We conducted a study testing the hypothesis that IL-18 predicts disease activity in patients with rheumatoid arthritis (RA), as measured by the number of swollen joints (40 joint counts), DAS28, the sedimentation rate and bone erosion. The study consisted of 57 patients with RA and 20 control subjects. The mean age of the patients was 51 years (range 29 to 75 years). The duration of the disease averaged 7.3 years (range 0 to 30 years). The numbers (means) of swollen and tender joints were 5.2 and 5.5 respectively. The mean activity disease score (DAS28) averaged 4.9 (range 0.6 to 11). Methotrexate and systemic glucocorticoids were administered to 40% and 30% of the patients, respectively. IL-18 was significantly increased in RA patients as compared to controls. The plasma concentration of IL-18 averaged 231 and 144 pg/ml in RA patients and controls (P < 0.001). DAS28, CRP, sedimentation rate and IL-18 positively correlated with the plasma IL-1Ra concentration (P < 0.001, P < 0.001, P < 0.001 and P < 0.002), respectively. Stepwise backward and multiple regression analyses demonstrated that the plasma concentration of IL-1Ra correlated positively with sedimentation rate, serum triglycerides and bone erosion (r=0.74; P < 0.001). Plasma IL-18 correlated positively with CRP (r=0.36; P < 0.007) and IL-1Ra (r=0.42; P < 0.002) but not with the overall disease activity (DAS 28) or bone erosions. S-triglycerides correlated positively with IL-1Ra and IL-18 (P < 0.002, P < 0.05), respectively. Although IL-18 correlated positively with CRP and IL-1Ra in RA, we were unable to demonstrate any significant correlation with clinical disease score in terms of DAS28, disease duration or bone erosion.