Noninvasive whole-body electrical bioimpedance cardiac output and invasivethermodilution cardiac output in high-risk surgical patients

Objective: To evaluate the reliability of whole-body impedance cardiography with two electrodes on either both wrists or one wrist and one ankle for t he measurement of cardiac output compared with the thermodilution method. Design: Prospective, clinical investigation Setting: Surgical intensive care unit of a university-affiliated community hospital. Patients: Simultaneous cardiac output measurements by noninvasive whole-bod y impedance cardiography (nCO) and invasive thermodilution (thCO) in 22 hig h-risk surgical patients scheduled for extended surgery requiring periopera tive pulmonary artery catheter monitoring. Interventions: None. Measurements and Main Results: A total of 109 sets of measurements consisti ng of 455 single comparison measurements between nCO and thCO were included in the analysis. The mean cardiac output difference between the two method s was 1.62 L/min with limits of agreement (2 SD) of +/- 4.64 L/min. The int er-measurement variance was slightly higher for nCO. The correlation coeffi cient between nCO and thCO was r(2) = 0.061 (p < .001) for single measureme nts and r(2) = 0.083 (p < .002) for sets of three to six measurements, The two most predictive factors for between-method differences were the absolut e thCO value (r(2) = 0.13; p < .001) and whether or not a continuous nitrog lycerin infusion was used (p < .05, Student's t-test). Conclusions: Agreement between whole-body impedance cardiography and thermo dilution in the measurement of cardiac output was unsatisfactory. Factors t hat can explain these differences are differences between the populations u sed for calibration of nCO and the study population, the influence of chang ing peripheral perfusion, and the effect of a supranormal hemodynamic state on the bioimpedance signal. Whole-body impedance cardiography cannot be re commended for assessing the hemodynamic state of high-risk surgical patient s as studied in this investigation. (Crit Care Med 2000; 28:2812-2818).