The Use of Treprostinil for Bronchopulmonary Dysplasia Associated Pulmonary Hypertension

Authors: Stephanie M. Tsoi, Claire Parker, Elizabeth Colglazier, Shannon Cheung, Mariam Taleb, Hythem Nawaytou, Elena Amin, Jeffrey R. Fineman, Roberta L. Keller

Abstract

Background

Treprostinil for the treatment of bronchopulmonary dysplasia‐associated pulmonary hypertension (BPD‐PH) has previously been described in small cohort studies, often used later in the course after failure to improve on other therapies.

Objective/Methods

We retrospectively describe the clinical course and outcomes of 18 infants (gestational age 26.3 ± 2.6 weeks) from 2012 to 2025 who received parenteral treprostinil to treat BPD‐PH, including changes in echocardiographic and cardiac catheterization parameters.

Results

All patients had moderate‐to‐severe BPD and PH, with a mean pulmonary arterial pressure of 45.6 ± 12.7 mmHg at cardiac catheterization prior to treprostinil. Treprostinil was initiated at a median postmenstrual age of 53.5 (IQR: 45.7, 62.6) weeks. Echocardiograms after 3 months of treatment showed improvement of PH severity. At repeat catheterization, mean pulmonary arterial pressure (delta −16.4 ± 12.2, p < 0.01) and indexed pulmonary vascular resistance (delta −4.2 ± 3.3, p < 0.01) significantly improved. Ten of 18 infants (55.6%) survived to discharge; BNP ≥ 35 pg/mL prior to treprostinil initiation demonstrates potential utility for mortality prediction with area under the receiver operator characteristic curve 0.87 ± 0.10 (95% CI: 0.67–1.00).

Conclusion

Our study shows a potential benefit of treprostinil use in moderate‐to‐severe BPD‐PH; larger studies are needed to validate our findings and guide decision‐making around treprostinil initiation and duration.

Introduction

Bronchopulmonary dysplasia (BPD) is a form of chronic lung disease that affects premature infants [1]. Twenty percent of infants with BPD will develop pulmonary hypertension (PH), and have a 4.7‐fold increased risk of mortality [2]. The pathogenesis of BPD‐associated PH (BPD‐PH) is characterized by exposure of the immature lung to prenatal and postnatal injurious conditions, such as infection, prolonged mechanical ventilation, and supplemental oxygen exposure, ultimately leading to impaired growth, structural remodeling of the pulmonary vasculature and increased pulmonary vascular resistance (PVR) [3, 4]. Without appropriate monitoring and treatment, BPD‐PH can lead to right ventricular failure and death [5].

Current strategies to diagnose and treat BPD‐PH are limited and based on expert opinion [6, 7]. Recommendations for the initial approach include involvement of a multidisciplinary team, screening echocardiograms for patients with a diagnosis of BPD, and optimization and treatment of comorbid cardiopulmonary conditions. Severity of BPD‐PH should be assessed via cardiac catheterization when it is safe and feasible to guide a targeted treatment approach. Treatment can range from judicious use of supplemental oxygen, inhaled nitric oxide (iNO), and/or PH‐targeted medications [7, 8, 9]. PH‐targeted drug use in this population is off‐label and highly dependent on institutional experience [8].

After supplemental oxygen, sildenafil and iNO remain the most widely used pulmonary vasodilator therapies for BPD‐PH, while neonatal data from randomized studies of these medications are in persistent pulmonary hypertension of the newborn (PPHN), within hours or days of birth [7, 9, 10]. Endothelin receptor antagonists, like bosentan, and prostacyclin analogues, such as treprostinil, target alternative pathways for pulmonary vasodilator therapy and have been used with limited evidence as adjunctive therapies when BPD‐PH persists, is severe, or life‐threatening.

Parenteral treprostinil requires specialty care to initiate and monitor, and stable access for continuous delivery. Three small cohort studies describe treprostinil as well‐tolerated and associated with improved echocardiographic markers of right ventricular function and PH in infants with BPD [11, 12, 13]. These publications show the feasibility of treprostinil use in BPD‐PH, but larger cohort studies are needed to validate the findings and expand outcomes, including cardiac catheterization data.

