OUP user menu

Relation between serum N-terminal pro-brain natriuretic peptide and prognosis in patients with hypertrophic cardiomyopathy

Caroline J. Coats , Mathew J. Gallagher , Michael Foley , Constantinos O'Mahony , Christopher Critoph , Juan Gimeno , Anne Dawnay , William J. McKenna , Perry M. Elliott
DOI: http://dx.doi.org/10.1093/eurheartj/eht070 2529-2537 First published online: 1 March 2013


Aims To determine the relation between serum concentrations of N-terminal pro-brain natriuretic peptide (NT-proBNP) and prognosis in patients with hypertrophic cardiomyopathy (HCM).

Methods and results In total, 847 patients (53 ± 15 years; 67% male) with HCM (28% with left ventricular outflow tract obstruction ≥30 mmHg at rest) were followed for 3.5 years (IQR 2.5–4.5 years). The median NT-proBNP concentration was 78 pmol/L (range <5–1817 pmol/L and IQR 31−183 pmol/L). Sixty-eight patients (8%) reached the primary endpoint of all-cause mortality or cardiac transplantation. NT-proBNP concentration predicted long-term survival from the primary endpoint [area under the receiver operating characteristic curve of 0.78 (95% confidence interval 0.73−0.84)]. A serum concentration of ≥135 pmol/L was associated with an annual event rate of 6.1% (95% CI 4.4−7.7). Three independent predictors of primary outcome were identified in a multivariable Cox model: New York Heart Association class III/IV (HR 2.10, 95% CI 1.21−3.65, P = 0.008), ejection fraction (HR 0.98, 95% CI 0.96−1.00, P = 0.035), log NT-proBNP (HR 2.04, 95% CI 1.56−2.66, P < 0.001). Log NT-proBNP was a significant predictor of heart failure (HF) and transplant-related deaths (n = 23; HR 3.03, 95% CI 1.99−4.60, P < 0.001) but not sudden death or appropriate implantable cardioverter defibrillator shock (n = 11; HR 1.54, 95% CI 0.91−2.60, P = 0.111). In patients with ejection fraction ≥50% (n = 673), log NT-proBNP remained an independent predictor of the primary outcome (HR 2.11, 95% CI 1.54−2.90, P < 0.001).

Conclusion In patients with HCM, elevated NT-proBNP concentration is a strong predictor of overall prognosis, particularly HF-related death and transplantation.

  • Hypertrophic cardiomyopathy
  • Prognosis
  • Heart failure
  • Brain natriuretic peptide
  • Risk stratification


Pro-B-type natriuretic peptide (proBNP) is a cardiac neurohormone synthesized and secreted mainly from the ventricles of the heart in response to ventricular wall stress. After secretion, proBNP is cleaved into BNP and N-terminal proBNP (NT-proBNP). In addition to a natriuretic action, BNP and NT-proBNP relax vascular smooth muscle, peripherally inhibit the sympathetic nervous and renin–angiotensin–aldosterone systems, and have direct lusitropic effects.1,2 In patients with chronic heart failure (HF), BNP is a useful diagnostic test and prognostic marker.3,4

In hypertrophic cardiomyopathy (HCM), BNP is expressed in ventricular myocytes with particularly high levels found in patients with left ventricular outflow tract (LVOT) obstruction and diastolic dysfunction.57 Clinical studies in children and adults with the disease show that BNP and NT-proBNP concentration correlate with markers of disease severity and exercise intolerance, respectively.810 Although BNP concentrations in tissue and plasma are associated with fibrosis11 and progression to systolic dysfunction,6 it has been suggested that BNP has a limited role in the diagnosis and management of HCM because concentrations only weakly correlate with clinical symptoms and markers of systolic and diastolic function.12 However, most studies are small and few have examined the relation of BNP concentration to prognosis.


