The Effectiveness of Pulmonary Hypertension Screening in Infants Born Preterm

Abstract

Objective:

To determine if pulmonary hypertension (PH) screening in at-risk infants born preterm reduces morbidity and/or NICU length of stay.

Study design:

This single-center retrospective cohort chart review compared infants born < 32 weeks gestational age (GA) before and after the implementation of an updated PH screening guideline. Screening eligibility and PH diagnosis were determined by applying the standardized criteria to patients in both epochs. NICU and post-discharge outcomes were determined by chart review.

Results:

Pre- (N=513) and post-screening (N=544) epochs had similar gestational age and demographic characteristics. More echocardiograms were obtained in post-screening infants resulting in more PH diagnoses (11.5% vs. 16.5%, p=0.02) at a younger median post-natal age [day of life 73 (28–193) vs. 55 (28–212), median and range, p=0.01]. PH+ infants in the post-screen epoch were discharged at a younger median post-natal age [127 (49–407) vs 113.5 (46–433) days, p=0.02] and corrected GA [43.6 (36.7–87.4) vs 41.7 (36.6–64.9) weeks, p=0.03].

Conclusion:

PH screening protocol with multidisciplinary team involvement may be associated with increased detection of PH in at risk infants yet a shorter NICU stay in infants with PH.

Introduction

Bronchopulmonary dysplasia (BPD) affects up to 68% of infants in the neonatal intensive care unit (NICU) born < 29 weeks’ gestational age (GA) and increases the risk of pulmonary hypertension (PH) with a resultant increased morbidity and mortality^1–4. Neonatal exposures (e.g. supplemental oxygen, mechanical ventilation) disrupt development of lung parenchyma and microvasculature, resulting in abnormal vascular tone, vessel remodeling, and increased vascular reactivity manifesting as BPD associated PH (BPD-PH)^1,5. There is up to a 20% prevalence of PH in infants with BPD and 12% in infants born preterm with and without BPD⁶. PH development has been associated with risk factors such as fetal growth restriction, postnatal growth failure, and necrotizing enterocolitis (NEC)^5,7,8. PH is associated with increased length of stay (LOS), tracheostomy, rehospitalization, and mortality (up to 5-fold) compared to BPD alone^2,5,6,9,10. Medications such as phosphodiesterase-5 inhibitors (e.g. sildenafil) are under investigation for prevention and treatment of PH, making early detection a matter of interest^11,12. It is possible, therefore, that detecting and addressing PH earlier in the hospital course could reduce the LOS.

Guidelines for PH screening in infants born preterm are limited⁵. In 2015, the American Thoracic Society (ATS) recommended screening infants with BPD and significant respiratory support at 36 weeks’ corrected GA (CGA) with echocardiogram (ECHO) but acknowledged the limitations of ECHO^13–15. The Pediatric Pulmonary Hypertension Network (PPHNet) expanded the guidelines to include clinical screening time points and multi-disciplinary team management¹⁶. In 2019, the University of Rochester Medical Center (URMC) PH team (a neonatologist with BPD and PH expertise, a pediatric interventional cardiologist with PH expertise, pediatric pulmonologists, and an adult PH specialist) established updated PH screening guidelines (Figure 1) to reflect current recommendations¹⁶ and emphasized BPD-independent risk factors^6,17,18. Prior to the implementation of the 2019 guidelines there was no standard protocol within the URMC NICU for PH screening. As such, investigation for the disease was typically reactionary in the setting of slow clinical improvement or acute decompensation. We report a comparison of pre- and post-screening guideline outcomes, hypothesizing that establishing screening criteria would provide an early diagnosis and treatment of PH, resulting in improved PH outcomes shown by a primary outcome of shortened NICU LOS in patients with PH.

Figure 1:

Guideline for Screening for Pulmonary Hypertension in Preterm Infants at URMC.

Echocardiogram parameters --

RVSP (right ventricular systolic pressure) estimates:

No patent ductus arteriosus (PDA)/ventricular septal defect (VSD): RV Systolic pressure = TR peak gradient

PDA present: PA systolic pressure (estimates RV pressure) = Systolic blood pressure minus PDA gradient

VSD present: RV systolic pressure = Systolic blood pressure minus VSD Gradient

Moderate to severe PH – RVSP ≥ ½ systemic but < systemic; tricuspid regurgitation gradient predicts RVSP ≥ ½ systemic but < systemic, elevated pulmonary regurgitation peak and/or end-diastolic gradient (≥16 mmHg), septal flattening in systole, RV (right ventricular) function normal;

Severe PH− RVSP near or ≥ systemic pressure; if present, PDA or VSD with predominant right to left shunting, septal flattening throughout the cardiac cycle and/or bowing into the left ventricle, any RV dysfunction attributed to elevated pulmonary artery pressure].

