The FDA has ordered every marketed GLP-1 receptor agonist to carry a warning about a rare eye condition that can cause sudden, painless vision loss, usually in one eye.
The drug safety communication, issued Nov. 9, applies to 12 medicines, including semaglutide (Ozempic, Wegovy, Rybelsus), tirzepatide (Mounjaro, Zepbound), liraglutide (Victoza, Saxenda), dulaglutide (Trulicity), and the newly approved oral orforglipron (Foundayo). Each manufacturer must add non-arteritic anterior ischemic optic neuropathy, or NAION, to the warning section of its label.
The change is a warning, not a restriction. The agency said the benefits of the drugs "continue to outweigh their risks when used as directed," and it did not remove or narrow any approved indication.
What NAION is
NAION is an injury to the optic nerve caused by reduced blood flow at the point where the nerve leaves the eye. It typically produces a sudden loss of vision in one eye, often noticed on waking, and it is painless. Some patients recover part of their sight; others do not. Both eyes can be affected over time, though the first episode is usually unilateral. No treatment has been proven to reverse the damage.
The condition is rare, but its risk factors are common in the population that uses GLP-1 drugs: diabetes, high blood pressure, sleep apnea, high cholesterol, and certain blood disorders. That overlap is one reason the evidence has been difficult to interpret. A patient taking a GLP-1 drug for weight loss may already have several conditions that raise the risk of NAION on their own.
The presentation is what makes the injury easy to miss and easy to mistake for something else. A patient may notice a shadow or a missing patch in one eye, or colors may look washed out. There is no pain, no redness, and often no obvious trigger. The optic disc is swollen in the acute phase, which an ophthalmologist can see, but a primary care visit may not distinguish it from a migraine or a retinal problem. The window for evaluation is short, which is why the label's first instruction is to seek urgent care rather than a routine referral.
The evidence is a small signal, not a large one
The FDA based the class-wide warning on a review of six data streams, including a meta-analysis of 126 randomized trials covering 134,253 patients and 288,879 person-years. The review found 15 probable NAION cases: 10 among patients taking GLP-1 drugs, five among those on placebo, and none among patients on other glucose-lowering medicines.
The incidence was 6.4 cases per 100,000 person-years on GLP-1 drugs against 3.8 on placebo or other comparators. In a separate Sentinel cohort of adults with type 2 diabetes from 2019 to 2024, the rates were higher across the board: 30 to 32 per 100,000 person-years for GLP-1 drugs, 31 for SGLT2 inhibitors, and 26 for DPP-4 inhibitors. Compared with SGLT2 inhibitors, the hazard ratio was 1.19, with a confidence interval that crossed one and was not statistically significant. Compared with DPP-4 inhibitors, the hazard ratio was 1.43, with a confidence interval from 1.02 to 2.01.
The agency also weighed the 2024 JAMA Ophthalmology study that first raised the signal, as NeurologyLive summarized it. That cohort study reported much higher 36-month incidence figures among semaglutide users than among patients on other diabetes drugs, with hazard ratios of 4.28 in type 2 diabetes and 7.64 in people with overweight or obesity. The difference between those estimates and the FDA's trial-based rates is large. Population, follow-up, and how cases were identified can all change the number.
Clinical trials are not designed to catch events this rare. A review with 134,000 patients can estimate a rate of six per 100,000, but the confidence interval around that estimate is wide, and the 15 probable cases are the entire basis for the comparison. Real-world cohorts such as Sentinel can see more events because they follow larger populations for longer, but they bring their own problems: patients on GLP-1 drugs differ from those on other diabetes medicines, and doctors may look harder for NAION in patients on a new, widely discussed drug. The FDA weighed both kinds of evidence and concluded the signal was sufficient for a warning.
The FDA found no animal mechanism for the injury, and the adverse-event reports it reviewed were hard to interpret because so many patients had overlapping risk factors. It judged the overall signal sufficient for a class-wide warning rather than for a contraindication.
What the label does
The warning tells patients to seek urgent medical advice for sudden vision loss in one eye, visual-field gaps, altered color perception, or central blurring. It tells clinicians to arrange urgent ophthalmology review when NAION is suspected. And it says to consider discontinuing the GLP-1 drug in patients who develop NAION if the benefits are not clearly greater than the harms.
For patients who have already had NAION in one eye, the label advises a conversation with the prescriber about whether to continue, because the other eye is at higher risk. It does not tell anyone to stop treatment on their own.