Study population

This was an observational, single-centre cohort study. Data were collected prospectively. The study conforms to the principles of the Helsinki declaration. All authors have read and agreed to the manuscript as written. The study cohort was recruited between January 2006 and July 2010 at The Heart Hospital, a specialist inherited heart muscle disease unit. The diagnosis of HCM was based on echocardiographic identification of hypertrophy [maximum left ventricular wall thickness (MWT) ≥15 mm] unexplained by abnormal loading conditions or in accordance with the criteria for the diagnosis of familial disease in patients with at least one first-degree relative with an unequivocal diagnosis.13,14 Maximum left ventricular wall thickness at the time of initial diagnosis was noted. Patients with Friedreich's ataxia, Noonan syndrome, and metabolic disorders were excluded. All patients were ≥16 years of age at evaluation.

Patient assessment

New York Heart Association (NYHA) class, current medication, and blood pressure were recorded. A 12-lead electrocardiogram was recorded at 25 mm/s and 10 mm/mV. Heart rate and rhythm were recorded. Transthoracic echocardiography was performed in the left lateral decubitus position (M3S Probe, Vivid i or Vivid 7; GE-Vingmed, Norway). Maximum LV wall thickness was defined as the greatest thickness in any single segment. LV end-diastolic diameter and end-systolic diameter were measured from M-mode and 2-D images obtained from parasternal long-axis views. Fractional shortening and Teicholz ejection fraction were calculated.15 Left ventricular outflow tract gradient was calculated from continuous-wave Doppler using the simplified Bernoulli equation. Significant LVOT obstruction was defined as a peak gradient ≥30 mmHg.16

Clinical risk stratification for sudden cardiac death (SCD) was performed in accordance with previously published methods.17 The following clinical risk factors were assessed: SCD attributable to HCM in one or more first-degree relatives, unexplained syncope, MWT ≥30 mm, LVOT ≥30 mmHg at rest, non-sustained ventricular tachycardia (NSVT) on Holter monitoring (≥3 beats at >120 b.p.m.), and an abnormal blood pressure response to exercise defined as a systolic blood pressure rise of <20 mmHg from baseline during exercise off medication.18,19

Serum N-terminal pro-brain natriuretic peptide measurement

Venous blood samples were drawn under resting conditions into serum separator tubes and sent to the laboratory for quantification the same day as routine sample analysis. Serum NT-proBNP was measured by a two-site electrochemiluminescence immunoassay on a Roche E170 analyser. Results were reported in SI units of pmol/L with a lower limit of detection of <5 pmol/L.20 Normal values for the assay were defined using the manufacturer's guidelines; they were derived from 1981 blood donors aged between 18 and 65 years and 283 individuals with no known cardiac disease aged between 50 and 90 years of age (see Supplementary material online, File S1). Patients with values above the 97.5th percentile for age and gender were considered to have an abnormal NT-proBNP.


Patients were followed up as outpatients at The Heart Hospital. Survival status and time to death were assessed using hospital attendance records and the National Health Service computer system. The last date of follow-up was 31 October 2012.

Statistical analysis

The primary endpoint was all-cause mortality or cardiac transplantation. The secondary endpoints were: (i) death due to HF or cardiac transplantation; (ii) SCD or appropriate shock from an implantable cardioverter defibrillator (ICD); (iii) all cardiovascular deaths, appropriate ICD shock, or transplantation. Death related to HF was defined as death preceded by signs and/or symptoms of HF of more than 1 h duration, and/or cardiogenic shock. Sudden cardiac death was defined as witnessed sudden cardiac death within 1 h of new symptoms or nocturnal death with no antecedent history of worsening symptoms.

