Pulse Oximetry and Normoxemia in Neonatal Intensive Care

Abstract

BACKGROUND: Conflicting results are found in the medical literature concerning the reliability of pulse oximetry in predicting hyperoxemia ( $P_{a{o}_2}$ > 90 torr [12 kPa]) and hypoxemia ( ${\text{P}}_{{\text{ao}}_{\text{2}}}$ < 45 torr [6.0 kPa]) in neonates. This study was designed to determine whether oximeter saturation limits could be established that would adequately predict normoxemia in our neonatal population. METHODS: Fifty-two infants in our intensive care unit were studied over a wide range of values for temperature, pH, and PaCO2. Three-hundred fifty-three arterial blood gas samples from umbilical or arterial catheters were drawn and simultaneous pulse oximeter saturation ( $S_{p{o}_2}$ ) values recorded using Nellcor oximeters. RESULTS: $S_{p{o}_2}$ values ranged from 61% to 100%, and Pao, from 15 to 421 torr [2.0 to 56.1 kPa]. Predictive value analyses were applied to determine the reliability of $S_{p{o}_2}$ ' in detecting the absence of hyperoxemia and hypoxemia. By comparing the sensitivity, specificity, and positive predictive value (PPV) for different limiting values of $S_{p{o}_2}$ , an upper limit of 96% proved to be optimal in preventing hyperoxemia. $S_{p{o}_2}$ ' values < 96% were associated with ${\text{P}}_{{\text{ao}}_{\text{2}}}$ < 90 torr [12 kPa] in 239 out of 246 samples, yielding a PPV of 97% and a 2.8% rate of false positives ( $S_{p{o}_2}$ <96% with ${\text{P}}_{{\text{ao}}_{\text{2}}}$ > 90 torr [12 kPa]). $S_{p{o}_2}$ values ≥ 92% were associated with Pao, values ≥ 45 torr [6.0 kPa] in 262 of 273 samples, producing a PPV of 96% with a 4% false-positive rate ( $S_{p{o}_2}$ > 92% with ${\text{P}}_{{\text{ao}}_{\text{2}}}$ < 45 torr [6.0 kPa]). Hypoxemia was avoided in 96% of the samples with the Spo, maintained at < 92%, and hyperoxemia was avoided in 97% of the samples with $S_{p{o}_2}$ ' maintained at < 96%. DISCUSSION: Although establishing $S_{p{o}_2}$ ' limits can reliably prevent hyperoxemia and hypoxemia, a considerable number of false negatives do occur in which a ${\text{P}}_{{\text{ao}}_{\text{2}}}$ of 45-90 torr [6.0-12 kPa] is associated with a $S_{p{o}_2}$ ' value outside the 92% to 96% range. Also, occasional aberrant $S_{p{o}_2}$ values can occur for unexplained reasons. CONCLUSION: This study demonstrates the ability of pulse oximetry to predict (with reasonable certainty) a clinically acceptable range of ${\text{P}}_{{\text{ao}}_{\text{2}}}$ values in neonates, under variable conditions. However, the range of saturations to protect against hyperoxemia and hypoxemia must be established for each brand or model of oximeter.