Uses of pulse oximetry

• Simple, portable "all-in-one" monitor of oxygenation, pulse rate and rhythm regularity, suitable for "field" use.
• As a safe, non-invasive monitor of the cardio-respiratory status of high-dependency patients - in the emergency department, during general and regional anaesthesia, postoperatively and in intensive care. This includes procedures such as endoscopy, where often frail patients are given sedative drugs such as midazolam. Pulse oximeters detect the presence of cyanosis more reliably than even the best doctors when using their clinical judgement.
• During the transport of patients - especially when this is noisy - for example in aircraft, helicopters or ambulances. The audible tone and alarms may not be heard, but if a waveform can be seen together with an acceptable oxygen saturation, this gives a global indication of a patient's cardio-respiratory status.
• To assess the viability of limbs after plastic and orthopaedic surgery and, for example, following vascular grafting, or where there is soft tissue swelling or aortic dissection. As a pulse oximeter requires a pulsatile signal under the sensor, it can detect whether a limb is getting a blood supply.
• As a means of reducing the frequency of blood gas analysis in intensive care patients- especially in paediatric practice where vascular (arterial) access may be more difficult.
• To limit oxygen toxicity in premature neonates supplemental oxygen can be tapered to maintain an oxygen saturation of 90% - thus avoiding the damage to the lungs and retinas of neonates. Although pulse oximeters are calibrated for adult haemoglobin, HbA, the absorption spectra of HbA and HbF are almost identical over the range used in pulse oximetry, so the technique remains reliable in neonates.
• During thoracic anaesthesia - when one lung is being collapsed down - to determine whether oxygenation via the remaining lung is adequate or whether increased concentrations of oxygen must be given.
• Fetal oximetry- a developing technique that uses reflectance oximetry, using LEDs of 735nm and 900nm. The probe is placed over the temple or cheek of the fetus, and needs to be sterile and sterilisable. They are difficult to secure and the readings are variable, for physiological and technical reasons. Hence the trend is more useful than the absolute value.

Limitations of pulse oximetry

• Not a monitor of ventilation A recent case report^(3,4) highlighted the false sense of security provided by pulse oximetry. An elderly woman postoperatively in the recovery room was receiving oxygen by face mask. She became increasingly drowsy, despite having an oxygen saturation of 96%. The reason was that her respiratory rate and minute volume were low due to residual neuromuscular block and sedation, yet she was receiving high concentrations of inspired oxygen, so her oxygen saturation was maintained. She ended up with an arterial carbon dioxide concentration of 280 mmHg (normal 40 mmHg) and was ventilated for 24 hours on intensive care. Thus oximetry gives a good estimation of adequate oxygenation, but no direct information about ventilation, particularly as in this case, when supplemental oxygen is being administered.
• Critically ill patients It may be less effective in very sick patients, because tissue perfusion may be poor and thus the oximeter probe may not detect a pulsatile signal.
• Waveform presence If there is no waveform visible on a pulse oximeter, any percentage saturation values obtained are meaningless.
• Inaccuracies Bright overhead lighting, shivering and motion artefact may give pulsatile waveforms and saturation values when there is no pulse.
  1. Abnormal haemoglobins such as methaemoglobinaemia, for example following overdose of prilocaine, cause readings to tend towards 85%.
  2. Carboxyhaemoglobin, caused by carbon monoxide poisoning, causes saturation values to tend towards 100%. A pulse oximeter is extremely misleading in cases of carbon monoxide poisoning for this reason and should not be used. CO-oximetry is the only available method of estimating the severity of carbon monoxide poisoning.
  3. Dyes and pigments, including nail varnish, may give artificially low values.
  4. Vasoconstriction and hypothermia cause reduced tissue perfusion and failure to register a signal.
  5. Rare cardiac valvular defects such as tricuspid regurgitation cause venous pulsation and therefore venous oxygen saturation is recorded by the oximeter.
  6. Oxygen saturation values less than 70% are inaccurate as there are no control values to compare them to.
  7. Cardiac arrhythmias may interfere with the oximeter picking up the pulsatile signal properly and with calculation of the pulse rate.
  NB. Age, sex, anaemia, jaundice and dark-skin have little or no effects on oximeter function.

