VA study finds doubled risk of rare optic nerve stroke with semaglutide

A large study of U.S. veterans with type 2 diabetes found that people who started semaglutide had roughly twice the risk of a rare optic nerve condition that causes sudden vision loss compared with people who started a different class of diabetes drugs. The findings were published February 12, 2026, in JAMA Ophthalmology ^[1][2].

The condition, nonarteritic anterior ischemic optic neuropathy (NAION), is a leading cause of sudden, painless and usually one-sided vision loss. It is sometimes called "eye stroke" because it results from interrupted blood flow to the optic nerve, and it often appears upon waking. Adults 50 and older are most at risk ^[2].

The researchers emulated a randomized target trial using Veterans Health Administration data from March 1, 2018, through March 1, 2025 ^[2]. The analytic cohort included 102,361 veterans with type 2 diabetes who were taking metformin and had not previously used a GLP-1 receptor agonist or an SGLT2 inhibitor. Of those, 11,478 started semaglutide and 90,883 started an SGLT2 inhibitor as a second-line diabetes therapy ^[2].

Participants had a mean age of 60, a mean body mass index of 37.8 and a mean hemoglobin A1c of 7.0%. Most, 85.5%, were male; 61.9% were non-Hispanic white, 20.7% were Black and 8.1% were Hispanic ^[2].

Over a maximum follow-up of 7.5 years, 173 NAION events occurred ^[1][2]. The incidence rate was 123 per 100,000 person-years among semaglutide initiators and 67 per 100,000 person-years among SGLT2 inhibitor initiators ^[2]. In 2.1 years of median follow-up, semaglutide initiators had a hazard ratio of 2.33 (95% CI, 1.54-3.54; P < .001), the researchers reported ^[2]. The overlap-weighted cumulative risk was 0.29% with semaglutide versus 0.13% with SGLT2 inhibitors, an average treatment difference of 0.16 percentage points ^[1][2].

The authors wrote: "While absolute incidence of NAION is low in patients initiating semaglutide, medical counseling about this potential complication, which can result in substantial loss of vision, may be warranted" ^[2]. They added that clinicians and patients should be counseled about the rare but evident increased risk, and that patients with visual symptoms should be evaluated promptly ^[2].

The biological mechanism linking GLP-1 drugs to NAION remains unclear. Proposed explanations include low blood pressure, fluid loss from gastrointestinal side effects, rapid improvement in blood sugar with temporary changes in small blood vessels, and impaired blood flow regulation at the optic nerve head ^[2].

Why it matters for patients

The absolute risk is small. Weighted cumulative risk was 0.29% over about two years with semaglutide, fewer than 3 in 1,000 initiators, against about 1.3 in 1,000 in the comparison group. The authors described the absolute incidence as low, roughly one additional case per several thousand treated patients ^[2]. Semaglutide is used by an estimated 15 million U.S. adults ^[2].

Several limits matter for reading the numbers. The comparison was semaglutide against SGLT2 inhibitors, not placebo, so the study does not measure NAION risk against taking no diabetes medication ^[1][2]. Whether risk varies by dose, by how long someone takes the drug, or among the different semaglutide products sold as Ozempic, Wegovy and Rybelsus has not been reported; the authors called for future studies of dose and duration relationships ^[2]. All participants had type 2 diabetes and were using metformin, so whether the findings extend to people taking semaglutide only for weight loss is not established ^[2].

What happens next

A correction to a figure in the paper has been published, and two comment letters raising questions about the data and the timing of risk have appeared in the journal ^[1]. A separate JAMA Network Open study of GLP-1 receptor agonists, SGLT2 inhibitors and NAION was published April 30, 2026 ^[1]. The authors said future per-protocol analyses should characterize dose and duration response relationships ^[2].