Functional versus fractional oxygen saturation readings: bias and agreement using simulated solutions and adult blood

The purpose of this study was to examine the bias and agreement between functional oxygen saturation (SO2) and fractional oxyhemoglobin (HbO2) using simulated quality control (QC) solutions and adult blood. Using a hemoximeter, 5 analysts performed at least 5 tests each on QC solutions with 3 different hemoglobin (Hb) levels and on adult blood samples of various oxygen saturation levels representing venous or arterial samples. Bias and the limits of agreement were determined using the technique of Bland and Altman. Using QC solutions with low, normal, and high Hb levels, the bias for SO2 against HbO2 was 20.82 +/- 0.50 (n = 66), 19.14 +/- 0.56 (n = 81), and 19.59 +/- 0.43 (n = 79), respectively, with SO2 reading consistently higher. The correlation between SO2 and HbO2 was -0.49, -0.69, and -0.68, respectively. Using adult blood, the bias for SO2 against HbO2 was 1.29 +/- 0.48 for venous samples (n = 62) and 1.9 +/- 0.19 for fully oxygenated samples (n = 36), and the correlation between SO2 and HbO2 was 1.0 and 0.68, respectively. These findings suggest that the consistency between the measurements of SO2 and HbO2 may be dependent on hemoglobin levels and oxygenation status. Thus, caution is warranted when assuming that the measurements of SO2 and HbO2 are interchangeable.