Real-World Outcomes of Treprostinil Treatment in Very-Low-Birth-Weight Infants with Pulmonary Hypertension: A Nationwide Cohort Study

Article information

Neonatal Med. 2026;33(1):56-65

Publication date (electronic) : 2026 May 31

doi : https://doi.org/10.5385/nm.26005

Jieun Jeong, MD¹, Haemin Kang, MD¹, Ho Jung Choi, MD¹, Sohee Oh, PhD², Jin A Lee, MD, PhD^,1

¹Department of Pediatrics, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul National University College of Medicine, Seoul, Korea

²Medical Research Collaborating Center, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul National University College of Medicine, Seoul, Korea

Correspondence to: Jin A Lee, MD, PhD Department of Pediatrics, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul National University College of Medicine, 20 Boramae-ro 5-gil, Dongjak-gu, Seoul 07061, Korea Tel: +82-2-870-2364 Fax: +82-2-831-2826 E-mail: ljinna@snu.ac.kr

Received 2026 March 28; Revised 2026 April 15; Accepted 2026 April 29.

Abstract

INTRODUCTION

Pulmonary hypertension (PH) is a critical complication in very-low-birth-weight infants (VLBWI) and contributes substantially to neonatal morbidity and mortality. Although the incidence of early neonatal PH in VLBWI is approximately 4.6%–8% [1], this rate increases sharply to 18%–40% among VLBWI with bronchopulmonary dysplasia (BPD) and extremely low birth weight infants [2,3]. Despite advances in neonatal care, PH in this population leads to severe respiratory failure and poor prognosis [4].

The management of PH in preterm infants remains challenging because of the limited number of evidence-based strategies available to treat or prevent this condition [5]. Although inhaled nitric oxide (iNO) is widely used as the initial therapy for neonatal PH, a substantial proportion of infants exhibit an incomplete response. Additional pulmonary vasodilators including sildenafil, endothelin receptor antagonists, and prostacyclin analogs are often used off-label in refractory cases [6]. Treprostinil, a prostacyclin analog that can be administered via continuous infusion, has shown potential benefits in small neonatal studies and case reports [7]. However, previous data regarding its use in preterm infants are insufficient, and the overall evidence remains limited owing to small sample sizes, single-center experiences, and heterogeneous underlying conditions [8-10].

Comprehensive real-world data regarding treprostinil use in VLBWI with severe PH, particularly regarding long-term growth and neurodevelopment, are lacking. Thus, we used data from the Korean Neonatal Network (KNN), a nationwide prospective multicenter registry of VLBWI, established in 2013. By analyzing more than a decade of KNN data, this study describes the clinical characteristics, in-hospital morbidities, specific causes of death, and long-term neurodevelopmental outcomes of VLBWI treated with treprostinil for PH. Because of the lack of an optimal comparator cohort, this study focused on presenting the real-world use and descriptive natural history of these infants rather than evaluating the specific efficacy of the drug.

MATERIALS AND METHODS

RESULTS

1. Treatment of PH among VLBWI in the KNN

Among the 21,645 VLBWI admitted to the NICU during the study period, 1,962 (9.1%) underwent pharmacological treatment for PH, after excluding infants with major congenital anomalies (n=741), undetermined sex (n=7), and those without PH (n=18,935) (Figure 1). iNO was the most frequently used agent (n=1,441 [73.4%]), whereas treprostinil was used in 39 infants (2.0%). Combination therapy with ≥2 agents was administered to 562 (28.6%) infants. All infants treated with treprostinil were also treated with other combined PH medications. The frequency and relative annual proportion of treprostinil use increased in recent years (Figure 2).

Figure 1.

Study population.

Figure 2.

Annual proportion of treprostinil use by birth year. Abbreviation: PH, pulmonary hypertension.

