If you lose weight by dieting, how much of it is muscle?

Does Dieting Burn Your Muscle? A 6-Month Trial Scanned 374 Adults

2026-10-31 · By Anthony Falcone

Verdict: Keep it, with a caveat. Less than the warnings suggest, at least in this trial. After 6 months of diet-only weight loss, lean tissue in the arms and legs, the best stand-in for muscle on a DXA scan, fell by 0.80 kg in women and 1.02 kg in men, and relative to their new body size people had slightly more of it, not less. The caveat: DXA doesn't measure muscle directly, there was no comparison group for this question, and the participants were aged 18–50 and lost a moderate amount.

What the researchers did

A common worry about dieting is that a big share of the weight you lose is muscle. Earlier diet studies put lean tissue at roughly 20% to 30% of the weight lost, but many of them used imperfect ways to measure it, and few asked whether the loss actually matters for someone who is now smaller.

This team at Stanford went back to DIETFITS, a large trial in which 609 adults aged 18–50 with overweight or obesity and no diabetes were randomly assigned to a healthy low-fat or a healthy low-carbohydrate diet for a year. Because of funding limits, DXA body scans were only offered from the second group of recruits onwards, so 374 people had a scan at the start and at 6 months, when most of the weight loss had happened. Their average age was 39.4, their average BMI 32.3, and 61% were women.

The original trial had found no significant difference in weight or body fat lost between the two diets after a year, so the authors pooled both groups. They analysed women and men separately, because their body composition differs.

Three ways of looking at "muscle"

A DXA scan doesn't see muscle. It splits the body into fat, bone and "lean soft tissue", which includes organs, connective tissue and water as well as muscle. So the authors looked at the scan three ways.

First, the arms and legs only. Lean tissue there is about 70% skeletal muscle, which makes it a better stand-in for muscle than the whole-body figure.

Second, a correction for fat tissue itself. Fat tissue is roughly 85% fat and 15% water, protein and minerals, and DXA counts that 15% as lean. When you lose a lot of fat, some "lean" loss is really the rest of the fat tissue going with it. The authors subtracted it with a simple formula (lean tissue minus 0.176 times fat mass).

Third, lean tissue relative to body size. They compared each person's arm and leg lean tissue with what would be expected for their height, weight and background, where 100% means exactly as expected.

What they found

Over 6 months both sexes lost about 6.9% of their starting weight: 5.9 kg on average for women and 7.18 kg for men. Most of it was fat, 4.28 kg in women and 4.92 kg in men.

On the whole-body scan, lean tissue fell by 1.61 kg in women and 2.26 kg in men, a median of 26.7% and 32.5% of the weight each person lost. That is in line with earlier diet studies. In the arms and legs, though, the drop was smaller: 0.80 kg in women and 1.02 kg in men, a median of 12.8% of the weight lost in both sexes. After the fat-tissue correction it shrank again, to 0.43 kg and 0.69 kg, or 6.7% and 8.2% of the weight lost.

Relative to body size, people didn't come out worse on average. Arm and leg lean tissue went from 100.5% to 101.4% of the expected value in women and from 100.2% to 101.1% in men, a small but statistically significant increase. In other words, after losing weight people had, if anything, slightly more lean tissue for their new size. The authors suggest some lean loss may simply be the body adapting to carrying less weight.

The diet didn't change the picture. Men on the low-carbohydrate diet lost more weight and more lean tissue in kilograms, but the share of weight lost as lean tissue was the same on both diets. Self-reported physical activity wasn't linked to how much lean tissue people lost.

Individuals varied: most lost both fat and lean tissue, but some gained lean tissue while losing fat. The authors also tested 242 blood proteins, and changes in 10 in women and 27 in men were linked to keeping more lean tissue, with one called DLK1 at the top. That part is exploratory and doesn't tell you anything to do yet.

What this means for you

Don't let fear of muscle loss stop you from losing weight with a sensible diet: in this trial most of what people lost was fat, and arm and leg lean tissue barely moved.
If a body scan or smart scale shows your lean mass dropping as you diet, read it with care. Part of that "lean" loss is water and the non-fat part of fat tissue, and a smaller body needs a bit less lean mass to carry it.
This study didn't test protein intake or strength training, so it can't tell you how much they add. It only shows that a moderate diet-based weight loss didn't cost these adults much lean tissue.
If you have a medical condition, take regular medication or are older, talk to your doctor before starting a weight-loss diet.

How much to trust it

The authors' own: DXA estimates lean tissue rather than measuring muscle, and MRI would be more accurate; the protein analysis covered only 242 blood proteins linked to heart health and inflammation; and without muscle biopsies the study can't say anything about muscle quality.
No comparison group for this question. Both diet arms were pooled, so this is a before-and-after look within one trial, not a randomised test of dieting against not dieting. It's also a secondary analysis, done after the trial.
Only 374 of the 609 participants are included, because scans were optional and not offered to the first group of recruits, so they may not represent the whole trial.
A specific group: adults aged 18–50 (average 39.4), mostly White, with a BMI of 28–40 and no diabetes, losing about 6.9% of their weight. The authors note their younger age may have protected against lean loss; older adults or bigger losses may look different.
Physical activity came from a questionnaire, and this paper doesn't report the participants' protein intake or any strength training, so we can't tell how much either contributed.
The authors report no funding for this analysis. One senior author declares founder, adviser or investor roles in many biotech and health-tech companies; the others report no conflicts of interest.

The study

Roberts AK, Panyard DJ, Hislop B, Ward CP, Snyder MP, Gardner CD, Haddad F. Diet-based weight-loss intervention is not associated with a meaningful change in lean soft tissue. The American Journal of Clinical Nutrition, 2026. doi.org/10.1016/j.ajcnut.2026.101251

Our Fat Loss Protocol sets your calorie, protein and weekly-loss targets, shows how to protect your muscle with protein and training, builds meals that keep you full and plans your first four weeks, plateaus included. See the protocol

Want to change how you eat first? The Keto Blueprint is a 14-day, research-checked plan.