Pediatric Cardiology

, Volume 35, Issue 5, pp 879–887

Factors Associated with Serum B-Type Natriuretic Peptide in Infants with Single Ventricles


    • Department of PediatricsMedical University of South Carolina
  • Victor Zak
    • New England Research Institutes
  • Daphne Hsu
    • Division of CardiologyChildren’s Hospital at Montefiore
  • James Cnota
    • Department of CardiologyCincinnati Children’s Hospital
  • Steven D. Colan
    • Department of CardiologyBoston Children’s Hospital
  • David Hehir
    • Department of PediatricsMedical College of Wisconsin
  • Paul Kantor
    • Department of CardiologyHospital for Sick Children
  • Jami C. Levine
    • Department of CardiologyBoston Children’s Hospital
  • Renee Margossian
    • Department of CardiologyBoston Children’s Hospital
  • Marc Richmond
    • Department of CardiologyChildren’s Hospital of New York
  • Anita Szwast
    • Department of CardiologyChildren’s Hospital of Philadelphia
  • Derek Williams
    • Department of CardiologyBrenner Children’s Hospital
  • Richard Williams
    • Department of PediatricsUniversity of Utah
  • Andrew M. Atz
    • Department of PediatricsMedical University of South Carolina
Original Article

DOI: 10.1007/s00246-014-0872-z

Cite this article as:
Butts, R.J., Zak, V., Hsu, D. et al. Pediatr Cardiol (2014) 35: 879. doi:10.1007/s00246-014-0872-z


Data regarding the value of B-type natriuretic peptide (BNP) measurements in infants with a single-ventricle (SV) physiology are lacking. This analysis aimed to describe the BNP level changes in infants with an SV physiology before and after superior cavopulmonary connection (SCPC) surgery. Levels of BNP were measured by a core laboratory before SCPC (at 5.0 ± 1.6 months) and at the age of 14 months during a multicenter trial of angiotensin-converting enzyme inhibition therapy for infants with SV. Multivariable longitudinal analysis was used to model the associations between BNP levels and three sets of grouped variables (echocardiography, catheterization, growth). Multivariable analysis was performed to assess associations with patient characteristics at both visits. Associations between BNP levels and neurodevelopmental variables were investigated at the 14 month visit because neurodevelopmental assessment was performed only at this visit. The BNP level was significantly higher before SCPC (n = 173) than at the age of 14 months (n = 134). The respective median levels were 80.8 pg/ml (interquartile range [IQR], 35–187 pg/ml) and 34.5 pg/ml (IQR, 17–67 pg/ml) (p < 0.01). A BNP level higher than 100 pg/ml was present in 72 subjects (42 %) before SCPC and in 21 subjects (16 %) at the age of 14 months. In the 117 patients who had BNP measurements at both visits, the median BNP level decreased 32 pg/ml (IQR, 1–79 pg/ml) (p < 0.01). In the longitudinal multivariable analysis, higher BNP levels were associated with a higher end-systolic volume z-score (p = 0.01), a greater degree of atrioventricular (AV) valve regurgitation (p < 0.01), a lower weight z-score (p < 0.01), and a lower length z-score (p = 0.02). In multivariable analyses, a higher BNP level at the age of 14 months was associated with arrhythmia after SCPC surgery (p < 0.01), a prior Norwood procedure (p < 0.01), a longer hospital stay after SCPC surgery (p = 0.04), and a lower Bayley psychomotor developmental index (p = 0.02). The levels of BNP decreases in infants with SV from the pre-SCPC visit to the age of 14 months. A higher BNP level is associated with increased ventricular dilation in systole, increased AV valve regurgitation, impaired growth, and poorer neurodevelopmental outcomes. Therefore, BNP level may be a useful seromarker for identifying infants with SV at risk for worse outcomes.


B-type natriuretic peptideBNPSingle-ventricle physiologySuperior cavopulmonary connection surgeryAngiotensin-converting enzyme inhibition therapySCPCVentricular dilation


Ventricular cardiac myocytes produce B-type natriuretic peptide (BNP) as a response to stretch [1]. Clinicians have used BNP to help differentiate between lung disease and cardiac disease [2, 3], to help predict successful weaning from mechanical circulatory support [4], and to help predict successful medical treatment of a patent ductus arteriosus in neonates [5]. Elevated serum BNP has been associated with poor outcomes in adults and children with heart failure [6, 7].

