Plasma B-Type Natriuretic Peptide Levels in Ambulatory Patients With Established Chronic Symptomatic Systolic Heart Failure
Background— The diagnostic and prognostic values of plasma B-type natriuretic peptide (BNP) testing are established. However, the range of plasma BNP levels present in the setting of chronic, stable systolic heart failure (HF) is unclear.
Methods and Results— We followed up 558 consecutive ambulatory patients with chronic, stable systolic HF (left ventricular ejection fraction <50%) treated at a specialized outpatient HF clinic between November 2001 and February 2003. Retrospective chart review was performed to determine clinical and functional data at the time of BNP testing (Biosite Triage). The clinical characteristics of patients with plasma BNP levels <100 pg/mL and those with ≥100 pg/mL were compared. In our cohort, 60 patients were considered asymptomatic, and their plasma BNP levels ranged from 5 to 572 pg/mL (median, 147 pg/mL). Of the remaining 498 symptomatic (NYHA functional class II–III) patients, 106 (21.3%) had plasma BNP levels in the “normal” diagnostic range (<100 pg/mL). Patients in this “normal BNP” subgroup were more likely to be younger, to be female, to have nonischemic pathogenesis, and to have better-preserved cardiac and renal function and less likely to have atrial fibrillation.
Conclusions— In the ambulatory care setting, both symptomatic and asymptomatic patients with chronic, stable systolic HF may present with a wide range of plasma BNP levels. In a subset of symptomatic patients (up to 21% in our cohort), plasma BNP levels are below what would be considered “diagnostic” (<100 pg/mL).
Received May 5, 2003; de novo received August 13, 2003; revision received October 24, 2003; accepted October 28, 2003.
Plasma B-type natriuretic peptide (BNP) measurement has been established as a helpful aid in the diagnosis of heart failure (HF).1 Preliminary evidence has demonstrated the potential usefulness of sequential plasma BNP measurements in monitoring patients with HF.2 However, variations in plasma BNP levels have been recognized both in the general population and in patients with chronic HF.3,4 Several confounders, including age, sex, renal function, cardiac rhythm, drug therapy, and body habitus, should be taken into account in the interpretation of plasma BNP results. With increasing clinical experience, we have identified the existence of a subgroup of symptomatic patients with established chronic systolic HF and “normal” levels of plasma BNP <100 pg/mL. The objective of this study is to examine the prevalence, clinical characteristics, and predictors of “normal” plasma BNP levels in symptomatic patients with HF in the clinical practice setting.
See p 2950
We studied consecutive patients with a clinical diagnosis of chronic systolic HF (>3 months’ duration on stable medical therapy, baseline left ventricular ejection fraction <50%, NYHA functional class I–III) in a specialized outpatient HF clinic between November 2001 and February 2003 who underwent plasma BNP testing at the time of their clinic visit. We excluded all patients with a history of congenital heart disease, cardiac transplantation, primary valvular disease, or active ischemia requiring urgent revascularization.
This is a retrospective, cross-sectional study approved by our Institutional Review Board. We performed standardized chart review from paper and electronic medical records to extract demographic, clinical, and echocardiographic data documented within 6 months of BNP testing. As part of routine care, all patients underwent diagnostic evaluation to establish the diagnosis of HF and to determine the severity (by echocardiogram and clinical history) and the underlying pathogenesis (by stress testing with perfusion imaging and/or coronary angiography) of HF. On the basis of these available data, the diagnosis of HF and the underlying pathogenesis were established by an HF specialist who was caring for the patient and was blinded to the study. In most cases, the HF specialist documented the NYHA functional class in the clinical notes. For patients without NYHA documentation in the clinical notes, 2 collaborating investigators of the study who were blinded to the results (J.P.G., M.J.L.) retrospectively reviewed the description of functional status from the clinical notes at the time of BNP testing to assign the corresponding NYHA class. All plasma BNP assays were performed on site during outpatient clinic visits by use of the point-of-care Biosite Triage Assay (Biosite Diagnostics Corp). If multiple tests were performed on the same day, the highest BNP value was recorded. A plasma BNP level of <100 pg/mL was considered to be in the “normal” diagnostic range.1 All patients were treated according to our HF clinical management guidelines.5
We performed standard univariate comparisons of patient characteristics between patients with normal and elevated plasma BNP levels. We then performed forward stepwise multivariable logistic regression to identify predictors for normal plasma BNP levels in symptomatic patients with HF using SPSS version 10.0. Variables included in the analysis were age, sex, diabetes mellitus, hypertension, HF pathogenesis, renal insufficiency, left ventricular ejection fraction, left ventricular end-diastolic dimension, atrial fibrillation, and NYHA functional class.
