CDH is a severe congenital anomaly that occurs in 1 in 2000 to 4000 live births [2, 3]. Despite advances in prenatal diagnosis and medical therapies, early mortality remains high at 30–60% of affected infants [3]. Respiratory failure secondary to impaired in-utero lung development resulting in pulmonary hypoplasia, pulmonary hypertension, and cardiac dysfunction is one of the leading causes of mortality [3, 4].

Postnatal medical management of CDH is directed toward providing respiratory and cardiovascular support until definitive surgical repair. Medical therapies include sedation and neuromuscular blockade, prevention of acidosis, use of pulmonary vasodilators, cardiac support with inotropes and vasopressors, and gentle ventilation [5]. About 30% of infants with CDH-associated cardiopulmonary compromise end up receiving ECLS, which is associated with 2.7 to 4 times higher mortality after adjusting for disease severity compared to those who do not receive ECLS [6]. As the use of ECLS is not without complications and is usually a last resort, optimizing other medical interventions to decrease the use of ECLS and mortality becomes crucial.

Multiple pulmonary vasodilators targeting different pathways involved in CDH-associated pulmonary hypertension (CDH-PH), such as inhaled nitric oxide, milrinone, and sildenafil, have previously been studied [7]. Recently, prostacyclin, a potent pulmonary arterial vasodilator, has been increasingly used in the neonatal population [4]. A limited number of small observational studies have evaluated the use of prostacyclin or its synthetic analogs in neonates and reported improvement in pulmonary hypertension [8, 9]. Prostacyclin has been commonly used in the pediatric and adult population to treat pulmonary arterial hypertension for the last 30 years [4]. Prostacyclin decreases pulmonary arterial pressure by acting on prostacyclin receptors of the endothelial and smooth muscle cells, leading to smooth muscle relaxation and vasodilation [4].

Prior to the current study, there were several small studies that focused on the efficacy and safety of prostacyclin in infants with CDH-PH. Carpentier et al. and Lawrence et al. both reported improvement in pulmonary hypertension [10, 11], while Skarda et al. reported no difference in mortality [12]. Currently, there is no consensus or guideline for the use of prostacyclin or its synthetic analogs in infants with CDH-PH, especially regarding timing of initiation and duration prior to ECLS cannulation [4].

This study by Ramaraj et al. is the first large study evaluating early intravenous prostacyclin use during the first week of life in infants with CDH and the outcomes of ECLS and mortality. It demonstrated a potential benefit of decreased ECLS use and duration with early prostacyclin therapy after adjusting for disease severity with propensity score matching. Overall, this is a well-designed retrospective cohort study that had many strengths, including robust statistical techniques that help to answer different questions regarding early prostacyclin use in the CDH population. Full cohort analyses examined the effects of early prostacyclin use and ECLS outcomes across all infants with CDH after adjusting for a priori selected previously described covariates associated with poorer outcomes. Propensity matching helped to adjust for disease severity and reduce bias due to differences in baseline characteristics between the exposed and unexposed groups. There were several important baseline imbalances in the exposed and unexposed groups, with infants who received early prostacyclin having more severe disease. Additionally, the finding of early prostacyclin use being associated with increased mortality in the full cohort but not in the matched cohort further suggests infants who received early prostacyclin were generally sicker and thus more likely to receive ECLS and have higher mortality. As infants who receive early prostacyclin therapy are more likely to receive ECLS, the finding of a decrease in odds of ECLS use to 40% of the baseline odds may suggest a meaningful reduction.

While the variables used in propensity score matching were selected based on previously described risk factors associated with ECLS in infants with CDH, there are missing variables that may affect the decision for ECLS use and outcomes, such as arterial partial pressure of oxygen, oxygen index, degree of hypoxemia, ventilator settings, use of vasopressors for blood pressure support, presence of cardiac dysfunction (which was excluded due to incomplete documentation), and presence of other comorbidities. Furthermore, propensity matched analysis can only identify patients similar to the treatment group and adjust for measured confounders. The consequences of unmeasured confounders remain difficult to ascertain. The E-values for receipt of ECLS and categorical ECLS duration were 2.58 and 2.42, respectively, suggesting that only a relatively small effect size of an unmeasured confounder would be needed to fully explain away the observed association in this study. Also interesting, based on the finding of an ICC of 0.24 with ECLS use, is that one-quarter of the total variation in whether an infant received ECLS depended upon the institution providing care after accounting for the other factors in the model, indicating differences between hospitals. Future randomized controlled trials (RCTs) may better evaluate the benefits of early prostacyclin therapy in infants with CDH-PH.