The effect of extra-articular manifestations on tumor necrosis factor-α inhibitor treatment duration in patients with ankylosing spondylitis: nationwide data from the Korean College of Rheumatology BIOlogics (KOBIO) registry

Tumor necrosis factor-α inhibitor (TNFi) therapy has shown to be remarkably effective for treating ankylosing spondylitis (AS); however, nearly 30% of AS patients every year either stop TNFi therapy or switch to a different TNFi due to inefficacy or adverse effects. The goal of this study was to identify predictors of TNFi treatment duration, including extra-articular manifestations, using a nationwide registry in Korea. Data obtained from the Korean College of Rheumatology Biologics (KOBIO) registry, a nationwide, multi-center database representing 58 tertiary care hospitals in Korea. Demographics, clinical features, laboratory findings, disease activity indices (BASDAI, ASDAS-ESR, ASDAS-CRP), peripheral arthritis, and extra-articular manifestations (uveitis, enthesitis, dactylitis, psoriasis, and inflammatory bowel disease) were studied in patients with AS during TNFi therapy. We also analyzed treatment duration outcomes for five TNFi agents (etanercept, infliximab, infliximab biosimilar, adalimumab, and golimumab), as well as factors associated with treatment duration, particularly in terms of extra-articular manifestations. Univariable and multivariable Cox regression analyses were performed to verify preliminary results. A total 1482 AS patients starting TNFi drug therapy between Dec. 2012 and Jan. 2017 were included. No differences in demographics, disease activity, or extra-articular manifestations were evident between continued and discontinued TNFi groups at baseline, though baseline differences were detected for gender distribution, CRP, platelet counts, and HLA-B27 positivity. During treatment period, the effects of extra-articular manifestations, including uveitis (unadjusted hazard ratio [HR] 0.92, 95% confidence interval [CI] 0.57 to 1.48, p = 0.74), enthesitis, dactylitis, psoriasis, and inflammatory bowel disease, on TNFi treatment duration were not statistically significant. By contrast, the occurrence of peripheral arthritis was significantly associated with shorter TNFi treatment duration (unadjusted HR 2.21, 95% CI 1.66 to 2.95; adjusted HR 1.38, 95% CI 1.01 to 1.88). Among disease activity indices, higher ASDAS-ESR levels were significantly associated with shortening of the TNFi treatment duration (unadjusted HR 1.87, 95% CI 1.73 to 2.03; adjusted HR 2.23, 95% CI 2.00 to 2.63). Among TNFi drugs, golimumab had a lower discontinuation rate than that of etanercept over a 3-year follow-up period (unadjusted HR 0.46, 95% CI 0.31 to 0.68; adjusted HR 0.65, 95% CI 0.43 to 0.99). In a nationwide KOBIO registry, extra-articular manifestations, including uveitis, were not associated with TNFi treatment duration. Among clinical cofactors, the development of peripheral arthritis during TNFi therapy was associated with a higher risk of TNFi treatment discontinuance in AS patients.