On November 9, 2026, the US Food and Drug Administration ordered a new safety warning across one of the most commercially important drug classes in medicine.
The agency is requiring manufacturers of all GLP-1 receptor agonist medicines to add information about a potential increased risk of a rare eye condition called non-arteritic anterior ischemic optic neuropathy, or NAION, to the Warnings and Precautions section of their prescribing information.
The decision covers medicines including Ozempic, Wegovy, Mounjaro, Zepbound and Lilly’s newer oral obesity drug Foundayo.
NAION can cause sudden, painless vision loss in one eye. The damage can be permanent.
The FDA’s conclusion is carefully worded. The agency identified a potential increased risk, not definitive proof that these medicines cause the condition.
Its analysis of 126 randomized controlled trials included more than 134,000 people and identified just 15 probable NAION events. Ten occurred among GLP-1 recipients and five among people receiving placebo.
Those numbers are small. The difference between them is why the new warning exists.
For patients, the announcement raises a practical question about a rare but potentially serious injury. For drugmakers, it arrives as hundreds of lawsuits involving alleged GLP-1-associated vision loss move through federal court.
Understanding the warning requires separating three things that can easily become confused: an observed safety signal, evidence of causation and an individual’s actual risk.
What the FDA Found About GLP-1 Vision Loss
NAION occurs when blood flow to the anterior portion of the optic nerve becomes insufficient, damaging nerve tissue responsible for transmitting visual information from the eye to the brain.
It typically causes sudden, painless loss of vision in one eye. Some people notice the change when they wake up, while others experience missing areas of vision, blurred central vision or a disturbance in their visual field.
The condition is uncommon, including among people who do not use GLP-1 medicines.
The FDA cites background incidence estimates of 2.3 to 10.2 cases per 100,000 people annually among adults older than 50.
The agency’s investigation followed a July 2024 paper by Hathaway and colleagues in JAMA Ophthalmology that reported an association between semaglutide exposure and NAION.
The FDA identified limitations in that study but considered the finding sufficiently concerning to warrant a broader review.
Regulators examined spontaneous adverse-event reports, clinical trial datasets, animal experiments, observational studies and other available evidence.
The centerpiece was a pooled analysis of 126 randomized controlled trials involving multiple GLP-1 receptor agonists.
Across those studies, 134,253 participants contributed approximately 288,879 person-years of follow-up.
Of the participants, 78,217 received a GLP-1 receptor agonist and 56,036 received either placebo or another glucose-lowering medication.
Investigators identified 15 probable NAION events: 10 among GLP-1-treated patients, five among placebo recipients and none among participants receiving the other glucose-lowering medicines in these trials.
Because the groups contributed different amounts of follow-up time, the FDA also calculated incidence rates.
They were 6.4 probable NAION events per 100,000 person-years among GLP-1 recipients and 3.8 per 100,000 person-years among comparator recipients.
That is an observed difference of 2.6 events per 100,000 person-years in the pooled trial dataset.
A person-year represents one person followed for one year. It is a statistical measure of exposure time, not a guarantee that every individual faces the same annual risk.
The analysis also covered multiple drugs, study populations, doses and treatment durations. Its overall incidence rate should not be presented as a separately established risk estimate for Ozempic, Zepbound or any single product.
Ten events on GLP-1 drugs. Five on placebo. More than 134,000 trial participants. The signal is worth investigating, but the numbers do not establish causation.
Why the Evidence Does Not Point in One Direction
Fifteen events are a narrow foundation for estimating a rare adverse effect, even when they come from a large collection of randomized trials.
The FDA therefore examined real-world medical data through its Sentinel surveillance system.
That analysis followed adults with type 2 diabetes who began treatment with a GLP-1 receptor agonist, an SGLT-2 inhibitor or a DPP-4 inhibitor between 2019 and 2024.
The observed incidence rates were higher than in the randomized trials: approximately 30 to 32 cases per 100,000 person-years among GLP-1 users, 31 among SGLT-2 inhibitor users and 26 among DPP-4 inhibitor users.
These higher numbers came from a different patient population and research design. They should not be interpreted as proof that the real-world study measured a stronger drug effect.
The comparison with SGLT-2 inhibitors produced a hazard ratio of 1.19, with a 95% confidence interval of 0.91 to 1.54.
Because that interval includes 1.0, the analysis did not demonstrate a statistically significant increase against this comparator.
The comparison with DPP-4 inhibitors was different.
The hazard ratio was 1.43, with a 95% confidence interval of 1.02 to 2.01, indicating a statistically significant association in that comparison.
That inconsistency matters. A conclusion can change depending on which other medicine provides the comparison.
NAION also has established or suspected risk factors that overlap with the conditions commonly found in people prescribed GLP-1 drugs, including diabetes, hypertension, sleep apnea, smoking, kidney disease and anemia.
