- Aug 15, 2018
- 602
- 346
- 63
We've had plenty of discussions here about cutting, maintaining strength, and what a bathroom scale fails to tell us. So let's tackle a topic that's generated some wild headlines lately: **does weight loss in GLP-1 trials mean people are losing muscle?**
The honest research answer is more complicated than either “all the weight lost is muscle” or “there's nothing to discuss.” Studies have measured declines in both fat mass and lean mass during substantial weight loss. But **lean mass is not the same thing as skeletal muscle**, and a scan showing less lean tissue doesn't automatically show reduced strength or physical function.
I want to keep this thread on published evidence and how we interpret it, not personal treatment experiences or anyone's instructions for using these medications.
**First, what does a body-composition scan measure?**
Many trial substudies use dual-energy X-ray absorptiometry, or DXA. It estimates fat mass, bone mineral content, and lean soft tissue. That last category includes skeletal muscle but also water, organs, blood and other non-fat tissue. A change in lean mass therefore cannot be reported accurately as if every gram came from contractile muscle fibres.
There's another distinction: the **absolute amount** of lean tissue can fall even while its **proportion of total body weight** rises. Imagine someone loses substantially more fat than lean mass. The final body can contain fewer kilograms of each, yet a greater percentage of its total weight will be lean. Both statements can be true simultaneously.
That doesn't answer whether strength changed. DXA isn't a direct measure of a squat, grip-strength test, or muscle performance. Those are different outcomes that need their own measurements.
**What STEP actually reported**
An exploratory DXA analysis of **STEP 1**, published in the *Journal of the Endocrine Society* in **2021**, evaluated body composition in adults with overweight or obesity during a semaglutide trial. It reported reductions in total body fat and in absolute lean mass, while the proportion of lean mass relative to total body mass increased in the studied group.
That's a useful example of why a headline saying “lean mass declined” can be accurate but incomplete. The composition of the weight lost matters, as do the method and study population. An exploratory analysis is also different from a trial built primarily to compare changes in muscle function.
More importantly for this forum, the STEP finding cannot tell us how a resistance-trained lifter with a particular training history would perform. The participants in a weight-management study aren't automatically representative of competitive strength athletes. That's not a flaw in the study; it's a limit on how far we can generalize.
**What the SURMOUNT-1 DXA substudy found**
A **2025** peer-reviewed analysis of **SURMOUNT-1** examined **160 participants** from the larger trial using DXA at baseline and week 72. In the tirzepatide group, average body weight declined **21.3%**, fat mass declined **33.9%**, and lean mass declined **10.9%** from their respective baselines. The corresponding changes in the placebo group were **5.3%**, **8.2%**, and **2.6%** decreases.
Notice the denominators: those percentages describe change in each separate compartment relative to its own starting value. They are **not** percentages of total weight loss that can simply be added together. Confusing those two types of percentages is one reason online explanations become misleading so quickly.
The substudy shows that the measured lean compartment declined alongside a larger proportional reduction in fat. It does not demonstrate that all lean loss was muscle tissue, that everyone had the same outcome, or that the drug independently caused muscle loss beyond changes associated with weight reduction. Those are separate research questions.
**Where resistance-training evidence fits—and where it doesn't**
A systematic review and meta-analysis published in *BMJ Open Sport & Exercise Medicine* in **2025** examined resistance exercise added to dietary weight-loss interventions among adults with overweight or obesity. The authors found better preservation of fat-free mass in the resistance-exercise groups, with differences in fat-mass and strength outcomes also assessed.
That's relevant background, but these were **diet-induced weight-loss comparisons**. They shouldn't be presented as if they were randomized trials specifically testing training against no training inside STEP or SURMOUNT. Evidence can be biologically relevant without answering exactly the same experimental question.
Protein research has a similar limitation. Morton and colleagues' **2018** *British Journal of Sports Medicine* meta-analysis pooled **49 resistance-training studies with 1,863 participants** and found modest additional average gains in fat-free mass and strength with protein supplementation. That supports investigating protein availability in training adaptation, but it does not provide a directly tested formula for preserving lean mass in every GLP-1 trial participant.
**Why the word “muscle” deserves caution**
There are at least three related but different questions: how much lean mass changes; how much actual skeletal muscle tissue changes; and whether strength or physical function changes. A single DXA estimate can't settle all three. Water shifts and body size changes can also complicate interpretation of lean-tissue measurements.
Some trials report overall body-composition averages, but subgroup conclusions can be much less certain. Age, initial body composition, weight-loss magnitude, activity patterns, and other factors may differ across participants. If a paper didn't measure training history or muscle performance well, we shouldn't fill that gap with assumptions.
My reading is that the published evidence supports concern for **measurement and research quality**, not a sensational claim that all weight loss is muscle wasting. The fat-versus-lean distinction matters. So do the difference between absolute and relative changes, the timing of measurements, and the absence of direct function testing in some substudies.
For anyone who wants to read the originals, look up the **STEP 1 DXA exploratory analysis (2021)**, **SURMOUNT-1 body-composition substudy (2025)**, the **2025 resistance-exercise weight-loss meta-analysis**, and **Morton et al. (2018)**. Each contributes to the picture, but none substitutes for the others.
I think this is a good subject for a lifting forum because we can be precise about what a body-composition number means without jumping into individual protocols. Let's keep the discussion centred on research quality and the claims the data can genuinely support.
1. When you see “lean mass loss” in a study headline, do you first look for DXA methods, absolute tissue changes, or actual strength and function tests?
2. What kind of randomized study would best separate the effect of substantial weight loss itself from any additional effect of a GLP-1-based intervention on muscle outcomes?
