Transesophageal echocardiography is becoming increasingly popular as a method of intraoperative monitoring because it can be performed continuously, does not transgress the sterile operative field, and provides data with regard to valve function, ventricular volumes, and contractility. Recently it was suggested that it can be used to measure cardiac output; however, controversy remains regarding its accuracy. Cardiac output was measured simultaneously by transesophageal echocardiography (using a 5-MHz pulse-wave Doppler, single-plane viewing probe) and by the thermodilution method in 21 patients undergoing open heart operations. The cardiac outputs measured by thermodilution correlated poorly (r = 0.45) with the transesophageal values derived from the left ventricular cross-sectional area, and the mean difference was 0.47 +/- 2.17 (standard deviation) L.min-1, giving limits of agreement of from -3.87 to +4.81 L.min-1. Cardiac outputs measured by thermodilution correlated well (r = 0.95) with transesophageal Doppler values derived from pulmonary artery flow velocity, with a mean difference of 0.12 +/- 0.45 L.min-1 and narrow limits of agreement of from -0.78 to +1.02 L.min-1. Based on our findings, transesophageal Doppler echocardiographic determination of cardiac output using pulmonary artery flow measurements can provide accurate hemodynamic data in patients undergoing cardiac operations.