Statistical analysis was performed using PAWS version 18.0 (SPSS, Inc., Chicago, IL, USA). Continuous variables are presented as mean ± standard deviation and median and inter-quartile range where appropriate, and categorical variables as frequencies (n) and percentages (%). Logarithmic transformation allowed NT-proBNP concentrations to be treated as a normally distributed variable. The Pearson correlation was used to determine correlation between log NT-proBNP and clinical variables. The association of the primary and secondary outcomes with individual demographic, echocardiographic, and clinical variables was assessed using the Cox proportional hazards model. The predictor variables were pre-specified on the basis of previously published literature and included age, gender, log NT-proBNP, atrial fibrillation, NYHA class, ejection fraction, LA size, family history of SCD, left ventricular MWT, LVOT gradient, and NSVT.

A multivariable model fitted using backward elimination with a significance level of 10% followed the univariable analysis. Results are presented as hazard ratios with 95% confidence intervals (CIs). Receiver operating characteristic (ROC) curve analysis was performed to test the ability of NT-proBNP to predict the study outcomes. Survival estimates were calculated by the Kaplan–Meier method and the log-rank test used for comparison between groups. The study cohort was divided into tertiles of NT-proBNP concentration on the basis of a pre-specified protocol; this approach has previously been taken by the Framingham Heart Study group.21 A two-sided P-value of <0.05 was considered to indicate statistical significance in all analyses.


Eight hundred and forty-seven patients were studied [53 ± 15 years old; 569 (67%) male]. Baseline characteristics of the study population are shown in Table 1. The median NT-proBNP concentration was 78 pmol/L (range <5–1817 pmol/L, IQR 31–183 pmol/L). N-terminal pro-brain natriuretic peptide concentrations were higher in women (median 121 pmol/L, IQR 64–288 pmol/L) than in men (median 63 pmol/L, IQR 23–145 pmol/L). Mean NT-proBNP concentrations were higher across all age groups compared with reference values (Figure 1). Six hundred and sixty-eight (79%) patients had an abnormal NT-proBNP concentration, of which 147 (17%) had a left ventricular EF <50%.

View this table:
Table 1

Characteristics of the study population

Figure 1

Bar chart showing mean N-terminal pro-brain natriuretic peptide concentrations in the study population alongside mean reference values for men and women. Normal values for the assay were derived from 1981 blood donors aged between 18 and 65 years and 283 individuals with no known cardiac disease aged between 50 and 90 years of age (see Supplementary material online, File S1).

Table 2 shows baseline characteristics according to tertiles of NT-proBNP. The upper tertile contained more females; patients were older (56 ± 16 years) with higher serum creatinine (95 ± 33 µmol/L), larger left atria (48 ± 9 mm), and higher LVOT gradients (38 ± 45 mmHg). They were more likely to have moderate or severe MR and paroxysmal or permanent atrial fibrillation.

View this table:
Table 2

Clinical characteristics according to N-terminal pro-brain natriuretic peptide tertile

Log NT-proBNP correlated positively with age (r = 0.23, P < 0.05) and creatinine (r = 0.16; P < 0.05) and inversely with weight (r = −0.22, P < 0.05). There was no relation between log NT-proBNP and symptoms of chest pain (P = 0.315) or palpitations (P = 0.06). There was a significant difference in NT-proBNP across NYHA classes (Figure 2).

Figure 2

Box and whisker plot showing the relation of N-terminal pro-brain natriuretic peptide to New York Heart Association class.

Study endpoints

Patients were followed for a median duration of 3.5 years (IQR 2.5– 4.5 years). During this time, 68 patients (8%) died or underwent cardiac transplantation. There were 8 SCDs (mean age 42 ± 13 years); 16 HF deaths (55 ± 14 years), and 7 cardiac transplants (45 ± 12 years). Seven other cardiac deaths were due to myocardial infarction (n = 1), valvular heart disease (n = 1), stroke (n = 1), and post-operative complications (n = 4). There were 12 non-cardiac deaths due to cancer (n = 7), intra-cranial bleed (n = 2), end-stage lung disease (n = 1), infection (n = 1), and trauma (n = 1). The cause of death could not be confirmed in 18 patients. Three patients received an appropriate ICD shock during the follow-up period. Of the 16 patients who died from HF, 6 had restrictive physiology and pulmonary hypertension, 3 died while on the transplant waiting list, and none had LVOT obstruction at the time of NT-proBNP measurement.

