Echocardiographic Predictors of Hospital Outcomes in Preterm Neonates with Patent Ductus Arteriosus and Clinical Suspicion of Pulmonary Hypertension

Abstract

Background

Preterm neonates often have an echocardiogram performed in the first few days of life for suspicion of pulmonary hypertension and patent ductus arteriosus. The usefulness of this echocardiogram in predicting outcomes in this population are unknown. The objective of this study was to investigate associations between initial echocardiographic assessment and hospital outcomes in preterm neonates with patent ductus arteriosus and clinical suspicion of pulmonary hypertension.

Methods

Premature neonates (<37 weeks gestational age) with patent ductus arteriosus and clinical suspicion of pulmonary hypertension born at our institution or transferred within 48 hours of life were included in this single center retrospective study. The primary outcome was in-hospital extracorporeal membrane oxygenation utilization or mortality.

Results

86 patients were included. Median age at echocardiogram was 2 days (interquartile range 1,7), gestational age was 27 weeks (interquartile range 25,32), and birth weight was 878 grams (interquartile range 650,1818). 15 patients (17%) met the primary outcome. Larger patent ductus arteriosus size (p = .03), patent foramen ovale flow that was bidirectional or right to left (p = .047), and right atrial volume (p = .04) were independently associated with the primary outcome.

Conclusion

Larger patent ductus arteriosus size, bidirectional or right to left flow at the patent foramen ovale, and lower right atrial volume are independently associated with in-hospital mortality. These findings on the initial echocardiogram of a preterm neonate can be used to risk stratify these patients for elevated risk for in-hospital extracorporeal membrane oxygenation utilization or mortality.

Keywords

Pulmonary hypertension premature neonates echocardiography

Introduction

Pre-term delivery affects 1 out of 10 neonates born in the United States.¹ This population has a higher incidence of pulmonary hypertension compared to term infants.² In practice, pulmonary hypertension is a clinical diagnosis characterized by hypoxemia, impaired cardiac output, and need for respiratory support disproportionate to their lung disease. Right heart catheterization is the diagnostic gold standard for the assessment of pulmonary artery pressure and pulmonary vascular resistance; however, it is not routinely performed in the neonatal population due to its invasive nature.³ Echocardiography has become a standard noninvasive tool to confirm clinical suspicion of pulmonary hypertension by estimating pulmonary artery pressure.^{4, 5, 6}

Preterm neonates often have clinical suspicion of pulmonary hypertension and are found to have a patent ductus arteriosus on their initial screening echocardiogram. The clinical paradigm of using echocardiography in the estimation of pulmonary artery pressure to identify patients with increased pulmonary vascular resistance/pulmonary hypertension is confounded in patients with an unrestrictive patent ductus arteriosus, as this physiology results in elevated pulmonary artery pressure even in the absence of elevated pulmonary vascular resistance. In this context, echocardiography’s ability to detect elevated pulmonary vascular resistance by estimating pulmonary artery pressure, and, the prognostic information that comes with it, are lost.

The initial echocardiogram in these patients is a potential opportunity to gain prognostic information outside of estimating pulmonary artery pressure. For example, the negative effects of elevated pulmonary artery pressure on myocardial function has been found to be predictive of poor outcomes in term neonates.^{7, 8} There is an important knowledge gap regarding echocardiography’s ability to predict hospital outcomes in preterm neonates with clinical suspicion of elevated pulmonary vascular resistance and patent ductus arteriosus. Identifying novel echocardiographic tools is imperative to identify high risk neonates and improve outcomes in this ever-growing population.

The objective of this study was to investigate the association between patients’ initial echocardiographic assessment of pulmonary vascular disease versus hospital outcomes in preterm neonates with patent ductus arteriosus and clinical suspicion of pulmonary hypertension. We hypothesized that both measures of right ventricular function and estimates of right ventricular pressure would be associated with hospital outcomes in this population.

