Hypoxemic and hyperoxemic likelihood in pulse oximetry ranges: NICU observational study

Objective Describe the likelihood of hypoxemia and hyperoxemia across ranges of oxygen saturation (SpO(2)), during mechanical ventilation with supplemental oxygenation. Design Retrospective observational study. Setting University affiliated tertiary care neonatal intensive care unit. Patients Two groups of neonates based on postmenstrual age (PMA): < 32 weeks (n= 104) and > 36 weeks (n= 709). Main measures Hypoxemia was defined as a PaO2 < 40 mm Hg, hyperoxemia as a PaO2 of > 99 mm Hg and normoxemia as a PaO2 of 50-80 mm Hg. Twenty-five per cent was defined as marked likelihood of hypoxemia or hyperoxemia. Results From these infants, 18 034 SpO(2)-PaO2 pairs were evaluated of which 10% were preterm. The PMA (median and IQR) of the two groups were: 28 weeks (27-30) and 40 weeks (38-41). With SpO(2) levels between 90% and 95%, the likelihoods of hypoxemia and hyperoxemia were low and balanced. With increasing levels of SpO(2), the likelihood of hyperoxemia increased. It became marked in the preterm group when SpO(2) was 99%-100% (95% CI 29% to 41%) and in the term group with SpO(2) levels of 96%-98% (95% CI 29% to 32%). The likelihood of hypoxemia increased as SpO(2) decreased. It became marked in both with SpO(2) levels of 80%-85% (95% CI 20% to 31%, 24% to 28%, respectively). Conclusions The likelihood of a PaO2 < 40 mm Hg is marked with SpO(2) below 86%. The likelihood of a PaO2 > 99 mm Hg is marked in term infants with SpO(2) above 95% and above 98% in preterm infants. SpO(2) levels between 90% and 95% are appropriate targets for term and preterm infants.