Assessment of accuracy of two pulse oximeters in infants with cyanotic and acyanotic congenital heart diseases.

Abstract

Peripherally measured oxygen saturation (SpO2) may often differ from arterial oxygen saturation (SaO2), measured by co-oximetry, especially within the lower range of oxygen saturations. This can potentially impact clinical decisions and therapy in children with congenital heart disease, as critical hypoxemia might remain unnoticed. Our aim was to investigate the accuracy of two different pulse oximeters compared to SaO2 in infants with congenital heart diseases. Simultaneous recordings of SpO2, measured by two different pulse oximeters (Philips IntelliVue X3 Monitor and Nellcor™ OxiMax™), were compared to SaO2 obtained by arterial blood gas analysis. A total of 153 measurements were performed in 44 infants with arterial oxygen saturation between 70 and 100%. We divided the measurements into 3 subgroups: group 1-SaO2 70.0%-85.0%, group 2-SaO2 85.1%-94.0%, group 3-SaO2 >94.1%. For Philipps, the median bias was 5.3 (IQR: 2.6-8.7) %, 2.3 (IQR: 0.9-6.0) % and 1.1 (IQR: -0.8-2.4) % in group 1, 2 and 3, respectively. For OxiMax™, the median bias was 2.7 (IQR: 0.5-5.1) %, 0.2 (IQR: -0.9-2.6) % and -0.5 (IQR: -1.3-0.6) % in group 1, 2 and 3, respectively. Regarding the accuracy of these oximeters, as evaluated with the Accuracy root mean squared index (Arms), it was 9.8 versus 4.5% in group 1, 4.5 versus 2.9% in group 2 and 2.4 versus 1.9% in group 3 for Philipps and OxiMax™, respectively. In lower range saturations between 70% and 85% the accuracy of both pulse oximeters exceeded the threshold of ≤3% recommended by the Food and Drug Administration (FDA). Therefore, peripheral pulse oximetry within the lower range of oxygen saturations should be interpreted with caution in infants with congenital heart diseases, taking into consideration its limitations. Direct co-oximetry should be the preferred method to support clinical decisions in children with cyanotic congenital heart diseases.