Background/Purpose:Treatment with anti‐TNF therapies (anti‐TNF) for polyarticular forms (extended oligo, Poly RF +/−) of JIA (PF‐JIA) results in >50% demonstrating clinically inactive disease (CID). The aims of this study were to determine the frequency, timing and predictors of flare upon withdrawal of anti‐TNF in PF‐JIA in CID.Methods:In 16 centers 137 children with PF‐JIA in CID on anti‐TNF were enrolled and followed for ≥14 mos. If CID was maintained for the first 6 study mos, then anti‐TNF was stopped. The primary outcome variable was a validated definition of disease flare within 8 months after stopping anti‐TNF. Background meds were stable. Blood for S100, DEK, DNA and RNA was drawn for current and future biomarker and genetic studies.Results:The study population included 18 (13%) extended oligarticular, 17 (12%) RF+ Poly and 102 (74%) RF‐ Poly JIA patients. At enrollment, age (mean/median/range) was 11.3/11.6/3.4–20.1 yrs; disease duration was 5.0/4.1/0.6–18.6 years; 103 (75%) were females and 64 (47%) were ANA+. Duration of CID at baseline was 1.2/0.5/1 day–12.1 yrs. Anti‐TNF was etanercept in 106 (77%), 25 (18%) adalimumab and 6 (5%) infliximab. 40% were on MTX at baseline (mean/median dose 0.4/0.4 mg/kg/week). Other meds: 1 leflunomide, 2 hydroxychloroquine, and 1 prednisolone.31 (23%) subjects were discontinued from the study in the first 6 mos: 23 (17%) due to loss of CID, 5 (4%) med noncompliance, 2 (1%) moved/LTF, 1 (1%) ILAR subtype changed (oligo to psoriatic). For the extended oligo, Poly RF− and Poly RF+ categories 94%, 82% and 60%, respectively, maintained CID for the first 6 months (c2 6.7, p 0.03). ANA status, MTX use, and type of anti‐TNF were not associated with the ability to maintain CID (c2 p values 0.48, 0.14, and 0.75, respectively).106 (77%) subjects maintained CID for the first 6 months and stopped anti‐TNF as per protocol. 67/106 (63%) maintained CID for ≥8 mos off anti‐TNF while 39 (37%) flared. Time without flaring after stopping anti‐TNF therapy was duration from the month 6 visit to the last study visit (mean/median/range for duration of followup was 249/250/126–322 days). The mean/median/range for time to flare was 108/105/7–271 days. Time to flare (days) for etanercept was 105/105/7–271, adalimumab 119/120/28–238 and infliximab 28/28/28. Flare was seen in 47% (8/17) extended oligo, 37% (30/80) poly RF– and 11% (1/9) poly RF+ ((c2 p‐value 0.19). In those on background MTX, 33% (13/40) flared at a mean of 90 days, while those not on background MTX, 39% (26/66) flared at a mean of 113 days ((c2 p‐value 0.48). Using univariate analysis of variance, only weak correlations of MTX dose, disease duration, and CID duration with flare/no flare were seen (Spearman correlations −0.03, −0.17, −0.19, respectively).Conclusion:In this prospective multicenter study, 77% of the PF‐JIA patients were able to maintain CID for the first 6 months on anti‐TNF. Discontinuation of anti‐TNF in PR‐JIA (who have demonstrated on average 1.8 years of CID) resulted in a flare rate of 37% within 8 mos. Clinical parameters had only minimal predictive ability. Ongoing work includes biomarker identification and continued follow‐up of the cohort.