PURPOSE: Non-arteritic anterior ischemic optic neuropathy (NAION) is the most common acute optic neuropathy in adults older than 50 years, with contralateral eye involvement occurring in 15–25% of patients within 5 years. However, risk factors predicting contralateral involvement remain incompletely characterized in large, real-world populations. This retrospective cohort study used electronic health record (EHR) data to assess 25 unique clinical conditions, plus diagnostic subtypes and groupings of these risk factors by clinical category, that are potentially associated with development of contralateral NAION.

METHODS: Risk factors previously reported to be associated with initial NAION incidence were identified through literature review, with medication risk factor analysis requiring ≥3 instances of documentation for inclusion. Using TriNetX, separate cohorts of patients with isolated left-eye or right-eye NAION were assembled. Within each cohort, Cox proportional hazards models were used to estimate the effect of each risk factor on developing contralateral NAION within 1 or 5 years, controlling for age at index, sex, essential(primary) hypertension, diabetes mellitus, lipid disorders, sleep apnea, overweight/obesity, and disorders of the lens. Hazard ratios (HRs) were pooled across right- and left-eye cohorts using inverse-variance weighting and reported with 95% confidence intervals (CI).

RESULTS: In the pooled 1- and 5-year analyses, none of the evaluated risk factors demonstrated a statistically significant association with contralateral NAION. Risk factors showing nonsignificant trends toward increased risk and HR >1.5 included (HR; 95% CI): primary thrombophilia at 5 years (2.10; 0.28–15.95), amiodarone at 1 year (2.04; 0.69–5.99) and 5 years (1.68; 0.65–4.32), optic disc drusen at 1 year (1.65; 0.32–8.43) and 5 years (1.57; 0.39–6.42), and semaglutide at 1 year (1.57; 0.28–8.83).

CONCLUSIONS: Risk factors previously associated with first-eye NAION were not confirmed as predictors of contralateral involvement in this large EHR-based analysis. The low incidence of contralateral NAION in this dataset limited statistical power to detect associations. A dataset with higher incidence of contralateral NAION or evaluation of alternative or additional risk factors may be required to better characterize predictors of contralateral NAION.