Semaglutide and Muscle Mass: What the STEP Trial Sub-Analyses Actually Show

Every weight-loss intervention — surgical, pharmaceutical, or dietary — takes some of the weight from lean tissue rather than fat. The interesting question is how much. For semaglutide, the answer matters because the drug is increasingly prescribed to populations where preserving muscle is medically important: older adults, post-menopausal women, and people with type 2 diabetes who already have lower baseline lean mass.

The headline number that has been circulating in popular coverage — that 30–40% of weight lost on semaglutide is lean body mass — is essentially correct, but it deserves unpacking. It comes from a DEXA sub-analysis of the STEP 1 trial, and the picture it paints is more nuanced than either the “semaglutide melts your muscles” takes or the defensive industry response that “all weight loss looks like this” would suggest. Both can be true at the same time.^[1]

What STEP 1 actually measured

STEP 1 was the pivotal phase 3 trial that put semaglutide on the weight-loss map. 1,961 adults with overweight or obesity (and without type 2 diabetes) were randomized 2:1 to weekly subcutaneous semaglutide 2.4 mg or placebo, both layered on top of lifestyle intervention. After 68 weeks, the semaglutide arm had lost a mean of 14.9% of body weight, versus 2.4% in the placebo arm.^[2]

Buried inside that trial was a body composition sub-study: 140 participants underwent DEXA (dual-energy X-ray absorptiometry) scans at baseline and again at week 68. DEXA is the gold standard non-invasive method for separating fat mass from lean mass. The sub-study results were published as a separate analysis a year after the main trial, and they are the source of every credible statistic on this topic.^[3]

Among DEXA participants in the semaglutide arm:

• Total body weight loss averaged 15.8 kg (about 35 lb).
• Of that loss, roughly 9.7 kg was fat mass (61%).
• And roughly 6.1 kg was lean body mass (39%).
• Visceral adipose tissue dropped sharply — the metabolically dangerous fat fell faster than subcutaneous fat, which is the finding that matters most for cardiovascular and diabetes outcomes.

The 39% lean-mass figure is what gets quoted. But the qualifier the original investigators added is important: the proportion of lost weight that comes from lean tissue during semaglutide therapy is comparable to what you see in any unstructured caloric restriction of similar magnitude. In other words, the issue isn't that GLP-1s are uniquely catabolic — it's that fast weight loss without resistance training and adequate protein causes lean-mass loss across the board, and GLP-1s simply make fast weight loss possible at scale for the first time.^[4]

How that compares to other weight-loss methods

For context, here's what the literature shows for the share of lost weight that comes from lean mass across several well-studied interventions:

• Bariatric surgery (Roux-en-Y). Roughly 25–30% of total weight lost is lean mass, with significant variation by patient age and post-op activity. The percentage tends to be lower in younger, more active patients.^[5]
• Very low calorie diets (VLCD, <800 kcal/day). 25–45% lean mass loss, depending on protein intake. High-protein VLCDs sit at the low end of that range.^[6]
• Intermittent fasting. Generally 20–40% lean-mass share, with the same protein-and-resistance-training modifiers.^[7]
• Semaglutide (STEP 1 sub-study). ~39% lean-mass share.
• Tirzepatide (SURMOUNT 1 sub-study). Similar range, with newer evidence suggesting comparable proportional composition despite the larger absolute weight loss.^[8]

The takeaway: semaglutide isn't catastrophically worse than other major interventions on lean-mass preservation, but it also isn't magically better. The percentages are similar; the difference is that semaglutide produces dramatically more total weight loss in routine clinical use, which means the absolute number of pounds of lean mass lost is larger even when the percentage is normal.

