Continuous thermodilution (CT) using a pulmonary artery (PA) catheter with a thermal filament has the potential for intraoperative on-line monitoring of cardiac output. Liver transplantation frequently requires rapid fluid administration and often includes the use of an extracorporeal veno-venous bypass. To assess the agreement between CT and bolus thermodilution (BT) in such a setting, we conducted a prospective intraoperative study in 14 liver transplant patients. Throughout the operation, CT cardiac output was recorded and paired with BT measurements taken every 30 min and whenever indicated for clinical reason. Corresponding data were assigned to acquisition periods when patients were on or off veno-venous bypass (flow rate 2.5 +/- 0.2 L/min) and were discriminated by the various range of intravenous infusion rates (< 150 mL/h, 150-1000 mL/h, 1000-2000 mL/h, and 2000-4000 mL/h) and the magnitude of cardiac output (< or = 7.5 L/min, 7.5-10.0 L/min, > 10.0 L/min). A total of 270 data pairs was obtained and examined by analysis of agreement (mean difference +/- SD), variance, error, and weighted regression. Trend analysis was performed for significant CT and BT cardiac output changes, defined as changes greater than 15%. Agreement of both methods was best at peripheral intravenous fluid infusion rates < or = 1000 mL/h and BT cardiac output > 10 L/min (0.0 +/- 0.6 L/min) and was unaffected by veno-venous bypass. Discrepancy was most evident at intravenous fluid infusion rates > 2000 mL/h and BT cardiac output < or = 7.5 L/min (2.1 +/- 1.7 L/min). Correlation of CT and BT cardiac output was excellent (r = 0.95, P < 0.001) for combined data from all patients. Changes in CT cardiac output > 15% (n = 116) correctly indicated the direction in 93% of BT cardiac output changes and were 74% sensitive and 75% specific for significant BT cardiac output changes. The thermal filament technique enhances the usefulness of PA catheterization during liver transplantation but reflects BT cardiac output with clinically acceptable error only at low peripheral intravenous fluid infusion rates.
Implications: Cardiac output determines organ perfusion. In clinical practice, it is measured by intermittent thermodilution using right heart catheterization. This intraoperative study compared the intermittent method with a technique based on continuous thermodilution. The new technique provides logistical advantages and challenges the accuracy of the intermittent method during liver transplantation.