New-Onset Nonarteritic Anterior Ischemic Optic Neuropathy and Initiators of Semaglutide in US Veterans With Type 2 Diabetes.

IMPORTANCE: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are considered safe, effective medications for type 2 diabetes (T2D) and weight loss, used by millions worldwide. While their cardiometabolic benefits are well established, emerging observations suggest a potential association between GLP-1RA use and new-onset nonarteritic anterior ischemic optic neuropathy (NAION). OBJECTIVE: To emulate a target trial evaluating the risk of NAION associated with initiation of semaglutide (GLP-1RA), compared with a sodium-glucose cotransporter-2 inhibitor (SGLT2i) as second-line therapy for T2D in a nationwide cohort of US veterans. DESIGN, SETTING, AND PARTICIPANTS: This study was conducted nationwide using data from the Veterans Health Administration health care system between March 1, 2018, and March 1, 2025. This active-comparator, new-user, target trial emulation used cause-specific hazard ratios (HRs) that were estimated using overlap weighting to account for confounding. Participants included US veterans with T2D, current metformin use, and no prior GLP-1RA or SGLT2i use. Data analysis was conducted from July 2025 through September 2025. EXPOSURE: Initiation of semaglutide or any SGLT2i. MAIN OUTCOME AND MEASURE: Incident NAION, identified using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision and Systematized Nomenclature of Medicine codes. RESULTS: A total of 102 361 US veterans met inclusion criteria, including 11 478 initiators of semaglutide and 90 883 initiators of an SGLT2i. Baseline characteristics were well balanced between treatment groups after overlap weighting (mean [SD] age, 60.1 [11.7] years; body mass index, 37.8 [6.7]; hemoglobin A1c, 7.0% [1.4]; 85.5% male and 14.5% female; 20.7% Black, 8.1% Hispanic, and 61.9% non-Hispanic White). Over a maximum follow-up of 7.5 years, 173 total incident NAION events occurred. The incidence rate of NAION was 123 per 100 000 person-years among semaglutide initiators and 67 per 100 000 person-years among SGLT2i. In 2.1 years of median follow-up, semaglutide initiators had a 2.33-fold higher risk than SGLT2i initiators (hazard ratio, 2.33; 95% CI, 1.54-3.54; P < .001). The overlap weighted incidence rate of NAION was 0.29% for semaglutide initiators and 0.13% for SGLT2i initiators, with a corresponding average treatment effect of 0.16 percentage points. CONCLUSIONS AND RELEVANCE: In this nationwide cohort of US veterans with T2D, semaglutide initiators had a 2-fold NAION risk than SGLT2i initiators, while the absolute risk was low. Clinicians and patients should be counseled on the rare but evident increased risk of NAION after semaglutide initiation.