Background Identifying predictors for early clinically inactive disease (CID) would allow stratified treatment decisions at diagnosis, minimising the burden of unnecessary therapies. JIA core outcome variables (COVs) are routinely collected and would therefore be convenient predictors. However, different groups of children are identified by current CID definitions (Wallace's preliminary criteria vs. clinical Juvenile Arthritis Disease Activity Score in 10 joints (cJADAS10)) and predictors may differ depending on which CID outcome is applied.

Objectives To assess whether baseline COVs predict achievement of CID according to Wallace's preliminary criteria or the cJADAS10 cut-off in patients with JIA.

Methods Children and young people enrolled to the Childhood Arthritis Prospective Study (CAPS), a UK multicentre inception cohort, before January 2011, were selected if diagnosed with oligoarticular, RF-negative or RF-positive polyarticular JIA.

At one year following initial presentation to paediatric rheumatology, children were classified as i) CID according to Wallace's preliminary criteria and ii) CID according to cJADAS10. Baseline COVs (active joint count, limited joint count, physician's global, parental global, functional ability (Childhood Health Assessment Questionnaire (CHAQ)) and ESR) were tested for predictive ability for these outcomes using univariate and forced-entry multivariate logistic regressions, adjusting for age and symptom duration at initial presentation, gender and ILAR subtype. Multiple imputation accounted for missing data.

Results Of 829 children included, 70% were female and the majority had oligoarticular JIA (68%). At one year, 28% had achieved CID according to Wallace's preliminary criteria and 38% according to the cJADAS10 (21% CID on both).

In univariate analyses, increased baseline CHAQ and physician's global assessment score predicted lower odds of achieving both CID states. In addition, increased active joints (OR: 0.97, 95% CI0.94, 0.99), and patient/parent global assessment scores (OR: 0.87, 95% CI 0.81, 0.93) predicted lower odds of CID on the cJADAS10 only. In multivariate analyses, one increased CHAQ point at baseline independently predicted 28% lower odds of CID on the cJADAS10 (95% CI 0.53, 0.98). However, no baseline COVs predicted CID on Wallace's preliminary criteria. No demographic variables were significantly predictive in any model.

Conclusions There were different predictors for CID on the cJADAS10 vs. Wallace's preliminary criteria. Children with poor functional ability at initial presentation are less likely to achieve CID on the cJADAS10. These children could be targeted with more aggressive treatment strategies to better control their disease.

Disclosure of Interest None declared