Of the 39 infants who received treprostinil, six (15.4%) were transferred to other departments or facilities, and one (2.6%) remained hospitalized for >1 year. After excluding these seven infants due to limited follow-up data after transfer or prolonged hospitalization, 32 patients were included in the final comparative analysis. Of these 32 infants, 14 (43.8%) survived to discharge, including one who was discharged with a subcutaneous treprostinil pump, and 18 (56.3%) died before discharge (Figure 1). Follow-up data were available for only 12 (92.3%) infants at FU1 and eight (88.9%) at FU2. Only seven infants were followed up at both FU1 and FU2.

2. Baseline and demographic characteristics

Maternal age was significantly lower in survivors than in non-survivors (31.1 years vs. 33.8 years, P=0.039). Cord blood pH was significantly higher in the survivor group (7.26 vs. 7.08, P=0.005). No statistically significant differences in other baseline characteristics, including GA, birth weight, sex, and SGA status were observed between the two groups (Table 1). Multivariate logistic regression analysis revealed that cord blood pH was the only independent predictor of mortality (P=0.016), and a decrease in cord pH was associated with a significantly increased risk of death (odds ratio, 0.924 per 0.01 increase in cord pH; 95% confidence interval, 0.812 to 0.989) (Supplementary Table 1).

Table 1.

Demographic and Baseline Characteristics of the Study Population

3. Pharmacological characteristics of treprostinil treatment

The postnatal day at the initiation of PH treatment did not differ significantly between survivors and non-survivors (P=0.203). The combination regimen differed significantly between the two groups (P=0.015). The most common regimen was iNO plus treprostinil (n=15 [46.9%]), which was used significantly more frequently by non-survivors than by survivors (21.4% vs. 66.7%, P=0.011) (Table 2).

Table 2.

Pharmacological Characteristics of Treprostinil Treatment in the Study Population

4. In-hospital outcomes

Compared with non-survivors, survivors exhibited significantly higher rates of moderate-to-severe BPD (85.7% vs. 16.7%, P<0.001), severe BPD (78.6% vs. 16.7%, P<0.001), steroid use for BPD (71.4% vs. 11.1%, P<0.001), PDA requiring treatment (64.3% vs. 11.1%, P=0.003), and ROP requiring surgery or anti-vascular endothelial growth factor therapy (85.7% vs. 11.1%, P<0.001). The duration of parenteral nutrition, hospital stay, invasive ventilator support, and total respiratory support were all significantly longer in survivors (all P<0.001). Six survivors (42.9%) were discharged with supplemental oxygen or home ventilation (Table 3). However, these higher incidences in the survivor group are primarily attributable to a time-at-risk bias, as nonsurvivors died earlier before these chronic complications could manifest.

Table 3.

Comparison of In-Hospital Outcomes between Survivors and Non-Survivors Treated with Treprostinil

6. Long-term outcomes of survivors

Growth restriction persisted throughout the follow-up period. Among the cohort, 58.3% had a body weight below the 10th percentile at the FU1 visit, and 87.5% had both weight and height below the 15th percentile at the FU2 visit (Tables 4, 5).

Table 4.

Growth Parameters from Discharge to Follow-up at 18–24 Months and 3 Years Corrected Age

Table 5.

Medical Treatment from Discharge to Follow-up at 18-24 Months and 3 Years Corrected Age

Oxygen therapy and PH medications were required by 58.3% and 36.4% of survivors, respectively, from discharge until the FU1 visit. These requirements decreased over time, and none of the patients remained on oxygen therapy or PH medication during the FU2 visit among those assessed. However, rehabilitation and language therapy remain substantial. For infants who completed both follow-up visits (n=7), the cumulative treatment rates were 71.4%, 85.7%, and 50.0% for oxygen therapy, rehabilitation, and speech therapy, respectively (Tables 4, 5).

Major neurodevelopmental impairments such as CP, blindness, or hearing loss requiring aid were uncommon; however, various degrees of developmental delay were identified. At the FU1 visit, cognitive, language, and motor domain delay was observed in patients with BSID-II and BSID-III. At the FU2 visit, 50% of infants assessed with the K-DST demonstrated significant gross motor delay (<−2 SD), and 33.3% exhibited delay in the fine motor, cognitive, language, social, and self-help domains. Overall, developmental delays were observed in more than half the survivors at FU1 (75.0%) and FU2 (50.0%) (Table 6).

