A note on method: this piece applies the Holistic United Discernment Method, our framework for justified judgment. It separates whether a risk exists from how large it is, asks what each comparison is made against, and aims for the strongest judgment the evidence permits, and no stronger. None of these checks is new: relative and absolute risk, the choice of comparator and confounding are long-standing tools of epidemiology and causal inference, which the Method aims to bring together rather than replace.
If you take one of these drugs, or someone in your family does, the FDA’s announcement on November 9 may have landed hard. Every GLP-1 drug in the United States will now carry a warning about a “potential increased risk” of an eye condition that can take away vision in one eye, sometimes permanently.
So what does that actually mean for the person at your kitchen table?
First, in our own words, not the FDA’s: do not start, stop or change any medication because of this article, or any headline, without talking to the clinician who prescribes it. This piece is about how to read the evidence, not about what anyone takes.
The condition is non-arteritic anterior ischemic optic neuropathy, or NAION. The FDA describes it as “a rare condition that can result in acute, painless, decreased vision in one eye, including permanent vision loss.” It is caused by reduced blood flow to the front of the optic nerve, “but the underlying cause is not completely understood.”
The FDA’s guidance for patients is specific. It lists the symptoms, “sudden, painless vision loss in one eye that patients may notice upon waking; blurred or missing portions of vision (visual defects such as blind spots or tunnel vision); distortion of colors; and reduced central sharpness,” and says: “Patients should contact their health care professional right away if they have any of these symptoms.” For people who have already had NAION, it says they “should talk with their health care professional about whether to continue taking their GLP-1 RA medication or to stop it.”
It does not tell everyone on these drugs to stop, or to carry on. The prescriber line above is ours.
Now the numbers. There are two ways to describe any risk, and they feel very different.
A relative figure says how much higher the rate is in one group than another: “twice the risk.” An absolute figure says how many extra cases there are among a given number of people over a given time. The Method treats these as separate questions. Evidence that something is a cause, it says, “is frequently insufficient to establish how much.” A risk can be real and small at the same time.
Here is what the FDA found when it pooled 126 clinical trials. Among people on GLP-1 drugs, NAION occurred at 6.4 cases per 100,000 person-years. Among people on comparison treatments, it was 3.8. A person-year is one person on the drug for one year.
Put plainly: for every 100,000 trial participants on these drugs for a year, about 6 cases, against about 4 among those on placebo or other treatments. That is roughly 2.6 extra cases per 100,000. In relative terms, about 1.7 times the rate. Both descriptions are true. Only one of them sounds alarming.
That trial figure is also thin. It rests on 10 cases against 5, and the FDA reports no confidence interval for it. In some large 2026 studies of people with type 2 diabetes, who start at higher risk, both rates are higher and so is the gap. Among US veterans with diabetes, one study (read at abstract level) found 123 against 67 cases per 100,000 person-years, about 56 extra. The European regulator, looking at semaglutide, put it at “approximately one additional case of NAION per 10,000 person-years of treatment,” and calls NAION a “very rare” side effect that “may affect up to 1 in 10,000 people taking semaglutide.”
So the added risk ranges from a few cases per 100,000 people a year to a few dozen, depending on who is being studied and how. For comparison, the FDA puts the background rate in adults over 50 at “2.3 to 10.2 cases per 100,000 persons per year.”
What about the “four to eight times the risk” figures? Those come from a 2024 study at one specialist eye clinic, whose own authors cautioned that their findings “may not be fully generalizable to other settings.” The patients had already been referred to a neuro-ophthalmologist, so those figures cannot be turned into a risk for everyone who takes the drug.
One more thing changes the answer: what the drug is compared against. In the FDA’s own real-world study of adults with diabetes, GLP-1 drugs showed “no statistically significant increase” in risk compared with one common class of diabetes drugs (SGLT-2 inhibitors), and an increase compared with another (DPP-4 inhibitors). Same data, different comparison, different headline. The Method’s question here is simple: relative to what?
There are two easy readings of all this, and each goes further than the evidence.
The alarmed reading says these drugs damage vision. It is right that the signal shows up in most large comparison studies and that regulators in the US and Europe acted on it. It goes too far when it treats causation as proven. The FDA calls the risk “potential,” and its review of animal data “did not identify a mechanism” for how the drugs could raise the risk.
The reassuring reading says the risk is tiny, so ignore it. It is right that NAION is rare and that some studies, including a large multinational one, found no clear increase. It goes too far when it leans on the small trial figure alone. Several large 2026 studies of people with diabetes found bigger gaps, including one that checked whether users were simply being examined more often and found they were not, and “rare” is not “zero.”
The Method asks that both readings meet the same standard. If one specialist clinic is too narrow a base for alarm, one trial pool with fifteen cases is too narrow a base for reassurance.
A few questions carry over to the next risk headline you see.
What is the drug being compared against?
How many extra cases, out of how many people, over how long?
Who was studied: specialist-clinic patients, trial volunteers or everyday users?
What else could explain the pattern, such as diabetes, sleep apnea or high blood pressure, all of which the FDA lists as risk factors?
And is the headline about how sure we can be, or about what to do? Those are different questions. The second depends on what a drug is doing for one particular person, and it belongs with that person and their clinician.
On sudden changes in vision, the FDA’s guidance is: “Patients should contact their health care professional right away if they have any of these symptoms.”
Every figure in this piece comes from the FDA’s notice and the studies listed below.
The paid paper works through the evidence in detail, including the comparison that flips the answer, the risk factors that travel with these drugs, the test for whether people were simply checked more often, and where the evidence stops.
References
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 Rare But Serious Eye Condition. Drug Safety Communication, November 9, 2026. https://www.fda.gov/drugs/drug-safety-communications/fda-requires-safety-labeling-changes-glucagon-peptide-1-receptor-agonists-glp-1-ras-due-potential
European Medicines Agency. PRAC concludes eye condition NAION is a very rare side effect of semaglutide medicines Ozempic, Rybelsus and Wegovy. June 6, 2025. https://www.ema.europa.eu/en/news/prac-concludes-eye-condition-naion-very-rare-side-effect-semaglutide-medicines-ozempic-rybelsus-wegovy
Hathaway JT, et al. Risk of nonarteritic anterior ischemic optic neuropathy in patients prescribed semaglutide. JAMA Ophthalmology. 2024;142(8):732-739. https://doi.org/10.1001/jamaophthalmol.2024.2296
Heberer K, et al. New-onset nonarteritic anterior ischemic optic neuropathy and initiators of semaglutide in US veterans with type 2 diabetes. JAMA Ophthalmology. 2026;144(3):259-264. https://doi.org/10.1001/jamaophthalmol.2025.6262 (abstract read)
Choi T, Al-Aly Z, Xie Y. GLP-1 receptor agonists or SGLT2 inhibitors and nonarteritic anterior ischemic optic neuropathy. JAMA Network Open. 2026;9(4):e269917. https://doi.org/10.1001/jamanetworkopen.2026.9917
Chou CC, et al. Association between semaglutide and nonarteritic anterior ischemic optic neuropathy: a multinational population-based study. Ophthalmology. 2025;132(4):381-388. https://doi.org/10.1016/j.ophtha.2024.10.030 (abstract read)



