INTRODUCTION — Oxygen supplementation is an important component of intensive neonatal care for the preterm infant. Ideally, oxygen administration provides adequate oxygenation for the metabolic needs of the neonate while avoiding the consequences of both hypoxemia and hyperoxia. However, it remains clinically challenging to define the optimal target levels of oxygen, especially in preterm infants.
Establishing neonatal target oxygen level for the preterm infant is discussed here. Oxygen delivery and monitoring and mechanical ventilation in the newborn are discussed separately. (See "Respiratory support, oxygen delivery, and oxygen monitoring in the newborn".)
SpO2 AND PaO2 — Normal values for hemoglobin oxygen saturation (SpO2) reach or exceed 80 percent within 10 minutes of birth in term and healthy preterm infants without supplemented oxygen [1]. In general, arterial partial pressure of oxygen (PaO2) values of 50 to 80 mmHg are adequate to meet metabolic demands of the neonate in part due to the greater proportion of fetal hemoglobin (HbF).
Use of SpO2 for targeting oxygen levels — For neonates who receive supplemental oxygen therapy, established neonatal target oxygen levels are based on hemoglobin oxygen saturation (SpO2) since most neonatal intensive care units (NICUs) use continuous noninvasive pulse oximetry, which measures SpO2, to monitor oxygenation. Periodic arterial blood gas (ABG) samples are obtained that measure arterial oxygen tension (PaO2) to further refine safety and efficacy of oxygen therapy and correlate with pulse oximetry measurements, especially for infants at risk for either hypoxemia or hyperoxemia.
However, ABG measurements require blood sampling, either through indwelling catheters or percutaneous puncture of a palpable artery. ABG measurements may change during percutaneous punctures as the infant responds to the procedure, so clinicians should be mindful of this in interpreting the results. (See "Respiratory support, oxygen delivery, and oxygen monitoring in the newborn", section on 'Pulse oximetry' and "Respiratory support, oxygen delivery, and oxygen monitoring in the newborn", section on 'Arterial blood gas measurement'.)
Correlation between PaO2 and SpO2 — In neonates, SpO2 levels between 85 and 95 percent generally correlate to PaO2 levels between 45 and 65 mmHg based on their high fetal hemoglobin (figure 1) [2,3]. However, the use of pulse oximeters and measurement of SpO2 are limited when there is significant hyperoxemia or hypoxemia. For example, at saturations above 96 percent, PaO2 values can continue to increase with little or no further corresponding change in SpO2.
In addition, the increased concentration of HbF and its increased affinity for oxygen is another factor to consider in establishing target SpO2 values in the neonate. HbF will shift the oxygen dissociation curve to the left due to its high affinity for oxygen, which may result in high oxygen saturation (eg, 85 percent) at PaO2 levels below 45 mmHg (figure 1). The proportion of HbF increases with decreasing gestational age (GA) so that the concentration of HbF in an infant born at 28 weeks gestation is approximately 90 percent.
Because of these complex interactions, preterm infants at high risk for hyperoxemia or severe hypoxemia, as well as those with rapidly changing cardiopulmonary status, require periodic ABG determination. (See "Anemia of prematurity (AOP)", section on 'Oxygen delivery'.)
OXYGEN TARGET LEVELS
Goals — An optimal target pulse oximetry saturation range provides oxygenation that meets the metabolic needs of the preterm infant, yet limits high concentrations of supplemental oxygen that are not needed, and avoids hyperoxia and hypoxia. Successful target oxygen levels avoid the adverse effects of:
●High concentrations of oxygen and hyperoxia – High concentrations of supplemental oxygen contribute to bronchopulmonary dysplasia (BPD) due to direct oxygen toxicity to the lungs. Retinopathy of prematurity (ROP) is a major complication of excessive tissue oxygenation (hyperoxia). (See "Bronchopulmonary dysplasia: Definition, pathogenesis, and clinical features", section on 'Oxygen toxicity' and "Retinopathy of prematurity: Pathogenesis, epidemiology, classification, and screening", section on 'Risk factors'.)
●Hypoxia (insufficient tissue oxygenation) is associated with metabolic acidosis, increased mortality, and possibly neurodevelopmental impairment. (See 'Clinical trials' below.)
High versus low SpO2 targets — However, the optimal SpO2 for preterm infants who receive supplemental oxygen therapy has not been fully established despite data provided by several clinical trials [4]. Nevertheless, based on the available evidence, we believe the most prudent target range for SpO2 in preterm infants is between 90 and 95 percent. This range minimizes both the low and high extreme oxygenation levels that have been associated with adverse outcomes and mortality.
Clinical trials — The initial clinical trial published in 2000, Supplemental Therapeutic Oxygen for Prethreshold Retinopathy of Prematurity (STOP-ROP) trial, demonstrated that an oxygen target goal of SpO2 between 96 to 99 percent in preterm infants was associated with greater morbidity (eg, greater risk of BPD and longer duration of hospitalization) than those assigned to a lower target range of 89 to 94 percent [5].
