Artifactual pulse-oximetry estimation in neonates.

Two sources of artifactual pulse-oximetry estimation were investigated in 20 neonates. Increased pressure on tissue due to inappropriate sensor fixation was mimicked with a blood pressure cuff. The error in arterial oxygen saturation (pSO2) exceeded 2 SD (> 3%) in 25% subjects at 50 mm Hg which in an ancillary experiment was produced by 11 of 26 nurses fixing the sensor. Venous congestion at 30 and 40 mm Hg permitted normal detection of pulse rate but induced errors in pSO2 over 2 SD in 15% and 30% of subjects, respectively. Pulse-oximeter values need to be scrutinised for these common errors.