5. Pulse Oximetry

This variability is compounded by factors such as the type and intensity of ambient light and the patient’s skin pigmentation. In addition, the generation of cyanosis is said to require around 50 g l–1 of deoxygenated haemoglobin in the capillary bed

Oxygenated haemoglobin (HbO2) and deoxygenated haemoglobin (Hb) have differential absorption spectra.
At a wavelength of 660 nm (red light), HbO2 absorbs less than Hb, hence its red colour.
At a wavelength of 940 nm (infrared light), this is reversed and Hb absorbs more than HbO2. At 800 nm – the isobestic point – the absorption coefficients are identical. The pulse oximeter uses two light-emitting diodes which emit pulses of red (660nm) and infrared (980nm) light every 5–10 μs from one side of the probe. The light is transmitted through the tissue to be sensed by a photocell on the other side.
The output is submitted to electronic processing, during which the absorption of the blood at the two different wavelengths is converted to a ratio, which is compared to an algorithm produced from experimental data. Oximetry aims to measure the saturation in arterial blood, and so the instrument detects the points of maximum and minimum absorption (during cardiac systole and diastole). It measures the pulsatile component and subtracts the non-arterial constant component before displaying a pulse waveform and the percentage oxygen saturation. Hence, strictly defined, it is measuring the Sp (plethysmographic) O2 rather than the Sa (arterial) O2.

Pulse oximetry is calibrated against healthy volunteers, which means that calibration against dangerously hypoxic values has not been possible. The instruments are less accurate at SpO2 values below 70%. (You can use this fact to reassure colleagues who are less composed than you in the face of a patient’s saturation that otherwise seems alarmingly low.) Interference by ambient light. This can occur if light is bright and direct, but the pulsed nature of the emissions is intended to allow detection of, and compensation for, any ambient light. Loss of the pulsatile component. This occurs in conditions of hypoperfusion, hypothermia and peripheral vasoconstriction, when there is a narrow pulse pressure, arrhythmias which distort the points of maximum and minimum absorption, or venous congestion. These are all common reasons for a poor signal. Movement artefact or electrical interference (neither are major problems).
Infrared absorption by other substances, such as nail varnish or nicotine staining

with absorption by abnormal haemoglobins and other compounds: — Carboxyhaemoglobinaemia: this is seen in heavy smokers or in CO poisoning. COHb has a similar absorption coefficient to HbO2 and will give an abnormally high SpO2 reading of about 96%. — Jaundice: bilirubin has a similar absorption coefficient to deoxygenated Hb and will give abnormally low saturation readings. — Methaemoglobinaemia: metHb has absorption similar at both wavelengths and gives a saturation reading of around 84%. — Dyes. Various dyes in the circulation such as methylene blue or disulphine blue give falsely low readings.

Pulse oximetry does not detect respiratory failure. A high FiO2 may mask ventilatory failure by ensuring high SpO2% readings despite a rising CO2.
The slope of the oxygen–haemoglobin dissociation curve means that there is a lag of 20 seconds or more between any drop in arterial oxygen tension and the resultant fall in oxygen saturation.
In very anaemic patients SpO2% readings may show high saturations, although oxygen delivery to the tissues may be impaired.
The amplitude of the pulse waveform is not a reliable indicator of the pulse volume. Many instruments automatically augment the trace to fill the display. As a general point of discussion (should there be time to spare), it can be argued that pulse oximetry may not actually be the most useful single monitor, were anaesthetists to be restricted to one. The amplitude of the pulse waveform is not a reliable indicator of the pulse volume. In contrast to end-tidal CO2 measurement, pulse oximetry gives information some of which can be obtained by clinical observation, although clinical recognition of cyanosis is unreliable. The examiner might ask what single, theoretical monitoring device you would use, were you to be allowed only one. An answer might be a device that reliably measured the state of cerebral oxygenation. (If cerebral oxygenation is maintained, then none of the other monitored indices really matter.) One such is the use of near infrared spectroscopy.