Simultaneous comparison of thoracic bioimpedance and arterial pulse waveform-derived cardiac output with thermodilution measurement.

OBJECTIVE: To compare the accuracy and reliability of thoracic electrical bioimpedance (TEB) and the arterial pulse waveform analysis with simultaneous measurement of thermodilution cardiac output (TD-CO) in critically ill patients.

DESIGN: Prospective data collection.

SETTING: Emergency department and critical care unit in a 2,000-bed inner-city hospital.

PATIENTS: A total of 29 critically ill patients requiring invasive hemodynamic monitoring for clinical management were prospectively studied.

INTERVENTIONS: Noninvasive cardiac output was simultaneously measured by a TEB device and by analysis of the arterial pulse waveform derived from the finger artery. Invasive cardiac output was determined by the thermodilution technique.

MEASUREMENTS AND MAIN RESULTS: A total of 175 corresponding TD-CO and noninvasive hemodynamic measurements were collected in 30-min intervals. They revealed an overall bias of 0.34 L/min/m2 (95% confidence interval, 0.24-0.44 L/min/m2; p < .001) for the arterial pulse waveform analysis and of 0.61 L/min/m2 (95% confidence interval, 0.50-0.72 L/min/m2; p < .001) for the TEB. In 39.4% (n = 69) of all measurements, the discrepancy between arterial pulse waveform analysis and TD-CO was >0.50 L/min/m2. The discrepancies of the arterial pulse waveform analysis correlated positively with the magnitude of the cardiac index (r2 = 0.29; p < .001). In 56.6% (n = 99) of all measurements, the discrepancy between TEB and TD-CO was >0.50 L/min/m2. The magnitude of the discrepancies of the TEB was significantly correlated with age (r2 = 0.17; p = .02). Measurements were in phase in 93.2% of all arterial pulse waveform analysis and in 84.9% of all TEB readings (p < .001).

CONCLUSIONS: The arterial pulse waveform analysis exhibits a greater accuracy and reliability as compared with the TEB with regard to overall bias, number of inaccurate readings, and phase lags. The arterial pulse waveform analysis may be useful for the monitoring of hemodynamic changes. However, both methods fail to be a substitute for the TD-CO because of a substantial percentage of inaccurate readings.