We previously demonstrated that high levels of IL-6/sIL-6R complexes are present in sera of patients with systemic juvenile idiopathic arthritis (s-JIA) and that the amount of IL-6 estimated in the IL-6/sIL-6R complexes is markedly higher than that measured by the B9 assay. Here, we show that two additional bioassays, employing human myeloma XG-1 cells and human hepatoma Hep3B cells, detected serum IL-6 levels similar to those measured by the B9 assay and approximately 10-fold lower than the IL-6 levels estimated to be present in the IL-6/sIL-6R complex. Using an assay for the measurement of the amount of circulating IL-6 complexed with the sIL-6R and available for binding to gp130 (gp130 binding activity), we show that the IL-6/gp130 binding activity is similar to that detected by the bioassays and again significantly lower than that estimated to be present in the IL-6/sIL-6R complex. Addition of recombinant human IL-6 (rhIL-6) to sera of patients or controls results in a markedly lower increase in the gp130 binding activity in patients than in controls. Moreover, sera from s-JIA patients inhibited in a dose dependent manner the gp130 binding activity assay. These results show that sera from patients with s-JIA contain a factor, or factors, that inhibit(s) the binding of the IL-6/sIL-6R complex to gp130. This inhibitory activity does not appear to be due to soluble gp130, C-reactive protein or autoantibodies to IL-6.