Nationwide health claims analysis of phenothiazine‐associated retinopathy risk in chlorpromazine and perphenazine users

Abstract

We conducted a nationwide, retrospective cohort study to quantify the risk of phenothiazine-associated retinopathy among new users of chlorpromazine (n = 35,955) and perphenazine (n = 224,790) in South Korea between January 1, 2015, and December 31, 2023. Adults without prior phenothiazine prescriptions or retinopathy who received at least one week of either drug were identified in the Health Insurance Review and Assessment database. We compared incidence rates of overall retinopathy, maculopathy, and pigmentary retinopathy/retinal degeneration before and after drug initiation, plotted cumulative incidence using Kaplan–Meier analyses, and used multivariable Cox regression to estimate hazard ratios (HRs) adjusted for age and sex. We also examined dose–response relationships across quartiles of cumulative exposure. By the end of follow-up, overall retinopathy and maculopathy occurred in 10.3% and 2.7% of chlorpromazine users and 19.5% and 6.2% of perphenazine users, respectively, with pigmentary retinopathy/retinal degeneration developed in 3.6% and 7.6%, respectively. Post- versus pre-exposure incidence rate ratios (IRRs) for overall retinopathy were 1.17 (95% CI 1.12–1.22) with chlorpromazine and 1.10 (95% CI 1.08–1.11) with perphenazine; IRRs for maculopathy were 1.18 (95% CI 1.08–1.29) and 1.11 (95% CI 1.08–1.14), indicating a modest increase in risk after initiation (all P < 0.05) Cumulative perphenazine exposure was associated with a slight increase in risk for pigmentary retinopathy/retinal degeneration (HR 1.06, 95% CI 1.02–1.11 in the highest quartile, P = 0.012), the highest chlorpromazine quartile showed reduced risks for overall retinopathy (HR 0.75, 95% CI 0.68–0.82), maculopathy (HR 0.53, 95% CI 0.44–0.63), and pigmentary retinopathy/retinal degeneration (HR 0.63, 95% CI 0.54–0.73; all P < 0.001). In multivariable models, perphenazine use (vs. chlorpromazine), older age, female sex, diabetes, hypertension, and hyperlipidemia were all significant predictors of the outcomes (all P < 0.05). These findings demonstrate that both drugs carry a modest risk of retinal toxicity—particularly perphenazine—and support the need for proactive ophthalmic screening in this population.