Rahul Sakhuja Cardiology Division and Department of Medicine, Massachusetts General Hospital, Harvard Medical School , James L Januzzi Cardiology Division and Department of Medicine, Massachusetts General Hospital, Harvard Medical School

Originally printed in:
» US Cardiology 2004 - August 2004

Uses o f  NT- proBNP in CHF

A significant body of evidence has developed to demonstrate that NT-proBNP levels correlate with diagnosis, clinical status and prognosis in congestive heart failure, and may be useful for the longitudinal management of patients with CHF.

Diagnosis of CHF

Bay and colleagues published one of the first large studies revealing the utility of NT-proBNP in predicting LV dysfunction. From 3,236 hospitalized patients with symptomatic and asymptomatic CHF, NT-proBNP had a sensitivity of 73%, specificity of 82%, and, most impressively, a negative predictive value of 98%.The diagnostic value to predict a left ventricular ejection fraction (LVEF) <40% as represented by the area under the receiver operating characteristic curve (AUC) was 0.85. As compared with BNP,NT-proBNP was an equivalent predictor of LVEF <30% as reflected by an AUC of 0.88 versus 0.85. Overall, the study revealed that NT-proBNP added significantly more diagnostic power to the clinical history.11 Subsequently, other studies revealed that not only was NT-proBNP elevated in CHF from LV dysfunction, but was also in forms of CHF with normal LV function (diastolic dysfunction), although the levels of NT-proBNP among patients with non-systolic CHF are typically lower than those with systolic dysfunction and CHF.(9)

Following such large-scale studies demonstrating the feasibility of detection of LV abnormalities, the use of NT-proBNP for the acute evaluation of dyspneic patients with possible CHF was then explored in three recent studies. In the first such study, Lainchbury and colleagues demonstrated NT-proBNP to be of value in the evaluation of patients with dyspnea and suspected acute CHF in the emergency department (ED).(12) Subsequently, Bayes-Genis and colleagues found that NT-proBNP levels were significantly higher in patients with decompensated CHF, and also demonstrated the value of the marker for identifying those patients with ‘masked’ heart failure, defined as those patients with LV dysfunction and concomitant pulmonary disease. Furthermore, Bayes-Genis and others demonstrated as the heart failure was treated, NT-proBNP levels fell in tandem.(13,14)

Most recently, more definitive data supporting the use of NT-proBNP in the ED were reported. In a blinded prospective analysis of 600 patients presenting with acute dyspnea, the ProBNP Investigation of Dyspnea in the Emergency Department (PRIDE) Study investigators demonstrated NT-proBNP levels to be markedly elevated among patients with decompensated CHF (see Figure 2).(15) NT-proBNP was highly sensitive and specific for the diagnosis of acute CHF, and correlated with the severity of CHF symptoms.Among all the factors evaluated, an elevated NT-proBNP proved to be the single strongest independent predictor for the final diagnosis of acute CHF. Lastly, in the PRIDE Study, NT-proBNP was superior to clinical assessment for the identification of acute CHF. However, the combination of NT-proBNP testing plus clinical assessment was the most superior tool for patient evaluation.

Figure 2:Figure 2: Median Levels of NT-proBNP Among Patients Presenting with Dyspnea
Due to Acute CHF Versus Other Causes.

Data from the PRIDE study.15

Limited head-to-head data exist comparing NTproBNP to BNP for patients with suspected or confirmed CHF. As a general rule, BNP is used for the detection of acute CHF, but exceptions may exist and, in many studies, NT-proBNP was the more sensitive marker. (12) Among patients with acute CHF in the PRIDE Study, NT-proBNP was more sensitive than BNP (90% versus 80% overall), a finding demonstrated in several other studies.(16–19) In most of these studies,NTproBNP was particularly superior to BNP when evaluating patients with mild-to-moderate structural heart disease, diastolic CHF, or chronic relatively compensated CHF. Indeed, relevant to the last category of patients,Tang and colleagues demonstrated that more than 20% of symptomatic out-patients with chronic CHF may have BNP levels in the normal range, a finding not frequently demonstrated with NTproBNP. (20) Thus, in certain scenarios, NT-proBNP may be more sensitive than BNP.

Prognosisin CHF

With a five-year mortality of 50% and a 10-year mortality nearing 90%, a marker that correlates with prognosis would be helpful in risk stratification, particularly if such a marker was also useful for patient management.(21) Fisher and colleagues measured the concentration of NT-proBNP in 87 patients emergently admitted with CHF caused by LV dysfunction and found that NTproBNP levels were a strong predictor of both death and CHF hospitalization.(22) A larger follow-up study validated these findings in 650 ambulatory patients, among whom NT-proBNP was the strongest independent predictor of mortality (hazard ratio=5.70, p <0.0001), hospital admissions for CHF (hazard ratio=13.83, p <0.0001), and other cardiac admissions. (23) Interestingly, in addition to predicting prognosis and risk for decompensated CHF after MI,NT-proBNP recently also predicted mortality and urgent transplantation among patients with advanced heart failure from all etiologies awaiting transplant. (24–26) Among a study of patients with severe advanced LV dysfunction, NTproBNP was the strongest predictor of adverse outcome, more so than LV ejection fraction, heart failure survival scores, and the usual ‘gold standard’ for predicting mortality in chronic CHF,maximal oxygen extraction.(26) These exciting studies have led to the concept that NTproBNP testing may be the new ‘gold standard’ for predicting long-term outcomes in CHF.