• Lag monitor This means that the partial pressure of oxygen can have fallen a great deal before the oxygen saturation starts to fall. If a healthy adult patient is given 100% oxygen to breathe for a few minutes and then ventilation ceases for any reason, several minutes may elapse before the oxygen saturation starts to fall. A pulse oximeter in these circumstances warns of a potentially fatal complication several minutes after it has happened. The pulse oximeter has been described as "a sentry standing at the edge of the cliff of desaturation." because of this fact. The explanation of this lies in the sigmoid shape of the haemoglobin / oxygen dissociation curve (figure 1).

• Response delay due to signal averaging. This means that there is a delay after the actual oxygen saturation starts to drop because the signal is averaged out over 5 to 20 seconds.
• Patient safety there have been one or two case reports of skin burns or pressure damage under the probe because some early probes had a heater unit to ensure adequate skin perfusion. The probe should be correctly sized, and should not exert excessive pressure. Special probes are now available for paediatric use.

  The penumbra effect re-emphasises the importance of correct probe positioning. This effect causes falsely low readings and occurs when the probe is not symmetrically placed, such that the pathlength between the two LEDs and the photodetector is unequal, causing one wavelength to be "overloaded". Repositioning of the probe often leads to sudden improvement in saturation readings. The penumbra effect may be compounded by the presence of variable blood flow through cutaneous pulsatile venules. Note that the waveform may appear normal despite the penumbra effect as it measures predominantly one wavelength only.

Alternatives to pulse oximetry?

• Bench CO-oximetry is the gold standard - and is the classic method by which a pulse oximeter is calibrated. The CO-oximeter calculates the actual concentrations of haemoglobin, deoxyhaemoglobin, carboxyhaemoglobin and methaemoglobin in the sample and hence calculates the actual oxygen saturation. CO-oximeters are much more accurate than pulse oximeters - to within 1%, but they give a 'snapshot' of oxygen saturation, are bulky, expensive and require constant maintenance as well as requiring a sample of arterial blood to be taken.
• Blood gas analysis - requires an invasive sample of arterial blood. It gives the 'full picture', including arterial partial pressure of oxygen and carbon dioxide, arterial pH, actual and standardised base excess and actual and standardised bicarbonate concentrations. Many blood gas analysers report a calculated saturation which is less accurate than that provided by the pulse oximeter. 

Summary Points

• Pulse oximeters give non-invasive estimation of the arterial haemoglobin oxygen saturation.
• Useful in: anaesthesia, recovery, intensive care (including neonatal), patient transport.
• 2 principles involved:
  1. Differential light absorption by haemoglobin and oxyhaemoglobin.
  2. Identification of pulsatile component of signal.
• No direct indication of a patient's ventilation, only of their oxygenation.
• Lag monitor - time delay between potentially hypoxic event such as respiratory obstruction and a pulse oximeter detecting low oxygen saturation.
• Inaccuracies: ambient light; shivering and vasoconstriction; abnormal haemoglobins; and alterations in pulse rate and rhythm.
• Advances in microprocessor have led to improved signal processing.

Further Reading

1. Stoneham MD, Saville GM, Wilson IH. Knowledge about pulse oximetry among medical and nursing staff. Lancet 1994:334:1339-1342.
2. Moyle JTB. Pulse oximetry. Principles and Practice Series. Editors: Hahn CEW and Adams AP. BMJ Publishing, London, 1994.
3. Davidson JAH, Hosie HE. Limitations of pulse oximetry: respiratory insufficiency - a failure of detection. BMJ 1993;307:372-373.
4. Hutton P, Clutton-Brock T. The benefits and pitfalls of pulse oximetry. BMJ 1993;307:457-458. 

---

© World Federation of Societies of Anaesthesiologists WWW implementation by the NDA Web Team, Oxford

---