A newly developed transtracheal Doppler (TTD) computer for cardiac output determination was studied in nine patients after open heart surgery (coronary artery bypass grafting, n = 4; mitral valve replacement, n = 5). The measurements were compared with those simultaneously obtained by thermodilation. Doppler signals were adequate in 78% of the patients studied. Limited correlation between both methods (r = 0.248; r2 = 0.0615; mean of difference, 1.714 +/- 1.67 L/min; limits of agreement, -1.6 to 5.0 L/min) was found. The large difference in cardiac output readings between TTD and thermodilation may be due to (a) false angles of the ultrasound beam in relation to the aortic wall and blood flow or (b) misplacement of the ultrasound head and underestimation of the aortic lumen. Patients must be completely sedated and paralyzed to prohibit artifacts. Routine patient care can interfere with continuous measurements. Cardiac output determinations by TTD are limited to the period during which the trachea is intubated with the special TTD tube. We conclude that the TTD system does not offer accurate cardiac output determinations and that the routine use of this device is not practical.