Evaluation of a new continuous thermodilution cardiac output monitor in cardiac surgical patients: a prospective criterion standard study.

Objective: To evaluate the accuracy of a new continuous cardiac output monitor in critically ill patients.

Design: Criterion standard study.

Setting: Cardiac surgery intensive care unit in a university hospital.

Patients: Twenty cardiac surgical patients requiring intensive care treatment with pulmonary artery catheters after surgery.

Interventions: None.

Measurements And Main Results: Cardiac output was monitored continuously with a modified pulmonary artery catheter that has a heating filament on the outside of the catheter. Four modes of data processing with different response times ("Normal," "Fast," "FastFilter," and "Urgent" modes) used by the monitoring system. A total of 240 determinations of cardiac output were performed using conventional bolus thermodilution technique; these results were compared with those obtained using three of the four continuous measuring modes available ("Normal," "FastFilter," and "Urgent"). Cardiac output ranged from 3.47 to 15.77 L/min (bolus thermodilution). The mean (bias) +/- SD of differences (precision) for all measurements was 0.40+/-1.26 L/min in the Normal mode (cardiac output <10 L/min: 0.34+/-0.66 L/min), 0.53+/-1.27 L/min in the FastFilter-mode (cardiac output <10 L/min: 0.60+/-0.75 L/min), and 0.63+/-1.34 L/min in the Urgent mode (cardiac output <10 L/min: 0.57+/-0.82 L/min).

Conclusions: Continuous cardiac output measurement using the thermodilution technique is reasonably accurate, reliable, and applicable in routine clinical practice. The values obtained using the Normal mode of the monitor agreed significantly better with the conventional thermodilution method than the results of the two other modes studied (FastFilter and Urgent). In addition, measurements in two patients with cardiac output values of >10 L/min did not agree with the results of the bolus thermodilution method.