Pulse oximetry versus transcutaneous pO2 in sick newborn infants.

A pulse oximeter (Ohmeda Biox 3700) and two transcutaneous systems (Radiometer TCM3) were applied simultaneously to 18 newborn infants with respiratory insufficiency. All infants had either an umbilical catheter placed in the mid thoracic aorta or a radial artery catheter. The average monitoring time was 2 hours. Arterial blood pO2, pCO2 and pH (Radiometer ABL300), arterial sO2, HbCO and metHb (Radiometer OSM3), erythrocyte 2,3 DPG concentration, and fetal hemoglobin fraction (alkali denaturation kinetic method) were measured. Using arterial sO2 and pO2 as reference, the analytical bias of pulse oximetry (-0.5 +/- 1.0%, mean +/- 1 SD) corresponded in magnitude, when converted to pO2, to that of transcutaneous - pO2 (0.6 +/- 1.4 kPa for combined O2-CO2 electrode and -0.1 +/- 2.3 kPa for single O2 electrode). Transcutaneous pCO2 showed the smallest bias (0.3 +/- 0.3 kPa). Both pulse oximetry and transcutaneous pO2 electrodes were good as trend monitors detecting rapid changes in the infants' oxygenation status. The pulse oximeter offers certain advantages in not requiring calibration or heating. The variations in the levels of fetal hemoglobin fraction (44 to 97%), pH (7.27 to 7.49), pCO2 (3.3 to 6.8 kPa) and 2,3 diphosphoglycerate concentration (1.6 to 5.9 mmol/l) between the infants studied, resulted in a variable pO2-sO2 relation (p50 2.5 to 3.5 kPa). This presents difficulties in interpreting sO2 values in sick newborn infants, and we therefore recommend caution in using a pulse oximeter to apply strict limits for avoiding hypoxia and hyperoxia in this population.