1995 OPEN FORUM Abstracts
ACCURACY COMPARISON OF BEDSIDE AND LABORATORY BLOOD GAS ANALYZERS.
Patricia A. Meyers, RRT, Dennis Bing, RRT, Edrie Murphy,BS,CLS, MBA, Kendra Smith, MD, Mark Mammel, MD. from Infant Pulmonary Research Center, Children's Health Care St. Paul, MN
Introduction: Bedside blood gas monitoring in the ICU is now available. Can these new devices replace conventional in- lab analysis? This study compares accuracy of two bedside blood gas analyzers to standard clinical laboratory measurement in a neonatal lung injury animal model. Method:We evaluated these systems: P-7 (Paratrend-7) intravascular continuous blood gas monitoring system (Biomedical Sensors, Malvern, PA), StatPal II, a portable bedside blood gas analyzer (PPG Industries, La Jolla, CA), and ABL 620 bench analyzer (Radiometer, Copenhagen, Denmark). 6 newborn piglets were sedated, intubated and ventilated at an FiO2 of 1.0. We induced lung injury by repeated saline lung lavage, producing a wide variety of arterial blood gas values. We placed in vitro calibrated P-7 sensors in the descending aorta via carotid artery cutdown. The StatPal II was calibrated just prior to each sample. The ABL 620 auto-calibrates on timed cycles. Saving current P-7 readings, we drew arterial blood samples for analysis by StatPal II and ABL, corrected for body temperature. We collected 41 comparative points. We analyzed all data by linear regression; and separately analyzed 7 data points with PO2 values greater than 300 torr.
Results: All three devices showed good correlation and slope-intercepts for pH and PCO2. Correlation for PO2 was lower, particularly at very high values.
ABL- P[cdot]7 ABL-Stat Stat-P[cdot]7
pH r= 0.988 0.988 0.973
slope 0.944 1.032 0.890
PCO2r=0.991 0.990 0.981
slope 0.972 0.954 0.999
PO2 r=0.956 0.983 0.951
(all) slope0.892 0.641 1.361
PO2 r=0.802 0.944 0.698
(>300) slope 0.674 1.082 0.511
Conclusion: The two bedside blood gas devices we evaluated are acceptably accurate for clinical management.