Despite a growing knowledge concerning the utility of BNP in the management of heart disease, little is known about the changes in BNP that occur in patients with a single-ventricle (SV) physiology undergoing staged palliation. Findings have shown that BNP helps to differentiate between ventricular dysfunction and cavopulmonary connection failure in patients with SV disease after superior cavopulmonary connection (SCPC) surgery or Fontan procedure [8].

Few data are available concerning the BNP levels in patients who are between their first-stage palliation and SCPC surgery. During this interval, patients are at the highest risk for both mortality and significant morbidity [911]. Prior studies investigating BNP between stage 1 and SCPC surgery have been limited to single-center, cross-sectional studies involving small numbers of patients [12, 13].

This study aimed to describe BNP trends before and after SCPC in a large cohort of infants with SV physiology. An additional objective was to search for associations between BNP and other variables known to be important in this patient population including patient characteristics, echocardiographic and catheterization variables, growth parameters, and neurodevelopmental outcomes.


This study was a secondary analysis of the National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI)-sponsored trial of angiotensin-converting enzyme (ACE) inhibition in infants with SV conducted by the Pediatric Heart Network. The study design and main results of the trial have been previously published [14, 15].

The study enrolled infants 7–45 days of age with SV physiology and stable pulmonary and systemic blood flow who were predicted to undergo a SCPC surgery (bidirectional Glenn, bilateral bidirectional Glenn, or hemi-Fontan procedure). The exclusion criteria ruled out prematurity (gestation age <35 weeks), small size for gestation age (weight <10th percentile for gestation age), systemic oxygen saturation lower than 65 %, creatinine level higher than 1.0 mg/dl, an absolute neutrophil count below 1,000 cells/ml, prior use of ACE inhibitor, and any other clinical situation that prevented the use of an ACE inhibitor.

Between August, 2003 and May, 2007, 10 centers throughout North America enrolled patients. The institutional review board or ethics board of each participating center approved the study protocol, and written informed consent was obtained from a parent or guardian.

Data Collection

Plasma BNP levels were obtained at the pre-SCPC visit and during the final study visit at the age of 14 months. The sample for BNP was drawn after the subject had been in a sitting/supine position in a quiet room. For the sample, 1 ml of whole blood was collected and placed into a pre-chilled lavender-topped ethylenediaminetetraacetic acid tube [16]. The resulting plasma was dispensed into vials for storage and frozen immediately at –20° to –80° C. The samples then were sent to the serology core laboratory (Mayo Central Laboratory for Clinical Trials) on dry ice.

Levels of BNP were determined using the Shionogi method [17]. Detailed anatomic diagnosis, gestation age, and birth weight were recorded. Height, weight, and head circumference were measured at seven different time points during the study (in this study, the pre-SCPC and 14 month measurements were used for analysis) and adjusted for age using the WHO criteria.

Echocardiograms were performed at the pre-SCPC and 14 month visits, then analyzed in a core laboratory by a single reader. Measurements of ventricular volume, mass, end-systolic volume, end-diastolic volume, and ejection fraction were calculated using a modified Simpson’s algorithm [18]. If cardiac catheterization was performed before SCPC surgery, the data also were included.

Neurodevelopmental testing with the Bayley scale of infant development (BSID)-II was performed at the 14 month visit. Analysis was performed for the following sections of the BSID-II: mental developmental index and psychomotor developmental index.

Data Analysis

The BNP values were log transformed (with natural base, logBNP) to attain an approximately normal distribution. All the patients with a BNP measurement were included in the analysis irrespective of type of surgery performed between the two time points. The logBNP was compared with other continuous variables using the Pearson correlation test. The mean logBNP was compared between categorical variables using Student’s t-test.

Longitudinal multivariable analysis was used for the variables collected at both time points (echocardiographic, catheterization, and growth measures). Multivariable analysis was performed for variables not assessed at multiple time points (patient characteristics, neurodevelopmental testing). Multivariable analysis was performed separately for each category of variables (echocardiographic, catheterization, growth, neurodevelopmental measures, and patient characteristics). All models used logBNP as the outcome.

