Neonatal pulse oximetry: accuracy and reliability

Pulse oximetry has gained widespread use in neonatal oxygen monitoring. However, because specific morbidity in neonates has been related to abnormal levels of oxygen partial pressure (PaO2), it is essential to demonstrate that pulse oxygen saturation values (SpO2) are highly accurate and reflect with precision the simultaneous PaO2. In this report, data is presented that describe the accuracy of the Ohmeda Biox 3700 pulse oximeter. SpO2 was highly correlated with arterial blood oxygen saturation (r = .99) measured on arterial (catheter) blood using a two-wavelength Radiometer OSM2 Hemoximeter. When compared with PaO2, SpO2 values of 92% +/- 3% (mean +/- range) excluded all of the measured PaO2 values less than 45 mm Hg and greater than 100 mm Hg. SpO2-tcPO2 (transcutaneous PO2) and SpO2-PaO2 (catheter or percutaneous arterial PO2) correlations showed that bronchopulmonary dysplasia, percutaneous arterial sampling, and nipple feeding skewed the tcPO2 but not the PaO2 correlations with SpO2, indicating that SpO2 is not sensitive to peripheral factors that affect tcPO2. SpO2-PaO2 correlation was not affected by gestational age. A high degree of accuracy of SpO2 values and SpO2 prediction of normal appearing PaO2 values are defined by these results.