Main

Infants born very preterm (<32 wk of gestation) commonly suffer respiratory failure that requires intervention to initiate and sustain pulmonary gas exchange (1). However, it is unclear how this can be achieved without injuring the immature lung and impeding the physiological processes that underpin the transition to postnatal life (2). Before birth, the future airways of the lung are filled with liquid that must be cleared at birth to allow air entry and the onset of pulmonary gas exchange (3,4). Recent imaging studies have demonstrated that airway liquid clearance is closely associated with inspiration and that transpulmonary pressures generated during inspiration play a major role (3,5). The air/liquid interface moves distally with each inspiratory effort and as the liquid leaving the airways is replaced with inhaled air, not all of the inhaled air is exhaled during expiration (3,5). Thus, there is a breath-by-breath increase in end-expiratory gas volumes leading to establishment of a functional residual capacity (FRC) (3). It is not surprising, therefore, that the application of a sustained inflation (SI) pressure to overcome the long time- constant required to move liquid through the airways into the distal air sacs can greatly facilitate airway liquid clearance (6,7). However, it is currently unclear whether a sustained increase in intrathoracic pressure during the SI will adversely affect the cardiovascular transition at birth.

Before birth, pulmonary vascular resistance (PVR) is high, and the majority (∼90%) of right ventricular output bypasses the lungs and flows from the pulmonary artery into the thoracic aorta (right-to-left shunting) via the ductus arteriosus. The onset of pulmonary ventilation after birth causes a large decrease in PVR, which greatly increases pulmonary blood flow (PBF), allowing it to accept the entire output from the right ventricle and a large contribution from the left ventricle via left-to-right shunting through the ductus arteriosus (i.e. from the systemic to pulmonary circulation); this contributes up to 50% of PBF immediately after birth (8). These cardiovascular changes at birth are central to the successful transition to postnatal life and are triggered by the reduction in PVR induced by lung aeration. However, it is well established that sustained increases in intrathoracic pressure increase PVR and decrease venous return leading to a decrease in cardiac output and cerebral vasodilation (9,10). Thus, it is unclear whether an initial SI at the onset of pulmonary ventilation would impede the cardiovascular transition immediately after birth. The primary aim of this study was to determine whether an initial SI would impede the increase in PBF and adversely influence carotid blood flow (CBF) immediately after birth. A secondary aim was to compare the effects of a SI with shorter inflation times (0.3 s) on the respiratory transition at birth, particularly tidal volume (VT) recruitment and lung compliance changes.

METHODS

All experimental procedures were approved by the relevant Monash University Animal Ethics Committee. Thirteen pregnant ewes underwent surgery at 122 ± 2 (mean ± SD) days of gestation (d: term is ∼147 d) for insertion of catheters into a fetal carotid artery (CA) and jugular vein, the left pulmonary artery (LPA), and the amniotic sac. A 4-mm ultrasonic flow probe (Transonic Systems, Ithaca, NY) was placed around the LPA and the noncatheterized CA, as previously described (11). Fetal well-being was monitored daily after surgery by measuring fetal arterial partial pressure of oxygen (PaO2), partial pressure of carbon dioxide (PaCO2), pH, and arterial oxygen saturation (SaO2; ABL30, Radiometer, Copenhagen, Denmark).

At 127 ± 4 d (mean ± SD), the ewe and fetus were anesthetized, and the fetal head and neck were exposed via cesarean section. The fetal trachea was intubated with a cuffed endotracheal tube (3.5 mm), and lung liquid was drained passively. A transcutaneous oximeter (Masimo, Irvine, CA) was attached around the right forelimb before the umbilical cord was clamped and cut. Lambs were delivered, dried, weighed, and placed under a radiant heater. Lambs received 5% dextrose (i.v.) and were sedated (alfaxane i.v.: 5–15 mg/kg/h; Jurox, East Tamaki, Auckland, New Zealand) to minimize spontaneous breathing during the experiment. Ewes were humanely killed (sodium pentobarbitone: ∼100 mg/kg i.v.) after delivery of the lamb.