Table 6.

Long-Term Neurodevelopmental Outcomes at 18–24 Months and 3 Years Corrected Age

DISCUSSION

To our knowledge, this is the first nationwide registry study to investigate the use of treprostinil for PH treatment in VLBWI. We comprehensively evaluated the clinical characteristics, in-hospital morbidities, and long-term neurodevelopmental outcomes of VLBWI who underwent treprostinil therapy for PH. Among 1,962 VLBWI treated for PH over a 10-year period, treprostinil was administered in only 2.0% of cases. The overall mortality rate of the treprostinil group was extremely high (56.3%). Furthermore, although the survivors eventually discontinued oxygen and PH medications, they experienced significant neonatal morbidities and continued to exhibit persistent growth failure and developmental delays during the long-term follow-up.

Based on its short-term hemodynamic and oxygenation benefits in both term and preterm infants, treprostinil is increasingly considered an adjunctive therapy for refractory PH, despite not being approved for neonatal use. The published literature addressing neonatal treprostinil use has gradually expanded over the past decade; however, the available evidence remains limited to case reports, small retrospective cohorts, and selected populations such as infants with BPD-associated PH [7-10]. Given these considerations, our nationwide analysis is meaningful because it revealed a temporal increase in treprostinil use in real-world practice and provided short- and long-term outcome data for preterm infants.

The high mortality rate observed in this cohort likely reflects the severity of the underlying cardiopulmonary disease rather than the adverse effects of the medication itself. Most deaths (83.3%) were attributed to cardiopulmonary causes, with PH accounting for 53.3% of all deaths. This finding is consistent with those of previous studies reporting a poor prognosis of PH in preterm infants unresponsive to iNO [15,16]. In the multivariate regression analysis, lower umbilical cord blood pH was the only potential prognostic factor for in-hospital mortality. By contrast, the 5-minute Apgar scores did not show a statistically significant difference between the two groups. This discrepancy may be partly attributable to the observer-dependent nature of Apgar scores and the fact that they may be immediately altered by resuscitative interventions in the delivery room, whereas cord blood pH may represent a more objective indicator of fetal hypoxia. A low cord blood pH reflects the degree of perinatal deterioration and fetal hypoxia at birth [17], and perinatal asphyxia is a well-known factor that induces irreversible pulmonary vascular remodeling and severe persistent PH in newborn [18]. However, given the small number of outcome events in our cohort, this multivariate model is susceptible to overfitting. Furthermore, the prognostic value of cord blood pH in VLBWI with PH requiring treprostinil therapy has not been previously reported. Therefore, it should be considered a potential prognostic factor that requires validation in larger prospective studies.

Treprostinil is typically administered as a combination therapy in a minority of infants receiving medication for PH. Thus, it was selected for infants with treatment-refractory diseases rather than as a first-line therapy. This therapeutic approach is consistent with the current global clinical practice for PH. In neonatal PH, iNO remains the standard initial therapy, whereas prostacyclin analogs, such as treprostinil, are generally considered adjunctive or step-up therapies in cases of inadequate response [6,7,19]. Consistent with these clinical practices, our data demonstrated an increasing trend in treprostinil use over the study period.

Treprostinil was administered to all infants in combination with other vasodilators, and a significant difference in the composition of the combination therapy was observed between survivors and non-survivors. The proportion of infants receiving a combination of iNO and treprostinil was significantly higher among non-survivors than among survivors. However, this difference should be interpreted cautiously because it likely reflects greater disease severity rather than an adverse effect of the treatment. In support of this, a previous retrospective cohort study from Korea [7] has reported no clinically significant adverse events after treprostinil therapy in neonates with persistent PH. Furthermore, the postnatal day at PH treatment initiation differed between the two groups (median, 6.5 days in survivors vs. 1.0 day in non-survivors). Although this difference was not statistically significant, likely because of the small sample size, it suggests potential phenotypic heterogeneity within our cohort. Early administration of medication in the non-survivor group may represent an early-onset PH phenotype, whereas the survivor group may predominantly reflect a late-onset BPD-associated PH phenotype. Consequently, in the non-survivor group, disease progression may have been so rapid and fulminant that death occurred before there was sufficient time to escalate the therapy stepwise or add additional pulmonary vasodilators such as sildenafil or treprostinil. By contrast, survivors likely experienced a relatively less severe or more gradual disease course, which provided the time necessary to initiate and optimize multidrug therapies.

