
Smart scale muscle mass on a GLP-1: how to read lean mass, body fat and DXA numbers
A smart scale weighs you accurately, but its muscle and body fat numbers are estimates that can be off by several kilograms. On a GLP-1, the "muscle" or "lean" number can also drop fast early on because it counts water, not only muscle. Weigh in the same way each time, read body composition as a 4-week trend, and let your strength log tell you whether you're keeping muscle.
Lean mass is not the same as muscle
Most scales and scan reports split your body into fat and everything else. "Everything else" is lean mass, also called fat-free mass: muscle, water, organs, skin, connective tissue and, depending on the report, bone. Some scales label part of it "muscle mass," but they can't see muscle directly. They estimate it.
Two facts explain most of the confusion:
- Lean mass is mostly water. In adults, about 73% of fat-free mass is water[1]. When your body's water shifts, your lean number shifts with it.
- Stored carbs carry water. Your body stores carbohydrate as glycogen in your liver and muscles, packed with about 3 to 4 parts water[2]. Eating less, which is what a GLP-1 helps you do, uses some of it up. That water leaves too, and most body composition methods read it as lost lean mass.
What GLP-1 studies actually measured
Lean mass really does fall during GLP-1 treatment. The question is how much of that is muscle. Here is what the main body composition studies found:
| Study | How they measured | What happened |
|---|---|---|
| STEP 1 substudy, semaglutide, 68 weeks[3] | DXA scan, 140 adults | Weight fell 15.0%, fat mass 19.3% and lean mass 9.7%. Because fat fell faster, lean mass made up a larger share of body weight at the end (up 3.0 percentage points) |
| SURMOUNT-1 substudy, tirzepatide, 72 weeks[4] | DXA scan, 160 adults | About 75% of the weight lost was fat and 25% was lean mass, the same split seen in the placebo group, who lost less weight with lifestyle changes |
| Review of GLP-1 body composition trials[5] | Mostly DXA | Lean mass made up about 15% or less of weight lost in some studies and 40% to 60% in others |
| STEP UP MRI substudy, semaglutide, 72 weeks[6] | MRI of the thigh, 55 adults | Thigh muscle volume fell about 10% in the semaglutide group, fat inside the muscle went down, and the average 30-second sit-to-stand count held up (about 13 at the start, 15 at the end) |
A 2024 review of the MRI studies concluded that muscle changes on GLP-1 medications look mostly adaptive: muscle shrank roughly as much as expected for the weight lost, while muscle quality improved. The authors still flagged older age and frailty as reasons for extra care, and called for tracking strength and function, not just mass[7].
What these studies did and didn't show
- DXA lean mass includes water and organs, so a 9.7% or 25% figure overstates how much actual muscle was lost[1].
- The MRI study measured thigh muscle directly, but only 55 people had usable scans and only 6 were on placebo[6].
- Most participants did lifestyle programs but not structured strength training, so the results don't show what happens when you lift and eat enough protein.
- None of these used smart scales. Lab scans are far more precise than a bathroom scale.
How accurate are smart scales?
Smart scales estimate body composition by sending a tiny electrical current through your feet (and hands, on some models) and plugging the result into a formula. Here is how they did when researchers checked them:
| Study | What they tested | What they found |
|---|---|---|
| Hospital study, 3 smart scales[8] | Each scale vs a DXA scan, about 50 patients per scale | Weight was accurate (median error 0 to 0.3 kg). Fat mass was off by a median of 2.2 to 4.4 kg, usually reading too low. The authors concluded smart scales are not accurate for body composition |
| 15 impedance devices, 14 consumer models[9] | Each vs a lab 4-part body composition model, 73 adults, retested 12 to 16 weeks later in 37 | Repeat readings were very consistent, but single body fat readings were off by an average of 3.1 to 7.5 percentage points depending on the device. Tracking change over time was better (1.7 to 2.6 points), with some full-body and foot-to-foot models doing best |
| Eating and drinking before a reading[10] | Foot-to-foot scale, 55 adults | A meal or an electrolyte drink nudged body fat readings up by about 1 percentage point |
| DXA scans across a normal day[11] | 5 scans over 2 days, 31 active adults | Daily activities and eating breakfast raised lean mass readings. Fasted morning scans with the same setup were the most reliable |
The takeaway: a smart scale is consistent but not exact. It can show you a direction over weeks. It can't tell you that you lost 1.4 lb of muscle on Tuesday.
