Patent Ductus Arteriosus and Bronchopulmonary Dysplasia–Associated Pulmonary Hypertension

Key Points Question Is patent ductus arteriosus (PDA) associated with a higher risk of developing bronchopulmonary dysplasia–associated pulmonary hypertension (BPD-PH) among very and extremely preterm infants? Findings In this meta-analysis of 32 studies with 8513 infants, the bayesian model-averaged meta-analysis found associations of BPD-PH with both PDA requiring surgery and prolonged PDA. Meaning These findings suggest that sustained patency of a hemodynamically significant ductal shunt may be a key contributor to pulmonary vascular disease in very and extremely preterm infants and highlight the need to monitor for PH in high-risk preterm infants and to incorporate PH risk into clinical decisions regarding PDA management.


Introduction
Bronchopulmonary dysplasia (BPD) is one of the most common complications of very and extremely preterm birth (ie, gestational age <32 weeks) and is an important contributor to pulmonary and nonpulmonary morbidity and mortality. 1BPD is often associated with pulmonary vascular disease and secondary pulmonary hypertension (PH).BPD-associated pulmonary hypertension (BPD-PH) is characterized by arrested vascular growth and abnormal remodeling of the pulmonary vasculature, resulting in high vascular tone and abnormal reactivity, which may lead to increased pulmonary vascular resistance and right heart failure. 2,3The pathogenesis of BPD-PH is multifactorial and involves maternal, placental, fetal, and postnatal factors that disrupt fetal and neonatal growth and development of the pulmonary vascular bed. 3,4BPD-PH is estimated to be present in 25% of infants with moderate-to-severe BPD, and its presence is associated with worse respiratory outcomes and higher hospital readmission and mortality rates. 1,3,4though there is no consensus on the diagnostic criteria for BPD-PH, a growing number of observational studies have been published comparing infants with PH, as assessed by echocardiography, with infants having BPD without PH.[7] Preclinical studies 8,9 have shown that hemodynamic stress impairs pulmonary vascular function, structure, and growth in animal models of perinatal PH, but the impact of a patent ductus arteriosus (PDA) on late clinical outcomes after preterm birth remains uncertain.The presence of or prolonged exposure to a hemodynamically significant PDA (hsPDA) has been suggested as a potential risk factor for developing BPD-PH in some recent cohorts. 10,11[7] Our objective for the current study was to conduct a systematic review and meta-analysis of the association of PDA with BPD-PH.Instead of the more commonly used frequentist statistics, we used a bayesian approach for the meta-analysis.In contrast to frequentist null hypothesis (H 0 ) significance testing, which focuses exclusively on H 0 , Bayesian hypothesis testing aims to quantify the relative plausibility of the alternative hypothesis (H 1 ) and H 0 .[14]

Sources and Search Strategy
A comprehensive literature search was undertaken using the PubMed, EMBASE, and Web of Science databases.The literature search was updated up to February 2023.The search strategy is detailed in the eAppendix in Supplement 1.Studies were included if they examined very and extremely preterm (GA <32 weeks) or very low birth weight (<1500 g) infants and reported primary data that could be used to measure the association of exposure to PDA with the development of BPD-PH. .When a study reported different periods of exposure to PDA, the longest period was selected.In addition to these 6 PDA groups, PDA exposure time was collected as a continuous variable.

Data Extraction, Definitions, and Quality Assessment
Methodological quality was assessed using the Newcastle-Ottawa Scale for cohort or casecontrol studies. 19Newcastle-Ottawa Scale scores of 7 or higher were considered high-quality studies (low risk of bias), and scores of 5 to 6 denoted moderate quality (moderate risk of bias). 19