This creates a confounding problem. Patients using different medicines may also differ in ways that independently affect their likelihood of developing an optic nerve injury.
Statistical adjustment can reduce that problem, but observational comparisons cannot always eliminate it.
The FDA also reported that its review of available animal experiments did not identify a biological mechanism explaining how GLP-1 receptor activation might increase NAION risk.
That does not prove a mechanism cannot exist. It means that one has not been established by the animal evidence the agency reviewed.
Meanwhile, published observational studies considered methodologically rigorous by the FDA generally supported an association between GLP-1 initiation and NAION in people with type 2 diabetes.
Regulators therefore had several imperfect but relevant evidence streams pointing toward a possible safety problem.
The resulting decision was to require disclosure of the potential risk rather than wait for a definitive mechanistic explanation.
A statistically significant association in one comparison is not the same as proof that a medicine caused a particular patient’s vision loss.
Europe Warned About Semaglutide First. The FDA Went Further
The US decision did not emerge in isolation.
In June 2025, the European Medicines Agency’s Pharmacovigilance Risk Assessment Committee concluded that NAION was a very rare side effect of semaglutide-containing medicines.
That assessment concerned Ozempic, Rybelsus and Wegovy.
The European regulator estimated that NAION might affect up to one in 10,000 people taking semaglutide. It also cited epidemiological studies suggesting an approximately twofold increase in risk among adults with type 2 diabetes.
The EMA recommended adding the condition to semaglutide product information as a very rare adverse effect.
The US FDA has now applied a precautionary labeling requirement across the broader GLP-1 receptor agonist class.
Its November 9 communication lists 12 historically approved branded entries, including semaglutide products, liraglutide products, dulaglutide, insulin combinations, tirzepatide and the oral small molecule orforglipron.
One detail in that list is easy to miss: the original Byetta branded application was withdrawn for commercial reasons in 2025, although generic exenatide remained available.
The FDA says NAION events have been reported with all currently marketed GLP-1 receptor agonist products.
But applying the warning across the class does not establish that each individual medicine has the same risk, nor that each has been shown to cause NAION independently.
That distinction is especially relevant to Foundayo, the orforglipron pill approved in 2026.
Its inclusion reflects the FDA’s class-wide action. The announcement does not present a separate trial-based NAION risk estimate for orforglipron.
The same caution applies to tirzepatide, a medicine that activates both GLP-1 and GIP receptors.
Regulators can decide that a warning is appropriate across a pharmacological class even while the evidence for differences between individual products remains incomplete.
The FDA’s wording is also not identical to the EMA’s conclusion about semaglutide. The US agency describes a potential increased risk across the class, rather than classifying NAION as a confirmed adverse effect of every individual GLP-1 medicine.
What the Vision-Loss Warning Means for Patients
The FDA’s practical advice is more important than the headline debate over whether the agency has proved causation.
Patients taking GLP-1 receptor agonists should promptly report sudden, painless vision loss, missing portions of vision, blind spots, tunnel vision, distorted color perception or a new blurry central area.
These symptoms may be noticed upon waking, but they can occur at other times.
Sudden changes in vision require urgent medical assessment because NAION is not the only serious condition capable of producing them.
For healthcare professionals, the FDA recommends promptly referring suspected NAION cases to an ophthalmologist.
It advises clinicians to consider discontinuing the GLP-1 medicine after NAION develops unless the benefits of continued treatment clearly outweigh the risks.
Patients who have already experienced NAION should discuss continued treatment with their prescriber.
A previous episode in one eye matters because it increases the risk of the condition developing in the other eye.
The announcement does not recommend that every patient taking Ozempic, Wegovy, Mounjaro or another GLP-1 medicine stop treatment because of the warning.
For many patients, these drugs provide important medical benefits, including improved glycemic control, substantial weight reduction and reductions in serious cardiovascular or kidney outcomes in eligible populations.
Some medicines in the class also have approved indications for obstructive sleep apnea or metabolic dysfunction-associated steatohepatitis.
Those benefits have to be weighed against a rare and incompletely understood potential adverse effect.
The appropriate response is to discuss individual risk with a healthcare professional, not to make an abrupt medication change based solely on an alarming headline.
For someone experiencing new, sudden visual symptoms, however, the priority is urgent assessment rather than waiting for the next routine appointment.
The warning does not tell every patient to stop taking a GLP-1 drug. It tells patients and clinicians which visual symptoms must not be ignored.
The Warning Also Changes the GLP-1 Lawsuit Story
The FDA announcement arrives during a period of expanding product-liability litigation involving GLP-1 medicines.
Grey Pepper recently covered the growing federal GLP-1 litigation dockets, including the separate proceedings focused on gastrointestinal injuries and NAION-related vision loss.
The NAION litigation is consolidated as multidistrict litigation number 3163 in the Eastern District of Pennsylvania, before Judge Karen Spencer Marston.
