Do GLP-1 Drugs Like Ozempic and Wegovy Cause Muscle Loss?

GLP-1 receptor agonists mimic a gut hormone that slows stomach emptying, dampens appetite and improves insulin response. Tirzepatide adds a second hormone pathway (GIP) on top of GLP-1, which is part of why it tends to produce somewhat larger average weight loss in head-to-head data. Both classes were developed for type 2 diabetes first and later approved specifically for chronic weight management at higher doses — which matters, because a lot of the safety data below comes from diabetes trials, and the risk picture can shift slightly when the same drug is used by people without diabetes, at a different dose, purely for weight loss.

The mechanism is not mysterious: any large, sustained calorie deficit — drug-assisted or not — pulls weight from both fat and lean tissue unless you specifically counteract it. The real question is the ratio, and here the trial data is more reassuring than the “Ozempic makes you weak” framing suggests.

A DXA-scan substudy of the SURMOUNT-1 trial (160 of the 2,539 participants) found that at 72 weeks, tirzepatide produced a 21.3% drop in body weight, made up of a 33.9% drop in fat mass and a 10.9% drop in lean mass — versus 5.3%/8.2%/2.6% on placebo (p<0.001 for every comparison).¹ Do the arithmetic and roughly three-quarters of the weight lost was fat, one-quarter was lean tissue — a split that is fairly typical of substantial weight loss in general, not something unique to the drug.

A companion MRI substudy of SURPASS-3, in people with type 2 diabetes, measured actual thigh muscle volume rather than whole-body DXA lean mass: tirzepatide reduced muscle volume by 0.64 L over 52 weeks (p<0.0001) — but it also reduced muscle fat infiltration by 0.36 percentage points, meaning the muscle that remained had less fat marbled through it, a marker of quality rather than pure shrinkage. Crucially, the size of the muscle-volume drop tracked almost exactly what population data (from UK Biobank) predicts for that amount of general weight loss — the loss wasn’t an unusual drug-specific effect, just proportional to how much weight came off.²

For semaglutide, a 2024 systematic review of six trials covering 1,541 adults found lean-mass loss ranging from close to 0% up to about 40% of total weight lost, with larger reductions more common in the biggest trials — though the share of lean mass relative to total body weight often still improved overall.³ The most encouraging data point comes from the prospective SEMALEAN study (n=106): weight fell 10% at 7 months and 13% at 12 months, lean mass dropped about 3 kg by 7 months and then stabilised, and by 12 months handgrip strength had actually improved by 4.5 kg on average, while the proportion of participants classified with sarcopenic obesity fell from 49% to 33%.⁴ Muscle mass and muscle function are not the same thing — this suggests that, managed well, the drop in scale-weight muscle doesn’t have to translate into a weaker body.

The countermeasure is the same one that works for any diet-induced weight loss: enough protein, distributed across meals, plus resistance training to give the body a reason to keep the muscle it has. We cover the specific numbers — roughly 1.2 g/kg/day or more, paired with strength training — in our proteinmaxxing guide, which was written for exactly this situation. Appetite suppression makes it easy to under-eat protein at the precise moment preserving it matters most, so this is not an optional add-on for people on these drugs; it is the main lever you actually control.

Nausea, vomiting, diarrhoea and constipation are the most common side effects of both drugs, are dose-dependent, and are usually mild-to-moderate and transient — though they are also the leading reason people stop treatment, with discontinuation for GI reasons roughly twice as likely as on placebo.

Gallbladder disease is the more concrete long-term concern, and it is driven partly by rapid weight loss itself (which raises cholesterol in bile and slows gallbladder emptying) rather than the drug alone. In 13 randomised trials of obese adults without diabetes (26,894 participants), semaglutide raised the risk of gallstones (cholelithiasis) by about 2.6 times versus placebo (95% CI 1.40–4.82); tirzepatide showed no significant extra biliary risk in that same comparison.⁵ A separate meta-analysis in type 2 diabetes and obesity trials found the combined gallbladder-or-biliary-disease outcome nearly doubled with tirzepatide (RR 1.97, 95% CI 1.14–3.42) versus placebo or basal insulin, while pancreatitis itself was not significantly increased (RR 1.46, 95% CI 0.59–3.61).⁶

Both drug classes carry an FDA boxed warning about thyroid C-cell tumours, based on a dose-dependent effect seen in rats, not humans — and it is the reason both are specifically contraindicated for anyone with a personal or family history of medullary thyroid carcinoma or MEN2 syndrome. The best current human evidence is reassuring for everyone else: a February 2026 systematic review and meta-analysis of 48 randomised placebo-controlled trials (94,245 participants) found GLP-1 receptor agonists probably have little or no effect on thyroid cancer risk (OR 1.37, 95% CI 0.82–2.31 — not statistically significant), and the same held for pancreatic, breast and kidney cancer.⁸ A rodent-specific warning that hasn’t shown up in nearly 100,000 human trial participants is a reason for the specific exclusion above, not a reason for general alarm.

In 2023, reports from Icelandic regulators triggered both FDA and EMA reviews into whether GLP-1 drugs raise the risk of suicidal thoughts or behaviour. The EMA concluded in April 2024 that the evidence didn’t support a link. The FDA went further: in a comprehensive meta-analysis of 91 placebo-controlled trials covering 107,910 participants, it found no increased risk of suicidal ideation or behaviour, or of other psychiatric adverse events such as anxiety, depression or psychosis — and in January 2026 formally requested that the warning be removed from GLP-1 weight-loss drug labels.⁹ If you started one of these medications while that warning was still on the label, this is the update: the best current evidence, at far larger scale than the original case reports, doesn’t back it up.