Secondary Logo

Journal Logo

Pediatric Circulatory Support

Adverse Outcomes in Neonates and Children with Pulmonary Artery Hypertension Supported with ECMO

Nasr, Viviane G.*; Faraoni, David*; DiNardo, James A.*; Thiagarajan, Ravi R.

Author Information
doi: 10.1097/MAT.0000000000000419


Pulmonary artery hypertension (PAH) is a rare chronic and severe condition associated with significant mortality and morbidity.1,2 Multiple primary (idiopathic, familial) and secondary (such as congenital heart disease and connective tissue disease) etiologies exist.3 The heterogeneity and complexity of the underlying disease make PAH difficult to treat. Medical therapy with endothelin-receptor antagonists, phosphodiesterase type 5 inhibitors, prostacyclin, and nitric oxide is largely off-label. Extracorporeal membrane oxygenation (ECMO) has been increasingly used to rescue neonates and children with life-threatening cardiac or respiratory failure, and other critical illnesses including PAH that are unresponsive to conventional medical therapies. To date, no study has examined the efficacy of ECMO in neonates and children diagnosed with PAH.

In this study, we sought to assess mortality and outcomes in neonates and children with PAH supported with ECMO and included in the Health Care Cost and Use Project (HCUP) Kid’s Inpatient Databases (KID).


We performed a retrospective analysis of the 2009 and 2012 HCUP KID, an administrative database derived from participating hospitals in 44 US States. It includes all patients ≤20 years of age and clinical and resource utilization information. The KID dataset includes >100 clinical and nonclinical variables for each hospital stay, available from discharge summaries. It reports a maximum of 25 diagnoses and 15 procedures for each hospitalization. Systematic random sampling is used to generate the dataset, including 80% of all pediatrics and adolescent hospital discharges, 80% of complicated in-hospital births, and 10% of uncomplicated in-hospital births from each participating state. Specialty hospitals, public hospitals, and academic medical centers are included in the database. The AHA’s (American Hospital Association, Annual Survey of Hospitals determined the hospital characteristics, location, teaching status, and size. Health Care Cost and Use Project assigns hospital bed-size categories as small (1–99 beds), medium (100–399 beds), and large (≥400 beds).

Neonates and children with PAH were identified using International Classification of Diseases, Ninth Revision, Clinical Modification codes (ICD-9-CM) diagnostic codes (416, 416.0). Children with congenital heart disease were excluded from the analysis by using a combination of ICD-9-CM procedure and diagnostic codes. Demographic and hospital characteristics, length of hospital stay, and in-hospital mortality recorded in the database were used for the purpose of these analyses. Patient’s age at admission was categorized into five age groups: neonates (<1 month of age), 1–12 months old, 1–6 years old, 6–12 years old, and ≥12 years old. Neonates and children treated with ECMO were identified using ICD-9-CM procedure codes (39.65). Thrombotic complications defined as intravascular thrombosis (venous, arterial, or intracardiac including shunts), cardioembolic strokes, or pulmonary embolism were identified using ICD-9-CM diagnostic codes as were the outcomes of acute kidney injury (AKI), dialysis, sepsis, and neurologic complications. For each eligible patient, we used a modification of the Elixhauser comorbidity score as a measure of comorbidities to generate a comorbidity score.4,5 All ICD-9-CM codes used in this study are available in the electronic supplementary material (see Table, Supplemental Digital Content, at

Statistical Analysis

Neonates and children with PAH were allocated to a group “ECMO” and a group “No ECMO.” Categorical variables are presented as number and percentage (%), and continuous variables are reported as median and interquartile range. Groups were compared using the χ2 test for categorical variables and the Wilcoxon rank-sum test for continuous variables.

We defined an a priori univariate cutoff value of p < 0.10 for variable inclusion in a propensity-matched analysis which ultimately included age group, elective admission, and the Elixhauser comorbidity score. We used a logistic regression model predicting ECMO support on the basis of the three covariates (age group, elective admission, and the Elixhauser comorbidity score) to determine the propensity score using the nearest neighbor approach and to match PAH patients supported with ECMO to PAH patients not supported with ECMO (Figure 1). Univariate logistic regression was applied to assess the relationship between the use of ECMO and outcomes using the matched cohorts.