For the longitudinal analysis of BNP, only the data of patients with BNP measured at both time points who also had an SCPC surgery were included. Longitudinal multivariable modeling was performed for each set of grouped variables collected at both time points (echocardiographic, catheterization, and growth measures) to identify associations with logBNP and included all available data.

When applicable, the parameter estimate (PE) of longitudinal analysis is reported. Multivariable analyses of both neurodevelopmental measures and patient characteristics (Table 1) were performed to assess associations with logBNP. The multivariable analysis of neurodevelopmental measures had only logBNP as the outcome at the 14 month visit because neurodevelopmental testing was performed only at that time. Multivariable modeling of logBNP at the pre-SCPC visit included only the patient characteristics present before the visit. Multivariable modeling of logBNP at the 14 month visit included all relevant patient characteristics regardless of time. Because some BNP measurements were missing, analysis was performed to determine potential bias in the patient sample.
Table 1

Patient characteristics evaluated

Patient factors

Birth weight

Gestational age




Anatomic diagnosis (HLHS: yes/no)

Dominant ventricular type (left/right/mixed)

Age at BNP test

Medical factors during neonatal hospitalization

Age at palliative surgery

Length of Hospital Stay

No. of discharge medications

Discharge oxygen saturation

No. of other cardiac surgical procedures

Type of palliative surgery (Norwood vs non-Norwood)

Cardiopulmonary bypass time

Aortic cross-clamp time

Use of deep hypothermic circulatory arrest

Medical factors during SCPC hospitalization

Age at SCPC Surgery

Length of hospital stay

No. of discharge medications

Discharge oxygen saturation

No. of other cardiac surgical procedures

Presence of chylous drainage

Presence of postoperative arrhythmia

HLHS hypoplastic left heart syndrome, BNP brain natriuretic peptide, SCPC superior cavopulmonary connection

The characteristics of the patients with and without BNP measurements at each time point were compared using Student’s t-test, Fisher’s exact test, Wilcoxon rank sum test, and Mantel–Haenszel test for linear trend, as appropriate. All analyses were performed using SAS statistical software version 9.3 (SAS Institute, Inc., Cary, NC, USA).



Among the 230 subjects enrolled in the trial, 200 had a pre-SCPC visit at 5.0 ± 1.6 months, whereas 28 patients withdrew before the pre-SCPC visit, and 2 patients did not have a pre-SCPC visit but remained in the trial. Of the 202 patients who continued in the study after the pre-SCPC visit, 185 had a 14 month visit, and 175 of these 185 patients had a SCPC procedure. The remaining 10 subjects had procedures physiologically different from a SCPC (5 had Kawashima; 3 had a systemic to pulmonary shunt; and 2 had no procedure). A total of 117 subjects underwent an SCPC surgery and had their BNP measured at both the pre-SCPC and 14 month visits.

Hypoplastic left heart syndrome (HLHS) was the most common diagnosis of the cohort, representing 107 (61.8 %) of the patients with pre-SCPC BNP data (n = 173) and 80 (59.7 %) of the patients with 14 month BNP data (n = 134). The dominant ventricle in the patients with pre-SCPC BNP data was the right ventricle (RV) in 120 patients, the left ventricle (LV) in 35 patients, and a mixed condition in 18 patients. In the patients with 14 month BNP data, the dominant ventricle was the RV in 94 patients, the LV in 28 patients, and a mixed condition in 12 patients.

To explore potential participation bias, of the 200 patients who had a pre-SCPC visit, the characteristics of the 173 patients with BNP data were compared with the characteristics of the 27 patients without BNP data. The two groups did not differ except that the patients with BNP measurements were more likely to have moderate or severe atrioventricular (AV) valve regurgitation than mild regurgitation or none (p = 0.04, Fisher’s exact test).

At the 14 month visit, the only significant difference between the 134 patients with BNP measurements and the 51 patients without BNP data was race. The subjects with a BNP measurement were more likely to be African–American and less likely to be in the “other” category (the proportion of whites was similar in the two groups, and represented 80 % of the total sample; p = 0.04).

