News|Articles|November 13, 2026

FDA Requires NAION Warning Across GLP-1 Receptor Agonist Class

Listen

0:00 / 0:00

Key Takeaways

FDA requires all GLP-1 RAs, including semaglutide, tirzepatide, and orforglipron, to add NAION risk language to Warnings and Precautions, irrespective of diabetes or obesity indication.
NAION presents as sudden, painless, typically unilateral vision loss with frequent permanence; prior NAION increases contralateral-eye risk, prompting urgent reporting and ophthalmology referral.
Evidence base combines low absolute incidence in pooled trials with increased relative risk in Sentinel versus DPP-4 inhibitors; confounding persists given shared NAION risk factors.
Neurology development continues despite mixed efficacy signals, prominent gastrointestinal intolerance, and rare serious events; ongoing CNS trials will now scrutinize ophthalmologic adverse events more closely.

The FDA is requiring new labeling across all GLP-1 receptor agonists to warn of a potential increased risk of a rare, vision-threatening optic nerve condition.

The FDA is requiring every glucagon-like peptide-1 receptor agonist (GLP-1 RA) to carry a new label warning about a potential increased risk of nonarteritic anterior ischemic optic neuropathy (NAION), a rare but serious eye condition. The agency announced the requirement on November 9, 2026, directing all application holders to add the language to the Warnings and Precautions section of labeling.¹

The requirement is classwide, covering exenatide, liraglutide, dulaglutide, lixisenatide, semaglutide, tirzepatide, and the newly approved orforglipron (Foundayo), regardless of whether the product is marketed for diabetes or weight management.

NAION results from decreased blood flow to the optic disc, the front of the optic nerve, though the agency said the underlying cause is not completely understood. It typically presents as sudden, painless vision loss in one eye, often noticed upon waking, with blurred vision, blind spots, or color distortion. Vision loss is frequently permanent, and a prior episode raises the risk of a future one in the other eye.

FDA's conclusion draws on several data sources. A pooled meta-analysis of 126 GLP-1 RA trials, covering 134,253 patients and 288,879 person-years, identified 15 probable NAION events, 10 on a GLP-1 RA and 5 on a comparator, for an incidence of 6.4 versus 3.8 per 100,000 person-years. A separate Sentinel System cohort study found an increased risk relative to DPP-4 inhibitors (hazard ratio, 1.43; 95% CI, 1.02-2.01) but not relative to SGLT-2 inhibitors (hazard ratio, 1.19; 95% CI, 0.91-1.54).¹

That analysis followed a July 2024 cohort study led by Jimena Tatiana Hathaway, MD, MPH, of Massachusetts Eye and Ear and Harvard Medical School, which first raised the signal. Among patients with type 2 diabetes, semaglutide was associated with a 36-month NAION incidence of 8.9% versus 1.8% with non-GLP-1 RA antidiabetic drugs (hazard ratio, 4.28). Among overweight or obese patients, incidence was 6.7% with semaglutide versus 0.8% with non-GLP-1 RA weight-loss drugs (hazard ratio, 7.64).²

FDA said its review of animal data did not identify a biological mechanism, and that adverse event reports were hard to interpret because NAION risk factors, including diabetes, hypertension, and sleep apnea, are common in patients prescribed these drugs.¹

The agency is advising clinicians to have patients report sudden, painless vision loss or other visual disturbances in one eye right away and to refer suspected cases promptly to an ophthalmologist, and to consider discontinuing the GLP-1 RA in patients who develop NAION unless benefits clearly outweigh the risk. Patients with a prior NAION episode face an increased risk of recurrence and should discuss continued use with their prescriber.

A safety signal against a widening neurology pipeline

The warning lands as GLP-1 RAs are being tested across a growing list of neurologic diseases for their anti-inflammatory and neuroprotective potential. None of the trials below has reported a NAION signal, but each carries its own safety data that will now be read alongside FDA's warning.

In Alzheimer disease (AD), Novo Nordisk's phase 3 EVOKE and EVOKE+ trials tested oral semaglutide in 3808 adults with mild cognitive impairment or mild dementia due to AD. Neither trial met its primary endpoint of slowing decline on the Clinical Dementia Rating-Sum of Boxes scale, though semaglutide reduced high-sensitivity C-reactive protein and showed nominal reductions in several cerebrospinal fluid tau markers.

Full results at CTAD 2025 showed adverse events in 91.2% of semaglutide-treated patients versus 84.8% on placebo, mainly decreased appetite, nausea, diarrhea, and vomiting, with no visual or ophthalmologic events reported above the 10% threshold.³

In Parkinson disease, three trials have tested different GLP-1 RAs with mixed results. A phase 2 trial of lixisenatide in 156 patients found a statistically significant slowing of motor disability progression at 12 months, but nausea occurred in 46% of treated patients versus 12% on placebo, and one patient developed treatment-related pancreatitis, prompting outside experts to caution that the drug was “not ready for prime time.”⁴

A larger phase 3 trial of weekly exenatide in 194 patients found no benefit on motor, cognitive, or imaging measures over 96 weeks, with serious adverse events (AEs) in 9% of exenatide-treated patients versus 11% on placebo.⁵ A smaller Swedish trial found exenatide altered motor-network imaging patterns without a measurable clinical benefit. None of the three reported eye-related AEs.

GLP-1 RAs have also drawn interest in idiopathic intracranial hypertension (IIH), where elevated brain pressure can itself damage the optic nerve through papilledema, a mechanism distinct from NAION's vascular origin. A retrospective study of more than 44,000 patients with IIH found that GLP-1 RA users had lower rates of papilledema (2.2% versus 11.5%) and vision loss or disturbances (7.0% versus 11.7%) than nonusers, along with fewer surgical interventions and lower mortality.⁶ The senior author cautioned the finding is hypothesis-generating, not proof that GLP-1 RAs protect vision in IIH, and that a placebo-controlled trial is the logical next step.

Taken together, these trials show a consistent gastrointestinal tolerability burden, and in one case a serious pancreatitis event, but none has yet reported the optic nerve signal now being added to every GLP-1 RA label. As GLP-1 trials in neurologic disease continue to enroll, investigators and treating neurologists will be watching for ophthalmologic adverse events with new attention.