In-hospital morbidities, including moderate-to-severe BPD, treated PDA, and ROP requiring intervention, were more prevalent among the survivors. However, these descriptive findings likely reflect a survival bias and do not represent direct comparative risks. Non-survivors died before these chronic complications fully developed, whereas survivors had sufficient time for them to manifest. Survivors also experienced an extended duration of parenteral nutrition, hospitalization, invasive ventilation, and total respiratory support. Given that prolonged positive-pressure ventilation and hyperoxia directly contribute to the pathophysiology of severe BPD and subsequent complications [20], such as PH and ROP [21], extended exposure to intensive respiratory support likely drove the higher morbidity burden observed in surviving infants. Although selection bias severely limits direct statistical comparisons, the exceptionally high rate of treated ROP in our survivors (85.7%) compared to the general VLBWI in Korea (11.5%) [22] descriptively highlights the profound morbidity burden of refractory PH.

Long-term follow-up of survivors revealed discrepancies in cardiopulmonary recovery, growth, and neurodevelopmental outcomes. The need for respiratory support decreased significantly over time, and all infants were successfully weaned off supplemental oxygen and PH medications by 3 years of age. This finding is consistent with that of a previous study investigating BPD-associated PH, which suggested that most patients could be weaned off oxygen and PH therapies by 5 years of age [23].

In contrast, our cohort exhibited high rates of growth failure (58.3% at 18–24 months; 87.5% at 3 years) and developmental delays, with 50% requiring speech therapy and >80% requiring continuous rehabilitation over 3 years. These morbidity rates were substantially higher than those of the general Korean VLBWI. Previous studies using KNN data have reported significantly lower rates of growth failure (20.3% at 18–24 months; 35.2% at 3 years) [24], rehabilitative support, and speech therapy (34.9%–37.6%, approximately 5% at 18–24 months) [25]. These findings are consistent with those reported in the existing literature, demonstrating that PH independently worsens growth and neurodevelopmental outcomes in preterm infants [26] and highlight the need for long-term systemic follow-up.

The present study has some limitations. First, as a descriptive, uncontrolled study without a comparator cohort of infants with PH who were not treated with treprostinil, the efficacy of treprostinil could not be determined. Second, treprostinil was selectively administered to infants with treatment-refractory disease as part of the combination therapy. Thus, our results are limited by a strong indication bias and confounding by severity. Third, the small sample size limits interpretation owing to low statistical power; thus, predictive findings, including cord pH, are strictly exploratory. Fourth, the 10-year study period introduces the potential for an ‘era effect’ as the overall neonatal intensive care and PH management strategies have changed considerably. Fifth, phenotypic heterogeneity within the cohort was likely, as suggested by the varying timing of treatment initiation. Owing to the limited sample size, we could not systematically stratify and analyze these subgroups independently. Finally, the absence of a standardized protocol and detailed registry data regarding the precise dosage, route, and duration of treprostinil administration limits the clinical and pharmacological interpretations of our results. Nevertheless, this nationwide study provided important real-world data regarding the use of treprostinil in VLBWI.

In conclusion, in this nationwide cohort, treprostinil was used as rescue therapy for refractory PH in VLBWI and was associated with high short-term mortality. Despite the successful weaning of survivors from oxygen and PH medications, the long-term impact is profound and is characterized by exceptionally high rates of severe neonatal morbidities, persistent growth failure, and substantial neurodevelopmental delays up to 3 years of age. These findings highlight the necessity for multidisciplinary long-term follow-up of this high-risk population and underscore the need for large prospective studies to develop standardized treatment guidelines.

Supplementary materials

Article information Continued