Left pulmonary and carotid arterial blood flows [PBF and CBF, respectively] were recorded continuously (Powerlab; ADInstruments, Castle Hill, NSW, Australia), along with pressures in the LPA and CA (DTX Plus Transducer; Becton Dickinson, Singapore) from before delivery until the end of the experiment. Lambs were randomly assigned to receive either an initial SI or conventional ventilation (non-SI). Lambs in the SI group (n = 7) received an initial SI for either 1 min or until 20 mL/kg (15 mL/kg FRC and 5 mL/kg VT) of air had entered the lungs. The SI was delivered using a Neopuff (Fisher & Paykel Healthcare, Panmure, Auckland, New Zealand) set to deliver a peak inspiratory pressure (PIP) of 40 cm H2O and positive end-expiratory pressure (PEEP) of 5 cm H2O. After the SI, lambs were connected to a mechanical ventilator (Babylog 8000+ ventilator; Drager, Lubeck, Germany) and ventilated in volume guarantee mode (VT 7 mL/kg) for 30 min (PEEP, 5 cm H2O; inspiratory time, 0.3 s; and expiratory time, 0.6 s). Ventilation was the same for lambs in the control group (non-SI; n = 6) but without the initial SI; the PIP was initially adjusted every 30 s (by a maximum of 5 cm H2O) to achieve the desired VT. We recorded the delivered VT and PIP simultaneously with arterial blood flow and pressure data. Lung compliance was calculated (12) according to the formula:

Ten minutes after the SI or the onset of ventilation in control lambs, the fraction of inspired oxygen (FiO2) increased from 21 to 100% for 10 min and then reduced to 21% for a further 10 min. These manipulations were performed to investigate the effects of oxygenation on PBF and CBF. Arterial blood gases were measured every 5 min. Thirty minutes after delivery, lambs were killed by anesthetic overdose (sodium pentobarbitone; 130 mg/kg i.v.).

Analytical methods.

Mean PBF, CBF, LPA, and CA pressures were averaged over 20-s epochs immediately before and after umbilical cord occlusion, and at 1, 2, 3, 4, 5, 10, 15, 20, 25, and 30 min after the onset of ventilation. In addition, to interrogate changes in these variables during the initial SI, average values over each 5 s were calculated during the first minute. VT and PIP were averaged for these same epochs. At times corresponding to when arterial blood gases were measured, we calculated carotid arterial blood O2 content to estimate cerebral O2 delivery (13) according to the formula:

where blood O2 content = (Hb concentration × SaO2 × 1.36) + (Pao2 × 0.003).

Statistical methods.

Data were analyzed using two-way repeated measures ANOVA with group (SI versus non-SI) and time as factors. Post hoc comparisons between groups and time points were performed using the Holm-Sidak test. Data are presented as mean ± SEM unless otherwise stated.

RESULTS

Fetal well-being before delivery, as indicated by fetal blood pH, PaCO2, PaO2, and SaO2 values were not different between groups (Table 1). The mean duration (± SD) of the initial SI was 40.1 ± 21.3 s. The average VT was initially below the targeted volume (7 mL/kg) in both groups of lambs but tended to be higher in SI lambs during the first 5 min of tidal ventilation (data not shown). Although VT were not significantly different between groups, the inspiratory pressures required to achieve these volumes were higher (by ∼9–15 cm H2O) in non-SI lambs compared with SI lambs (p < 0.001; Fig. 1). As a result, lung compliance was greater in SI lambs compared with non-SI lambs throughout the 30-min ventilation period (Fig. 1).

Table 1 Fetal arterial pH, PaCO2, PaO2, and SaO2 before delivery
Full size table
Figure 1
figure 1

Inspiratory pressure (A) and lung compliance (B) in lambs receiving a SI (○) or no SI (bull;) with the initiation of ventilation (designated as time 0). Light shading indicates ventilation with air; dark shading indicates ventilation with 100% O2. Data are mean ± SEM. ‡p < 0.05, SI vs No SI; *p < 0.05 vs “1 min” time point for both groups.

