Objective: Evaluation of continuous cardiac output monitoring based on the thermodilution technique in the critically ill.

Design: Prospective clinical investigation.

Setting: A surgical intensive care unit of a university hospital.

Patients: Thirty-five critically ill patients (trauma and/or sepsis patients), who needed pulmonary artery catheterization. The patients were prospectively studied according to the following groups: a) patients with a heart rate of >120 beats/min; b) those patients with a cardiac output of >10 L/min; c) patients with a cardiac output of <4.5 L/min; d) patients with a rectal temperature of >39.0 degrees C; and e) patients with a pulmonary artery catheter inserted for >4 days.

Interventions: Therapies were carried out according to modern intensive care medicine protocols by physicians who were not involved in the study.

Measurements: Cardiac output was monitored continuously using a new, modified pulmonary artery catheter. This catheter has a heating filament by which energy is transmitted to the circulating blood (modified thermodilution technique). A bedside microprocessor calculated cardiac output using a new algorithm. Standard bolus thermodilution technique (10 mL of ice-cold saline solution) was used to compare the continuous cardiac output measurement with the intermittent bolus cardiac output measurement.

Main Results: A total of 404 pairs of intermittent (bolus) cardiac output and continuous cardiac output measurements were obtained from the 35 patients. The bias (mean difference between bolus cardiac output measurement and continuous cardiac output measurement) of all measurements was 0.03 +/- 0.52 L/min and the 95% confidence limit (mean difference +/-2 SD) was -1.01/1.06 L/min. Also, continuous cardiac output measurement agreed closely with bolus cardiac output measurement (bias was 0.16 +/- 0.51 L/min in the cardiac output of >10 L/min group; bias was -0.17 +/- 0.50 L/min for the cardiac output of <4.5 L/min group). Increased temperature and prolonged length of stay did not influence the agreement of continuous cardiac output measurement with bolus cardiac output measurement (bias was 0.09 +/- 0.51 L/min in the >39 degrees C rectal temperature group).

Conclusions: Continuous monitoring of cardiac output using a modified pulmonary artery catheter with a heated filament has proven to be accurate and precise in the critically ill patient when compared with the ''standard'' intermittent bolus thermodilution technique. The continuous monitoring technique enhances our armamentarium for more intensive monitoring of these patients under a variety of circumstances.