Study wrapper · #169
Weight-loss dynamics with tirzepatide versus semaglutide.
Editor's note
This large real-world retrospective cohort (10,339 matched pairs from electronic health records) compares weight-loss dynamics for tirzepatide versus semaglutide, using AI-assisted note curation for adverse events. Researchers reported greater mean weight reduction with tirzepatide (14.7% vs 10.8%), nearly double the rate of high responders (≥15% in year 1: 42.6% vs 21.6%), faster weight-loss velocity, and a lower recorded prevalence of gastrointestinal and systemic adverse events. Because both cohorts drew on routine care with 1:1 propensity matching, the comparison is more directly head-to-head than pooled trial data — a genuine strength. But confounding by indication, dosing differences, adherence, and AI-based adverse-event extraction from free text are real limitations, and demographic disparities in response (by sex and race) warrant caution about generalising. This aligns with the pattern seen in randomised comparisons like SURMOUNT-5. Weigh it as strong, direction-consistent real-world evidence that tirzepatide is associated with greater weight loss, while treating the safety comparison as exploratory.
Plain-language abstract
Using real-world medical records, researchers compared two popular medications — tirzepatide and semaglutide — for weight loss. They matched 10,339 people on each drug so the groups were similar in age, sex, diabetes status, and starting weight, then tracked results over two years. On average, people taking tirzepatide lost more weight (about 14.7%) than those on semaglutide (about 10.8%), and the share of 'high responders' who lost at least 15% in the first year was nearly double with tirzepatide (about 43% vs 22%). Tirzepatide users also lost weight faster and, based on an AI review of clinical notes, had fewer recorded stomach-and-gut and whole-body side effects. Results varied by group: women and White patients were more common among the biggest responders, while Black and Hispanic patients were more common among those who lost the least. Because this used routine-care data rather than a randomised trial, it shows strong associations but cannot fully rule out other explanations.