The picture, before it has a name, is a sudden shadow or a loss of vision in one eye. The records used to confirm it describe the loss as painless, and as accompanied, in the acute phase, by edema of the optic nerve head. The injury is called nonarteritic anterior ischemic optic neuropathy, an ischemic injury at the front of the optic nerve, distinct from the arteritic form that follows inflamed arteries. Whether a peptide can produce that picture is a narrower question than the one asked about peptides as a class. The studies that measure it measure semaglutide, the glucagon-like peptide-1 receptor agonist in Ozempic and Wegovy. In a matched cohort published in JAMA Ophthalmology in 2024, Jimena Tatiana Hathaway and colleagues found that, among people already seen by neuro-ophthalmologists at one academic center, a semaglutide prescription was associated with a higher risk than prescriptions for other drugs used for the same purpose: a hazard ratio of 4.28, with a 95 percent confidence interval from 1.62 to 11.29, in type 2 diabetes, and a hazard ratio of 7.64, from 2.21 to 26.36, in overweight or obesity. They called the result an association, and wrote that future study would be required before causality could be assessed. Most research peptides have no eye data of this kind.
A ratio of 4.28 whose lower bound is 1.62 fits a modest increase and a large one, and the weight comparison, 20 events against 3, runs from just over a doubling to a hazard more than twenty-six times the comparison group. A hazard ratio from a service that already gathers this injury is not a rate among everyone who is prescribed the drug. Semaglutide is a peptide, and it is the compound these cohorts contain. Whether peptides can damage kidneys is a separate question, with its own trials, and it does not settle this one.
A hazard ratio from a service that already gathers this injury is not a rate among everyone who is prescribed the drug.
The association was counted in a clinic that already gathers this injury
Hathaway and colleagues used the neuro-ophthalmology registry at Massachusetts Eye and Ear, from December 1, 2017, through November 30, 2023: 16,827 patients with no prior history of the injury. Among them, 710 had type 2 diabetes (median age 59), 194 prescribed semaglutide and 516 a diabetes medicine that was not a GLP-1 receptor agonist, and 979 were overweight or obese (median age 47), 361 prescribed semaglutide and 618 another weight-loss medicine. A hand review had to confirm painless vision loss and optic-nerve-head edema in the acute phase.
The hazard ratios rest on the matched comparisons, which is where the imbalance matters. In the unmatched diabetes cohort, obesity was recorded in 81 percent of the semaglutide patients and 34 percent of the others, so an unmatched count would have mixed the drug with a condition it is used to treat. After propensity-score matching, the diabetes analysis held 17 events on semaglutide against 6 on the other diabetes drugs: a 36-month cumulative incidence of 8.9 percent (4.5 to 13.1 percent) against 1.8 percent (0 to 3.5 percent). In overweight or obesity the matched events were 20 against 3, and the cumulative incidence 6.7 percent (3.6 to 9.7 percent) against 0.8 percent (0 to 1.8 percent). The fall was steepest in the first year, when cumulative incidence on semaglutide reached 6.5 percent in diabetes (2.7 to 10.2 percent) and 5.5 percent in overweight or obesity (2.7 to 8.3 percent). A stricter match gave hazard ratios of 4.35 (1.37 to 13.81) and 7.28 (1.59 to 33.34).
The authors place those figures in a service that sees a large share of the region's cases, and they say the findings may not generalize. They give its population incidence as 2 to 10 cases per 100,000 persons, and they call it a significant cause of blindness among adults. They could not show that every prescription was taken, only that a dose had been dispensed when the injury occurred. Of the 16,827 patients, 5.7 percent were recorded as Black or African American, against 22.5 percent of people who identified that way in the greater Boston area in 2022; the authors note that Black individuals generally have a lower risk, and they ask that the result be read with caution beyond that clinic. The paper states that it does not inform a mechanism and could not establish cause. Hand review also showed why a larger coded study would be hard: 40 percent of records coded as ischemic optic neuropathy were not this injury, but arteritic disease from giant cell arteritis or some other optic neuropathy.
