Janus kinase inhibitor use and incident dry eye disease in rheumatoid arthritis: a real-world cohort study

Purpose: To examine the association between systemic Janus kinase (JAK) inhibitor use and the incidence of newly diagnosed dry eye disease (DED) in patients with rheumatoid arthritis (RA). Methods: This multi-institutional retrospective cohort study used de-identified electronic health records from the United States Collaborative Network. Adults aged 18 years or older with RA were categorized according to systemic JAK inhibitor use or non-use. A secondary active-comparator analysis compared JAK inhibitor users with rituximab-treated patients. Propensity score matching was performed to balance demographic characteristics, comorbidities, medication history, ocular conditions, and healthcare utilization. The primary outcome was newly diagnosed DED. Time-to-event analyses were conducted using Kaplan-Meier methods and Cox proportional hazards models. Results: A total of 25,149 JAK inhibitor users and 150,250 non-users with RA were identified. After matching, each cohort contained 23,343 patients with balanced baseline characteristics. JAK inhibitor use was associated with a lower incidence of newly diagnosed DED compared with non-use within 1 year of follow-up (hazard ratio [HR], 0.633; 95% CI, 0.545-0.736). In the secondary active-comparator analysis, 6,637 JAK inhibitor users were matched to 6,637 rituximab-treated patients, and JAK inhibitor use remained associated with a lower incidence of newly diagnosed DED (HR, 0.613; 95% CI, 0.476-0.791). Individual-agent analyses showed directionally consistent associations for tofacitinib and upadacitinib. Findings were consistent across sensitivity analyses using varying follow-up durations and landmark definitions. Conclusion: In this large real-world cohort study, JAK inhibitor use was associated with a lower incidence of newly diagnosed DED among patients with RA, with similar findings in a rituximab active-comparator analysis. However, the observational design, residual confounding, and coding-based outcome definition preclude causal inference. Prospective studies incorporating standardized ocular surface assessments and direct RA disease activity measures are needed to clarify the clinical relationship between systemic JAK inhibition and DED.