The Muscle Question: What Large-Scale Data Shows About GLP-1 Drugs and Lean Mass

One of the most common concerns raised about GLP-1 medications is how much of the weight lost is muscle rather than fat, and this study tackled that question at a genuinely large scale. Researchers built a pipeline that uses AI-based language extraction to pull body-composition data directly out of electronic health records, then applied it to 670,422 people starting GLP-1 therapy for the first time: 456,742 on semaglutide and 213,680 on tirzepatide. Across this routine-care population, tirzepatide users showed a greater decline in lean body mass than semaglutide users. This is an important, practically useful finding precisely because it reflects real-world use rather than a controlled trial setting. That said, it is worth being clear this is a preprint, posted on medRxiv and not yet through formal peer review, and the novel EHR-extraction method, while promising, has not been independently validated at this scale before. Read it as an early, large, and thought-provoking signal rather than a settled clinical fact.