Materials and Methods

We retrospectively identified premature neonates (<37 weeks gestational age) with patent ductus arteriosus and a clinical diagnosis of pulmonary hypertension born at our institution or transferred within 48 hours of life. Patients with a major congenital or genetic abnormality, had surgery for a congenital anomaly, or, those with congenital heart disease other than a patent ductus arteriosus were excluded. Demographic information was collected including gestational age, race, sex, birth weight, body surface area, age at time of echo, blood pressure at time of echo, mode of delivery, and whether infant was born at the home institution. Other potentially important clinical factors such as need for positive pressure ventilation, supplemental oxygen, initial hemoglobin, cord pH, APGARs, and presence of necrotizing enterocolitis were recorded. We also recorded medications administered to each neonate. The primary outcome was in-hospital extracorporeal membrane oxygenation utilization or mortality. Secondary outcomes included hospital length of stay and ventilator days. The study was approved by the Medical University of South Carolina Institutional Review Board (Pro00075177).

Echocardiographic Analysis

All studies were performed using Phillips ie33 (Andover, MA) or GE E95 (Waukesha, WI) echocardiography carts. Patients’ initial echocardiograms were identified and analyzed offline by a single blinded reviewer (KG) using an Xcelera workstation (Philips, Andover, MA). Estimated right ventricular pressure was assessed by tricuspid regurgitation gradient.⁹ Left ventricular end systolic eccentricity index was calculated to evaluate for septal flattening.¹⁰ Patent ductus arteriosus size was measured as the patent ductus arteriosus to left pulmonary artery diameter ratio. Direction of shunting of the patent ductus arteriosus was recorded.¹¹

Previously described estimates of elevated pulmonary vascular resistance and elevated filling pressures were assessed by several parameters. Pulmonary artery acceleration time,¹² right ventricle ejection time, and their ratio (pulmonary artery acceleration time/right ventricle ejection time) were used to assess for evidence of elevated pulmonary vascular resistance.¹² To assess for evidence of elevated filling pressures, right atrial volume was measured using the area length method¹³ and indexed to body surface area and presence of a patent foramen ovale and direction of shunting were recorded.¹⁴

Right ventricular function was assessed using percent fractional area change,^{15, 16} two-dimensional tricuspid annular plane systolic excursion indexed to body surface area,¹⁷ systolic and diastolic time ratio measured via the tricuspid regurgitant jet Doppler pattern,¹⁸ spectral Doppler tricuspid inflow E, A waves and their ration (E:A), tissue Doppler of the right ventricular free wall (e’, a’, and s wave peak velocities),¹⁹ and two-dimensional speckle-tracking derived longitudinal strain of the right ventricle from the apical four chamber view.²⁰

Statistical Analysis

Comparisons between patients with and without in-hospital extracorporeal membrane oxygenation or death were performed using Fisher’s Exact test or Mann-Whitney U test for categorical and continuous variables as appropriate. Independent variables missing > 10% data were excluded from analysis. Univariable and stepwise multivariable logistic and linear regression were used to determine associations between patient clinical characteristics and the primary and secondary outcomes. Variables were initially included in the multivariable regressions if they displayed a p value < .2 on univariable analysis. Values were kept in the final multivariable model if p < .05 or if p < .10 and the effect size upon removal was greater than a 3% change in variance. Tricuspid regurgitation velocity was not analyzed due to 35% missing data. All other variables had less than 5% missing data that was likely not random; therefore, missing data was not imputed. Intraobserver (KG) and interobserver (SC) variability was assessed on a subset of 10% of echocardiograms and was quantified using intraclass correlation coefficients measuring absolute agreement. A p value < .05 was considered statistically significant. Analyses were performed using SPSS v.25 (IBM, Armonk, NY).

Results

A total of 86 total patients were assessed, 39 male and 47 female. The primary outcome of in-hospital extracorporeal membrane oxygenation or death was met in 15 patients (17%). Of note, only one patient in our analysis group went on in-hospital extracorporeal membrane oxygenation, and this patient went on to suffer in-hospital mortality. Unsurprisingly, use of a β-mimetic (odds ratio [OR] 16.7, p < .01), hydrocortisone (OR 11.3, p < .01), and inhaled nitric oxide (OR 9.2, p < .01) were all associated with the primary outcome.

