A comparison of hemodynamic parameters derived from transthoracic electrical bioimpedance with those parameters obtained by thermodilution and ventricular angiography

Abstract

Objective: To determine the limits of agreement between the cardiac output and volumetric data estimated by impedance cardiography with the cardiac output determined by thermodilution and the left ventricular ejection fraction and end-diastolic volume estimated from left ventriculography.

Design: A prospective study.

Setting: The cardiac catheterization laboratory of a university-affiliated teaching hospital.

Patients: Twenty-four patients with coronary artery disease undergoing elective left- and right heart catheterization.

Interventions: Cardiac output was measured by the thermodilution method and the ejection fraction and left ventricular volumetric data were determined by ventriculography. These same measurements were obtained by simultaneously performed impedance cardiography using a commercially available bioimpedance device.

Measurements and main results: The patients' mean cardiac output was 4.6 +/- 1.7 L/min by bioimpedance and 5.0 +/- 1.1 L/min by thermodilution. The limits of agreement between the two methods was -4.1 to 3.5 L/min. The 95% confidence intervals for the lower and upper limits of agreement were -2.7 to -5.5 L/min and 2.1 to 4.9 L/min, respectively. The mean ejection fraction was 63 +/- 8% by bioimpedance and 53 +/- 15% by ventriculography. The limits of agreement between the ejection fraction estimated by bioimpedance and ventriculography was -35% to 37%. The 95% confidence intervals for the lower and upper limits of agreement were -22% to -48% and 24% to 50%, respectively. The mean left ventricular end-diastolic volume was 108 +/- 47 mL, as estimated by bioimpedance, and 121 +/- 35 mL, as estimated by ventriculography. The limits of agreement between the left ventricular end-diastolic volume as estimated by bioimpedance and ventriculography was -139 to 113 mL. The 95% confidence intervals for the lower and upper limits of agreement were -184 to -94 mL and 68 to 158 mL, respectively.

Conclusions: The 95% confidence range defining the limits of agreement between cardiac output and volumetric data estimated by bioimpedance, with the cardiac output measurement by thermodilution and the volumetric data estimated from left ventriculography, were wide, making the degree of agreement clinically unacceptable. In the opinion of the authors, impedance cardiography should not replace invasive hemodynamic monitoring at this time.