Statistical Analysis
The effect size of dichotomous variables (PDA exposure) was expressed as log odds ratio (OR), and the effect size of continuous variables (time of exposure to PDA) was expressed using the Hedges g value.Values of log OR or Hedges g and the corresponding SEs of each individual study were calculated using Comprehensive Meta-Analysis statistical software version 4.0 (Biostat).The results were further pooled and analyzed by using bayesian model-averaged (BMA) meta-analysis. 12,13We performed the BMA in JASP version 0.17.3 (JASP Team 2023), which uses the metaBMA R package. 20,21BMA uses Bayes factors (BFs) and bayesian model averaging to evaluate the likelihood of the data under the combination of models assuming the presence vs the absence of the metaanalytic effect and heterogeneity. 12,13The BF 10 is the ratio of the probability of the data under H 1 over the probability of the data under H 0 and was interpreted using the evidence categories suggested by Lee et al. 22 The evidence in favor of H 1 (BF 10 >1) was categorized as weak or inconclusive (BF 10 1 to <3), moderate (BF 10 3 to <10), strong (BF 10 10 to <30), very strong (BF 10 30 to <100), and extreme (BF 10 >100).The evidence in favor of H 0 (BF 10 <1) was categorized as weak or inconclusive (BF 10 1/3 to <1), moderate (BF 10 1/10 to <1/3), strong (BF 10 1/30 to <1/10), very strong (BF 10 1/100 to <1/30), and extreme (BF 10 <1/100).The BF rf is the ratio of the probability of the data under the random-effects model over the probability of the data under the fixed-effect model.The categories of strength of the evidence in favor of the random effects or the fixed effect were similar to those described already for BF 10 .We used the empirical prior distributions based on the Cochrane Database of Systematics Reviews transformed to log OR; log OR ~t (μ = 0, σ = 0.78, ν = 5), and τ ~inverse γ (k = 1.71, θ = 0.73). 12,13We used robust bayesian meta-analysis 23,24 to assess the robustness of the results to the potential presence of publication bias.

JAMA Network Open | Pediatrics
Patent Ductus Arteriosus and Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension

Description of Studies and Quality Assessment
The PRISMA flow diagram of the search process is shown in eFigure 1 in Supplement 1.Of 186 potentially relevant studies, 32 studies were included. 10,11, Thes studies included 8513 infants.
Characteristics of the studies are summarized in eTable 1 in Supplement 1. Risk of bias assessment according to the Newcastle-Ottawa Scale is depicted in eTable 1 in Supplement 1.All studies received a score of at least 7 points, indicating a low risk of bias.The criteria used in the various studies for the diagnosis of BPD-PH are summarized in eTable 2 in Supplement 1.
The adjusted and unadjusted effect sizes are shown in eTable 5 in Supplement 1. Robust bayesian

Discussion
To our knowledge, this is the first bayesian meta-analysis pooling the published data on the association between PDA and BPD-PH as diagnosed by echocardiogram.Bayesian meta-analysis showed that the evidence for an association with BPD-PH was extreme for surgically treated PDA and for the continuous variable duration of PDA exposure.In addition, the evidence was moderate for hsPDA and strong for the dichotomous variable prolonged exposure to PDA.Taken together, the present data support the hypothesis that sustained patency of a hemodynamically significant ductal shunt may be a key contributor to the pathogenesis of BPD-PH. 10,11,54r some clinicians, PDA has acquired the status of a benign event, with treatment resulting in more complications than benefits.6][57] As a result, an increasing number of preterm infants are and will be potentially exposed to a prolonged ductal shunt. 58,59As pointed out by El-Khuffash et al, 60 it is unlikely that a true hsPDA with chronic left-to-right shunting does not play a role in pulmonary  vascular remodeling during the first few weeks of life.Accordingly, the highest evidence for an association between PDA and BPD-PH came from the group of analyses that considered the duration of ductal shunting as a continuous or categorical variable (Figure 4).
It could also be speculated that surgical ligation of PDA is a surrogate for prolonged exposure to hsPDA.This speculation is based on the fact that ligation is usually reserved for those infants in whom conservative or pharmacological treatment of PDA has failed or is contraindicated. 61In a recent retrospective study 62 of a cohort of 2418 very-low-birth-weight infants who underwent invasive PDA closure, the median age at the time of the procedure was approximately 30 days.However, it cannot be overlooked that in several neonatology centers, surgical ligation of PDA has been the preferred first-line treatment for small preterm infants. 61Only 2 of the studies 42,49 included in the metaanalysis provided information on the timing of invasive PDA closure.Interestingly, Lagatta et al 42 observed that the median time at which infants underwent PDA ligation was approximately 2 weeks longer in the BPD-PH group than in the group without BPD-PH.In contrast, Sheth et al 49 did not find any association of early (<28 days) vs late surgical closure of PDA with development of BPD-PH.
The association of PDA with BPD is complex, with a number of seemingly contradictory findings. 63Preclinical studies 64 in baboons showed that exposure for 2 weeks to a moderate-to-large PDA led to a decrease in alveolar surface area and accentuated the arrest in alveolar development that characterizes BPD.However, smooth muscle abundances around terminal bronchioles and their neighboring pulmonary arteries were not affected by exposure to ductal shunt. 64Several recent single-center observational studies 62,[64][65][66][67][68][69] have shown that infants with small PDA shunts do not appear to be at increased risk for developing BPD.6][67][68][69][70] Interestingly, the duration of exposure to mechanical ventilation appears to play an important role in the interaction between PDA and BPD.Thus, Clyman et al 63,71 showed that infants exposed to mechanical ventilation for more than 10 days developed the most severe forms of BPD, particularly when they had concurrent exposure to a moderate-to-large PDA for at least 7 to 14 days.
A reliable assessment of the impact of PDA on BPD risk requires randomized clinical trials (RCTs) in which enrollment is limited to infants with the highest risk of abnormal outcomes and in whom   70 This would ensure that ductal shunt exposure is of sufficient magnitude and duration.Conversely, the control group should achieve a high rate of ductal closure. 70In the last few years, several RCTs have compared early (within the first 24-72 hours after birth) pharmacologic treatment of PDA vs placebo [72][73][74][75] or vs an expectant attitude toward PDA. 76wever, all these RCTs focused exclusively on the incidence of BPD, without consideration of the impact of PDA on pulmonary vascular development or risk of PH.Four trials [72][73][74][75] found no effect of early PDA treatment on the incidence of BPD, and another 76 even reported a higher incidence of moderate-to-severe BPD in the group receiving treatment.Nevertheless, it should be noted that spontaneous ductal closure by the end of the first week occurred in as many as 20% to 50% of infants in the control group of these RCTs.3][74][75][76] Therefore, there was substantial overlap between the intervention and control groups in terms of PDA exposure.
The absence of association between a potential risk factor and BPD is not an indication that it cannot have an association with BPD-PH.The current definition of BPD, which is based on the need for supplemental oxygen and/or respiratory support at 36 weeks postmenstrual age, is an umbrella term that cannot discriminate the contribution of airway, parenchymal lung, interstitium, or pulmonary vascular disease to the respiratory difficulty. 1,3,77In recent years, considerable effort has been devoted to the characterization of various BPD phenotypes.Among the potential phenotypic subgroups, the so-called vascular phenotype is characterized by the associated presence of PH. 1,3,4,77 A pathological condition such as PDA may have a specific effect on the development of the vascular phenotype of BPD without an effect on the overall incidence of BPD.An example of this was found when we recently analyzed the association between hypertensive disorders of pregnancy and BPD.
Although hypertensive disorders of pregnancy were not associated with moderate-to-severe BPD, they were associated with an increased risk of developing BPD-PH.