*based on systolic blood pressure at time of echocardiogram

Methods

We conducted a retrospective cohort study of infants born < 32 weeks’ GA who survived to ≥ 28 days and were admitted to the URMC Level IV NICU within 72 hours of birth from January 1, 2016 to December 31, 2022. Infants were divided into two epochs, either admitted pre-PH screening guideline from January 1, 2016 – April 30, 2019 or post-PH screening guideline from August 1, 2019 – December 31, 2022. Infants with major co-morbidities independent of prematurity (i.e. congenital diaphragmatic hernia, tracheoesophageal fistula, cleft lip/palate, intestinal atresia, chromosomal anomalies, and complex congenital heart disease) were excluded.

Ethics and Data Storage

This study was approved by the URMC institutional review board (IRB) (#00007499). To improve post-NICU data collection, study approval was also provided by the Rochester Regional Health (RRH) medical system IRB (#3008 B). Infants are frequently transferred from the NICU at URMC to convalesce at the RRH Level II Special Care Nursery. The URMC and RRH medical systems are nearly exclusive providers of pediatric hospitalizations and subspeciality care in the midwestern- New York region. The URMC Clinical and Translational Science Institute (CTSI) provided Secure Research Data Analytics (SERDA) for patient data storage.

Data Collection: Initial Hospitalization and Re-admission

NICU data, (demographics, perinatal, admission data, neonatal complications, respiratory support, medications received, LOS, and death), were extracted by query from a local database (NeoData, Isoprime Corporation, Lisle, IL, USA) and manually from the electronic medical record (EMR) (Tables 1 and 3). The URMC and RRH EMR were manually reviewed for infants that were transferred to another unit during their initial hospitalization to complete ascertainment of total LOS, death, and discharge outcomes. URMC is a participating center in sildenafil trials currently being conducted under the Best Pharmaceuticals in Children’s Act by the National Institute of Child Health and Human Development (NICHD)-funded Pediatric Trials Network (PTN), to investigate drug safety and utility to prevent PH in infants with BPD. No clinical trials at URMC investigate sildenafil efficacy for PH treatment. Therefore, infants with “sildenafil/study drug” orders were recorded as not receiving sildenafil as any study drug received was not for identified PH. The effect of recording all study subjects as receiving sildenafil was also considered. Rehospitalization data were obtained via the URMC CTSI facilitated query of the EMR. Unplanned rehospitalizations up to 6 months after discharge were recorded. Rehospitalizations for planned procedures (gastrostomy tube, ROP procedures, etc.) were removed. Infants that did not have re-admission at URMC were counted as having not been readmitted as URMC is the major admitting children’s hospital in the area for children with complex medical needs.

Table 1:

