Semaglutide Body Composition: What the STEP Trials Really Showed (It’s Not What You Think)
TL;DR
- The semaglutide body composition story isn’t the 14.9% headline — it’s what that weight loss was made of, and the honest number is about 25% lean tissue, not the scarier figures floating around.
- The STEP 1 DXA substudy showed absolute lean mass fell 9.7%, but lean mass as a proportion of the body actually rose — fat dropped faster than muscle.
- The robust cross-trial estimate, pooled across 22 studies, is ~25% of weight lost being lean. That’s the number to plan around.
- These outcomes came from people with no preservation strategy. They’re a baseline to beat, not a ceiling.
- Run the three levers and you can flip the result entirely — I gained lean mass across a 41-week deficit.
The headline from the STEP 1 semaglutide trial is 14.9% mean weight loss at 68 weeks. That number gets cited everywhere — news articles, prescribing conversations, before-and-after posts. What rarely gets cited is what that weight loss was made of, and that’s where most people get the wrong impression in both directions.
This article goes behind the headline to the actual body composition data — what the DXA substudy showed, what the pooled analysis shows, and what it means in pounds for a real person. If you’d rather skip to the protocol that changes these outcomes, the free GLP-1 Starter Framework lays out the three levers.
The STEP 1 Body Composition Data
The primary STEP 1 trial enrolled 1,961 non-diabetic adults with BMI ≥30 (or ≥27 with a comorbidity), randomized to semaglutide 2.4mg/week or placebo for 68 weeks (Wilding JPH et al. NEJM. 2021;384:989-1002). Mean weight loss in the semaglutide arm: 14.9% of body weight.
To understand what that loss was composed of, you need the DXA substudy (140 participants). It found two things that have to be read together: lean body mass fell 9.7%, but lean mass as a proportion of total body weight rose about 3.0 percentage points (Wilding JPH et al. Journal of the Endocrine Society. 2021). In plain terms — absolute muscle dropped, but fat dropped faster, so the body got leaner as a ratio even while losing some lean tissue.
For the better estimate of how much of the weight lost is lean, the strongest data is a pooled analysis of 22 randomized trials and 2,258 participants (Karakasis P, Patoulias D, Fragakis N, Mantzoros CS. Metabolism. 2025;164:156113). It found lean mass at approximately 25% of total weight lost — a lean reduction of about 0.86 kg against 3.55 kg of total loss — and that the percentage of the body that was lean did not worsen. That ~25% is the figure to plan around.
Estimates vary across the literature by population, dose, trial length, and measurement method, and some trials report higher fractions (Neeland IJ, Linge J, Birkenfeld AL. Diabetes Obes Metab. 2024;26 Suppl 4:16-27). But the central, robust number is roughly a quarter — not the “nearly half” figure that circulates from smaller exploratory analyses. The full muscle-loss picture across GLP-1 trials covers the range and its caveats.
What That Means in Pounds
For a 200-pound man who loses 30 pounds on semaglutide (15% of starting weight), at the 25% pooled estimate: approximately 7.5 pounds of lean mass lost alongside 22.5 pounds of fat.
That 7.5 pounds is not a rounding error — it’s roughly the lean tissue equivalent of an arm, and it shows up as reduced strength, less definition, and a lower metabolic baseline. These aren’t reasons to avoid semaglutide. They’re reasons to treat lean mass preservation as a primary objective alongside fat loss, not an afterthought.
Frame it as information, not alarm: that 7.5-pound figure reflects what happens with no specific preservation strategy. It’s a baseline, not a ceiling — and the back half of this article is about what beating it looks like.
The Body Proportion Improvement — And Why It’s Not the Whole Story
Worth acknowledging: even with absolute lean mass declining, the proportion of the body that’s lean typically increased in STEP participants — the +3.0 proportion points from the substudy. Fat dropped faster than lean. The lean-to-fat ratio improved.
That’s a real positive. For someone focused on health markers, metabolic function, and disease risk, the proportion is the relevant outcome.
For a man who trains, cares about strength and physical performance, and doesn’t want to spend the next year rebuilding muscle — the absolute lean mass number matters more than the ratio. Losing several pounds of lean mass isn’t trivial, even if the proportion improved.
Both things are true. Overall body composition improved. The absolute lean loss was meaningful. Which metric matters most depends on your goals — and what “body recomposition” actually means takes that distinction apart in full.
How to Use These Numbers
The STEP data and the pooled figure are baseline references, not predictions about you specifically. They describe average outcomes in populations given no structured preservation intervention. Three things follow.
First: get a DEXA scan. Baseline before significant loss begins, retest at 12 weeks. Commercial services run $40–75 in most metro areas. The trend between scans — fat dropping while lean holds — is the real report card. Scale weight can’t give you that.
Second: treat the lean mass fraction as a tracking target. Lose 20 pounds over three months with 5 pounds of lean loss and you’re at 25% — the trial average, achieved with no deliberate strategy. If you’re training and hitting protein, you should beat it. If you’re not, that’s diagnostic.