Limitations
The main limitation of the included studies, and therefore of our meta-analysis, is the difficulty in assessing both exposure (PDA) and outcome (BPD-PH).This may account for the high heterogeneity we found in some analyses.With regard to exposure, a majority of the included studies were retrospective and did not report the time of diagnosis or treatment of PDA or the rate of response to treatment.9][80] It is important to note that the hemodynamic meaning of PDA is affected by factors such as the transductal diameter, the balance between pulmonary and systemic vascular resistance, and the compensatory capacity of the immature myocardium. 69These factors may vary over time and with clinical evolution, but this complexity is not reflected in the dichotomization of the data required by metaanalysis.
Regarding the outcome, cardiac catheterization is the criterion standard test for the diagnosis of PH, but the procedure is generally reserved for a limited group of infants because of its invasiveness. 81,82Echocardiography, on the other hand, has the advantage of being noninvasive and allows for serial clinical evaluation.6][87] The studies included in the present meta-analysis used different combinations of these echocardiographic measurements (see eTable 2 in Supplement 1).However, none of these individual metrics is free from criticism regarding accuracy in the diagnosis of PH and limited ability to differentiate between high flow, elevated pulmonary vascular resistance (PVR), pulmonary venous hypertension from left ventricular diastolic dysfunction, or pulmonary vein stenosis as contributors of PH. 60,[85][86][87][88] The only study included in the meta-analysis that determined the presence of PH by cardiac catheterization was Philip et al. 46 Although that study included a very selective group of infants, its findings are very relevant to differentiate the relative role of high flow vs high PVR in BPD-PH when a ductal shunt is present.They retrospectively analyzed a series of 100 infants with GA less than 27 weeks who underwent hemodynamic evaluation before transcatheter PDA closure.In 64 infants, they observed an elevated pulmonary arterial systolic pressure (defined as pulmonary-to-systemic ratio >0.5).However, only 20 of these 64 infants had an increase in PVR.In the remaining 44 infants, the elevated pulmonary arterial systolic pressure was secondary to increased pulmonary blood flow from the PDA.Interestingly, of the 20 infants with elevated PVR, 17 had been exposed to PDA for more than 8 weeks. 46These data suggest that prolonged exposure to PDA often leads to pulmonary vascular disease.

