Early Echocardiography Predicts Intervention Need in Antenatal Suspicion of Coarctation of the Aorta

Abstract

Background

Coarctation of the aorta (CoA) is challenging to diagnose in early postnatal life. We aimed to describe the resource utilization and predictors for the need of intervention in an antenatal suspicion of CoA.

Methods

A retrospective study of infants with an antenatal suspicion of CoA born at ≥37 weeks was performed. Those not requiring intervention (normal) were compared with those who required cardiac surgery (CoA). Strain was measured using speckle-tracking echocardiography.

Results

A total of 51 newborns were included; 40 (78%) were considered normal and 11 (22%) underwent intervention. Echocardiography occurred within the first day of life for both groups. Right ventricular (RV) predominance was present in the CoA group, as demonstrated by the left ventricular (LV) end-systolic eccentricity index (1.60 [0.28] vs 2.16 [0.45]; P < 0.001) and by a larger RV end-diastolic area (EDA) in apical 4-chamber (A4C) relative to LV-EDA—with a ratio of 1.56 [0.23] vs 1.02 [0.2]; P < 0.001. An RV/LV EDA ratio in A4C ≥1.3 had a high probability for CoA (area under the curve = 0.97). Newborns with CoA had a lower RV deformation (peak systolic strain rate: −0.98 [0.17] vs −0.83 [0.2]; P = 0.02). Intraclass correlation coefficient for the EDA ratio revealed a good inter-rater agreement (0.76; 95% confidence interval: 0.55-0.87). Analysis with rater #2 revealed that the EDA ratio ≥1.3 predicted 100% of CoA.

Conclusions

The majority of those with an antenatal suspicion of CoA did not require intervention but were high consumers of resources. Within the first day of life, the ventricular EDA ratio in A4C may help predicting those with true CoA requiring intervention.

Graphical abstract

Coarctation of the aorta (CoA) is a ductal-dependent congenital heart disease (CHD) that is challenging to confirm in the prenatal and early postnatal life. Its prevalence is 1.8-3.4 per 10,000 live births, and it represents 5% of all CHD.^1,2 A CoA is a local narrowing of the aortic lumen, most commonly located at the level of the aortic isthmus. Because of patency of the ductus, a coarctation may be missed during fetal life, as well as during the early neonatal period. Without early recognition and upon ductal closure, these infants are at risk of cardiovascular collapse and end-organ damage, which may lead to death.^3,4 Studies have described various indirect fetal echocardiographic findings raising suspicion for underlying CoA; however, the false-positive rate remains high.5, 6, 7, 8 Thus, an antenatal suspicion for CoA may require admission to the neonatal, paediatric, or cardiac intensive care unit (ICU), delays parental bonding, postpones achievement of full enteral feeds, increases the length of hospital stay, consumes medical resources, and increases the risk for iatrogenic complications.^9,10 In clinical practice, many of these infants are exposed to serial echocardiography scans, as well as a prolonged stay in the ICU, awaiting complete ductal closure to confirm or exclude the CoA diagnosis. In this study, we sought to describe early postnatal echocardiography predictors for CoA requiring a neonatal intervention. Further, we sought to describe resource utilization in this population of infants, such as the duration of hospital stay, the need for multiple echocardiography scans, the use of feeding assistance, and the exposure to intravenous therapy. Our hypothesis was that newborns with a prenatal suspicion of coarctation with a need of postnatal intervention, compared with those without a need of postnatal intervention, would have on their first postnatal echocardiography: (a) signs of right ventricular predominance and (b) decreased right ventricular peak longitudinal strain by speckle-tracking echocardiography (STE).

Methods

This is a single-centre retrospective observational cohort study. Infants born at ≥37 weeks of estimated gestational age (GA), between January 2014 and March 2018, and with an antenatal diagnosis of CoA requiring admission to our neonatal intensive care unit (NICU) were included. Antenatal suspicion for the diagnosis of CoA or hypoplastic aortic arch was based on fetal echocardiography performed by an expert in fetal cardiology. Recommendation for postnatal admission to our NICU was based on the size of the arch prenatally, or concomitant other indirect markers such as right ventricular predominance and the presence of a left-sided persistent superior vena cava. We excluded infants who were admitted after 24 hours of age, had other major cardiac anomalies (except an atrial or ventricular septal defect, a patent ductus arteriosus [PDA], or a bicuspid aortic valve), genetic disorder, and other complex malformation. We applied a similar echocardiography methodology for data extraction, as reported by a previous publication by our group.¹¹ First available clinically acquired echocardiography was reviewed offline for data extractions using a Syngo Dynamics workstation (Siemens Medical Solutions, Erlangen, Germany). Experts extracting data from echocardiographic images were masked to the outcome of the patient, as well as to the values of the other data extractor. However, both extractors were not blinded to the study question. Images were acquired using a GE Vivid E9 (Wisconsin). Images were acquired and parameters were measured according to the American Society of Echocardiography guidelines¹² (see Supplemental Appendix S1). The study was approved by the McGill University Health Centre research ethics boards.

