The present study included data from three marathon races to investigate the hypothesis that a relationship exists between running intensity and elevated concentrations of interleukin (IL)-6 in plasma. The study included a total of 53 subjects whose mean age was 30.6 [95% confidence interval (CI) 1.4] years, mean body mass 77.7 (95% CI 2.0) kg, mean maximal oxygen uptake (VO2max) 59.3 (95% CI 1.4) ml x min(-1) x kg(-1), and who had participated in the Copenhagen Marathons of 1996, 1997 or 1998, achieving a mean running time of 206 (95% CI 7) min. Running intensity was calculated as running speed divided by VO2 max. The concentration of IL-6 in plasma peaked immediately after the run. There was a negative correlation between peak IL-6 concentration and running time (r = -0.30, P<0.05) and a positive correlation between peak IL-6 concentration and running intensity (r = 0.32, P<0.05). The IL-1 receptor antagonist (IL-1ra) plasma concentration peaked 1.5 h after the run and there was a positive correlation between the peak plasma concentrations of IL-6 and IL-1ra (r = 0.39, P<0.01). Creatine kinase (CK) plasma concentration peaked on the 1st day after the run, but no association was found between peak concentrations of IL-6 and CK. In conclusion, the results confirmed the hypothesized association between plasma IL-6 concentration and running intensity, but did not confirm the previous finding of a connection between IL-6 plasma concentration and muscle damage.