Racial effects on Masimo pulse oximetry: a laboratory study

Abstract

Recent publications have suggested that pulse oximeters exhibit reduced accuracy in dark-skinned patients during periods of hypoxemia. Masimo SET® (Signal Extraction Technology®) has been designed, calibrated, and validated using nearly equal numbers of dark and light skinned subjects, with the goal of eliminating differences between pulse oximetry saturation (SpO2) and arterial oxygen saturation (SaO2) values due to skin pigmentation. The accuracy concerns reported in dark-skinned patients led us to perform a retrospective analysis of healthy Black and White volunteers. Seventy-five subjects who self-identified as being racially Black or White underwent a desaturation protocol where SaO2 values were decreased from 100 to 70%, while simultaneous SpO2 values were recorded using Masimo RD SET® sensors. Statistical bias (mean difference) and precision (standard deviation of difference) were − 0.20 ± 1.40% for Black and − 0.05 ± 1.35% for White subjects. Plots of SpO2 versus SaO2 show no significant visible differences between races throughout the saturation range from 70 to 100%. Box plots grouped in 1% saturation bins, from 89–96%, and plotted against concomitant SaO2 values, show that occult hypoxemia (SaO2 < 88% when SpO2 = 92–96%) occurred in only 0.2% of White subject data pairs, but not in any Black subjects. There were no clinically significant differences in bias (mean difference of SpO2-SaO2) found between healthy Black and White subjects. Occult hypoxemia was rare and did not occur in Black subjects. Masimo RD SET® can be used with equal assurance in people with dark or light skin. These laboratory results were obtained in well-controlled experimental conditions in healthy volunteers—not reflecting actual clinical conditions/patients.

A review of the effect of skin pigmentation on pulse oximeter accuracy

“…According to regulations, 15% of the participant pool should be darkly pigmented. This requirement was met in (Wong et al 2021, Baker andWilson 2022) (White pool = 65.5%, Black pool 29.6%, White pool = 55.44%, Black pool = 44.56%), but not in other studies (Guber et al 2019, Michael et al 2020, Valbuena et al 2021, Wiles et al 2021, Bangash et al 2022, Wiles et al 2022. Despite this, the inaccuracies seen in Black or darkly skinned subjects mentioned above did not change, although it was difficult to make conclusive statements in studies with a small number of volunteers (Gabrielczyk and Buist 1988).…”

State of the art: Monitoring of the respiratory system during veno-venous extracorporeal membrane oxygenation

“…A few specific points merit highlighting:(1) A systematic overestimation of SpO 2 in hypoxic patients of black skin pigmentation occurs with many SpO 2 monitors 60,61 (although at least one device has been demonstrated to be accurate regardless of skin tone). 62 (2) Haemolysis, which inevitably occurs on ECMO, can generate endogenous carbon monoxide release leading to carboxyhaemoglobinaemia. Indeed, it has been demonstrated that SpO 2 may become less accurate over time during prolonged ECMO runs due to a high prevalence of mild (>3%) carboxyhaemoglobinaemia which should be measured on serial blood gases.…”

“…(1) A systematic overestimation of SpO 2 in hypoxic patients of black skin pigmentation occurs with many SpO 2 monitors 60,61 (although at least one device has been demonstrated to be accurate regardless of skin tone). 62…”

Skin Pigmentation Effects on Pulse Oximetry Accuracy Need a Prospective Study

“…The studies rarely mentioned or assessed other well-known sources of Sp o 2 measurement error and artifact that must be managed to obtain accurate Sp o 2 readings, including ambient light, motion, conditions causing venous pulsations in the tissue bed, electromagnetic interference, poor peripheral perfusion, sensor placement and application to the skin, optical shunting, colored dyes, nail polish, and other hemoglobins (carboxyhemoglobin and methemoglobin). The degree of success in managing many of these issues is device dependent and varies between manufacturers, despite compliance with regulatory standards, as was demonstrated in a recent follow-up investigation 4 . However, one might expect that many of the factors described previously appear as random effects in analysis, with limited effect on results.…”