FAERS analysis finds semaglutide optic neuropathy signal far above tirzepatide's

A comparison of reports filed with the FDA Adverse Event Reporting System (FAERS) found a far stronger safety signal for optic ischemic neuropathy with semaglutide than with tirzepatide, and no signal for three other incretin-based drugs. The analysis was published June 8, 2026, in the journal Diabetes, Obesity and Metabolism ^[1].

Researchers led by Ashish Kumar Kakkar at the Postgraduate Institute of Medical Education and Research in Chandigarh, India, extracted reports from each drug's approval date through the third quarter of 2025 using the tool OpenVigil 2.1 ^[1]. They limited the analysis to reports naming the drug as the primary suspect and used the regulatory term "optic ischaemic neuropathy" to identify cases. Signals were measured with a reporting odds ratio (ROR) and a proportional reporting ratio (PRR), applying Evans' criteria ^[1].

Semaglutide, sold as Ozempic, Wegovy and Rybelsus, had 355 reports of optic ischemic neuropathy and an ROR of 94.45 (95% confidence interval 83.02 to 107.45), with a PRR of 93.77 (95% CI 82.48 to 106.61) ^[1]. Tirzepatide, sold as Mounjaro and Zepbound, had 49 reports and an ROR of 2.94 (95% CI 2.20 to 3.93), with the same figure for the PRR ^[1]. Liraglutide had 13 reports and an ROR of 4.58 (95% CI 2.65 to 7.91) ^[1].

The study found no signal for dulaglutide, exenatide or lixisenatide ^[1].

Among the semaglutide reports, the mean age was 59 years, disability was documented in 18% of cases and hospitalization in 11% ^[1]. The abstract reports those details only for semaglutide; characteristics of the liraglutide and dulaglutide reports appear in supplementary tables that accompany the paper ^[1].

The study notes that non-arteritic anterior ischemic optic neuropathy (NAION) had already become a safety concern with semaglutide, prompting a formal regulatory review and action by the European Medicines Agency ^[1]. Less was known about how the condition appears across the wider group of incretin-based therapies, which is what the new analysis set out to compare ^[1].

The authors concluded that the absence of a signal for several other GLP-1 receptor agonists argues against a uniform class effect ^[1]. They wrote that prospective studies with diagnoses confirmed by neuro-ophthalmologists are needed to establish causality and to quantify absolute risk ^[1].

Why it matters for patients

The numbers in this study come from spontaneous reports to a federal database, not from confirmed medical diagnoses, and a reporting odds ratio is not a measure of how likely a person is to develop a condition. It describes how often a drug appears in reports of that condition relative to other reports in the same database ^[1]. The authors state directly that their method does not establish cause and does not quantify absolute risk ^[1].

That distinction matters when reading the gap between semaglutide's ratio of 94.45 and tirzepatide's 2.94. The two drugs are not interchangeable products, and the analysis does not compare their users' underlying health, the doses involved, how long people took the drugs or how many people took each drug overall. None of those figures is reported in the paper ^[1].

The study also does not say whether any of the reported cases were confirmed by an eye specialist, and it offers no guidance on monitoring or on which symptoms might matter ^[1].

What happens next

The paper was first published on June 8, 2026, and covers reports through the third quarter of 2025 ^[1]. The authors call for prospective studies using neuro-ophthalmologist-confirmed diagnoses, but no such study has been reported, and no timetable for one has been disclosed ^[1]. Whether regulators will revisit labeling for any of the drugs named in the analysis has not been reported ^[1].