What each number on your scale really means
| Reading | What it is | How much to trust it |
|---|---|---|
| Weight | Your actual body weight | High. Look at weekly averages, since daily weight swings with food, salt and water |
| Body fat % | An estimate from electrical resistance plus your height, age and sex | Low for one reading, fair as a 4-week trend under the same conditions |
| Muscle mass or lean mass | Fat-free mass estimate, including water. Brands define it differently | Low for one reading. Early drops on a GLP-1 are often water and glycogen |
| Body water % | An estimate of total water | Moves with what you drank, sweat and ate. Useful only as context |
| Bone mass | A formula estimate, not a bone scan | Very low. Only a DXA bone density scan measures bone |
How to weigh in for cleaner numbers
Clinical guidelines for impedance devices ask for an empty bladder, no recent food, drink or exercise, and the same time of day for repeat measurements[12]. At home, that becomes:
- Weigh in the morning, after the bathroom, before coffee, food or water
- Do it before any workout, not after
- Bare, dry feet on the electrodes, scale on a hard floor in the same spot
- Use the same scale every time. Different brands use different formulas
- Write down a weekly average, and compare body fat and lean mass as 4-week averages
- Expect odd readings after a salty dinner, a hard workout, a rough shot day with little food or fluid, or around a menstrual period
If you get a DXA scan
Ask for the same machine each time, book it first thing in the morning before eating, skip exercise beforehand and wear similar clothes[11]. Look at fat mass, total lean mass and the lean mass in your arms and legs, which is mostly muscle. Space scans a few months apart so real change has time to show up beyond the normal wobble.
4 better ways to tell if you're keeping muscle
| Check | How | How often |
|---|---|---|
| 1. Strength log | Write down the weight or band and the reps for each exercise. Holding steady or climbing while you lose weight is a strong sign you're keeping muscle | Every session |
| 2. 30-second chair stand[13] | Sit in a sturdy chair with no arms (about 17 inches high), arms crossed at your chest. Count how many times you fully stand up in 30 seconds. The CDC lists below-average scores for ages 60 and up, for example fewer than 12 for women and 14 for men aged 60 to 64 | Once a month |
| 3. Everyday tasks | Stairs, carrying groceries, getting up off the floor or out of a low chair. Note if they feel easier, the same, or harder | Once a month |
| 4. Scale trend, done right | Same conditions as above, compared as 4-week averages | Weekly average, monthly comparison |
If strength is holding up and daily life feels easier, a drop in the scale's muscle number is much less worrying. If you're getting weaker for several weeks in a row, that matters more than any single reading.
When the muscle number drops: what to do
| What you see | Most likely reason | What to do |
|---|---|---|
| A big drop in the first few weeks | Water and glycogen leaving as you eat less | Keep training and eating protein. Recheck the 4-week trend |
| A drop after a rough shot day | Less food and fluid that day | Ignore it. Weigh in again once you're eating and drinking normally |
| Body fat % went up while weight went down | Normal reading noise, often from hydration | Compare monthly averages instead |
| A steady drop for 2 months or more, and your strength log is sliding | Some real muscle loss is possible | Check your protein, add a training day or move up a band, and tell your care team |
Use your numbers for protein, not panic
Here is one good use for a body fat reading: setting your protein target. Our house range is 0.7 to 1 g of protein per pound of lean body mass a day. In a 2024 meta-analysis of adults trying to lose weight, eating less than about 1.0 g per kg of body weight a day was linked to a higher risk of losing muscle[14].
Example: a 200 lb person whose scale reads 40% body fat has about 80 lb of fat and 120 lb of lean mass, so they'd aim for 84 to 120 g of protein a day, or about 25 to 30 g at each of 3 or 4 meals. And if the scale is 5 points off? Lean mass shifts by about 10 lb, which moves the target by only 7 to 10 g, roughly the protein in 1 to 2 eggs. Protein math is forgiving. The protein calculator takes your body fat % if you have one, or estimates lean mass from your height and weight with the Boer formula[15] if you don't.
A worked example
Marcus, 58, on Zepbound, watched his smart scale's muscle number drop 5 lb in his first 3 weeks and started worrying it was all muscle. He changed three things. He weighed in every morning after the bathroom and before coffee, and only compared 4-week averages. He started a 2-day resistance band plan and wrote down his reps. And he did the chair stand test on the first Sunday of each month: 11 stands at the start, 15 by month three. The muscle number leveled off after the first month, his band rows moved up two bands, and he used the scale's body fat reading in the protein calculator to set a target he could hit with 4 small meals.
Talk to your prescriber
Body composition numbers are for your own tracking, not a diagnosis. Check in with your prescriber or care team if:
- You have a pacemaker, defibrillator or other implanted electronic device. Scale makers such as Withings advise turning off the body composition feature, since it sends a small current through your body[16]. Weight-only mode is fine
- Everyday tasks like stairs, getting out of a chair or carrying bags have been getting harder for several weeks
- You've fallen, feel unsteady, or get dizzy or lightheaded when you stand up
- You have trouble keeping food or fluids down, which skews scale readings and can leave you short on energy and protein
- You take insulin or another medicine that can cause low blood sugar and you're adding exercise
- You're over 65 or have been told you have low muscle mass or frailty, and want a scan or a plan to protect strength
Don't change how you take your medication based on a scale reading. Bring your trends, your strength log and your questions to your next visit instead.
