To compare cardiac index (CI) values obtained by pulmonary artery thermodilution (CI_PA), arterial thermodilution (CI_TD), and arterial pulse contour analysis (CI_PC) during rapid fluid administration, as accurate and rapid detection of CI changes is critical during acute preload changes for guiding volume and vasopressor therapy in critically ill patients, and the accuracy of CI_PC during acute changes in loading condition is currently unknown.

Prospective clinical study.

Cardiac surgical intensive care unit of a university hospital.

Seventeen American Society of Anesthesiologists (ASA) physical status II and III patients, aged 32 to 76 years, with normal left ventricular function during the early postoperative period after elective coronary artery bypass graft surgery.

After baseline determinations of CI_PA, CI_PC, and CI_TD were made, fluid loading was performed using 10 mL times body mass index of hydroxyethyl starch 6%. CI_PA, CI_PC, and CI_TD were determined, and changes in CI (ΔCI) were calculated. Fluid load was repeated until no increase in stroke volume index (ΔSVI <10%) was achieved.

Regression analysis between CI_PA/CI_PC, CI_PA/CI_TD, and CI_PC/CI_TD revealed r² = 0.92, r² = 0.92, and r² = 0.98. Regression analysis between ΔCI_PA/ΔCI_PC, ΔCI_PA/ΔCI_TD, and ΔCI_PC/ΔCI_TD revealed r² = 0.57, r² = 0.67, and r² = 0.74, respectively. Bland-Altman analysis was used to determine accuracy and precision of the 3 methods compared. The mean differences (m) and SD between ΔCI_PA/ΔCI_PC, ΔCI_PA/ΔCI_TD, and ΔCI_PC/ΔCI_TD resulted in m = −1.01%, SD = 6.51%; m = −0.83%, SD = 5.80%; and m = −0.33%, SD = 4.65%, respectively.

Compared with pulmonary artery thermodilution, arterial pulse contour analysis reflects relative changes in CI during rapid changes of preload with clinically acceptable accuracy.