Key Points:
- Small retrospective studies of critically ill adults have shown pulse oximetry levels different by race
- No large prospective study has shown possible difference in pulse oximetry among darker skinned individuals
- The EquiOx study found that among critically ill adults, there were higher rates of overestimated pulse oximetry among darker skinned individuals
- This is the first large study to demonstrate findings of pulse oximetry bias, and future studies may benefit from expansion to other relevant populations
Previous retrospective studies of critically ill adults have demonstrated the pulse oximetry levels differ by race. However, previous studies investigating controlled desaturation among healthy participants have been too small to show possible differences in pulse oximetry among individuals those with darker skin pigments. There has not yet been a large-enough prospective real-world study investigating the relationship between skin pigment and pulse oximeter bias.
The EquiOx study is a single-center FDA-funded real-world observational study that enrolled 635 adults with arterial lines from a mixed ICU population at a universal-affiliated safety-net hospital. Research staff observed paired co-oximeter (SaO2) and pulse oximeter (SpO2) and noted probe position, stability, and patient movement. Skin pigment was measured using the standardized Monk Skin Tone Scale and individiual typology angle (ITA) with a spectrophotometer. This study aimed to test for difference in pulse oximeter bias (SpO2 – SaO2) between skin pigment categories after adjusting for possible confounders, including perfusion index (PI), SaO2, and vasopressor administration.
Participants were commonly noted to have chronic medical comorbidities, including hypertension (63%), systolic heart failure (22%), and diabetes (32%). By demographics, 25% identified as White, 21% Black, 20% Asian, and 20% Hispanic. Skin pigment by ITA was categorized as 32% light (ITA > 30), 54% medium (ITA 30 to –30), and 14% dark (ITA < -30).
Although pulse oximeter bias on average was negative for all people, it was less negative in the darkly pigmented people than in the people with lighter pigment, meaning that pulse oximeters do not perform the same across different skin pigment categories. However, 20% of observations showed a positive bias in darker skin pigments, showing that pulse oximetry was found to slightly overestimate true arterial saturation, and ultimately demonstrating a less negative pulse oximetry bias in darker skin pigment individuals. In addition, the study showed that pulse oximetry performance is worse in critically ill adults compared to controlled studies: performance measured by ARMS was 3.9 in the full cohort, whereas the FDA recommends ARMS < 3%.
The authors acknowledge that the study may have limitations including prospective study design and equipment differences. However, the EquiOx study is the largest prospective study of pulse oximeter bias across skin pigmentation, demonstrating underestimation of arterial oxygen saturation concentrated among darker skinned individuals. Future studies would benefit from expanding to patients with stable hypoxemia, analyzing even darker skinned individuals, and possibly pediatric patients.