Maternal and neonatal characteristics of all infants, Risk +, and PH+ infants

Characteristics	All Infants Pre-Screen (n=513)	All Infants Post-Screen (n=544)	p-value	Risk+ Pre-Screen (n=270)	Risk+ Post-Screen (n=273)	p-value	PH+ Pre-Screen (n=59)	PH+ Post-Screen (n=90)	p-value
Maternal
IUGR	13 (2.5%)	20 (3.7%)	0.30	9 (3.33%)	14 (5.13%)	0.39	1 (1.7%)	5 (5.6%)	0.40
Pre-eclampsia	46 (9.0%)	50 (9.2%)	0.92	20 (7.41%)	24 (8.79%)	0.64	2 (3.4%)	7 (7.8%)	0.32
Clinical Chorioamnionitis	139 (27.1%)	87 (16.0%)	<0.001**	84 (31.11%)	48 (17.58%)	<0.001**	17 (28.8%)	19 (21.1%)	0.33
Prenatal steroids (any)	491 (100%) n=491	496 (95.4%) n=520	<0.001**	256 (100%)	251 (96.54%)	0.004**	52 (100%)	79 (95.2%)	0.30
Oligohydramnios*	0	0		0	0		0	0
Neonatal
GA, wk*	28.6 (2.4)	28.5 (2.5)	0.97	27.4 (2.5)	27.1 (2.5)	0.22	26.0 (2.4)	26.1 (2.2)	0.84
Birthweight, kg*	1.16 (0.39)	1.14 (0.40)	0.46	0.981 (0.35)	0.931 (0.35)	0.10	0.84 (0.35)	0.79 (0.24)	0.35
Sex (Female)	245 (47.8%)	273 (50.2%)	0.43	133 (49.3%)	136 (49.8%)	0.93	21 (35.6%)	43 (47.8%)
Mechanical ventilation days	2 (0–260)	1 (0–433)	0.01**	9.5 (0–260)	11 (0–433)	0.61	51 (0–260)	31 (0–433)	<0.001**
NRN 2019 BPD (mod-severe)*	67 (13.1%) n=513	70 (12.9%) n=542	0.93	67 (24.81%)	70 (25.8%)	0.84	29 (49.2%)	31 (34.4%)	0.09
IVH (≥ Grade 3)	21 (4.1%)	23 (4.2%)	1.00	17 (6.3%)	17 (6.23%)	1.0	7 (11.9%)	9 (10.0%)	0.79
Airway obstruction	2 (0.4%)	0	0.24	2 (0.74%)	0	0.25	1 (1.69%)	0	0.40
GERD (requiring medication)	46 (9.0%)	60 (11.0%)	0.31	30 (11.1%)	44 (16.1%)	0.10	12 (20.3%)	16 (17.8%)	0.83
ROP (any stage)	142 (27.7%)	192 (35.3%)	0.01**	110 (40.7%)	152 (55.7%)	<0.001**	36 (61.0%)	58 (64.4%)	0.73
PDA on problem list	81 (15.8%)	179 (32.9%)	<0.001**	65 (24.1%)	143 (52.4%)	<0.001**	29 (49.2%)	64 (71.1%)	0.01**
PDA Ligation	24 (4.7%)	25 (4.6%)	1	22 (8.2%)	25 (9.2%)	0.76	12 (20.3%)	11 (12.2%)	0.25
Sepsis type			0.51			0.65			0.22
Late onset	39 (7.6%)	52 (9.6%)		35 (13.0%)	42 (15.4%)		16 (27.1%)	16 (17.8%)
Early onset	4 (0.8%)	5 (0.9%)		2 (0.74%)	3 (1.1%)		0	2 (2.2%)
SGA*	50 (9.8%)	71 (13.1%)	0.10	35 (13.0%)	54 (19.8%)	0.04**	7 (11.9%)	16 (17.8%)	0.36
Post Natal Growth Failure**	39 (7.6%)	32 (5.9%)	0.27	36 (13.3%)	29 (10.6%)	0.36	16 (27.1%)	9 (10.0%)	0.01**
ASD/PFO*	48 (9.4%)	160 (29.4%)	<0.001**	40 (14.8%)	125 (45.8%)	<0.001**	13 (22.0%)	53 (58.9%)	<0.001**
NEC ≥ 2A*	37 (7.2%)	25 (4.6%)	0.09	26 (9.6%)	21 (7.7%)	0.45	5 (8.5%)	4 (4.4%)	0.48
> 1 Steroid course*	10 (2.0%)	3 (0.6%)	0.05	10 (3.7%)	2 (0.73%)	0.02**	7 (11.9%)	1 (1.1%)	0.01**
Culture positive sepsis*	43 (8.4%)	55 (10.1%)	0.34	37 (13.7%)	49 (18.0%)	0.12	16 (27.1%)	19 (21.1%)	0.43
NICU Medications
Dexamethasone	66 (12.9%)	103 (18.9%)	0.01	57 (21.1%)	96 (35.2%)	<0.001**	28 (47.5%)	46 (51.1%)	0.74
Prednisolone	0	20 (3.7%)	<0.001**	0	20 (7.3%)	<0.001**	0	10 (11.1%)	0.01**
Inhaled steroid	150 (29.2%)	177 (32.5%)	0.26	129 (47.8%)	158 (57.9%)	0.02**	46 (78.0%)	70 (77.8%)	1
Any PH med	18 (3.5%)	21 (3.1%)	0.73	16 (5.9%)	19 (5.9%)	1	13 (22.0)	14 (12.2%)	0.11
Sildenafil	4 (0.8%)	9 (1.7%)	0.39	4 (1.5%)	9 (3.3%)	0.26	4 (6.8%)	7 (7.8%)	1
Furosemide	332 (64.7%)	294 (54%)	<0.001**	225 (83.3%)	218 (79.9%)	0.32	59 (100%)	83 (92.2%)	0.04**
Thiazide	195 (38.0%)	135 (24.8%)	<0.001**	151 (55.9%)	122 (44.7%)	0.010**	50 (84.8%)	60 (66.7%)	0.02**
Any Diuretic	332 (64.7%)	294 (54.0%)	<0.001**	225 (83.3%)	218 (79.9%)	0.32	59 (100%)	83 (92.2%)	0.04**

Data presented as n(%)

^**p < 0.05

^*PH risk factors included in URMC PH Screening Guideline (GA < 26 weeks, BW < 750 grams)

PH = pulmonary hypertension, IUGR = intrauterine growth restriction, GA = gestational age,, NRN = neonatal research network, BPD = bronchopulmonary dysplasia, IVH = intraventricular hemorrhage, GERD (requiring medication) = gastroesophageal reflux disease (requiring proton pump inhibitor or histamine -2 blocker at discharge), ROP = retinopathy of prematurity, PDA = patent ductus arteriosus, SGA = small for gestational age, ASD = atrial septal defect, PFO = patent foramen ovale, NEC = necrotizing enterocolitis

**problem list query: growth failure, failure to thrive, poor weight gain, weight loss

Table 3:

Discharge outcomes of all infants, Risk +, and PH positive infants that survived to discharge.

