Randomized Trial of Oxygen Saturation Targets during and after Resuscitation and Reversal of Ductal Flow in an Ovine Model of Meconium Aspiration and Pulmonary Hypertension

Neonatal resuscitation (NRP) guidelines suggest targeting 85–95% preductal SpO2 by 10 min after birth. Optimal oxygen saturation (SpO2) targets during resuscitation and in the post-resuscitation management of neonatal meconium aspiration syndrome (MAS) with persistent pulmonary hypertension (PPHN) remains uncertain. Our objective was to compare the time to reversal of ductal flow from fetal pattern (right-to-left), to left-to-right, and to evaluate pulmonary (QPA), carotid (QCA)and ductal (QDA) blood flows between standard (85–94%) and high (95–99%) SpO2 targets during and after resuscitation. Twelve lambs asphyxiated by endotracheal meconium instillation and cord occlusion to induce MAS and PPHN were resuscitated per NRP guidelines and were randomized to either standard (85–94%) or high (95–99%) SpO2 targets. Out of twelve lambs with MAS and PPHN, six each were randomized to standard and high SpO2 targets. Median [interquartile range] time to change in direction of blood flow across the ductus arteriosus from right-to-left, to left-to-right was significantly shorter with high SpO2 target (7.4 (4.4–10.8) min) compared to standard SpO2 target (31.5 (21–66.2) min, p = 0.03). QPA was significantly higher during the first 10 min after birth with higher SpO2 target. At 60 min after birth, the QPA, QCA and QDA were not different between the groups. To conclude, targeting SpO2 of 95–99% during and after resuscitation may hasten reversal of ductal flow in lambs with MAS and PPHN and transiently increase QPA but no differences were observed at 60 min. Clinical studies comparing low and high SpO2 targets assessing hemodynamics and neurodevelopmental outcomes are warranted.


Introduction
Successful transition of the fetus to extrauterine life involves a rapid increase in pulmonary blood flow (Q PA ) during the first few breaths after birth, allowing the lungs to establish as the site of gas exchange. A failure of this transition can lead to persistent pulmonary hypertension of the newborn (PPHN), characterized by sustained elevation of pulmonary vascular resistance (PVR), right-to-left shunting of blood across the foramen ovale and ductus arteriosus, and reduced Q PA [1]. These newborns experience severe respiratory distress and labile hypoxemia soon after birth.
Oxygen (O 2 ) mediates decrease in PVR after birth and has been used to correct the hypoxemia in PPHN, along with strategies to improve lung inflation with respiratory support [2,3]. Current neonatal resuscitation guidelines recommend initiating ventilation with 21% O 2 with subsequent O 2 titration to target goal preductal pulse oximetry O 2 saturation (SpO 2 ) ranges corresponding to the minute of life, to achieve 85-95% SpO 2 by 10 min after birth. This strategy has been associated with optimal hemodynamics and gas exchange during resuscitation [4]. Rawat et al. reported that targeting 95-99% SpO 2 in the post-resuscitation period lowered PVR and improved cerebral O 2 delivery, while targeting 85-89% SpO 2 increased PVR, and decreased Q PA and cerebral O 2 delivery in a term ovine model of meconium aspiration syndrome (MAS) with PPHN [5]. However, flow across the ductus arteriosus (Q DA ) was not evaluated in this study and ductal flow is a major contributor to Q PA [6]. The optimal preductal SpO 2 target range during resuscitation and post-resuscitation period that hastens reversal of shunting across the patent ductus arteriosus (PDA) from the fetal pattern to the postnatal pattern of left-to-right remains unknown.
We hypothesized that the time to reversal of shunting across the PDA is shorter with high SpO 2 target of 95-99% compared to standard SpO 2 target of 85-94%. Our objective was to compare the time to reversal of shunt across the PDA from the fetal (right-to-left) to the postnatal (left-to-right) pattern between standard (85-94%) and high (95-99%) SpO 2 target ranges during resuscitation and the post-resuscitation period in a term ovine asphyxiated model of MAS and PPHN. We also evaluated the changes in Q PA , Q DA and carotid blood flow (Q CA ), and gas exchange at 10 min and 60 min after birth between low and high SpO 2 targets as secondary outcomes.

