Across 41 trials, tirzepatide cut the most fat, and the most lean mass

Concern that GLP-1 and dual GIP/GLP-1 drugs cause meaningful loss of lean muscle mass alongside fat loss has become a prominent thread in the broader conversation about these drugs, though most individual trials weren't designed to directly compare body-composition effects across the full range of antidiabetic drug classes. A network meta-analysis published this month in Diabetes, Obesity and Metabolism pools 41 randomized trials to make that comparison directly [s1].

The design

Researchers searched PubMed, Web of Science, and Scopus from inception through March 2025 for randomized controlled trials reporting body composition outcomes [s1]. A frequentist random-effects network meta-analysis was used to estimate mean differences and confidence intervals for changes in fat mass and lean body mass across drug classes — a method that allows indirect comparison between drugs even when they haven't been tested head-to-head in the same trial, by triangulating through trials that share a common comparator [s1]. Forty-one trials involving 2,906 participants were included [s1].

What it found

Among GLP-1 and dual GIP/GLP-1 receptor agonists, tirzepatide produced the largest fat mass reduction compared with placebo: mean difference −10.70 kg (95% CI −13.42 to −7.99) [s1], followed by liraglutide combined with exercise. Semaglutide and liraglutide alone produced more moderate fat mass reductions [s1]. In contrast, insulin glargine and alogliptin were associated with fat mass gain relative to metformin and exenatide [s1].

On lean body mass, tirzepatide again showed the largest effect — but in the loss direction: mean difference −4.40 kg (95% CI −7.58 to −1.22) [s1]. Liraglutide alone was also associated with significant lean mass loss (mean difference −1.54 kg, 95% CI −2.55 to −0.52) [s1]. SGLT2 inhibitors caused minor lean mass losses, while metformin, insulin regimens, and DPP4 inhibitors showed neutral effects on both fat and lean mass [s1].

The exercise finding is the most clinically actionable detail here

Among the specific comparisons pooled, liraglutide combined with exercise ranked just behind tirzepatide for fat mass reduction [s1] — and the analysis's discussion frames exercise as mitigating liraglutide's lean-mass loss specifically [s1], suggesting that pairing a GLP-1 drug with structured exercise may address, at least partially, the very trade-off (fat loss coming with a lean-mass cost) that has drawn the most criticism. This is the kind of modifiable-factor finding that matters more to a treating clinician than the drug rankings alone, since it points toward exercise as a companion intervention rather than treating body-composition trade-offs as an unavoidable feature of the drug itself.

Reading the tirzepatide numbers side by side

Tirzepatide's fat mass reduction (−10.70 kg) is roughly 2.4 times larger than its lean mass reduction (−4.40 kg) — meaning the majority of weight lost on tirzepatide in this pooled analysis was fat rather than muscle, even though the absolute lean-mass loss is not trivial. That ratio is a useful frame for interpreting concerns about GLP-1-related muscle loss: some lean mass reduction does accompany treatment, consistent with what's typically observed during any substantial weight loss regardless of method, but it doesn't appear to represent the majority of the weight lost.

What this doesn't establish

Network meta-analysis relies on indirect comparisons across trials that differ in population, duration, dosing, and measurement methods for body composition (DXA scans, bioelectrical impedance, and other methods aren't necessarily equivalent, and the paper's summary doesn't specify which methods dominated across the 41 pooled trials). This is also a comparison of average effects pooled across trials, not a single head-to-head trial directly comparing all these drugs in the same population under the same conditions. The paper's own authors call for further research on how dose, treatment duration, and lifestyle factors influence these body-composition effects [s1] — an acknowledgment that this analysis characterizes broad patterns rather than settling every open question about the mechanism or modifiability of GLP-1-related lean mass loss.

What to watch

Whether further trials specifically test exercise or resistance training as an adjunct to GLP-1 and dual-agonist therapy to preserve lean mass, building on the liraglutide-plus-exercise signal found here. This article is not medical advice.

Sources

1. Comparative Effects of Antidiabetic Drugs on Body Composition: A Systematic Review and Network Meta-Analysis — Diabetes, Obesity and Metabolism, 16 June 2026