Abstract
Bronchopulmonary dysplasia (BPD) is the most common chronic lung disease in premature infants and is frequently complicated by pulmonary hypertension (PH), which worsens prognosis and increases healthcare utilization. Long-term outcomes beyond 1 year for extremely premature infants (< 28 weeks gestation) with BPD and PH remain poorly characterized. We conducted a single-center retrospective review of infants born < 28 weeks gestation between January 2012 and April 2016, diagnosed with BPD and discharged on home oxygen. Patients were stratified by the presence or absence of PH, defined by echocardiographic criteria. Outcomes assessed through 5 years post-NICU discharge included hospital readmissions, emergency department (ED) visits, duration of home oxygen, and mortality. Among 175 infants (mean gestational age 25.3 weeks), 84 (48%) had PH. Compared to those without PH, infants with PH had lower birth weight (677 g vs. 772 g, P = 0.0006), longer NICU stay (median 109.5 vs. 88 days, P = 0.002), and more frequent tracheostomy (20.2% vs. 8.8%, P = 0.03). At 5 years, PH patients were more likely to have ≥ 1 hospitalization (71.4% vs. 48.4%, P = 0.002) and had higher median readmissions (2 vs. 0, P < 0.0001). ED visit frequency and mortality did not differ significantly. Most PH resolved by one year post-discharge.
|
What is known: • In patients with BPD, PH has been associated with increase hospitalizations by one year of age compared to non-PH infants. |
|
What is new: • This study follows premature infants with BPD and PH for 5 years to decribe their clinical outcome.Premature infants with BPD and PH requrie more healthcare resources than those without PH. There is an ongoing need for tragetedoutpatient strategies to reduce long-term morbidity. |
Conclusions: Extremely premature infants with BPD and PH experience greater healthcare utilization up to 5 years despite PH resolution in most cases. These findings highlight the need for targeted outpatient strategies to reduce long-term morbidity in this high-risk population.
Keywords: Bronchopulmonary dysplasia, Pulmonary hypertension, Premature Infants
Introduction
Bronchopulmonary dysplasia (BPD) is the most common chronic lung disease in infants and occurs most often in premature infants [1, 2]. BPD is a major contributor to prolonged hospitalizations, morbidity, late mortality, and the utilization of healthcare resources post-neonatal intensive care unit (NICU) discharge in premature infants [2]. Pulmonary hypertension (PH), defined as a mean pulmonary artery pressure (PAP) greater than 20 mmHg, is one of the most common co-morbidities associated with BPD [3, 4], occurring in 18–37% of patients with BPD [1, 4]. It is known that PH worsens the prognosis of BPD and the infant mortality for patients with both BPD and PH has been estimated to be between 14 and 48% [1, 5, 6]. In patients with BPD, PH has also been found to be associated with increased frequency of readmissions at 1 year of age [7]. In addition, up to 25% of patients with BPD will continue to have respiratory complications into young adulthood despite the fact that most infants with BPD will no longer require oxygen beyond 2 years of age [8].
For BPD infants born at less than 32 weeks gestation who have evidence of PH, outcomes past 1 year corrected age compared to those without PH have not been well documented [7]. While it is known that premature infants with BPD and PH have a higher rate of readmissions by 1 year as compared to those without PH, it is unknown if this difference persists. Previous studies on outcomes for premature infants with BPD and PH have not focused exclusively on extremely premature infants, who are infants born at less than 28 weeks gestation, and have not evaluated post-NICU hospital visits as an outcome. The overall goal of this study was to describe and compare post-NICU outcomes and hospitalizations in extremely premature infants who have BPD with and without PH up to 5 years of age, including outcomes related to rates of emergency department visits, inpatient hospital admissions, co-morbidities of prematurity, and rates of mortality.