Materials and Methods
The protocol was approved by the Institutional Animal Care and Use Committee (IACUC, protocol #20267) at the University of California Davis, CA, USA. This protocol involves a perinatal model of MAS and PPHN in term newborn lambs that has been extensively described previously [5,7]. All experiments were performed in accordance with animal ethical guidelines (ARRIVE) [8]. Time-dated near-term (139-141 days) gestation pregnant ewes from Van Laningham Farm (Arbuckle, CA, USA) underwent cesarean section following overnight fasting, after endotracheal intubation under general anesthesia with IV diazepam and ketamine, and inhaled 2% isoflurane, as previously described [9].

Fetal Instrumentation
The fetal lamb was partially exteriorized and intubated with a 4.5-mm cuffed endotracheal tube (ETT), the lung fluid was passively drained by gravity, and the ETT was occluded to prevent entry of air. The lamb was instrumented under maternal anesthesia after subcutaneous bupivacaine infiltration. Catheters were inserted into the right carotid artery and right jugular vein for preductal arterial blood draws, invasive blood pressure and heart rate monitoring, and IV access respectively. A flow probe (Transonics, Ithaca, NY, USA) was placed around the left carotid artery to measure blood flow. A left thoracotomy was performed, and flow probes were placed around the left pulmonary artery and ductus arteriosus to measure blood flows. Subsequently, the thoracotomy and neck incisions were closed in layers. The baseline hemodynamics were recorded and arterial blood gases were obtained.

Meconium Instillation, Asphyxia and Resuscitation
Fetal lambs were asphyxiated following instrumentation by umbilical cord occlusion (by manual compression) for 5 min or until heart rate decreased below 40 beats per minute. Meconium (5 mL/kg of 20% meconium suspended in ewe amniotic fluid) was simultaneously instilled into their endotracheal tube as previously described [7,10]. During asphyxiation, the lambs gasped and spontaneously aspirated the meconium into their lungs. The cord compression was released for 2 min to allow hemodynamic recovery, followed by another 5-min cord occlusion.
Lambs were randomized to standard SpO 2 target (85-94%) or high SpO 2 target (95-99%) using opaque envelopes prior to the beginning of the study (incision for cesarean section). The lambs were then delivered and ventilated with peak inflation pressures (PIP) of 30-35 cm H 2 O, PEEP of 5 cm H 2 O, rate of 40 breaths per minute and inspired O 2 of 21% that was then titrated based on preductal SpO 2 per Neonatal Resuscitation Program (NRP) guidelines during resuscitation to achieve a goal of 85-94% (standard target arm) [11] or titrated to achieve a goal of 95-99% SpO 2 after delivery (high target arm). The titration of oxygen was proportional to the difference between observed SpO 2 and target SpO 2 and was performed every minute. The endotracheal tube was connected to a ventilator and PaCO 2 was targeted in the 40-60 mm Hg range to allow permissive hypercapnia. The resuscitators were not blinded to the intervention. Hemodynamics were continuously monitored, and arterial blood gases were obtained at baseline, 10-min and subsequently at 15-min intervals. Lambs were monitored for up to 60 min and were finally euthanized using IV pentobarbital (Fatal-Plus, Vortech Pharmaceuticals, Dearborn, MI, USA).

Primary and Secondary Outcomes
Primary outcome measures were time to reversal of ductal shunt (from right-to-left to left-to-right) from the time of delivery.
Secondary outcome measures were changes in Q PA , Q DA and Q CA , and gas exchange at 10 min and 60 min after birth between standard and high SpO 2 targets. Cerebral oxygen delivery (mL/kg/min) was calculated by multiplying carotid artery oxygen content (CaO 2 = (1.34 × Hemoglobin in g/dL × SaO 2 %/100%) + (partial pressure of O 2 in mm Hg × 0.0031)) and left carotid artery blood flow (mL/kg/min).

