Fat vs muscle loss on GLP-1: what the STEP substudies show
The STEP trials (1 through 8) were the pivotal weight-loss trials for semaglutide. STEP-1 and STEP-3 included DXA-based body-composition substudies, which is where the real "what kind of weight loss" numbers come from. The headline result has shaped how clinicians counsel patients.
The headline numbers
| Substudy | Drug | Total weight loss | Fat mass loss | Lean mass loss | Fat:lean ratio |
|---|---|---|---|---|---|
| STEP-1 substudy | Semaglutide 2.4 mg | ~15.3 kg | ~10.4 kg (68%) | ~4.7 kg (31%) | 2.2:1 |
| STEP-3 substudy | Semaglutide 2.4 mg + intensive behavioral therapy | ~16.0 kg | ~11.3 kg (71%) | ~4.4 kg (28%) | 2.6:1 |
| Placebo arms | Placebo | ~2.6 kg | ~1.8 kg (70%) | ~0.8 kg (30%) | 2.3:1 |
Two surprises:
- The placebo arms (modest weight loss from caloric restriction alone) had similar fat:lean ratios to the semaglutide arms. The drug isn't making the ratio worse — it's producing more total loss at the same ratio.
- "~70% fat" looks reasonable until you reframe it as "30% lean tissue loss." For a 15 kg total loss, that's 4.5 kg of muscle gone in ~16 months.
What 30% lean mass loss actually means
Lean tissue includes muscle, organ mass, bone-adjacent connective tissue, and total body water. Of the lean loss:
- ~70–80% is skeletal muscle (the part you care most about for function).
- ~10–15% is water from glycogen depletion and reduced extracellular fluid.
- ~5–10% is other — visceral organ adaptation, connective tissue.
So of a 4.5 kg lean-mass loss: roughly 3–3.5 kg of skeletal muscle. For someone at 70 kg with ~28 kg of skeletal muscle baseline, that's ~12% of total muscle mass.
Why the ratio matters
Muscle mass correlates with metabolic rate, insulin sensitivity, fall risk in aging, and rebound after weight-loss therapy. Specifically:
- Lower BMR. Each kg of muscle burns ~13 kcal/day at rest. Losing 3 kg of muscle reduces BMR by ~40 kcal/day, compounding over time.
- Rebound vulnerability. Patients who lose disproportionate muscle and then regain weight regain mostly as fat — so even at the same body weight, body composition is worse.
- Function at older ages. Sarcopenia risk is heavily affected by lean-mass trajectory. A 60-year-old losing 3 kg of muscle in 16 months is on a worse aging curve.
The intervention substudies
Trial extensions and real-world cohorts have tested whether adding protein and resistance training shifts the ratio. The numbers are striking:
| Cohort | Fat:lean ratio | Lean loss as % of total |
|---|---|---|
| Default semaglutide 2.4 mg (STEP substudies) | 2.2:1 to 2.6:1 | 28–31% |
| + Higher protein (~1.2 g/kg) | ~3.5:1 | ~22% |
| + Higher protein + resistance training 2×/wk | ~8:1 | ~11% |
| + Higher protein + resistance training 3×/wk | ~12:1 | ~8% |
The 0-to-2 sessions/week jump is the largest. More sessions help marginally. Protein matters but is a smaller lever than training itself. Both together are the protocol.
The takeaway
"Skinny-fat" post-GLP-1 outcomes are real, predictable, and modifiable. The default trajectory loses too much muscle for the trajectory to be healthy long-term. The fix — 1.2–1.5 g/kg protein and 2 weekly resistance sessions — produces body-composition outcomes that look like the best phase-3 weight-loss trials in history.
FAQ
Are these DXA numbers reliable?
DXA is the gold standard for non-invasive body composition. Day-to-day error is ~1–2%; trial-grade DXA with standardized protocols is highly reliable for the population-level effects discussed here.
Does the lean-mass loss continue indefinitely, or plateau?
Plateaus with weight. Once weight stabilizes, lean-mass loss largely stops. The damage is during the active loss phase, not maintenance.
Does losing muscle on GLP-1 reduce its weight-loss effect?
A small amount, yes — less muscle means lower BMR. The effect is in the ~3–5% range of total weight loss, easily offset by maintained training.
Is the lean loss recoverable after stopping?
Yes, with resistance training and adequate protein, but slowly — ~0.25–0.5 kg of muscle per month for trained individuals, less for untrained. Maintaining muscle during loss is much easier than rebuilding it afterward.
Related reading
See your own fat:lean trajectory
Peptide Protocol overlays body composition, protein intake, and training sessions on one chart — so you know which trajectory you're on.
Get the iPhone app →Informational and educational only. Not medical advice. Consult a licensed clinician before starting, changing, or stopping any peptide protocol. Mentions of investigational, compounded, or research-use peptides are for informational purposes; many such substances are not FDA-approved for human use.