OUP user menu

Usefulness of N-Terminal-Pro-B-Type Natriuretic Peptide as a Screening Tool for Identifying Pediatric Patients With Congenital Heart Disease

Emmanuel Jairaj Moses BSc(Hons), MSc, Sharifah A.I. Mokhtar MBBCh (Hons), MMed, Amir Hamzah MD, MMed, Basir Selvam Abdullah BSc, Narazah Mohd Yusoff MBBS, MMed, PhD
DOI: http://dx.doi.org/10.1309/LMW0U87COTHXGELF 75-80 First published online: 1 February 2011


Background: The aim of this study was to evaluate the usefulness of N-terminal-pro-B-type natriuretic peptide (NT-proBNP) as a screening tool in the diagnosis of congenital heart disease (CHD) among pediatric patients.

Methods: N-terminal-pro-B-type natriuretic peptide concentrations were analyzed in 119 pediatric patients with CHD and 33 healthy pediatric patients.

Results: N-terminal-pro-B-type natriuretic peptide levels in normal patients (mean, 120; range: 60–380 pg/mL) differed significantly from CHD patients with acyanotic heart diseases (mean, 372; range: 60–3000 pg/mL, P<0.001) and cyanotic heart diseases (mean, 1023; range: 182–3000 pg/mL, P<0.001). The diagnostic performance of NT-proBNP to differentiate patients with and without CHD was high with an area under curve of 0.79. At a cut-off value of 98 pg/mL, the sensitivity was 82% and the specificity was 46%.

Conclusion: N-terminal-pro-B-type natriuretic peptide measurement may be a valuable tool in screening pediatric patients for CHD.

  • N-terminal-pro-B-type natriuretic peptide
  • NT-proBNP
  • congenital heart disease
  • pediatrics

Two types of natriuretic peptide hormones, namely atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP), are secreted by the heart. Both of these peptides exhibit high structural homologies, and their genes are located in close proximity on Chromosome 1.1 These peptides regulate extracellular fluid volume and blood pressure. Further roles include induction of natriuresis, diuresis, and vasodilation. Both ANP and BNP also act specifically to counter the effects of stress hormones produced by the renin-angiotensin-aldosterone systems (RAAS), adrenergic systems, and endothelin.2 Myocyte cardiac stretch is the primary stimulus for the release of ANP and BNP.3 Atrial natriuretic peptide is primarily secreted from the cardiac atria, and BNP is primarily secreted from the ventricles.4

Brain natriuretic peptide is synthesized as a 108-amino acid prohormone (proBNP). This prohormone is cleaved by proteolytic enzymes known as furin and corin upon secretion and yields a C terminal biologically active peptide (BNP, amino acid 77–108) and an N terminal prohormone (NT-proBNP, amino acid 1–76), which is biologically inactive. Both of these components are found in the plasma.1,4,5,6 Plasma levels of natriuretic peptides are high in many cardiac diseases; therefore, they can serve as biochemical markers for heart disease.4 However, recent studies have shown that BNP and NT-proBNP are superior to ANP as biochemical markers.7,8

Brain natriuretic peptide and NT-proBNP levels were found to be elevated in adult cardiac patients who had left ventricular (LV) systolic and diastolic dysfunction,2,3,9 ischemic heart disease, hypertrophic cardiomyopathy,10 and other cardiac diseases.4 Generally, pediatric populations exhibited similar findings as adults.4 However, it should be noted that clinical signs and symptoms of heart disease in pediatric populations may lack sensitivity and specificity. Furthermore, they can be obscured by coexisting illnesses, such as lung diseases.1113

Therefore, this study was conducted to evaluate the usefulness of NT-proBNP levels in screening for pediatric patients with congenital heart diseases (CHDs). From the data obtained, it is hoped that we would be able to gauge the reliability of this test in terms of sensitivity and specificity.

