Mortality Patterns in Pediatric Pulmonary Vein Stenosis: Insights Into Right Ventricular Systolic Pressure Associations

Authors: Daiji Takajo, Paul J. Critser, Michelle Cash, Melissa Magness, Russel Hirsch

Abstract

Background

Pulmonary vein stenosis in children is associated with a poor prognosis. However, the cause and risk factors for mortality remain uncertain.

Methods

This retrospective, single‐center study identified children with primary and secondary pulmonary vein stenosis through a cardiac catheterization database. Kaplan‐Meier analysis, log‐rank tests, and Cox regression analysis were performed to assess outcome and identify significant predictors of mortality.

Results

Among 56 children with pulmonary vein stenosis (33 male children, 59%), 20 (36%) died at a median age of 10 months (interquartile range, 4–24 months). All patients underwent cardiac catheterization, with 45 (80%) undergoing at least 1 interventional procedure. Causes of death included multiorgan failure (35%), progressive respiratory failure (20%), and sudden cardiac death (15%). Prematurity, chronic lung disease, a genetic syndrome, or the number of affected pulmonary veins did not significantly correlate with mortality. However, right ventricular (RV) systolic pressure greater than half systemic pressure was associated with mortality (hazard ratio [HR], 5.5 [95% CI, 2.2–14.1]; P<0.001). The final predictive model for mortality included RV systolic pressure greater than half systemic pressure (HR, 4.0 [95% CI, 1.6–10.4]; P=0.004), moderately or severely diminished RV systolic function (HR, 3.6 [95% CI, 1.1–11.5]; P=0.032), and the presence of congenital heart disease (HR, 2.4 [95% CI, 0.9–6.7]; P=0.084).

Conclusions

This report is the first to indicate that RV systolic pressure and RV dysfunction are significant independent predictors of mortality in children with pulmonary vein stenosis. A greater understanding of mortality in this population is necessary, particularly in those with RV systolic pressure less than half systemic.

Clinical PerspectiveWhat Is New? Elevated right ventricular systolic pressure and right ventricular dysfunction were studied as independent predictors of mortality in children with pulmonary vein stenosis.Mortality in pediatric pulmonary vein stenosis is multifactorial, often occurring within a complex clinical context beyond direct pulmonary vein disease. What Are the Clinical Implications? Elevated right ventricular systolic pressure and right ventricular dysfunction should prompt closer monitoring and tailored interventions in children with pulmonary vein stenosis.

Pulmonary vein stenosis (PVS) is a rare condition in the pediatric population, occurring in ≈0.03% of congenital heart diseases (CHDs). ^1^ , ^2^ However, there is a growing recognition that PVS in children and infants manifests differently from the indolent course observed in adults (usually after radiofrequency ablation). In the pediatric population, PVS can exhibit a relentless and progressive course, characterized by neointimal obstruction of the pulmonary veins, ultimately leading to disruption of normal cardiopulmonary interactions, such as secondary pulmonary hypertension (PH), right heart failure, and ventilation/perfusion mismatch.

Historically, PVS among children has generally been regarded as a devastating disease with a poor prognosis. The reported mortality rate ranges from ≈50% to 60%, according to previous studies. ^3^ , ^4^ , ^5^ , ^6^ Identified predictors for mortality include an earlier age at diagnosis, multiple and/or bilateral pulmonary vein involvement, restenosis after surgery, and disease progression. ^4^ , ^5^ , ^6^ , ^7^ , ^8^ , ^9^ Given the increasing number of premature birth survivors and the growing complexity of neonatal intensive care, it remains unclear whether the high mortality rate among children with PVS is solely explained by deaths secondary to right‐sided heart failure resulting from severe PVS and subsequent PH, or if the high mortality rate reflects the complexity of the neonatal clinical course, inclusive of PVS (ie, dying due to pulmonary vein stenosis versus dying with pulmonary vein stenosis).

Although there has been substantial progress, there is still a lack of clinical practice consistency across different institutions for the surveillance and treatment of recurrent disease. Consequently, accurate data on mortality rates, the incidence of catheter‐based and/or surgical interventions, and the occurrence of lung transplantation for this diagnosis are scarce.

Therefore, the primary objective of our study was to identify the incidence of mortality among children with PVS in our institution. The secondary objectives examined the causes of death among children with PVS and identified predictors of mortality.

