The wavelength dependence of pulse oximetry

Abstract

Although the `AC/DC' ratios (except in transmission below 600nm) have the same basic form as the HbO, spectra (as would be expected from pulse oximeter theory since the arterial blood in a healthy subject is mainly HbO,) the quantitative agreement, using simple models, is very poor. In addition, although the graphs above 600 nm in reflection and transmission modes are similar they are clearly not the same, demonstrating that there is indeed a difference between the 2 modes of operation. It is interesting to note that the ratios of the `AC/DC' ratios at 660 nm and 940 nm (at which many commercial devices operate) are very similar (approximately 0.5) in both reflection and transmission modes. This means that using the same pulse oximeter with either reflection or transmission probes will produce reliable results. However, if wavelengths other than 660 and 940 nm were employed the authors' figure suggests that this would not longer be so. A further finding of significance to the operation of commercial pulse oximeters was that the `AC' signal in reflection mode was at its lowest, and prone to noise, at 660 nm. Consequently, the worst `red' wavelength to use in a reflection pulse oximeter would be 660 nm. Finally, these results demonstrate that it is possible to produce `AC/DC' ratios below 600 nm, that is in the visible part of the spectrum which has been largely ignored for pulse oximetry. Although the high absorption of light by blood in this part of the spectrum may cause some problems it is clear that pulse oximeters operating in the visible could be constructed. Potential advantages are that: the actual `AC/DC' ratios are larger; there is more `structure' to the absorption spectra of HbO, which could be used for verification of data; and that other haemoglobin compounds also possess distinctive spectra in this region which should allow their presence to be monitored. (3 pages)