Methods
We conducted a single-center retrospective chart review of patients born between January 2012 and April 2016 and hospitalized in the NICU at Children’s Wisconsin. Inclusion criteria included < 28 weeks gestational age (GA), BPD diagnosis, and NICU discharge on home oxygen therapy. BPD was defined per the National Institute of Child Health and Human Development (NICHD) Workshop definition [9]. Patients who had moderate BPD (oxygen or positive pressure requirement) were reviewed for this study. A BPD echocardiography screening protocol for PH at our institution was used during this study period in which all patients were screened every 3 months while inpatient and every 6 months while outpatient if still on oxygen with additional echocardiograms if clinically indicated. The presence of PH was considered probably or likely present by echocardiography if there was a tricuspid regurgitation gradient > 30 mmHg and/or other indirect findings of PH such as systolic interventricular septal flattening, right ventricular (RV) dilation or RV hypertrophy, post-tricuspid shunt gradient predicting elevated RV pressures, and notching of pulmonary artery Doppler signals [10]. The presence of a patent PDA was noted if present on echocardiogram ≥ 7 days of life.
Home oxygen therapy was provided via nasal cannula, tracheostomy collar, or home ventilator. Patients were excluded if they had congenital diaphragmatic hernia or complex congenital heart disease. Patients with simple cardiac defects including a ventricular septal defect (VSD), atrial septal defect (ASD), or patent ductus arteriosus (PDA) were included in the study. For patients transferred to Children’s Wisconsin after birth, records were collected from their initial hospitalization. Study data were collected and managed using REDCap electronic data capture tools. This study was approved by the Children’s Wisconsin Institutional Review Board.
Demographic and clinical variables were obtained from the electronic medical record. Patient information including sex, gestational age at the time of birth, and birth weight was collected. Associated medical conditions such as the presence of necrotizing enterocolitis (NEC), retinopathy of prematurity (ROP), intraventricular hemorrhage (IVH), or sepsis were also collected and correlated with outcomes between PH and no PH groups.
Outcomes including hospital re-admissions, emergency department only visits, duration of home oxygen, duration of PH, and death were evaluated at 5 years follow-up after NICU discharge. The duration of initial hospitalization was calculated from birth to initial discharge. The time until first hospital re-admission was calculated from the date of discharge to the date of next hospital admission.
Statistical analysis
Comparisons were made between patients with PH vs without (no PH) and also across groups based on the presence of PDA ≥ 7 days. The results are described as a mean with standard deviations. Mean (STD) (for symmetric data) or median (IQR) (for skewed data) were reported for the continuous variables. Categorical variables (N (%) reported) were tested using chi-square/Fisher’s exact test and continuous variables using t-test/Mann–Whitney/ANOVA/Kruskal–Wallis tests as appropriate. The number of hospital re-admissions and emergency department only visits were tested across the groups using a generalized linear model assuming negative binomial (NB) distribution. In the case of multiple comparisons, P-values were adjusted using the false discovery rate. P < 0.05 was considered significant. Statistical analysis was performed using SAS V9.4 (SAS Institute Inc., Cary, NC).
Results
A total of 175 BPD patients were included in the study, with a mean gestational age of 25.3 (SD 1.3) weeks. Of the patients included, 84 patients (48%) had pulmonary hypertension demonstrated on at least one echocardiogram. There were no significant differences between sex and gestational age between the two groups (P = 0.32 and P = 0.28, respectively). The PH group had a significantly lower average birth weight as compared to those patients without PH (677.1 g vs. 771.6 g, P = 0.0006) as shown in Table 1.
Table 1.
Demographics of infants by presence of pulmonary hypertension
| Variables | PH (n = 84) | No PH (n = 91) | P-value |
|---|---|---|---|
| Female sex, N (%) | 46 (54.8) | 43 (47.2) | 0.32a |
| Gestational age (weeks) | |||
| Mean (SD) | 25.3 (1.3) | 25.5 (1.3) | 0.28b |
| Birth weight (grams)* | |||
| Mean (SD) | 677.1 (184.2) | 771.6 (171.5) | 0.0006b |
| Non-Hispanic or non-Latino, N (%) | 76 (90.5) | 78 (85.7) | 0.82a |
aChi-square test. bt-test. * indicates statistical significance with P < 0.05
Patients with PH had a longer median length of initial stay in the NICU (109.5 days with IQR 78.0–142.0 vs. 88.0 days with IQR 62.0–106.0, P = 0.002) and were intubated for a longer period of time (57.0 days with IQR 33.0–75.0 vs. 32 days with IQR 15.0–56.0, P = 0.0005) as compared to patients without PH. In addition, patients with PH were more likely to require respiratory support with an oscillator (65.5% vs. 37.4%, P = 0.0002).
