Risk Factors for Hydroxychloroquine Retinopathy and Its Subtypes

Abstract

Importance: The major toxic effect of hydroxychloroquine is retinopathy. Thus, current guidelines recommend limiting the dose and screening annually for retinopathy among all long-term users, but individual patient factors may be associated with retinopathy risk.

Objective: To identify risk factors beyond hydroxychloroquine dose and duration of use for hydroxychloroquine retinopathy.

Design, setting, and participants: This cohort study of 4677 patients in the Kaiser Permanente Northern California integrated health network who initiated hydroxychloroquine, continued treatment, and underwent retinopathy screening after 5 years of use was conducted from July 1, 1997, to December 31, 2020, with up to 15 years of follow-up. Statistical analysis was performed in August 2023.

Exposure: Candidate risk factors included age at hydroxychloroquine initiation, sex, race and ethnicity, indications, chronic kidney disease (CKD), liver disease, diabetes, tamoxifen use, and medications that interact with hydroxychloroquine metabolism. Hydroxychloroquine dose was assessed from pharmacy dispensing records.

Main outcome and measures: Incident hydroxychloroquine retinopathy was adjudicated from masked review of guideline-recommended screening studies and classified as parafoveal or pericentral pattern. Multivariable Cox proportional hazards regression was used to assess potential risk factors for hydroxychloroquine retinopathy within 15 years of initiation.

Results: Of 4677 long-term hydroxychloroquine users (mean [SD] age at initiation, 52.4 [14.1] years; 3877 women [82.9%]), 125 patients developed hydroxychloroquine retinopathy within 15 years (102 parafoveal, 23 pericentral). Older age at time of hydroxychloroquine initiation was associated with retinopathy risk, with adjusted hazard ratios (HRs) of 2.48 (95% CI, 1.28-4.78) for those aged 45 to 54 years, 3.82 (95% CI, 2.05-7.14) for those aged 55 to 64 years, and 5.68 (95% CI, 2.99-10.79) for those aged 65 years or older compared with those younger than 45 years. The risk of retinopathy was higher among females than males (HR, 3.83 [95% CI, 1.86-7.89]), among patients with CKD stage 3 or greater (HR, 1.95 [95% CI, 1.25-3.04]), and among individuals with tamoxifen use (HR, 3.43 [95% CI, 1.08-10.89]). The likelihood of pericentral retinopathy was higher among Asian patients (HR, 15.02 [95% CI, 4.82-46.87]) and Black patients (HR, 5.51 [95% CI, 1.22-24.97]) compared with non-Hispanic White patients.

Conclusions and relevance: This study suggests that increasing age, female sex, CKD stage 3 or greater, and tamoxifen use were associated with a higher risk of hydroxychloroquine retinopathy, whereas being younger than 45 years at hydroxychloroquine initiation and male sex were associated with a lower risk. Race and ethnicity were also associated with the pattern of retinopathy. These factors should be incorporated into hydroxychloroquine dosing decisions.

Figure 1.. Restricted Cubic Spline Curve of Estimated Glomerular Filtration Rate and the Hazard Ratio for Hydroxychloroquine Retinopathy

The curve was constructed using a multivariable Cox proportional hazards regression model, adjusting for age, sex, race and ethnicity, hydroxychloroquine weight-based dose, and cumulative hydroxychloroquine dose in first 5 years of use. Estimated glomerular filtration rate was assessed at 5 years of hydroxychloroquine use. Hydroxychloroquine retinopathy was assessed through 15 years of follow-up. The smoothed curve was fitted with restricted cubic splines with knots at 40, 60, and 90 mL/min/1.73 m² or more. The reference estimated glomerular filtration rate was 90 mg/min/1.73 m². The horizontal dashed line indicates a hazard ratio of 1.0. The shaded region indicates the bounds of 95% CIs for the restricted cubic spline curve.

Figure 2.. Weight-Based Hydroxychloroquine Dose and the Risk of Hydroxychloroquine Retinopathy After 15 Years of Use

The curve was constructed using a multivariable Cox proportional hazards regression model, adjusting for age, sex, race and ethnicity, indication for hydroxychloroquine use, and chronic kidney disease. Hydroxychloroquine dose was assessed at 5 years of hydroxychloroquine use. Hydroxychloroquine retinopathy was assessed through 15 years of follow-up. The smoothed curve was fitted with restricted cubic splines with knots at 4, 5, and 6 mg/kg/d. The reference weight-based dose was 5 mg/kg/d. The horizontal dashed line indicates a hazard ratio of 1.0. The shaded region indicates the bounds of 95% CIs for the restricted cubic spline curve.