Differences between patients who did and did not meet the primary outcome are reported in Table 1. Patients who reached the primary outcome had a lower median birth weight (635 g [interquartile range (IQR) 544, 902] versus 900 g [IQR 670, 1850], p = .01). There was no difference in sex, gestational age, race, mode of delivery, and where born between the infants who lived and those who went on to meet the primary endpoint. Differences in echocardiographic estimates of pulmonary vascular resistance, filling pressure, right ventricular function, and patent ductus arteriosus size are reported in Table 2. Intraobserver variability intraclass correlation coefficient was > 0.88 and interobserver variability intraclass correlation coefficient was > 0.83 for all echocardiographic measurements suggesting strong reproducibility of the measurements.

Table 1.

Demographic Differences Between Patient Who Did and Did Not Meet the Primary Outcome.

Demographics	No ECMO/Death (n = 71)	ECMO/Death (n = 15)	p-Value
Female sex, n (%)	39 (55%)	8 (53%)	.91
Gestational age (weeks)	27.4 (25.3, 32.1)	26.4 (24.3, 28.7)	.17
Birth weight (grams)	900 (670, 1850)	635 (544, 902)	.01
BSA at echocardiogram (m²)	0.09 (0.07, 0.15)	0.07 (0.06, 0.09)	.02
Race, white, n (%)	34 (48%)	7 (47%)	.93
Vaginal delivery, n (%)	19 (28%)	4 (27%)	.92
Born in house, n (%)	66 (93%)	13 (87%)	.60
Age at time of echo (days)	2 (1,7)	2 (1,6)	.97
SBP (mmHg)	58 (47, 65)	56 (50, 67)	.79
DBP (mmHg)	30 (24, 40)	30 (26, 38)	.85
1-minute APGAR	4.0 (1.3, 5.8)	3.5 (2.0, 6.0)	.76
5-minute APGAR	6.0 (4.0, 8.0)	6.0 (5.0, 8.0)	.67
SGA, n (%)	22 (31%)	7 (47%)	.24
Hemoglobin (g/dL)	15.4 (10.7, 18.8)	14.8 (9.9, 18.1)	.55
Cord arterial pH	7.26 (7.10, 7.38)	7.23 (7.05, 7.30)	.31
Positive pressure respiratory support at echo, n (%)			.18
IPPV	55 (77%)	15 (100%)
NIPPV	13 (18%)	0 (0%)
None	3 (4%)	0 (0%)
Supplemental oxygen FiO2 > 21% at echo, n (%)	61 (86%)	14 (93%)	.68
NEC, n (%)	8 (11%)	2 (13%)	1.00

Source: The data obtained from patient echocardiograms and medical records.

Abbreviation: BSA, body surface area; DBP, diastolic blood pressure; IPPV, invasive positive pressure ventilation; NEC, necrotizing enterocolitis; NIPPV, non-invasive positive pressure ventilation; SBP, systolic blood pressure.

Note: Data reported as counts (percentages) or median (interquartile range).

Table 2.

Echocardiographic Differences Between Patient Who Did and Did Not Meet the Primary Outcome.

Echocardiographic Parameters	No ECMO/Death (n = 71)	ECMO/Death (n = 15)	p-Value
RV FAC (%)	25 (19, 32)	23 (18, 37)	.72
TAPSE (cm/m²)	2.7 (1.6, 4.3)	3.2 (1.6, 4.6)	.49
RV S:D ratio	1.1 (0.6, 1.6)	0.8 (0.1, 1.8)	.63
RV longitudinal strain (%)	–12.6 (–16.3, –9.4)	–10.3 (–15.5, –8.3)	.26
RA volume/BSA (cm³/m²)	11.0 (7.0, 15.5)	7.7 (5.9, 10.4)	.10
Patent foramen ovale, n (%)	6 (94%)	14 (100%)	.60
Atrial shunt left to right, n (%)	41 (62%)	4 (29%)	.04
LV eccentricity index	1.4 (1.2, 1.5)	1.3 (1.3, 1.5)	.70
PAAT/RVET	0.23 (0.18, 0.30)	0.23 (0.13, 0.30)	.82
PDA flow left to right, n (%)	40 (56%)	5 (33%)	.16
PDA size (PDA:LPA)	0.90 (0.67, 1.10)	1.00 0.83, 1.30)	.05

Source: The data obtained from patient echocardiograms and medical records.

Abbreviation: FAC, fractional area change; LV, left ventricular; PAAT/RVET, pulmonary artery acceleration time/right ventricular ejection time; RA, right atrial; RV, right ventricular; S:D ratio, systolic to diastolic ratio; TAPSE, tricuspid annular plane systolic excursion.

