Skin Pigmentation and Accuracy of Pulse Oximetry Values

First introduced in the 1970s, pulse oximetry has been a mainstay in patient monitoring and respiratory management for more than 50 years.1 Using sensors placed on the fingertip, earlobe, or forehead, this simple, low-cost noninvasive technology uses the principle of photoplethysmography. With photoplethysmography, the intensity of light traveling through tissue is modulated by the absorption of light by pulsatile blood.2 Oxygenated and deoxygenated hemoglobin absorb different amounts of certain wavelengths of light. Specifically, oxygenated hemoglobin absorbs more infrared light whereas deoxygenated hemoglobin absorbs more red light. Pulse oximeters calculate oxygen saturation values by comparing the ratio of red to infrared light absorption in arterial blood against manufacturers’ empirically derived calibration curves mapped to known arterial oxygen saturation (Sao2) values.1,3While the COVID-19 pandemic broadened the use of pulse oximetry to assess blood oxygenation rapidly in hospital and home settings, the pandemic also further heightened concerns about the accuracy of pulse oximetry in patients with darker skin pigmentation. Thus, the PICO (patient/problem, intervention, comparison, outcome) question for this evidence synthesis was “Do differences in skin pigmentation affect detection of hypoxemia by pulse oximetry, initiation of treatment, and disease outcomes?”The strategy included searching PubMed and CINAHL. Key words included critical care, skin pigmentation, oxygen saturation, pulse oximetry, and accuracy. The search was limited to original research from the past 5 years.Table 1 outlines findings of 10 studies. Of these, 8 were retrospective cohort studies, 1 was a cross-sectional study, and 1 was a systematic review. All studies showed that pulse oximetry overestimated oxygenation in patients with darker skin pigmentation. Rates of occult hypoxemia, defined as Sao2 less than 88% when oxygen saturation shown by pulse oximetry (Spo2) is 92% to 96%, increased as Sao2 levels declined, indicating that pulse oximetry is less accurate at lower oxygen saturation levels. One study also showed higher respiration rates among dark-skinned persons.13 Sjoding et al4 noted a nearly 3 times higher frequency of occult hypoxemia in Black persons. Additionally, Wong and colleagues5 found that Hispanic, Black, and Asian patients were significantly less likely to receive arterial blood gas measurement (1.9%, 2.8%, and 3.4%, respectively) compared with White patients (5.6%).In relation to treatment, a large cohort study of more than 43 000 patients revealed a significantly lower likelihood of hospital admission for dark-skinned persons.7 Fawzy et al9 reported that more than half of the 23% of patients whose treatment eligibility was never recognized were Black. Delays in recognition of treatment eligibility ranged from 5 hours (in Hispanic and White patients) to more than 7 hours in Asian and Black patients.9 Once patients were admitted, delivery of supplemental oxygen7,8 and treatment with dexamethasone7 were used less often and after longer delays in Black than White patients.7 Hypoxemia was also associated with higher rates of organ dysfunction and in-hospital mortality in one large study.5 This body of evidence suggests that the lower accuracy of pulse oximetry in dark-skinned persons may unintentionally contribute to health disparities, threatening health equity for access to supplemental oxygen or more intensive support.4,6,8,14 As Sjoding et al4 noted, relying on pulse oximetry for triage and to adjust supplemental oxygen in dark-skinned persons placed those persons at increased risk for hypoxemia.The best available evidence (levels A-C, Table 2) indicated that pulse oximetry overestimated arterial oxygenation with higher rates of hidden hypoxemia in dark-skinned persons. These findings may be related to melanin in the outer skin layer absorbing more red light—a variable that is not controlled for in current pulse oximetry technology.1,3 These studies, however, have major limitations, namely, their retrospective nature and their reliance on self-report of racial background as an indicator of degree of skin pigmentation. Future research must move to prospective, controlled designs comparing blood sampling and pulse oximeter readings in near real time as Sao2 is prone to fluctuations over short periods of time.3,16 Additionally, these of Spo2 bias. Objective skin color assessment with instruments like the Fitzpatrick scale or the Munsell chart will ensure better distribution for adequate sample representation of individuals representing a full range of skin pigmentation.3,4,12,16 Beyond skin pigmentation, other factors can affect the discordance between Sao2 and Spo2 levels, including dark nail polish, finger size, temperature, variations in breathing, anemia, low perfusion, tissue edema, excessive motion, and incorrect sensor placement. These factors must be measured and accounted for in real-world pulse oximetry studies.3,12,16–19Indeed, the US Food and Drug Administration20 issued a safety communication about the limitations of pulse oximetry in February 2021 and later convened a Medical Device Advisory Committee. In this report, the committee stated that the sensitivity, specificity, and positive and negative predictive values of pulse oximeters are not well described.18 As a result, the committee will investigate methods to improve the equitable accuracy of pulse oximeters. Once the mechanism involved in Spo2 overestimation is identified, effective design of improved pulse oximeters will require collaboration among device manufacturers, clinicians, and regulators. New-generation devices may involve calibrations based on skin tone (rather than race-based adjustments, which can contribute to health inequities) or sensors that can objectively assess variables that can alter Spo2 levels such as skin pigmentation and perfusion index.1,17 Such noninvasive tech nologies that incorporate skin tone measurements are currently being tested.21 New manufacturing standards are also imperative. Pulse oximeters must be shown to be accurate for a variety of specific skin tones. Accuracy of pulse oximetry for a specific skin tone cannot be determined using data from broad racial samples that include lighter and darker skinned persons.16,17Until improvements in the physical design of pulse oximeter technology become available, some researchers have suggested parameters for pulse oximeter levels. Chesley et al10 recommended targeting an Spo2 goal between 94% and 98% for all patients to minimize occult hypoxemia and limit hyperoxia (Pao2 > 110 mm Hg). Wong et al5 suggested specific Spo2 goals of 92% for Hispanic patients, 93% for White and Asian patients, and 96% for Black patients to reduce the risk of occult hypoxemia to less than 10%. Using parameters such as these can help ensure judicious use of supplemental oxygen, lessening the risk of hyperoxia and its negative effects.3Gallagher19 reminds us of many challenging clinical conditions that affect the accuracy of pulse oximetry. For instance, vasodilatation may result in false low Spo2 readings due to venous pulsation in capillary beds, whereas low perfusion or vasoconstriction can reduce the strength and quality of the pleth waveform and thus Spo2 accuracy. Dysfunctional hemoglobin such as carboxyhemoglobin with smoke inhalation as well as anemia can also affect Spo2 accuracy. Therefore, correlation of clinical findings with Spo2 readings is key for patient management. Pulse oximetry is one tool to use in conjunction with other clinical data, such as vital signs, skin color, dyspnea, and mentation, to reduce possible unintentional bias while caring for patients. When in doubt, rely on clinical judgment.19