Statistical analysis

Results are described as mean with standard deviation or median with interquartile range (IQR) for continuous variables and counts with proportions for categorical variables. Newborns with an antenatal suspicion of CoA who were free of arch anomalies after ductal closure (controls) were compared with those who required a surgery in the postnatal period (CoA). The Fisher exact test was used to assess differences in categorical characteristics, whereas the Student t-test and Wilcoxon-Mann-Whitney test were used for continuous variables with parametric and nonparametric distributions, respectively. Receiver operating characteristic (ROC) curves were generated and the Youden index was used to determine the optimal cutoff value for the right ventricular (RV)-left ventricular (LV) end-diastolic area (EDA) ratio (Supplemental Fig. S1), the aortic isthmus diameter, and the RV–peak longitudinal strain in apical 4-chamber (pLS) associated with a CoA. Intraclass correlation coefficient, using a 2-way random-effects model, was used to assess inter-rater variability on a random sample of 44 infants (inclusive of all infants who required intervention).¹³ Statistical analyses were done with Stata SE (Version 14.2; StataCorp, College Station, TX) and RStudio (International Open-Source Collaborative). A P value of <0.05 was considered statistically significant.

Results

A total of 216 infants were identified with a suspicion of CoA or hypoplastic aortic arch from a review of an institutional echocardiography database, and 51 met inclusion criteria (Fig. 1). Of these, 40 (78%) had an arch with nonsignificant obstruction on postnatal echocardiography after ductal closure and 11 (22%) eventually underwent a neonatal intervention for a confirmed CoA. Baseline characteristics were similar between both groups (Table 1). GA at first fetal cardiology encounter was similar between groups (31.7 [5.4] vs 28.2 [6.4] weeks; P = 0.1), as well as GA at birth (39.2 [1.1] vs 39.7 [1.0] weeks; P = 0.2). At NICU admission, traditional clinical signs of left-sided obstruction, as documented by the clinical team, were not different between both groups. As such, there was no difference in the quality of the perceived femoral pulses (described as “poor” or “adequate”) (P = 0.30), the presence of a pre- to postductal systolic blood pressure difference of $\ge$ 10 mm Hg (P = 0.34), or the presence of a difference in the pre- to postductal oxygen saturation $\ge$ 10% (P = 0.38) (Table 1). Among those who required an intervention, 5 underwent a coarctation repair, 4 had both an aortic arch augmentation and coarctectomy, and 2 underwent only an aortic arch repair. The median age at first intervention was of 10 (IQR: 7-18) postnatal days. Of those, 2 infants (18%) had a recurrence and required a second intervention. Only 1 infant, who received a complete neonatal repair, died at 15 months from a sudden cardiac event in the community.

Figure 1.

Consort diagram. A total of 216 infants were identified with a suspicion of CoA or hypoplastic aortic arch from our database. Thirteen infants had genetic disorder or significant anomalies, such as trisomy 21, 13, 18, VACTERL association, and Dandy-Walker syndrome. Twelve infants had complex congenital heart disease (hypoplastic left heart syndrome = 2, atrioventricular septal defect = 2, double outlet right ventricle = 2, transposition of great arteries = 1, tetralogy of Fallot = 1, total anomalous pulmonary venous return = 2, severe aortic stenosis = 1, and large unrestrictive ventricular septal defect with atrial septal defect and patent ductus arteriosus = 1). Therefore, a total of 52 infants born after a pregnancy complicated with antenatal suspicion of CoA were identified. One infant was excluded because of a transfer from an outside centre at 15 days of life. CoA, coarctation of the aorta; DOL, days of life.

Table 1.