Subsequent trials (Canadian Oxygen Trial [COT]; Surfactant, Positive Pressure, and Oxygenation Randomized Trial [SUPPORT] study; and Benefits of Oxygen Saturation Targeting trial [BOOST]-II study) using lower oxygen target goals were conducted to determine the optimal SpO2 target range for extremely preterm infants (gestational age [GA] <28 weeks) [6-8]. These trials recruited thousands of infants and had similar protocols comparing target saturation ranges between high SpO2 target range (91 to 95 percent) and low SpO2 target range (85 to 89 percent) [6-8].
Systematic reviews of these trials have reported the following findings [9-12].
●There was a higher rate of mortality before the postmenstrual age (PMA) of 36 weeks, before hospital discharge, and prior to reaching a corrected age of 18 to 24 months in infants assigned to the low SpO2 target range [9,11,12].
●There was a higher incidence of necrotizing enterocolitis (NEC) associated with the low SpO2 target range [9,10,12].
●There was a greater risk of patent ductus arteriosus requiring surgical ligation for infants assigned to the low SpO2 target range [12].
●There was no difference in the combined outcome of death and major disability at 24 months corrected age [9,10,12]. There was also no difference in neurodevelopmental outcome, blindness, severe hearing loss, or cerebral palsy [9].
●There was a lower incidence of ROP requiring treatment associated with the low SpO2 target range [9,12].
●There was a lower risk of receiving supplemental oxygen at PMA of 36 weeks associated with the low SpO2 target range [12].
However, results from these meta-analyses need to be interpreted with caution because there was significant overlap in the actual exposed oxygen saturation between the two groups despite the design of the trials specifying separation between the target ranges. In addition, these data were obtained in a research setting, and it may be challenging to implement targeted oxygen saturation goals in a non-research setting, as it often is difficult to maintain an infant requiring oxygen within a narrow saturation range. (See 'Adherence to oxygen target goals' below.)
Changes also occurred while the BOOST-II study, which combined data from trials in Australia and the United Kingdom, was in progress. Oximeters were revised to correct a calibration-algorithm artifact. However, subsequent analysis showed that the increased risk of death still remained higher for the low target oxygen group [12-14].
There have been several follow-up publications of these trials that include the following:
●A post-hoc analysis of the COT showed prolonged hypoxemic episodes (defined as SpO2 <80 percent for ≥1 minute) during the first two to three months after birth were associated with late mortality or neurodevelopmental impairment at 18 months corrected age [15]. Of note, the association between exposure to prolonged hypoxemic episodes and the primary outcome of death and neurodevelopmental impairment was stronger for infants randomly assigned to the high oxygen saturation target range (91 to 95 percent) than for those assigned to the low target range (85 to 89 percent). The authors speculated that these episodes of prolonged hypoxemia in infants assigned to the high oxygen target levels represented greater drops of saturations from baseline values resulting in more severe disturbance of oxygen homeostasis.
●A post-hoc analysis of the SUPPORT study reported increased mortality before discharge for small for gestational age (SGA) compared with appropriate for gestational age (AGA) infants (38.5 versus 16.4 percent) [16]. There was no difference in mortality between the low and high target oxygen groups for AGA infants (17.6 versus 15.2). However, for SGA infants, mortality (defined as death by discharge) was greater for the infants assigned to a low target oxygen level (56.1 versus 25.5 percent). This disadvantage manifested by SGA versus AGA infants persisted, although the difference narrowed when all trials were included [12]. Of note, SGA infants in the lower target group were more likely to have lower achieved oxygen saturation than AGA infants. Additionally, SGA infants with the highest number of hypoxic events greater than 20 seconds duration in the first three days of life had the lowest survival [17].
●In a follow-up study of the SUPPORT cohort, there was no difference in growth between the low and high target oxygen groups during hospitalization (birth to 36 weeks postmenstrual age [PMA]), and at 18 to 22 months corrected age [18].
Our approach — Data regarding target oxygen ranges are from clinical trials conducted in extremely preterm infants (GA <28 weeks) who are the most vulnerable to the effects of both high and low oxygenation levels. Based on the available evidence, we recommend a target oxygen range from 90 to 95 percent resulting in minimizing extreme oxygenation levels for all preterm infants [19,20]. Although data are lacking in more mature infants, this target range appears to be safe for preterm infants ≥28 weeks gestation. In the most mature preterm infants (gestational age >34 weeks), the risk of ROP decreases, and the upper limit can be increased to 97 percent [21].
There is also a paucity of data regarding oxygen target ranges as the preterm infant advances in age. However, by two to three weeks postnatal age, the risk of intermittent hypoxia increases, which may aggravate ROP by enhancing retinal proliferation [22]. As a result, with advancing age, we consider raising target saturation to >95 percent if the infant still needs supplemental oxygen when the corrected PMA is >32 weeks.
