Patients who were on tirzepatide before rotator cuff surgery came back for a repeat repair about half as often as patients on semaglutide, 2.83 percent versus 5.67 percent over two years. Both drugs are in the same family. That they split apart is the interesting part.
A rotator cuff repair reattaches torn tendons in the shoulder, and a meaningful fraction of repairs fail and have to be redone. Earlier work had found that patients on semaglutide did better after the operation than patients on no weight-loss drug at all. Nobody had put the two leading drugs head to head for this.
The study ran in the Orthopaedic Journal of Sports Medicine ↗ this month. A University of Pennsylvania team led by Praneeth Thota and John Kelly pulled TriNetX records for adults aged 40 to 80 who had a first arthroscopic rotator cuff repair. They sorted patients by whether they had been on tirzepatide or semaglutide in the year before surgery, then ran two matched comparisons: tirzepatide against semaglutide, and either drug against patients on neither.
In the head-to-head, the tirzepatide group's two-year rate of a repeat repair was half the semaglutide group's, a relative risk of 0.50. The same group filled fewer opioid prescriptions afterward (48.2 versus 55.7 percent) and used less physical therapy (36.5 versus 42.5 percent). In the second comparison, patients on either drug had fewer complete re-tears than patients on neither, holding at 90 days, one year, and two years.
The catch sits in what got counted. The repeat-surgery figure is a billing code, a second instance of the procedure code for arthroscopic cuff repair, not a re-tear confirmed on a scan. A lower code rate can mean the tendon held. It can also mean the patient did not come back, was never re-imaged, or had the second operation recorded outside the network. The authors say it directly. The study describes differences in procedure coding and health-care use, not confirmed structural healing. It is a retrospective records study, level of evidence 3, which sits well below a trial.
So why would two drugs in the same class come apart? The leading guess is the receptors each one hits. Semaglutide ↗ works on the GLP-1 receptor ↗ alone. Tirzepatide ↗ adds the GIP receptor ↗ on top of it, and whether that second arm does anything for tendon or soft-tissue recovery is unknown. A records study can flag the split; it cannot test the cause.
The same dual-versus-single pattern showed up in our coverage of tirzepatide edging dulaglutide on heart failure ↗, where the drug again looked better than a single-receptor agonist on an outcome nobody had set out to measure. Two signals, same shape: tirzepatide pulling ahead of a single-receptor cousin on something the study was not built to confirm. The second receptor is a plausible reason. A plausible reason is not a healed tendon.