To investigate whether the release of endotoxin during cardiopulmonarybypass (CPB) is determined by perfusion-related factors, endotoxinconcentrations were determined before, during, and after CPB in 21 malepatients (age range 45-75 years) undergoing elective coronary artery bypassgrafting. Hemodynamic parameters and oncotic pressure were also measured.Significant increases in endotoxin concentrations were observed after thestart of CPB (P < 0.005), before aortic cross-clamp release (P <0.05), and after aortic cross-clamp release (P < 0.05). The medianendotoxin concentration after cessation of CPB was 0.264 EU/ml (range <0.036-0.480 EU/ml). Endotoxin concentrations derived from the primesolutions were not contributory. Positive correlations were found betweenarterial pressure after the start of CPB and the endotoxin concentration 10min after (r = 0.58, P < 0.01) and between the duration of aorticcross-clamping and the endotoxin concentration after the cessation of CPB(r = 0.64, P < 0.005). Arterial pressure after the start of CPB, theduration of aortic cross-clamping, and decrease in oncotic pressureappeared to be independent variables in a forward variable selection modelthat predicted endotoxin concentrations after CPB. We conclude that inpatients undergoing elective coronary artery bypass grafting, an earlyphase of endotoxin release during CPB could be demonstrated, and that thisis due to vasoconstriction. The endotoxin concentrations after thecessation of CPB were determined by early vasoconstriction, duration ofaortic cross- clamping, and hypo-oncotic hemodilution.