The FDA Adds an Optic Nerve Warning to All 12 GLP-1 Drug Labels and Stops Short of Europe's Advice

Every glucagon-like peptide-1 receptor agonist sold in the United States will carry a warning about non-arteritic anterior ischemic optic neuropathy (NAION), the Food and Drug Administration said in a Drug Safety Communication on Nov. 9.

The agency is requiring application holders to add the potential risk to the Warnings and Precautions section of the labels. The communication lists 12 products, from Byetta (exenatide), first approved in 2005, to Foundayo (orforglipron), approved this year, and includes the dual agonist tirzepatide in both Mounjaro and Zepbound. “NAION cases have been reported for all currently marketed GLP-1 RA products,” the FDA wrote. NAION reduces blood flow to the front of the optic nerve and can cause sudden, painless vision loss in one eye that may be permanent.

What evidence did the FDA weigh?

The FDA did not conclude that the drugs cause NAION. It described a “potential increased risk of rare NAION events” after reviewing adverse event reports, animal data, clinical trials and observational studies. The review was prompted by a July 2024 study by Hathaway and colleagues in JAMA Ophthalmology, which the FDA said had several limitations.

The agency found its adverse event reports hard to interpret, because NAION risk factors such as diabetes, hypertension and sleep apnea are common in the people who take these drugs. Animal studies identified no mechanism. Its meta-analysis of 126 randomized trials covered 134,253 patients and 288,879 person-years, and found 15 probable NAION events: 10 among patients on a GLP-1 drug, 5 on placebo and none on other glucose-lowering drugs. That is 6.4 cases per 100,000 person-years on GLP-1 drugs and 3.8 on comparators.

A retrospective cohort in the FDA's Sentinel system, covering adults with type 2 diabetes from 2019 to 2024, found NAION incidence of 26 to 32 per 100,000 person-years across groups. Compared with SGLT2 inhibitors, the hazard ratio for GLP-1 drugs was 1.19 (95% CI, 0.91 to 1.54), not statistically significant. Compared with DPP-4 inhibitors it was 1.43 (95% CI, 1.02 to 2.01). The FDA said the more rigorous published observational studies generally pointed the same way.

The absolute risk is small and the evidence is not causal, but the advice to clinicians is specific: ask about vision, refer quickly, and reconsider the drug if NAION occurs.

How do the FDA's numbers compare with Europe's?

The two regulators reached similar conclusions 521 days apart, by our count, with different scope and different language. The European Medicines Agency's safety committee concluded on June 6, 2025 that NAION is a “very rare side effect” of semaglutide, affecting “up to 1 in 10,000 people taking semaglutide,” and estimated about one extra case per 10,000 person-years of treatment.

Our side-by-side of the two decisions:

Scope. EMA: semaglutide only (Ozempic, Rybelsus, Wegovy). FDA: all 12 GLP-1 products in its table, including liraglutide, dulaglutide, tirzepatide, lixisenatide and orforglipron.

Frequency. EMA: “very rare,” with an estimated one extra case per 10,000 person-years, or 10 per 100,000. FDA: no frequency category in the communication; its trial meta-analysis implies an excess of 2.6 cases per 100,000 person-years (6.4 minus 3.8), or about one extra case per 38,000 person-years by our arithmetic.

Advice if NAION occurs. EMA: semaglutide treatment “should be stopped” if NAION is confirmed. FDA: “Consider discontinuing the GLP-1 RA in patients who develop NAION,” unless the benefits clearly outweigh the risk.

The trial-based excess is about a quarter of Europe's estimate. The observational data sit closer to it. As a rough illustration of ours, if a DPP-4 inhibitor group had NAION at the low end of Sentinel's range, 26 per 100,000 person-years, a hazard ratio of 1.43 would add about 11 cases per 100,000 person-years. The FDA put background incidence at 2.3 to 10.2 cases per 100,000 adults over 50 each year, and the type 2 diabetes cohort ran well above that, which is why the comparator matters.

What should prescribers tell patients?

The FDA asks clinicians to tell patients to report sudden painless vision loss, visual field defects or other visual disturbances in one eye right away, and to refer promptly to an ophthalmologist when NAION is suspected. Patients who have already had NAION are at increased risk of recurrence, and the agency lists smoking, diabetes, high blood pressure, sleep apnea, anemia and kidney disease as known risk factors. Its patient advice notes that vision loss is often noticed on waking.

The advice in the communication centers on symptoms, referral and the decision after a diagnosis. The change goes into Warnings and Precautions, not Contraindications, so it is a counseling point rather than a bar to prescribing.

What it means in clinic

Add one question to GLP-1 starts and refills: any prior episode of sudden vision loss in one eye. Document prior NAION before prescribing, since it raises the risk in the other eye, and make the counseling about one-eyed vision changes part of the same conversation as nausea and pancreatitis. For patients with diabetes, keep the comparison honest: the FDA's own cohort found no significant difference against SGLT2 inhibitors. Clinicians managing type 2 diabetes outside insulin therapy will meet this question most often. Label updates from each manufacturer are the next documents to watch, along with any adverse event case counts the agency publishes. For other recent eye-related approvals, see our report on satralizumab for thyroid eye disease.

Share this storyImages sized for Facebook, Instagram, TikTok, X and link previews

Download all imagesCopy link

This article is for professional education and does not replace clinical judgment. Treatment decisions should be based on the individual patient and current guidelines.