Discharge Outcomes	All Infants Pre-Screen (n=501)	All Infants Post-Screen (n=535)	p-value	Risk+ Pre-Screen (n=258)	Risk+ Post-Screen (n=264)	p-value	PH+ Pre-Screen (n=54)	PH+ Post-Screen (n=86)	p-value
Discharge CGA, wks	38.4 (34.0–87.4)	38.7 (34.4–64.9)	0.45	40.4 (34.9–87.4)	40.7 (35.3–64.9)	0.23	43.6 (36.7–87.4)	41.7 (36.6–64.9)	0.02**
Length of stay, days	65.0 (28–407)	67.0 (20–433)	0.53	90.5 (35–407)	97.0 (28–433)	0.07	127 (49–407)	113.5 (46–433)	0.03**
Growth failure at discharge or death (weight Z ≤ −1.28)	147/513 (28.7%)	151/544 (27.8%)	0.75	93/270 (34.4%)	86/273 (31.5%)	0.47	22/59 (37.3%)	24/90 (26.7%)	0.17
NICU Death	12/513 (2.3%)	9/544 (1.7%)	0.51	12/270 (4.4%)	9/273 (3.3%)	0.51	5/59 (8.5%)	4/90 (4.4%)	0.48
Medications
Sildenafil	3 (0.6%)	9 (1.7%)	0.15	3 (1.2%)	9 (3.4%)	0.14	3 (5.6%)	7 (8.1%)	0.74
Inhaled steroids	21 (4.2%)	37 (6.9%)	0.06	21 (8.1%)	36 (13.6%)	0.05**	14 (25.9%)	22 (25.6%)	1
Furosemide	12 (2.4%)	17 (3.2%)	0.46	11 (4.3%)	15 (5.7%)	0.55	7 (13.0%)	9 (10.5%)	0.79
Thiazide	113 (22.6%)	75 (14.0%)	<0.001**	94 (36.4%)	67 (25.4%)	0.008**	38 (70.4%)	36 (41.9%)	0.002**
Spironolactone	108 (21.6%)	75 (14.0%)	0.002**	90 (34.9%)	67 (25.4%)	0.02**	36 (66.7%)	35 (40.7%)	0.003**
Reflux medications	93 (18.6%)	91 (17.0%)	0.52	69 (26.7%)	68 (25.8%)	0.84	21 (38.9%)	29 (33.7%)	0.59
Respiratory support			0.15			0.15			0.02**
Room air	470 (93.8%)	504 (94.2%)		227 (88.0%)	233 (88.3%)		32 (59.3%)	67 (77.9%)
Low flow O2	22 (4.4%)	28 (5.2%)		22 (8.5%)	28 (10.6%)		14 (25.9%)	16 (18.6%)
Tracheostomy	9 (1.8%)	3 (0.56%)		9 (3.5%)	3 (1.1%)		8 (14.8%)	3 (3.5%)
Feeding			<0.001**			0.004**			0.95
All PO	396 (79.0%)	380 (71.0%)		169 (65.5%)	145 (54.9%)		21 (38.9%)	33 (38.4%)
PO + Tube	85 (17.0%)	105 (19.6%)		69 (26.7%)	75 (28.4%)		20 (37.0%)	34 (39.5%)
Tube only	20 (4.0%)	47 (9.4%)		20 (7.6%)	44 (16.7%)		13 (24.1%)	19 (22.1%)
Readmission within 6 months	92 (18.4%)	68 (12.7%)	0.01**	56 (21.7%)	40 (15.4%)	0.06	16 (32.7%)	16 (19.5%)	0.10

Outcome statistics excludes NICU deaths. Data presented as n(%), median(range)

^**p < 0.05

PH = pulmonary hypertension, CGA = corrected gestational age, NICU = neonatal intensive care unit

ECHO Data and PH Determination

The Syngo ECHO system was queried for all patients between birth and 2 years of age for 1) PH diagnosis by report-generating cardiologist and 2) PH relevant ECHO parameters included in the URMC PH Screening Guideline [See Figure 1]. Infants were characterized as having PH if calculations based on these report measurements predicted moderate-to-severe pulmonary hypertension as outlined on the Screening Guideline (Figure 1). For outcome analysis, infants were defined as PH Positive (PH+) if they had evidence of PH on ≥ 1 ECHO at ≥ 28 days of life per URMC guideline criteria for PH. A subset (n=27) of ECHOs were overread by the PH team cardiologist for internal study validation.

Screening Criteria

The guideline (Figure 1) was retroactively applied to study epochs to determine screening eligibility at ≥ 28 postnatal days. Eligible infants were characterized as Risk Positive (Risk+). Infants were determined to have met criteria for a screening pulmonary ECHO to evaluate for PH at 28 days if they had a birthweight (BW) < 750 grams and remained mechanically ventilated or remained on non-invasive respiratory support longer than expected for CGA (Supplemental Table 1). Two different cut off criteria for prolonged length of respiratory support were made for < 28 weeks GA and ≥ 28 weeks to allow for some additional weaning time for younger GAs prior to determining eligibility. For example, infants that were born < 28 weeks qualified for prolonged CPAP if they remained on CPAP at 33 weeks GA and 32.5 weeks if born > 28 weeks (Supplemental Table 1).The length of respiratory support determined to be prolonged for the study is based on URMC experience and protocols which are based on a standard of maintaining bubble CPAP until 32 weeks’ gestation in this population^19–21. Infants also met criteria if they had ≥ 2 specific risk factors including GA < 26 weeks, birthweight < 750 grams, small for gestational age (SGA) (BW < 10%tile), oligohydramnios, post-natal growth failure (problem list query for failure to thrive, growth failure, poor weight gain), culture-proven sepsis, ≥ Grade 2A necrotizing enterocolitis (NEC), ASD/PFO, and/or > 1 postnatal steroid course. At 36 weeks’ post-menstrual age, the 2019 Neonatal Research Network (NRN 2019) BPD definition, shown to correlate with long term outcomes³ and tracked internally by the URMC NICU²², was also used to determine if infants met criteria for screening (≥ Grade 2).

