GLP-1s and Muscle Loss: Protecting Lean Mass While Losing Weight

You have probably seen the headline: a third or more of the weight you lose on a GLP-1 is muscle. The number is real and it comes from the trials. It is also one of the most misread statistics in obesity medicine. What the scans measured was lean mass, which is not the same thing as muscle — and when researchers compared people on the drug against people losing weight without it, the ratio came out the same.

That does not make the concern imaginary. It relocates it. The question is not whether GLP-1s strip muscle from everyone. It is who is genuinely at risk, and what the evidence shows you can do about it.

How much of GLP-1 weight loss is lean mass?

Two large trials measured body composition directly with scans rather than estimating it. In the STEP 1 DXA substudy, participants taking semaglutide lost 13.6 kg on average — 8.36 kg of fat and 5.26 kg of lean body mass, so lean tissue made up about 39% of the total. In SURMOUNT-1, tirzepatide produced a 21.3% fall in body weight, a 33.9% fall in fat mass and a 10.9% fall in lean mass over 72 weeks, which works out to roughly 75% fat to 25% lean.

Lean mass is not the same as muscle

This is where most coverage goes wrong. On a DXA scan, lean body mass means everything that is not fat and not bone mineral: skeletal muscle, but also water, stored glycogen, connective tissue and the internal organs.

Each gram of glycogen is stored with roughly three grams of water alongside it. Someone who sharply reduces how much they eat sheds several kilograms of glycogen and the water bound to it within the first weeks — and every bit of that registers as lean mass lost on the scan. The skeletal-muscle share of that number is meaningfully smaller than the headline suggests.

The placebo group is the number that matters

The SURMOUNT-1 body-composition analysis found that about 75% of the weight lost was fat and 25% was lean mass in both the tirzepatide group and the placebo group. People losing weight through lifestyle change alone gave up lean tissue in the same proportion as people on the drug.

Tirzepatide was not associated with disproportionate loss of lean tissue. It was associated with far more total weight loss, and the composition of that loss followed the ordinary pattern. This is what physiology would predict: losing weight in an energy deficit has always cost some lean tissue, whether the deficit comes from a diet, from bariatric surgery, or from a medicine that blunts appetite.

Muscle quality improved, not just quantity

The most informative evidence here is SURPASS-3 MRI, a post-hoc analysis published in The Lancet Diabetes & Endocrinology in April 2025. It used MRI rather than DXA, which meant it could measure actual thigh muscle volume and, separately, the fat infiltrating the muscle itself.

Across 52 weeks, thigh muscle volume did fall — but it fell in line with what population data predicts for that amount of weight loss. Muscle fat infiltration, known as myosteatosis, dropped by considerably more than expected. Fat stored inside muscle is associated with weakness and insulin resistance, so less of it points toward better-functioning muscle rather than worse.

The amount of muscle volume changes with tirzepatide appears to be in line with muscle volume changes seen in the population for similar differences in weight.

That is Professor Naveed Sattar of the University of Glasgow, who led the analysis.

Did body composition get better or worse?

Better, on the measure that counts. In STEP 1, lean mass fell 9.7% in absolute terms — but lean mass as a share of total body mass rose by 3.0 percentage points. Participants finished with bodies that were proportionally more lean tissue than when they started. That is the direction you want body composition to move.

So who should actually worry?

Not the average trial participant. The legitimate concern is the person who starts out with too little muscle in the first place.

• Older adults. Muscle mass declines with age regardless of treatment, and a substantial energy deficit layered on top of that compounds the loss. This is the group clinicians raise most often.
• People with sarcopenic obesity — high body fat combined with low muscle mass and reduced muscle function. It affects an estimated 10–20% of older adults and is often missed, because a high body weight hides it.
• Anyone cycling on and off treatment. Weight regained after stopping tends to return disproportionately as fat, while lean tissue does not automatically come back. Repeated cycles can leave someone fatter and less muscular than before they started.
• People who are frail, undernourished, or eating very little because of nausea or heavy appetite suppression at higher doses.

For these groups the risk-benefit calculation genuinely is different, which is why the clinical press has begun commissioning coverage on prescribing GLP-1s to patients who already have sarcopenia.

What actually protects muscle

The evidence converges on two levers, and together they work considerably better than either does alone.

1. 1.Protein. The target cited across this literature is 1.2 to 1.6 g per kilogram of body weight per day, spread evenly across meals rather than concentrated in one. Even distribution matters because muscle protein synthesis responds to a threshold dose per meal. Anyone with kidney disease needs this figure set by their clinician.
2. 2.Resistance training. Loading the muscle is the signal that tells your body to keep the tissue. Programmes in this research typically use two to four sessions a week covering the major muscle groups. Walking and other cardio are worth doing but do not substitute for it.
3. 3.A deficit you can sustain. The sharper the energy deficit, the more lean tissue leaves alongside the fat. Titrating at a measured pace, and eating enough at each dose step, costs less muscle than racing to the top dose.

The effect size is not trivial. In a six-month prospective study of 200 adults starting semaglutide or tirzepatide who were coached on resistance training and protein intake, participants lost about 13% of body weight while losing roughly 3% of muscle mass — a far better ratio than untrained trial averages.

Drugs built to keep the muscle

This has moved from theory into active development. The BELIEVE trial, a phase 2 study of 507 adults published in Nature Medicine, paired semaglutide with bimagrumab, an antibody that blocks a signalling pathway which normally limits muscle growth.

At 48 weeks, lean mass changed by −5.3% to −7.9% with semaglutide alone, by +2.3% to +2.7% with bimagrumab alone, and by only −1.1% to −2.6% with the combinations. The highest-dose combination produced 17.8 kg of weight loss against 14.2 kg for semaglutide alone — more weight gone, with most of the lean tissue kept.

Phase 2 is not approval, and bimagrumab is not available for this purpose. What the trial establishes is that the lean-mass share of weight loss is modifiable rather than fixed.

How to tell if you are losing too much

The scale will not tell you and neither will the mirror. Track function instead.

• Strength falling faster than your body weight is. Some decline is expected while you are eating less; a steep drop is a signal worth raising.
• Everyday tasks getting harder — stairs, carrying shopping, standing up from a low chair without pushing off with your hands.
• Grip strength, the standard clinical proxy for whole-body muscle function, and cheap to measure.
• Fatigue that does not track with your sleep.

If you want a measurement rather than an impression, ask about a DXA scan at baseline and again after six months. Bioelectrical impedance scales are widely sold but are sensitive to how hydrated you are, which makes them unreliable for precisely the thing you are trying to track.