Full size image

The PaCO2 was significantly lower and PaO2 significantly higher (p < 0.05; Fig. 2) in SI lambs compared with non-SI lambs during the entire experimental period, regardless of FiO2 (Fig. 2). In SI lambs, the PaO2 and SaO2 were both increased (p < 0.05) from fetal levels by ventilation with air and the PaO2 markedly increased further (to 210.2 ± 44.4 mm Hg) by ventilation with 100% O2. Although the SaO2 increased in the SI group to 100% with the increase in FiO2, this was not statistically significant as the SaO2 levels were already high (88.5 ± 5.3%) during the previous ventilation in air. In contrast, we found it very difficult to maintain oxygen levels in non-SI lambs using conventional ventilation in air, despite having an almost identical VT to that used in SI lambs. During ventilation in air, both the PaO2 and SaO2 in non-SI lambs were significantly reduced from fetal levels (Fig. 2) and were only increased in response to ventilation with 100% O2. Ventilation with 100% O2 increased the PaO2 and SaO2 from 11.0 ± 0.1 mm Hg and 16.0 ± 1.5% to 58.1 ± 9.9 mm Hg and 89.5 ± 6.1%, respectively, within 10 min (Fig. 2). Returning the ventilation of non-SI lambs to air rapidly (within 5 min) reduced both the PaO2 and SaO2 (p < 0.05) to values measured before the increase in FiO2 (Fig. 2).

Figure 2
figure 2

PaCO2 (A), PaO2 (B), and SaO2 (C) in lambs receiving a SI (○) or no SI (bull;) measured before umbilical cord occlusion (fetal) and after ventilation onset (designated as time 0). Light shading indicates ventilation with air; dark shading indicates ventilation with 100% O2. Data are mean ± SEM. ‡p < 0.05, SI vs No SI; *p < 0.05 vs “fetal” time point for both groups.

Full size image

At 20 s to 3 min after ventilation onset, PBF in SI lambs increased above fetal levels (p < 0.05) and was significantly greater than in non-SI lambs. In non-SI lambs, PBF did not increase above fetal levels until 2 min after ventilation onset (p < 0.05). Throughout the remainder of the experiment, PBF was similar in both groups. Pulmonary arterial pressure was transiently increased from fetal levels after umbilical cord occlusion in both groups (p < 0.05; Fig. 3). Pulmonary arterial pressure was not different between groups at any time point. Compared with values before ventilation, pulmonary arterial pressure was lower from 15 min in the SI group and from 20 min in the non-SI group.

Figure 3
figure 3

Left pulmonary arterial blood flow (A), pulmonary arterial pressure (B), carotid arterial flow (C), and carotid arterial pressure (D) in lambs receiving a SI (○) or no SI (bull;) measured before (fetal) and after umbilical cord occlusion and after ventilation onset (designated as time 0). Light shading indicates ventilation with air; dark shading indicates ventilation with 100% O2. Data are mean ± SEM. ‡p < 0.05, SI vs No SI. *p < 0.05 vs “fetal” time point for both groups. †p < 0.05 vs “fetal” time point for No SI group. §p < 0.05 vs “fetal” time point for SI group.

Full size image

In SI lambs, CBF tended to increase transiently with the onset of ventilation and then decreased over time (Fig. 3). In contrast, in non-SI lambs, CBF was significantly (p < 0.05) increased with ventilation onset over the first 5 min and remained elevated compared with SI lambs throughout the remainder of the experimental period (p < 0.05; Fig. 3). Carotid arterial pressure was higher in SI lambs compared with non-SI lambs (p < 0.05) from 15 s to 1 min after ventilation onset (i.e. during the SI period; Fig. 3). At 25 and 30 min after ventilation onset, carotid arterial pressure was lower in non-SI lambs than SI lambs (p < 0.05).

In SI lambs, cerebral O2 delivery did not vary significantly through the experimental period (Fig. 4). In contrast, cerebral O2 delivery in non-SI lambs tended to decrease, relative to fetal levels, during ventilation with air (Fig. 4) and was markedly increased, ∼4-fold (p < 0.05), in response to ventilation with 100% O2. Cerebral O2 delivery was ∼2-fold higher (p < 0.05) in non-SI lambs than in the SI group during ventilation with 100% O2. The resumption of ventilation with air lowered cerebral O2 delivery to levels seen before ventilation with 100% O2 in non-SI lambs but had no effect in SI lambs.