Outcomes

Outcomes were analyzed for the entire cohort, Risk+ patients, and PH+ patients. The primary outcome of LOS was evaluated by CGA and day of life at discharge. Secondary outcomes evaluated at discharge were respiratory support, feeding support, BPD/PH specific medications, and death. Growth failure at discharge was defined as discharge Z score ≤ −1.28. Post-discharge outcome was unplanned hospitalizations within 6 months of discharge.

Statistical Analysis

All statistical analyses were conducted using the SAS System for Windows (version 9.4). Prior to testing hypotheses, descriptive statistics were evaluated by two sample t-test (continuous variables) or chi-square and Fisher’s exact tests (categorical variables) to compare baseline demographics, pregnancy data and clinical factors between the two cohorts as well as between subgroups of Risk+ and PH+. Variables with distributions violating assumptions of normality were compared using Wilcoxon Rank Sum and/or Mann-Whitney tests. Hypothesis-driven comparisons were made to control the family-wise type I error rate at 0.05. Total length of stay includes age at death in infants that died.

Results

Maternal and Neonatal Characteristics

1057 infants (Figure 2, Table 1) were identified who met criteria for PH screening. Maternal and neonatal characteristics for all infants, the Risk+, and PH+ infants are shown in Table 1. There were no differences in GA or BW between groups in either epoch. Prenatally, the total cohort and Risk+ infants in the post-screen guideline epoch had decreased chorioamnionitis and exposure to prenatal steroids. These were not different in the PH+ infants.

Figure 2:

Study infants pre- and post- screening guideline epochs divided into all infants, Risk +, and PH + groups

All study infants divided into groups defined as all infants, Risk + based on pulmonary hypertension (PH) screening guideline, and PH + based on echocardiographic findings before and after implementation of the screening guideline.

Risk Factors for PH

Clinical NICU factors for all infants, the Risk+, and PH+ infants are shown in Table 1. There was no inter-epoch difference in Grade 2–3 BPD in the total cohort or in Risk+ infants but there was a trend for lower grade BPD in the PH+ infants (49.2% vs 34.4%, p=0.09) in the post-screen epoch. The only difference identified between epochs in clinical risk factors for PH (Figure 1) was the presence of a PFO or ASD which was higher in the post screen epoch (9.4% vs 29.4%, <0.001). In looking at respiratory support eligibility criteria, more infants in the pre-screen epoch met eligibility criteria for PH screening based on prolonged NIPPV requirement (28.3% vs 19.5%, p < 0.001). There were no differences in eligibility between epochs based on need for prolonged HFNC, CPAP, or FiO2 (Table 2).

Table 2:

PH Screening outcomes

PH Screening	All Infants Pre-Screen (n=513)	All Infants Post-Screen (n=544)	p-value	Risk+ Pre-Screen (n=270)	Risk+ Post-Screen (n=273)	p-value	PH+ Pre-Screen (n=59)	PH+ Post-Screen (n=90)	p-value
Meets criteria based on BW <750g and mechanical ventilation at 28 days	64 (12.5%)	62 (11.4%)	0.64
Meets criteria based on risk factors	119 (23.2%)	143 (26.3%)	0.25
Meets criteria for prolonged respiratory support:
HFNC	106 (20.7%)	120 (22.1%)	0.60
CPAP	35 (6.8%)	43 (7.9%)	0.56
NIPPV	145 (28.3%)	106 (19.5%)	< 0.001**
Last FiO2 > 30	140 (27.3%)	178 (32.8%)	0.06
Grade 2–3 BPD	67 (13.1%)	70 (12.9%)	0.92
Meets criteria for screening	270 (52.6%)	273 (50.2%)	0.46				56 (94.9%)	83 (92.2%)	0.74
≥ 1 echo at day of life ≥ 28	179 (34.9%)	255 (46.9%)	<0.001**	144 (53.3%)	196 (71.8%)	<0.001**
PH+ on ≥ 1 echo ≥ day of life 28	59 (11.5%)	90 (16.5%)	0.02**	56 (20.7%)	83 (30.4%)	0.01**
DOL of 1st PH + echo ≥ DOL 28							73 (28–193)	55 (28–212)	0.01

Data presented as n(%), median (range). HFNC – High Flow Nasal Cannula ≥ 2 LPM, CPAP – Continuous Positive Airway Pressure, NIPPV – Nasal Intermittent Positive Pressure Ventilation

Outcomes from applying the PH screening algorithm to the epochs are shown in Table 2 for the groups. When all eligibility criteria were combined (Figure 1), the number of infants that were Risk + (52.6% vs 50.2%, p = 0.46, Table 2), did not differ between the epochs. the epochs.

