Congestive heart failure (CHF) is a significant clinical problem and is a leading cause of hospitalization in North America. Dyspnea, however nebulous, is often the only presenting complaint, and the time to diagnosis of CHF, assessment of severity, and treatment is thus variable. This is concerning as the costs associated with the disease are astounding and we now have treatment options that result in marked symptom relief and improved survival. Currently, the most common approach to diagnosing heart failure includes a history and physical exam, electrocardiography, echocardiography, and radionuclide ventriculography. However, none of these are entirely reliable, reproducible, or realistic for everyday practice. It is not surprising then that there is widespread interest in identifying good markers for CHF, and that one of the natriuretic peptides (NPs) has been brought to the forefront as a promising option.

The NP family consists of three structurally similar but genetically distinct peptides: atrial (or A-type) natriuretic peptide (ANP), brain (or B-type) natriuretic peptide (BNP), and C-type natriuretic peptide (CNP) (Table 1). There are many conditions that are associated with an increase in NPs (Table 2).[1] The final common pathway in the vast majority of these disease states is left ventricular (LV) dysfunction, which results in a process of neurohormonal activation—a response from the sympathetic nervous system, the renin-angiotensin-aldosterone system, and the endothelin pathways[1] that can ultimately lead to myocardial apoptosis and fibrosis. [2,3] In contrast, activation of the NP system can play an important role in maintaining the compensated state of CHF through its broad effects at multiple sites in the cardiovascular system (Table 3).[4-5]

Ventricular dysfunction includes both diastolic and systolic dysfunction. Systolic dysfunction refers to an impaired ability to eject blood from the heart, whereas diastolic dysfunction refers to impaired cardiac relaxation, or the inability to fill appropriately with normal systolic function. It is important to remember that the two are not mutually exclusive.

Of the natriuretic peptides, BNP seems to be the most sensitive and specific indicator of ventricular dysfunction.[6-9] PreproBNP is broken down into BNP (a 32-amino-acid peptide) and NT-proBNP and released from myocytes. ANP is less specific for ventricular dysfunction, and can be elevated in hypertension, stress, or exercise situations. In contrast, BNP levels have been demonstrated to correlate with LV function [9,10] left ventricular end diastolic pressures, [11] and New York Heart Association classification. [7,12] Given the above, clinical studies have focused on plasma BNP concentrations.

There are two different assays for BNP and both are easy to perform with results available in a reasonable time frame. The first is a point-of-care test, which measures BNP directly and provides results in 15 minutes. This test is reflective of current ventricular status as it measures BNP, a biologically active peptide that has a half-life of 18 minutes. The second assay is an immunometric assay that measures NT-proBNP, a biologically inactive molecule that has a half-life of 1 hour[12] and has a longer invitro stability. Results of NT-proBNP studies show that it is comparable to BNP for detection of LV dysfunction. BNP and NT-proBNP values are highly correlative (r = 0.88).[13-14]

In terms of cost, the Biosite point-of-care BNP assay is $33 per assay. The cost of the Roche NT-ProBNP assay is not currently listed.

 BNP in the diagnosis of CHF

There have been many studies that have recognized BNP as a useful marker for CHF in primary care settings.[15-20] For example, a study of 122 consecutive patients with suspected new CHF found that a BNP level of 76 pg/mL has a negative predictive value of 98%, specificity of 84%, and a sensitivity of 97% for diagnosing CHF.[15] Point-of-care testing of BNP in the emergency department setting has also been shown to differentiate between pulmonary and cardiac causes of dyspnea. Maisel and colleagues studied 1586 patients presenting to the emergency department with acute dyspnea.[16] In this study, a cut-off BNP level of 100 pg/mL was recommended with a sensitivity of 90%, specificity of 76%, and an accuracy of 83% in distinguishing CHF from other causes of dyspnea. This was more accurate than both the National Health and Nutrition Examination Survey criteria (67%) and the Framingham criteria (73%). In multivariate analysis, BNP was the strongest independent predictor of CHF, with an odds ratio of 29.60. The study concluded that the use of BNP in conjunction with clinical information is useful in diagnosing and excluding CHF from the diagnosis when patients present to an emergency setting with dyspnea.

