Abstract Introduction GLP-1 receptor agonists (GLP-1RAs) reduce cardiovascular mortality in type 2 diabetes (T2DM) while improving obstructive sleep apnea (OSA) severity. Whether individual GLP-1RAs differ in mortality benefit, and whether OSA modifies these agent-specific effects, remains unknown. Methods Using the U.S. Collaborative Network (122M patient records across 70 U.S. healthcare organizations) on the TriNetX data platform, we identified patients (using encounter ICD-10-CM diagnostic codes) between 10/2015 – 10/2024 with a history of T2DM who were prescribed metformin with or without GLP-1RA (using RxNorm codes). We additionally stratified these patients as having a history of OSA (2 or more encounter diagnostic codes) and compared 1-year all-cause mortality from time of tirzepatide, semaglutide, dulaglutide, or liraglutide prescription (versus metformin if not prescribed GLP-1RAs). Cox proportional hazards models compared mortality between those patients prescribed GLP-1RAs and those not, as well as in head-to-head comparisons between agents, adjusting these models for age, sex, race, as well as BMI and HbA1c category. Results We identified 2,471,348 patients with T2DM, of which 7141, 91 were prescribed at least one GLP-1RA: most commonly semaglutide, followed by dulaglutide, liraglutide, and tirzepatide. The GLP-1RA class reduced mortality by 59% in OSA (HR 0.41, 95% CI 0.39 – 0.43), versus 55% in non-OSA patients (HR 0.45, 95% CI 0.44 – 0.47). A consistent potency hierarchy emerged: tirzepatide (OSA: HR 0.24; non-OSA: 0.23), semaglutide (OSA: HR 0.36; non-OSA: 0.33), dulaglutide (OSA: HR 0.47; non-OSA: 0.50), and liraglutide (OSA: HR 0.50; non-OSA: 0.54), all p 0.001. Older agents (dulaglutide, liraglutide) were associated with enhanced mortality benefit in OSA, while tirzepatide and semaglutide showed equivalent benefit regardless of OSA status. Head-to-head analyses confirmed tirzepatide superiority over semaglutide in both populations (OSA: HR 0.69; non-OSA: HR 0.73). Dulaglutide and liraglutide were equivalent in OSA (HR 0.90, p=0.06) but not in non-OSA (HR 0.86, favoring dulaglutide, p=0.006). Conclusion GLP-1RAs demonstrate substantial mortality reduction in T2DM, with enhanced class effect in OSA patients. A clear potency hierarchy exists, with tirzepatide consistently showing the greatest mortality reduction, followed by semaglutide. The class effect observed in OSA appeared driven primarily by older agents, with tirzepatide and semaglutide yielding similar benefit regardless of OSA status. Support (if any)
Song et al. (Fri,) studied this question.