A comparison between direct and calculated oxygen saturation in intensive care

We investigated the discrepancy between calculated and spectrophotometrically determined oxygen saturation, and the corresponding effect of this difference on calculated oxygen uptake in 46 arterial-venous sample pairs from 28 critically ill patients. The range of discrepancy between the two methods showed limits of agreement (mean +/- 2SD) of -2.26 to +0.70% for arterial samples, and -5.52 to +4.96% for the corresponding venous samples. The effect of this variation on oxygen uptake showed limits of agreement of -43.2 to 36.0 ml/min when the discrepancy between oxygen uptake, calculated using the direct measure of saturation, was compared to that using the derived value. Multiple regression analysis showed that PCO2, temperature and 2,3 diphosphoglycerate were significantly related to saturation discrepancy with an R-squared value of 0.64 (P < 0.0001) for a subgroup of 25 venous samples. The precision of the PO2 electrode was also found to be a major contributory component to the discrepancies, particularly at venous PO2 values. Thus the use of calculated oxygen saturation may result in clinically significant inaccuracies in the assessment of some oxygen flux variables.