Targeted SpO2 levels of infants with congenital heart disease, BPD, or pulmonary hypertension are individualized based upon the clinical status of the neonate due to the paucity of data regarding optimal oxygenation for these disorders. (See "Diagnosis and initial management of cyanotic heart disease in the newborn", section on 'General supportive care' and "Bronchopulmonary dysplasia: Management", section on 'Respiratory support'.)
Adherence to oxygen target goals — Guidelines need to consider the ability to adhere to oxygen target goals with currently available technology and provide, as best as possible, a safe range of oxygen target levels that can be reliably used in the practice setting. However, maintaining infants within the targeted range remains challenging, as illustrated by the following [4,23-25]:
●In a prospective Australian study of 45 preterm infants (mean GA 30 weeks) receiving continuous positive airway pressure (CPAP) and supplemental oxygen, continuous oxygenation monitoring was conducted over 24 hours [24]. Results showed infants were in the target SpO2 range for only 31 percent of the total recording time. The incidence of severe hyperoxia (defined as SpO2 ≥98 percent) was 48 episodes per 24 hours and hypoxia (defined as SpO2 <80 percent) was 9 episodes per 24 hours. In this study, the mean number of adjustments of the fraction of inspired oxygen (FiO2) was 25 times per 24 hours.
●In another study, although the lower alarm limit was usually set correctly, the upper alarm limit was set correctly only 23 percent of the time [25].
These results highlight the difficulty of keeping infants in the desired target range, and need to be considered when guidelines are developed for oxygen targeting in the preterm infant.
Emerging technology using automated systems with closed loop feedback may enhance consistency of adhering to oxygen saturation targets compared with the current method of manual control of FiO2 including for infants receiving nasal high flow oxygen [26-31]. Although encouraging, data are only from short-term observational studies and do not provide information on the impact of these new systems on patient outcomes [32].
Potential tool: Near-infrared spectroscopy — Near-infrared spectroscopy (NIRS) can provide noninvasive continuous monitoring of cerebral oxygenation, which is influenced by pulse oximeter saturation (SpO2), hemoglobin content, and cerebral blood flow. Observational studies have shown periods of low cerebral oxygenation detected by NIRS monitoring were associated with poor outcome (mortality, radiologic evidence of neurologic injury, necrotizing enterocolitis colitis, and neurodevelopmental impairment at two and three years of age) [33-35]. It has been proposed that using NIRS with a target cerebral oxygenation level may improve long-term outcome for preterm infants. One small study of 127 preterm infants <32 weeks gestation reported levels below NIRS cerebral saturation levels <67 percent during the first 72 hours of life was associated with an unfavorable outcome (death or neurodevelopmental impairment at 2 years) [36]. Further studies are needed to determine whether monitoring and identifying target cerebral oxygenation levels using NIRS will improve neonatal outcomes.
SUMMARY AND RECOMMENDATIONS
●Monitoring oxygenation using SpO2 – For neonates who receive supplemental oxygen, established neonatal target levels are based on hemoglobin oxygen saturation (SpO2) because continuous noninvasive pulse oximetry, which measures SpO2, is primarily used to monitor neonatal oxygenation. Periodic arterial blood gas (ABG) samples, which measure arterial oxygen tension (PaO2), are obtained to further refine safety and efficacy of oxygen therapy, and correlate with pulse oximetry measurements especially for preterm infants at-risk for either hypoxemia or hyperoxemia. (See 'Use of SpO2 for targeting oxygen levels' above.)
●Correlation between PaO2 and SpO2 – In neonates, SpO2 levels between 85 and 95 percent generally correlate to PaO2 levels between 45 and 65 mmHg. However, pulse oximetry measurements are limited when there is significant hyperoxemia or hypoxemia (figure 1). In addition, the increased concentration of fetal hemoglobin (HbF) with its high affinity for oxygen must be considered in establishing target SpO2 levels, as it shifts the oxygen dissociation curve to the left. The concentration of HbF increases with decreasing gestational age (GA). (See 'Correlation between PaO2 and SpO2' above.)
●SpO2 target levels – An optimal target pulse oximetry saturation range provides adequate oxygenation that meets the metabolic needs of the neonate while avoiding high concentrations of oxygen, hyperoxia, and hypoxia. Data from clinical trials of preterm infants using pulse oximetry showed a target goal of SpO2 above 96 percent was associated with an increased risk of bronchopulmonary dysplasia (BPD), whereas a low target goal between 85 and 89 percent was associated with increased mortality compared with a target goal of 91 to 95 percent. (See 'Clinical trials' above.)
•Based on the available data, we recommend a target oxygen range for SpO2 between 90 and 95 percent for preterm infants who receive supplemental oxygen and are continuously monitored by pulse oximetry (Grade 1B). (See 'Oxygen target levels' above.)
●Adherence – Maintaining infants within the targeted range is challenging as adhering to target oxygen range is difficult. As a result, guidelines for oxygen target ranges need to consider the ability to adhere to goals and need to provide as best as possible a safe range of oxygen target levels that can be reliably used in the practice setting. (See 'Adherence to oxygen target goals' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges James Adams, Jr., MD, who contributed to an earlier version of this topic review.
