BACKGROUND: Inflammatory cytokines such as tumor necrosis factor alpha (TNFalpha) and interleukin (IL)-6 play an important role in pathophysiology of rheumatoid arthritis (RA). We investigated the possibility whether TNFalpha and IL-6 could be used as an objective marker reflecting treatment response in RA. METHODS: Erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) and rheumatoid factor (RF) together with TNFalpha and IL-6 were measured in 159 specimens obtained from 95 RA patients. RA patients were divided into pre-treatment, methotrexate (MTX) and non-MTX groups by treatment regimen and into inactive and active groups by disease activity. The agreement between changes in marker levels and treatment response, and the correlation between each marker were analyzed. RESULTS: IL-6 was higher in active than in inactive group of patients in all three different treatment subgroups, but TNFalpha was not different between the two groups. IL-6 showed a better agreement with treatment response (MTX group, K=0.58; non-MTX group, K=0.21) than ESR or CRP, whereas TNFalpha did not show an agreement with treatment response. IL-6 was correlated with both ESR (r=0.22) and CRP (r=0.54), but TNFalpha was correlated only with ESR (r=0.21). CONCLUSIONS: Unlike TNFalpha, IL-6 reflects disease activity of RA and shows a better agreement with treatment response than ESR or CRP, indicating that it has an association with clinical features of RA. Therefore IL-6 could be used as an additional marker in the evaluation of treatment response when markers like ESR or CRP show results discordant from clinical features.