BACKGROUND: It has been suggested that measurement of continuous cardiac output (CCO) is an advancement in the management of critically ill patients. Our objective was to determine the accuracy of CCO during the rapid hemodynamic changes induced by hemorrhage and resuscitation. METHODS: In 12 anesthetized dogs (20.2+/-0.9 kg), pulmonary artery blood flow, our "gold standard" cardiac output, was measured with an sonographic flowprobe, whereas CCO, intermittent bolus cardiac output (ICO), and mixed venous oxygen saturation were measured with a thermodilution fiberoptic pulmonary artery catheter with a thermal filament. A graded hemorrhage (20 mL/min) was produced to a mean arterial pressure of 40 mm Hg, which was maintained at this level for 30 minutes. Total shed blood volume (701+/-53 mL) was retransfused at a rate of 40 mL/min, over 30 minutes, after which a massive hemorrhage (100 mL/min) was produced over 10 minutes. RESULTS: Hemorrhage induced significant decreases in mean arterial pressure, mixed venous oxygen saturation, and oxygen delivery, which were all restored during early resuscitation. However, CCO showed a delayed response after hemorrhage and resuscitation, compared with pulmonary blood flow, throughout the study (r = 0.549), matching only at baseline and at the end of both graded hemorrhage and resuscitation periods. There was a good correlation between ICO and pulmonary artery blood flow (r = 0.964) and no significant differences between them throughout the study. CONCLUSION: CCO has a delayed response during acute hemodynamic changes induced by hemorrhage and resuscitation. When sudden changes in mean arterial pressure or in mixed venous oxygen saturation are detected, cardiac output must be estimated by the standard bolus thermodilution technique, not by CCO.