Prior to technological advances such as pulse oximetry, detection of hypoxaemia was reliant upon observation methods such as cyanosis. While cyanosis remains an important sign, it is not always reliable, with multiple factors such as skin pigmentation, lighting, and inter-observer differences affecting detection. Cyanosis is also often a late sign so having access to a device that can provide early recognition of hypoxia is a useful tool. Pulse oximetry does have its limitations but knowing how the technology works and understanding the limitations of its use can reduce errors.
Pulse oximetry has been considered the fifth vital sign for in-hospital care for nearly 30 years (Tierney et al, 1997) and its use has become the standard of care in developed emergency medical services across the world. Its role is to measure the oxygen saturation in the blood – enabling clinicians to detect hypoxia – as well as the response of a patient to oxygen therapy.
Pulse oximetry
Pulse oximetry is based on the principle that oxyhaemoglobin (O2Hb) and deoxyhaemoglobin (HHb) absorb red and near-infrared (NIR) light differently. O2Hb absorbs greater amounts of infrared (IR) light and lower amounts of red light than does HHb; HHb, on the other hand, absorbs more red light. The fact that it is red and NIR light is fortuitous as these two wavelengths just happen to penetrate tissues well whereas far-IR (FIR), green, blue and yellow are all more highly absorbed by non-vascular tissues and water (Sinex, 1999; Chan et al, 2013). Pulse oximeters emit two wavelengths of light (red and NIR) from a small light-emitting diode located in one arm of the probe. The light is transmitted through the finger (usually) and then detected on the opposite arm of the probe. The oximeter uses the different absorption properties of O2Hb and HHb to determine the proportion of haemoglobin that is bound to oxygen (Chan et al, 2013).
There are numerous pulse oximetry products available within healthcare settings, and a range of pulse oximetry probes that are designed for differing areas of the body including the finger, forehead, nose and the earlobe. These sites tend to be used owing to the relatively high vascular density compared with other areas of the skin such as on the chest wall (note: sensors for the forehead use reflectance technology to measure the SpO2). Each practice area will have a preferred pulse oximeter manufacturer and probe configuration; however, the principles remain the same regardless of make and model.
Indications for pulse oximetry
Pulse oximetry forms part of the National Early Warning Scores so should be routinely applied in prehospital emergency care. Specific indications for pulse oximetry include:
- Respiratory disorders such as chronic obstructive pulmonary disease (COPD), asthma or acute respiratory infection
- To enable titration of oxygen to meet current guidelines
- To assess effectiveness of interventions (e.g. oxygen, bronchodilators)
- To gain baseline oxygen saturations prior to a procedure; for example, sedation and anaesthesia.
Limitations of pulse oximetry
The use of a pulse oximeter is not without limitations and sources of error. A variety of potential limitations have been identified and need to be considered when assessing oxygen saturations through these methods.
Skin pigmentation
There is evidence that pulse oximetry consistently overestimates oxygen saturations in adults with hypoxaemia who have high levels of skin pigmentation (Jamali et al, 2022; Shi et al, 2022). There is no correction factor that would allow clinicians to account for the pulse oximeter bias and, likewise, there is no single correction factor that could be applied based on skin-tone. This leads to obvious issues for clinicians and for which there is no simple answer. The risk of failing to recognise hypoxia is likely to be higher in people with darker skin, but simply treating for suspected hypoxaemia creates the risk of hyperoxia and its consequences. Given that these issues have been known about for decades, it is concerning that no medical technology company has yet found a solution!
Nail polish
There has been growing concern over the validity of SpO2 measurements achieved in patients who are wearing nail polish. Common practice as a result has been to remove nail polish prior to measurement, which obviously has time and expense implications. A number of large-scale studies on healthy and sick people have noted that differences do occur with dark colour nail polishes such as black, brown and blue, though these changes may not be clinically significant. A recently published systematic review by Aggarwal et al (2023) captured 21 studies on the effects of nail polish on pulse oximetry readings. They found that the application of fingernail polish can marginally reduce SpO2 readings in groups of patients, and that in a very few cases, it may not be possible to measure SpO2 from the painted nails at all. The dark blue, brown and black nail polishes were associated with the maximum mean reductions in SpO2, but even these did not appear to be clinically important. In conclusion, Aggarwal et al (2023:1279) stated that ‘based on our analysis, we do not recommend routine removal of fingernail polish prior to pulse oximetry’.
Carbon monoxide poisoning
Carbon monoxide (CO) has a higher affinity for haemoglobin than does oxygen, thus the presence of CO will result in the formation of carboxyhaemoglobin (COHb). COHb and O2Hb have similar absorption properties for red light and similarly low absorption rates of NIR light such that a pulse oximeter that emits only red and NIR light cannot differentiate between O2Hb and COHb (Chan et al, 2013). It is likely that many pulse oximeters used in paramedic practice are unreliable in suspected CO poisoning and readings should be treated with caution.
