Pulse Oximeters Shown to Overestimate Oxygen Readings in Darker Skin Tones

Daniel Martin, OBE

Credit: EBPOM

Readings of arterial hemoglobin oxygen saturation (SaO₂) through pulse oximetry can result in overestimation of true SaO₂ among patients with darker skin tones, according to new findings.¹

This research was conducted to address concerns regarding pulse oximeters’ limitations in different circumstances, as well as the regulatory systems which are set up to govern the devices. The study was led by Daniel Martin, OBE, from University of Plymouth’s Peninsula Medical School in the United Kingdom.

Martin and colleagues acknowledged suspicions that pulse oximeters may overestimate arterial oxygen saturation among individuals with darker skin types, though they added that evidence had been mixed in recent studies.²

“The purpose of this review was to gather and synthesise all evidence to date that describes original research evaluating the effect of ethnicity, skin tone, or both on the accuracy of pulse oximetry to estimate SaO₂,” Martin and colleagues wrote. “In doing so, the aim was to determine whether there is a robust body of evidence to support the suggestion that pulse oximetry overestimates SaO₂ in people with darker skin tone.”

Background and Methods

The investigators conducted a systematic review, with criteria for study inclusion being studies which involved human subjects and covered original research that compared pulse oximetry readings to reference standard for the purposes of assessing accuracy. The index test was pulse oximetry-derived oxygen saturation (SpO₂), and the reference standard was decided to be SaO₂ measured through the use of arterial blood gas (ABG) co-oximetry.

The research team used both prospective and retrospective designs. The team additionally considered research designed to evaluate pulse oximeter accuracy based on skin tones or patient ethnicity.

Criteria for exclusion of research included studies that looked at animals or did not involve human subjects as well as research that had lacked a reference standard deemed acceptable. The investigators did not apply geographic or setting restrictions to the eligibility of studies.

The research team implemented the MEDLINE, Embase, CINAHL (EBSCO), and Web of Science databases from their inception until March 2023 for their searches. The team sought studies from the reference lists of selected articles and then looked at review articles.

The investigators recorded study subjects’ ethnicities and skin tones. Their primary outcome assessed over the course of the study was the tested pulse oximeters’ accuracy, and this would then be assessed through many different metrics such as precision, bias, and overall accuracy.

The research team also considered ‘occult hypoxemia,’ a case in which true hypoxaemia remains undetected through pulse oximetry SpO₂ owing to an overestimation of SaO₂. Other data that the team looked at included pulse oximeter details, settings for studies, demographics of subjects, and reference standard type.

Findings

Overall, the investigators ended up with 44 studies in total, as well as 222,644 participants and 733,722 paired SpO2-SaO₂ measurements. There was wide variation in the study methodologies, many of which were prospective clinical, laboratory, and retrospective clinical research.

The research team found that 64% of the studies they looked at had exhibited a high risk of bias, reporting that there was notable diversity in data analysis, research design, and reporting standards. Only 25% of the included research had looked into skin tone, and among only 1.1% of the participants.

The majority of the studies reported ethnicity of subjects, and the team reached the conclusion that 31.0% of subjects had been non-White ethnicity or had non-light skin tones. Most of the research the investigators looked at suggested an overestimation of SaO₂ thanks to pulse oximetry in subjects with darker skin tones or those of ethnicities presumed to have darker skin tones.

Despite these findings, some still were shown not to have such inaccuracies. Any meta-analyses of the investigators’ data were not deemed to be feasible.

“...(During) the COVID pandemic, and to some extent since, it was necessary to put thresholds in place which meant that people were only admitted to hospital if their levels fell to a certain point,” Martin said in a statement. “If those levels are being overestimated – so, for example, if a device is telling someone their oxygen saturation is 98% whereas it is in fact significantly lower – it could realistically mean people are missing out on treatments they need."³

References

1. ^{Martin D, Johns C, Sorrell L, et al. Effect of skin tone on the accuracy of the estimation of arterial oxygen saturation by pulse oximetry: a systematic review. Br J Anaesth. Published online February 17, 2024. doi:10.1016/j.bja.2024.01.023.}
2. ^{Improving pulse oximetry accuracy in dark-skinned patients: technical aspects and current regulations. Br J Anaesth. 2023; 131: 640-644.}
3. ^{Detailed study demonstrates how pulse oximeters significantly overestimate oxygen readings in people with darker skin tones. British Journal of Anaesthesia. February 29, 2024. https://www.eurekalert.org/news-releases/1036004. Date accessed: March 5, 2024.}