Does Mounjaro or Zepbound cause muscle loss? What the SURMOUNT-1 scans showed

Does Mounjaro cause muscle loss?

Tirzepatide users lose lean mass as well as fat, and the SURMOUNT-1 body scans put lean mass at about 26% of the weight lost.

Mounjaro and Zepbound are brand names for the same drug, tirzepatide. The scan data come from a substudy of the SURMOUNT-1 trial, which gave tirzepatide to adults with obesity or overweight and no diabetes for 72 weeks. Lean mass on a scan is not the same as muscle. It also counts water, organs and skin.

The drug has not been shown to attack muscle directly. Large weight loss commonly includes some lean-mass loss, so the useful questions are how much, and whether the share is unusual. For tirzepatide, the share matches placebo.

What did the SURMOUNT-1 scans show?

The SURMOUNT-1 DXA analysis included 160 adults, and lean mass fell 10.9% on tirzepatide.

DXA is dual-energy X-ray absorptiometry, a low-dose scan that splits the body into fat, bone and everything else, called lean mass. Scans were taken at baseline and at week 72 or early discontinuation; three missing week-72 measurements were imputed. The substudy covered 160 of the 255 people enrolled at the DXA sites. Average starting weight was 102.5 kg (226 lb) and BMI 38.0.

Item	Tirzepatide (5, 10, 15 mg pooled)	Placebo
Participants scanned	124	36
Body weight change	-21.3%	-5.3%
Fat mass change	-33.9% (-15.9 kg, 35 lb)	-8.2% (-3.6 kg, 8 lb)
Lean mass change	-10.9% (-5.6 kg, 12 lb)	-2.6% (-1.2 kg, 3 lb)
Starting fat mass	46.6 kg	49.4 kg
Starting lean mass	50.8 kg	53.0 kg

Source: Look et al., Diabetes, Obesity and Metabolism, 2025 (checked 2026-10-26). All participants also received lifestyle advice.

How much of the weight lost on Mounjaro is muscle?

About 26% of the DXA-measured tissue lost on tirzepatide was lean mass. The scans cannot say how much of that was skeletal muscle.

Here is the arithmetic using the kilogram figures from the table:

1. Fat lost: 15.9 kg.
2. Lean lost: 5.6 kg.
3. Total DXA tissue lost: 15.9 + 5.6 = 21.5 kg.
4. Lean share: 5.6 ÷ 21.5 = 0.26, or 26%.

The paper reports 74% fat and 26% lean on tirzepatide, and 75% fat and 25% lean on placebo. The 21.5 kg total is my own sum, since the paper gives no direct kilogram figure for total weight lost.

The share barely moved with dose in the paper’s reported dose analysis. Fat made up 75%, 72% and 75% of weight lost at 5, 10 and 15 mg. By age, it ranged from 74% to 76%. These are post hoc subgroups, some very small.

Is tirzepatide muscle loss different from semaglutide?

In the scan data, tirzepatide lost a smaller share as lean mass than semaglutide, about 26% against about 40%, but the absolute lean loss was similar.

Measure	Tirzepatide, SURMOUNT-1 (72 weeks)	Semaglutide 2.4 mg, STEP 1 (68 weeks)
Scanned on drug	124	95
Fat lost	15.9 kg	10.40 kg
Lean lost	5.6 kg	6.92 kg
Total tissue lost	21.5 kg	17.32 kg
Lean share of loss	26%	40%
Lean change from baseline	-10.9%	-9.7%

Sources: Look et al. 2025; STEP 1 exploratory analysis (Journal of the Endocrine Society, 2021) for the percentage; Diabetes & Metabolism Journal review (2026) for the STEP 1 kilogram figures. STEP 1 kilograms are from the trial-product estimand.

Lean mass fell by a similar amount in both trials, and slightly more as a percentage on tirzepatide. The lean share is lower on tirzepatide because the fat loss was larger, which makes the denominator bigger. A lower share does not mean less lean mass was lost.

Two separate trials with different participants, scanners and lengths cannot rank the drugs. The ratio and the absolute kilograms tell different stories, and both are needed. For the semaglutide side, see Does Ozempic cause muscle loss?. The wider evidence is collected in GLP-1 muscle loss studies.

Is lean mass the same as muscle?

No. DXA lean mass is all tissue that is neither fat nor bone, so it includes water, organs and skin as well as skeletal muscle.

A 2024 review by Tinsley and Heymsfield gives fat-free mass as about 74% water, 19% protein and 6.5% mineral. That is why a drop in lean mass can overstate how much muscle tissue was lost. It also means a scan cannot say whether you got weaker.

One MRI substudy looked at muscle directly. In SURPASS-3, adults with type 2 diabetes on tirzepatide for 52 weeks lost 0.6 L (6.9%) of fat-free muscle volume in women and 0.7 L (5.5%) in men. A 2025 systematic review described these falls as small and within UK Biobank reference ranges, and reported lower fat inside the muscle. Both substudies were Lilly-sponsored.

Is Mounjaro different from Zepbound for muscle?

Mounjaro and Zepbound contain the same drug, so the body-composition evidence applies to both, though the trials enrolled different groups.

SURMOUNT-1 enrolled adults with obesity or overweight and no diabetes. The SURPASS-3 MRI work enrolled adults with type 2 diabetes. If you take Mounjaro for diabetes, the MRI figures are the closer match.

What do newer tirzepatide body-composition papers add?

As of 2026-10-26, newer papers add little certainty. A 2026 meta-analysis found the evidence too thin to say whether GLP-1 medicines spare or cost muscle.

The Diabetes, Obesity and Metabolism meta-analysis (published 2026-08-17, searches to May 2026) included 19 studies but only two comparative studies on muscle quantity. It called the results highly imprecise, with low to very low certainty. It also said lean-mass reductions should not be read as sarcopenia, and that the data cannot rule out clinically relevant muscle loss in susceptible people.

A 2026 real-world study of 122 adults on tirzepatide for 24 weeks, with a low-carbohydrate diet and tailored activity, used bioimpedance. Weight fell from 95 to 80 kg, fat mass fell 35% and fat-free mass fell 5%. The authors reported skeletal muscle mass as preserved. There was no control group, and bioimpedance estimates are sensitive to hydration, especially during active weight loss.

What the tirzepatide scans cannot tell you

The SURMOUNT-1 scans measured how much mass was lost, not whether anyone became weaker or less able.

• The DXA substudy included post hoc exploratory analyses with small subgroups, and Lilly funded it. Several authors work for Lilly.
• Activity and food intake were not measured, so the data cannot say whether protein or training changed the result.
• DXA estimates body composition and cannot separate muscle from water or organs.
• Only 36 people on placebo were scanned.
• Objective strength and physical-performance tests were not reported in these body-composition analyses. The overall SURMOUNT-1 trial reported improved patient-reported physical function.

The summary: some lean mass is lost, the scans cannot say how much of it is muscle, and function matters more than a scan number. To protect it, build your plan around protein and resistance work: see how much protein to eat on a GLP-1 and can you build muscle on Ozempic, or browse the guides index. See a clinician if you notice weakness, falls or trouble climbing stairs.