3. Do you think the public discussion would improve if researchers and journalists consistently separated lean mass, skeletal muscle mass, and measured performance?
— OLYMPIC
The honest research answer is more complicated than either “all the weight lost is muscle” or “there's nothing to discuss.” Studies have measured declines in both fat mass and lean mass during substantial weight loss. But **lean mass is not the same thing as skeletal muscle**, and a scan showing less lean tissue doesn't automatically show reduced strength or physical function.
I want to keep this thread on published evidence and how we interpret it, not personal treatment experiences or anyone's instructions for using these medications.
**First, what does a body-composition scan measure?**
Many trial substudies use dual-energy X-ray absorptiometry, or DXA. It estimates fat mass, bone mineral content, and lean soft tissue. That last category includes skeletal muscle but also water, organs, blood and other non-fat tissue. A change in lean mass therefore cannot be reported accurately as if every gram came from contractile muscle fibres.
There's another distinction: the **absolute amount** of lean tissue can fall even while its **proportion of total body weight** rises. Imagine someone loses substantially more fat than lean mass. The final body can contain fewer kilograms of each, yet a greater percentage of its total weight will be lean. Both statements can be true simultaneously.
That doesn't answer whether strength changed. DXA isn't a direct measure of a squat, grip-strength test, or muscle performance. Those are different outcomes that need their own measurements.
**What STEP actually reported**
An exploratory DXA analysis of **STEP 1**, published in the *Journal of the Endocrine Society* in **2021**, evaluated body composition in adults with overweight or obesity during a semaglutide trial. It reported reductions in total body fat and in absolute lean mass, while the proportion of lean mass relative to total body mass increased in the studied group.
That's a useful example of why a headline saying “lean mass declined” can be accurate but incomplete. The composition of the weight lost matters, as do the method and study population. An exploratory analysis is also different from a trial built primarily to compare changes in muscle function.
More importantly for this forum, the STEP finding cannot tell us how a resistance-trained lifter with a particular training history would perform. The participants in a weight-management study aren't automatically representative of competitive strength athletes. That's not a flaw in the study; it's a limit on how far we can generalize.
**What the SURMOUNT-1 DXA substudy found**
A **2025** peer-reviewed analysis of **SURMOUNT-1** examined **160 participants** from the larger trial using DXA at baseline and week 72. In the tirzepatide group, average body weight declined **21.3%**, fat mass declined **33.9%**, and lean mass declined **10.9%** from their respective baselines. The corresponding changes in the placebo group were **5.3%**, **8.2%**, and **2.6%** decreases.
Notice the denominators: those percentages describe change in each separate compartment relative to its own starting value. They are **not** percentages of total weight loss that can simply be added together. Confusing those two types of percentages is one reason online explanations become misleading so quickly.
The substudy shows that the measured lean compartment declined alongside a larger proportional reduction in fat. It does not demonstrate that all lean loss was muscle tissue, that everyone had the same outcome, or that the drug independently caused muscle loss beyond changes associated with weight reduction. Those are separate research questions.
**Where resistance-training evidence fits—and where it doesn't**
A systematic review and meta-analysis published in *BMJ Open Sport & Exercise Medicine* in **2025** examined resistance exercise added to dietary weight-loss interventions among adults with overweight or obesity. The authors found better preservation of fat-free mass in the resistance-exercise groups, with differences in fat-mass and strength outcomes also assessed.
That's relevant background, but these were **diet-induced weight-loss comparisons**. They shouldn't be presented as if they were randomized trials specifically testing training against no training inside STEP or SURMOUNT. Evidence can be biologically relevant without answering exactly the same experimental question.
Protein research has a similar limitation. Morton and colleagues' **2018** *British Journal of Sports Medicine* meta-analysis pooled **49 resistance-training studies with 1,863 participants** and found modest additional average gains in fat-free mass and strength with protein supplementation. That supports investigating protein availability in training adaptation, but it does not provide a directly tested formula for preserving lean mass in every GLP-1 trial participant.
**Why the word “muscle” deserves caution**
There are at least three related but different questions: how much lean mass changes; how much actual skeletal muscle tissue changes; and whether strength or physical function changes. A single DXA estimate can't settle all three. Water shifts and body size changes can also complicate interpretation of lean-tissue measurements.
Some trials report overall body-composition averages, but subgroup conclusions can be much less certain. Age, initial body composition, weight-loss magnitude, activity patterns, and other factors may differ across participants. If a paper didn't measure training history or muscle performance well, we shouldn't fill that gap with assumptions.
My reading is that the published evidence supports concern for **measurement and research quality**, not a sensational claim that all weight loss is muscle wasting. The fat-versus-lean distinction matters. So do the difference between absolute and relative changes, the timing of measurements, and the absence of direct function testing in some substudies.
For anyone who wants to read the originals, look up the **STEP 1 DXA exploratory analysis (2021)**, **SURMOUNT-1 body-composition substudy (2025)**, the **2025 resistance-exercise weight-loss meta-analysis**, and **Morton et al. (2018)**. Each contributes to the picture, but none substitutes for the others.
I think this is a good subject for a lifting forum because we can be precise about what a body-composition number means without jumping into individual protocols. Let's keep the discussion centred on research quality and the claims the data can genuinely support.
1. When you see “lean mass loss” in a study headline, do you first look for DXA methods, absolute tissue changes, or actual strength and function tests?
2. What kind of randomized study would best separate the effect of substantial weight loss itself from any additional effect of a GLP-1-based intervention on muscle outcomes?
3. Do you think the public discussion would improve if researchers and journalists consistently separated lean mass, skeletal muscle mass, and measured performance?
— OLYMPIC