Tirzepatide Body Composition: SURMOUNT-1 & SURPASS-3 Imaging

If you searched for the “SURMOUNT-MORPH” tirzepatide imaging trial, here is the honest answer up front: there is no clinical trial literally named SURMOUNT-MORPH. The real, peer-reviewed evidence on what tirzepatide (Mounjaro, Zepbound) actually does to your body composition — visceral fat, subcutaneous fat, liver fat, and lean muscle — comes from two named imaging substudies: the SURMOUNT-1 DXA body-composition substudy in people with obesity (Look 2025 ^[1]) and the SURPASS-3 MRI substudy in people with type 2 diabetes (Gastaldelli 2022 ^[2]), plus a 2025 muscle-composition re-analysis of SURPASS-3 (Sattar 2025 ^[3]). The headline that matters: in the SURMOUNT-1 DXA substudy, tirzepatide cut visceral fat mass by about 40% versus roughly 7% on placebo, and about 74% of the weight lost was fat mass (only ~26% lean), while the SURPASS-3 MRI substudy showed large drops in liver fat and abdominal fat depots ^[1][2]. This deep-dive walks through every number these imaging studies reported and what they mean for the “am I losing fat or muscle?” question.

The honest summary

• The body-composition data are from DXA and MRI substudies, not a standalone “MORPH” trial. DXA (dual-energy X-ray absorptiometry) measured fat vs lean mass in SURMOUNT-1; MRI measured liver fat and abdominal fat depots in SURPASS-3 ^[1][2].
• Most of the weight lost was fat. In the SURMOUNT-1 DXA substudy, about 74% of the weight lost on tirzepatide was fat mass and ~26% was lean mass — essentially the same fat-to-lean ratio as placebo (75%/26%), despite far greater total weight loss on tirzepatide ^[1].
• Visceral fat dropped the most. Visceral fat mass fell ~40% on tirzepatide vs ~7% on placebo — a disproportionately large reduction in the metabolically dangerous fat around the organs ^[1].
• Liver fat fell sharply (SURPASS-3 MRI). Pooled higher-dose tirzepatide reduced liver fat content by about 8 absolute percentage points (roughly a 30%+ relative drop from a ~16% baseline), versus a small reduction on insulin degludec ^[2].
• Lean mass goes down too — proportionally. Muscle volume fell roughly in line with overall weight loss in SURPASS-3, while muscle fat infiltration (myosteatosis) actually improved — a quality-of-muscle gain (Sattar 2025 ^[3]).
• Bottom line: the imaging evidence is reassuring on fat-loss quality but is a strong argument for protecting muscle deliberately — through resistance training and adequate protein — because some lean loss is real.

Why imaging substudies exist at all

A number on the scale can't tell you what you lost. Two people can both lose 20 kg and have completely different outcomes if one lost mostly fat and the other lost a worrying amount of muscle. That is the entire reason drug developers run imaging substudies: the scale measures total body weight, but DXA and MRI can separate that weight into fat mass, lean (mostly muscle) mass, visceral fat (the fat packed around your abdominal organs), subcutaneous fat (under the skin), and even liver fat and fat inside the muscle. For GLP-1 and dual-agonist drugs like tirzepatide, the loudest public worry has been muscle loss — so the morphology data are exactly what people are searching for.

Two named trials carry most of this evidence. SURMOUNT-1 was the pivotal obesity trial of tirzepatide — 2,539 adults with obesity or overweight, 72 weeks, the trial that produced up to ~21% mean weight loss and led to Zepbound's approval (Jastreboff 2022 ^[4]). A subset of those participants had whole-body DXA scans at baseline and week 72, and that became the SURMOUNT-1 DXA body-composition substudy (Look 2025 ^[1]). Separately, SURPASS-3 studied tirzepatide in type 2 diabetes, and its MRI substudy imaged liver fat and abdominal fat depots (Gastaldelli 2022 ^[2]), with a later post-hoc analysis zooming in on the muscle (Sattar 2025 ^[3]).

SURMOUNT-1 DXA substudy: the fat-vs-lean walkthrough

The DXA substudy enrolled 255 participants from SURMOUNT-1; 160 had usable baseline and week-72 scans (Look 2025 ^[1]). In that subgroup, mean total body weight changed by about −21.3% on tirzepatide versus −5.3% on placebo — consistent with the parent trial. The substudy's job was to break that weight change into its tissue components using whole-body DXA.

Read the table carefully, because the framing matters. Tirzepatide reduced fat mass by an estimated treatment difference (ETD) of about 25.7% more than placebo, while lean mass fell by an ETD of about 8.3% more than placebo ^[1]. So lean mass did decline — that is real and worth taking seriously. But the proportion of total weight lost that came from fat (~74%) versus lean (~26%) was essentially identical to placebo (about 75%/26%), even though people on tirzepatide lost roughly four times as much total weight ^[1]. In plain terms: the drug did not shift the mix of tissue lost toward muscle; it just produced much more of the same kind of weight loss you'd see with diet alone. That 25%-ish lean fraction is also broadly in line with what's seen in non-pharmacologic weight loss.