Where to start
Set your daily target with the free protein calculator, take the 1-minute muscle check to see where to focus, and start a strength log with the 8-move resistance band workout. For the bigger picture, read what the studies say about GLP-1 muscle loss.
Free: Protect Your Muscle on GLP-1s: Starter Guide
A free printable starter guide for anyone on GLP-1 medications: strength and protein basics, a simple first week, and questions to ask your provider.
Frequently asked questions
Is my smart scale accurate for muscle mass?
For body weight, usually yes. For muscle and body fat, not very. In a hospital study comparing 3 smart scales with DXA scans, the scales got weight right to within a fraction of a kilogram, but fat mass was off by a median of about 2 to 4 kg, and the authors concluded smart scales are not accurate for body composition. Use the muscle and body fat numbers as rough trends, measured the same way each time, not as exact amounts.
Why does my smart scale say I'm losing muscle on Ozempic or Zepbound?
Smart scales and DXA scans count water as lean mass. When you eat less, your body uses up stored carbohydrate (glycogen), which is held with about 3 to 4 parts water, so the lean number can fall quickly in the first weeks even if little muscle is lost. In GLP-1 trials, lean mass did fall as part of the weight lost, which is why strength training and protein still matter.
Is lean mass the same as muscle?
No. Lean mass, or fat-free mass, is everything that isn't fat: muscle, water, organs, skin, connective tissue and, on some reports, bone. About 73% of fat-free mass is water. So a drop in lean mass is not the same as an equal drop in muscle.
Should I get a DXA scan while on a GLP-1?
It can help if your care team offers it, but it isn't required. If you do get scans, make them comparable: same machine, in the morning before eating, before any workout, and in similar clothes. In one study, daily activities and eating breakfast raised DXA lean mass readings, so an unplanned afternoon scan can mislead you.
How can I tell if I'm keeping muscle without a scan?
Track strength. Log your reps and weights or bands for each exercise, do a 30-second chair stand test once a month, and notice whether stairs and carrying groceries feel easier or harder. Strength holding steady or rising while you lose weight is one of the best everyday signs you're protecting your muscle.
Sources
- Wang et al. Hydration of fat-free body mass: review and critique of a classic body-composition constant. Am J Clin Nutr, 1999
- Kreitzman et al. Glycogen storage: illusions of easy weight loss, excessive weight regain, and distortions in estimates of body composition. Am J Clin Nutr, 1992
- Wilding et al. STEP 1 body composition (DXA) substudy. Journal of the Endocrine Society, 2021
- Look et al. Body composition changes during weight reduction with tirzepatide in SURMOUNT-1. Diabetes, Obesity and Metabolism, 2025
- Neeland et al. Changes in lean body mass with GLP-1-based therapies. Diabetes, Obesity and Metabolism, 2024
- Hjelmesæth et al. Effect of semaglutide on thigh skeletal muscle volume, fat infiltration and physical function in people with obesity: secondary analysis of STEP UP. Diabetes Obes Metab, 2026
- Linge, Birkenfeld and Neeland. Muscle mass and glucagon-like peptide-1 receptor agonists: adaptive or maladaptive response to weight loss? Circulation, 2024
- Frija-Masson et al. Accuracy of smart scales on weight and body composition: observational study. JMIR mHealth and uHealth, 2021
- Siedler et al. Assessing the reliability and cross-sectional and longitudinal validity of fifteen bioelectrical impedance analysis devices. British Journal of Nutrition, 2023
- Androutsos et al. Impact of eating and drinking on body composition measurements by bioelectrical impedance. J Hum Nutr Diet, 2015
- Nana et al. Effects of daily activities on dual-energy X-ray absorptiometry measurements of body composition in active people. Med Sci Sports Exerc, 2012
- Kyle et al. Bioelectrical impedance analysis, part II: utilization in clinical practice (ESPEN). Clinical Nutrition, 2004
- CDC STEADI: Assessment, 30-Second Chair Stand
- Kokura et al. Enhanced protein intake on maintaining muscle mass in adults with overweight/obesity: systematic review and meta-analysis. Clinical Nutrition ESPEN, 2024
- Boer P. Estimated lean body mass as an index for normalization of body fluid volumes in humans. American Journal of Physiology, 1984
- Withings Support: Body+ frequently asked questions about safety (pacemakers and implanted devices)
Not medical advice. This article is general information from published research and public guidelines, not a diagnosis or a plan. Talk with your prescriber or healthcare provider before changing how you eat, move or take your medication.
More in Muscle & strength
- GLP-1 muscle loss: what the studies actually measured, and how to keep yours
- Can you build muscle on a GLP-1 like Ozempic or Zepbound?
- Cardio vs strength training on a GLP-1: which matters more?
- Does Ozempic cause bone loss? What the research shows, and how strength training helps
- All Muscle & strength articles