Design and validation of pulse oximetry for low saturation.

Abstract

As pulse oximetry migrates into new areas of patient care, it becomes increasingly important to rigorously test performance in the environment of use before claiming success. If evaluated only under tightly or ideally controlled laboratory or bench-top conditions, results may have little relationship to real-world behavior. As an example, fetal pulse oximetry sensors are subject to challenges not typically found with air-breathing patients. The "presenting part" is an attractive site for fetal oximetry, however the presence of resolvable pulses by itself does not ensure accurate readings. This study explores the effects of physiological and mechanical perturbations at low saturation expected during use on fetuses with such sensors. Numerical modeling and animal studies demonstrate that at 40% true arterial saturation (SaO2), calculated saturation values (SpO2) may err by more than 30%, even in the presence of otherwise normal pulse oximetry signals.