The Relative Accuracy of Three Transcutaneous Dual Electrodes at 45 °C in Adults

Abstract

Conventional practice dictates that the results of any study comparing transcutaneous partial pressures of oxygen and carbon dioxide (tcPo₂ and tcPCo₂) with arterial blood gas values (PaO₂ and PaCo₂) should be presented in the form of a scattergram (Fig. 1). The relationship between the two is usually represented by a linear regression equation of the form

$$ {\rm{y}}({\rm{tcP}}{{\rm{O}}_2}) = {\rm{a}} + {\rm{bx}}({\rm{Pa}}{{\rm{O}}_2}) $$

and a correlation coefficient, r, calculated (Godfrey, I985). The probability that the transcutaneous and arterial blood gas values are related is given by a probability value, P (Table I). Does this statistical analysis give us the most useful information from transcutaneous monitoring for clinical purposes? We suggest that a recently proposed alternative method of analysis (Bland and Altman, I986) may be more useful and we demonstrate its practical application in evaluating combined tcPo₂ and tcPCo₂ electrodes against endtidal oxygen and carbon dioxide (ETPO₂ and ETPCO₂) values in normal adults.