BNP Measurements

The BNP level was higher in patients at the pre-SCPC visit (median, 80.8 pg/ml; interquartile range [IQR], 35–187 pg/ml; n = 173) than at the 14 month visit (median, 34.5 pg/ml; IQR, 17–67 pg/ml; n = 134) (p < 0.01)) (Fig. 1). The median drop in BNP for the 117 patients who underwent SCPC surgery and had a BNP measurement at both visits was 32 pg/ml (IQR, 1–79 pg/ml; p < 0.01). In the pre-SCPC sample (n = 173), a weak correlation between greater age and lower logBNP was observed (r = −0.16; p = 0.01). However, no association was observed between age and logBNP after SCPC surgery (r = 0.03; p = 0.9; n = 134).
Fig. 1

a The log-transformed brain natriuretic peptide (logBNP) level was significantly higher at the pre-superior cavopulmonary connection (Pre-SCPC) visit (at 5.0 ± 1.6 months) than at the 14 month visit (4.4 ± 1.3 vs 3.5 ± 1.7; p < 0.01). The raw median values of BNP were 80.8 pg/ml (IQR, 35–187 pg/ml; n = 173) at the pre-SCPC time point and 34.5 pg/ml (IQR, 17–67 pg/ml; n = 134) (p < 0.01) and 14 month time point. b The raw BNP values are presented (6 Pre-SCPC outlying values between 1,100 and 4,100 pg/ml are excluded to improve the resolution)

In the multivariable longitudinal analysis that focused on age and gender, no associations were observed between age or race and logBNP. A BNP level higher than 100 pg/ml was present in 72 (42 %) of the subjects before SCPC surgery and in 21 (16 %) of the subjects at 14 months.

Associations Between BNP and Echocardiography

In the univariate analysis, higher logBNP at both time points was associated with increased end-systolic and end-diastolic volume z-scores, ventricular mass, lower ejection fraction z-score, and moderate or severe AV valve regurgitation versus mild regurgitation or none. The multivariable longitudinal analysis of logBNP (290 observations) that examined the echocardiographic measurements as correlates found higher logBNP and greater end-systolic volume z-score associated with the degree of AV valve regurgitation (Table 2).
Table 2

Echocardiogram measures and log-transformed brain natriuretic peptide (logBNP)

Echocardiogram measure

Pre-SCPC visit (r) (p value)

14 month visit (r) (p value)

Multivariable longitudinal model (PE)

End-diastolic vol z-score






End-systolic vol z-score






Ejection fraction z-score






Mass z-score






AV valve regurgitationa


None/mild: pg/ml (range)

77 (30–180)


29 (13–64)



Moderate/severe: pg/ml (range)

107 (60–270)

52 (32–127)

SCPC superior cavopulmonary anastomosis, PE parameter estimate, vol, volume, AV atrioventricular

aDisplayed values for AV valve regurgitation are median (interquartile range)

Associations Between BNP and Catheterization

At the pre-SCPC visit, the univariate analysis of catheterization measures found correlations between higher pre-SCPC logBNP and higher end-diastolic pressure (PE 0.1; p < 0.01) as well as lower superior vena cava saturation (PE −0.3; p = 0.01). In the multivariable analysis of logBNP that examined catheterization variables as correlates, the only association was a positive correlation between mean pulmonary artery pressure and logBNP levels at the pre-SCPC visit (R2 = 0.04; β = 0.07 ± 0.02; p = 0.01). LogBNP measured at the 14 month visit had no statistically significant associations with pre-SCPC catheterization measures.

