Prenatal prediction of neonatal haemodynamic adaptation after maternal hyperoxygenation

Studies have shown that fetal pulmonary vasculature reacts to maternal hyperoxygenation (MH) [4, 11, 12]. Oxygen induces the release of several vasodilators including endothelium-derived nitric oxide and prostacyclin, resulting in a decrease in PVR and an increase in pulmonary blood flow [13]. Following maternal oxygen administration, a decrease in the fetal PVR, as demonstrated by the pulmonary artery (PA) Doppler, is deemed to indicate vasoreactivity in the pulmonary vascular bed [14]. The reactivity of the PA to changes in fetal oxygen tension can be detected by noninvasive Doppler ultrasound techniques between 31 and 36 weeks GA [4]. Nomograms of pulmonary reactivity induced by hyperoxia during gestation have been established [13]. The measurement of pulmonary velocity waveforms before and after MH may therefore help in predicting how the fetus will adapt to the extra-uterine environment and transition to neonatal life. If there is a failure of the normal circulatory transition in the early newborn period, persistence of the fetal circulation occurs, resulting in pulmonary hypertension, low oxygen levels and may result in right-to-left shunting of blood in the newborn heart [2]. Some degree of pulmonary hypertension complicates the course of more than 10% of all neonates with respiratory failure [15]. Approximately 10% of neonates will require some form of clinical intervention at birth with 1% requiring more extensive resuscitation [16]. It is vitally important for clinicians to understand the changes that occur during the transition to neonatal life and to predict which neonates may have difficulty transitioning [16]. Prediction of neonatal pulmonary hypertension may influence the delivery planning for particularly high-risk cases and help guide the pharmacological and neonatal intensive care strategies that optimise postnatal survival. The ability to predict the fetal transition to neonatal life by a method that is non-invasive and reproducible would be beneficial for obstetric management, for parental counselling and for determining optimal neonatal management.

In this explorative prospective study, we hypothesise that the degree of the fetal pulmonary vascular bed response to maternal hyperoxia assessed using pulmonary artery Doppler ultrasound at ≥31 weeks’ gestation can predict transition during the early neonatal period.

To evaluate changes in the fetal pulmonary artery Doppler following maternal hyperoxygenation during the third trimester of pregnancy and to correlate those findings with neonatal echocardiograph indices indicative of increased pulmonary systolic pressures, including estimates of right ventricular systolic pressures, PA acceleration time in addition to measurements of neonatal ejection fraction and left ventricular output.

Image-directed pulsed and colour Doppler equipment (Voluson E8) was used with a 5-MHz sector probe. All Doppler recordings were obtained using the lowest high-pass filter level (100 Hz), and the spatial peak temporal average power output for colour and pulsed Doppler was kept at < 100 mW/cm [27]. An angle of ≤15° between the vessel being studied and the Doppler beam was deemed acceptable and used for analysis. All participants underwent a standard ultrasound examination for estimation of fetal weight, amniotic fluid volume measurements, fetal heart rate and umbilical (UA) and middle cerebral artery (MCA) Doppler assessment. A fetal echocardiogram was performed according to the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) guidelines for fetal echocardiography [28] on fetuses between 31 and 40 weeks gestation. This involved a sequential segmental analysis of the atria, ventricles, and great arteries and their connections. Specific echocardiographic Doppler studies included those of the distal main PA and ductus arteriosus (DA) (Figs. 1 and 2). The following measurements specific to the PA Doppler waveform were recorded: The peak systolic velocity (PSV), end diastolic velocity (EDV), time-averaged velocity (TAV), pulsatility index (PI), resistance index (RI), ejection time (ET) and acceleration time (AT). The ductus arteriosus waveform was obtained in the traditional longitudinal ductal arch view. Reported values were averaged from three consecutive waveforms. Oxygen was administered to the subjects while in a semi-recumbent position in the hospital ultrasound department, at a rate of 12 L/min for a duration of 10 min via a partial non-rebreather mask. Immediately following MH, a repeat fetal echocardiogram was performed, and all Doppler recordings were repeated. Each fetus served as its own control. Recordings were stored on the ultrasound machine for further analysis and for data safety monitoring purposes. The hyperoxygenation test was considered positive when the PI of the fetal PA decreased by ≥10% from its baseline (responders). Where the fetal PA PI did not decrease by at least 10%, cases were classified as non-responders. The fetal PA Doppler measurements were kept in a specific study file and not in the subject’s hospital file. The calculation of fetal PA reactivity was performed after the ultrasound assessment, when the subject had left the ultrasound room. The subject was not made aware of the prenatal PA reactivity results.