Demographic and clinical characteristics

	Normal ECHO (N = 40)	Intervention (N = 11)	P value
Gestation age (GA) at birth (wk)	39.2 (1.1)	39.7 (1.0)	0.20
Male	22 (55)	9 (82)	0.10
Vaginal delivery	29 (73)	9 (82)	0.42
Apgar score at 5 min	9 (9-9)	9 (7-9)	0.34
Birth weight (g)	3343 (552)	3337 (578)	0.97
GA at fetal diagnosis	31.7 (5.4)	28.2 (6.4)	0.10
Duration of hospital stay (d)	4 (2-6)	35 (27-109)	<0.0001
Mortality	0 (0)	1 (9)	0.22
Exposed to PGE infusion	6 (15)	11 (100)	<0.0001
Duration of PGE (d)	0 (0-1)	10 (6-18)	<0.0001
Differential in pre- and postductal BP ≥10 mm Hg on admission∗	13 (35)	4 (50)	0.34
Poor femoral pulses as per clinician on admission∗	6 (15)	3 (27)	0.30
Differential of saturation in pre- and postductal limb ≥10% on admission∗	1 (3)	1 (10)	0.38

Expressed in count (%) and median (interquartile range).

BP, blood pressure; ECHO, echocardiography; GA, gestational age; PGE, prostaglandin E.

^∗Blood pressure (by the oscillometric method) and femoral pulses were performed at the first physical examination on admission.

Medical resource utilization in those without confirmation of a coarctation

Only 1 infant without coarctation was intubated because of respiratory compromise from a meconium aspiration and was extubated to room air the following day. Otherwise, all the other infants remained in room air during their admission. Prostaglandin E (PGE) was initiated immediately after birth in 6 (15%) infants without coarctation, before postnatal echocardiography, and based on the fetal evaluation. The duration of PGE infusion was 0 (IQR: 0-1) day (0 corresponds to an initiation and cessation on the same day) in these newborns. PGE infusion was discontinued after the first postnatal echocardiography, which demonstrated reassuring aortic arch measurements. These infants remained admitted to the NICU for a median of 4 (IQR: 4-6) days until ductal closure confirmation by cardiology (median of 2 echocardiography scans [IQR: 2-3]). During their admission, 38 (95%) infants were on intravenous fluids (median duration of 1 [range: 0-4] day) and 9 (23%) required gavage feeds. The median age at full oral feed was of 2 (range: 0-8) days.

Echocardiographic findings

Echocardiography occurred within the first day of life for both groups. RV predominance was found to be more pronounced in the CoA infants. Indeed, in the apical 4-chamber (A4C), the absolute ventricular EDA of each ventricle indicated an increased RV area (EDA-RV: 4.24 [0.90] vs 4.95 [0.91] cm²; P = 0.03), relative to the LV area (EDA-LV: 4.26 [1.00] vs 3.23 [0.71] cm²; P = 0.002). As such, a larger EDA ratio was observed (1.02 [0.2] vs 1.56 [0.23]; P < 0.001). Furthermore, in the parasternal short axis view, the ratio of the perpendicular diameter of the RV to the perpendicular diameter of the LV at the end of systole at the mid-papillary level indicated RV predominance (D2/D2 ratio; P = 0.008). In addition, the LV end-systolic eccentricity index (Supplemental Fig. S2) was increased (1.60 [0.28] vs 2.16 [0.45]; P < 0.001), indicating septal deformation towards the LV cavity (Table 2). The aortic arch measurements were significantly smaller in the CoA group (ascending: 0.76 [0.12] vs 0.65 [0.12] cm, P = 0.01; proximal transverse arch: 0.58 [0.10] vs 0.38 [0.08] cm, P < 0.001; distal transverse arch: 0.48 [0.09] vs 0.33 [0.07] cm, P < 0.001; isthmus: 0.36 [0.07] vs 0.26 [0.07] cm, P < 0.001), despite the larger ductal diameter in those with a CoA (2.91 [1.25] vs 4.24 [1.79] cm; P = 0.008). The gradient in the descending aorta was higher in the infants with a CoA despite ductal patency (7.3 [4.1] vs 12.7 [7.0] mm Hg; P = 0.007). LV ejection fraction by Simpson’s disc method was similar and indicated adequate function in both groups (64 [11] vs 65 [11] %; P = 0.94). The systolic RV function by tricuspid annular plane systolic excursion and the fractional area change method was also within normal. Deformation analysis (Table 3) indicated decreased RV systolic and diastolic performance in the CoA group by RV-pLS, RV–peak longitudinal systolic strain rate, and RV–early diastolic longitudinal strain rate, but preserved LV functional parameters.

