When a study published in JAMA Ophthalmology reported that semaglutide users developed a rare form of sudden vision loss at roughly seven to eight times the rate of comparable patients on other therapies, it sent a jolt through the ophthalmology world. The condition, nonarteritic anterior ischemic optic neuropathy (NAION), cuts blood flow to the optic nerve and is sometimes called an “eye stroke.” It can cause permanent vision loss in the affected eye, and there is no proven treatment once it strikes.
Now, with more than 10 million Americans estimated to be taking semaglutide under the brand names Wegovy and Ozempic, a string of follow-up studies has sharpened the question without settling it. As of May 2026, no regulatory agency has added NAION to the drug’s warning label, and the evidence remains a patchwork of conflicting findings. Here is what the research actually shows, where it falls short, and what patients should watch for.
The study that raised the alarm
The initial signal came from a team led by Joseph Hathaway at Massachusetts Eye and Ear, a Harvard-affiliated neuro-ophthalmology clinic. Their retrospective cohort study, published by Hathaway et al. in JAMA Ophthalmology in 2024, covered patients seen between December 2017 and November 2023 and compared those prescribed semaglutide against matched patients on non-GLP-1 therapies. Among patients with overweight or obesity, the semaglutide group had a roughly 8.9-fold higher rate of NAION. Among those with type 2 diabetes, the rate was about 4.3 times higher.
Those relative risk numbers are striking, but context matters. NAION is rare in the general population, occurring in an estimated 2 to 10 people per 100,000 each year. The study identified only a small number of actual cases, and because it drew from a single specialty referral center, the patient pool skewed toward people already being evaluated for unusual eye problems. “These findings require replication in larger, more diverse populations,” Hathaway and colleagues wrote, cautioning against drawing causal conclusions from a single-center retrospective design.
A second team pointed to Wegovy specifically
Researchers at the University of Toronto took a different approach, mining the FDA’s Adverse Event Reporting System (FAERS) to compare signals across semaglutide formulations. Their disproportionality analysis, published in the British Journal of Ophthalmology, found a stronger reporting odds ratio (ROR) for Wegovy than for Ozempic, with further breakdowns by sex and dose level. The ROR is a standard pharmacovigilance metric that quantifies whether a specific drug-event combination appears more often than expected in a safety database; a higher ROR for Wegovy suggests the weight-loss formulation accounted for a disproportionate share of NAION reports relative to other drugs in the system.
That distinction matters because Wegovy is prescribed at higher doses for weight management (up to 2.4 mg weekly) than Ozempic is for type 2 diabetes (up to 2 mg weekly). If the association is real, a dose-dependent vascular effect could explain why the higher-dose formulation appears more frequently in adverse-event reports. No study has yet mapped a formal dose-response curve, so this remains a plausible hypothesis rather than an established fact.
FAERS data is publicly available and released in quarterly extracts, which allows independent researchers to replicate signal-detection work. But the database has a well-known limitation: reports are voluntary, and a drug generating heavy media attention will attract more filings regardless of whether the actual risk has changed. Reporting odds ratios can flag a signal worth investigating; they cannot, on their own, prove a drug causes harm.
A large replication effort found no elevated risk
The most statistically powerful counterpoint came from a study led by Jiyeon Choi and colleagues, who used a nationally representative electronic health record and insurance claims dataset covering approximately 66 million patients. The team ran multiple sensitivity analyses, restricting the sample to type 2 diabetes patients only, obesity patients only, and those with documented eye exams. They also compared semaglutide head-to-head against other GLP-1 receptor agonists. Across every analysis, the Choi et al. study did not find an elevated NAION risk for semaglutide users.
That result directly contradicts the Harvard-affiliated single-center findings. A subsequent systematic review has cataloged the divergence, noting that differences in study design, patient selection, and outcome coding likely explain much of the gap. Electronic health records carry their own blind spots: NAION diagnoses that are miscoded or never formally entered will not appear in claims data, potentially masking a real but small signal.
Why the question is so hard to answer
Several factors make this particular safety signal unusually difficult to pin down.
First, the patients most likely to take semaglutide, those with type 2 diabetes and obesity, already face elevated rates of vascular events affecting the optic nerve. Separating a drug effect from the underlying disease burden requires large, well-matched populations and long follow-up periods that no existing study has fully achieved.
Second, no one has identified a clear biological mechanism by which semaglutide would damage optic nerve blood supply. Without a plausible pathway, statistical associations remain vulnerable to alternative explanations. Those include unmeasured cardiovascular risk factors, obstructive sleep apnea, and the vascular stress of rapid weight loss, all of which are common in the same patient population.
Third, regulatory responses have lagged behind the published research. The current U.S. prescribing information for Wegovy, available through DailyMed, does not list NAION as an adverse reaction or warning. The European Medicines Agency’s Pharmacovigilance Risk Assessment Committee (PRAC) initiated a signal assessment in late 2024. As of early 2026, no formal label change has resulted from that review based on publicly available agency communications. That gap leaves clinicians without harmonized guidance from regulators, meaning discussions about this risk are being driven largely by specialty societies and individual experts.
What patients should know right now
For people currently taking semaglutide, the available evidence does not support stopping the medication solely out of concern about NAION. The condition is rare in absolute terms, and the number of reported cases remains small relative to the millions of active prescriptions. Semaglutide has demonstrated substantial benefits for blood sugar control, weight reduction, and cardiovascular outcomes in high-risk patients. Those benefits are documented in large randomized controlled trials, while the NAION signal rests on less definitive observational data.
That said, patients should know the warning signs of NAION: sudden, painless vision loss in one eye; a dark or blurry patch in the visual field; or a noticeable drop in vision upon waking. Anyone who experiences these symptoms should seek urgent evaluation, ideally from an ophthalmologist or neuro-ophthalmologist, and should mention all current medications, including GLP-1 receptor agonists.
For patients with pre-existing optic nerve conditions, a history of NAION in one eye, or anatomically crowded optic discs (a known structural risk factor), a more cautious conversation with their prescribing physician may be warranted. That conversation should weigh the proven metabolic and cardiovascular benefits of semaglutide against the small, uncertain possibility of serious eye harm. It may also include discussion of alternative weight-loss or diabetes therapies.
Where the research goes from here
Researchers across multiple institutions are now calling for prospective studies designed to answer the question more definitively. One priority is embedding standardized eye assessments into ongoing and future semaglutide clinical trials so that sudden vision changes are systematically captured rather than left to passive reporting. Another is harmonizing how NAION is defined and coded across pharmacovigilance and claims databases, which would make it possible to compare results more reliably across studies. Collaborative registries pooling cases from multiple neuro-ophthalmology centers could also help clarify whether specific patient characteristics or dosing patterns account for most of the reported cases.
Until those studies deliver results, the semaglutide and NAION question will remain open. The current body of research supports treating NAION as a potential but unproven safety signal: one that warrants awareness and monitoring, not alarm. For the millions of people relying on these medications, the most practical step is straightforward. Know the symptoms, report them quickly if they appear, and keep the conversation going with your doctor as new data emerge.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
