Comparing the risk of developing uveitis in patients initiating anti-tumor necrosis factor therapy for ankylosing spondylitis: an analysis of a large US claims database

Abstract

Objective:To compare the risk of developing uveitis in patients initiating anti-tumor necrosis factor (anti-TNF) agents (adalimumab, etanercept, and infliximab) for ankylosing spondylitis (AS).Methods:Anti-TNF-naive patients with a diagnosis of AS and without a history of uveitis (N = 2115) who subsequently initiated anti-TNF therapy for AS were identified in a large claims database (2005 to 2011). A multivariate Cox proportional-hazards model was used to compare the risk of uveitis in patients who received etanercept or infliximab vs adalimumab.Results:The median number of days to the first occurrence of uveitis after initiation of anti-TNF was 191. Among the three anti-TNF groups, the median time to event of uveitis was longest in patients taking adalimumab (243 days), followed by etanercept (182 days) and infliximab (144 days). The incidence rate for uveitis over 1 year was lowest for patients who received adalimumab (2.4%, N = 717), highest for patients who received etanercept (4.5%, N = 1087), and intermediate for patients who received infliximab (3.2%, N = 311). The risk of uveitis was 1.9 times higher in patients receiving etanercept compared with those taking adalimumab (hazard ratio [HR]: 1.91, 95% confidence interval [CI]: 1.1 to 3.31). For patients taking infliximab, the risk of uveitis was not statistically significantly different (HR: 1.35, 95% CI: 0.62 to 2.95) compared to adalimumab.Conclusion:The results indicated that initial adalimumab therapy is associated with a significantly lower risk of developing uveitis compared to initial etanercept therapy in patients diagnosed with AS and no prior history of uveitis; however, the risk was not different between adalimumab and infliximab. Limitations to consider when interpreting this conclusion include that disease-level clinical data, such as disease duration, were not available for inclusion in the model and that risk of uveitis beyond 1 year was not evaluated.