INTRAARTERIAL BLOOD-GAS MONITORING-SYSTEM - MORE ACCURATE VALUES CAN BE OBTAINED

Objective. To compare values measured by a continuous intra-arterial b lood gas monitoring system with those measured by conventional blood g as analyzer for the assessment of the clinical performance of a new de vice for measurement of PaO2, PaCO2, and arterial pH. Methods. Forty-s ix patients undergoing cardiopulmonary bypass were enrolled in this st udy. All patients had a continuous intra-arterial sensor (PB 3300) pla ced into the radial artery through a 20-gauge catheter. A total of 319 arterial blood gas and pH values were obtained for comparison with a conventional blood gas analyzer. The measurements were performed every 12 hrs after the initial in vitro calibration of the sensor for each patient. Results. Measurements were made over a range of 12 to 192 hrs . The overall bias and precision determined by the two methods were 4. 5 and 17.1 mmHg for PaO2; 4.5 and 6.2 mmHg for PaCO2; and 0.009 and 0. 035 for pH, respectively. For the range of PO2 less than 150 mmHg, the bias and precision improved to 4.2 and 9.5 mmHg. The sensor-derived P CO2 value, PCO2(IABG), increased significantly more than the conventio nal blood gas analysis value, PCO2(ABG), even within 72 hrs (2.8 and 4 .1 mmHg). The relationship between the two measurements can be describ ed as: PCO2(IABG)/PCO2(ABG) = 1 + 0.0026 . t where t is the time perio d of use (in hours). By correcting the PCO2(IABG) value using this for mula, the overall bias and precision of the values measured by two met hods decreases to -0.4 and 3.6 mmHg. Conclusions. The PO2 and pH value s derived from an intra-arterial blood gas monitoring system agreed we ll with the values measured by a conventional blood gas analyzer. Howe ver, the PCO2 value must be corrected due to an increase of drift, esp ecially with extended use for more than 72 hours.