semaglutide (reporting odds ratio 2.39, 95% CI 1.63–3.52) and tirzepatide (ROR 1.69, 95% CI 1.14–2.50), while exenatide (0.26, 0.12–0.55) and liraglutide (0.27, 0.09–0.83) showed significantly lower reporting odds [s4]. The authors state directly that these findings should be read as signal detection rather than evidence of causality, and call for studies using objective measures of muscle mass and function [s4]. Spontaneous reporting databases reflect what gets reported, which tracks media attention and prescribing volume as much as biology.

What a drug designed for the problem showed

The clearest demonstration that the lean mass loss is real, measurable and modifiable came from a phase 2 trial of an antibody built to target it. BELIEVE randomised 507 adults with obesity to nine groups — placebo, bimagrumab at 10 or 30 mg/kg intravenously every 12 weeks, semaglutide at 1.0 or 2.4 mg weekly, and combinations — for 48 weeks, followed by an open-label extension to week 72 [s3].

At week 48, least-squares mean weight change was −9.3 kg with bimagrumab 30 mg/kg, −14.2 kg with semaglutide 2.4 mg, −17.8 kg with the high-dose combination and −3.3 kg with placebo, all significant versus placebo at p < 0.001 [s3]. The body composition split is the point. DXA lean mass changed by −0.9% with placebo, +1.1% with bimagrumab 30 mg/kg, −6.9% with semaglutide 2.4 mg and −2.3% with the combination at week 48; by week 72 those figures were +2.5%, −7.4% and −2.9% respectively [s3]. Expressed as the proportion of weight loss coming from fat, week 48 values were 100% for bimagrumab, 71.1% for semaglutide alone and 92.3% for the combination [s3].

The trial's limits are stated by its authors. Semaglutide was open-label because a matched placebo was unavailable at initiation. Bimagrumab was given intravenously every 12 weeks, which may have contributed to early laboratory abnormalities and adverse events. Muscle was assessed by DXA rather than the MRI used in earlier bimagrumab studies. And there was no significant improvement in patient-reported outcomes or grip strength, which the authors attribute in part to the breadth of the population studied [s3]. Discontinuations due to adverse events ranged from 3.6% on placebo to 21.4% on bimagrumab 10 mg/kg, with muscle spasms, diarrhoea and acne common in bimagrumab groups [s3]. This is a phase 2 trial of an investigational drug, not an approved option.

The gap for people who train

Every one of these datasets was generated in populations selected for obesity, not for training status. The evidence closest to the question of what happens to someone lifting seriously while on these drugs is thin enough to describe individually.

A case series of three patients from a virtual clinic, with DXA scans across treatment, reported one case losing 8.7% of total weight as lean soft tissue, one gaining 2.5% lean soft tissue despite 26.8% weight loss, and one gaining 5.8% despite 13.2% weight loss [s6]. All three performed resistance training three to five days per week and consumed 0.7 to 2.3 g of protein per kilogram per day [s6]. The same paper summarises the reference trials: in STEP 1, roughly 40% of weight lost was lean soft tissue (about 6.9 kg lean against 10.4 kg fat), and in SURMOUNT-1 roughly 26% (about 5.6 kg against 15.9 kg) [s6]. Three patients, no control group, self-reported nutrition and exercise, and scans performed at different facilities — the authors say plainly that the findings are not generalisable [s6].

The trial designed to answer this is still enrolling. LEAN-PREP, a single-centre six-month randomised controlled trial at the Dasman Diabetes Institute, plans to enrol 232 adults with obesity randomised equally to control, resistance exercise, protein supplementation, or both, alongside semaglutide or tirzepatide therapy [s5]. Resistance exercise is home-based, three sessions weekly, progressing from one to three sets; protein supplementation targets 1.6 g/kg/day [s5]. The primary outcome is MRI-measured quadriceps cross-sectional area — a direct muscle measurement rather than a DXA lean-mass proxy [s5]. It is registered as NCT06885736 [s5].

What to watch

Three things. Whether LEAN-PREP's MRI-based primary outcome confirms or contradicts the DXA-based literature [s5]. Whether the pharmacovigilance signal survives contact with objective measurement, as its own authors request [s4]. And whether bimagrumab or a similar agent progresses beyond phase 2, which would turn a research finding about the composition of weight loss into a clinical decision [s3].

Sources

1. [s1] Eisa N, Barood O. Lean Mass Changes With Incretin Therapy Versus Lifestyle Intervention: A Systematic Review and Meta-Analysis of Randomised Controlled Trials. Diabetes, Obesity and Metabolism, published online 24 March 2026. https://doi.org/10.1111/dom.70666
2. [s2] Langer HT, Gilmore NK, Hayden CMT, et al. Weight loss with GLP-1 medicines does not result in a disproportionate loss of muscle mass or function in obese mice and humans. Cell Reports Medicine, March 2026. https://pmc.ncbi.nlm.nih.gov/articles/PMC13006392/
3. [s3] Heymsfield SB, Aronne LJ, Montgomery P, et al. Bimagrumab plus semaglutide alone or in combination for the treatment of obesity: a randomized phase 2 trial. Nature Medicine, 2 March 2026. https://pmc.ncbi.nlm.nih.gov/articles/PMC13004672/
4. [s4] Kwan ATH, Lakhani M, McIntyre RS. Muscle atrophy associated with glucagon-like peptide-1 receptor agonists: A population-based observational study. Clinical Nutrition, published online 4 March 2026. https://doi.org/10.1016/j.clnu.2026.106620
5. [s5] Alawadhi AA, Alroudhan D, Alsaeed DJ, et al. LEAN-PREP study protocol. BMJ Open, 22 April 2026. https://pmc.ncbi.nlm.nih.gov/articles/PMC13110620/
6. [s6] Preservation of lean soft tissue during weight loss induced by GLP-1 and GLP-1/GIP receptor agonists: A case series. SAGE Open Medical Case Reports, 16 October 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12536186/