The lawsuits raise questions about whether the medicines caused particular injuries, whether product warnings were adequate and whether manufacturers bear legal responsibility.
Those allegations have not been established merely because the cases were filed.
A new FDA labeling requirement is relevant to the broader debate about warnings, but it does not automatically settle earlier claims.
Courts must consider the applicable law, the evidence available at relevant times, the wording of earlier labels and the individual facts of each claim.
A warning required in November 2026 does not by itself establish what manufacturers legally should have warned about years earlier.
Nor does it establish that a specific person’s NAION was caused by semaglutide, tirzepatide or another GLP-1 medicine.
That distinction is central to the scientific and legal questions now being examined.
The manufacturers have also pushed back against conclusions of causation. Reuters reported on November 9 that Novo Nordisk maintained the available evidence does not establish a causal link, while Lilly said it was evaluating the FDA’s action.
The broader pharmaceutical industry has reasons to pay attention beyond the litigation.
Prescribing information influences discussions between doctors and patients, insurance coverage decisions, advertising and the assessment of a drug’s benefit-risk profile.
Still, it would be premature to claim that this warning will meaningfully reduce sales or determine the outcome of pending lawsuits.
The FDA has also demonstrated that warnings can move in the other direction.
In January 2026, it requested the removal of suicidal-behavior and ideation warning language from Saxenda, Wegovy and Zepbound after its review found no increased risk associated with GLP-1 receptor agonists.
That is a reminder that labeling is not a permanent verdict on every suspected adverse effect. It changes as regulators evaluate the available evidence.
A different kind of label debate continues around thyroid tumors. Grey Pepper recently examined why semaglutide carries a thyroid warning rooted in rodent findings, and why some researchers now question whether that warning remains justified.
The NAION decision illustrates another side of the same process: regulators may require stronger precautions when evidence raises concern, even while important causal questions remain open.
What the New FDA Warning Still Cannot Tell Us
Three scientific questions remain particularly important.
The first is mechanism.
Researchers still need to determine whether GLP-1 receptor activation can directly contribute to optic nerve ischemia, whether changes associated with treatment might affect susceptible patients or whether some observed associations arise from underlying differences between treatment populations.
The FDA’s review of animal studies did not identify a mechanism, leaving that part of the explanation unresolved.
The second is whether the risk differs by medicine, dose or duration.
The pooled trial analysis included multiple GLP-1 drugs, while the Sentinel analysis compared treatment classes in adults with type 2 diabetes.
Neither analysis, as presented in the November 9 communication, establishes whether higher doses carry more risk or whether risk changes substantially over longer treatment periods.
Nor does it establish identical risk across every drug included in the class-wide warning.
The third question is who is most susceptible.
Some NAION risk factors are already recognized, but their interaction with GLP-1 treatment has not been fully resolved.
The FDA did not publish a validated tool that could reliably predict which individual patient might develop the condition after starting treatment.
That matters because a low population-level incidence can conceal meaningful differences in susceptibility, while a relative risk increase can sound much more alarming than the underlying absolute numbers justify.
Both problems become especially important when discussing medicines taken by millions of people.
For now, the most defensible interpretation is straightforward.
The FDA has identified enough evidence of a possible association to require a clear warning across the GLP-1 drug class.
The clinical trial signal involves very few events, real-world comparisons are not entirely consistent and the biological mechanism remains uncertain.
Those limitations do not make the concern meaningless. They define how carefully it should be communicated.
The most important unanswered question is not whether the warning sounds frightening. It is whether future research can establish causation, quantify risk for individual medicines and identify which patients are most vulnerable.
The FDA has settled the need for a warning. It has not settled why NAION occurs in some patients using GLP-1 medicines, or how much risk is attributable to the drugs themselves.
Sources and Further Reading
- US Food and Drug Administration — FDA Requires Safety Labeling Changes for GLP-1 Receptor Agonists Due to Potential Increased Risk of Rare But Serious Eye Condition. November 9, 2026.
- European Medicines Agency — PRAC Concludes Eye Condition NAION Is a Very Rare Side Effect of Semaglutide Medicines. June 2025.
- World Health Organization — The Use of Semaglutide Medicines and Risk of Non-Arteritic Anterior Ischemic Optic Neuropathy. June 27, 2025.
- Hathaway JT, et al. Risk of Nonarteritic Anterior Ischemic Optic Neuropathy in Patients Prescribed Semaglutide. JAMA Ophthalmology, 2024.
- Reuters — US FDA Orders New Eye Risk Warnings for Ozempic, Zepbound. November 9, 2026.
- US District Court, Eastern District of Pennsylvania — MDL 3163: GLP-1 Receptor Agonists NAION Products Liability Litigation.
- US Food and Drug Administration — FDA Requests Removal of Suicidal Behavior and Ideation Warning From GLP-1 Receptor Agonist Medications. January 13, 2026.
