Does GLP-1 Weight Loss Cause Muscle Loss? What the Lean-Mass Data Shows for Semaglutide and Tirzepatide

GLP-1 and GIP/GLP-1 receptor agonists produce weight loss primarily by suppressing appetite and reducing caloric intake. When the body is in a sustained energy deficit, it draws on both fat and lean tissue. This is not unique to these drugs - it is the expected physiology of any large, relatively rapid weight loss, including from bariatric surgery or aggressive caloric restriction.

The catch is in how the number is reported. When a headline says '40% of the weight lost was muscle,' the underlying measurement is almost always lean body mass on DXA (dual-energy X-ray absorptiometry, an imaging method that partitions body weight into fat, lean soft tissue, and bone). Lean body mass is a broad bucket. It includes skeletal muscle but also water, glycogen-bound water, organ mass, and connective tissue. A person losing weight sheds intracellular and extracellular water and shrinks enlarged organs alongside actual muscle. So the DXA 'lean' figure sets an upper bound on muscle loss, not a precise measurement of it. The literature is consistent that some genuine muscle is lost; it is far less consistent on how much.

The pivotal semaglutide obesity trial, STEP 1, reported a mean body-weight reduction of roughly 15% at 68 weeks in adults with overweight or obesity. A body-composition substudy within STEP 1 used DXA to partition that loss. The substudy reported that both fat mass and lean body mass declined, with lean mass making up a substantial minority of the total - a proportion commonly summarized as being in the range of about 40%. Importantly, because both compartments shrank, the ratio of fat mass to total mass actually improved: participants ended with a lower percentage of body fat than they started with.

That last point is why researchers caution against reading the lean-mass percentage in isolation. Losing lean mass while improving the fat-to-lean ratio is the normal signature of weight loss in people carrying excess adiposity. The STEP 1 substudy was not powered or designed to measure muscle strength or physical function, so it cannot tell us whether the lean-mass change translated into any functional decline in the studied population.

SURMOUNT-1, the pivotal obesity trial of the dual GIP/GLP-1 agonist tirzepatide, reported mean weight reductions of roughly 15% to 21% across doses over 72 weeks - the largest reductions seen in a Phase 3 incretin obesity trial at the time. A DXA body-composition substudy again partitioned the loss.

The tirzepatide substudy reported that the majority of weight lost was fat mass, with lean mass making up a smaller share of the total than has typically been cited for semaglutide - figures often summarized as a fat-to-lean loss ratio of roughly 3 to 1. As with STEP 1, total fat percentage fell, indicating a favorable shift in composition despite absolute lean-mass loss. Whether the somewhat more favorable ratio reflects a real pharmacologic difference between the two drug classes or simply differences in trial populations, magnitude of loss, and DXA substudy design remains an open research question. Head-to-head body-composition data are limited.

The table below summarizes reported DXA body-composition findings. All percentages are approximate, drawn from substudies with modest sample sizes, and describe lean body mass rather than verified skeletal muscle. They are best read as directional, not precise.

The reference row matters: a lean-mass share in the 20% to 40% band is broadly what the older weight-loss literature would predict for any intervention producing this magnitude of loss over this timeframe. The GLP-1 numbers are not obviously outside that historical envelope.

If the research question is 'can lean mass be preserved during a large energy deficit,' the GH axis (the growth-hormone / IGF-1 signaling system that regulates protein synthesis and lipolysis) is the most-studied pharmacologic lever. A 2026 narrative review in the Journal of Clinical Medicine surveyed drug classes being investigated specifically to preserve lean body mass during weight loss, naming candidates across several mechanisms - the myostatin/activin pathway (bimagrumab), selective androgen receptor modulators (enobosarm), and the GH axis (tesamorelin).

Tesamorelin is a synthetic analogue of GHRH (growth-hormone-releasing hormone, the hypothalamic peptide that stimulates the pituitary to release GH). By raising endogenous GH and IGF-1 in a pulsatile pattern, GHRH analogues are studied for their effects on lipolysis and on lean-tissue anabolism. The mechanistic rationale for pairing a GH-axis peptide with a GLP-1 agent is straightforward on paper: the GLP-1 drives the caloric deficit and fat loss, while the GH-axis signal is hypothesized to bias the tissue that is retained toward lean mass. That rationale is a hypothesis under study, not an established clinical result.

Tesamorelin has the deepest clinical dossier of the GH-axis peptides, but that dossier was built in a different context. Its Phase 3 program studied HIV-associated visceral fat accumulation, where daily 2 mg tesamorelin reduced visceral adipose tissue (the metabolically active fat packed around the abdominal organs) by about 15% versus an increase on placebo, alongside rises in IGF-1. Subsequent research extended the visceral-fat and liver-fat findings.

What that body of work does not include is a trial of tesamorelin as a muscle-sparing add-on to semaglutide or tirzepatide. The lean-mass-preservation hypothesis for GH-axis peptides in the GLP-1 setting is drawn from mechanism and from adjacent trial data, not from a completed head-to-head study in people losing weight on incretins. This is the single most important caveat for any researcher evaluating the compound for this use case.

Several limitations bound every claim above.

DXA is not myography. The 'lean mass' number conflates skeletal muscle with water and organ tissue. None of the cited trials biopsied muscle or measured cross-sectional muscle area directly, so the true fraction of contractile muscle lost is unknown and likely lower than the headline lean-mass share.

Function was not the endpoint. The pivotal trials measured weight and metabolic markers, not grip strength, gait speed, or muscle power. Whether the observed lean-mass changes matter functionally, and for whom (younger vs. older, sedentary vs. resistance-training subjects), has not been established in these datasets.

Preservation agents lack outcome trials in this setting. Tesamorelin, bimagrumab, and enobosarm are at various research stages for lean-mass preservation. As of this review, no GH-axis peptide has a completed, adequately powered trial demonstrating that it preserves function-relevant muscle when added to GLP-1 therapy. Mechanistic plausibility is not efficacy.

RUO context. The peptides discussed here are research compounds. Nothing in this article should be read as a claim that any compound prevents muscle loss, treats disease, or is safe or effective for human use.

The defensible summary of the current literature is narrow. GLP-1 and GIP/GLP-1 weight loss does reduce lean body mass, at a share of total loss that appears consistent with the general physiology of rapid weight reduction rather than a drug-specific toxicity. The clinical significance of that lean-mass loss - whether it degrades strength or function - is not yet resolved by the pivotal trials.

The most active research frontier is the second question: whether combining an incretin agent with a lean-mass-preservation agent shifts the fat-to-lean ratio of the loss. The GH axis, and tesamorelin specifically, is a leading mechanistic candidate for that role, but the confirmatory combination trials do not yet exist. For laboratory and research audiences, that gap - a strong mechanistic hypothesis paired with an absence of outcome data - is precisely the space where careful, well-controlled preclinical and translational work is most warranted.

Explore the compound referenced in this review and related body-composition research:

Related research articles: