Analysis of the Errors of a Bubble Method for Estimation of Pco₂ and Po₂ in Whole Blood

¹ From the Departments of Environmental Medicine, Medicine and Biostatistics, Johns Hopkins University, Baltimore, Maryland

The errors of a bubble method for estimating Pco₂ and Po₂ in whole blood were studied by analyzing samples of blood prepared by classical tonometry. There were systematic differences between ‘bubble’ and ‘tonometer’ Pco₂ and Po₂ whose magnitude varied with the initial tension of the blood. There was no significant systematic difference between analyses of Pco₂ or Po₂ by two operators. There was no systematic difference between analyses of Po₂ in different sets of apparatus, but there was a 1.2 mm systematic difference between analyses of Pco₂ in different apparatus. The pattern of error of the method is such that averaging ‘the best two out of three’ analyses when duplicates fail to check may impair accuracy rather than improve it. A rejection procedure is outlined which should avoid this loss of accuracy without including a significant number of grossly erroneous values.