Note: Data reported as counts (percentages) or median (interquartile range).

Upon univariable logistic regression, patent ductus arteriosus to left pulmonary artery ratio (OR 6.4 [95% confidence interval (CI) 1.4-29.5], p = .02) and patent foramen ovale flow that was bidirectional or right to left (OR 4.1 [95% CI 1.2-14.5], p = .03) were associated with the primary outcome. Patent ductus arteriosus flow direction, birth weight, gestational age, and right atrial volume all displayed p < .20 and were included in the initial multivariable regression. After stepwise multivariable logistic regression analysis (Pearson chi-square = 86.5, p = .19, Nagelkerke R² = 0.29), increased patent ductus arteriosus to left pulmonary artery ratio (OR 6.9 [95% CI 1.2-40.3], p = .03), patent foramen ovale flow that was bidirectional or right to left (OR 3.9 [95% CI 1.02-15.2], p = .047), and right atrial volume (OR 0.86 [95% CI 0.74-0.99], p = .04) were independently associated with death.

Median hospital length of stay was 85 days (interquartile range 27, 127). Upon univariable linear regression, length of stay was associated with lower gestational age (p < .01), lower birth weight (p < .01), lower right atrial volume (p < .01), lower tricuspid annular plane systolic excursion (p < .01), non-white race (p = .01), patent ductus arteriosus directionality that was not left to right (p = .02), and lower right ventricular percent fractional area change (p = .02). Presence of a patent foramen ovale, right ventricular longitudinal strain, and place of birth all displayed p < .20 and were included in the initial multivariable regression. Upon stepwise multivariable regression (R² = 0.18), only gestational age was independently associated with the length of stay (β = –0.42, p < .01).

Median number of days ventilated was 24 days (IQR 4, 55). Upon univariable linear regression, number of ventilator days was associated with lower gestational age (p < .01), lower birth weight (p < .01), mode of delivery other than vaginal (p < .01), non-white race (p < .01), lower tricuspid annular plane systolic excursion (p = .03), presence of a patent foramen ovale (p = .02), and lower right ventricular percent fractional area change (p = .03) were associated with number of days ventilated. Longitudinal strain, right atrial volume, and place of birth all displayed p < .20 and were included in the initial multivariable regression. Upon stepwise multivariable regression (R² = 0.19), race other than white (p < .01), presence of a patent foramen ovale (β = 2.1, p = .04), and tricuspid annular plane systolic excursion (p = .04) were independently associated with days ventilated.

Discussion

Identifying pre-term neonates with patent ductus arteriosus and pulmonary hypertension who are at high risk of in-hospital mortality may lead to changes in their management and, ultimately, their outcomes. This study investigated potential echocardiographic parameters that may be used to identify such patients. The results of our study indicate that increased patent ductus arteriosus size, bidirectional or right to left flow at the patent foramen ovale, and lower right atrial volume on preterm neonates’ initial echocardiogram are independently associated with their in-hospital mortality.

The presence of a patent ductus arteriosus in preterm neonates is known to increase patients’ hospital morbidity and mortality.²¹ This study shows that, at the initial echocardiogram, the size of the patent ductus arteriosus in preterm neonates with clinical suspicion of pulmonary hypertension is predictive of outcomes with bigger defects being associated with poor outcomes. Our findings are consistent with those by Sellmer et al, who found that larger patent ductus arteriosus size on day of life 3 was associated with increased morbidity and mortality in the preterm neonates born prior to 28 weeks gestation.²²

Atrial shunting through a patent foramen ovale is commonly seen in preterm neonates. Choi et al found an association between left to right atrial shunting in preterm infants with bronchopulmonary dysplasia and increased mortality.²¹ In contrast, our study found that bidirectional or right to left atrial shunting was associated with in-hospital mortality. We posit that shunting in these directions is a measure of elevated atrial pressure and right ventricular diastolic heart failure, potentially related to elevated pulmonary artery pressures, which predisposes these preterm neonates to morbidity and mortality.

