Proinflammatory cytokines, such as tumor necrosis factor-alpha, interleukin-6, and interleukin-8, and anti unflammatory cytokines, such as interleukin-10, may play an important role in patient responses to cardiopulmonary bypass. We sought to define whether the myocardium and the lungs serve as important sources of these cytokines under conditions of cardiopulmonary bypass. Ten patients (age 64 +/- 3 years, mean +/- standard error of the mean) undergoing elective coronary artery bypass grafting were monitored with an arterial catheter, a coronary sinus catheter, and pulmonary artery catheter. Plasma levels of tumor necrosis factor-alpha, interleukin-6, interleukin-8, and interleukin-10 were measured simultaneously in peripheral arterial blood, coronary sinus blood, and mixed venous blood before heparin administration, 1 minute before aortic crossclamping, 5 minutes after aortic declamping, and at 0.5, 1, 1.5 and 2 hours after aortic declamping. The durations of cardiopulmonary bypass and aortic crossclamping were 114 +/- 9 and 64 +/- 5 minutes, respectively. Levels of tumor necrosis factor-alpha and interleukin-6 were significantly higher in coronary sinus blood than in arterial blood after aortic declamping. Tumor necrosis factor-alpha and interleukin-6 levels were also higher in mixed venous blood than in arterial blood within 1 hour after declamping. There were no significant differences among the three sampling sites with respect to interleukin-8 and interleukin-10 levels. In one patient who had postoperative myocardial infarction, however, interleukin-8 levels were three times as high as in coronary sinus blood than in arterial blood. These data indicate that the myocardium is a major source of tumor necrosis factor-alpha and interleukin-6 in patients undergoing cardiopulmonary bypass. The lungs may consume rather than release proinflammatory cytokines in the early phase of reperfusion. The source under these conditions of the antünflammatory cytokine interleukin-10 remains to be determined.