A functional neonatal echocardiogram was performed within the first 24 h of life. Neonates were divided into those that responded to MH in utero (responders) and those that did not respond (non-responders). Echocardiography was carried out using a Vivid S6 echocardiography machine and a 7 MHz neonatal probe (GE Medical, Milwaukee, USA). Studies were conducted during a resting state in accordance to recent guidelines and congenital heart disease was excluded during the first scan [29]. Data were stored as raw DICOM images in an archiving system (EchoPac, General Electric, version 112 revision 1.3) and analysis of all the echocardiography parameters was carried out by a single investigator who was blinded to the results of fetal Doppler ultrasounds (N.B). The following echocardiography parameters were also measured using previously described methods [30], left ventricular (LV) length measured at end diastole; mitral value annual diameter; left ventricular output (LVO, ml/kg/min); ejection fraction using Simpson’s Biplane method (EF, %); mitral valve inflow velocities and velocity time index, patent ductus arteriosus (PDA) diameter (mm) measured in 2D at the pulmonary end; diastolic and systolic flow velocity across the PDA; flow pattern across the duct; pulmonary artery acceleration time (PAAT); and right ventricular (RV) end systolic pressure measured using the tricuspid valve regurgitant jet.

Sample sizes of between 24 and 50 have been recommended variously for pilot studies [31‐38]. Following these recommendations, we chose a recruitment sample size of 40–60 which would allow for a moderate dropout rate. A significant dropout rate (e.g. 40%) would reduce the pilot sample size to below a minimum of 24, in which case a possible larger study would be called into question in the first place, having possible external validity issues, pragmatic or ethical concerns. Data analysis was performed using SPSS software (version 24.0; IBM Corporation, Armonk, NY). Continuous data were expressed as means and standard deviations or medians and interquartile ranges as appropriate. Independent data were compared using the independent Student t-test or the Mann-Whitney U test as appropriate. Paired data were compared using the paired Student t-test or the non-parametric equivalent as appropriate. Categorical data were compared using the Chi Square test or the Fisher Exact test. Data were deemed statistically significant at a p-value < 0.05.

Intraobserver and interobserver variability of fetal PA Doppler indices were assessed using a subset of 10 patients. One reader (A.M) repeated measurements at a time temporally remote from the initial assessment. To assess interobserver variability, a second reader (F.B), blinded to the original data, repeated PA Doppler measurements. Intraobserver and interobserver variability was assessed by calculation of mean percent error, defined as the absolute difference between observations divided by the mean of observations and using the intraclass correlation coefficient (ICC) and 95% confidence intervals (95% CI). To assess the repeatability of the measured values, the mean and SD of differences and the repeatability coefficient of the two repeated tests within subjects were calculated. The PA Doppler indices that are reported in the results were taken by a single operator (A.M). The study was not powered as a predictive tool and therefore results may not be reflective of a cause and effect relationship.

To our knowledge, this is the first study to compare the effects of MH with neonatal echocardiography evaluation, in normal pregnancies. All fetal Doppler measurements included in the study results were performed by a single operator (A.M). All neonatal echocardiography measurements were carried out by a single investigator who was blinded to the results of the fetal Doppler assessments (N.B).

We acknowledge the small sample size in our study. This warrants further exploration in a larger cohort to establish the ability of the fetal response to MH to predict postnatal adaptation of the pulmonary vascular circulation during the transition to neonatal life.