We describe our institution's experience using parenteral treprostinil to treat infants with BPD‐PH, and report baseline conditions and clinical outcomes after treprostinil initiation, including hemodynamics at cardiac catheterization, echocardiographic parameters, and B‐type natriuretic peptide (BNP) levels (as a predictor of outcome).

Methods

This is a single‐center, retrospective cohort study from the University of California San Francisco, an academic institution with a quaternary children's hospital that provides specialized referral‐based care in BPD and PH. Infants were included if they had PH secondary to BPD, had treprostinil initiated during their initial neonatal admission (including following transfer to our center), and were either discharged or deceased at the time of data collection in June, 2025 (birth dates 2012–2023). Congenital heart disease was not an exclusion criteria, although there were no infants with complex congenital heart disease treated with treprostinil. The primary outcome was survival to discharge. Secondary outcomes included duration of treprostinil use, cardiac catheterization hemodynamics, echocardiographic findings, and BNP levels.

BPD severity was classified per Jensen et al. criteria based on level of respiratory support received at 36 weeks' postmenstrual age (PMA): Grade 1—nasal cannula ≤ 2 L/min, Grade 2—nasal cannula > 2 L/min or noninvasive positive pressure ventilation, and Grade 3—invasive mechanical ventilation [1]. PH was clinically evaluated by echocardiography, with diagnosis confirmed by cardiac catheterization. All echocardiograms were overread for the purposes of the study by a cardiologist blinded to clinical outcomes (MT) using previously established echocardiographic tricuspid regurgitant (TR) jet velocity, pulmonary regurgitant (PR) jet velocity, ventricular septal defect (VSD) and patent ductus arteriosus (PDA) shunt direction and flow velocity (if present), interventricular septal (IVS) wall motion and systolic eccentricity index (EI), and tricuspid annular plane systolic excursion (TAPSE, a marker of systolic right ventricular function, with normal values for age corrected for prematurity [14, 15]). We classified infants into three categories of PH based on the degree of elevation of the estimated right ventricular pressure (RVP): mild (< 50% systemic RVP), moderate (≥ 50% systemic RVP, but less than systemic RVP), and severe (≥ systemic RVP) [16]. Echocardiographic findings from two timepoints were at time of initiation of treprostinil (baseline) and 3 months ± 2 weeks after initiation of treprostinil (3 months). Cardiac catheterization parameters for PH diagnosis included mean pulmonary arterial pressure (mPAP) greater than 20 mmHg and indexed PVR (PVRi) greater than or equal to 3.0 WU/m [2].

Treprostinil initiation is guided by a multidisciplinary PH team using a comprehensive evaluation of each patient (including echocardiography, cardiac catheterization, BPD disease severity and course, functional class and current therapies) to determine treatment approach [17]. Respiratory support, diuretics, and adjunctive pulmonary vasodilatory therapy (i.e. modest provision of supplemental oxygen, iNO, and sildenafil or bosentan if started at prior institutions) were optimized before treprostinil initiation.

Descriptive data are reported as means with standard deviation, medians with interquartile range, or counts and frequencies. Comparisons across groups were performed for categorical data via chi‐squared or Fisher's Exact test, as appropriate. Continuous parametric data were compared using t‐test and paired t‐test, and non‐parametric data were compared using Wilcoxon Rank Sum test and Wilcoxon Signed‐Rank test. Patients with missing values for a particular variable were retained in the cohort, but excluded for that particular analysis. We modeled the relationship between BNP before initiation of treprostinil and mortality using a receiver operating characteristic (ROC) curve and the area under the curve (AUC) to report discrimination.

For all analyses, statistical significance was set at a p‐value ≤ 0.05. All analyses were performed using Stata SE (Statacorp. 2021. Stata Statistical Software: Release 17. College Station, TX: Statacorp LLC). The study protocol was approved by the institutional review board of University of California, San Francisco (IRB Protocol #23‐39071), which waived requirement for informed consent due to the retrospective nature of the study.

