Estimation of cardiac output by noninvasive echocardiographic techniques in the critically ill subject.

We evaluated the accuracy of cardiac output estimations by three transthoracic echocardiographic techniques in critically ill subjects. This study was a prospective comparison study carried out in a general intensive care unit of a teaching hospital. The subjects had a broad range of diagnoses including pulmonary embolus, cardiogenic shock, septic shock, Legionnaire's disease and perioperative myocardial infarction. All patients requiring pulmonary artery catheterization underwent echocardiographic cardiac assessment with comparison of findings to those obtained by thermodilution techniques. Nineteen studies on eighteen patients were performed, with cardiac output calculated by the two-chamber Simpson's, four-chamber Simpson's, and left ventricular outflow tract (LVOT) Doppler methods. Acceptable data was obtained in those patients without mitral regurgitation. There was good correlation between the thermodilution technique and Simpson's two-chamber method (r = 0.91), but less so with the Simpson's four-chamber method (r = 0.77). All studies were included in the LVOT Doppler method with a good correlation (r = 0.94). A plot of differences between methods using the Bland and Altman statistical method indicated that only the LVOT Doppler method demonstrated acceptable agreement with a mean of 0.2 litres/minute, standard deviation of 0.82 litres/minute and 95% limits of agreement of -1.5 to +1.9 litres/minute. We concluded that the LVOT Doppler method was the only one which demonstrated acceptable agreement between the thermodilution method and echocardiographic techniques in all critically ill patients studied.