Sickle cell disease
Sickle cell disease is complex, and the pathophysiology is beyond the remit of this article. What is important to note is that hypoxaemia and hypoxia are central elements of the disease pathophysiology, and disease-related morbidity and mortality. The accurate assessment of the patient's oxygen saturation is key to management and inaccuracies could have significant and detrimental consequences for the patient. Fortunately, numerous studies have been conducted on adults and children in different settings to assess how well SpO2 correlates with fractional oxyhaemoglobin (FO2Hb), which is the most accurate reflection of oxygenation status (Lucas et al, 2023)s. Lucas et al collated and reviewed the evidence and provide the following useful summary:
- SpO2 unanimously overestimates FO2Hb as dyshaemoglobins are not accounted for. Dyshaemoglobin is where the haemoglobin molecule has been functionally altered and is therefore unable to carry oxygen
- The accuracy of pulse oximetry is generally limited in critically ill patients, and acute phases/acute exacerbations are frequent in sickle cell disease
- Pulse oximetry measurements have repeatedly been shown to provide falsely higher SpO2 in. As sickle cell disease frequently affects black individuals, this represents an important additional source of bias.
Using pulse oximetry may provide false reassurance of the patient's oxygenation status, so caution needs to be applied when interpreting the readings.
Poor peripheral circulation
The pulse oximeter is reliant upon pulse waves so poor perfusion owing to conditions such as cold or hypotension could result in intermittent drop-outs or an ability of the oximeter to read SpO2. Poorzargar et al (2022) published a systematic review comparing SpO2 values recorded by different pulse oximeters to arterial blood gas-derived values in patients with poor perfusion. It seems apparent that the ability of the oximeter to perform accurately in patients with poor perfusion is down to the equipment. Traditional oximeters struggle to eliminate venous pulsations, which result in inaccurate recordings; whereas some of the more recent oximeters use different algorithms to calculate the figure and can better account for artefact. If you have a more modern pulse oximeter with updated algorithms, it is likely to be more accurate than older, more traditional models.
The Poorzargar review suggested that earlobe sensors were more sensitive and had greater accuracy than fingertip sensors in patients with poor peripheral perfusion; however, this was not consistent across all the studies they reviewed. This is unhelpful but is the reality of research in the area.
Motion artefact
The excessive motion of fingers from tremor, seizure or shivering can interfere with signal detection or interpretation (Petterson et al, 2007). However, manufacturers have been working on motion-resistant technologies and claim many breakthroughs. There still appears to be work needed to fully eliminate motion artefact, as studies still highlight variable performances under motion artefact conditions (Ganesh Kumar et al, 2022; Giuliana et al, 2023). Of additional importance to many paramedics is the impact of transportation, which could also induce artefact.
As there remains uncertainty in the effectiveness of new technologies for motion artefact, it is advised that placement of the probe is still important, alongside calming of the patient wherever possible.
Different lighting
Different lighting has been shown in very old studies to have impacted on readings owing to the infiltration of light into the probe (Brooks et al, 1984; Amar et al, 1989). In the Amar study, saturations displayed on the oximeter dropped significantly when a handheld fluorescent light was used to observe the patients. This is unlikely to be a factor in paramedic practice so evidence from other studies may be more helpful. In a more recent (but still old) study by Fluck et al (2003), pulse oximetry readings were taken under five different lighting conditions; however, no statistically significant differences were found. The authors argued that the differences were so small that even had they been statistically significant, they would have been clinically unimportant.
Inconsistent pulse
Oximeters require a steady pulse signal; therefore, conditions that affect the consistency of a pulse may reduce accuracy. A common example is the presence of cardiac arrhythmias such as atrial fibrillation.
Perfusion index
The perfusion index (PI) is defined as the ratio of the pulse wave of the pulsatile portion (arteries) to the non-pulsatile portion (venous and other tissues) (Sun et al, 2024). It represents a continuous and non-invasive measurement of peripheral perfusion and provides useful information when selecting a monitoring site. In a patient where peripheral perfusion falls below the minimum required for tissue oxygenation, the PI alerts the clinician to consider another monitoring site. The best monitoring sites are those with a relatively high and stable PI. It is notable that the PI is highly individual and varies according to multiple factors.
Typically, the middle finger is used in clinical trials and has been advocated as the optimal point for clinical measurements. However, this is not yet conclusive and further studies are required to validate measurements taken at different sites.
Other factors
High bilirubin levels in hepatitis and cirrhosis of the liver and some blood dyes used in angiography may indirectly or directly affect pulse oximetry
Conclusion
Pulse oximetry is one of the key baseline measurements and is considered to be the fifth vital sign. While it provides valuable information on the patient's condition, there are many factors that can impact on the accuracy of the reading and care should be taken to ensure that the readings are interpreted carefully in light of those limitations. Of particular concern is the impact of skin pigmentation on the accuracy of the results and the risk that hypoxia may be missed.