Results

Eighteen infants with BPD‐PH used treprostinil, with an average gestational age of 26.3 ± 2.6 weeks and average birth weight of 718 ± 300 g. Seventeen (94.4%) infants were transferred from an outside hospital at an average PMA of 52.3 ± 13.9 weeks. Seven (38.9%) infants were small‐for‐gestational age and 5 (27.8%) were products of multiple gestation. Ten (55.6%) had moderate (Grade 2) BPD and 8 (44.4%) had severe (Grade 3) BPD. PH was diagnosed by echocardiography at an average PMA of 43.0 ± 14.3 weeks. A total of 12 (72.2%) patients had a PDA, of which 8 needed it procedurally closed. 5 (27.8%) had pulmonary vein stenosis, 3 of whom developed it before treprostinil initiation; 1 of these 3 underwent balloon angioplasty during the pre‐treprostinil catheterization (Table 1).

Ten (55.6%) infants survived to discharge. Of survivors, the average PMA at discharge was 95.8 ± 30.8 weeks (Table 1). For non‐survivors, average PMA at death was 72.2 ± 17.1 weeks. Cause of death included sepsis (n = 2), pulmonary vein stenosis (n = 1), acute cardiac event (n = 2), and PH crisis (n = 1) (Table S1). When categorized by treatment approach, 14 infants were treated with upfront treprostinil therapy based on diagnosis of significant pulmonary hypertension by cardiac catheterization (upfront therapy) and 4 (28.6%) of those infants died. In contrast, 4 infants were treated with treprostinil in response to acute signs and symptoms of worsening cardiopulmonary disease (salvage therapy) and all of these infants died (100%, p‐value = 0.02, Table S2). We also evaluated mortality before and after 2018, when our institution changed its approach of mechanical ventilation for infants with severe BPD receiving chronic invasive ventilation. The mortality of infants with treprostinil initiated after this change (n = 11) was 27.3% compared to mortality of 71.4% prior to the ventilatory strategy change.

Prior to treprostinil initiation, 6 (35.3%) infants were treated with bosentan and 10 (58.8%) patients were treated with sildenafil (5 were on dual therapy). Treprostinil was initiated at average 55.8 ± 12.1 weeks' PMA and 15.2 ± 15.7 days after cardiac catheterization, with 14 (77.8%) infants receiving initial subcutaneous administration. The average peak treprostinil dose was 60.4 ± 13.2 ng/kg/min amongst survivors. Treprostinil discontinuation in survivors was at average 136.0 ± 38.1 weeks' PMA, with total duration of treprostinil 81.1 ± 33.7 weeks (Table 2). There was no significant mortality difference seen in patients who had treprostinil started before 50 weeks' PMA (4/8, 50%) compared to at or after 50 weeks' PMA (4/10, 40%, p‐value 0.67).

Cardiac catheterization was performed in 17 (94.4%) infants before treprostinil initiation (Table 3). Of these infants, 11 (61.1%) underwent a second catheterization at an average of 42.4 ± 43.0 weeks after treprostinil initiation; only 2 of the 11 were non‐survivors. The change in mPAP with treprostinil was −16.4 ± 12.2 mmHg (p < 0.01) and change in PVRi was −4.2 ± 3.3 WUi (p < 0.01).

All infants (n = 18) had an echocardiogram prior to starting treprostinil, and 13 had repeat echocardiograms at 3 months of treprostinil initiation (Table 4). Six infants had either an ASD or VSD that may have affected echocardiograph parameters, and 13 had a PDA. All patients had moderate (66.7%) or severe (33.3%) PH before treprostinil initiation by echocardiogram, and at 3 months, only 15.4% of patients remained with severe PH. Median baseline EI was 1.42 (IQR: 1.35, 1.53; n = 16) and at 3 months, median EI was 1.38 (IQR: 1.19, 1.49; n = 12) [18]; by repeated measures analysis, this change was not significant (p‐value 0.46). Average TAPSE at baseline was 1.12 ± 0.26 cm (n = 14) and at 3‐months 1.40 ± 0.32 cm (n = 8); 3 survivors and 1 non‐survivor had a baseline TAPSE below the normal range for age, which normalized for the 1 survivor and 1 non‐survivor that had follow‐up echocardiograms at 3 months. Infants who survived tended to have less severe PH at baseline, with only 20% severe PH compared to 50% severe PH for non‐survivors, consistent with a somewhat lower EI [median: 1.42 (IQR: 1.34, 1.53) vs. 1.49 (IQR: 1.36, 1.65)].