Table 2.

Echocardiography results

	Normal ECHO (N = 40)	Intervention (N = 11)	P value
Age at echocardiography (d)	0 (0-1)	0 (0-1)	0.90
Bicuspid aortic valve	3 (8%)	6 (55%)	0.002
Ventricular septal defect	7 (18%)	5 (45%)	0.10
Measurement in PLAX view (cm)
Aortic opening	0.65 (0.09)	0.58 (0.12)	0.05
Aortic root	0.92 (0.13)	0.83 (0.16)	0.07
STJ	0.74 (0.12)	0.63 (0.08)	<0.05
Ascending aorta	0.85 (0.10)	0.71 (0.13)	0.001
Measurement in the A4C view
EDA RV	4.24 (0.90)	4.95 (0.91)	0.03
EDA LV	4.26 (1.00)	3.23 (0.71)	0.002
EDA RV/LV ratio	1.02 (0.20)	1.56 (0.23)	<0.001
4C-EF (%)	64.30 (10.99)	64.60 (10.83)	0.94
MV annulus (cm)	0.90 (0.15)	0.66 (0.09)	<0.001
TV annulus (cm)	0.98 (0.17)	0.97 (0.14)	0.87
TV/MV annulus ratio	1.28 (0.27)	1.76 (0.16)	<0.001
Measurement in PSAX
Eccentricity index	1.60 (0.28)	2.16 (0.45)	<0.001
D2 RV	1.55 (0.27)	1.89 (0.47)	0.006
D2 LV/D2 RV	0.65 (0.16)	0.46 (0.19)	0.003
PDA size (mm)	2.91 (1.25)	4.24 (1.79)	0.008
Measurement at arch view (cm)
AA	0.76 (0.12)	0.65 (0.12)	0.01
PTA	0.58 (0.10)	0.38 (0.08)	<0.001
DTA	0.48 (0.09)	0.33 (0.07)	<0.001
AI	0.36 (0.07)	0.26 (0.07)	<0.001
Isthmus to PDA ratio	1.50 (0.71)	0.79 (0.62)	0.006
Gradient in descending aorta (mm Hg)	7.25 (4.06)	12.72 (7.04)	0.007

Expressed as N (%), mean (standard deviation), or median (interquartile range).

AA, ascending aorta; AI, aortic isthmus; A4C, apical-4-chamber view; DTA, distal transverse aorta; D2, ventricular diameter perpendicular to the septum at the peak of systole; ECHO, echocardiography; EDA, end-diastolic area; EF, ejection fraction by Simpson’s disc method in the apical 4-chamber view; LV, left ventricle; MV, mitral valve; PDA, patent ductus arteriosus; PLAX, parasternal long axis view; PSAX, parasternal short axis view; PTA, proximal transverse aorta; RV, right ventricle; STJ, sinotubular junction; TV, tricuspid valve.

Table 3.

Strain analysis by speckle-tracking echocardiography in apical 4 chamber view

	Normal ECHO (N = 40)	Intervention (N = 11)	P value
RV pLS	−20.0 (3.5)	−17.5 (4.3)	0.06
RV pLSR	−0.98 (0.17)	−0.83 (0.20)	0.02
RV early diastolic SR	1.05 (0.20)	0.85 (0.26)	0.01
LV pLS	−25.1 (4.6)	−23.8 (4.0)	0.39
LV pLSR	−1.28 (0.24)	−1.21 (0.22)	0.35
LV early diastolic SR	1.31 (0.30)	1.23 (0.31)	0.44

Expressed as mean (standard deviation).

ECHO, echocardiography; LV, left ventricle; pLS, peak longitudinal strain in A4C; RV, right ventricle; SR, strain rate.

The ROC analysis indicated that there was an association between the EDA ratio, aortic isthmus, and RV-pLS with the diagnosis of CoA (Fig. 2). An EDA ratio greater than 1.29 (Youden’s index) demonstrated the highest probability for diagnosis (Fig. 2, 100% sensitivity, 86% specificity, and area under the curve = 0.97). Further, the EDA ratio was found to have good agreement among 2 raters blinded to each other’s value, with an intraclass correlation coefficient of 0.76 (95% confidence interval: 0.55-0.87).¹³ ROC analysis, for the values of the second rater, revealed that a ratio of 1.24 was 100% specific and 97% sensitive for the requirement of intervention, with an area under the curve of 0.98 (Supplementary Fig. S3).