Materials and Methods

Subjects were obtained prospectively from pediatric patients who had been newly referred to the Pediatric Cardiology Unit in Penang General Hospital, Malaysia, between 2008 and 2010. A total of 152 pediatric subjects between the ages of 5 days and 12 years were enrolled, including those (n=119) with various types of CHDs and those (control subjects; n=33) with no evidence of CHD or with an innocent heart murmur. Patients with acquired heart diseases and other chronic illnesses such as Down syndrome were excluded in this study. The history (biological data and symptoms of CHD) of each patient was recorded. Physical examinations, which included growth parameters as well as general and cardiovascular examinations, were carried out. Further investigations such as electrocardiogram and echocardiogram, which were part of the standard cardiac assessment, were completed. All of the above procedures were performed by qualified pediatric cardiologists. Echocardiography was the main tool of investigation for confirming the presence or absence of CHD. Baseline characteristics of the patients are shown in Table 1. Parental consent was obtained prior to blood sample collection. Blood samples were drawn with the patients at rest. This study was approved by the Clinical Research Center (CRC), Malaysian Ministry of Health (MOH), and Advanced Medical and Dental Institute, Universiti Sains Malaysia (AMDI, USM).

A laboratory evaluated a point-of-care system known as the Cardiac Reader (Roche Diagnostics, Mannheim, Germany), and it was used to determine NT-proBNP levels in the blood samples.14 The cardiac reader used a monoclonal and polyclonal antibody directed against amino acids 27–31 and 39–50 of the NT-proBNP molecule, respectively. The gold-labeled monoclonal antibody and biotinylated polyclonal antibody formed a sandwich structure with the NT-proBNP molecule in the middle. This complex was attached to a streptavidin molecule and coupled to the base of the test strip. The color intensity of the sandwich structure was then measured optically by the Cardiac Reader. This point-of-care assay used 150 μL heparinized venous whole blood and provided a turnaround time of ∼12 minutes for an NT-proBNP result.15

N-terminal-pro-B-type natriuretic peptide levels among the study population were analyzed using log transformed values of NT-proBNP (Figure 1A and Figure 1B). The influence of 5 factors (variables) on NT-proBNP levels was studied. These factors included history of the patients (symptomatic or asymptomatic for CHD), size of heart chambers (based on echocardiograph results), classification of CHD (cyanotic or acyanotic), and severity of the disease and gender. All statistical analysis was performed using SPSS for Windows software (SPSS 12.0, Chicago, IL).


Congenital heart disease patients who were symptomatic and asymptomatic had significantly higher values of NT-proBNP (mean=1361 pg/mL, P<0.001 and mean=238 pg/mL, P=0.005), respectively, compared to healthy patients (mean=120 pg/mL). N-terminal-pro-B-type natriuretic peptide levels were much higher in CHD patients with dilated left heart chambers (mean=676 pg/mL, P<0.001) or dilated right heart chambers (mean=660 pg/mL, P<0.001). Congenital heart disease patients with non-dilated heart chambers had lower NT-proBNP values (mean=251 pg/mL). However, this level was still significantly higher (P=0.039) than the value obtained for healthy patients. Patients with acyanotic heart diseases had high NT-proBNP levels (mean=372 pg/mL, P<0.001) compared to normal patients. The levels in patients with cyanotic heart disease were higher still (mean=1023 pg/mL, P<0.001). There was a non-significant increase of NT-proBNP levels in patients with mild CHD (mean=151 pg/mL; P=0.567). However, this changed drastically as the disease progressed to moderate (mean=512 pg/mL, P<0.001) and severe (mean=2512 pg/mL). There was no significant difference of NT-proBNP levels between healthy male and female subjects (P=0.392). A similar observation was made between male CHD patients and female CHD patients (P=0.085).

View this table:
Table 1

Baseline Characteristics of the Study Population

VariablePercentage (%)
Gender (N=152)
Male (N=79)52.0
Female (N=73)48.0
Age Categories (N=152)
<1 month (N=21)13.8
1 month-1 year (N=52)34.2
>1 year-<7 years (N=36)23.7
7 years-12 years (N=43)28.3
Race (N=152)
Malay (N=109)71.7
Chinese (N=28)18.4
Indian (N=14)9.2
Others (N=1)0.7
Classification of CHD* (N=152)
None/Normal (N=33)21.7
ACNHD* (VSD, ASD, PDA)(N=101)66.4
CNHD* (TOF, TGA, PA, PS) (N=18)11.8
Severity of the lesion (N=152)
None/Normal (N=33)21.7
Mild (N=54)35.5
Moderate (N=36)23.7
Severe (N=29)19.1
  • * CHD, Congenital Heart Disease; ACNHD, Acyanotic Heart Disease; CNHD, Cyanotic Heart Disease.

