1999 OPEN FORUM Abstracts

INACCURACIES IN CO₂ MEASUREMENTS OF THE i-STAT BLOOD GAS ANALYZER WHEN COMPARED TO THE AVL OMNI BLOOD GAS ANALYZER

RW Kraemer CRTT, M Siobal BS RRT, RH Kallet MS RRT, JD Marks MD PhD, Department of Anesthesia. Valerie Ng MD, Clinical Laboratories, San Francisco General Hospital, UCSF

Introduction: We report inaccuracies in PaCO₂ using the i-STAT Portable Clinical Analyzer with Thermal Control (i-STAT Corp, East Windsor, NJ). During routine weekly comparisons of i-STAT vs. AVL Omni blood gas instrument (AVL Scientific Corp. Roswell, GA), the i-STAT value for PaCO₂ was 20 mmHg lower than the AVL Omni value. In the previous month, a PaCO₂ discrepancy of 15 mmHg occurred. In both instances, commercial liquid controls performed on the i-STAT were within the accepted range. Because of these discrepancies, we performed a number of i-STAT vs. AVL comparisons in order to determine the extent of the problem.

Methods: First, three levels of liquid controls were performed on all instruments. Then 1.5 - 2.5 ml of blood was drawn from 11 critically ill patients using the SIMS line draw sampling kit (Part # 4043W-2, SIMS Portex, Keene, NH). Analysis of the sample was performed on the AVL Omni. AVL 995, and two i-STAT instruments providing 4 separate measurements for each sample.

Results: Liquid controls performed on all instruments were within the manufacturers acceptable range. In 3 of 11 patient samples (6 of 22 i-STAT vs. Omni comparisons), the i-STAT PaCO₂ was > 5 mmHg lower than the AVL with a maximum variance of 11.6 mmHg.