Data Collection and Analysis
Hemodynamic variables were continuously monitored and recorded using BIOPAC systems data acquisition software (Goleta, CA, USA). Blood gases were analyzed using a blood gas analyzer (Radiometer ABL90 FLEX, Denmark). By convention, the right-to-left direction of Q DA was labeled as negative and left-to-right direction of Q DA was labeled as positive. Categorical data were analyzed using chi-squared test with Fisher's exact test as appropriate, parametric continuous data were analyzed using unpaired t-test, and changes in Q DA and Q PA over time were compared using repeated measures ANOVA. The median time to reversal of shunting (non-parametric) was compared between standard and high SpO 2 targets using Wilcoxon rank sum test. Statistical significance was defined as p < 0.05.

Results
Out of twelve near-term lambs that were asphyxiated, six were randomized to standard SpO 2 target and the remaining six were randomized to high SpO 2 target. Hemodynamic and arterial blood gas characteristics at fetal baseline prior to asphyxia were similar between the study groups (Table 1).

Time to Reversal of Shunt across the PDA
Median (interquartile range) time to transition from right-to-left to exclusive leftto-right shunting across the ductus arteriosus was significantly shorter with high SpO 2 target (7.4 (4.4-10.8) min) compared to standard SpO 2 target (31.5 (12-66.2) min, p = 0.03, Figure 1). The mean Q DA (left-to-right) flow was increased with the high SpO 2 targets from 0.5 to 10 min by ANOVA repeated measures. The Q DA significantly increased from fetal baseline to 5-min after birth in both standard and high SpO 2 targets (p < 0.05) and decreased significantly by 60 min in high SpO 2 target (p < 0.05). Table 1. Comparison of fetal baseline hemodynamic and arterial blood gas parameters and end-asphyxia hemodynamic parameters in a near-term ovine model of meconium aspiration syndrome (MAS) and persistent pulmonary hypertension (PPHN) randomized to standard (85-94%) and high (95-99%) preductal SpO 2 target groups.

Time to Reversal of Shunt across the PDA
Median (interquartile range) time to transition from right-to-left to exclusive left-toright shunting across the ductus arteriosus was significantly shorter with high SpO2 target (7.4 (4.4-10.8) min) compared to standard SpO2 target (31.5 (12-66.2) min, p = 0.03, Figure 1). The mean QDA (left-to-right) flow was increased with the high SpO2 targets from 0.5 to 10 min by ANOVA repeated measures. The QDA significantly increased from fetal baseline to 5-min after birth in both standard and high SpO2 targets (p < 0.05) and decreased significantly by 60 min in high SpO2 target (p < 0.05). However, by 60 min after delivery, the Q DA was not different between the two SpO 2 targets. The QDA significantly increased in both high and standard SpO 2 targets from fetal baseline to 5-min after birth, and significantly decreased by 60 min after birth in the high SpO 2 target.

Comparison of Hemodynamics and Arterial Blood Gas Parameters at 5 and 10 min after Birth
During the first 10 min after delivery, the Q DA (left-to-right, Figure 1, p = 0.002) and Q PA (Figure 2, p = 0.048) were significantly higher with high SpO 2 target compared to standard SpO 2 target by ANOVA repeated measures. The Q PA significantly decreased from 5-min to 60-min after birth with high SpO 2 target. The hemodynamic parameters at 5 min and at 10 min time points after delivery are depicted in Tables 2 and 3 respectively and were not different. The arterial blood gas parameters were not different at 10-min after delivery (Table 3). Although the PaO 2 and SpO 2 were higher in the high SpO 2 target group, these differences did not reach statistical significance.

