To determine whether transesophageal echocardiography could be used to estimate intraoperative cardiac output, the authors studied 35 consecutive patients undergoing cardiovascular surgery (coronary artery disease [n = 22], aortic valve disease [n = 5], mitral valve stenosis [n = 5], peripheral vascular disease [n = 3]). Two-dimensional echocardiographic and pulsed-wave Doppler signals of the pulmonary artery and mitral valve flow velocity were obtained simultaneously with thermodilution measurements of cardiac output. Cardiac output derived from pulsed Doppler imaging of pulmonary artery systolic flow velocity modestly correlated with the thermodilution-derived cardiac output (r = 0.65), but output determined from the mitral valve diastolic flow velocity did not (r = 0.24). Transesophageal echocardiography of pulmonary artery systolic flow satisfactorily detected intraoperative increases in cardiac output greater than 15% (sensitivity, 71%; specificity, 82%) but not decreases (sensitivity, 54%; specificity, 90%). Although this technique identifies increases in cardiac output greater than 15%, it does not detect decreases as accurately as those detected by thermodilution measurements. At this time, therefore, transesophageal Doppler echocardiography has significant limitations as an off-line monitor of cardiac output.