Figure 2.

Receiver operating characteristic (ROC) curve for prediction of coarctation requiring intervention. ROC curve for first postnatal echocardiography thresholds associated with coarctation of the aorta requiring neonatal intervention. A4C, apical 4-chamber; AUC, area under the curve; CoA, coarctation of the aorta; EDA, end-diastolic area; LV, left ventricular; RV-pLS, right ventricular–peak longitudinal strain in A4C.

Discussion

In this cohort of patients with an antenatal suspicion of the coarctation, a majority of them (78%) were without a need for postnatal intervention, while still requiring an NICU admission for a median of 4 days. Most (95%) were exposed to intravenous support, whereas 23% required gavage feeds due to the delayed introduction of enteral nutrition. Clinical signs on admission were not predictive for CoA. Initial echocardiography demonstrated significant differences, including smaller arch diameters, increased markers of RV predominance, and altered indicators of RV performance by STE in those with CoA. An EDA ratio of the RV to the LV of 1.3 or more in the A4C view was the most sensitive marker predictive of CoA. Identification of early markers of reassurance is of importance in the context of antenatal suspicion of CoA, as these may allow for early initiation of postnatal feeds, decrease potential for iatrogenic interventions, reduction in hospitalization cost, and early transfer to a nursery or for rooming in with the family.

In our centre, infants with an antenatal suspicion of CoA are admitted to the NICU for monitoring and investigations, as per American Heart Association recommendations.¹⁴ These infants remain nil per os and are exposed to intravenous support until confirmation of an adequate aortic arch calibre and, often, ductal closure. A previous study reported similar results, outlining that among 50 infants identified prenatally, 47 (94%) were false positives, 38 (76%) were admitted to their NICU, and the median hospital duration was 4 days.¹⁰ Upon admission, traditional clinical signs of left-sided obstruction were not discriminative of those with eventual CoA. The unreliability of clinical findings, often occurring later upon ductal closure, has been reported in other studies.^3,15,16 Hence, these patients should be monitored in the context of intensive care to avoid cardiac failure and end-organ damage. At our centre, an NICU admission is 16,311 CAD$/d, whereas the admission to the nursery (in the mother’s room) is 2757 $/d.¹⁷ Markers predictive of CoA in the presence of an open ductus may identify those with a low diagnostic suspicion, encourage a transfer from the NICU to the nursery for ongoing arch monitoring until ductal closure (potentially decreasing the costs), and promote the introduction of early enteral nutrition (and breastfeeding) in these infants.

Right ventricular dominance

Ventricular disproportion has been described as a fetal marker of CoA.¹⁸ However, half of those with fetal cardiac ventricular disproportion are not found to have a postnatal CoA.^19,20 In addition, the RV appears larger than the LV even in normal foetuses towards the end of the pregnancy.²¹ Aortic arch obstruction during fetal life alters biventricular development and function, making ventricular disproportion an imperfect prenatal predictor of CoA.²² A prospective study on fetal suspicion of CoA described an RV/LV ratio >1.5, using the transversal diameter under the level of the atrioventricular valves at the end of diastole, as predictive of postnatal CoA when combined with other markers (aortic isthmus and ductal measurements).²³ Another group identified several markers of RV/LV disproportion during fetal life markedly different in those with postnatal CoA compared with controls.²² Indeed, during fetal life, those with CoA had a flatter LV and a more globular RV, decreased LV width, increased RV transverse width, and RV and LV depressed markers of contractility. Fetal and postnatal ventricular disproportion are often described subjectively. As such, formal quantifying methods in the postnatal setting are necessary. Our study demonstrated that on their first postnatal echocardiography, infants with CoA presented larger EDA-RV and smaller EDA-LV in A4C, as well as increased RV relative to LV diameter in the parasternal short axis. These findings could be secondary to a relative underlying LV hypoplasia, associated with left-sided anomalies. As such, CoA newborns in our cohort had concomitantly smaller mitral valve (MV) annulus and smaller MV to tricuspid valve (TV) diameter ratio. Further, the ventricular growth during fetal life may be impacted by loading conditions and fetal flows, possibly leading to RV overgrowth relative to the LV. RV width, TV diameter, and pulmonary artery diameter were reported as increased in a cohort of CoA foetuses.²¹ Recently, a report evaluating early postnatal echocardiography in antenatal suspicion of CoA did not find a difference in the RV:LV length ratio and in the TV:MV annulus ratio, but did not evaluate ventricular disproportion using the EDA ratio.²⁴ Another group did not find a difference in the RV and LV estimated volumes in those with and without CoA. These authors used the 5/6 area-length to estimate LV diastolic volume and the 2/3 area-length to estimate RV diastolic volume using retrospectively retrieved 2D echocardiography scans of infants diagnosed with possible but not definitive CoA, in the presence of a PDA by echocardiography.²⁵ However, the EDA ratio was not measured, and 48% of their cohort had an associated CHD.