Comparison of Hemodynamics and Arterial Blood Gas Parameters at 5 and 10 min after Birth
During the first 10 min after delivery, the QDA (left-to-right, Figure 1, p = 0.002) and QPA (Figure 2, p = 0.048) were significantly higher with high SpO2 target compared to standard SpO2 target by ANOVA repeated measures. The QPA significantly decreased from 5-min to 60-min after birth with high SpO2 target. The hemodynamic parameters at 5 min and at 10 min time points after delivery are depicted in Tables 2 and 3 respectively and were not different. The arterial blood gas parameters were not different at 10-min after delivery (Table 3). Although the PaO2 and SpO2 were higher in the high SpO2 target group, these differences did not reach statistical significance.  . Delivery of lamb is indicated by the dashed, vertical line. The QPA was significantly higher from 0.5 min to 10 min after delivery with high SpO2 target compared to standard SpO2 target (p = 0.048). However, by 60 min after delivery, the QPA was not different between the two SpO2 targets. * p < 0.05.  . Delivery of lamb is indicated by the dashed, vertical line. The Q PA was significantly higher from 0.5 min to 10 min after delivery with high SpO 2 target compared to standard SpO 2 target (p = 0.048). However, by 60 min after delivery, the Q PA was not different between the two SpO 2 targets. * p < 0.05.

Comparison of Hemodynamics and Arterial Blood Gas Parameters at 60 min after Birth
There was no significant difference in mean Q CA , Q DA , or Q PA blood flows at 60 min after birth (Table 4). One lamb randomized in the standard SpO 2 group maintained a bidirectional Q DA shunt throughout the study period. There were no significant differences in inspired O 2 concentration, arterial O 2 content, cerebral O 2 delivery or PaO 2 between the standard and high SpO 2 target groups at 60 min after birth. However, the pH was higher (p = 0.046) and PaCO 2 was lower (p = 0.042) with high SpO 2 target at the 60 min timepoint.
An illustration summarizing the hemodynamics and gas exchange at 60 min after birth is presented as Figure 3A,B.

