Three-Year Outcomes and Latent Class Trajectory Analysis of the Childhood Arthritis and Rheumatology Research Alliance Polyarticular JIA Consensus Treatment Plans Study.

OBJECTIVE To assess the impact of differences in the timing of initial biologic DMARD (bDMARD) therapy using the CARRA STOP-JIA polyarticular JIA (pJIA) consensus treatment plans (CTPs) study on outcomes through 3 years. METHODS STOP-JIA participants with CARRA Registry follow-up through 3 years were eligible. Outcomes included ACR clinically inactive disease (CID) off glucocorticoids, clinical Juvenile Arthritis Disease Activity Score (10 joints) inactive disease (cJADAS-10-ID: cJADAS-10≤2.5), clinical remission on medication (CRM), and proportion of time in CID and/or cJADAS-10-ID. Latent class trajectory modeling (LCTM) was applied to identify participant subgroups sharing similar disease courses over this period. RESULTS 297 participants were included (n=190 Step Up [SU], 76 Early Combination [EC], 31 Biologic First [BF]). At the 3-year visit, CID was achieved by 35%, 42% and 44%, respectively (p=0.35 for SU versus EC). EC was superior to SU in achieving CRM (59.9% vs. 40.6%; p=0.012), time spent in CID (38% versus 30%; p=0.04) and cJADAS-10-ID (51% versus 41%; p=0.02). Slow, moderate, and rapid improvement trajectories were identified by LCTM; bDMARD initiation within the first (OR 5.33) or second (OR 2.67) month from baseline was associated with the rapid improvement trajectory. CONCLUSIONS EC was superior to SU across 3 years for outcomes reflecting time spent in lower disease activity. Starting bDMARD within 2 months predicted rapid improvement and maintenance of inactive disease. These results support the hypothesis that early bDMARD initiation reduces overall disease burden in pJIA.