Pulse oximeter performance during desaturation and resaturation: a comparison of seven models

To compare pulse oximeter performance during induced hypoxemia.

Prospective investigation in human volunteers.

Laboratory facility at a university medical center.

8 unanesthetized, healthy ASA physical status I volunteers.

We evaluated the accuracy and response times of seven popular pulse oximeters during induced hypoxemia. Arterial blood fractional oxygen saturation (SaO2) measurements were performed simultaneously and considered a gold standard.

All oximeters were accurate (+/-2%) while subjects were breathing room air. During maximal hypoxemia (induced by breathing a FIO2 = 10% in nitrogen), large differences were noted between oxygen saturation as measured by pulse oximetry (SpO2) and SaO2 values, with pulse oximeters consistently underreporting SpO2 when actual SaO2 values were 75% or less. The Ohmeda 3740 (Ohmeda, Boulder, CO) using an ear probe was the first to detect desaturation (change in SpO2 > 3%) in 4 of 8 subjects (p < 0.05), and the Nellcor N200 reflectance oximeter (Nellcor, Inc., Pleasanton, CA) was first in 3 of 8 subjects (p < 0.05). During resaturation (after administering 100% oxygen), the Novametrix Oxypleth (Novametrix, Wallingford, CT) was significantly faster than other oximeters (p < 0.05) to return to baseline (SpO2 = 98%).

Most models of oximeters tested performed well when hemoglobin oxygen saturation was high, but all were inaccurate when SaO2 was approximately 75%. During induced hypoxemia, there were significant differences in the response times of oximeters tested, with no model demonstrably superior to others in all measures of performance.