BACKGROUND: The aim of this study was to compare the accuracy of measured cardiac output using the newly developed esophageal doppler device with that of the thermodilution method using a pulmonary artery catheter. METHODS: In 15 patients undergoing off-pump coronary artery bypass surgery, cardiac outputs were measured at four episodes of surgery; (1) after induction of anesthesia, (2) during dissection of the internal mammary artery, (3) during anastomosis of the left anterior descending artery, and (4) after closure of the pericardium. RESULTS: The bias between the two methods was 0.52 +/- 1.09 L/min. Analysis of the changes in cardiac output from sample episode 1 to 2, from sample episode 2 to 3 and from sample episode 3 to 4, expressed as percent change values, shows no significant differences between the two methods (P > 0.05). CONCLUSIONS: The esophageal doppler accurately reflects changes in cardiac output with time when compared with that of the thermodilution.