Conclusions
The findings summarized in this bayesian meta-analysis suggest that prolonged exposure to PDA may be associated with increased risk of pulmonary vascular disease in extremely preterm infants.

Figure 1 .
Figure 1.Summary of Bayesian Model-Averaged Meta-Analysis of the Association of Patent Ductus Arteriosus (PDA) With Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension

Figure 2 .
Figure 2. Bayesian Model Averaged Meta-Analysis of the Association of Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension (BPD-PH) With Patent Ductus Arteriosus (PDA)

Figure 3 .
Figure 3. Bayesian Model-Averaged Meta-Analysis of the Association of Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension (BPD-PH) With Surgically Ligated or Catheter Occluded Patent Ductus Arteriosus (PDA)

1 . 2 . 3 . 4 . 5 .
This emphasizes the necessity of monitoring PH in high-risk preterm infants with prolonged exposure to PDA and incorporating the risk of developing PH into clinical decisions regarding definitive PDA closure by surgery or catheter occlusion.In addition, PH should be included as important outcome in clinical trials of PDA management.JAMA Network Open | PediatricsPatent Ductus Arteriosus and Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension SUPPLEMENT 1. eAppendix.Supplemental Methods eFigure 1. Flow Diagram of the Systematic Search eFigure 2. Bayesian Model Averaged Meta-Analysis on the Association Between Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension and (A) Medically Treated Patent Ductus Arteriosus (PDA), and (B) Medically or Surgically Treated PDA eTable Characteristics of the Included Studies eTable Criteria for Echocardiographic Assessment of Pulmonary Hypertension in the Different Studies eTable Data on Heterogeneity of the Bayesian Model-Averaged Meta-Analysis (BMA) eTable Analysis of Publication Bias by Robust Bayesian Meta-Analysis (RoBMA) eTable Patent Ductus Arteriosus and Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension JAMA Network Open.2023;6(11):e2345299. doi:10.1001/jamanetworkopen.2023.45299(Reprinted) November 28, 2023 15/15 Downloaded from jamanetwork.comby guest on 12/22/2023

JAMA Network Open | Pediatrics
Because this meta-analysis did not involve animal subjects or personally identifiable information on human participants, ethics review board approval and patient consent were not required, in accordance with Dutch law regarding human medical scientific research, which is enforced by the Central Committee on Research Involving Human Subjects.The study was JAMA Network Open.2023;6(11):e2345299. doi:10.1001/jamanetworkopen.2023.45299(Reprinted) November 28, 2023 2/15 Downloaded from jamanetwork.comby guest on 12/22/2023 performed and reported according to Preferred Reporting Items for Systematic Reviews and Metaanalyses (PRISMA) and Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines.The population, exposure, comparison, and outcome question was, do very and extremely preterm infants (population) exposed to PDA (exposure) have a higher risk of developing BPD-PH (outcome) than preterm infants with no history of exposure (comparison)?

Table .
Bayesian Model-Averaged Meta-Analysis of the Association of PDA With BPD-PH is the ratio of the probability of the data under the random-effects model over the probability of the data under the fixed-effect model.
a K denotes the number of studies included in the analysis.bBayesfactor (BF) 10 denotes the ratio of the probability of the data under the alternative hypothesis (H 1 ; association of PDA with BPD-HP) over the probability of the data under the null hypothesis (H 0 ).cCalculated by random effects frequentist meta-analysis.dBFrf e Data are Hedges g values.JAMA Network Open | PediatricsPatent Ductus Arteriosus and Bronchopulmonary Dysplasia-Associated Pulmonary Hypertension JAMA Network Open.2023;6(11):e2345299. doi:10.1001/jamanetworkopen.2023.45299(Reprinted) November 28, 2023 5/15 Downloaded from jamanetwork.comby guest on 12/22/2023 meta-analysis found no evidence for or against publication bias in any of the meta-analyses (eTable 4 in Supplement 1).
Graph shows data for any PDA.B, Graph shows data for hemodynamically significant PDA.The size of the diamonds denotes the 95% credible interval (CrI).OR indicates odds ratio.
6Graph shows PDA as categorical variable.B, Graph shows PDA as continuous variable.The sizes of the diamonds denote the 95% credible interval (CrI).OR indicates odds ratio. BA,