Increased right atrial area has been shown to be associated with clinical worsening in adults with idiopathic pulmonary hypertension.²³ Surprisingly, our findings suggest a negative correlation, with smaller indexed right atrial volumes portending poor primary and secondary outcomes. The reasons for this may be multiple. We have shown an association with bidirectional or right to left flow at the atrial level with increased in-hospital mortality, and it follows that these same neonates could have decreased right atrial volumes in the setting of volume unloading due to right to left atrial shunting.²⁴ Additionally, patients with more severe disease often display low intravascular volume in the early days following pre-term delivery, contributing to decreased right atrial volumes. Lastly, echocardiograms in the first few days of life will not yet reflect chronic signs of elevated right ventricular filling pressures that are associated with atrial dilation as seen in previous studies.

We found no independent associations between measures of right ventricular systolic function or estimates of right ventricular pressure/pulmonary vascular resistance with the primary outcome. This may be related to the early acquisitions of echocardiograms in these patients. Dynamic changes in transitional physiology, medication changes, and fluid balance may confound the association between these measures and the primary outcome early in their course. In addition, in patients with unrestrictive patent ductus arteriosus, such as those in our study, the right ventricular pressure is equal to the systemic pressure, so echocardiographic estimates of right ventricular pressure have little utility in this cohort as the right ventricular pressure is known. It appears evidence of right ventricular diastolic heart failure (bidirectional or right to left flow at the patent foramen ovale and associated lower right atrial areas) are more predictive of outcomes on early echocardiograms than convention measures of pulmonary vascular resistance or right ventricular systolic function.

Study Limitations

This is a single center, retrospective study and is therefore subject to institutional biases regarding usage of echocardiography and management of pulmonary hypertension. One variable, tricuspid regurgitation gradient, displayed a large amount of missing data due to its inability to be measured on each echocardiogram, which limited our ability to fully assess its associations with outcomes. In addition, other measures that may influence atrial filling, such as markers of ventricular diastolic function could not be assessed due to missing data. Our cohort was relatively small and should be validated in larger samples. The groups were not optimally matched for birth weight, though we attempted to control for this in our multivariable analysis. While data is available regarding medication administration globally over the whole hospitalization, medication administered during the time of the echocardiogram was not available and may have influenced our results. Our findings are limited to early echocardiographic findings in the first few days of life and cannot be translated to echocardiograms performed later in life.

Conclusion

On the initial echocardiogram of preterm neonates with patent ductus arteriosus and clinical suspicion of pulmonary hypertension, larger patent ductus arteriosus size in relation to the left pulmonary artery, bidirectional or right to left flow at the patent foramen ovale, and lower right atrial volume are all independently associated with in-hospital mortality. These echocardiographic parameters can be used to identify high risk patients. Future studies should investigate if altering treatment plans based on the measures may result in improved outcomes in this fragile population.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical Approval

The study was approved by the Medical University of South Carolina Institutional Review Board (Pro00075177).

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research received no specific grant from any funding agency, commercial or not-for-profit sectors. Dr. Chowdhury is supported by NIH HL133447.

Informed Consent

Not applicable

ORCID iD

Shahryar M. Chowdhury

References

Ferre

, Callaghan

, Olson

, Sharma

, Barfield

Effects of maternal age and age-specific preterm birth rates on overall preterm birth rates—Unites States, 2007 and 2014. MMWR Morb Mortal Wkly Rep . 2016;65:1181–1184.

Kumar

, Hutchinson

, Lakshminrusimha

, Morin

3rd , Wynn

, Ryan

RM.

Characteristics of pulmonary hypertension in preterm neonates. J Perinatol . 2007;27:214–219.

Levy

, Patel

, Groh

, . Pulmonary artery acceleration time provides a reliable estimate of invasive pulmonary hemodynamics in children. J Am Soc Echocardiogr . 2016;29: 1056–1065.

Surinder

, Shahidi

, Gin

, Swiston

Diagnostic accuracy of echocardiography for pulmonary hypertension: a systematic review and meta-analysis. Heart . 2001;97:612–622.

Mourani

, Sontag

, Younoszai

, Ivy

, Abman

SH.

Clinical utility of echocardiography for the diagnosis and management of pulmonary vascular disease in young children with chronic lung disease. Pediatrics . 2008;121(2):317–325.

, Peng

, Tsai

CH.

Persistent pulmonary hypertension of the newborn: echocardiographic assessment. Acta Paediatr Taiwan . 2001;42(4):218–223.

