Home  /  Blog

Muscle Loss on GLP-1s: What the Research Actually Says

6 min read · Evidence-based nutrition for life on GLP-1s

Key takeaway: In GLP-1 trials, lean mass typically accounts for about 25% to 39% of total weight loss, a share similar to other forms of rapid weight loss. DEXA-based "lean mass" overstates true muscle loss because it includes water, glycogen, and organs. In a SURMOUNT-1 substudy of tirzepatide, about 75% of weight lost was fat and 25% was lean mass, close to the split seen with placebo.

Few topics in the GLP-1 world generate more anxiety, and more confident misinformation, than muscle loss. You'll hear "these drugs melt your muscles" and "it's totally fine, don't worry about it." Neither is what the research shows. Here is what the trials found.

In the STEP 1 trial of semaglutide, a body-composition substudy of about 140 participants, published in JAMA Network Open in 2021, found total body weight fell 15.0% while fat mass dropped 19.3% and lean body mass dropped 9.7%. Lean mass as a share of total body weight actually rose by 3.0 points, because fat fell faster than lean.

At the American Diabetes Association's 2026 Scientific Sessions, obesity researcher Dr. Samuel Klein described a reanalysis correcting for the fat-free mass naturally lost from shrinking adipose tissue. That brought STEP 1's lean share of weight loss from about 39% down to about 30%. Klein also noted that across GLP-1 trials where participants lost at least 15% of body weight, the lean share ranged from about 25% with tirzepatide to about 39% with semaglutide, a normal range for weight loss generally.

In the SURMOUNT-1 trial of tirzepatide, a 2025 DEXA substudy led by Dr. Jennifer Look and colleagues (160 participants, week 72) found body weight down about 21.3%, fat mass down about 33.9%, and lean mass down about 10.9%. The split was about 75% fat and 25% lean, nearly identical to the placebo group. The lean-loss pattern tracks the weight loss itself, not something unique to the drug.

Then there's the ENDO 2025 presentation from Massachusetts General Hospital. In a small prospective observational study of 40 adults over three months, the semaglutide group lost about 6.3% of body weight, and nearly half (47.5%) was lean mass, versus 35.7% in the lifestyle group. Lower protein intake was associated with greater lean mass loss in the semaglutide group (R = 0.52). Older age and female sex were also associated with more lean loss.

That 47.5% number was widely shared without context. The context: 40 people, three months, early in treatment when weight loss is fastest and fluid shifts are largest. The group difference wasn't statistically significant. The protein finding is still the one I keep coming back to: lower protein intake was associated with greater lean mass loss. That's a lever you can actually pull.

The measurement problem nobody talks about

DEXA "lean mass" is not the same as skeletal muscle. It includes body water, glycogen (which is stored with water), organs, and other non-fat tissue. During weight loss, glycogen stores shrink, fluid balance shifts, and food volume in the gut drops. All of that shows up as "lean mass loss" on a DEXA scan without a single muscle fiber disappearing.

The honest summary: some true muscle loss happens during rapid weight loss, and the DEXA numbers overstate it by an amount nobody can precisely pin down. Reviews of the literature conclude the pattern is proportional to total weight loss and similar to caloric restriction alone. The drug isn't selectively attacking your muscles. The speed and size of the calorie deficit is doing most of the work, so the countermeasures are the same ones that have always worked.

Who's most at risk

Older adults, who lose muscle naturally with age and have less reserve. Postmenopausal women, flagged in the ENDO 2025 data. People starting with low muscle mass. And anyone whose protein intake has collapsed along with their appetite.

What you can actually do

Eat enough protein. The RDA is 0.8 grams per kilogram per day, but during weight loss, clinical guidance in the literature commonly lands higher, around 1.0 to 1.5 grams per kilogram for moderately active adults. Personalize your target with your doctor or dietitian.

Do resistance training. Bodyweight exercises, resistance bands, or light dumbbells two to three times a week give your body a reason to hold onto muscle.

Lose at a reasonable pace. Faster isn't better if a disproportionate share is lean tissue. If your weight is dropping extremely fast and protein is low, that's worth a conversation with your doctor.

Don't chase the scale alone. Strength, energy, and how your clothes fit tell you things the scale can't.

Frequently asked questions

How much muscle do you lose on a GLP-1?

In major trials, lean mass accounted for about 25% to 39% of total weight loss. DEXA-measured "lean mass" includes water, glycogen, and organs, so true skeletal muscle loss is lower. A reanalysis of STEP 1 accounting for fat-free mass in adipose tissue brought the semaglutide estimate to about 30%.

Do GLP-1s cause muscle loss, or is it just the weight loss?

The evidence points to the weight loss itself. In the SURMOUNT-1 substudy, the placebo group showed nearly the same fat-to-lean split as the tirzepatide group.

Can protein and strength training prevent muscle loss on GLP-1s?

Nothing is proven to prevent it completely, but protein is the best-supported lever. In the ENDO 2025 study, lower protein intake was associated with greater lean mass loss on semaglutide. Resistance training provides the mechanical signal that tells your body to keep muscle.

Keep your strength isn't just a tagline. It's the whole strategy.

Keep your strength.

Upkeep is being built around this exact challenge, starting with a gentle clear whey protein designed for small appetites. Join the waitlist at liveupkeep.com for early access.

Join the waitlist

This post is educational, not medical advice. Talk to your doctor or a registered dietitian before changing your diet or exercise routine, especially if you have medical conditions or take other medications.

References

  1. Wilding JPH, et al. JAMA Network Open. 2021. STEP 1 DEXA analysis, n = 140, 68 weeks: weight -15.0%, fat mass -19.3%, lean body mass -9.7%; lean mass share of body weight rose 3.0 points.
  2. Look M, et al. Diabetes Obes Metab. 2025. PMID: 39996356. SURMOUNT-1 DEXA substudy, n = 160, week 72: weight -21.3%, fat mass -33.9%, lean mass -10.9%; about 75% fat, 25% lean.
  3. Klein S. ADA 2026 Scientific Sessions (reported by Medscape). STEP 1 reanalysis: lean share of weight loss from about 39% to about 30%; across GLP-1 trials, lean body mass about 25% to 39% of total weight loss.
  4. Haines MS, et al. ENDO 2025, San Francisco. Prospective study, n = 40, 3 months: 6.3% weight loss with semaglutide, 47.5% as lean mass; lower protein intake associated with greater lean mass decrease (R = 0.52).
  5. Neeland IJ, et al. Diabetes Obes Metab. 2024. Review of lean body mass changes with GLP-1 therapies, measurement limits, and mitigation strategies.
  6. Mechanick JI, et al. Obes Rev. 2025. Mitigation strategies: nutrition, exercise, supplementation, monitoring.
  7. Wilding JPH, et al. N Engl J Med. 2021;384:989-1002. STEP 1 parent trial: 1,961 adults, 68 weeks; weight reduction about 14.9% vs 2.4% placebo.