Figure 1.
Figure 1.:
Flow chart.

A p value <0.05 was considered statistically significant for all tests. All reported values in this study are absolute values measured from the dataset. Statistical analyses were performed using STATA version 14.0 for Mac OS (StataCorp, College Station, TX).


From the 6,386,928 children included in the 2009 and 2012 KID HCUP databases, we identified 9,355 neonates and children (0.15%) with PAH. The incidence of ECMO utilization in this population was 1.4% (132/9,355).

Table 1 describes univariate comparisons between children in the ECMO and No ECMO groups. Children supported with ECMO were younger (p < 0.001), were less likely to have an elective hospital admission (p = 0.037), and had a higher Elixhauser comorbidity score (12 [7–20] vs. 7 [7–12], p < 0.001). Before propensity-matched analysis (Table 2), neonates and children supported with ECMO had higher incidences of dialysis (odds ratio [OR]: 6.06, 95% confidence interval [CI]: 3.47–10.57, p < 0.001), AKI (OR: 5.38, 95% CI: 3.67–7.88, p < 0.001), neurologic complications (OR: 5.54, 95% CI: 2.84–8.87, p < 0.001), sepsis (OR: 5.21, 95% CI: 3.59–7.57, p < 0.001), and thrombotic complications (OR: 4.59, 95% CI: 2.68–7.85, p < 0.001). The incidence of mortality was 39% (51/132) in the patients supported with ECMO and 3% (313/9,223) in the control group (OR: 17.92, 95% CI: 12.40–25.89, p < 0.001).

Table 1.
Table 1.:
Demographics Before and After Matching
Table 2.
Table 2.:
Outcomes of Patients with PAH Before Matching

After propensity-matched analysis, 130 neonates and children were included in each study group (Table 3). No difference was observed in univariate characteristics between the matched cohorts (Table 1). Neonates and children supported on ECMO had higher incidences of AKI (OR: 2.41, 95% CI: 1.30–4.47, p = 0.005), neurologic complications (OR: 7.11, 95% CI: 1.57–32.18, p = 0.011), sepsis (OR: 2.69, 95% CI: 1.46–4.96, p = 0.002), and thrombotic complications (OR: 2.90, 95% CI: 1.10–7.67, p = 0.032). The incidence of mortality was 39% (51/130) in the ECMO group and 8% (11/130) in the No ECMO group (OR: 6.98, 95% CI: 3.43–14.21, p < 0.001).

Table 3.
Table 3.:
Outcomes for Patients with PAH After Matching


This retrospective study evaluated the incidence of complications and mortality in neonates and children with PAH supported with ECMO. In this cohort of patients with identical Elixhauser comorbidity score and propensity matching, we observed a significant association between the use of ECMO in patients with PAH and increased risk of both complications and mortality.

The incidence of mortality in patients with PAH has been assessed in the Registry to Evaluate Early and Long-Term PAH Disease Management (REVEAL), the UK Pulmonary Hypertension Service for Children, and the Netherlands PH registry.6–8 Even though a trend toward a reduction in mortality rate has been observed, the overall mortality remains relatively high. In a retrospective review of the 2000, 2003, 2006, and 2009 Kids’ Inpatient Database, the overall mortality rate in PAH patients aged <20 years was high with a 15-fold higher mortality rate in PAH-related hospitalizations compared with hospitalizations unrelated to PAH.9

To date, there is a paucity of studies evaluating outcomes in children with PAH supported with ECMO. After propensity-matched analysis, our study revealed a hospital mortality rate of 39% for patients with PAH supported with ECMO compared with 8% in PAH patients without ECMO treatment. Smaller studies have shown 100% survival on ECMO in children with PAH associated with congenital heart disease or persistent neonatal PAH.10,11 Outcomes of patients with pulmonary hypertension (PH) on ECMO vary on the basis of the etiology of the disease. ECMO may be beneficial for specific subsets of patients with PH such as persistent PH of the newborn, and patients with congenital heart disease with higher survival rates compared with other conditions such as prelung transplant, where mortality is 60%.10–13