Jain

, McNamara

PJ.

Persistent pulmonary hypertension of the newborn: Advances in diagnosis and treatment. Semin Fetal Neonatal Med . 2015;20(4):262–271.

Malowiz

, Forasha

, Smith

, Cotton

, Barker

, Tatum

GH.

Right ventricular echocardiographic indices predict poor outcomes in infants with persistent pulmonary hypertension of the newborn. Eur Heart J Cardiovasc imaging . 2015; 16:1224–1231.

De Scordilli

, Pinamonti

, Albani

, . Reliability of noninvasive hemodynamic assessment with Doppler echocardiography: comparison with the invasive evaluation. Journal of Cardiovascular Medicine . 2019;20(10):682–690.

10.

Abraham

, Weismann

CG.

Left ventricular end-systolic eccentricity index for assessment of pulmonary hypertension in infants. Echocardiography . 2016;33(6):910–915.

11.

Malowitz

, Forsha

, Smith

, . Right ventricular echocardiographic indices predict poor outcomes in infants with persistent pulmonary hypertension of the newborn. Eur Heart J Cardiovasc Imaging . 2015;16:1224–1231.

12.

Levy

, Patel

, Groh

, . Pulmonary artery acceleration time provides a reliable estimate of invasive pulmonary hemodynamics in children. J Am Soc Echocardiogr . 2017; 29(11):1056–1065.

13.

Lopez

, Colan

, Frommelt

, . Recommendations for quantification methods during the performance of a pediatric echocardiogram: a report from the pediatric measurements writing group of the American Society of Echocardiography Pediatric and Congenital Heart Disease Council. J Am Soc Echocardiogr . 2010;23(5):465–495.

14.

Giesinger

, More

, Odame

, . Controversies in the identification and management of acute pulmonary hypertension in preterm neonates. Pediatric Reseach . 2017;82(6): 901–914.

15.

Levy

, Dioneda

, Holland

, . Right ventricular function in preterm and term neonates: reference values for right ventricle areas and fractional area of change. J Am Soc Echocardiogr . 2015;28(5):559–569. doi:10.1016/j.echo.2015.01.024

16.

Hoette

, Crueze

, Gunther

, . RV fractional area change and TAPSE as predictors of severe right ventricular dysfunction in pulmonary hypertension: A CMR study. Lung . 2018;196(2): 157–164.

17.

Breatnach

, Levy

, James

, Franklin

, El-Khuffash

Novel echocardiography methods in the functional assessment of the newborn heart. Neonatology . 2016;110:248–260.

18.

Aggarwal

, Natarajan

Echocardiographic correlates of persistent pulmonary hypertension of the newborn. Early Hum Dev . 2015;91(4):285–289.

19.

Jain

, Mohamed

, El-Khuffash

, . A Comprehensive echocardiographic protocol for assessing neonatal right ventricular dimensions and function in the transitional period: Normative data and Z scores. J Am Soc Echocardiogr . 2017; 27:1293–1304.

20.

Levy

, El-Khuffash

, Patel

, . Maturational patterns of systolic ventricular deformation mechanics by two-dimensional speckle-tracking echocardiography in preterm infants over the first year of age. J Am Soc Echocardiogr . 2017;30(7): 685–698.e1. doi:10.1016/j.echo.2017.03.003

21.

Choi

, . The impact of atrial left-to-right shunt on pulmonary hypertension in preterm infants with moderate or severe bronchopulmonary dysplasia. Pediatr Neonatol . 2015;56(3): 317–323.

22.

Sellmer

, . Morbidity and mortality in preterm neonates with patent ductus arteriosus on day 3. Arch Dis Child Fetal Neonatal Ed . 2013;98(6):F505–F510.

23.

D’Alto

, D’Andrea

, Di Salvo

, . Right atrial function and prognosis in idiopathic pulmonary arterial hypertension. Int J Cardiol . 2017;248:320–325.

24.

Giesinger

, McNamara

PJ.

Hemodynamic instability in the critically ill neonate: An approach to cardiovascular support based on disease pathophysiology. Semin Perinatol . 2016 Apr; 40(3):174–88. doi: 10.1053/j.semperi.2015.12.005. Epub 2016 Jan 14. PMID: 26778235.