How are brain natriuretic peptide (BNP) and N-terminal prohormone of brain natriuretic peptide (NT-proBNP) levels used in the evaluation of left ventricular (LV) function?

Updated: Jan 08, 2018
  • Author: Donald Schreiber, MD, CM; Chief Editor: Erik D Schraga, MD  more...
  • Print


Initial studies demonstrated that brain natriuretic peptide (BNP) levels were highly correlated with increasing congestive heart failure (CHF) symptoms. [2, 3, 4, 5, 6] These studies showed that BNP levels were strongly related to impaired left ventricular (LV) function, as measured using the LV ejection fraction (LVEF) during echocardiography and angiography. Studies compared BNP and NT-proBNP assays as well, finding little diagnostic difference between them; however, the studies tended to be underpowered with regard to detecting such a difference.

Seino and colleagues, using BNP and NT-proBNP assays, compared 105 patients having chronic heart failure with 67 healthy control subjects and found that the peptide levels were significantly correlated with CHF symptoms. The levels were measured using the New York Heart Association (NYHA) classification of heart failure. [7]

Furthermore, receiver operating characteristic (ROC) curves were calculated to detect an LVEF of less than 40%, with resulting area under the curve (AUC) values of 0.75 and 0.77 for NT-proBNP and BNP measurements, respectively. (The ideal test has an AUC of 1.0, and a test based on chance alone has an AUC of 0.5.) When adjusted to detect an LVEF of less than 50%, AUC values increased to 0.82 and 0.79 for NT-proBNP and BNP, respectively.

In another study, of 180 inpatients, Mueller et al concluded that BNP and NT-proBNP were equally sensitive for diagnosing an exacerbation of CHF but that NT-proBNP may be most useful for detecting early LV dysfunction without overt clinical CHF. [8] Heart failure was diagnosed in 43 patients, with no difference in the AUC detected (0.92 and 0.93 for BNP and NT-proBNP, respectively). However, among 56 patients with asymptomatic structural heart disease and 81 patients without structural heart disease, NT-proBNP measurements resulted in an AUC of 0.84, a statistically significantly greater value than the AUC of 0.74 found for BNP measurements.

In a small study in which 43 symptomatic patients with echocardiographic abnormalities were compared with 137 asymptomatic control subjects with normal echocardiograms, the AUC to detect significant LV dysfunction was 0.92 for NT-proBNP and 0.93 for BNP. These results were slightly superior to those obtained in 339 hospitalized patients referred for diagnostic coronary angiography. In this patient population, NT-proBNP measurements provided an AUC of 0.83 for identifying severe systolic dysfunction and 0.81 for identifying any LV dysfunction.

The relative performance of BNP and NT-proBNP was also compared, and no significant difference in performance characteristics was detected. However, NT-proBNP measurements tended to improve sensitivity in detecting low levels of LV dysfunction.

In a large study by McDonagh et al in which the authors defined an abnormal NT-proBNP level as one greater than the 95th percentile for patients without LV dysfunction, NT-proBNP levels aided the diagnosis of CHF with a sensitivity of 75% and a negative predictive value of 99%. [9]

In this study, NT-proBNP levels were compared with LV function in 3051 patients pooled from 3 large European data sets. About 10% of these patients had significant LV dysfunction, whereas 3% had acute CHF. Median values for patients with normal LV function, LV dysfunction, and acute CHF were 20, 117, and 270pg/mL, respectively. AUCs for diagnosing clinical CHF and for detecting any LV dysfunction were 0.85 and 0.69, respectively.

Rather than selecting a specific cutoff with optimal performance characteristics, the authors, as stated, defined an abnormal NT-proBNP level as one greater than the 95th percentile for patients without LV dysfunction, corrected for age and sex. By using this cutoff strategy, NT-proBNP levels had the above-mentioned sensitivity of 75% and negative predictive value of 99% in the diagnosis of CHF

These data confirm the correlation between natriuretic peptide levels and degree of LV dysfunction, using either a BNP or the NT-proBNP assay. Furthermore, subjects with echocardiographically determined LV dysfunction or structural LV heart disease (eg, valvular heart disease), but without clinical CHF exacerbation, have natriuretic peptide levels between those of persons with a normal heart and normal LV function and the levels of patients who have been diagnosed with acute clinical CHF.

Although these studies demonstrated moderate discriminatory power of natriuretic peptide testing in detecting LV dysfunction, each study revealed substantial overlap in BNP and NT-proBNP levels between patients with and those without LV dysfunction, symptomatic or not. These results suggest that BNP or NT-proBNP might not be useful in screening for CHF or LV dysfunction.

However, with appropriate cutoffs and in certain high-risk populations, such as patients with dyspnea in the emergency department, BNP and NT-proBNP test characteristics provide sufficient negative predictive value to effectively rule out LV dysfunction.

Did this answer your question?
Additional feedback? (Optional)
Thank you for your feedback!