In addition, we observed an increased incidence of complications including dialysis, AKI, neurologic complication, and thrombotic complication in PAH patients treated with ECMO compared with PAH patients not treated with ECMO but with similar Elixhauser comorbidity scores. The Elixhauser comorbidity classification based on ICD-9 coding has been shown to be significantly associated with hospital outcomes and discriminative for mortality.4,14,15

This study has several limitations. It is a retrospective study, and the analyses were performed using a large administrative dataset. Consequently, miscoding of diagnosis and procedures may occur. Overlapping codes or missing data may limit the validity of the analyses in large datasets. We elected to include all codes for PAH in this study. However, with this database, it is not possible to determine what test modality (echocardiogram, cardiac catheterization, magnetic resonance imaging) was used to establish the PAH diagnosis or to determine who established the diagnosis. In addition, the results of laboratory tests and imaging studies that can be used to determine the severity of PAH after diagnosis were not available to us. Determination of the timing of initiation of ECMO and type of ECMO (veno-venous or veno-arterial) is not available in the KID database.

In conclusion, we found that neonates and children with PAH supported with ECMO have a higher mortality rate and a higher incidence of complications as compared with PAH patients with comparable comorbidities not supported with ECMO. Future studies are needed to determine the indications and timing for ECMO initiation in patients with PH.


1. Takatsuki S, Ivy DDCurrent challenges in pediatric pulmonary hypertension.Semin Respir Crit Care Med201334627644
2. Ivy DD, Abman SH, Barst RJ, et al.Pediatric pulmonary hypertension.J Am Coll Cardiol20136225 supplD117D126
3. Abman SH, Hansmann G, Archer SL, et al.American Heart Association Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation; Council on Clinical Cardiology; Council on Cardiovascular Disease in the Young; Council on Cardiovascular Radiology and Intervention; Council on Cardiovascular Surgery and Anesthesia; and the American Thoracic SocietyPediatric pulmonary hypertension: Guidelines from the American Heart Association and American Thoracic Society.Circulation201513220372099
4. van Walraven C, Austin PC, Jennings A, Quan H, Forster AJA modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data.Med Care200947626633
5. Quan H, Sundararajan V, Halfon P, et al.Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.Med Care20054311301139
6. Barst RJ, McGoon MD, Elliott CG, Foreman AJ, Miller DP, Ivy DDSurvival in childhood pulmonary arterial hypertension: Insights from the registry to evaluate early and long-term pulmonary arterial hypertension disease management.Circulation2012125113122
7. Haworth SG, Hislop AATreatment and survival in children with pulmonary arterial hypertension: The UK Pulmonary Hypertension Service for Children 2001–2006.Heart200995312317
8. van Loon RL, Roofthooft MT, Hillege HL, et al.Pediatric pulmonary hypertension in the Netherlands: Epidemiology and characterization during the period 1991 to 2005.Circulation201112417551764
9. Frank DB, Crystal MA, Morales DL, et al.Trends in pediatric pulmonary hypertension-related hospitalizations in the United States from 2000-2009.Pulm Circ20155339348
10. Dhillon R, Pearson GA, Firmin RK, Chan KC, Leanage RExtracorporeal membrane oxygenation and the treatment of critical pulmonary hypertension in congenital heart disease.Eur J Cardiothorac Surg19959553556
11. Karimova A, Brown K, Ridout D, et al.Neonatal extracorporeal membrane oxygenation: Practice patterns and predictors of outcome in the UK.Arch Dis Child Fetal Neonatal Ed200994F129F132
12. Puri V, Epstein D, Raithel SC, et al.Extracorporeal membrane oxygenation in pediatric lung transplantation.J Thorac Cardiovasc Surg2010140427432
13. Gazit AZ, Sweet SC, Grady RM, Huddleston CBFirst experience with a paracorporeal artificial lung in a small child with pulmonary hypertension.J Thorac Cardiovasc Surg2011141e48e50
14. Elixhauser A, Steiner C, Harris DR, Coffey RMComorbidity measures for use with administrative data.Med Care199836827
15. Poses RM, McClish DK, Smith WR, Bekes C, Scott WEPrediction of survival of critically ill patients by admission comorbidity.J Clin Epidemiol199649743747

outcomes; extracorporeal membrane oxygenation; pulmonary artery hypertension; children; neonates

Supplemental Digital Content

Copyright © 2016 by the ASAIO