A prediction rule for polyarticular extension in oligoarticular-onset juvenile idiopathic arthritis

Abstract

To search for predictors of polyarticular extension in children with oligoarticular-onset juvenile idiopathic arthritis (JIA) and to develop a prediction model for an extended course. The clinical charts of consecutive patients with oligoarticular-onset JIA and ≥2 years of disease duration were reviewed. Predictor variables included demographic data, number and type of affected joints, presence of iridocyclitis, laboratory tests including antinuclear antibodies, and therapeutic interventions in the first 6 months. Joint examinations were evaluated to establish whether after the first 6 months of disease patients had persistent or extended course (i.e. involvement of 4 or less, or 5 or more joints). Statistics included univariable and multivariable analyses. Regression coefficients (β) of variables that entered the best-fitting logistic regression model were converted and summed to obtain a "prediction score" for an extended course. A total of 480 patients with a median disease duration of 7.4 years were included. 61.2% had persistent oligoarthritis, whereas 38.8% experienced polyarticular extension. On multivariable analysis, independent correlations with extended course were identified for the presence of ≥2 involved joints and a CRP >0.8 mg/dl in the first 6 months. The prediction score ranged from 0 to 6 and its cut-off that discriminated best between patients who had or did not have polyarticular extension was >1. Sensitivity and specificity were 59.6 and 79.8, respectively. The number of affected joints and the CRP level in the first 6 months were the strongest predictors of polyarticular extension in our children with oligoarticular-onset JIA.