  • VSD, Ventricular Septal Defect; ASD, Atrial Septal Defect; PDA, Persistent Ductus Arteriosus.

  • TOF, Tetralogy of Fallot; TGA, Transposition of Great Arteries; PA, Pulmonary Atresia; PS, Pulmonary Stenosis

Figure 1

(1A) Distribution of NT-proBNP levels in the study population. (1B) Distribution of NT-proBNP levels in the study population using log transformed values. Log transformed values yielded a more normal distribution compared to actual values. Therefore, all statistical analyses were performed using log transformed values, which were then converted back to the corresponding actual values.

Multiple linear regression analysis was carried out to determine the strongest predictors of NT-proBNP levels. A good regression model was obtained (adjusted R2=0.780, F5,146=107.93, P<0.001). Based on this model, the strongest factors that influenced NT-proBNP levels are the severity of the CHD (standardized β coefficient=0.788, P<0.001) and presence of CHD in general (standardized β coefficient=0.372, P<0.001).

Receiver operating characteristic (ROC) curves were constructed (Figure 2A2D) to evaluate the diagnostic performance of this test in differentiating the various types of CHD. Based on the curves, a suitable cut-off value for NT-proBNP levels was chosen to select for pediatric patients with various types of CHD. An NT-proBNP cut-off value of 91 pg/mL could differentiate an Acyanotic Heart Disease (ACNHD) patient (n=101) from a healthy sample population (n=33) with a sensitivity of 84% and specificity of 42% (Figure 2A). On the other hand, congenital nonspherocytic hemolytic disease (CNHD) patients (n=18) could be differentiated from a healthy sample population (n=33) at an NT-proBNP cut-off value of 318 pg/mL with 94% sensitivity and 97% specificity (Figure 2B). Likewise, an NT-proBNP value of 408 pg/mL was 83% sensitive and 57% specific in differentiating patients with ACNHD (n=101) from patients with CNHD (n=18) (Figure 2C). Generally, a patient with CHD (n=119) could be differentiated from a healthy sample population (n=33) at a NT-proBNP cut-off value of 98 pg/mL with 82% sensitivity and 46% specificity (Figure 2D).

Similarly, another set of ROC curves was constructed (Figure 3) to analyze the diagnostic performance of NT-proBNP in predicting the severity of the heart lesions. An NT-proBNP cut-off value of 64 pg/mL could differentiate patients with mild CHD (n=54) from the control patients (n=33) with a sensitivity of 83% and specificity of 30% (Figure 3A). Patients with moderate CHD (n=33) could be differentiated from the control patients (n=33) at an NT-proBNP cut-off value of 178 pg/mL with 92% sensitivity and 70% specificity (Figure 3B). This test can perfectly (sensitivity 100%, specificity 100%) discriminate between patients with severe CHD (n=29) from the healthy sample population (n=33) at an NT-proBNP level of 440 pg/mL.

Figure 2

(2A) Receiver operating characteristic curve showing the ability of NT-proBNP to identify patients with ACNHD (AUC [95%CI]: 0.75[0.67–0.83]). (2B) Receiver operating characteristic curve showing the ability of NT-proBNP to identify patients with CNHD (AUC [95%CI]: 0.98[0.94–1.01]). (2C) Receiver operating characteristic curve showing the ability of NT-proBNP to differentiate patients with ACNHD and CNHD (AUC [95%CI]: 0.74[0.63–0.84]). (2D) Receiver operating characteristic curve showing the ability of NT-proBNP to identify patients with CHD in general. (AUC [95%CI]: 0.79[0.72–0.86]).


The results show that NT-proBNP levels do not follow normal Gaussian distribution, but the log transformed values do. This observation is echoed in several studies as well.4,16,17 N-terminal-pro-B-type natriuretic peptide levels were also able to differentiate between asymptomatic CHD patients and healthy patients with relatively high diagnostic accuracy (ie, AUC∼0.8). The high NT-proBNP levels in CHD patients with dilated left or dilated right chambers could be due to LV systolic dysfunction, right ventricular pressure, or volume overload.13,16,18,19 Patients with acyanotic heart disease showed higher levels of NT-proBNP due to the significant left-to-right shunt usually caused by ventricular septal defect (VSD), atrial septal defect (ASD), or persistent ductus arteriosus (PDA).18 The extremely high NT-proBNP levels observed in our patients could be due to the fact that cyanotic heart lesions are the most complex congenital malformations and tend to be more complicated and severe. This statement could be substantiated by the results in this study showing 55.6% of the CHD patients had a severe form of this disease. Hopkins and colleagues20 showed that adult patients with various forms of cyanotic heart disease including Eisenmenger syndrome (VSD with pulmonary hypertension and cyanosis) had higher NT-proBNP levels compared with the controls. There were no gender-dependent changes of NT-proBNP levels observed in this study, which is in good agreement with the results obtained by others.21,22

