Noninvasive estimation of the arterial partial pressure of oxygen in NICU infants

The Oxygen-hemoglobin dissociation curve can be used to calculate the pressure of arterial oxygen (PaO₂) from peripheral oxygen saturation (SpO₂) in adults. Our objective was to evaluate whether this calculation could be performed in newborn babies. We obtained blood gas measurements from three infant groups, two from the University of Iowa Hospitals and Clinics (UIHC) neonatal intensive care unit (NICU) and one from Harvard Medical school. We plotted PaO₂ against SaO₂ and compared the resultant curve to published curves for adults and newborns. UIHC infants were more likely to store oxygen on hemoglobin when compared to published curves for newborns and adults. We tested the additional hypothesis that this was because of the large proportion of premature infants in our cohort, but age did not predict the shift.

Next, we calculated how much blood acidity, and infant age contributed as sources of error in using this curve to estimate PaO₂ from SpO₂. We found that changes from ideal acidity contributed a limited source of error, respectively, to the calculation of PaO₂ using our cohort’s curve. We demonstrated that data from published cohorts may not be universally useful for the calculation of PaO₂ in neonates.

We validated the UIHC curve in additional patients with well characterized clinical phenotypes. We found promising accuracy in simulations that quantified the additive error due to acidity, and age in the calculation of PaO₂ in neonates. In future studies we would like to perform prospective analysis on patient data that limit potential variation in SpO₂ measures.