Associations Between BNP and Patient Characteristics

In the univariate analysis, higher logBNP before SCPC surgery was associated with a diagnosis of HLHS, stage 1 Norwood procedure, and RV morphology. Higher logBNP at the 14 month visit was associated with the diagnosis of HLHS, stage 1 Norwood procedure, RV morphology, post-SCPC arrhythmia, lower birth weight, and longer post-SCPC hospital stay in the univariate analysis (Table 3).
Table 3

Significant univariate associations between patient variables and log-transformed brain natriuretic peptide (logBNP)

Patient variable


p value


Hypoplastic left heart







Norwood procedure







Ventricle type










14 month BNP

Postoperative SCPC arrhythmia







Prior Norwood







Ventricle type










Hospital stay after SCPC (days)


PE = 0.2; p = 0.02

SCPC superior cavopulmonary connection

In the multivariable model of logBNP at the pre-SCPC visit, only the diagnosis of HLHS was associated with higher logBNP (R2 = 0.08; β = 0.75 ± 0.2; p < 0.01). At the 14 month visit, the presence of post-SCPC arrhythmia (β = 0.92 ± 0.33; p < 0.01), log-transformed hospital length of stay after SCPC surgery (β = 0.66 ± 0.22; p < 0.01), and stage 1 Norwood procedure (β = 0.26 ± 0.12; p = 0.04) were associated with higher logBNP in the multivariable model of logBNP (R2 = 0.16; p < 0.01). All the associations remained statistically significant when degree of AV regurgitation was introduced into the model.

Associations Between BNP and Growth

In the univariate analysis at both visits, LogBNP had a negative correlation with weight and length z-scores (Table 4). The head circumference z-score had a negative correlation with logBNP at the pre-SCPC visit. However, this association did not persist at the 14 month visit. The weight-for-length z-score was negatively correlated with logBNP at the 14 month visit only.
Table 4

Growth measures and log-transformed brain natriuretic peptide (LogBNP)

Growth measure

Pre-SCPC visit (r) (p Value)

14 month visit (r) (p value)

Multivariable longitudinal model (PE) (p value)

Weight for age z-score







Length for age z-score






Head circumference for age z-score






Weight for length z-score







SCPC superior cavopulmonary anastomosis

In the longitudinal multivariable analysis of logBNP (306 observations) that examined growth variables as correlates, lower logBNP was associated with a higher weight z-score, a higher length z-score, and the BNP sample obtained at the 14 month visit versus the pre-SCPC visit (Table 4).

Associations Between BNP and Neurodevelopment

Of the 170 patients who had neurodevelopmental testing at the 14 month visit, 149 had a sample for BNP drawn at the pre-SCPC visit, and 126 had a sample for BNP drawn at the 14 month visit. LogBNP at the pre-SCPC visit correlated negatively with the mental developmental index z-score (r = −0.21; p = 0.01I) and the psychomotor developmental index (r = −0.22; p = 0.01). LogBNP at the 14 month visit correlated negatively with the psychomotor developmental index (r = −0.22; p = 0.01) but not with the mental developmental index (r = −0.12; p = 0.20). In the multivariable analysis of LogBNP, only the psychomotor developmental index was weakly associated with logBNP measured at the 14 month visit (R2 = 0.05; p = 0.04).


This study was unique in that the series analyzed represented the largest group of patients to date with SV heart disease who had a systematic collection of BNP during the first 14 months of life. The description of BNP values and variation in this population is a valuable reference for the medical professionals caring for children with SV. The large study population allowed for a robust evaluation of the association between BNP and many variables including echocardiographic, catheterization, neurodevelopmental, and growth factors. It also allowed for longitudinal analysis at two distinct time points that represented two different physiologic conditions. The analysis showed that BNP declines after SCPC surgery and that elevated BNP levels are associated with echocardiographic measures of worse ventricular function as well as poorer growth and developmental outcomes in infants with SV physiology.

We found that BNP values decreased significantly over time in this multicenter cohort of SV subjects to levels closer those of normal pediatric patients. However, the median BNP level at the age of 14 months in this study remained slightly higher (4–10 pg/ml) than the previously published 95th percentile of plasma BNP in pediatric control subjects [19].

The Pediatric Heart Network (PHN) reported previously on the BNP values for 510 pediatric patients with a Fontan circulation [20]. In that study, involving subjects between the ages of 6 and 18 years, the BNP values were generally low (median, 13.0 pg/ml; IQR, 7.1–25.9 pg/ml) and associated with both age and gender [20].

Our study showed no association of BNP values before SCPC surgery or at 14 months with gender. Previous reports of healthy children have shown that gender differences in BNP are not present until after the age of 10 years [19]. A weak correlation with age was observed, but this was only for the pre-SCPC samples.

