J Prim Care Community Health. 2026 Jan-Dec;17:21501319261485526. doi: 10.1177/21501319261485526. Epub 2026 Aug 28.
ABSTRACT
ObjectivesPulse-oximeter oxygen saturation (SpO2) is widely used in outpatient respiratory care because it is fast and noninvasive, but arterial blood gas (ABG) testing provides a more direct measure of arterial oxygen saturation (SaO2). Prior studies show that pulse oximeters may overestimate oxygen saturation in individuals with darker skin tones, potentially leading to missed recognition of clinically meaningful hypoxemia. However, limited research has examined measurement agreement in outpatient occupational lung disease settings.MethodsA retrospective agreement analysis was conducted using paired ABG-derived SaO2 and pulse-oximeter SpO2 measurements from the same outpatient visit at a rural West Virginia Black Lung Clinic. Paired measurements were defined as SaO2 and SpO2 readings recorded during the same clinical encounter. Agreement was evaluated using the Bland-Altman approach, which quantifies the average difference between two methods (mean bias) and the range within which most individual differences are expected to fall (limits of agreement). Hidden hypoxemia was defined as SaO2 ≤ 88% with concurrent SpO2 ≥ 90% or ≥ 92%.ResultsThirty-five participants were included; all male and self-identified as African American. Mean age was 70.7 years (SD 8.9). At rest, pulse oximetry overestimated SaO2 by an average of 1.29 percentage points (SD 3.74), with 95% limits of agreement from -6.04 to +8.62 percentage points. Following exercise, mean bias was +0.75 percentage points (SD 3.68), with limits of agreement from -6.46 to +7.96. Although average differences were modest, individual discrepancies were sometimes several percentage points higher or lower than ABG values. ABG-defined hypoxemia occurred in 1/35 participants at rest and 1/34 post-exercise; in both cases, SpO2 remained ≥ 90% and ≥ 92%, consistent with hidden hypoxemia. Regression analysis did not demonstrate statistically significant proportional bias at either timepoint.ConclusionsIn this outpatient Black Lung Clinic sample, pulse oximetry averaged only modest overestimation of SaO2, but individual readings varied widely. This matters because oxygen and disability-related decisions are made one patient at a time, and differences of several percentage points can change classification near common thresholds. These findings support cautious interpretation and use of pulse-ox values in outpatient occupational lung disease care and justify larger, multi-site studies.
PMID:42664382 | DOI:10.1177/21501319261485526
