GLP-1 drugs at obesity doses cut lean mass in kilograms but raise its share of body weight

GLP-1 receptor agonists now produce weight loss of 5% to 21%, but rapid weight loss can strip lean tissue and raise the risk of sarcopenic obesity, which is linked to falls, fractures, cardiovascular events and mortality. Earlier meta-analyses reported modest lean mass losses, but they pooled trials using the lower doses approved for diabetes. That may have underestimated the effect at the higher doses used for obesity.

This PROSPERO-registered review searched PubMed, Embase and LILACS through April 2025. It included placebo-controlled RCTs in adults with obesity (BMI above 30, with or without type 2 diabetes) using liraglutide 3 mg daily, semaglutide 2.4 mg weekly or tirzepatide 5 to 15 mg weekly. Body composition had to be measured by DXA, CT, MRI or bioimpedance. Seven RCTs with 821 participants qualified: five on liraglutide, one on semaglutide and one on tirzepatide. All were rated low risk of bias, and the authors used random-effects models and GRADE.

Lean mass as a proportion of total weight rose by 1.81% with GLP-1 treatment (95% CI 1.1 to 2.52; I² = 7%). In subgroups, liraglutide for under 52 weeks gave +1.51% and semaglutide gave +2.95%. Absolute lean mass fell by 1.74 kg versus placebo (95% CI -3.04 to -0.45) and percentage lean mass fell by 3.06% (95% CI -5.10 to -1.02), both with I² = 98%. Semaglutide showed the largest losses: -5.44 kg and -9.9%. Liraglutide lost -0.41 kg under 52 weeks and -2.65 kg beyond 52 weeks.

Tirzepatide could not be pooled because the SURMOUNT-1 DXA subgroup data were not available by dose. The trial reported lean mass changes of -10.9% versus -2.0% with placebo, a treatment difference of -8.3%. Fat mass fell roughly three times as much as lean mass. Across the included trials, lean mass accounted for 14% to 22% of weight lost with liraglutide, up to 45% with semaglutide and about 26% with tirzepatide. An MRI analysis from SURPASS-3 showed muscle volume falling by 0.64 L.

The authors conclude that lean mass loss should not be treated as a barrier to these drugs, but they stress pairing treatment with resistance exercise and adequate protein. Placebo participants also lost lean mass, probably from age-related decline, while many gained weight and fat, so untreated obesity worsened body composition. Gastrointestinal adverse events were common: semaglutide 74.2% vs 47.9% with placebo, liraglutide 53% vs 8%, tirzepatide up to 81.8% vs 72%. Discontinuations for adverse events ran 7% to 9%. Serious events were rare.

Several caveats limit these conclusions. No included trial measured strength or physical function, and lean mass is only a proxy for muscle. Heterogeneity was very high for the absolute and percentage outcomes (I² = 98%), and the semaglutide and long-term liraglutide results each rest on one study. Tirzepatide data were incomplete, participants were young and mostly non-diabetic, and trials included diet and activity run-ins. The authors themselves note the 1.8-point gain in lean proportion may not be clinically meaningful. A funnel plot was not possible with fewer than ten studies.