Evolving utilization of bariatric surgery since the rise of semaglutide and tirzepatide

Semaglutide and tirzepatide have transformed obesity treatment, but recent changes to bariatric surgery utilization are not well understood.

Epic's nationwide Cosmos database was queried for patients who underwent primary sleeve gastrectomy or gastric bypass between 2018 and 2025. Patient characteristics including preoperative semaglutide or tirzepatide dispense history were compared using chi-squared and Wilcoxon rank sum tests. Modified Poisson regression was used to identify factors independently associated with pre-surgery GLP-1RA use. Multilevel models were used to examine hospital- and state-level variation in pre-surgery GLP-1RA use.

Bariatric surgery utilization increased after Q3 2018, peaked in Q4 2022, and subsequently decreased 39% through Q4 2025. Between Q4 2018 and Q4 2025, the proportion of Hispanic bariatric surgery patients increased (8.1% vs. 16.8%, p < 0.001), and the proportion of patients who received pre-surgery GLP-1RA increased (0.2% vs. 35.3%, p < 0.001). Factors associated with receiving pre-surgery GLP-1RA were year, private insurance, White race, type 2 diabetes (RR 2.94 [2.88-3.00]), older age, sleep apnea, and metabolic dysfunction-associated steatotic liver disease. Factors associated with receiving surgery upfront were Hispanic ethnicity, Black race, and public or no insurance. After adjusting for patient characteristics and year, there was a 15-fold difference in pre-surgery GLP-1RA use between the highest and lowest hospitals (RR 0.19-2.88).

In the Epic Cosmos database, bariatric surgery utilization decreased from 2022 to 2025, and those who underwent bariatric surgery increasingly received GLP-1RA before surgery. Patients who received pre-surgery GLP-1RA were older, White, privately insured, with diabetes and other weight-related comorbidities, while patients who received surgery upfront were Hispanic, Black, and publicly insured. Pre-surgery GLP-1RA use was also driven by center-specific non-clinical factors.