Abnormally high and low levels of carbon dioxide and oxygen in the first few hours or days after birth have been associated with inpatient mortality and morbidities in very preterm babies (1–3). Over a decade ago, Tracy et al. reported hypocarbia and hyperoxia during the resuscitation of preterm infants with severe respiratory distress in the delivery room (DR) (4). At that time, respiratory support in the DR often included (i) self-inflating bags, (ii) no positive end-expiratory pressure (PEEP), (iii) early intubation, and (iv) prophylactic surfactant administration (5, 6). In addition, infants were resuscitated with 100% oxygen, and inspired oxygen concentration was not titrated against oxygen saturation (5, 6). Invasive respiratory support can cause lung injury in preterm infants through several mechanisms including high pressures (barotrauma), hyperinflation and overdistension (volutrauma), alveolar instability due to repeated expansion and collapse (atelectrauma), and the release of inflammatory mediators (biotrauma) (7). Added exposure to high oxygen concentration causes oxidant-mediated lung injury (8). As a result, a variety of lung protective strategies have been introduced in the DR (7, 9).

Practice changes in the DR have included (i) using PEEP during positive-pressure ventilation (PPV) (10–12), (ii) early use of non-invasive respiratory support with continuous positive airway pressure (CPAP) (13), (iii) elective and selective intubation for surfactant administration (14), (iv) initial use of a low inspired oxygen concentration (21–40%) rather than starting with 100% oxygen (15), and (v) targeted oxygen delivery using oxygen saturation reference ranges (16, 17). A recent randomized controlled trial including 88 preterm infants reported less oxygen exposure and oxidative stress with decreased respiratory morbidities when a limited oxygen strategy was used in DR resuscitation, compared to that of a high oxygen strategy (18).

We hypothesize that improved resuscitation practices have led to less variation in blood gas values and thus a reduced contribution of these values to adverse inpatient outcomes. We describe early blood gas measurements (in the first hour after birth) in preterm neonates (<33 weeks gestational age) receiving current DR stabilization techniques to control oxygenation and ventilation. In addition, we evaluate the association of initial partial pressures of carbon dioxide (PCO₂) and oxygen (PO₂) and their correlation with inpatient morbidities and mortality.

Demographics of study infants are presented in Table 1. Blood gases were performed in 170 infants (arterial n = 75, venous n = 95) within first hour (a mean of 47 min) after birth. The infants had mean (SD) gestational age of 28 (3) weeks and birth weight of 1,111 (403) g. Overall, there was a greater number of arterial blood gases collected from preterm infants who were <27 weeks gestation (46 arterial and 7 venous), had lower 1- and 5-min Apgar scores, and were less likely to have delayed cord clamping, compared to infants ≥27 weeks gestation (29 arterial and 88 venous) (p < 0.05). There were no differences in gender or Cesarean section rate.

In the whole cohort, 43% had normocarbia, 57% had hypercarbia, and 3% had hyperoxia.

Our study observed arterial and venous blood gases in a cohort of very preterm babies in the first hour after birth. By using current resuscitation and stabilization techniques that monitor and control oxygenation and ventilation, we found a low rate of hyperoxia and hyperventilation (hypocarbia), both of which are implicated in tissue injury. We were unable to demonstrate an association of abnormal blood gases with major inpatient morbidity or death.

There are limited data on PCO₂ and PO₂ values in preterm infants immediately after birth. A decade ago, Tracy et al. reported that 26% of preterm infants <34 weeks gestation had hypocarbia, 73% had normocarbia, and none had hypercarbia at NICU admission (4). In addition, 38% of infants had hypoxia, and 20% were both hypocarbic and hyperoxic. Kong et al. reported median cord PCO₂ values of 50 mmHg, with a wide range of PCO₂ values after NICU admission (23).

The optimal PCO₂ goal in clinical practice has not been determined; however, both hypocarbia and hypercarbia have been associated with severe intraventricular hemorrhage, periventricular leukomalacia, and bronchopulmonary dysplasia (1–3, 24–26). Both hypocarbia and hypercarbia cause changes to cerebral perfusion–reperfusion, which are associated with brain injury (27). However, these mechanisms may be absent during the first postnatal day, with cerebral blood flow reactivity to PCO₂ being re-established by the second or third postnatal day in very preterm neonates (27). We examined arterial and venous umbilical blood gases in 170 infants within an hour after birth. We did not observe any hypocarbia and observed normocarbia in 43% and hypercarbia in 57%. In particular, infants with hypercarbia were significantly more likely to be intubated and receive surfactant compared to those with normocarbia (Table 2). We speculate that the significant reduction in hypocarbia was associated with the increased use of CPAP (13) and elective (rather than prophylactic) intubation for surfactant administration (14).

Furthermore, we observed hyperoxia in 3%, hypoxia in 22%, and normoxia in 75% of preterm infants. This is of clinical relevance because oxygen administration at birth leads to an increase in oxidative stress markers (18, 28). The guidelines on the use of oxygen during neonatal resuscitation were revised in 2010, recommending room air (21% oxygen) as the initial gas. This practice was based on animal studies as well as meta-analyses of multiple human studies (16). The situation is a little less clear for preterm infants, even though the potential for harm is greater: cautious use has been recommended in anticipation of future randomized controlled trials. Many centers like ours initiate resuscitation with 21–30% oxygen to avoid hyperoxia (29). Using the published SpO₂ nomogram of newborn infants in the DR, inspired oxygen concentration is currently titrated to target the oxygen saturation ranges during neonatal resuscitation and stabilization (16). Preterm infants are at an increased risk for oxidative stress and its related injury due to the immaturity of the antioxidative defense mechanisms. Saugstad coined the term “Oxidative Radical Disease of Neonatology” to refer to several free radical-related morbidities in preterm infants including intraventricular hemorrhage, periventricular leukomalacia necrotizing enterocolitis, retinopathy of prematurity, and chronic lung disease (30). Our multiple regressions did not find significant differences in these outcomes in infants with hypoxia versus normoxia. Other clinical practices, such as delayed cord clamping, permissive hypercapnia, and developmentally sensitive care may influence the changes in SpO₂ and PCO₂ in the first minutes after birth. These factors should be considered when establishing standard approaches for resuscitation of preterm infants. In addition to the changes in neonatal resuscitation practices, factors such as regionalization of perinatal care and improvements in obstetric care will inevitably influence rates of hypoxia, hyperoxia, hypocarbia, and hypercarbia after birth. It is important that these trends and their potential impact on neonatal outcomes are understood. We believe our study provides interesting and valuable insights into the impact of current practices.

Measurement and confirmation of hypoxia, hyperoxia, hypocarbia, and hypercarbia after very preterm birth is an inexact science, and prevalence may vary based on clinical practices. Current practices lead to low rates of hyperoxia or hypocarbia. Given the study limitations, no associations were found between early blood gas measurements up to 1 h of age and inpatient morbidity or death.

HH was involved in data collection, analysis, and interpretation; wrote the first draft; performed revisions of the drafted article; and approved the final version of the manuscript for submission. P-YC, MO, KA, and GS were involved in the conception and design of the study; involved in data collection, analysis, and interpretation; performed critical revisions of the drafted article; and approved the final version of the manuscript for submission. SO and AS were involved in data collection, analysis, and interpretation; performed critical revisions of the drafted article; and approved the final version of the manuscript for submission.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.