Average BNP before initiation of treprostinil was 55.9 ± 79.0 pg/mL (n = 15). BNP level before treprostinil initiation was predictive of mortality at a threshold value ≥35 pg/mL with 71.4% sensitivity and 100.0% specificity (AUC: 0.87 ± 0.10, 95% CI: 0.67–1.00, Figure 1). At time of repeat catheterization, average BNP had decreased to 14.0 ± 9.3 pg/mL (n = 7, p‐value 0.26).

Discussion

To date, this study describes the largest cohort of premature infants with BPD‐PH treated with treprostinil that includes echocardiographic and cardiac catheterization data. All infants had moderate or severe BPD (Grade 2 or 3) and PH by echocardiography and cardiac catheterization, with a baseline median PVRi of 8.2 (6.5, 9.8). Treprostinil was typically started at 6–7 months chronological age (PMA 53.5 weeks (45.7, 62.6)) via subcutaneous administration. Repeat echocardiograms 3 months after initiation of treprostinil in the context of optimizing usual care showed improvement in PH severity, with 85% of echoes with mild‐to‐moderate PH. Repeat cardiac catheterization data showed substantial improvement in mPAP (11.5 mmHg (10, 18)) and PVRi (3.7 WUi (2.4, 5.7)). Although average duration of treprostinil therapy in survivors was 81.1 ± 33.7 weeks, median age at discontinuation among survivors was less than 2 years corrected age. These findings suggest that treprostinil therapy in selected infants with BPD‐PH with optimized respiratory support and overall neonatal growth may play a role in improving hemodynamics and pulmonary vascular resistance.

The mortality rate of our cohort was 44.4%, which is similar to a prior “natural history” study of infants with BPD‐PH, while those infants overall had less severe BPD, with none supported by invasive ventilation at 36 weeks' PMA [19]. Although co‐morbidities of prematurity were common in our cohort, we found no significant differences between survivors and non‐survivors. However, there are several factors that may influence the mortality rate in our cohort. First, our sample includes both infants who were treated with treprostinil as part of aggressive upfront therapy following cardiac catheterization (n = 14, 28.6% mortality), and infants who received treprostinil as salvage therapy after signs and symptoms of worsening disease (n = 4, 100% mortality). The comparison of various markers of PH severity and illness (Table S2) suggests that the addition of treprostinil has more potential to influence outcomes when initiated as upfront therapy in infants with severe pulmonary vascular disease who are not in crisis, rather than as rescue during acute decompensation. The second factor is our institution's change in ventilator management in 2018 for infants with severe Grade 3 BPD, to align with best practices toward achieving stability and limiting hyperoxia exposure in infants receiving chronic ventilation [20]. Our data compare favorably to that of Arjaans et al. among the cohort of infants on positive pressure at 36 weeks' PMA (CPAP only), as their mortality was 75% (6 of 8 infants) [19]. Overall, these two factors confound evaluation of the positive effect of treprostinil that we observed in our small sample size, and emphasize the need for larger, multi‐center reports to better understand the specific effect of treprostinil treatment on infants with Grade 2 or 3 BPD.

Previous studies have shown utility in using BNP as a biomarker to predict mortality and disease severity in pediatric patients with pulmonary hypertension [21, 22, 23]. Specific to BPD, Konig et al. showed that higher BNP values at 36 weeks corrected age (median: 54.5 pg/mL) were associated with evidence of tricuspid regurgitation on echocardiogram [24]. Cuna et al. evaluated the association of BNP and mortality in 36 infants with BPD‐PH and found that a peak BNP greater than 220 pg/mL at any time had 90% sensitivity and 65% specificity for predicting mortality [25]; none of these infants were treated with treprostinil. We found that BNP was significantly lower in infants treated with treprostinil as upfront therapy rather than salvage therapy, and that a BNP value greater than or equal to 35 pg/mL prior to treprostinil initiation was predictive of mortality. Only one infant with BNP measured prior to treatment who received upfront treprostinil therapy had a BNP greater than or equal to 35 pg/mL and that patient did not survive. Our cutoff is lower than previously reported BNP levels associated with survival in this population, and may be a reflection of our small sample size. Before clinical application of BNP to aid in decision‐making regarding treprostinil initiation, further studies of greater power should be completed.