Discussion
Neonatal PPHN is commonly associated with parenchymal lung disease such as MAS [1]. Mortality rates of 4-33% have been reported for neonatal PPHN [12]. Rightto-left shunting via the PDA and PFO resulting in labile hypoxemia is characteristic of PPHN. Supplemental O 2 can decrease PVR thus improving Q PA , hastening reversal of extrapulmonary shunting to left-to-right. We demonstrate quicker reversal of ductal shunting to left-to-right with higher (95-99%) compared to the lower (85-94%) SpO 2 target during resuscitation and in the post-resuscitation period.
The American Academy of Pediatrics (AAP) NRP Textbook of Neonatal Resuscitation recommends initiating resuscitation with 21% O 2 , titrating inspired O 2 to achieve the target SpO 2 range for every minute, and finally 85-95% by 10 min after birth. Maintaining SpO 2 within the range recommended by NRP by actively titrating the inspired O 2 led to effective oxygenation and Q PA in an ovine model of MAS [4]. However, there is wide variation among neonatologists in their preference of SpO 2 targets immediately after the initial resuscitation period [13,14]. In a recent animal study by Rawat et al., higher SpO 2 target of 95-99% resulted in lower PVR, higher Q PA , higher cerebral O 2 delivery with lower lactate levels, but was associated with higher inspired O 2 and higher lung 3-nitrotyrosine, a marker of oxidative stress compared to lower ranges of target SpO 2 [5]. However, the authors did not evaluate direction of ductal flow and Q DA in that study.
During neonatal resuscitation, O 2 needs to be titrated judiciously to avoid hypo or hyperoxia. Kapadia et al. reported that post-resuscitation hyperoxia with perinatal acidemia is associated with higher incidence of moderate to severe hypoxic ischemic encephalopathy [15]. Thus, it is prudent to decrease supplemental O 2 in a timely manner to avoid hyperoxemia. We show that the benefits of targeting higher SpO 2 are transient and Q PA is identical between the two study groups by 60 min after birth. We also observed lower PaCO 2 in the high SpO 2 target group at 60 min after birth. We speculate that the higher SpO 2 target may have caused a transient surge in Q PA (Figure 2) that might have contributed to better gas exchange and lower PaCO 2 ( Figure 3A,B). Additionally, the initial increase in left-to-right Q DA and increase in Q PA were not persistent at 60 min after birth (Figures 1 and 2). We speculate that reduction in Q PA could be secondary to transient effect of increased inspired O 2 on PVR or ductal constriction limiting Q DA leading to lower Q PA . As the ductal flow remained steady between 5 min and 60 min after birth (Figure 1), we suspect that the PDA remained patent at 60 min in these lambs. We speculate that higher SpO 2 target is associated with earlier ductal narrowing and less cardiac dysfunction with quicker cardiovascular recovery from the asphyxial insult resulting in better mean arterial pressures (MAPs). Whereas in the standard SpO 2 target group, there may be slower ductal constriction and cardiovascular recovery from the asphyxial insult resulting in lower MAPs. Higher volume of shunt from the aorta-to-pulmonary artery is likely to expose ductal tissue to higher PaO 2 and hasten narrowing of the ductus. Furthermore, the diastolic BP were higher with high SpO 2 target at 60 min after delivery when compared to standard SpO 2 target (65 ± 13 vs. 45 ± 6, p = 0.02), possibly due to ductal constriction, thus contributing to higher MAPs.
Our study has several limitations. The severity of asphyxia was mild to moderate. Despite this, we did demonstrate significant degree of hypoxemia, low Q PA and rightto-left shunting across the PDA following delivery in all the lambs (Tables 1 and 2 and Figure 1). We did not measure the pulmonary artery and left atrial pressures, and hence could not evaluate the PVR. In addition, we did not evaluate the shunting across the PFO.
We had a small sample size and larger number of lambs might have led to different results. The markers of oxidative stress have not been evaluated. We did not recover the lambs and assess for long term outcomes. Although we targeted to achieve preductal SpO 2 within a narrow set range, the achieved SpO 2 was 88 ± 7% and 95 ± 3% in the standard and high SpO 2 target groups, respectively. This difficulty in achieving a set SpO 2 target range has previously been demonstrated in randomized trials in preterm infants [16,17].
To our knowledge, this is the first comparison of time to reversal of ductal shunting in PPHN between standard and high SpO 2 targets during resuscitation and in the postresuscitation period. The novel aspect of this study is the measurement of Q DA along with the direction of flow in a large mammalian model of MAS and PPHN closely mimicking human cardiopulmonary physiology. The lambs underwent meconium aspiration by spontaneous aspiration with negative pressure in the perinatal period during transition rather forceful instillation of meconium in the postnatal (1-3 days, often the case in piglet models) period [18]. Although an ovine model of PPHN can be induced by prenatal ductal ligation, we are unable to assess Q DA in that model. Finally, the randomized study design is a major strength of this study.

Conclusions
In this term lamb model of MAS and PPHN, targeting SpO 2 at a higher range (95-99%) during resuscitation and in the immediate post-resuscitation period, led to a quicker transition to left-to-right shunting across the PDA but did not result in sustained increase in pulmonary blood flow. These findings support the current NRP recommendations to target preductal SpO 2 between 85-95% by 10 min. Clinical trials evaluating hemodynamics and long-term neurocognitive outcomes in term neonates comparing current recommended range to higher SpO 2 targets are warranted in patients at risk of PPHN such as MAS with asphyxia and congenital diaphragmatic hernia.

Data Availability Statement:
The data presented in this study are available in this article.

Conflicts of Interest:
The authors declare no conflict of interest. S.L. is a member of the AAP NRP steering committee. The views expressed in this article are his own and does not represent the official position of AAP or NRP. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.