Variability in estimating shunt from single pulse oximetry measurements

Area of Science:

• Physiology
• Medical Instrumentation
• Biomedical Engineering

Background:

• Virtual shunt quantifies oxygen loss from alveolar gas to arterial blood.
• Clinicians use pulse oximetry (SpO2) to estimate virtual shunt, but precision varies.
• The non-linear relationship between virtual shunt and SpO2 is often overlooked.

Purpose of the Study:

• To develop a model analyzing virtual shunt estimation from inspired oxygen (FiO2) and SpO2.
• To highlight the impact of the Hb-O2 dissociation curve's non-linearity.
• To incorporate pulse oximeter measurement variability into the model.

Main Methods:

• Utilized basic physiological equations to model virtual shunt estimation.
• Incorporated the non-linear Hb-O2 dissociation curve.
• Accounted for SpO2 measurement variability from pulse oximeters.
• Validated the model with experiments in a hypoxia chamber on healthy subjects.

Main Results:

• Model-derived SpO2 probability distributions were estimated.
• Observed a variable bias (1.2-2.1%) between pulse oximeter brands.
• Tested pulse oximeters remained within manufacturer tolerances.
• Experimental results aligned with model probability distributions.

Conclusions:

• Single SpO2 measurements are insufficient for accurate virtual shunt estimation with pulse oximeters having 2% tolerance.
• Clinical decisions require understanding the limitations of pulse oximetry.
• The non-linearity of the Hb-O2 dissociation curve significantly impacts virtual shunt estimation.