It has been evident for more than a decade that different instrument models can yield model dependent results for Po2, Pco2, and pH. This has been demonstrated using aqueous or fluorocarbon containing emulsion (FCE) proficiency testing materials as well as aliquots of tonometered blood. Logically, instruments that are technologically similar should reveal near-identical results for Po2, Pco2, and pH intensities along their entire range, from low to high values, whereas differences in electrodes, calibration, sample flow, or signal processing might cause two instrument models to give differing results. We sought to determine the extent of these differences among the 20 most commonly used analyzers, whether made by the same or different manufacturers, by analyzing FCE proficiency testing data employing a wide range of analyte intensities from >900 instruments. This analysis leads to the conclusion that measurements made by blood gas analyzer models differ not only between manufacturers, but also within manufacturers.
Materials and Methods
Data Acquisition and Tabulation
Every 4 months, the American Thoracic Society-California Thoracic Society Proficiency Testing Survey sends out ampules of identical composition from five lots of FCE proficiency testing material to approximately 400 laboratories with >900 enrolled instruments. Within a month, data are received from each enrolled instrument. From these responses, the mean and SD for each lot for each model are determined. Results are reported back to enrollees and, as necessary, to governmental agencies. We performed a retrospective analysis of these data for six successive periods from late 1994 to mid 1996, in which all instrument mean (AIM) lot values ranged from about 40 to 180 mm Hg for Po2, 20 to 75 mm Hg for Pco2, and 7.15 to 7.60 for pH units (U).