Protein & Muscle
Muscle Loss on GLP-1s: What the Research Actually Says
Published July 19, 2026
“Ozempic makes you lose muscle” has become a headline genre of its own. The truth is more nuanced — and more actionable — than the scary version.
What the trials measured
Body-composition sub-studies give us the clearest numbers we have:
| Study | Drug | Lean mass share of total loss |
|---|---|---|
| STEP 1 sub-study (NEJM 2021) | Semaglutide 2.4 mg | ~39% |
| SURMOUNT-1 sub-study (NEJM 2022) | Tirzepatide | ~25% (reported range across analyses) |
Two things to hold at once:
First, this is not unique to the drugs. Weight loss from diet alone, or from bariatric surgery, typically shows lean-mass losses in a similar 20–40% band. The medication isn’t “eating your muscle” — rapid weight loss with inadequate protein and no strength stimulus is, regardless of how the weight loss happens.
Second, the absolute numbers are big because the weight loss is big. Losing 15% of body weight means the lean-mass share is a meaningful number of pounds — for a 220 lb person losing 33 lb, a 30% lean share is ~10 lb of lean tissue. For adults over 50, who are already losing muscle to age, that’s worth taking seriously.
What “lean mass” does and doesn’t mean
DEXA’s “lean mass” includes water, organ tissue, and glycogen — not just contractile muscle. Some early lean-mass drop is water and glycogen and partially reverses or stabilizes. Function is what matters: grip strength, chair stands, stair climbs. Some small studies show preserved strength despite lean-mass declines; others flag real functional loss in older adults. The research frontier is active, and anyone who claims certainty here is ahead of the data.
The two protective factors with the best evidence
- Adequate protein — see our protein target guide for the numbers. Low protein intake is the single most consistent correlate of worse lean-mass outcomes in deficit studies.
- Resistance training — in non-GLP-1 weight-loss research, adding structured resistance exercise reliably cuts lean-mass loss, sometimes by more than half. Trials testing this specifically on GLP-1s are running now; there is little reason to wait for them before picking up dumbbells twice a week.
Questions worth asking your prescriber
- Should I get a baseline body-composition scan (DEXA) before or early in treatment?
- What protein intake do you recommend for my kidneys, age, and goals?
- Can you refer me to a dietitian or physical therapist familiar with GLP-1 patients?
The medication handles the appetite. Muscle is the part of the journey that’s still yours to defend — and it responds to remarkably ordinary tools: enough protein, twice-weekly strength work, and patience.
Frequently asked questions
Is muscle loss from GLP-1 weight loss permanent?
Not necessarily. Muscle responds to training and adequate protein at any age, and some of the early lean-mass decline on DEXA scans is water and glycogen rather than muscle tissue. The best strategy is prevention — protein plus resistance training during the loss — but rebuilding afterward is also possible.
Do I need a DEXA scan before starting?
It's not required, but a baseline body-composition measurement makes your progress much more informative than the scale alone, especially for adults over 50. Ask your prescriber whether it's worth it in your case; some gyms and clinics offer DEXA scans affordably.
Is muscle loss worse on semaglutide or tirzepatide?
There's no rigorous head-to-head body-composition trial, so honest answer: nobody knows. The sub-study numbers (roughly 39% and 25% lean-mass share) come from different trials with different populations and methods, so comparing them directly is unreliable.
Will losing weight more slowly protect my muscle?
Plausible but unproven for GLP-1s specifically. In general weight-loss research, very rapid loss with low protein is the worst combination for lean mass. If weight is falling extremely fast and you're struggling to eat protein, that's a conversation with your prescriber about titration pace.
Written by Max, founder of Well Basecamp — see our editorial policy for how we source and verify what we publish.