Secondary Logo

Journal Logo

Aspects of Cardiothoracic Anaesthesia

Inaccuracy of cardiac output determination by transoesophageal echocardiography: 125

Bettex, D. A.1; Hinselmann, V.1; Hellermann, J. P.2; Jenni, R.2; Schmid, E. R.1

Author Information
European Journal of Anaesthesiology: June 2004 - Volume 21 - Issue - p 20-21

Introduction: Determination of cardiac output (CO) by transoesophageal echocardiography (TOE) has revealed varying degrees of accuracy [1,2]. The purpose of this randomized, single blinded, double controlled study was to compare and establish prospectively the best TOE methods to determine CO in patients after cardiac surgery.

Method: Thirty consecutive patients undergoing coronary artery bypass grafting were included. Measurements were done postoperatively, after stabilization in the intensive care unit (ICU). TOE measurements were initiated and analysed on-line by a single observer, blinded to the thermodilution CO (TCO) values. A second analysis was later performed off-line by the initial as well as a second blinded observer. TOE and TCO measurements were executed in triplicates and averaged. TOE CO was obtained in randomized order through the aortic, mitral, and pulmonary valves, right and left ventricular outflow tracts, trans-gastric surface areas of the left ventricle (LV) and LV bi-dimensional (2D) volumes (Simpson's rules). An “eye-ball guessing” was obtained off-line from the 2D ventricular views. Pairs of values were compared according to the method of Bland and Altman. Bias, standard deviation (SD) and percent of error defined as 2SD/average CO were given.

Results: Complete measurement series were obtained in 19 (63.3%) patients; one measurement failed in 10 (33.3%) patients and two in 1 (3.3%) patient. Mean TCO was 5.43 ± 1.11 L/min. The best results were achieved by trans-aortic measurements with the use of the triangular shape assumption of opening surface (bias −0.36 to 0.5; percent error 37.0 to 42.5%) but bias values deviated considerably from zero, and none of these approaches reached the limit of agreement set at 30%. Eye-ball guessing was comparable with the best TOE measurements (bias −0.12 to −0.8; percent error 36.0 to 44.0%). Bland-Altman of most relevant data are reported in figure 1. No significant difference in the Bland-Altman test was found between on-line and off-line measurements, with respect to inter-observer variability.

Figure 1.
Figure 1.:
Bland and Altman plots of matched TOE and TCO trans-aortic (A) and eye-ball guessing (B) data in L/min.

Discussion: Our results show that in a post cardiac surgery ICU, TOE is an unreliable tool for CO determination. Although the measurements through the aortic valve, considering the aortic valve opening as a triangular surface [1], offered slightly better results, they remain in the range of clinical unreliability and are not better than eye-ball guessing. Off-line inter- and intra-observer control of our data revealed no significant difference.


1 Darmon PL, Hillel Z, Mogtader A, et al. Cardiac output by transesophageal echocardiography using continuous-wave Doppler across the aortic valve. Anesthesiology 1994; 80: 796-805.
2 Muhiudeen IA, Kuecherer HF, Lee E, et al. Intraoperative estimation of cardiac output by transesophageal pulsed Doppler echocardiography. Anesthesiology 1991; 74: 9-14.
© 2004 European Society of Anaesthesiology