To the Editors:
We would like to comment on the recently published study by Muñoz et al. 1, which analysed accuracy and precision of pulse oximetry at different carbon dioxide arterial tension (Pa,CO2) levels. Although the paper investigates an interesting hypothesis, significant limitations should be addressed.
The most important limitation is that the reported mean Pa,CO2 of 45.3±8.0 mmHg does not exactly reflect hypercapnia. As 45.0 mmHg is an accepted boundary value for Pa,CO2, the conclusions drawn by Muñoz et al. 1 are not adequately supported by their results. While the authors excluded blue nail polish, other colours, such as dark green or black, may have even more pronounced effects 2, 3.
One significant statement Muñoz et al. 1 make is that Pa,CO2 might directly affect pulse oximetry readings. They emphasise that carbohaemoglobin may be the cause of this measurement bias. The sole effect of carbohaemoglobin was neither analysed nor measured in the study by Muñoz et al. 1, nor is their hypothesis supported by any other published study. However, hypercapnia has been proven to have an indirect effect due to vasoconstriction and consecutive alterations of perfusion 4, giving a more likely explanation for the effects observed.
The calculation of bias by Muñoz et al. 1 was not performed as usual or as suggested by Bland and Altman 5: the gold standard should be subtracted from the measured value e.g. bias equalled arterial O2 measured by pulse oximetry (Sp,O2), minus arterial O2 saturation (Sa,O2). For data interpretation, the use of a regression line for the cloud of data points does not seem adequate. Removal of one single data point (∼64 mmHg on the x-axis; -13% on the y-axis; see figs 1a or 2c of 1), might lead to a horizontal regression line similar to the lines in the other figures.
Although pulse oximetry is an established technique used regularly in the clinical setting, newer and improved devices require continued re-evaluation of limitations and influencing factors. We wholeheartedly agree with Muñoz et al. 1 that arterial blood gas analysis remains the gold standard, supported but not replaced by pulse oximetry readings.
Statement of interest
None declared.
- © ERS Journals Ltd