The ROC curve analysis shows that the NT-proBNP test can discriminate CNHD patients from the healthy sample population effectively (AUC=0.98). However, the effectiveness of the test is somewhat lesser in ACNHD patients (AUC=0.75). This could be attributed in part to the severity of the lesion as 84% of CNHD patients had a more complex form of CHD while only 49% of the ACNHD patients had a rather severe form of heart lesions. In addition, the ability of this test to discriminate between patients with ACNHD and CNHD was not too evident (AUC=0.735).

Figure 3

(3A) Receiver operating characteristic curve showing the ability of NT-proBNP to identify patients with mild CHD (AUC [95%CI]: 0.57[0.45–0.70]). (3B) Receiver operating characteristic curve showing the ability of NT-proBNP to identify patients with moderate CHD (AUC [95%CI]: 0.94 [0.88–0.99]). (3C) Receiver operating characteristic curve showing the ability of NT-proBNP to identify patients with severe CHD (AUC: 1.00).

Receiver operating characteristic curve analysis for the severity of the heart lesion showed that this test is rather weak in discriminating patients with mild CHD from the normal population (AUC=0.57). This could be due to the large overlap in NT-proBNP values between the 2 groups. A similar observation was made by Pfister and colleagues23 where there was a huge overlap of NT-proBNP levels in patients with a lesser degree of LV dysfunction. Nevertheless, these values changed significantly in patients with moderate CHD compared to the healthy population (AUC=0.94), and patients with severe CHD compared to the controls (AUC=1.0).

The ROC curve analyses and other results show that elevated NT-proBNP levels (above 98 pg/mL) can discriminate between patients with CHD in general from the healthy sample population with a relatively high diagnostic accuracy (AUC∼0.8). The high level of sensitivity achieved (>80%) shows this test may serve as an important screening tool and provide a more convincing selection criterion in identifying pediatric patients with CHD who are candidates for echocardiography compared to the conventional approach of listening for heart murmurs, which can be rather subjective.

Because of the relatively small size of our subgroups of patients with acyanotic or cyanotic heart disease, a larger study may be useful in determining if NT-proBNP levels can discriminate with high diagnostic accuracy between pediatric patients with these types of heart diseases and others (eg, dilated cardiomyopathy, myocardial infarction, and Kawasaki Disease [mucocutaneous lymph node syndrome]). Furthermore, this test could be evaluated in other pediatric clinical settings, such as intensive care units, and preoperative monitoring before surgeries. The results obtained might be useful in the management of pediatric cardiac patients.


N-terminal-pro-B-type natriuretic peptide levels may be a valuable screening tool in identifying pediatric patients with CHD and also serve as an indicator regarding the severity of the heart lesion.


This work was funded by the Research University Grant (No. 1001/CIIPPT/812013) from Universiti Sains Malaysia (USM). The authors would like to thank research officers Abdul Rahman Azhari and Nur Farhayu Omar for their help in collecting and analyzing some of the samples. The invaluable assistance of the pediatric cardiologists, medical officer, medical assistant, and nurses in the Pediatric Cardiology Unit, Penang General Hospital, is greatly appreciated. The authors would also like to thank the staff of the Advanced Medical and Dental Institute.


N-terminal-pro-B-type natriuretic peptide
congenital heart disease
atrial natriuretic peptide
brain natriuretic peptide
renin-angiotensin-aldosterone system
Clinical Research Center
Ministry of Health
Advanced Medical and Dental Institute, Universiti Sains Malaysia
receiver operating characteristic
acyan-otic heart disease
congenital nonspherocytic hemolytic disease
ventricular septal defect
atrial septal defect
persistent ductus arteriosus
left ventricular


View Abstract