Eighteen Caucasian (white, Middle East and Asian) children diagnosed by paediatric rheumatologists in the UK and France as having systemic juvenile idiopathic arthritis (sJIA) were enrolled in this open label, single dose trial. All patients had evidence of continued symptoms and disease activity for at least three months while receiving >0.2 mg/kg/day of prednisolone, or its equivalent, prior to recruitment. Twelve patients also received methotrexate (< or =20 mg/m2/week). The patients were divided into three groups receiving 2, 4 or 8 mg/kg of MRA (tocilizumab) by intravenous infusion. No evidence of dose-limiting toxicity was observed and there were no dose-limiting safety issues. MRA appeared to be dramatically effective, with clinical and laboratory responses observed by 48 h post infusion, and these improvements continued well after serum MRA was undetectable. Eleven patients achieved the JIA definition of improvement (at least 3 of 6 core set criteria with a 30% improvement and no more than one worsened by 30%) and eight achieved > or =50% improvement. There were no observable differences with age. Clinical improvement in these children was observed for up to eight weeks, supporting the hypothesis that IL-6 is a key cytokine in the upregulation of genes crucial in the inflammation processes of sJIA, and the possibility of sequestration of MRA in the extra-vascular compartment needs to be considered.