Scientific deep-dive
Muscle Volume Versus Muscle Function on a GLP-1
Psoas muscle volume fell 9.3% over 24 weeks while chair-rise time and walking speed both edged better. Mass and function moved in opposite directions in the same people, and only one of them is what anybody wants.
The standard worry about GLP-1 drugs is that they take muscle along with fat, and it is a real effect that shows up in every body-composition study. A 2025 analysis measured both halves of the question in the same people and found something the alarming version leaves out: psoas muscle volume fell 9.3% over 24 weeks, while chair-rise time and walking speed both moved slightly better.[1]
Mass and function are different measurements
This is the distinction the whole debate turns on, and it is usually collapsed. Muscle mass is what a scan measures. Muscle function is whether you can stand up from a chair and walk across a room. They are correlated but not the same, and only one of them is what anybody wants.
In this study, participants received semaglutide for 24 weeks, titrated to 1 mg a week. Psoas muscle volume was measured on MRI and fell by 9.3%, a highly significant change. Physical function was assessed by a ten-time chair rise and a four-meter gait speed test. Chair rise improved by 1.27 seconds and gait speed by 0.05 meters per second. Neither change was statistically significant, but both pointed the opposite way from the scan.[1]
| Measure | Change | Significant? |
|---|---|---|
| Psoas muscle volume | −9.3% (95% CI −13.4 to −5.2) | Yes, P<0.001 |
| Psoas muscle fat fraction | −0.42% (95% CI −1.00 to 0.17) | No, P=0.16 |
| Ten-time chair rise | 1.27 seconds faster | No |
| Four-meter gait speed | 0.05 m/s faster | No |
Why the muscle loss is expected, not alarming on its own
Losing weight of any kind costs lean tissue. It happens with dieting, with bariatric surgery and with illness, and it happens because a smaller body needs less muscle to move itself. Since lean mass always falls, the question that matters is whether the person ends up weaker.
On that question, this study is reassuring and thin. Our article on what the trials measured covers the body-composition side in the larger programs, and the practical response, resistance training and adequate protein, has not changed and is not controversial.
The comparison that answers the question best
A larger MRI study went further, and it did the one thing that turns muscle numbers into an answer: it compared them against what would be expected anyway. SURPASS-3’s imaging substudy scanned 246 people with type 2 diabetes at baseline and week 52, randomized to tirzepatide at three doses or to titrated insulin, and then set the results against longitudinal MRI data from UK Biobank participants.[2]
| Measure | Change | 95% CI |
|---|---|---|
| Muscle fat infiltration | −0.36 percentage points | −0.48 to −0.25 |
| Muscle volume | −0.64 L | −0.74 to −0.54 |
| Muscle volume Z score | −0.22 | −0.29 to −0.15 |
Read alone, that is the alarming version: muscle volume down, and down relative to what body size would predict. All three changes reached p<0.0001 and all three differed significantly from the insulin arm, where body weight and muscle volume both rose modestly.
So the muscle that went was about what you would expect to go. The muscle that stayed was less infiltrated with fat than expected. Fat infiltration is a measure of muscle quality, and it moved in the favorable direction beyond what the weight loss alone would account for.
The insulin arm makes the same point from the other side. Those participants gained weight and gained muscle volume, and their muscle fat infiltration did not significantly improve. They ended up with more muscle of no better quality, which is why volume on its own is not the thing to want.
Two limits apply. This was a post-hoc exploratory analysis in people with type 2 diabetes and fatty liver, not a general population. And a UK Biobank comparison is a modeled expectation rather than a matched control group. It is a good model, but not the same as randomizing people to lose weight another way.
How much weight this can carry
For the semaglutide study, not much. It was a single-arm study with no control group. Everyone received the drug, so a functional improvement cannot be attributed to it. People who enroll in a study, lose weight and get tested twice may walk faster for reasons that have nothing to do with semaglutide.
- 51 participants, with muscle measures in 46. Small enough that the non-significant functional results may be underpowered rather than flat.
- 24 weeks, and titrated only to 1 mg a week, below the 2.4 mg used for weight management. Both the duration and the dose are modest.
- A specific population: people with HIV and metabolic dysfunction-associated steatotic liver disease, with a mean age of 50 and a mean BMI of 35.5.
- A secondary analysis. The parent study, SLIM LIVER, was designed to ask about liver fat. Muscle was not its primary question.
A single-arm study cannot tell you what a drug did. It can tell you what happened, which is sometimes enough to correct a bad assumption.
That is its value. It does not prove GLP-1 drugs preserve function. It does show that a substantial fall in measured muscle volume can happen alongside stable or slightly improved physical performance. So a scan showing lost lean mass is not, by itself, evidence that someone has been harmed. The ten papers we rely on for the muscle question are in our muscle reading list.
Frequently Asked Questions
References
- 1.Ditzenberger GL, Lake JE, Kitch DW, et al. Effects of Semaglutide on Muscle Structure and Function in the SLIM LIVER Study Clinical Infectious Diseases. 2025. PMID: 39046173.
- 2.Sattar N, Neeland IJ, Dahlqvist Leinhard O, et al. Tirzepatide and muscle composition changes in people with type 2 diabetes (SURPASS-3 MRI): a post-hoc analysis of a randomised, open-label, parallel-group, phase 3 trial The Lancet Diabetes & Endocrinology. 2025. PMID: 40318682.
Compare the sellers
Related research
GLP-1 Drugs After Spinal Cord Injury: Three People Treated, Two Not
In a controlled case series of five people with chronic spinal cord injury, three on weekly semaglutide lost weight, fat mass and visceral fat over 26 weeks while the two untreated participants gained on every measure.
5 min read
GLP-1 Drugs and HIV-Associated Lipohypertrophy: The Randomized Evidence
A phase 2b randomized trial found once-weekly semaglutide cut abdominal visceral fat by about 30.6% over 32 weeks in people with HIV-associated lipohypertrophy, alongside a 9.3% fall in psoas muscle volume that did not cost measurable function.
6 min read
Protein Targets and the Lab Panel
At least 1.2 g of protein per kg daily, 0.3–0.4 g/kg per meal, and a short lab panel: vitamin D, B12, iron studies, folate and zinc, plus thiamin in high-risk patients. The upper protein target excludes chronic kidney disease.
5 min read
Exercise Plus a GLP-1: What Each One Actually Changes
In a four-arm randomized trial, the drug alone produced no improvement in physical fitness. Exercise did, with or without the drug, and the strength result needs reading carefully.
6 min read
Bone Density on a GLP-1: The Trial That Compared It Against Exercise
A four-arm randomized trial found liraglutide alone lost hip and spine bone density relative to exercise, while the combined arm had the greatest weight loss and kept bone density level with placebo.
7 min read
MK-677: A Real Trial in Eight People, and a Liver Report
A 1998 randomized crossover trial found MK-677 reversed diet-induced protein catabolism, moving nitrogen balance from -2.67 to +0.31 g/day. It enrolled eight people and measured a proxy. A 2025 case report describes liver injury after two months of use.
5 min read
Where to get semaglutide online, safely: sellers our editors have checked
These are telehealth sellers our editors have checked. For each one we hold a price, the form the drug comes in, and the states it reaches.
No insurance needed · vetted by our editors
Some of the links on this page earn us money. If you sign up with a provider after following one, that provider may pay GLP Watchdog a commission. Learn more
Collective
From $59/mo + membership
Get started →YourEra
From $99/mo
Get started →Wellorithm
From $247/mo
Get started →