Predictors of changes in disease activity among children with juvenile dermatomyositis enrolled in the Childhood Arthritis and Rheumatology Research Alliance (CARRA) Legacy Registry.

Determinants of changes in disease activity among patients with juvenile dermatomyositis (JDM) are unknown. Our objective was to develop predictive models to predict changes in disease activity using the CARRA Legacy Registry. The CARRA Legacy Registry included 658 subjects with definite or probably JDM with 297 subjects with a one follow-up visit after baseline, and we studied the 65 subjects with active disease at baseline. Linear regression models were used to build risk scores for changes in disease activity adjusted for baseline disease activity, age, sex, and disease duration. Disease activity improved from baseline to 6-month follow-up as measured by patient/parent global health score (median 4; p = 0.008), patient pain score (median 2; p = 0.014), physician global (median 4; p < 0.001), and Childhood Myositis Assessment Scale (CMAS) (median 41, p < 0.001). Anti-nuclear antibodies (p = 0.013) and hydroxychloroquine use (p = 0.045) were significant predictors of less improvement in patient/parent global and baseline patient/parent global. Anti-nuclear antibodies (p = 0.001) and V/shawl sign (p = 0.005) were significant predictors of less improvement in patient pain (R-square improved from 0.29 for adjustors alone to 0.46 for the full model). Small joint arthritis (p < 0.01) predicted less improvement and dysphagia/dysphonia (p = 0.033) predicted greater improvement in CMAS and baseline CMAS (R-square improved from 0.73 for adjustors alone to 0.86 for the full model). Disease characteristics can help identify patients who are less likely to improve over time. Risk scores to predict future changes in disease activity could be used to trigger more aggressive treatment earlier in the disease course.