BPD Management

In the total cohort, infants had less diuretic exposure in the post screen epoch (Table 1). Thiazide diuretic and spironolactone use (not shown) in the NICU decreased but not furosemide in Risk+ infants. There was decreased exposure to any diuretic in the PH+ infants in the post-screening guideline epoch. In the post-screen epoch, infants in the total cohort and Risk+ infants, had increased dexamethasone and prednisolone exposure (Table 1). Significantly more PH+ infants in the post PH-screening guideline epoch were also exposed to prednisolone (0% vs 11.1%, p=0.01) with no difference in dexamethasone exposure. Discharge outcomes are shown in Table 3. Infants in the total cohort were less often discharged on thiazide or spironolactone yet with a trend towards increased inhaled steroids. Risk+ infants in the post-screen guideline epoch were more frequently discharged on inhaled steroids but less frequently on thiazide and spironolactone with no difference in furosemide. The PH+ infants had decreased thiazide and spironolactone but no difference in furosemide or inhaled steroids at discharge in the post screen epoch.

PH Diagnosis and Management

More infants who met criteria for screening in the post-screen guideline epoch had ≥ 1 ECHO at or beyond day of life 28 ( 53.3% vs. 71.8%, p<0.001, Table 2). In the post-screen guideline epoch, more infants were diagnosed with PH (11.5% vs 16.5% p=0.02) and the PH+ diagnosis was made at an earlier day of life (DOL) (73 [28–193] vs 55 [28–212] days, median and range, p=0.01) (Table 2). There were no differences observed in any group in pooled PH medications (inhaled nitric oxide, epoprostenol, bosanten, sildenafil) or in sildenafil use alone while in the NICU. When randomized sildenafil study infants were considered as having received sildenafil, more infants in the total cohort and Risk + infants, but not the PH+ group, in the post-screen epoch received sildenafil (6.8% vs 12.2%, p=0.4) and a higher total number of infants in the total cohort were documented as having received sildenafil (n=9) than PH+ infants (n=7). There was no significant increase in sildenafil at discharge in the post guideline epoch in any group (Table 3). There were also no differences in respiratory needs at discharge in the total cohort or in the Risk+ infants. The respiratory needs at discharge in PH+ infants post-screening were however improved, seen as increased room air (59.3% vs 77.9%), decreased low flow oxygen (pre-screen 25.9% vs post-screen 18.6%), and decreased tracheostomy (pre-screen 14.8% vs post-screen 3.5%). There was decreased PO feeding in the post-screen epoch in all infants (79.0% vs 71.0%, p <0.001) and Risk + infants (65.5% vs 59.4%, p = 0.004.) but no differences in feeding methods at discharge were observed in the PH+ only infants between epochs.

Hospital Length of Stay, Mortality, and Readmission

Length of stay, mortality, and readmission are listed in Table 3. Despite no differences in other measures of severity of illness, there was a significant reduction in median length of stay in the post-PH screening guideline for PH+ infants in both corrected GA at discharge (43.6 (36.7–87.4) vs 41.7 (36.6–64.9) weeks, p=0.02) and in post-natal LOS (127 (49–407) vs 113.5 (46–433), p=0.03). There were no differences observed in NICU mortality. No difference was found between the epochs in readmission rates.

Discussion

Our single center study of a large sample of 1057 premature infants at risk for the development of PH revealed that a risk-based screening guideline for PH, with involvement by a multidisciplinary PH team, was associated with an increased detection of moderate-to-severe PH and reduced LOS in PH+ infants. To our knowledge, this is the first study to evaluate the impact of a PH screening protocol on clinical outcomes in infants born premature.

In 2017 Weismann at al. first reported a study for prospective screening to evaluate the prevalence of PH at 36 weeks in infants < 32 weeks’ GA with ECHO if they had BPD (NIH Criteria²³), were born < 750 grams (±BPD), or had clinical concerns for PH⁵. However, this study did not report a comparison of prevalence or outcomes prior to the initiation of the screening program making it difficult to evaluate the effect of screening. They found multiple BPD-independent clinical risk factors [NEC, early onset sepsis, etc.] that were associated with PH which were integrated into the screening guideline at URMC and likely assisted in our increased detection of PH prior to 36 weeks⁵. Additionally, studies have found that the persistence of a moderate to large PDA and a history of a PDA ligation have significant associations with the development of PH^24–26. Our results showed that there were more PH+ infants in the post-screen epoch with PDA on their problem list. Although the requirement of medical PDA treatment was beyond the scope of this study, this may indicate an association of a symptomatic PDA with PH. However, we did not find any differences in the need for PDA ligation. Although the presence of the PDA is not part of our screening algorithm, the PDA is discussed during the PH team meetings to determine if the PDA plays a flow mediated role in ECHO measurements suggestive of PH or if the PDA is hemodynamically significant and contributing to intra-pulmonary vascular dysfunction. Our guideline utilizes commonly tracked clinical characteristics which when pooled showed no difference between the groups in the number of infants at risk for PH development yet more were diagnosed and at a younger DOL suggesting the screening guideline increases diagnosis sensitivity. This increased sensitivity and earlier detection that we found in conjunction with our shorter LOS suggests that making a diagnosis gives the clinical team information to act upon earlier with respect to PH which may lead to a decreased LOS. Additional studies are needed to determine if making the early diagnosis of PH in otherwise undetected infants benefits long term outcomes.

