Muscle Loss: Drug Versus Dieting

“These drugs make you lose muscle” is the most repeated criticism of GLP-1 medicines, and it is almost never accompanied by the comparison that would make it meaningful. A meta-analysis of 20 randomized trials in 15,782 people supplied it: the share of weight lost as lean mass on incretin drugs is broadly the same as on dieting.^[1]

The comparison

The comparison between incretin drugs as a group and lifestyle intervention returned p=0.42 — no significant difference.^[1] Losing weight costs lean tissue. It costs about the same proportion whether the weight comes off through a drug or through dieting.

The criticism is real and it is not specific to the drugs. It is a property of losing weight.

The row that actually helps

Adding resistance training brought the lean share down to 17.5% — the most favorable profile in the analysis, and the only intervention here that changes the answer rather than describing it.^[1]

That is the practical conclusion, and it is the same one arrived at from a completely different direction in the exercise trial, where the drug alone did nothing for physical fitness and training did. Two independent lines of evidence, one instruction: lift something.

The difference between the drugs is not nothing

It would be convenient to flatten this into “all the same”, and the data do not allow it. Semaglutide sits at 35.2% and tirzepatide at 25.4%, and those confidence intervals do not overlap. Within the class, they differ.

What that means is genuinely unclear from this analysis. It could reflect the drugs, the speed of loss they produce, the trial populations, or how body composition was measured in each study — heterogeneity across the pooled trials was moderate. It is a real difference in the numbers and not yet an explained one, and it is not a reason to choose between them.

What lean mass does and does not tell you

Every figure above is a scan measurement. Whether the person can do more or less is a different question, and one that scans answer badly — we found a study where muscle volume fell 9.3% while chair-rise and walking speed both edged better, covered in muscle volume versus muscle function.

So the full picture is three-part: lean mass falls with any weight loss, at a similar proportion drug or diet; resistance training substantially reduces that share; and the mass number alone does not establish that anyone has been harmed.

The pooled figure, across every drug class

A 2026 network meta-analysis pooled 41 randomized trials and 2,906 participants to place every major diabetes and obesity drug against a common reference on fat and lean mass at once.^[2] That structure is what makes the numbers comparable, rather than a collection of separate trials read side by side.

Do the arithmetic on the tirzepatide row: 4.40 kg of lean against 10.70 kg of fat means lean tissue made up roughly 29% of what was lost. That is the best pooled estimate available, and it sits close to what is generally expected of substantial weight loss by any means.

The other direction is worth seeing too. Insulin glargine was associated with fat mass gain, which is the mirror image of what these drugs do and a reminder that “a diabetes drug” is not one thing.