Patients with PH were more likely to undergo tracheostomy prior to discharge from the NICU as compared to patients without PH (20.2% vs. 8.8%, P = 0.03). The rates of co-morbidities of prematurity, IVH, NEC, sepsis, and ROP were not significantly different between the two groups (Table 2). A PDA was present on at least one echocardiogram in 77.3% of patients with PH and 64.5% in those without PH, which was not statistically significant. Patients with PH and the presence of a PDA were more likely to have surgical closure of their PDA as compared to patients without PH (83.0% vs. 50.0%, P = 0.004).
Table 2.
Pulmonary vasodilator medication use during NICU hospitalization
| Number of different pulmonary vasodilator medications | Number of patients |
|---|---|
| 1 | 36 |
| 2 | 14 |
| 3 | 2 |
| Type of medication | Number of patients |
| Sildenafil | 27 |
| Tadalafil | 3 |
| Treprostinil | 1 |
| Epoprostenol | 1 |
| iNO | 38 |
Of patients with PH, 52 (47.6%) patients received pulmonary vasodilators during their NICU admission. Most patients only received one pulmonary vasodilator (N = 36). The most common pulmonary vasodilator used was inhaled nitric oxide (iNO) (38/52 patients). Of the 38 patients that received iNO, 13 patients also received sildenafil (34.2%), while 25 patients received iNO alone (65.8%). The second most common pulmonary vasodilator used was sildenafil (27/52 patients) (Table 3). Of the 52 patients receiving at least one PH medication, 27 (51.9%) patients were started on an additional medication of either a PDE5 inhibitor, prostacyclin, or endothelial receptor antagonist while hospitalized in the NICU, and 88.9% of those patients were discharged home on a pulmonary vasodilator. The median duration of pulmonary vasodilator medication use (excluding iNO) was 335.0 days. The median duration of pulmonary vasodilator medication use (excluding iNO) after NICU discharge was 205.0 days.
Table 3.
Co-morbidities of prematurity and NICU course of infants by presence of pulmonary hypertension
| Variables | PH (n = 84) | No PH (n = 91) | P-value |
|---|---|---|---|
| IVH, N (%) | 41 (48.8) | 32 (35.2) | 0.07a |
| NEC, N (%) | 16 (19.0) | 13 (14.3) | 0.40a |
| Blood culture proven sepsis, N (%) | 25 (29.8) | 19 (20.9) | 0.18a |
| Retinopathy of prematurity, N (%) | 64 (76.2) | 74 (81.3) | 0.41a |
| Length of NICU stay (days) | |||
| Median (IQR)* | 109.5 (78.0–142.0) | 88.0 (62.0–106.0) | 0.002b |
| Length of intubation (days) | |||
| Median (IQR)* | 57.0 (33.0–75.0) | 32.0 (15.0–56.0) | 0.0005b |
aChi-square test. bMann-Whitney test. * indicates statistical significance with P < 0.05
Of the 84 patients with PH demonstrated on at least one echocardiogram, 73 patients had data available to determine the timing of PH resolution based on echocardiogram findings. The median time from initial PH diagnosis to resolution on echocardiogram was 59.0 days. There were 32.9% (N = 24) of patients that still had PH on their echocardiogram prior to NICU discharge. Of the patients who had PH by the time of NICU discharge, 91.7% (N = 22) showed PH had resolved on echocardiogram by 1 year, one additional PH patient resolved by 3 years, and the last PH patient had resolved by 5 years after initial NICU discharge.
