Long-term evaluation of a continuous intra-arterial blood gas monitoring system in patients with severe respiratory failure.

Abstract

The purpose of this study was to evaluate the reliability and accuracy of a new continuous intra-arterial blood gas monitoring system (IABG; PB3300, Puritan Bennett) over a prolonged period of time (> 7 days). Prospective criterion standard study. Anesthesiological intensive care unit in a university hospital. 11 sensors were tested in 10 mechanically ventilated patients with severe respiratory failure. PO2, PCO2, and pH measured using IABG were compared to values obtained from 2 conventional blood gas analyzers. The quality of blood pressure tracings was assessed using a scoring system consisting of 5 grades. The median study period was 205h/sensor (range: 169-506h). 320 blood samples were obtained. The ranges of measured parameters were: PO2 = 46-433 mmHg, PCO2 = 25-79 mmHg, pH = 7.25-7.55. The mean (SD) differences for the whole study period were: -4.3 (11.9) mmHg for PO2, for the clinically important range (PO2 < 150 mmHg) -1.9 (5.4) mmHg, -2.8 (4.5) mmHg for PCO2, and -0.03 (0.04) for the pH value. The MD (SD) in relation to the sensor lifetime were for days 1-3: -1.1 (5.1) mmHg for PO2, -0.4 (3.9) mmHg for PCO2, and -0.01 (0.03) for the pH value; for days 4-6: -1.5 (6.0) mmHg for PO2, -3.3 (4.0) mmHg for PCO2, and -0.03 (0.03) for the pH value; for days 7-9: -2.5 (4.7) mmHg for PO2, -5.1 (4.6) mmHg for PCO2, and -0.04 (0.04) for the pH value; for days > 9: -4.9 (4.4) mmHg for PO2, -5.3 (4.1) mmHg for PCO2, and -0.05 (0.03) for the pH value. The IABG reliably measured blood gases and pH values with acceptable clinical performance based on the overall results. There was, however, a decline in the agreement of the sensors and conventional values with increasing sensor lifetime. The mean differences (bias) and the standard deviation of differences (precision) of PO2, PCO2 and the pH values were acceptable for clinical purposes up to day 6. The arterial blood pressure tracings and blood withdrawal were not adversely affected. No side effects due to the sensors occurred. In summary, a prolonged sensor use for a period of up to 6 days appears to be reasonable. This system offers on-line information on oxygenation, ventilation, and acid-base status and allows immediate detection of acute and potentially life-threatening events.