Figure 4
figure 4

Carotid arterial oxygen delivery in lambs receiving a SI (○) or no SI (bull;) measured before umbilical cord occlusion (fetal) and after ventilation onset (designated as time 0). Light shading indicates ventilation with air; dark shading indicates ventilation with 100% O2. Data are mean ± SEM. ‡p < 0.05, SI vs No SI; †p < 0.05 vs “fetal” time point for No SI group.

Full size image

DISCUSSION

This study has demonstrated that an initial SI markedly improved both the cardiovascular and respiratory transition at birth in lambs delivered preterm. Our primary aim was to demonstrate that the cardiovascular transition was not impeded by an initial sustained increase in intrathoracic pressure. We found that PBF was higher in SI lambs compared with non-SI lambs during the SI and remained similar to that measured in non-SI lambs throughout the remainder of the experimental period. CBF tended to decrease in SI lambs after ventilation and was markedly lower than that measured in non-SI lambs. A postventilation decrease in CBF is a well-described feature of well-oxygenated newborns and reflects an oxygen-mediated decrease in cerebral blood flow (14). In addition, despite having similar VT, the peak inflation pressures required to achieve these volumes in SI lambs were significantly less than those required in non-SI lambs. As a result, a SI significantly increased lung compliance for at least the first 30 min after ventilation onset. Combined with the very large differences in blood oxygenation and PaCO2 levels, these data indicate that a SI markedly improved and did not impede the physiological transition to newborn life.

Previous studies have demonstrated that an initial SI facilitates uniform lung aeration and FRC recruitment in preterm rabbits (6,7) and, in combination with subsequent nasal CPAP, reduces the need for intubation and the risk of moderate to severe bronchopulmonary dysplasia in preterm human infants (15). As lung aeration at birth involves the movement of liquid through the airways (5,16), the high resistance to liquid movement (compared with air) must increase the time required for the liquid to move distally into the gas exchange regions (3,4). Logically, therefore, a SI provides the time required for the relatively slow flowing liquid to move through the airways ahead of the initial entry of air. Furthermore, as the rate of liquid movement must vary across the lung depending on airway size, if the SI is of sufficient duration, most of the lung can be aerated by the initial inflation. This feature has been demonstrated clearly in imaging studies, which showed that an initial SI can fully recruit FRC and uniformly distribute gas within the lung during subsequent tidal ventilation (6,7).

Initially, we found that VT tended to be higher in lambs following a SI, but we were targeting a set VT so it is not surprising that these differences were not significant and diminished with time. However, the PIP required to achieve these volumes were markedly higher in non-SI lambs indicating that SI lambs had lungs that were significantly more compliant. In view of previous imaging studies (6,7), we consider that this resulted from a much greater degree of lung aeration in SI lambs, which is consistent with the markedly higher PaO2 and lower PaCO2 levels in these SI lambs. Indeed, we found it very difficult to ventilate non-SI lambs in air and were not able to achieve adequate oxygenation without increasing the FiO2, despite a high mean PBF in these lambs (see below). Thus, it is likely that only small parts of the lung were successfully aerated in non-SI lambs. As a result, because we were delivering a set VT (7 mL/kg), a gas volume that was appropriate for the entire lung was likely delivered to a small portion of the lung, probably causing overdistension and injury in those regions.

In non-SI lambs, increasing FiO2 to 1.0 markedly improved blood oxygenation but had no effect on ventilation parameters (e.g. lung compliance), PaCO2, or PBF. As such, the increase in oxygenation must have been achieved by simply increasing the partial pressure gradient for oxygen diffusion across the air/blood barrier. Other potential contributors, such as improved lung aeration, increased pulmonary perfusion, or an increase in ventilation/perfusion matching, would have led to physiological changes such as an increase in lung compliance, an increase in PBF, and a decrease in PaCO2 levels. Thus, in terms of respiratory function, increasing the FiO2 had little physiological benefit other than increasing blood oxygenation levels, which in this context may be of dubious benefit (see below). Increasing FiO2 to 1.0 in SI lambs, however, did not significantly alter most cardiorespiratory parameters examined. As these lambs had mean SaO2 of 88.5 ± 5.3% during the preceding period of ventilation in air, increasing the FiO2 to 1.0 simply increased the SaO2 to 100%, which was not a statistically significant increase.

