GLP-1 and Working Out: Strength Training, Muscle Loss and Protein

Roughly a quarter to two-fifths of the weight you lose on a GLP-1 comes off as lean tissue if you do nothing about it. That is the honest headline, and it comes from DXA scans inside the actual trials. The part almost nobody says out loud is that none of those trials ran a resistance-training programme. The participants were taking a drug and losing weight, and the body-composition numbers describe what happens when a large energy deficit meets no mechanical reason to keep muscle. Where people have trained and eaten enough protein, the number falls hard. In two documented cases it went negative, meaning lean soft tissue went up while body weight fell by 13.2% and 26.8%. This page is about the gap between those two outcomes and which levers sit inside it.

How much of GLP-1 weight loss is lean mass, in the trial data

Three numbers anchor this. In the STEP 1 DXA substudy, participants on semaglutide 2.4 mg over 68 weeks lost about 6.9 kg of lean soft tissue against 10.4 kg of fat, so roughly 40% of the total came from lean tissue. In the SURMOUNT-1 DXA substudy on tirzepatide, lean loss was about 5.6 kg against 15.9 kg of fat, or about 26%. A network meta-analysis across the class landed near 25%.

Those figures disagree by a factor of more than one and a half, which tells you how much the surrounding conditions matter. The stack-builder on this site uses 32% as its no-training central estimate, with a 25% to 40% band, because the trial values scatter across that range.

One more STEP 1 finding is worth holding onto, because it cuts the other way. Absolute lean mass fell, but lean mass as a proportion of total body mass rose by about 3 percentage points. Fat came off faster than lean tissue did. A person finishing that trial was lighter and, by body-composition percentage, leaner than when they started. Both statements are true at once and people tend to only hear one of them.

Why lean mass falls at all

There is no established direct catabolic action of these drugs on skeletal muscle. Semaglutide and tirzepatide are not eating your quads. What they do is suppress appetite, and everything downstream follows from that.

A large energy deficit lowers muscle protein synthesis because amino acids and energy are both less available. At the same time, protein intake usually falls further than total intake does, because protein-dense food is the most filling category on the plate and it tends to be the first thing people stop finishing. Steak becomes hard work. Yoghurt and toast do not. So the deficit arrives with a protein shortfall stapled to it.

The third input is loading. Muscle is metabolically expensive and the body keeps it in proportion to the mechanical demand placed on it. Remove the demand while the deficit and the protein shortfall are running, and lean tissue is the predictable casualty. Every one of those inputs is something you can act on, which is why the number moves so much between populations.

What resistance training does to the fraction

The randomised evidence here is S-LITE. Adding a supervised exercise programme to liraglutide after a low-calorie diet preserved fat-free mass and increased fat loss compared with the drug on its own, and the combination arm had the best body-composition result of the four groups tested. Exercise participants completed a median of about 2.65 sessions per week, roughly 108 minutes of moderate-to-vigorous activity. That volume is modest and it is the volume with actual randomised evidence behind it in a GLP-1 population.

At the other end sits a case series. Patients who resistance-trained 3 to 5 days a week and ate 1.6 to 2.3 g of protein per kg of fat-free mass saw lean soft tissue account for as little as 8.7% of weight lost. Two of them gained lean soft tissue, by 2.5% and 5.8%, while losing 26.8% and 13.2% of their body weight. That is a case series, not a trial. It has no control group and the people in it were doing everything right. Read it as proof that the ceiling is much higher than the trial averages suggest, not as an expected outcome.

Between those two poles the evidence thins out. The table below gives the numbers the stack-builder runs on, with the honesty about each one attached.

The protein arithmetic when appetite is suppressed

This is where the plan usually breaks. Protein research in energy deficit clusters at 1.2 to 1.6 g per kg of body weight per day, well above the 0.8 g/kg RDA, and the benefit for resistance training plateaus around 1.6 g/kg. No GLP-1-specific randomised trial establishes an optimal target, so those numbers are borrowed from the general energy-restriction and resistance-training literature.

Put real numbers on it. A person at 90 kg working from body weight needs 108 to 144 g of protein a day. Across three meals that is 36 to 48 g per sitting, roughly a chicken breast each time. Now suppress the appetite that would normally get that down. Many people on a GLP-1 report finishing half a normal portion.

The case-series figures use a different denominator, which matters. Those patients ate 1.6 to 2.3 g per kg of fat-free mass, not body weight. Somebody at 90 kg with 35% body fat carries 58.5 kg of fat-free mass, so the same target reads as 94 to 135 g a day. If a calculator quietly swaps the denominators it can hand you a number that is 50% too high. Check which one you are looking at.

Existing kidney disease changes the recommendation entirely, and that is a conversation with the clinician who knows your renal function, not a number to pick off a page.

• Body-weight target, 90 kg person: 108 to 144 g a day (1.2 to 1.6 g/kg)
• Fat-free-mass target, 90 kg person at 35% body fat: 94 to 135 g a day (1.6 to 2.3 g/kg FFM)
• Above roughly 1.6 g/kg body weight, the added benefit for resistance training flattens out
• The 0.8 g/kg RDA is a floor for sedentary people who are not in a deficit, and it is not the relevant number here
• Many clinicians suggest putting the protein first in each meal while appetite is suppressed, so it is not the thing you run out of room for. Ask yours what target fits your body size and kidney function

Strength versus size, and what happens to force output

Scan numbers and gym numbers can disagree, and the disagreement is informative. The SEMALEAN prospective study followed 106 completers with a mean BMI of 46.3. Lean mass fell by about 3 kg by month 7 and then stabilised. Handgrip strength rose by 4.5 kg at 12 months. Prevalence of sarcopenic obesity in the group dropped from 49% to 33%.