Finally, elevated BNP levels also have a relationship with diastolic filling abnormalities on echocardiogram, reinforcing the diagnosis of diastolic dysfunction.[21-23] Redfield and colleagues studied 657 patients with normal systolic function and found that BNP levels were higher in those with isolated diastolic dysfunction.[21] BNP alone cannot differentiate between diastolic and systolic dysfunction, but is strongly predictive in combination with clinical CHF and normal LV function.

 BNP in monitoring and optimizing treatment for CHF

Current practice for treating heart failure involves institution of therapy and titration of medications to target doses for symptomatic improvement. However, there is no optimal way to evaluate therapeutic efficacy on a neurohormonal or hemodynamic level. BNP not only reflects neurohormonal status, but correlates with a change in pulmonary artery wedge pressure, which is an independent predictor of mortality and morbidity in CHF. This was demonstrated by a pilot study of 20 patients, which showed a significant correlation between percentage change in wedge pressure from baseline per hour and a concomitant percentage change of BNP from baseline (r = 0.73, P < .05).[24]

Troughton and colleagues studied 69 patients and randomly assigned them to treatment guided by either NT-proBNP levels or clinical assessment in an outpatient setting.[25] In the NT-proBNP-guided therapy group, NT-proBNP levels fell 79 pmol/L below baseline at 6 months, as compared with 3 pmol/L in the clinically guided group. Patients in the NT-proBNP-guided therapy group had a decreased incidence of the primary composite endpoint: cardiovascular death, readmission, or new episodes of decompensated heart failure (P = .01). Other studies have also shown that titration of CHF therapy to BNP levels results in sustained symptomatic improvement and a decrease in hospital readmission rates and cardiac mortality.[26-28] For example, Cheng and colleagues studied 72 patients with decompensated CHF requiring hospital admission and showed that an increase in BNP during treatment was a strong predictor of mortality and early hospital readmission.[29]

It is clear that optimal treatment of CHF should go beyond symptom control and blind titration of medications to study target doses. Serial measurements of BNP would be useful in determining the proper dosage of CHF drugs and for monitoring the success of therapy.

 BNP as a prognostic tool

A large amount of data supports extending the use of BNP from a diagnostic marker to a prognostic marker. Harrison and colleagues looked at BNP levels in 325 patients presenting to the emergency department with dyspnea and determined that the admission BNP levels were highly predictive of cardiac events over the next 6 months.[30] Patients with BNP levels greater than 480 pg/mL had a 51% 6-month cumulative probability of a CHF event. BNP has also been shown to be a powerful predictor of functional status deterioration.[30] More recently, Berger and colleagues studied 452 patients with a left ventricular ejection fraction (LVEF) less than or equal to 35%, and multivariate analyses found BNP to be the strongest independent predictor of sudden death.[31] This has been demonstrated in other studies as well.[26]

BNP has also been looked at as a prognostic marker in acute coronary syndrome (ACS). De Lemos and colleagues measured BNP levels 40 ± 20 hours after the onset of acute ischemia in 2525 patients with ACS in the TIMI 16 study.[32] Baseline BNP levels were correlated with the risk of death, heart failure, and myocardial infarction at 30 days and 10 months.

Sabatine and colleagues found that elevations in troponin, creatine-reactive protein (CRP), and BNP were each an independent predictor of the composite endpoint of death, myocardial infarction, or CHF.[33] Troponin and CRP are also unique markers and reflect myocardial injury and inflammation, respectively.

