The obesity drug semaglutide (Wegovy) conferred a significantly higher risk of ischaemic optic neuropathy (ION) than any other GLP-1 agonist, including semaglutide for diabetes (Ozempic), an analysis of US Food & Drug Administration (FDA) data has showed.
This comes as the drugs’ manufacturer was last week hauled over the coals for not reporting some other serious side effects, including deaths, linked to its products.
MedPage Today reports that in a review of 31 774 semaglutide-related reports in the FDA Adverse Event Reporting System (FAERS), researchers found that Wegovy was associated with a nearly fivefold higher risk of ION versus Ozempic, with an adjusted odds ratio (aOR) of 4.74 (95% CI 2.54-8.77).
Men had a threefold higher ION risk versus women (aOR 3.33, 95% CI 1.89-5.88).
Given the higher maximum recommended dose for Wegovy (2.4 mg vs 2 mg for Ozempic), the findings suggest a dose-dependent safety issue that should be evaluated prospectively, reported Edward Margolin, MD, of the University of Toronto and colleagues in the British Journal of Ophthalmology.
“Semaglutide, in any formulation, was the only agent significantly associated with ION,” the authors noted. “These findings extend our prior global analysis and, whereas previous studies identified only an agent-specific association, this study provides the first evidence of a formulation- and dose-dependent ION risk, with the strongest association observed for Wegovy.”
The reporting odds ratio (ROR) for any semaglutide product was 21.4 (95% CI 17.40-26.65), was 18.8 (95% CI 14.11-25.07) for Ozempic, and reached 74.9 (95% CI 51.79-108.29) with Wegovy. A sex-stratified analysis showed the highest ION signal for Wegovy in men (ROR 116.37) and for Ozempic in women (ROR 26.86).
FDA labelling for semaglutide products does not list ION as a risk, but the European Medicines Agency has indicated that non-arteritic anterior ischaemic optic neuropathy (NAION) is a very rare side effect of the GLP-1 receptor agonist.
Limitations, key message
Interest in the association between semaglutide and ION has increased steadily since the earliest reports, said Rudrani Banik, MD, of the Icahn School of Medicine at Mount Sinai in New York City. The broad scale of the database and the large number of ION cases linked to semaglutide are particular strengths of the current analysis as compared with previous reports.
Nonetheless, owing to limitations of the FAERS database, the authors could not account for a number of potential confounding factors, beyond the dosage difference.
“The indications are different for the drugs, Ozempic for diabetes and Wegovy for weight loss and also cardiovascular health,” Banik, who was not involved in the study, told MedPage Today.
“There may be different patient populations on each of these drugs. If a patient has diabetes, what is the level of diabetic control? If the patient has comorbid sleep apnoea, which is a risk factor for ION, how might that have influenced the risk? There are a lot of gaps that are not answered within this type of study.”
The anatomy of the optic nerve can also influence the risk of ION, she continued, noting that a small cup-to-disc ratio is associated with increased risk. Nocturnal hypotension is another risk factor. Obesity per se has not been linked to ION but often is associated with sleep apnoea, so a potential impact cannot be ruled out.
Prospective studies are needed to provide more informative answers, said Banik. Studies to date have all been retrospective and observational in nature. A randomised, controlled clinical trial would be ideal but difficult to conduct, requiring a large number of patients to investigate an uncommon event.
In the meantime, the key message for ophthalmologists is awareness.
“I actually have patients ask me about this,” said Banik. “They have heard about the association and they might be considering going on one of these drugs, and they want to know about the risk (for ION). So, the public is aware, and ophthalmologists need to be prepared to offer guidance.”
ION and semaglutide
Vision recovery after ION, sometimes called an eye stroke or optic nerve stroke, is variable, she added. Overall, about 40% of patients have some degree of recovery of central visual acuity. However, peripheral vision loss may not improve.
“Even in the 40% of patients who have some spontaneous recovery, it never returns to 100% of where they were before,” said Banik.
Several recent studies have identified an association between semaglutide and ION, with severe vision loss in as many as 15% of cases and complete blindness in as many as 5%. The reports prompted an international safety review, Margolin and colleagues noted.