Losing lean tissue while getting functionally stronger sounds contradictory until you remember what got lost. Some was water and glycogen. Some was tissue supporting a much larger body.

The practical version: your working weights are a better early-warning system than the scan is. If load is progressing, or even holding, while body weight falls, the deficit is being pointed at fat. If working weights are dropping month over month at the same time as scale weight, that is the pattern worth raising with a clinician or a coach.

The SEMALEAN timing is also worth noting. Lean mass fell for about seven months and then stabilised, so the steepest part of the curve is early. That argues for getting training and protein in place at the start rather than waiting to see whether it becomes a problem, because by the time a scan shows it the bulk of the loss has already happened.

Older adults and sarcopenic obesity

Lean mass matters more the older you are, because you started with less of it and have less capacity to rebuild it. The SEMALEAN result is encouraging here, with sarcopenic obesity prevalence falling from 49% to 33%, though that population had a very high starting BMI and a great deal of fat to lose first.

The risk that concerns clinicians in older patients is functional rather than cosmetic. Difficulty rising from a chair or climbing stairs, and new unsteadiness, are the changes that predict falls. A fall threatens an older adult's independence far more than a few kilograms either way, so any of those appearing during weight loss belongs in front of a clinician promptly.

What to measure instead of scale weight

Scale weight cannot tell you what the weight was made of, which makes it the least useful number in this whole discussion despite being the one everybody watches.

A repeat body-composition scan every 3 to 6 months gives you the split directly. Grip strength is a cheap functional proxy. Your logged working weights are free and you already have them. Rate of loss is worth tracking too, since faster loss and larger deficits are associated with a larger lean fraction, which is one reason slower titration comes up in clinical conversations. Whether to slow anything is your prescriber's decision, not yours and not this page's.

• DXA or a comparable body-composition scan every 3 to 6 months
• Grip strength with a dynamometer, tracked against your own first reading
• Logged working weights, watched for direction over months rather than sessions
• Rate of loss as a percentage of body weight per week, not kilograms
• Protein actually eaten, weighed honestly for a week, rather than protein intended

Programming around dose days and titration

No trial has tested when in the injection cycle to train. What is documented is that GI adverse events concentrate in the escalation phase and after each dose increase, with nausea reported by about 44% of semaglutide 2.4 mg participants in STEP 1 and 31 to 43% of tirzepatide participants in SURMOUNT-1. Delayed gastric emptying is also largest after the first dose and after each escalation.

That gives you a shape rather than a rule. Many people find the 24 to 48 hours after an injection are the worst window during titration and settle into scheduling the demanding sessions elsewhere in the week. Whether that fits your dosing schedule is a question for your prescriber.

The window narrows as doses stabilise, since the gastric-emptying effect is largest after the first dose and each escalation and diminishes over time. A pattern that made training miserable in week three of a step-up is often unremarkable two months later at the same dose.

Recovery deserves a mention. Training hard is a stress paid for with food and sleep, and one of those is now harder to come by. The S-LITE volume, two to three sessions a week, is not a compromise. It is the dose with the evidence.

A first-person account, for context

Everything above is drawn from trial data and prescribing information. What that evidence cannot give you is what a GLP-1 block actually feels like week to week in someone training seriously.

Thomas Prommer, a competitive endurance athlete, has written up holding lean mass through a GLP-1 block while training for endurance racing. Read it as one person's experience rather than as evidence. It is n=1, it is uncontrolled, and it describes an athlete population the trials on this page did not study. It is useful for exactly that reason and for no other.

The evidence, one row per claim

Questions people ask

Sources

1. Preservation of lean soft tissue during weight loss induced by GLP-1 and GLP-1/GIP receptor agonists (PMC)
2. Effect of GLP-1 receptor agonists and co-agonists on body composition: systematic review and network meta-analysis
3. Muscle Mass and GLP-1 Receptor Agonists: Adaptive or Maladaptive Response to Weight Loss? (Circulation)
4. SEMALEAN study, Diabetes Obesity and Metabolism
5. Morton et al., systematic review and meta-regression of protein supplementation and resistance training (PMC)

Keep reading

• GLP-1 and Protein: How Much, and Why It Is Hard to EatThe protein evidence predates GLP-1s and is solid. The new problem is not the target, it is reaching one when
• GLP-1 and Running: Endurance Training, Pace and FuellingBody mass falls, which helps. Glycogen availability falls and resting heart rate rises, which hurts. Aerobic f
• GLP-1 and Doping: WADA Status of Semaglutide and TirzepatideThe question has two answers and most coverage gives only the first. An approved GLP-1 is not banned. An unapp
• GLP-1 and Energy: Fatigue, Gym Performance and What Causes ItFatigue on a GLP-1 usually has a findable cause. Separate the fixable from the expected instead of accepting i
• GLP-1 and Creatine: Does It Help Preserve Muscle?No GLP-1-specific research exists. The honest page covers creatine's own evidence, the water-weight artefact a
• GLP-1 and Muscle CrampsNot a labelled adverse reaction. The plausible drivers are all downstream: fluid and electrolyte losses, reduc