During a 5-year post-NICU follow-up, patients with PH on echocardiogram at any time were more likely to require repeat hospitalization after initial discharge compared to patients without PH (71.4% vs. 48.4%, P = 0.002). Patients with PH also had a significantly higher median number of re-admissions per patient (2 with IQR 0–6 vs. 0 with IQR 0–2, P < 0.0001). Of the patients with at least one hospitalization, patients with PH had a higher median (IQR) number of hospitalizations (4 with IQR 2–8 vs. 2 with IQR 1–4, P = 0.0006). In addition, the median number of days from initial hospital discharge to first re-admission was significantly shorter for patients with PH (53.5 with IQR 15.5–145.4 vs. 155.5 with IQR 42.0–331.0, P = 0.03) (Table 4). Patients with PH were more likely to require hospitalization for respiratory, cardiac, and “other” reasons (Fig. 1). However, the percentage of patients with PH who had any ED-only visits was not significantly different compared to patients without PH (79.8% vs. 75.8%, P = 0.53). Additionally there was no significant difference between the median number of ED visits between the two groups (6 with IQR 3–12 vs. 5 with IQR 2–9, P = 0.47).
Table 4.
Hospitalizations by presence of pulmonary hypertension
| PH (n = 84) | No PH (n = 91) | P-value | |
|---|---|---|---|
| At least one post-discharge hospitalization, N (%) | 60 (71.4) | 44 (48.4) | 0.002a |
| Number of hospitalizations all patients (median (IQR)) | 2 (0–6) | 0 (0–2) | < 0.0001b |
| Number of hospitalizations in patients with at least one hospitalization (median (IQR)) | 4 (2–8) | 2 (1–4) | 0.0006b |
|
Days from discharge to first re-hospitalization, median (IQR) |
53.5 (15.5–145.5) | 155.5 (42.0–331.0) | 0.03c |
aChi-square test. bGeneralized linear model. cMann-Whitney test. * indicates statistical significance with P < 0.05
Fig. 1.
Reasons for hospital readmissions for infants by the presence of pulmonary hypertension. * indicates statistical significance with P < 0.05. Note: hospital readmission for cardiac reasons was not noticed in the subjects with no PH
Patients with PH had a significantly longer median (IQR) duration of home oxygen use as compared to patients without PH (189 with IQR 121–276 vs. 134 with IQR 98–250, P = 0.02). There was no significant difference in death at 5 years’ follow-up between the two groups (4.8% vs. 4.4%, P > 0.9).
Discussion
In this retrospective single center study of extremely premature infants with BPD, we found that those infants with BPD and PH are more likely to need earlier hospital admission following discharge from the NICU and had a higher number of hospital re-admissions by 5 years follow-up as compared to infants with BPD but without PH. However, there was no difference in ED only visits or death. Lagatta et al. has shown that in patients with BPD, the presence of PH was associated with several adverse outcomes, including readmission before 1 year of corrected age [7]. We have shown that this difference persists at 5 years as well. Because there was no difference in the rates of co-morbidities of prematurity outside of birth weight between the two groups in our study, the difference in hospitalizations between the two groups seems unlikely to be related to other comorbidities of prematurity other than being small for gestational age.
Infants with PH may have more hospital readmissions due to medical fragility since this group is more likely to have hospital admissions for other causes, such as neurologic, ENT, and injury. This group of patients required supplemental home oxygen for a longer duration of time implying delayed lung maturation. Patients in the PH group also had a lower average birth weight, which has been shown to be associated with higher rates of hospital readmissions in preterm infants [11].
Other studies have shown that pulmonary hypertension accompanies BPD in 10–30% of patients [12]. Within our patient population, we found that 48% of patients with BPD also had PH. This is likely due to a combination of a broad definition of PH, as patients in our study were considered to have PH if it was diagnosed on any single echocardiogram. Also, due to the echo screening protocol in place at our institution, more patients with PH may have been identified compared to institutions without such a screening protocol. However, the true incidence of PH across centers is difficult to know given that the definition of PH can vary. In addition, less than half of PH patients required pulmonary vasodilator therapy during their NICU hospitalization and even less required pulmonary vasodilator therapy upon discharge from the NICU. Most patients had resolution of their PH on echocardiogram by 1 year after NICU discharge. This argues that increased hospitalizations over 5 years in patients with PH are unlikely to be related to PH alone and are more likely related to having more significant physical immaturity.
There are limitations to this study. The severity of PH was unable to be assessed and standardized, so the study population could be skewed towards more patients with mild PH. In addition, infants may have had an emergency department visit or have been readmitted at an outside institution where records were unavailable for our review. Readmissions to a hospital outside of Children’s Wisconsin may be more common in less severe illnesses.
