Measuring carbon dioxide tension in saline and alternative solutions: quantification of bias and precision in two blood gas analyzers

OBJECTIVE: To quantify the precision and bias for two standard blood gas analyzers when measuring PCO2 in five different test solutions (blood, normal saline, human albumin solution, succinylated gelatin 4%, and Hartmann's solution).

DESIGN: A comparative laboratory study of two blood gas analyzers measuring known PCO2 in test solutions prepared with a thin film tonometer.

MEASUREMENTS: The PCO2 was determined simultaneously in both blood gas analyzers. The precision and bias for both machines were calculated from the measured PCO2.

RESULTS: Analyzer 1 showed a negative bias of between 66.4% and 39.2% when measuring PCO2 in saline. Under the same conditions, analyzer 2 showed a negative bias of between 17.7% and 17.9% for PCO2 in the mid- and high ranges of PCO2, but a positive bias in the low range. Both machines measured PCO2 with satisfactory bias and precision in blood. The use of succinylated gelatin 4% as a test solution improved the bias and precision of both machines.

CONCLUSIONS: Because of the large bias and precision, neither of these machines is suitable for measuring PCO2 in saline solution. The use of succinylated gelatin 4% as a test solution improves the bias and precision of both machines and we favor this solution for nasogastric tonometry. Even with succinylated gelatin 4%, there is a large negative bias with analyzer 1 and our preference is to use analyzer 2. We recommend that other workers review critically the performance of blood gas electrode systems when measuring PCO2 in saline solution. Because of varying performance of "standard" machines, we also recommend that each institution determine its own reference range for intramucosal pH.