Estimation of PaCO2 by two noninvasive methods in the critically ill newborn infant

Simultaneous measurements of arterial, transcutaneous, and peak expired carbon dioxide were obtained in 24 newborn infants receiving mechanical ventilation during the first week after birth. Two calibration algorithms designed to estimate PaCO2 from the noninvasive measurements were then examined. Both approaches entailed finding a statistical relationship by which future noninvasive measurement could be used to estimate the arterial value rather then measuring it directly. The first utilized the difference between the initial paired measurements (an in vivo calibration); the second used the mean difference between all measurements in the population. Adjusted tcPCO2 measurements by either the in vivo calibration or by the population-based factor led to estimates of PaCO2 with 95% confidence limits of +/- 6 to +/- 8 torr. In contrast, this degree of precision for the peak expired CO2 measurement was only possible using the in vivo calibration method. The use of an airway adaptor for PCO2 measurement led to CO2 retention in more than half of the infants. Transcutaneous monitoring had no significant effects on the infants, but was hampered by excessive drift and erratic sensitivity of the electrodes.