Prediction of arterial partial pressure of oxygen with pulse oxygen saturation measurements

Objectives: We studied 22 neonates with gestational ages from 26 to 40 weeks to determine how accurately pulse oxygen saturation (SpO(2)) could predict arterial partial pressure of oxygen (PaO2) and how much time and effort would be required to achieve and maintain a desired SpO(2) value. Study design: SpO(2) was maintained at 90%, 92%, 94%, 96%, and 98% by adjustment of FIO2. PaO2 was correlated with SpO(2) by nonlinear mixed effects regression analysis. Results: PaO2 (y) correlated positively with SpO(2) (x): y = (0.03) . e(0.08(x)). Variation of PaO2 increased from a range of +/-10 Torr at SpO(2) = 90% to +/- 28 Torr at SpO(2) = 98%, principally as a result of decreasing hemoglobin-oxygen affinity, with no effect of advancing gestational age, subject temperature, pH, or PaCO2. Considerable time (18 +/- 14 SD minutes) and effort (5 +/- 3 SD FIO2 adjustments/min) were required to achieve and maintain a desired SpO(2). Conclusions: PaO2 varied broadly with SpO(2) in a clinically acceptable range of SpO(2) (90% to 98%), and the variability increased at higher SpO(2) values. Considerable time and effort were required to achieve and maintain desired SpO(2) values.