Multicenter trial of a new thoracic electrical bioimpedance device for cardiac output estimation

To evaluate the capacity of a new thoracic electric bioimpedance system to estimate cardiac output compared with the conventional thermodilution method. Prospective, multicenter study. A university-run county hospital, a university-run U.S. Veterans Affairs hospital, and a university-affiliated U.S. military hospital. A series of 68 critically ill patients whose conditions required pulmonary artery catheter insertion. A total of 842 simultaneous pairs of cardiac output estimations by conventional thermodilution and a new thoracic electric bioimpedance system that uses an improved signal processing technique based on an all-integer-coefficient filtering technology, using a time-frequency distribution that provides a high signal/noise ratio were evaluated. The r value was .86, r2 = .74, and p < .001 by regression analysis; the mean difference between the two methods relative to their average value was 16.6 +/- 12.9 (SD) %; the precision was 1.4 L/min or 0.8 L/min/m2; the bias was -0.013 L/min. The mean difference between successive pairs of thermodilution measurements was 8.6 +/- 0.6 (SD) %, which was about half the difference between simultaneous pairs of measurement by the two methods. The changes in impedance estimations were close to simultaneously measured changes in thermodilution estimates of cardiac output. The new bioimpedance system satisfactorily estimated cardiac output as measured by the thermodilution technique. The difference between the two estimations is more than made up for by the continuous noninvasive capability of the impedance system.