The warning section is the FDA's standard place for a risk that requires attention but does not outweigh a drug's benefits. A contraindication would bar use in a group of patients. A boxed warning would put the risk at the top of the label. NAION went into neither category.
A class-wide warning also reaches products that are no longer actively promoted. Exenatide's brand was withdrawn, but generics remain on the market; combination products such as Xultophy and Soliqua are covered as well. Each manufacturer is responsible for updating its own label, and the FDA's communication is the mechanism that makes the change uniform across the class.
The absolute risk is small
The numbers are easy to misread. A 6.4 per 100,000 person-year rate is not a 6.4% risk. It is about six cases for every 100,000 people treated for a year, against about four in the comparison group. The difference is roughly two to three additional cases per 100,000 person-years.
In the Sentinel cohort, where rates were higher, the comparison between GLP-1 drugs and SGLT2 inhibitors was not statistically significant. The signal was stronger against DPP-4 inhibitors, a class that includes sitagliptin. That pattern is one reason some clinicians have described the evidence as suggestive rather than conclusive.
Dr. Zilbermint, an endocrinologist quoted in Healthline's coverage, said the overall risk is still very small, while adding that decisions have to be made patient by patient.
The market context
The warning covers a class that has become one of the largest in medicine. The 12 products include the leading obesity and diabetes brands from Novo Nordisk and Eli Lilly, along with older medicines whose patents have expired or whose use has declined. Several are among the most widely prescribed treatments in the country, and a label change that reaches all of them is a different kind of event from a warning on a single product. A class-wide labeling change applies to all of them, including generics and combination products.
Analysts at BMO described the change as limited risk to obesity coverage and said the FDA had formalized a previously recognized safety signal. The firm noted that the update went into the Warnings and Precautions section rather than restricting approved indications.
The agency's action follows years of case reports and observational studies. The FDA's review covered the whole class, not just semaglutide, which is why the label change reaches products that had never been the focus of the original case series.
What patients should do
The practical advice is straightforward. Anyone taking a GLP-1 drug who notices sudden vision loss, a new blind spot, or a change in color vision in one eye should seek medical attention immediately rather than waiting for a routine appointment. They should not stop the medicine on their own. A clinician can assess whether the symptoms fit NAION, whether another cause is more likely, and whether the drug should be paused.
The harder question is what to do after a diagnosis. NAION in one eye raises the risk of NAION in the other, and the label gives the prescriber discretion rather than a rule. For a patient whose diabetes or weight is well controlled on the drug, the benefit may still be worth the risk. For a patient with prior NAION and other options, it may not be.
For patients who develop NAION, the alternatives depend on why they were taking the drug. Someone using it for type 2 diabetes may be able to switch to an SGLT2 inhibitor, a DPP-4 inhibitor, or insulin. Someone using it for obesity has fewer options, which is why the label leaves the decision to the prescriber rather than imposing a rule.
The warning does not resolve that calculation. It makes the trade-off visible, and it tells doctors to look for the injury quickly enough to document it and to consider stopping the drug.
What the warning changes
The FDA's decision formalizes a signal that had been building in the literature since 2024. It does not establish that GLP-1 drugs cause NAION in every case, and it does not quantify the risk for any individual patient. What it does is put a rare but serious outcome on the label, where doctors and patients can see it, and where a report of sudden vision loss becomes a reason to act rather than a coincidence to explain away.
For most patients on these drugs, the calculation does not change. The absolute risk remains small, the benefits remain substantial, and the agency has said so. For the smaller group with prior NAION, overlapping risk factors, or symptoms that fit the condition, the label now provides a clear instruction: get the eye examined, and revisit the prescription with the clinician who wrote it.
The label is not a verdict on causation. It is a prompt, and for a condition with no proven treatment, prompt recognition is the main thing medicine can offer.
Primary sources
- US Food and Drug Administration, "FDA requires safety labeling changes for glucagon-like peptide-1 receptor agonists (GLP-1 RAs) due to potential increased risk of NAION," Drug Safety Communication, Nov. 9, 2026.
- BioSpace, "FDA orders class-wide GLP-1 label update to include potential risk for rare eye disease," Nov. 2026.
- NeurologyLive, "FDA Requires NAION Warning Across GLP-1 Receptor Agonist Class," Nov. 2026.
- Healthline, "Ozempic: FDA Orders GLP-1 Labeling Changes Due to Eye Disease Risk," Nov. 2026.