The results of our analyses combined with the results from the PHN Fontan study demonstrate that serum BNP generally declines with each stage of surgical SV palliation. Therefore, when the serum BNP level in SV heart disease is interpreted, stage of palliation is important.

The decrease in BNP levels after SCPC surgery seen in the current study confirms similar results reported in smaller studies [2125]. This decrease in BNP after SCPC surgery supports previous findings that demonstrate a physiologic benefit of SCPC surgery over stage 1 palliation for SV physiology [26, 27].

In biventricular congenital heart disease, BNP levels increase in conjunction with ventricular dilation [2830]. The decline in BNP after SCPC surgery seen in this study likely resulted from the volume unloading of the SV. A decrease in ventricular volume results in a decrease in atrial and ventricular stretch and leads to decreased release of BNP. The association of higher BNP levels with moderate or severe systemic AV regurgitation at both the pre-SCPC and the 14 month time points in this study also supports this theory and is consistent with similar findings reported for Fontan patients [23].

The degree and size of ventricular hypertrophy have been associated with BNP in patients with and without congenital heart disease [3133]. In our study, which included a heterogeneous group of SVs, showed results similar to those of previous reports. Ventricular mass, end-systolic volume, and diastolic volume all were related to BNP in the univariate analysis, with end-systolic volume a predictor of BNP in the multivariate analysis. Therefore, an SV with an anatomy that leads to elevated ventricular mass or volume likely will have higher BNP levels.

Despite similar exercise capacity and echocardiographic measures of contractility, RV morphology in SV patients has been associated with worse outcomes [27, 34, 35]. Prior studies have reported higher BNP levels in patients with RV morphology than in patients with LV morphology after SCPC and total cavopulmonary connection surgery [24]. Our study found that RV morphology is associated with a higher BNP before but not after the SCPC surgery.

The association between RV dominance and higher BNP before SCPC surgery persisted after control was used for degree of AV valve regurgitation. This indicates that the observed BNP elevation in RV-dominant SV patients is not fully explained by worse AV valve function alone.

Weak correlations were observed between BNP and poorer growth and neurodevelopmental outcomes such as height z-score, weight z-score, and the Bayley psychomotor developmental index. Other patient characteristics associated with higher BNP at 14 months were arrhythmia and length of hospital stay after SCPC surgery. The BNP level showed only weak associations with catheterization values in this population.

Studies comparing catheterization measurements and BNP in pediatric heart disease are lacking. One study showed a correlation between BNP and right heart catheterization measurements in pediatric transplant recipients [36]. However, another study of 59 pediatric patients with atrial septal defect or ventricular septal defect showed no association between Qp:Qs and BNP [37].

Study Limitations

Only 173 of the 230 patients enrolled in this study had BNP measurement at the pre-SCPC visit, and only 134 had BNP measurement at the 14 month visit. The 173 patients with BNP measurement at the pre-SCPC visit had a higher degree of AV valve regurgitation than the patients with missing BNP data. Therefore, it is possible that if these 27 patients had undergone BNP measurement, the median BNP measured before SCPC surgery would have been lower. The results reported here reflect BNP values measured at only two time points (at 5 months and at 14 months) and at times when the subjects generally were doing well clinically.


After SCPC surgery, BNP has a predictable decline. Elevations in BNP levels are negatively correlated with important echocardiographic measures of ventricular function as well as growth and developmental outcomes in infants with SV physiology. Therefore, BNP may be a useful seromarker in helping to determine overall cardiac health in this population and may aid in identifying infants with SV at high risk for poor outcomes.


This study was supported by U01 grants from the National Heart, Lung, and Blood Institute (HL068269, HL068270, HL068279, HL068281, HL068285, HL068292, HL068290, HL068288, HL085057) and from the FDA Office of Orphan Products Development. It also was supported by a T32 grant from the National Heart, Lung and Blood Institute (5T32 HL007710). The contents of this report are solely the responsibility of the authors and do not necessarily represent the official views of NHLBI or NIH.

Copyright information

© Springer Science+Business Media New York 2014