Overall, our findings show that in extremely premature infants with BPD, patients with PH are more likely to have more readmissions at 5 years’ follow-up even though PH generally resolves. This data may provide useful prognostic information for both providers and parents. Moving forward, it will be important to evaluate the types of outpatient resources utilized by this patient population and whether certain resources or interventions lead to fewer hospitalizations or ED visits long term. This could help to guide further outpatient care in these patients. In addition, determining whether treatment of infants with PH medications affects long-term outcomes can also help to shape treatment strategies in the future.
Author contributions
AP and EK designed, wrote and edited the manuscript. SH co-wrote and edited manuscript. APan and MD performed and wrote the statistical analysis, tables and figures of the manuscript. All authors reviewed the manuscript.
Funding
No funding was utilized for this study.
Data availability
Data is derived from retrospective chart review and has PHI that is not suited for public use without IRB approval.
Declarations
Ethics approval
This study was approved by the IRB at Children’s Wisconsin and conducted in accordance with the Declaration of Helsinki.
Consent to participate
This is a retrospective study and IRB approval was granted without the need for direct consent.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.del Cerro, Maria Jesús et al (2014) “Pulmonary hypertension in bronchopulmonary dysplasia: clinical findings, cardiovascular anomalies and outcomes.” Pediatr Pulmonol vol. 49(no. 1):pp. 49–59. 10.1002/ppul.22797
- 2.Davidson Lauren, Berkelhamer Sara (2017) Bronchopulmonary dysplasia: chronic lung disease of Infancy and Long-Term Pulmonary Outcomes. J Clin Med 6(1):4. 10.3390/jcm6010004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Khemani, Ekta et al (2007) Pulmonary artery hypertension in formerly premature infants with Bronchopulmonary Dysplasia: Clinical features and outcomes in the surfactant era pp. 1260–69. 10.1542/peds.2007-0971
- 4.Varghese Nidhy, Rios Danielle (2019) Pulmonary hypertension associated with bronchopulmonary Dysplasia: A Review. Pediatr Allergy, Immunol, Pulmonol 32(4):140–48. 10.1089/ped.2018.0984 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mourani Peter M, Abman Steven H (2015) Pulmonary hypertension and vascular abnormalities in Bronchopulmonary Dysplasia. Clin Perinatol 42(4):839–55. 10.1016/j.clp.2015.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.An, Hyo Soon et al (Mar. 2010) “Pulmonary hypertension in preterm infants with bronchopulmonary dysplasia.” Korean Circ J pp. 131–36
- 7.Lagatta Joanne M et al (Dec.2018) The impact of pulmonary hypertension in preterm infants with severe bronchopulmonary dysplasia through 1 Year. The J Pediatr 203:218–24. 10.1016/j.jpeds.2018.07.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lapcharoensap W, Lee HC, Nyberg A, Dukhovny D (2018Apr) Health care and societal costs of Bronchopulmonary Dysplasia. Neoreviews. 19(4):e211–e223. 10.1542/neo.19-4-e211. (PMID: 33384574; PMCID: PMC7773141) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Jobe AH, Bancalari E (2001) Bronchopulmonary dysplasia. Am J Respir Crit Care Med vol. 163(7):1723-9. 10.1164/ajrccm.163.7.2011060
- 10.Kirkpatrick EC (2013) Echocardiography in pediatric pulmonary hypertension. Paediatr Respir Rev 14(3):157-164. 10.1016/j.prrv.2012.12.008
- 11.Yüksel B, Greenough A (1994) Birth weight and hospital readmission of infants born prematurely. Arch pediatr adolesc med 148(4):384–8. 10.1001/archpedi.1994.02170040050008 [DOI] [PubMed] [Google Scholar]
- 12.O’Connor MG, Cornfield DN, Austin ED (2016Jun) Pulmonary hypertension in the premature infant: a challenging comorbidity in a vulnerable population. Curr Opin Pediatr 28(3):324–30. 10.1097/MOP.0000000000000355. (PMID: 27043088; PMCID: PMC4894759) [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data is derived from retrospective chart review and has PHI that is not suited for public use without IRB approval.