None of the earlier reports examined formulation- or sex-specific patterns of semaglutide-associated ION, limitations addressed in the current study.
The investigators retrieved 30 668,520 FAERS reports filed from December 2017 to December 2024.
The total included 31 774 reports involving semaglutide. The semaglutide reports were analysed by individual brand (Ozempic, Wegovy, Rybelsus, and generic semaglutide) and as a pooled category. Additionally, tirzepatide was analysed as a pooled category that included Mounjaro, Zepbound, and generic tirzepatide and by indication (Mounjaro for diabetes and Zepbound for obesity).
For comparison, the authors analysed ION cases for metformin, insulin, and generic tirzepatide.
The results showed 3 070 adverse events attributed to Wegovy and 20 608 attributed to Ozempic, a difference reflecting the time between FDA approval of the different formulations.
The data showed 28 cases of ION in patients on Wegovy, 40 in those on Ozempic, and 85 with any form of semaglutide.
No ION events were attributed to Rybelsus, and none of the comparator drugs had a clear signal for ION: tirzepatide (ROR 0.56), Mounjaro (ROR 1.02), Zepbound (0), metformin (ROR 1.35), and insulin (ROR 1.61).
The dual agonistic effect of tirzepatide (GLP-1/GIP) may explain the lack of association with ION, the authors noted. GIP agonism may buffer GLP-1-mediated fluid shifts and stabilise vascular tone to reduce ischemic vulnerability.
“Despite greater HbA1c and weight reductions, tirzepatide showed no signal for ION, suggesting that the risk of NAION may be GLP-1-specific rather than purely metabolic,” Margolin and co-authors stated. “Its dual receptor activity, gradual titration, and broad dosing range (5-15 mg) produce more moderate GLP-1 stimulation, potentially limiting gastrointestinal fluid loss and orthostatic hypotension.
“By blunting haemodynamic shifts, these features may preserve vascular tone and autonomic stability, reducing ischaemic vulnerability, and may further lower NAION risk indirectly by improving sleep apnoea and oxygenation through fat loss,” the researchers added.
More problems, and warning letter
Meanwhile, last week the FDA issued a warning letter to Novo Nordisk for failing to report and investigate potentially serious side effects associated with its GLP-1 medications.
The agency cited the company for “serious violations” of post-marketing adverse drug experience (PADE) reporting requirements going back to an inspection in early 2025, saying these failures suggest “systemic failures” that compromise the FDA’s ability to monitor drug safety.
The letter detailed three deaths of patients taking semaglutide (Ozempic, Wegovy) that weren’t properly investigated or reported to the FDA. In one case, a physician reported that a patient taking semaglutide was depressed and committed suicide, but the company failed to document any attempt to investigate the case.
In another instance involving a patient death on semaglutide, the company closed the case without investigation because it claimed it had not obtained consent from the reporter.
However, obtaining consent isn’t a requirement for a PADE investigation, the FDA noted. According to the agency’s rules, all serious and unexpected adverse events must be reported within 15 days, regardless of whether the company believes they were drug-related or not.
Another case involved a report of suicidal ideation in a patient taking semaglutide that sat in medical review for two months, and was only submitted to the agency after FDA inspectors discovered it.
The investigation also turned up a case of a patient left disabled by a stroke while on liraglutide (Victoza, Saxenda). This case went unreported because the consumer didn’t think the stroke was drug-related.
“Patient safety is a top priority for Novo Nordisk,” the drugmaker stated in a response letter. “Since the inspection in early 2025, Novo Nordisk has been working to resolve the observations in the Form FDA 483 through its corrective and preventative action plan, and has kept the FDA informed on its progress, including through an initial response and seven updates.”
The investigation only looked at cases associated with products with active ingredients including semaglutide, liraglutide, nedosiran sodium (Rivfloza), and estradiol. But the FDA hasn’t ruled out that other products in the company’s portfolio could have similar reporting issues.
Novo Nordisk has 15 business days to notify the FDA of actions it will take to prevent future violations.
See more from MedicalBrief archives:
Nordisk slams study linking rare blindness to semaglutide
US study finds serious side effects from weight-loss drugs
Lawsuits blaming GLP-1s for blindness gather momentum