Predictors of response to etanercept in polyarticular-course juvenile idiopathic arthritis.

OBJECTIVE The aim of this study was to determine demographic, clinical, articular and laboratory variables predicting etanercept response in JIA. METHODS A total of 863 patients registered in the BIKER registry were included. Predictors of response for ACR paediatric criteria (PedACR) 30, 70 and 90 were analysed by bivariate analysis and logistic regression. RESULTS After 6 months, 81.9% of the included patients fulfilled the PedACR30, 55.2% the PedACR70 and 31.3% the PedACR90 criteria. In bivariate analyses, factors positively correlated with a PedACR70/90 response were extended oligoarthritis JIA category, lower age at the start of treatment, shorter disease duration until treatment start, no concomitant use of corticosteroids, duration of morning stiffness, parents' global assessment of overall well-being, lower Childhood Health Assessment Questionnaire (CHAQ) and higher ESR and ANA. Correlation between the JIA category systemic arthritis and response was significantly negative. Predictive factors for PedACR70 in multivariate logistic regression analysis were CHAQ [odds ratio (OR) = 0.70], ESR (OR = 1.02) and concomitant treatment with corticosteroids (OR = 0.68). Age at treatment start (OR = 0.94) and the systemic arthritis JIA category (OR = 0.28) were predictive for a PedACR70. We found no sufficiently distinctive models for PedACR30 or 90. CONCLUSION Parameters predicting a high-grade response to a 6-month course of etanercept were identifiable and included age at treatment start, disease duration before treatment, JIA category other than systemic arthritis, CHAQ, parents' global assessment and concomitant corticosteroids.