[Estimation of PaO2 using pulsatile oximetry and the HbO2 dissociation curve in the premature infant].

An estimation of arterial PO2 from pulse oximetry using oxyhemoglobin dissociation curve. Forty blood specimens were drawn from umbilical arteries of the neonates having a pulse sensor attached to either foot or hand. The sensor was connected to a Nellcor-N100C in twenty cases and to an Ohmeda Biox-3700 for the remaining. Mean birth weight was 1,513 g, mean gestational age 32.4 weeks, and foetal hemoglobin ranged from 54.8 to 100%. Oxyhemoglobin dissociation curves were drawn for each arterial sample to obtain P50 and estimated PO2. The estimated PaO2 was the PO2 corresponding, on the oxyhemoglobin dissociation curve, to the pulse oximeter saturation reading. Mean difference between blood saturation and pulse oximeter saturation of 1.2% was obtained with Nellcor and of 3.9% with Ohmeda. Arterial and estimated PO2 comparison gave a mean + S.E.M. difference of 4.6 +/- 1.3 torr for all cases. This represents less than 8% error in estimating PaO2 from pulse oximetry. We suggest the following polynomial equation Y = 4250 - 10.4X + 0.069X2 for using pulse oximetry (X) to evaluate PaO2 (Y) in premature infants.