In 2023, Yung et al. reported a multidisciplinary, systematic approach to BPD-PH in children with severe BPD, with resolution of BPD-PH in 87% of subjects prior to discharge²⁸. We did not investigate PH at time of discharge however our decreased LOS suggests an improved clinical status. Studies of the various phenotypes of PH associated with prematurity (early vs late) suggest that late PH, established after 4 weeks of age, is associated with longer need for respiratory support²⁹. Our results showing decreased mechanical ventilation days (Table 1) and respiratory support at discharge (Table 3) may suggest an improvement in this aspect of the phenotype in part from the screening and PH team. In the post-screen epoch, the PH team provides biweekly in-person reviews with the medical care team of growth, fluid balance, respiratory support, ECHO results, aspiration risk, and chest radiographs of each eligible child, especially those with chronic, worsening, or non-resolving lung disease. N-terminal pro-brain natriuretic peptide (NTproBNP - a marker for PH¹³) is tracked if disease is worsening. This provides longitudinal continuity of care as the PHTN team members attend each meeting while the medical care team rotates on service time. ECHO evidence of PH, intracardiac and PDA shunts, pulmonary venous stenosis and left heart dysfunction, are viewed and reviewed together by the team cardiologist, neonatologist, and pulmonologist facilitating a patient-centered, longitudinal assessment of functional cardiography, and promotes discussions of etiology and intervention beyond binary PH diagnosis and treatment. This longitudinal care allows for identification of infants progressing slowly in weaning respiratory support to identify contributory indolent PH as well as other causes. We feel that this early identification at least partially explains the reduced LOS identified post screening. Prior to the implementation of the screening guideline, a clinical team may have ordered an echocardiogram to evaluate for PH in babies with severe BPD, especially in the setting of difficulty weaning respiratory support, but there was no preemptive screening or standard timing. There was also no longitudinal discussion of clinical progress and standard follow up for progression of disease in infants at risk of PH or with PH prior to the initiation of the screening program. These points highlight probable value of a stable, proactive, multidisciplinary PH team.

We did not detect increased PH-directed pharmacotherapy in the PH+ infants in the post-guideline epoch. Given the low number of PH+ infants, we were likely underpowered for this outcome. Treatment of infants with BPD-PH with sildenafil may improve pulmonary arterial pressures but with an unclear impact on mortality or morbidity^30–32. Our study had a similar rate of sildenafil use to Yung at el.’s who reported 8% of their infants with PH having initiated sildenafil with their multidisciplinary team²⁸. In our NICU, sildenafil is exclusively used in consultation with the PH team. The PH team follows the ATS and PPHNet guidelines surrounding the management of BPD-PH which include aggressive treatment of underlying lung disease prior to the initiation of PH-targeted therapies unless PH is severe^13,16. This makes it difficult to differentiate the effect of screening and intervention for PH from BPD care. The increased steroid use in the post-screen epoch likely reflects BPD practice changes over time (Supplemental Table 2, Plan-Do-Study Act Cycles of URMC BPD Workgroup). Diuretic use was also decreased in the post-epoch group; likely secondary to a unit-wide initiative to reduce diuretics in infants with BPD making it difficult to draw conclusions about their use. Although diuretics are often used in BPD their value in affecting outcomes is not been shown³³. In BPD-PH, diuretics may have value in reducing documented cardiac strain and improving pulmonary vascular resistance and ventricular function. Close monitoring however is necessary due to the risk of reducing preload to a potentially hypertrophied, stiff, volume dependent right ventricle ^13,34,35. The URMC PH team tends to recommend diuretic use when there is evidence of left atrial or left ventricular enlargement, especially if complicated by a hemodynamically significant PDA or with elevations in NTproBNP. With combined efforts to be more selective and evidence driven, our study detected decreased diuretic use overall despite more targeted use in infants with PH which may have contributed to shortened LOS in these children. PPHNet also recommends supplemental oxygen to maintain oxygen saturation (SpO₂) 92–95% due to increased pulmonary artery pressure with intermittent hypoxia¹⁶. Our NICU standard of care, reinforced by the PH team, is to increase the SpO₂ targets when moderate-to-severe PH is identified (Figure 1). Although ordered SpO₂ targets were not available for study, this may have played a role in reducing progression of disease and in the decreased LOS. Although no more infants were discharged on supplemental oxygen (Table 3), it is possible that infants were able to wean off their respiratory support following improvement in their disease with the recommended increased oxygen for a period of time. As stated above, the PH team also plays a role in the treatment of a persistent hemodynamically significant PDA to help determine if it is contributing to intra-pulmonary vascular dysfunction. If lung disease is worsening or severe with over-circulation shown typically by left atrial dilation and pulmonary edema, and evidence of flow-mediated elevation in pulmonary arterial pressure, PDA closure may be recommended versus pulmonary vasodilator if PDA flow is restricted, bidirectional or low velocity, with signs of PH. It is also within scope for the PH team to also direct the care team to revisit nutrition plans for the infant’s linear growth. Given the individualized discussion of the PH team and the various strategies that may be employed, it is difficult to find one specific treatment difference to account for the reduced LOS in the PH+ infants post screening. However, given that the rates of Grade 2–3 BPD in Risk + infants and the LOS in Risk + infants between the epochs (which includes the PH+ infants) did not change (Table 1, Table 3) suggests that the LOS did not decrease for infants with BPD without PH. We feel this supports that the PH team improved the care of infants with PH, as evidenced by decreased LOS. Our study suggests that there is an advantage to being diagnosed at a lower day of life from the implementation of PH screening. Studies of each specific intervention, including randomized control trials with an increased number of PH+ infants, may be needed to identify which interventions are most impactful.

