Non-diabetic, obese patients taking a GLP-1 drug were about twice as likely to be readmitted to the hospital within 90 days of a shoulder replacement as similar patients not taking one. Two years out, that gap had closed. Infection, fracture, and the need for a second operation all came out the same. The split comes from a claims-database study of 4,839 shoulder-replacement patients published in the Journal of Shoulder and Elbow Arthroplasty ↗.
The narrow slice the study isolates
Most research on GLP-1 drugs and surgery has looked at people with diabetes, because that is who took these drugs first. This analysis deliberately cut diabetes out. Using the PearlDiver Mariner insurance database, the authors found adults with a body mass index of at least 30 who had an anatomic or reverse shoulder replacement between 2010 and 2023, excluded anyone with diabetes, cancer, infection, or a pathological fracture, and kept only patients who filled a GLP-1 prescription within three months of surgery. That left 462 GLP-1 users, matched against 4,377 non-users of the same age, sex, and comorbidity profile. The GLP-1 class is the appetite-and-glucose family that includes semaglutide ↗ and tirzepatide ↗; here the drugs were being used for weight, not blood sugar.
At 90 days, the GLP-1 group had higher odds of readmission (odds ratio 2.03, roughly double) and higher overall complication rates (odds ratio 1.47, roughly 50 percent higher). Every other 90-day measure, and every two-year measure the authors checked, showed no difference between the groups.
Early and transient, not structural
The shape of that result matters more than either number alone. A risk that spikes in the first three months and then vanishes by two years does not look like a drug weakening bone or wrecking the joint. It looks like something happening around the operation itself. The authors point at readmissions as the driver and suggest the drugs are probably reasonable to use around surgery, with effort put into keeping patients out of the hospital in those first weeks.
What a billing database cannot tell you is why. Readmission is a blunt endpoint. It could be nausea or vomiting, dehydration, or the aspiration risk that has driven anesthesiologists to debate whether patients should pause GLP-1 drugs before going under. These drugs slow stomach emptying, which is part of how they curb appetite, and a stomach that empties slowly is a stomach that can still hold food when a surgeon expects it to be empty. The American Society of Anesthesiologists told doctors to hold the drugs before elective surgery in 2023, then softened that in 2024 to let most patients keep taking them. This study lands in the middle of that unsettled call. The study counts the readmissions. It cannot name the mechanism, the same limit that dogged an earlier records study where weight-loss surgery beat semaglutide on heart outcomes and the database could not say why ↗.
The signal also runs against some other orthopedic data. GLP-1 users came through spine surgery clean and fused better ↗ in one recent analysis, and in a shoulder comparison tirzepatide users needed half the repeat rotator-cuff surgery of semaglutide users ↗. Shoulder replacement adds a wrinkle to that picture: a real but short-lived bump in early readmissions that leaves no mark on the joint two years later.