 Other possible roles for BNP

BNP may eventually have a role as a treatment agent, given that short-term infusions of ANP and BNP in CHF patients have shown beneficial effects in improving central hemodynamics and increasing stroke volume.[4,34,35] However, there has been little success with orally active analogues of NPs. Therefore, more focus has been on investigating neutral-endopeptidase (NEP) inhibitors in the hopes that they will improve the morbidity and mortality associated with CHF.[36,37] NEP degradation is one mechanism by which NPs are cleared. The NEP inhibitors have been shown to heighten NP levels and reduce the proportion of angiotensin II.[38]

Recently, the OVERTURE trial randomly assigned 557 patients with severe CHF, already optimized on standard heart failure treatment, to enalapril or omapatrilat (a combination ACE inhibitor and NEP inhibitor).[36] Mortality and CHF/MI hospitalizations were equivalent in both groups but there was a higher incidence of dizziness, angioedema, and hypotension in the omapatrilat group. The benefit of treatment with NPs or a NEP inhibitor is currently not well established. Further work is needed in this area before any definitive treatment can be put into practice.

BNP may also eventually have a role as a screening tool, given the nature of left ventricular systolic dysfunction (LVSD), which is present in up to 3% of individuals over the age of 65. Because up to 50% of these patients are asymptomatic,[39] current CHF consensus is to treat all symptomatic and asymptomatic patients with an LVEF less than 40%.[40] The data for this were derived mainly from the original Studies of Left Ventricular Dysfunction (SOLVD), showing that treatment with enalapril in asymptomatic LVSD (prevention arm) decreased the composite endpoint of mortality or development of CHF, and in symptomatic LVSD (treatment arm) decreased mortality at 4-year follow-up.[41] A 12-year follow-up of the original SOLVD study (XSOLVD, presented at the European Society of Cardiology Congress 2002), demonstrated a 6% reduction in mortality in the enalapril group (P < .001). Mortality benefits were demonstrated in both prevention and treatment arms.

Many studies have looked at the relationship between LVSD and BNP. A large community-based prospective cohort of 3177 patients from the Framingham study specifically addressed this issue and looked at BNP and NT-ANP for detection of elevated LV mass and LVSD.[42] Discrimination limits based on high specificity (0.95) only identified less than one-third of patients who had elevated LV mass or LVSD, suggesting limited usefulness of NPs as mass screening tools. Currently, the use of BNP as a screening tool for asymptomatic LVSD in the community is not recommended.

 Conclusion

BNP levels that accurately reflect a patient’s neurohormonal and hemodynamic status can be attained by a simple blood test. Even though the role of BNP as a treatment agent and screening tool for asymptomatic LVSD is unclear at this time, a vast amount of data supports its use in differentiating dyspnea in both outpatient and urgent care settings, predicting symptomatic LV dysfunction (both systolic and diastolic), tailoring and optimizing CHF treatment, and evaluating prognosis in both CHF and ACS. Widespread use of BNP measurements is sure to facilitate rapid diagnosis and prompt treatment of CHF, and has the potential to reduce patient morbidity and mortality by guiding optimal treatment of CHF and accurate assessment of CHF prognosis. With these benefits in mind, the Canadian Cardiovascular Society has included a reference to BNP in its 2003 practice guideline (available at www.ccs.ca [10] under CSS 2003 Consensus HF Update, pg. 12-13). A clinical tip advises that the measurement of BNP levels serves as an adjunct to a thorough clinical assessment when the etiology of dyspnea remains unclear (Grade A: Level 1 evidence). And in accordance with this clinical tip, the BC Congestive Heart Failure Advisory Committee has proposed a guideline for care that includes the use of BNP in a diagnostic algorithm.[43]

 Competing interests
None declared.

Table 1. Natriuretic peptides.

 Table 2. Diseases associated with increased natriuretic peptide levels.

 Table 3. Effects of natriuretic peptide (ANP/BNP) activity. 

Jasmine Grewal, MD, Mann Chandavimol, MD, FRCPC, and Andrew Ignaszewski, MD, FRCPC

Dr Grewal is a third-year resident in internal medicine at the University of British Columbia. Dr Chandavimol is co-chief cardiology resident at UBC. Dr Ignaszewski is an associate professor in the Division of Cardiology at UBC and the Healthy Heart Program at St. Paul’s Hospital.