On 10 November 2026, the US Food and Drug Administration ordered the makers of four medicines to add a warning to their labels.
The US Food and Drug Administration ordered the makers of four medicines — Ozempic, Wegovy, Mounjaro and Zepbound — to add a warning to their labels. The warning concerns a rare condition of the optic nerve that can leave a patient blind in one eye. The order changed no dose, no molecule and no approval. It changed what a patient is told to watch for.
A Warning That Travels by Leaflet
The instrument was a drug safety communication, the FDA’s public channel for announcing that a label must change. The condition is shortened to NAION — nonarteritic anterior ischemic optic neuropathy — a blocked-blood-supply injury to the optic nerve; ischemic means a shortage of blood, neuropathy means damage to a nerve. The drugs belong to a class called GLP-1 receptor agonists — drugs that mimic a gut hormone to lower blood sugar and reduce appetite — and are prescribed for type 2 diabetes and for obesity. Semaglutide — the active molecule in Ozempic and Wegovy — is a key ingredient in many GLP-1 drugs. None of that chemistry changes when the label does. The drug dispensed this week is the drug that was dispensed last month; only the paper around it is different.
The mechanism the warning describes is a blocked blood supply at the back of the eye. Vessels supplying blood to the optic nerves become blocked, and the nerves are deprived of oxygen. The symptom is painless decreased vision in one eye, and it can be permanent. Painless is the part that matters, because nothing hurts while it happens.
Rarity is a further obstacle. NAION occurs in somewhere between two and 10 people in every 100,000 of the general population, and those small numbers make it harder to establish with certainty whether a link to a drug exists. A handful of cases can be a coincidence or a real effect, and counting them does not decide which.
This is not the first line of small print these drugs carry. Their labels already warn about pancreatitis, an inflammation of the pancreas, and about gallbladder disease. The most common symptoms for semaglutide — nausea, vomiting and stomach pain — are listed as well, because frequency is part of what a label has to convey. That is how a rare event in the back of the eye comes to be printed beside a list of ordinary stomach upsets.
Where the
Regulator and the Makers Part The FDA did not act on a single report. Its conclusion followed a detailed review of reports of NAION in people using GLP-1 drugs and a review of other studies. The finding, in the agency’s own phrasing, is an “increased risk of rare NAION events” among people taking the drugs. [1] A label requirement is a regulatory act, not a proof of causation. The agency leaves the mechanism open: the FDA said it was not known why those taking GLP-1 drugs may be more at risk for the eye condition.
The methodological difficulty is stated plainly. Some of the research was hard to review because people taking these drugs for obesity and type 2 diabetes may already carry risk factors for NAION that are separate from the medicine. The list is long: diabetes, cardiovascular disease, high blood pressure, high cholesterol, metabolic syndrome, obstructive sleep apnea. Every one of those factors is common among the patients these drugs are prescribed to. That is the knot at the centre of the case. The agency concedes the remaining gap without decoration: it is not known why those taking GLP-1 drugs may be at higher risk.
Novo Nordisk, the manufacturer of Wegovy, does not accept the causal reading. Its statement is explicit that the data “do not establish a cause-and-effect relationship.” [1] The company then says this: “We recognize that NAION can have a profound impact on patients and their families. Patient safety is a top priority for Novo, and we continue to monitor the safety of our medicines.” The harm is acknowledged, the inference from the data is refused, and a promise to keep watching is entered. It is not a denial that the cases exist. It is a claim about what the cases can prove.
The American Academy of Ophthalmology, whose members treat this condition, occupies a third position. In its words: “any added risk appears small, and it has not been established that these medications cause NAION.” The academy sets that against the benefit it sees in the drugs: “GLP-1 medications offer important benefits for people with diabetes, obesity, and heart disease.” Its advice to patients is then as practical as medicine gets. “Don’t stop your medication without talking to your doctor, and if you notice sudden, painless vision loss in one eye, see an ophthalmologist promptly.” That instruction pushes the decision down to the individual case, where a prescription and a symptom meet, and it is the academy’s own phrasing, not the regulator’s.
What the Numbers Can and Cannot Say
The study that turned the question from anecdote into argument was published in 2024 in JAMA Ophthalmology by Harvard University researchers. It followed 710 patients with diabetes and 979 patients with overweight or obesity. Within each group, some took semaglutide and others took antidiabetic or weight-loss drugs that were not GLP-1 agonists. The comparison ran for 36 months. Among the patients with diabetes, 8.9 percent of those on semaglutide developed NAION, against 1.8 percent in the comparison group. Among those with overweight or obesity, 6.7 percent on a GLP-1 drug developed it, against 0.8 percent. Gaps that wide are hard to dismiss and hard to generalise.
One of the study’s authors is careful about what those numbers mean. Dr Joseph Rizzo, an ophthalmology professor at Harvard Medical School, described the findings as “significant but tentative” and said more research was needed on the relationship in a more diverse population. [1] The design explains the caution. All the patients came from a single referral centre that handled most of the region’s NAION cases. The incidence of NAION among the patients studied was far higher than the rate seen in the general population, which suggests possible selection bias — the group was not a random sample — and limits how far the results travel. The study did not account for how severe the patients’ diabetes was, and severity influences the risk. It also offered no hint of a mechanism, and researchers remain uncertain of one.
In the general population, NAION occurs in somewhere between two and 10 people in 100,000. Last year, a safety arm of the European Medicines Agency — the side-effect monitoring committee inside a medicines regulator — estimated that up to 1 in 10,000 people taking semaglutide may be affected. One in 10,000 is ten in 100,000. That is the top of the general-population range, not beyond it. The European estimate covered one ingredient rather than four brand names. None of the arithmetic decides the question of cause. All of it describes a risk small enough to be missed in ordinary practice and real enough to be printed.
There is no known treatment to reverse the vision loss NAION causes. According to the American Academy of Ophthalmology, it is the most common cause of acute optic nerve injury in people over 50. The new paragraph on the leaflet will change neither fact: the warning tells a patient what to watch for, not what will cure it.
Sources
Mentioned organisations (context, not sources)
- US Food and Drug Administration — Organisation (homepage)
- American Academy of Ophthalmology — Organisation (homepage)
- JAMA Ophthalmology — Organisation (homepage)
- Harvard University — Organisation (homepage)
- Harvard Medical School — Organisation (homepage)
- European Medicines Agency — Organisation (homepage)