In our cohort, those with confirmed CoA had a larger PDA diameter at echocardiography, possibly secondary to the increased PGE exposure. Loading conditions in that setting may be altered but should not significantly influence diastolic estimates of RV dimensions relative to the LV. However, this may explain the increased RV to LV diameter ratio at the end of systole found in the parasternal short axis, as well as the increased LV-eccentricity index. Furthermore, although intrinsic RV properties may be different in those newborns with CoA, altered RV deformation by STE (pLS, peak longitudinal systolic strain rate, and early diastolic SR) detected in our cohort may be secondary to the differences in loading conditions relative to the unrestricted ductus. In our cohort, we did not detect any LV functional anomalies (by STE or EF). Seguela et al.²⁶ reported lower LV strain in neonates with CoA, but after ductal closure. As such, arch obstruction significantly increases LV afterload and may alter performance. Furthermore, another study reported reduced LV deformation in CoA infants when evaluated just before intervention, at a median of 4 (1-42) days, which gradually improved after surgical correction.²⁷

This retrospective analysis is limited by the small number of newborns and its single centre format. Echocardiography assessment was limited by the use of multiple professionals acquiring the images and the limited views in some patients. Some of the echocardiography measures were not analyzed for intrareader or inter-reader variability, although we did evaluate the EDA ratio, and our methodology has been validated in previous reports.^11,28 Previous studies described strain assessment to have high inter-reader and intrareader agreement.²⁹ Reports have described a high reproducibility of STE-strain measurements by the same reader.³⁰ In our cohort, 8 infants were born at a birth weight of less than the 10th percentile according to the Fenton growth charts, of whom 2 met criteria for coarctation requiring intervention in the postnatal setting. Infants with intrauterine growth restriction may be born small for GA and may have had abnormal Doppler flows during prenatal life due to growth restriction. Even though these foetuses may have fetal markers similar to coarctation during the fetal setting, we decided to keep them in the analysis as some of them may truly have a diagnosis of coarctation in the postnatal life and were suspected as such by our cardiology experts. Many pregnancies were referred to our centre after evaluation in the community. The first fetal echocardiography was performed at an average of 31 weeks for the nonintervention group and 28 weeks for the intervention group. As such, it is possible that an earlier fetal echocardiography during pregnancy may confer a lower false-positive rate for true coarctation than our cohort. Although our catchment area tends to concentrate around the greater Montreal area, we did not have the information about the exact addresses of these families. As such, in those living remotely from our centre, the recommendation for a delivery and evaluation at our centre might have stemmed from excessive cautiousness, potentially inflating the false-positive fetal echocardiography rate in our cohort. However, previous cohorts have reported similar rates to those described here.¹⁰ Finally, inter-reader reproducibility was not explored. Reproducibility of our findings regarding this ratio in other cohorts, as a predictor of postnatal intervention, will be of high interest for neonatal providers.

Conclusions

Antenatal suspicion of CoA is associated with a high false-positive rate, leading to consumption of numerous medical resources and an increase in the duration of ICU-hospital stay. Echocardiographic markers in early postnatal allow for identification of infants with a low likelihood for CoA upon ductal closure. These markers include ventricular disproportion by the EDA ratio in A4C (1.3 or more), RV-pLS by STE, and aortic isthmus dimension. Future studies should validate these findings in a prospective multicentric cohort. In the meantime, infants with reassuring postnatal signs should be considered for an early transfer from a critical care unit to a nursery, with ongoing monitoring of the arch until ductal closure.

Data Sharing

Derived data generated will be shared on reasonable request to the corresponding author.

Ethics Statement

This research was done in accordance to the Declaration of Helsinki for performance of medical research. The research ethics board of McGill University Health Centre approved this research project.