METHODS

This retrospective medical record review study involving human participants was in accordance with the ethical standards of the institutional and national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The Human Investigation Committee (Institutional Review Board) of Cincinnati Children's Hospital Medical Center approved this study.

This study was a retrospective, single‐center study approved by the Institutional Review Board of Cincinnati Children's Hospital Medical Center. A waiver of informed consent was granted. The data that support the findings of this study are available from the corresponding author on reasonable request. The study included children with PVS who underwent cardiac catheterization at Cincinnati Children's Hospital Medical Center between 2015 and 2023. Patients with single‐ventricle physiology/anatomy were excluded from the study.

Eligible patients were identified through a query of electronic databases in the catheterization laboratory. At our institution, suspicion of PVS typically arises from ECG findings. Subsequently, all patients with suspected PVS undergo further imaging, including computed tomography scans and quantitative lung perfusion scans. As per an internally derived clinical protocol, cardiac catheterization was then performed to establish the definitive diagnosis of PVS (Figure 1). The number of pulmonary veins involved was determined by both computed tomography and angiography performed during the catheterizations. Comprehensive catheterization data, encompassing pulmonary vein angiography, gradient across pulmonary veins, the occurrence of interventions, right ventricular systolic pressure (RVSP), and left ventricular systolic pressure, were systematically collected. In cases where patients underwent interventional catheterization on pulmonary veins with either balloon dilation or stent placement, RVSP measured before each intervention was used for the analysis. The right ventricle/systemic systolic pressure ratio was calculated as RVSP divided by left ventricular systolic pressure. If left ventricular systolic pressure data were unavailable, either invasively or noninvasively measured aortic pressure was used for ratio computation. Echocardiographic data, including right ventricular (RV) systolic function, were extracted from clinical reports and reviewed independently for consistency. The ECG performed closest to the last catheterization was used for subsequent analysis.

Statistical Analysis

Statistical analysis was performed using SPSS, version 29 (IBM SPSS Inc, Chicago, IL). Continuous variables are reported as median and interquartile range (IQR), whereas categorical variables are denoted by number and percentage. The entire cohort was divided into 2 survivors and nonsurvivors. All prespecified demographic and catheterization parameters between the groups were analyzed using Mann‐Whitney U test and χ^2^ test, as appropriate. Kaplan‐Meier analysis with log‐rank test, along with univariate and multivariate Cox regression analysis, was performed to identify significant variables associated with mortality. Model selection for multivariate regression used backward selection, with an initial inclusion criterion based on a P value of <0.05 in univariate analysis. A threshold of P<0.05 was considered statistically significant.

RESULTS

Between 2015 and 2023, 56 children with PVS underwent cardiac catheterization at our catheterization laboratory, comprising 33 male children (59%). Table 1 provides a description of the demographics of the entire cohort and a comparison between survivors and nonsurvivors. The median gestational age at birth was 37 weeks (IQR, 29–38 weeks), with 26 patients (46%) born prematurely (<37 weeks). An underlying genetic syndrome, most commonly trisomy 21 and DiGeorge syndrome, was present in 17 patients (30%). Among 28 patients (50%), PVS was the primary cardiac diagnosis in the context of prematurity and/or bronchopulmonary dysplasia/chronic lung disease. Of these, 2 patients had congenital diaphragmatic hernia. In contrast, the remaining 28 patients (50%) had other significant CHDs, such as tetralogy of Fallot, balanced atrioventricular septal defect, or coarctation of the aorta, in addition to PVS. Of these 28 patients, 26 underwent surgical repair for their CHD or orthotopic heart transplantation, and 18 were diagnosed with PVS after their primary cardiac surgery. Among the 28 patients with CHD, 6 were born prematurely, although only 1 was born very preterm (29 weeks), whereas the others were born at 33 to 36 weeks. Primary PVS was observed in 40 patients (71%), whereas secondary (postsurgical) PVS was seen in 16 patients (29%), occurring after repair of total or partial anomalous pulmonary venous return.