We also analyzed outcomes at and beyond discharge. In comparison to the Yung study²⁸, we showed decreased tracheostomy, supplemental oxygen, and feeding tubes at discharge which may reflect differences in PH management strategies, BPD severity (Yung at el. only included infants with severe BPD), or discharge practice differences. We did not observe an increase in infants with BPD-PH discharged on low flow oxygen therapy despite recommendations to increase SpO₂ goal when PH is present. In fact, our data showed a decrease in low flow oxygen and tracheostomy in the post-screen epoch PH+ infants. We attribute these differences to the longitudinal follow up from the multidisciplinary PH team and the work of the BPD workgroup (Supplemental Table 2)²². It is also possible that infants were able to wean off their respiratory support following improvement in their PH. We did not see a difference in tube-only feeds at discharge despite PH team recommendations for dysphagia or reflux evaluation and tube-only feeding in the setting of aspiration which can exacerbate lung disease and PH¹³. PH is a risk factor for readmission following discharge². Although there was a trend towards decreased readmission in the Risk+ infants there was no effect in PH+ infants. The PH team provides recommendations for outpatient cardiology follow up and increases involvement of the subspecialists which may strengthen relationships for follow-up for infants with or at risk for PH and may in part be an explanation for the trend in avoiding hospitalization seen in the Risk+ patients.

There are limitations of our retrospective study. All outcomes can only be viewed as associations rather than causality of improvements from the PH team. Two factors with probable opposite effects on PH risk, chorioamnionitis, shown to increase the risk of PH²⁷ and risk lowering prenatal steroids, were increased in the pre-screen epoch (all infants and Risk +, Table 1), potentially modifying baseline risk for PH between epochs and leading to mis-interpretation of screening benefits. As a single center retrospective study, readmission may be impacted by loss to follow up. Detection bias may also influence our results if less severe PH in the post-screen epoch artificially reduced LOS. The study start time of 2016 coincides with the creation of BPD reduction workgroup²² and some improvements may be attributable to their implemented changes (Supplemental Table 2). In addition, this was a retrospective review of ECHO reports rather than the images. It is well documented that diagnosis of PH by ECHO is imperfect and limited by interobserver reliability^1,13,16. The method of using echocardiographic variables to calculate presence of PH rather than expert cardiologist reading has limitations. The large number of ECHOs in this study did not allow for over-reading of each image by a cardiologist. Additionally, although our full guideline does suggest screening for PH prior to 28 days in the setting of severe hypoxemic cardiorespiratory disease (not shown), we did not examine the effects of screening for very early PH, in the first 7–10 days, that has been associated with worse outcomes^29,36. Lastly the power to assess readmission, mortality, and differences in PH treatments was limited by the number of PH+ infants.

Conclusions

We found that a BPD-PH screening protocol with multidisciplinary team involvement was associated with increased detection of echocardiographic signs of PH in at risk infants, yet a shorter NICU stay in PH+ infants. This suggests a benefit to this screening approach in infants born preterm at risk for PH. Further studies are needed to determine the most efficacious strategies for prevention and treatment of the various phenotypes of BPD-PH. Further comparison of the infants in the two epochs with respect to clinical course prior to 28 days in relation to disease development, outpatient neurodevelopmental, and cardiopulmonary follow-up are possibilities for future study.

Highlights.

• First study evaluating outcomes of pulmonary hypertension (PH) screening

• A standardized PH screening algorithm was associated with increased detection of PH

• PH screening was associated with reduced length of stay in infants with PH

• No association with reduced re-hospitalization following PH screening identified

Abbreviations

BPD

bronchopulmonary dysplasia

NICU

neonatal intensive care unit

PH

pulmonary hypertension

GA

gestational age

CGA

corrected gestational age

ECHO

echocardiogram

ATS

American Thoracic Society

URMC

University of Rochester Medical Center

RRH

Rochester Regional Health System

PPHNet

Pediatric Pulmonary Hypertension Network

SGA

small for gestational age

BW

Birthweight

RVSP

right ventricular systolic pressure

RV

right ventricle

NEC

necrotizing enterocolitis

IRB

institutional review board

ICU

intensive care unit

EMR

electronic medical record

CTSI

Clinical and Translational Science Institute

HFNC

high flow nasal cannula

CPAP

continuous positive airway pressure

NIPPV

nasal intermittent positive pressure ventilation

DOL

day of life

NT-ProBNP

N-terminal pro-brain natriuretic peptide

LOS

Length of Stay

Data Sharing Statement:

Deidentified individual participant data will not be made available.