Third: don’t let the proportion improvement obscure the absolute number. The ratio improving is metabolically real; losing several pounds of muscle to get there still has consequences. Both are true.
What Changes These Outcomes
The STEP trials enrolled participants with general lifestyle guidance and no specific training or protein targets. The body composition outcomes are what they are under those conditions. They don’t reflect what a deliberate protocol produces.
Here’s the contrast that matters to me. Across roughly 41 weeks of deficit, I didn’t lose lean mass — I gained it, at about 0.10 pounds per week, for something close to four pounds of lean added while I was actively cutting fat. That’s DEXA-tracked, though I treat it as a hovering estimate rather than a precise figure, and the rate is slow enough that it’s measured in patience, not in any given week. Your mileage may vary — body recomposition is easier earlier in a training career and with more fat to lose. But it’s the difference between the trial baseline and what the three levers can do: the STEP participants lost a quarter of their weight as lean tissue, and a tracked protocol can move that number to zero or past it.
A case series by Tinsley GM and Nadolsky S (Sage Open Med Cases. 2025) is directionally consistent: three GLP-1 users training three to five days per week at 1.6–2.3g protein per kg of fat-free mass all preserved or gained lean mass. Three people is a case series, not a trial — but the direction matches what muscle physiology predicts when the inputs are in place.
The three levers — protein, resistance training, and rate of loss — are what change these numbers. How tirzepatide compares on the same metrics looks at the other major drug. The trial data tells you what happened without the levers. The full framework is where they become a protocol.
Get the Framework
Download the free GLP-1 Starter Framework — the three-lever system for losing fat without losing muscle. It’s the protocol that moves you off the trial baseline.
Download the free GLP-1 Starter Framework →
Want the full evidence base — every trial behind these numbers, sourced and caveated in one place? That’s GLP-1 & Body Composition: What the Research Actually Says.
FAQ
How much of semaglutide weight loss is muscle?
The most robust estimate, pooled across 22 trials, is about 25% of total weight lost — roughly 0.86 kg of lean against 3.55 kg of total loss on average (Karakasis et al., Metabolism, 2025). A smaller STEP 1 substudy showed absolute lean mass falling 9.7% while lean proportion rose. Individual results vary widely with protein, training, and rate of loss; people running a preservation protocol routinely come in well below 25%.
What did the STEP 1 trial show about body composition?
STEP 1’s DXA substudy of 140 participants found absolute lean body mass fell 9.7%, but lean mass as a share of total body weight rose about 3 percentage points — fat dropped faster than muscle (Wilding et al., Journal of the Endocrine Society, 2021). So body composition as a ratio improved even though some muscle was lost. The participants received only general lifestyle guidance, with no structured training or protein targets, so the result reflects default conditions.
Does semaglutide improve body composition?
As a ratio, yes — in the trials, the proportion of the body that was lean increased because fat dropped faster than lean mass. That’s metabolically meaningful. But absolute lean mass still declined, which matters for strength and physique. So “improves body composition” is true on a percentage basis and incomplete on an absolute basis. For someone who trains, protecting absolute lean mass with protein and resistance training is what makes the improvement complete.
Is the muscle loss in the STEP trials a concern?
It’s a calibrated concern, not a reason to avoid the drug. Roughly 25% of weight lost being lean tissue, in people doing nothing to prevent it, tells you the default outcome is meaningful muscle loss. But it’s a baseline, not a fixed result. The trials enrolled participants with no preservation strategy. With adequate protein, resistance training, and a managed rate of loss, that fraction drops sharply — and can reach zero.
How do you avoid the STEP trial body composition outcome?
By doing the three things the trial participants weren’t asked to do: hit a protein target of roughly 1.6–2.2g per kg daily, train against resistance at least twice a week near failure, and keep the rate of loss in the 0.5–1.0% of bodyweight per week range. Track with a baseline and 12-week DEXA scan so you can see whether lean mass is holding rather than guessing from the scale.
Nothing on this site constitutes medical advice. I’m not a physician, and this blog documents my own research and experience. Consult a qualified healthcare provider for decisions about medication, dosing, or treatment.
— Ryan Mercer | MetabolicMale.com | ryanmercer@metabolicmale.com
Citations:
Wilding JPH et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity. NEJM. 2021;384:989-1002.
Wilding JPH et al. STEP 1 body composition substudy. Journal of the Endocrine Society. 2021.
Karakasis P, Patoulias D, Fragakis N, Mantzoros CS. Effect of glucagon-like peptide-1 receptor agonists and co-agonists on body composition: Systematic review and network meta-analysis. Metabolism. 2025;164:156113.
Neeland IJ, Linge J, Birkenfeld AL. Changes in lean body mass with glucagon-like peptide-1-based therapies and mitigation strategies. Diabetes Obes Metab. 2024;26 Suppl 4:16-27.
Tinsley GM, Nadolsky S. Preservation of lean soft tissue during weight loss induced by GLP-1 and GLP-1/GIP receptor agonists: A case series. Sage Open Med Cases. 2025.
