HAND-HELD BLOOD-GAS ANALYZER IS ACCURATE IN THE CRITICAL CARE SETTING

Objective: To determine the accuracy of a new, portable, battery-power ed blood gas analyzer when used by nonlaboratory-trained clinicians in the critical care setting. Design: Prospective analysis of blood samp les from critically ill patients. Patients: Heterogeneous group of med ical and surgical critically ill patients.Interventions: None. Measure ments and Main Results: Two hundred thirty nine split blood samples fr om intensive care patients were analyzed by clinicians in the critical care environment using a new, portable, battery-powered blood gas ana lyzer (Immediate Response Mobile Analyzer [IRMA(TM)], Diametrics Medic al, St. Paul, MN). Near-patient measurements were compared with measur ements obtained by laboratory technologists using an IL-1312 blood gas analyzer (Instrumentation Laboratories, Lexington, MA) in an establis hed near patient critical care laboratory. Precision and coefficients of variation were also determined using repeated testing of quality co ntrol samples at three levels of pH, Po-2, and Pco(2). There was good agreement between IRMA determinations and the laboratory. Correlation coefficients ranged from 0.96 to 0.99. Bias and precision (+/- 2 SD), respectively, were 0.02 and 0.036 units for pH, -0.3 torr (-0.04 kPa) and 7.2 torr (0.96 kPa) for Pco(2), and -3.9 torr (-0.52 kPa) and 13.8 torr (1.8 kPa) for Po-2. Precision on repeated testing of quality con trol samples ranged from 0.022 to 0.04 units for a pH of 7.2 to 7.6, 1 .2 to 4.6 torr (0.16 to 0.61 kPa) for a Pco(2) of 20 to 60 torr (2.7 t o 8 kPa), and 3.0 to 7.4 torr (0.40 to 0.99 kPa) for a Po-2 of 70 to 1 60 torr (9.3 to 21.3 kPa). Coefficients of variation ranged from 0.15% to 0.28% for a pH of 7.2 to 7.6, 2.0% to 3.7% for a Pco(2) of 20 to 6 0 torr (2.7 to 8.0 kPa), and 1.7% to 3.6% for a Po-2 of 70 to 160 torr (9.3 to 21.3 kPa). Mean turnaround time was 16.5 +/- 10.1 mins for th e near-patient laboratory and 2 +/- 0.5 mins for IRMA. Conclusions: IR MA is accurate and reproducible when used in the clinical setting by n onlaboratory-trained individuals. Nonlaboratory trained individuals ca n obtain laboratory results in the critical care setting comparable wi th the results obtained by trained laboratory technologists. Bedside l aboratory testing decreases turnaround time compared with a near-patie nt laboratory.