Diagnostic or interventional catheterization occurred at least once in all patients. Catheterization demographics and findings are summarized in Table 2. The median age at first catheterization was 6 months (IQR, 4–12 months). Forty‐one patients (73%) underwent at least 1 balloon dilation in at least 1 pulmonary vein, and 33 patients (59%) underwent at least 1 stent placement in at least 1 pulmonary vein. At the last catheterization, the median RV/systemic systolic pressure ratio was 45% (IQR, 37%–64%). Among nonsurvivors, death occurred within a median of 25 days (IQR, 8–82 days) after the last catheterization. A comparison between survivors and nonsurvivors revealed a significantly higher RV/systemic systolic pressure ratio among nonsurvivors compared with survivors (64% versus 44%; P<0.001). The rate of catheterization, defined as the total number of catheterizations divided by the interval between the first and last catheterization (per year), was significantly higher among nonsurvivors compared with survivors (10 versus 2; P=0.002).

Pulmonary vasodilators were used in 15 patients (26%) with PVS. Of these, 8 patients were already taking sildenafil for PH associated with bronchopulmonary dysplasia, and this treatment was continued after the diagnosis of PVS. The remaining 7 patients were started on sildenafil following their PVS diagnosis. One patient was treated with sirolimus and imatinib, with these therapies initiated at another institution.

Mortality

During the median follow up period of 43 months (IQR, 6–70 months), 20 deaths occurred, with a median age at death of 10 months (IQR, 4–24 months). The causes of death were categorized as multiorgan failure in 7 patients (35%), progressive respiratory failure in 4 patients (20%), sudden cardiac death in 3 patients (15%), infection at outpatient settings (with limited information available) in 3 patients (15%), and withdrawal of support in 1 patient (5%). Detailed information on the cause of death was not available for 2 patients as they were transferred back to referring institutions. Demographics were not significantly different between survivors and nonsurvivors (Table 1).

Of the 18 patients with detailed information, 17 were mechanically ventilated at the time of death. Twelve patients were mechanically ventilated for >2 weeks before death, including 2 patients who had a tracheostomy and were ventilated at baseline. Among the 17 inpatient deaths, 3 patients had recent urinary tract infections before death. All 17 patients received empirical broad‐spectrum antibiotics because of clinical deterioration or worsening respiratory status, although blood or respiratory cultures were negative.

Causes of death in association with echocardiographic RV systolic function and catheterization‐driven RVSP are described in Table 3. Univariable Cox regression analysis revealed that RVSP greater than half systemic pressure at the last catheterization (hazard ratio [HR], 5.52 [95% CI, 2.16–14.12]; P<0.001), moderately or severely diminished RV systolic function (HR, 4.14 [95% CI, 1.37–12.52]; P=0.012), and the presence of CHD (HR, 2.88 [95% CI, 1.10–7.50]; P=0.031) were significant predictors of mortality (Table S1). In the multivariable analysis, the final model included RVSP greater than half systemic pressure at the last catheterization, moderately or severely diminished RV systolic function, and the presence of CHD as predictors. In this model, RVSP greater than half systemic pressure (HR, 4.03 [95% CI, 1.57–10.36]; P=0.004) and diminished RV systolic function (HR, 3.59 [95% CI, 1.12–11.51]; P=0.032) remained significant predictors, whereas the presence of CHD was not significant (HR, 2.44 [95% CI, 0.89–6.69]; P=0.084). The Kaplan‐Meier survival curves illustrating survival after the definitive diagnosis of PVS are presented in Figure 2A through 2D, where the entire cohort was divided into patients with RVSP less than half systemic pressure, between half and full systemic pressure, and more than systemic pressure, as well as accounting for normal versus diminished RV function.

DISCUSSION

This report indicates that both elevated RVSP and RV systolic dysfunction emerged as significant and independent predictors of mortality among pediatric patients with PVS. However, patients with PVS often died within a complex clinical context, not directly attributable solely to pulmonary vein disease. Therefore, it is crucial to acknowledge the multifactorial nature of mortality in this population.

Historically, PVS has been recognized as a rare but devastating cardiac disease in children. PVS may lead to PH, RV failure, and death. Because of the rarity of the disease, most previous publications are case reports, case series, or single‐center studies with a limited number of patients. The first systematic review in 2018 by Backes et al reported a mortality rate of 58.5% among 185 infants with PVS. ^3^ Although this included the largest number of patients with PVS at the time, the study included data before 2010 when protocolized, frequent, aggressive catheterization‐based intervention was not widely available, and the cause of death was not specified.

