A real-world study, conducted at Zuckerberg San Francisco General Hospital, demonstrated that pulse oximetry measurements of oxygen saturation are often lower than arterial oxygen saturation, with the magnitude of this discrepancy varying by skin pigmentation.

The EquiOx study, presented by Carolyn Hendrickson, MD, at the American College of Cardiology Annual Scientific Session in Chicago, aimed to assess pulse oximeter accuracy among 631 critically ill patients receiving care in the intensive care unit between 2022 and 2024. Mean patient age was 62. Skin pigment was considered light in 33%, medium in 53% and dark in 14%. 

Simultaneous SpO2 and SaO2 measurements were obtained, while skin pigmentation was evaluated using both the Monk Skin Tone Scale and spectrophotometric melanin assessments.

Key Findings

• Median pulse oximeter bias was negative across all skin pigment categories, with an overall median difference of -1.7%.

• The accuracy root mean square error was 3.9%.

• Bias was less negative among individuals with darker skin pigmentation.

• Adjusted bias attributed to dark versus light skin pigmentation, based on objective colorimetry, was 0.5%.

• Of observations, 20% demonstrated a positive bias, with higher likelihoods noted among patients with darker skin pigmentation, Black race, diabetes, hypertension, peripheral vascular disease, tobacco use, hypoxemia, hypothermia, and low perfusion index.

Interpretation

The study confirms that pulse oximetry accuracy is influenced by skin pigmentation, with darker skin pigmentation associated with a less negative bias. However, when SpO2 readings exceed SaO2 values, positive bias was more frequently observed in individuals with darker skin. This discrepancy raises concerns about the reliability of pulse oximetry in diverse patient populations.

Conflicts of interest were not reported.

Source: ACC.25