ConexiantSemaglutide Linked to NAION Risk

ℹ️ Observational Association Only

Researchers analyzed TriNetX data from approximately 3.35 million patients with diabetes, using propensity score matching to compare 174,584 semaglutide users with 174,584 non-GLP-1 RA users. NAION occurred in 45 semaglutide-treated patients versus 29 controls — absolute event rates of 0.026% and 0.017% — with a hazard ratio of 2.22 across the full study period.

Clinical Considerations

• Risk was not statistically significant at 1 year; elevated hazard emerged at 2 years and persisted through 4 years
• Elevated relative risk in subgroup analyses among patients aged 40 to 64, female patients, White patients, and those with hypertension — but small event counts and no multiple-comparison adjustment make these hypothesis-generating only
• Increased NAION risk was observed with injectable Ozempic; associations were not statistically significant for oral semaglutide or weight-management semaglutide prescriptions, though sparse event counts limit interpretation
• Residual confounding, ICD-10 outcome misclassification, and potential detection bias in more health-system-engaged semaglutide users are acknowledged limitations; causality cannot be established

Practice Applications

• Consider discussing NAION as a low-absolute-risk but elevated-relative-risk signal with patients on long-term injectable semaglutide, particularly those with existing optic nerve vulnerability
• Recognize that absolute event rates remain very low (0.026%) and that current evidence does not support treatment discontinuation based on this association alone
• Monitor for updated guidance from the AAO and retinal societies as this signal accrues further observational data
• Interpret subgroup findings cautiously pending larger, prospectively designed studies with NAION as a prespecified outcome