Survival analysis

Patients with an abnormal NT-proBNP had a seven-fold increased risk of death or transplantation (relative risk 6.7, 95% CI 1.7–27.2, P = 0.0074) compared with those with a normal NT-proBNP. N-terminal pro-brain natriuretic peptide levels predicted the primary endpoint with an area under the ROC curve of 0.78 (95% CI 0.73–0.84). A cut-off of <100 pmol/L had a 96% negative predictive value (Table 3).

View this table:
Table 3

Operating characteristics of N-terminal pro-brain natriuretic peptide thresholds to predict death or transplantation in patients with hypertrophic cardiomyopathy

N-terminal pro-brain natriuretic peptide concentration was higher in patients who were transplanted or died from HF (median 274 pmol/L, IQR 153–446) than in those who died suddenly or received an appropriate ICD shock (median 184 pmol/L, IQR 62–277) during the follow-up period. Univariable Cox regression found Log NT-proBNP was a significant predictor of HF and transplant-related deaths (n = 23; HR 3.03, 95% CI 1.99–4.60, P < 0.001) but not sudden death or appropriate ICD shock (n = 11; HR 1.54, 95% CI 0.91–2.60, P = 0.111). The event rate did not allow further multivariable analysis. The area under the ROC curve was 0.83 (95% CI 0.77–0.88) for HF and transplant deaths and 0.65 (95% CI 0.51–0.79) for SCD or appropriate ICD shock.

The overall event rate was 6.1% (95% CI 4.4–7.7%) per year in the highest tertile (NT-proBNP ≥135 pmol/L), 1.4% per year in the middle tertile (NT-proBNP 45–134 pmol/L), and 0.4% per year in the lowest tertile (NT-proBNP ≤44 pmol/L). Survival rates evaluated by a Kaplan–Meier lifetime analysis were lower in patients with abnormal NT-proBNP (Figure 3). This relation was significant across all strata (log-rank P = 0.028 for lowest and middle tertiles, and log-rank P < 0.0001 for middle and highest tertile).

Figure 3

Kaplan–Meier analysis showing cumulative rates of survival in 847 patients with hypertrophic cardiomyopathy, stratified into tertiles according to serum N-terminal pro-brain natriuretic peptide concentration.

Relation of N-terminal pro-brain natriuretic peptide to clinical variables

Table 4 shows the clinical predictor variables and their relation to survival. Age, symptoms, female gender, left atrial enlargement, low ejection fraction, atrial fibrillation, and high NT-proBNP were significant predictors of the primary outcome; family history of SCD, left ventricular wall thickness, LVOT gradient, and NSVT were not statistically significant.

View this table:
Table 4

Univariable Cox regression models for whole cohort (n = 847) according to primary and secondary endpoints

Multivariable Cox proportional hazards model analysis (Table 5) identified three independent predictors of all-cause mortality or transplantation: NYHA class III/IV (HR 2.10, 95% CI 1.21–3.65, P = 0.008), ejection fraction (HR 0.98, 95% CI 0.96–1.00, P = 0.035), and log NT-proBNP (HR 2.04, 95% CI 1.56–2.66, P < 0.001). When NT-proBNP was excluded from the model, gender, left atrial size, NYHA class, and ejection fraction were significant predictors of the primary outcome (P < 0.05). Considering only patients with ejection fraction ≥50% (n = 673), 42 reached the primary endpoint, and NT-proBNP remained an independent predictor of outcome (HR 2.11, 95% CI 1.54–2.90, P < 0.001) (Table 6).

