1996 OPEN FORUM Abstracts
ARE POINT-OF-CARE BLOOD GAS ANALYZERS ACCURATE ALTERNATIVES TO IN-LAB ANALYZERS?
Patricia A. Meyers. RRT, Dennis Bing, RRT, Edrie Murphy, BS, CLS, Mark Mammel, MD. from Infant Pulmonary Research Center, Children's Health Care St. Paul, MN
Introduction: Point-of-care blood gas analysis in the ICU is now available. Can these new devices be used in place of conventional in-lab analysis? This study compares results from two bedside blood gas analyzers and a standard clinical laboratory measurement in a neonatal lung injury animal model. Method: We compared 2 point-of-care blood gas analyzers, IRMA, (Diametrics Medical Inc., St. Paul, MN), and StatPal II, (SenDx Medical, Inc., Carlsbad, CA), and one in-lab analyzer, ABL 620 (Radiometer, Copenhagen, Denmark). 10 newborn piglets were sedated, intubated and ventilated at an FiO2 ranging from .50 to 1.0. We induced lung injury by repeated saline lung lavage, producing a wide variety of arterial blood gas values. The StatPal II and IRMA were calibrated just prior to each sample. The ABL 620 auto-calibrates on timed cycles. Using Smooth E arterial blood gas syringes (Radiometer, Copenhagen, Denmark) we drew samples for analysis by StatPal II, IRMA and ABL, corrected for body temperature. We collected 48 comparative points of pH, PCO2, and PO2, and separately analyzed 41 data points with PO2 values less than 300 torr. We assessed agreement using the Bland-Altman comparative technique. Results: Means of the differences for each data point (bias) and % of data within 2 standard deviations of the mean are reported below.
ABL-IRMA ABL-Stat IRMA-Stat
pH mean difference 0.04 0.01 -0.03
% within 2 SD 93.7 95.8 95.8
PCO2 mean difference -3 0 3
% within 2 SD 93.7 97.9 97.9
PO2 mean difference -8 -25 -17
(all) % within 2 SD 93.7 93.7 93.7
PO2 mean difference -3 -9 -6
( < 300)% within 2 SD 97.6 95.1 92.7
We found close agreement between the point-of-care analyzers and the in- lab analyzer for pH and PCO2. Mean differences for all PO2 data were wide. We analyzed PO2 data separately in a more physiologic range ( < 300 torr) and found bias to be acceptable. Conclusion: The two point-of-care instruments evaluated in this study, IRMA and StatPal, are accurate alternatives to in-lab analyzers in common clinical practice.