This study assessed the agreement between three methods of cardiac output (CO) measurement, thermodilution, the current clinical standard, and two transoesophageal echocardiographic techniques. Measurements were performed in 37 patients using thermodilution, continuous wave Doppler across the aortic valve and pulsed wave Doppler positioned in the left ventricular outflow tract. The aortic valve area was measured by direct planimetry, and the left ventricular outflow tract area was calculated from its diameter. Weighted least products regression analysis was employed to detect bias, and standard deviation of the difference (SDdiff) was calculated. There was no fixed bias but there was proportional bias between continuous wave Doppler and thermodilution methods (SDdiff 0.92 l/min). There was fixed bias but not proportional bias between pulsed wave and thermodilution methods (SDdiff 1.1 l/min). There was neither fixed nor proportional bias between pulsed wave and continuous wave Doppler methods (SDdiff 1.1 l/min). The transoesophageal Doppler methods described can be clinical alternatives to thermodilution cardiac output measurement.