View this table:
Table 5

Multivariable Cox regression models for whole cohort (n = 847) to predict all-cause mortality or transplantation (n = 68 events)

View this table:
Table 6

Multivariable Cox regression models for the subgroup where ejection fraction ≥50% (n = 673) to predict all-cause mortality or transplantation (n = 42 events)


This study shows that NT-proBNP is a significant predictor of HF and transplant-related deaths but not sudden death or appropriate ICD shock in patients with HCM. This suggests that raised NT-proBNP might identify patients who would benefit from pharmacological strategies to prevent progressive left ventricular dysfunction before the onset of clinically apparent HF.

Previous studies of brain natriuretic peptide in hypertrophic cardiomyopathy

Cardiomyocytes are the major source of BNP and its related peptides in the circulation. The main stimulus to its secretion is increased ventricular wall stress, modulated by other neurohormones via paracrine and possibly endocrine mechanisms.22 Numerous cross-sectional studies have shown that BNP levels are increased in patients with HCM compared with normal individuals and that they correlate with symptoms of HF,9 severity of hypertrophy,9,12 Doppler echocardiographic signs of LV diastolic dysfunction,7 fibrosis detected by cardiac magnetic resonance,11 and myocardial ischaemia.23 Patients with LVOT obstruction have particularly high levels of natriuretic peptides, and relief of outflow tract obstruction has been shown to reduce plasma concentrations.24,25

Very few studies have examined the relation between survival and BNP concentration in HCM (Table 7). In the largest study to date (167 patients), more cardiovascular events were observed in patients with BNP ≥200 pg/mL,26 corroborating findings in a smaller study of 80 patients in which patients with NT-proBNP levels <1500 pg/mL had significantly less cardiovascular death and hospitalization for HF.27 In contrast, a study examining the relation between BNP and the combined outcomes of sudden death, HF, and stroke due to atrial fibrillation in 130 patients found that BNP did not predict outcome.28

View this table:
Table 7

Summary of previously published studies

Clinical implications of this study

In the general population, natriuretic peptide levels are powerful markers of morbidity and mortality, and in HF populations a reduction in NT-proBNP is associated with better outcome.21,29,30 The findings in this study show that NT-proBNP concentration also predicts survival in patients with HCM, particularly with respect to death from HF. This observation is important because most contemporary studies suggest that HF has become the predominant mechanism of premature mortality in patients with HCM and there is, therefore, a need for sensitive early biomarkers that can be used to target early preventative therapies.

It is well established that BNP concentrations are higher in women than in men and rise with age, findings reproduced in this study. There are relatively few studies specifically examining the effect of sex on prognosis in HCM, but the finding that females have with a poorer outcome with respect to overall mortality has been observed previously.31

Although there was an association between NYHA functional class and NT-proBNP concentration, there was considerable overlap in NT-proBNP levels between classes. These results are consistent with most studies of symptomatic patients with cardiovascular disease that show close relation between clinical, echocardiographic, and biological disease markers. Nevertheless, severe symptoms and NT-proBNP were independent predictors of adverse outcome, suggesting that they provide complementary information on disease severity and progression.


The primary endpoint was all-cause mortality or cardiac transplantation; this was chosen to eliminate any bias arising from incorrect classification of the cause of death. Wall thickness measurements are slightly lower than those in previously reported cohorts. This may reflect the resolution of modern echocardiography equipment. The Cox proportional hazards model measures the relative effect of predictive factors on outcome by assuming that this relation is constant over time. There are currently no data on serial NT-proBNP measurements in patients with HCM.


Serum NT-proBNP concentration predicts the risk of death and progression to end-stage HF in patients with HCM.


This work was undertaken at University College London Hospitals/University College London, which receive a proportion of funding from the Department of Health's NIHR Biomedical Research Centre funding scheme. The British Heart Foundation supports C.J.C. (FS/10/027/28248) and C.C. (FS/10/66/28489).

Conflict of interest: none declared.


We are grateful to Joanna Mander and Myra O'Donovan for their support.


View Abstract