The results of this study indicate clearly that a SI does not impede the pulmonary transition at birth, particularly the increase in PBF. Indeed, we found that during the SI, PBF was higher in SI lambs than non-SI lambs, which was surprising considering the sustained increase in intrathoracic pressure. However, as venous return is reduced by ∼50% immediately before ventilation onset, because of the occlusion of the umbilical cord (8), it is possible that any further reduction caused by a sustained increase in thoracic pressure is insignificant in comparison. Aeration of the lungs underlies the increase in PBF after birth, so the increased PBF in the SI group is presumably the consequence of more uniform lung aeration. As the non-SI lambs were very hypoxic and it is likely that their lungs were not completely aerated, we are surprised that the PBF response to ventilation onset was not reduced in these lambs. This suggests that these lambs had a large ventilation/perfusion mismatch, compared with SI lambs, which is consistent with the finding of severe hypoxemia and hypercapnia despite a large increase in PBF and normal minute ventilation. Furthermore, our results demonstrate that oxygenation status of the newborn had little or no overall effect on PBF.

Numerous previous studies have demonstrated that, after birth, CBF gradually decreases, which is considered to be an autoregulatory process in response to an increase in oxygenation (14,17,18). This response was observed in SI lambs, with CBF gradually decreasing from 40.2 ± 7.8 mL/min/kg at 1 min after birth to 19.9 ± 4.3 mL/min/kg by the end of the experiment. However, in non-SI lambs, CBF remained elevated and, although it tended to decrease during ventilation with 100% O2, it was significantly greater than in SI lambs. It is highly likely that the elevated CBF in non-SI lambs is an autoregulatory response to the substantial hypoxemia these lambs were suffering, with the goal of sustaining oxygen delivery to the brain (O2 delivery is a product of blood flow and O2 content). Indeed, although oxygen delivery tended to decrease and be lower in non-SI lambs than in SI lambs during ventilation in air, these differences were not significant.

An interesting consequence of the higher CBF in non-SI lambs was that a rapid increase in oxygenation caused by increasing the FiO2 dramatically increased cerebral oxygen delivery in these lambs. Although the consequences of such a large and rapid increase in cerebral oxygen delivery are unknown, the potential to cause hyperoxia-induced brain injury, particularly in the presence of significant acidaemia, is high. In contrast, an increase in FiO2 in SI lambs did not affect cerebral oxygen delivery because the arterial blood was already highly saturated with oxygen before the FiO2 increased. Thus, the small increase in O2 content could be accounted for by a small decrease in CBF. It is important to recognize that we did not directly measure CBF or oxygen delivery in this study. Although measurements of carotid flow provide a valid (19,20) and commonly used (21,22) means of measuring relative changes in total brain blood flow, it provides no information on regional blood flow or oxygenation. The use of near-infrared spectroscopy to directly measure CBF and oxygenation or microspheres to assess regional blood flow would provide more detailed information on the changes in cerebral hemodynamics and tissue oxygenation. These data have important implications for the resuscitation of human neonates. They emphasize the dangers that are inherent in rapidly increasing the inspired FiO2 in response to clinical signs of cyanosis or low saturation measurements as indicated by pulse oximetry. Clinicians need to be cognizant of the physiological adaptations that have occurred in response to hypoxemia in the first minutes of life before instituting changes in treatment, such as rapidly alternating the inspired O2 content.

In summary, the findings of this study clearly demonstrate that a SI does not impede the PBF increase and markedly improves the respiratory adaptation to the onset of pulmonary ventilation after birth. The improvement in respiratory function in SI lambs was likely due to a much greater degree of lung aeration with fewer regions with airway liquid retention. Improved aeration increases lung compliance (7) and enhances the lung's gas exchange potential by increasing the surface area available for gas exchange and improving ventilation/perfusion matching. Furthermore, our studies demonstrate that physiological adaptations, such as an increase in CBF, can minimize the effect of hypoxemia on cerebral oxygen delivery immediately after birth. Thus, simply increasing the FiO2 to correct hypoxemia can potentially have injurious effects on the brain caused by hyperoxia. The improved oxygenation we observed after an initial SI suggests that adopting such a strategy might reduce the need for such interventions in human neonates and may also provide protection against neonatal brain injury.