Multivessel involvement has also been noted as a risk factor for mortality in previous studies. ^3^ , ^10^ In their study, Backes et al reported poor survival in cases of multivessel and bilateral involvement compared with single PVS. Notably, a less aggressive approach to PVS was used during that era, with catheterization intervention performed in only 8.3% of the patients in that series.

Contrary to previous publications, ^4^ , ^5^ , ^6^ , ^7^ , ^8^ , ^9^ multivessel involvement was not a significant predictor of mortality in our study, suggesting the positive impact of more aggressive percutaneous pulmonary vein rehabilitation and surgical intervention. We would further postulate that the observed improvement in overall mortality compared with previous studies is reflective of advances in neonatal and cardiac intensive care, as well as the acceptance of more aggressive programs for pulmonary vein rehabilitation. Although not the focus of our study, the escalating use of drug‐eluting stents and adjuvant therapies may further contribute to improved survival rates.

The most notable findings in our review of mortality were the impact of elevated RVSP and RV dysfunction on survival. RVSP greater than half systolic systemic pressure (measured at catheterization) and decreased RV function (by echocardiography) were both significant independent predictors of mortality. The association between PVS and elevated RVSP, and impact on survival, was first reported by Sykes et al in 2018. ^11^ Although PVS can cause pulmonary venous hypertension and subsequent RV hypertension, the specific mechanisms need further investigation. Multivessel disease is conceptually deemed more pathologically impactful, and single‐vessel disease is generally better tolerated, but exceptions do occur.

Because our study identified elevated RVSP and decreased RV function as independent predictors of mortality, more attention should be paid to vasodilator testing and treatment in this subgroup, especially when isolated PVS appears to be present. Right ventricular hypertension (with resultant RV dysfunction) reflects either PVS progression or secondary pulmonary hypertension, both of which require some form of treatment (mechanical or medical). ^12^ Despite theoretical challenges posed by acute pulmonary edema concerns, reports suggest the safe utility of pulmonary vasodilator testing with NO in the presence of PVS. ^13^ Hence, we would suggest that any patient with PVS who meets individual institutional criteria for catheterization should undergo cautious pulmonary vasodilator challenge during that procedure, if elevated RV pressure and pulmonary resistance are present. A positive vasodilator response could certainly assist with additional pharmacologic interventions.

Nevertheless, the question of why elevated RVSP is a predictor of mortality even among patients with preserved RV systolic function remains unclear. It is well established that even mild PH may exacerbate underlying respiratory failure in the context of prematurity and chronic lung disease, and that PH in the presence of bronchopulmonary dysplasia is associated with higher mortality compared with the bronchopulmonary dysplasia without PH. ^14^ PVS may thus be negatively synergistic when overlaid on a predisposing substrate for RV hypertension in this subpopulation.

Further prospective study will be required to determine if a proactive programmatic approach to PVS intervention, with the express aim of decreasing RVSP to less than half systemic pressure, will impact mortality in this disease.

Limitations

Our study has several limitations, including the retrospective, single‐center design and small sample size. Additionally, the limited time period covered by our study means that newer techniques and adjuvant therapies, such as drug‐eluting stents and chemotherapeutic agents introduced during the study period, may have impacted outcomes. Their effect cannot be reliably assessed with the available data. Furthermore, throughout the study period, our institution's formal protocol for managing PVS gradually evolved. Initially, decisions concerning surgical versus catheter‐based interventions, the frequency of cardiac catheterizations, and the administration of pulmonary vasodilators were primarily guided by the discretion of the individual cardiologist overseeing each patient's care. Over time, these practices were formalized into the current protocol shown in Figure 1. The increasing recognition of the importance of frequent cardiac catheterizations may have influenced the frequency of procedures during the study period, although we did not specifically analyze an era effect. Although lung perfusion scan data were available for most patients, the timing of these scans did not consistently align with the cardiac catheterization procedure. Indexed pulmonary vascular resistance was not reported because relative pulmonary blood flow was not quantified during catheterization for many patients. Although lung perfusion scans can provide insights into the clinical impact of PVS, we opted not to present these results to avoid inaccuracies associated with this variability.

CONCLUSIONS

This report reviews the various causes of mortality in infants and children with PVS and highlights the impact of elevated RVSP and RV dysfunction as significant independent predictors of poor outcome. Further understanding of mortality in this population is necessary. Multicenter studies with larger patient numbers will be important for elucidating mortality in this condition.

Sources of Funding

None.

Disclosures

None.

Supporting information