A multidisciplinary team is the foundation of our institutional approach to treatment of BPD‐PH, including the importance of early recognition and multi‐pronged management of PH. We recommend transferring infants with echocardiographic evidence or signs and symptoms of PH to a specialized center of excellence with the ability to perform cardiac catheterization and initiate treprostinil, if indicated by the evaluation. As this patient population is heterogeneous, careful individualized assessment of the infant should be considered by a team of neonatologists, pulmonologists, cardiologists and pulmonary hypertension specialists before treprostinil initiation. Further, given that pulmonary vein stenosis is frequently missed by echocardiography in former preterm infants [26], particular attention should be given to the pulmonary veins during the comprehensive assessment, as pulmonary vasodilation with treprostinil in the setting of distal obstruction can be harmful.

Initiating treprostinil starts by placing a subcutaneous site, if size and condition of the patient allow, in order to avoid the need for long‐term secure intravenous access and risk of central line infection. A typical starting dose is 1–2 ng/kg/min, with initial increase by 1 ng/kg/min every 6–12 h to target dose of 50–80 ng/kg/min. Uptitration is slowed if there are side effects (i.e. hypotension, emesis, diarrhea). We monitor clinical progress (i.e. ability to tolerate intercurrent illness, recovery from respiratory events, weaning of other vasoactive medications and iNO) as well as interval monthly echocardiograms and BNP to assess RV pressure estimate and function.

Although this is a single‐center study, it is the largest cohort of this type to date. As an academic referral center, accredited as a pulmonary hypertension care center by the Pulmonary Hypertension Association with a multidisciplinary team of experts in PH and BPD, we were able to describe the cardiac catheterization data and changes with therapy in our cohort. However, the specialized nature of our referral population may introduce bias and limit generalizability of our findings. Further, the small sample size of 18 infants is a limitation to adjusting for various differences between the survivor and non‐survivor groups. This includes limits on informing optimal timing of treprostinil initiation and other criteria that might suggest a positive response to treprostinil, such as a precise BNP cutoff and hemodynamic and clinical characteristics. Finally, our cohort spans 10 years which includes practice variation across the years, as noted. Multi‐center studies with larger sample sizes and controlled trials will allow us to better understand the additive improvement ascribed to treprostinil, beyond the natural trajectory of pulmonary vascular disease in moderate‐to‐severe BPD, with ongoing lung growth and development, contemporary ventilatory, and optimal supportive practices.

In conclusion, we described the clinical course and outcomes of 18 infants with moderate to severe BPD‐PH treated with treprostinil via a multidisciplinary approach. Treprostinil administration was associated with improved grade of PH by echocardiography, while significantly decreasing mean pulmonary arterial pressure and PVRi on cardiac catheterization. Future studies are needed to create data‐driven clinical practice guidelines that utilize treprostinil.

Author Contributions

Stephanie M. Tsoi: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, project administration, data curation, resources. Claire Parker: conceptualization, investigation, writing – review and editing, data curation. Elizabeth Colglazier: conceptualization, investigation, writing – review and editing, data curation. Shannon Cheung: data curation, writing – review and editing. Mariam Taleb: writing – review and editing, data curation. Hythem Nawaytou: supervision, writing – review and editing. Elena Amin: writing – review and editing, data curation. Jeffrey R. Fineman: writing – review and editing, supervision, funding acquisition, resources. Roberta L. Keller: conceptualization, investigation, methodology, validation, writing – review and editing, formal analysis, supervision, resources.

Ethics Statement

This study was approved by the University of California, San Francisco IRB (23‐39071), which approved waiver of consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Previous Presentation

Western Society for Pediatric Research, Carmel‐by‐the‐Sea CA (January 16‐18, 2025).

Supporting information