Why this matters more for some patients than others

Not every kilogram of lean mass loss is equal. Lean body mass is an aggregate that includes skeletal muscle, organ tissue, bone mineral, and intracellular water. When weight drops fast, intracellular water shifts can account for a meaningful chunk of early “lean” loss without representing actual muscle atrophy. That's the strongest argument for not panicking about the headline number.^[9]

But the patients where the concern is real and not theoretical are:

• Adults over 65. Sarcopenia — age-related muscle loss — starts in the fourth decade and accelerates past 60. An older patient losing 6 kg of lean mass on top of natural sarcopenia can plausibly cross the threshold into functional impairment (slower gait speed, reduced grip strength, higher fall risk).^[10]
• Post-menopausal women. The estrogen drop at menopause already accelerates muscle and bone loss. Layered on top of GLP-1 weight loss without a resistance program, this population shows the largest decrements in DEXA lean mass.
• Patients with sarcopenic obesity at baseline. A subset of obese patients already have low muscle mass relative to their body size before starting any weight-loss program. They're the highest-risk group, and ironically the group most likely to benefit from the metabolic effects of weight loss. Resistance training matters most for them.^[11]
• People with type 2 diabetes. Diabetes itself accelerates muscle loss through insulin resistance and chronic inflammation. The combined effect of diabetes plus rapid weight loss can be larger than either alone.

What the trials show actually preserves lean mass

The good news is that resistance training and adequate protein intake during semaglutide therapy have been studied directly, and the results are encouraging. A 2024 randomized trial paired semaglutide with a structured resistance-training program in healthy adults with overweight; the training arm lost the same amount of total weight as the semaglutide-only arm but preserved roughly twice as much lean body mass at 32 weeks.^[12]

The interventions that have actually been tested in trials and show meaningful effect:

1. Resistance training, 2–3 sessions per week, progressive overload. Doesn't need to be a gym program — bodyweight protocols and resistance bands have been shown to work in older adults. The dose-response relationship favors heavier loads when tolerable. Volume matters more than session length.
2. Protein intake of 1.2–1.6 g/kg body weight per day. Higher than the standard 0.8 g/kg RDA. For a 75 kg person, that's roughly 90–120 g per day. Spread across meals because GLP-1s slow gastric emptying and you can run into satiety problems trying to hit the target in one sitting.^[13]
3. Maintaining intake even during early appetite suppression. The first 8–12 weeks on semaglutide are when appetite drops fastest. Patients who passively let calorie intake fall to 800–1,000 kcal/day during this window are the ones who show the worst lean-mass outcomes. Forcing protein intake even when not hungry is the single most actionable lever.
4. Vitamin D and creatine. Smaller effects, both well-tolerated, both supported by reasonable trial evidence in older adults losing weight.^[14]

Notice what isn't on the list: branched-chain amino acid supplements, “anabolic” nutraceuticals, leucine powders, or any of the marketing categories sold around GLP-1s. None of these have trial-grade evidence for preserving lean mass during pharmaceutical weight loss.

The clinical bottom line

The fair summary of the evidence: semaglutide produces lean-mass loss in roughly the same proportion as other rapid weight-loss methods, but the absolute amount is larger because the absolute weight loss is larger. For most patients under 60 with normal baseline muscle mass and any kind of physical activity, the clinical significance of the lean-mass loss is small relative to the cardiovascular, metabolic, and quality-of-life benefits of the weight loss itself.

For older patients, post-menopausal women, and anyone with sarcopenic obesity, the picture changes — and the evidence-based response is not to avoid GLP-1s but to layer resistance training and protein intake on top of them. That combination has been studied, and it works.

The thing nobody quotes: the same STEP 1 DEXA sub-study found that visceral adipose tissue (the metabolically dangerous fat around the organs) fell roughly twice as fast as subcutaneous fat on semaglutide. That's the finding that drives the cardiovascular outcome benefits seen in SELECT and other post-marketing trials. From a longevity and metabolic-health standpoint, the trade is heavily favorable for almost every patient who would clinically qualify for the drug.

If you're weighing semaglutide and the muscle-loss concern is what's holding you back, the right question isn't “is this drug catabolic?” It's “am I willing to lift weights twice a week and eat enough protein while I'm on it?” If the answer is yes, the data suggests you'll come out the other side leaner, healthier, and not meaningfully weaker. If the answer is no, the trade is less obviously favorable — and that's a conversation worth having with the prescribing clinician before starting.