Later counts were smaller, and a regulator called the effect very rare
Cindy X. Cai and colleagues, writing in JAMA Ophthalmology in 2025, studied 37.1 million adults with type 2 diabetes across 14 databases, including 810,390 new semaglutide users, from December 1, 2017, through December 31, 2023. No code means this injury and nothing else, so they used the broader code for ischemic optic neuropathy under two rules: once, the sensitive definition, and twice, the specific. Incidence among new semaglutide users was 14.5 per 100,000 person-years on the sensitive definition and 8.7 on the specific. They set those rates against Hathaway's cumulative incidences of 8.9 percent and 1.8 percent, and they assign the gap to setting — a referral clinic in one study, national records in the other.
The association shrank, and it depended on the comparison. On the sensitive definition, semaglutide was not different from empagliflozin (hazard ratio 1.44; 0.78 to 2.68), sitagliptin (1.30; 0.56 to 3.01), glipizide (1.23; 0.66 to 2.28), or dulaglutide (0.93; 0.46 to 1.91). The specific definition was higher against empagliflozin (2.27; 1.16 to 4.46) and still not different from dulaglutide (1.28; 0.57 to 2.88). A self-controlled series, comparing exposed time with unexposed time in the same people, did find an increase: an incidence rate ratio of 1.32 (1.14 to 1.54) on the sensitive definition and 1.50 (1.26 to 1.79) on the specific one. Exenatide showed an increase only on the specific definition (1.62; 1.02 to 2.58). Cai and colleagues call the semaglutide result a modest increase, smaller than previously reported, with no known mechanism and no settled cause.
In June 2025 the European Medicines Agency's safety committee finished a review of Ozempic, Rybelsus, and Wegovy. After nonclinical studies, clinical trials, postmarketing surveillance, and the literature, the committee concluded that the injury is a very rare side effect of semaglutide, a frequency the announcement defines as one that may affect up to 1 in 10,000 people taking the medicine. Several large epidemiological studies, it wrote, suggest an approximately twofold increase in adults with type 2 diabetes compared with people not taking semaglutide, corresponding to approximately one additional case per 10,000 person-years of treatment. The announcement states that sudden loss of vision, or rapidly worsening eyesight, during treatment should lead the patient to contact a doctor without delay, and that treatment should be stopped if the injury is confirmed.
On the Food and Drug Administration's table of potential signals for October through December 2024, GLP-1 receptor agonists — Ozempic, Wegovy, and Rybelsus among them, together with other approved drugs in the class — stand beside non-arteritic anterior ischemic optic neuropathy. The note, as of March 6, 2025, is that the agency is evaluating the need for regulatory action.
The label's eye warning describes a different injury
The Ozempic prescribing information revised in May 2026 does address the eye, and the warning is diabetic retinopathy, a retinal complication of diabetes, rather than an ischemic injury of the optic nerve. In a two-year trial of patients with type 2 diabetes and high cardiovascular risk, diabetic retinopathy complications occurred in 3.0 percent on Ozempic and 1.8 percent on placebo. The absolute increase was larger among patients who already had diabetic retinopathy, 8.2 percent against 5.2 percent, than among patients without a known history, 0.7 percent against 0.4 percent. The label states that rapid improvement in glucose control has been associated with a temporary worsening of diabetic retinopathy, that the effect of long-term glycemic control with semaglutide on these complications has not been studied, and that patients with a history of the disease should be monitored for progression. Prescribers are told to inform patients to contact their physician if vision changes during treatment. In the medication guide, blurred vision is listed among the signs of low blood sugar. The prescribing information does not name this optic-nerve injury.
Read together, the documents describe two injuries and one drug. Diabetic retinopathy complications are the labeled trial result. The sudden shadow or loss in one eye is the injury in the observational studies, uncommon at 2 to 10 cases per 100,000, and placed by the European committee at about one additional case per 10,000 person-years. The strong counts are counts of semaglutide. Research peptides outside that literature have no comparable eye data, and a result about this medicine is not a result about peptides as a class.