Pooled analysis estimates about one extra NAION case per 7,000 semaglutide patients a year

A pooled analysis of observational studies published May 21 in PLOS Medicine reports that semaglutide (sold as Ozempic, Wegovy and Rybelsus) was associated with a higher relative risk of nonarteritic anterior ischemic optic neuropathy (NAION) in people with type 2 diabetes, while the absolute risk stayed low — about one extra case per 7,000 patients treated for one year ^[1].

The paper describes NAION as "a rare but vision-threatening condition." The analysis pooled five primary studies and added seven more for sensitivity testing, all observational and largely registry-based. It compared patients with type 2 diabetes aged 12 and older taking semaglutide against those taking other glucose-lowering medicines. Database searches covered January 2023 through November 2025 ^[1].

On the relative measures, semaglutide use was tied to a hazard ratio for NAION of 2.17 (95% confidence interval 1.73 to 2.74; p < 0.001) compared with non-semaglutide glucose-lowering regimens. Against regimens that excluded other GLP-1 receptor agonists, the hazard ratio was 2.13 (95% CI 1.60 to 2.83; p < 0.001), and against users of SGLT2 inhibitors it was 1.96 (95% CI 1.28 to 2.99; p = 0.002) ^[1].

In absolute terms, the researchers estimated a risk difference of 0.014% (95% CI 0.005% to 0.023%; p = 0.002), which corresponds to roughly one additional NAION case per 7,000 semaglutide-treated patients each year. The authors note this sits close to how regulators have framed the issue: the European Medicines Agency and the World Health Organization classify NAION as a "very rare" side effect of semaglutide, defined as up to 1 in 10,000 users per year ^[1].

The paper compares its results with earlier syntheses. It says its findings line up with other meta-analyzes of observational studies, while meta-analyzes of randomized controlled trials "consistently reported fewer NAION events" ^[1].

The authors grade the certainty of the evidence as low to very low and assess the included studies as having moderate to serious risk of bias. Because every included study was retrospective and registry-based, NAION cases were identified using diagnosis codes, which the authors say could introduce misclassification and confounding by indication — meaning underlying differences between patients who received semaglutide and those who did not could shape the results. The paper says more high-quality prospective studies using standardized NAION diagnostic criteria are needed ^[1].

Why it matters for patients

NAION was already a concern for people taking semaglutide, and this analysis tries to put a number on how often it occurs. Both framings point the same way: the relative risk roughly doubles, but the condition remains rare in absolute terms at about one extra case per 7,000 people per year ^[1].

Several things are not established. The findings come from people with type 2 diabetes in these registries; whether the same estimate applies to people taking semaglutide for weight loss has not been reported in this analysis. The paper also does not report whether risk varied by dose, by brand, or by how the drug was taken, and the mechanism behind any possible link is not established. The authors write that clinicians should be aware of the potential adverse event, particularly for people at increased baseline risk of optic neuropathies, and say the results support current recommendations to discontinue semaglutide in patients diagnosed with NAION ^[1].

The publication includes competing-interest disclosures: two authors report lecture honoraria from Novo Nordisk, which makes semaglutide, with one also reporting a consulting fee from the company, and other authors report honoraria from additional drugmakers. Funding came from Polish science and education programs, and the authors state the funders had no role in the study ^[1].