SURPASS-3 MRI substudy: liver fat and abdominal depots

Where SURMOUNT-1 used DXA in people with obesity, the SURPASS-3 MRI substudy used magnetic resonance imaging in 296 adults with type 2 diabetes, comparing tirzepatide (5, 10, 15 mg weekly) against once-daily insulin degludec over 52 weeks (Gastaldelli 2022 ^[2]). MRI is the gold standard for quantifying liver fat content and for measuring visceral (VAT) and abdominal subcutaneous (ASAT) adipose tissue volumes precisely.

Participants started with a mean liver fat content of about 15.7% — well into fatty-liver territory (normal is under ~5%). Pooled higher-dose tirzepatide (10 mg + 15 mg) reduced liver fat by an absolute −8.09%, versus −3.38% on insulin degludec — a treatment difference of about −4.71 percentage points (95% CI −6.72 to −2.70; p<0.0001) ^[2]. Relative to that ~16% baseline, that is roughly a 30%+ proportional reduction in liver fat. Critically, both VAT and ASAT volumes fell on tirzepatide while they rose on insulin degludec, and the liver-fat reduction correlated with the reductions in visceral and subcutaneous fat ^[2]. This is the imaging signature of genuine fat redistribution, not just water or temporary weight loss.

What about the muscle? The SURPASS-3 muscle re-analysis

The muscle-loss worry deserved its own analysis, so a 2025 post-hoc study re-examined the SURPASS-3 MRI scans specifically for thigh muscle volume, a muscle-volume Z-score (which normalizes for sex, height, weight and BMI), and muscle fat infiltration — fat stored inside the muscle, a marker of muscle quality known as myosteatosis (Sattar 2025 ^[3]).

The findings were nuanced and, on balance, reassuring. Muscle volume did decrease, but the decline appeared proportional to overall weight loss rather than disproportionate — and the muscle-volume Z-score (the height/weight-adjusted metric) did not show the kind of catastrophic muscle wasting some feared. Meanwhile, muscle fat infiltration improved versus population-based expectations — meaning the muscle that remained was, by this imaging marker, of better quality (less fatty) (Sattar 2025 ^[3]). A companion editorial framed it as muscle composition staying stable or improving without significantly compromising skeletal muscle. The upshot: some loss of muscle volume is real and expected with large weight loss, but the imaging does not support the idea that tirzepatide is uniquely destructive to muscle.

What this means for you

• The fat-loss quality is good. About three-quarters of the weight you lose on tirzepatide is fat, and the riskiest fat — visceral and liver fat — falls disproportionately ^[1][2]. That is the metabolically meaningful kind of weight loss.
• Lean loss is real but not abnormal. The ~26% lean fraction matches diet-based weight loss; tirzepatide doesn't make the muscle-to-fat loss mix worse, but it doesn't make it better either ^[1].
• Protect muscle on purpose. Because some lean and muscle-volume loss is expected, resistance training plus adequate protein are the evidence-aligned countermeasures. See our guides on preventing GLP-1 muscle loss and pairing exercise with a GLP-1 to preserve lean mass.
• Consider supplemental strategies cautiously. Evidence on adjuncts like creatine is still emerging — see creatine and lean-mass preservation on a GLP-1 — and dedicated muscle-preserving drugs like bimagrumab are still investigational (bimagrumab evidence).
• Don't over-read a single substudy. DXA and MRI substudies enroll subsets of the main trial, so confidence intervals are wider and the populations (obesity vs. type 2 diabetes) differ. Treat the numbers as directionally strong, not precise to the decimal.

Bottom line

There is no trial called SURMOUNT-MORPH, but the real tirzepatide imaging evidence tells a clear story. In the SURMOUNT-1 DXA substudy, tirzepatide reduced visceral fat mass by about 40% and produced weight loss that was roughly 74% fat and 26% lean — the same favorable fat-to-lean ratio as placebo, at four times the magnitude (Look 2025 ^[1]). The SURPASS-3 MRI substudy added that liver fat fell by about 8 absolute percentage points and visceral and subcutaneous fat depots shrank, while they grew on insulin (Gastaldelli 2022 ^[2]). Muscle volume declines proportionally with weight loss, but muscle quality (fat infiltration) improves (Sattar 2025 ^[3]). Net: the imaging supports tirzepatide as high-quality fat loss, with a clear, actionable caveat — train and eat to protect the muscle you keep.

This article is educational and is not medical advice. There is no trial literally named “SURMOUNT-MORPH”; the data above are drawn from the SURMOUNT-1 DXA body-composition substudy, the SURPASS-3 MRI substudy, and its muscle-composition post-hoc analysis. Every PMID and figure was verified against the live PubMed database and the published abstracts before publication. Discuss your own body-composition goals and muscle-preservation plan with your prescriber.