Cardiac index by thermodilution or transthoracic echocardiography in dogs at different hemodynamic states

Proper monitoring of cardiac index (CI) in critically ill patients requires accurate and minimally invasive methods. The aim of this study was to compare the CI values obtained by thermodilution or echocardiography using different methods in dogs in different hemodynamic states. Nine dogs weighing 19.6 +/- 1.3kg were anesthetized with isoflurane at 1.4V% (Baseline) and subjected to mechanical ventilation (MV), a hypodynamic state (Hypo) with isoflurane at 3.5V% and hyperdynamic state (hyper) with dobutamine infusion at 5 mu gkg(-1)min(-1). CI analysis was performed by thermodilution (TD) and using the modified Simpson's method, aortic velocitytime integral (A-VTI) method and pulmonary VTI (P-VTI) method. We performed Pearson's correlation and Bland-Altman analysis. The CI values (Lm(-2)min(-1)) of the animals in the Baseline, MV, Hypo and Hyper states were 4.3 +/- 1, 3.6 +/- 0.7, 2.9 +/- 0.66 and 6.1 +/- 2, for TD; 2.8 +/- 0.7, 2.4 +/- 0.3, 1.7 +/- 0.7 and 4.4 +/- 1.2, for Simpson's method; 3.4 +/- 0.9, 3.1 +/- 0.7, 2.6 +/- 3.4, 6.1 +/- 1.8 for A-VTI; and 3.6 +/- 0.8, 3.6 +/- 0.8, 2.7 +/- 0.6 and 6.2 +/- 1.5, for P-VTI. The CI values using Simpson's method were lower than those obtained by TD in all states, and it was observed a significant correlation in the Hypo (r=0.89) and Hyper (r=0.76) groups. In addition, the percent error in the Hypo group using Simpson's method was 26% relative to TD, which allowed for the identification of the different hemodynamic states. With respect to the other methods and states, there was no agreement or correlation between the methods and TD. We concluded that none of the tested echocardiography methods exhibited acceptable agreement with thermodilution at different hemodynamic states.