In our cohort of 662 consecutive patients, 104 met exclusion criteria and were excluded. Among the remaining 558 patients, 166 (30%) had plasma BNP levels of <100 pg/mL. The plasma BNP levels of the 109 asymptomatic patients ranged from 5 to 572 pg/mL (median, 87 pg/mL). Of the remaining 449 symptomatic patients, plasma BNP levels ranged from 6 to >1300 pg/mL (median levels, 165 pg/mL for NYHA II, 409 pg/mL for NYHA III). One hundred six symptomatic patients with chronic HF (21%) had plasma BNP levels in the “normal” diagnostic range (<100 pg/mL). The validity of our findings was confirmed when 45% of “normal” plasma BNP values were rechecked within 6 hours, and all repeat measurements were within 15% variance of the initial value. Patient characteristics of the “normal” and elevated plasma BNP subgroups are illustrated in Table 1.
Patients presenting with “normal” plasma BNP levels were more likely to be younger, to be female, and to have a nonischemic pathogenesis, better-preserved cardiac and renal function, and less atrial fibrillation (Table 1). Treatment regimens were similar between the 2 groups. However, only age, cardiac and renal function, and HF pathogenesis were independent predictors for “normal” plasma BNP level (Table 2). Overall, patients with nonischemic cardiomyopathy had lower mean plasma BNP levels than patients with ischemic cardiomyopathy (267±378 versus 579±452 pg/mL, P<0.01), although the former had less atrial fibrillation (13% versus 26%, respectively) and renal dysfunction (8% versus 19%, respectively) despite similar echocardiographic and functional measurements.
The primary finding in this cross-sectional study is that more than 1 in 5 patients with systolic HF and chronic symptoms seen at our outpatient HF clinic had plasma BNP <100 pg/mL—levels that are often considered “normal” in the contemporary diagnostic use of BNP testing. Therefore, it is prudent to recognize that the usefulness of BNP is specific to the clinical context and the patient population being tested. It is equally important to recognize that our findings in no way contradict the reliable diagnostic role for plasma BNP testing in the setting of acute exacerbation of HF, because mean plasma BNP levels have been shown to be invariably elevated in such patients.1
What are the potential mechanisms for relatively low BNP levels in a cohort of patients with symptomatic chronic stable systolic HF? Plasma BNP is considered a direct counterregulatory response to myocardial stress and increased left ventricular filling pressures in the failing myocardium.6 Significant reduction in plasma BNP levels has been observed after optimal therapy with diuretics and ACE inhibitors, which often reduce filling pressures. The reliable prognostic value of elevated plasma BNP in symptomatic systolic HF has been consistent in various clinical settings,7 and we believe that the prognostic value of plasma BNP in our study cohort will remain robust. It is therefore a logical deduction that patients with HF and relatively low plasma BNP levels may represent a subgroup of patients who simply respond well to optimal therapy.
If BNP generation were predominantly load-dependent with a constant relationship between intracardiac filling pressures and plasma BNP levels, one would postulate that the majority of asymptomatic (NYHA class I) patients with chronic HF would have low plasma BNP levels, and the majority of symptomatic patients may have elevated plasma BNP levels. Yet, our data indicate that the plasma BNP levels in all asymptomatic patients ranged from 5 to 572 pg/mL, whereas >20% of symptomatic patients had plasma BNP levels <100 pg/mL. This wide variability of plasma BNP levels in chronic stable systolic HF patients suggests that plasma BNP is a continuous and not a dichotomous variable. We believe that a better fundamental understanding of the mechanisms underlying this intraindividual heterogeneity of plasma BNP values is necessary before we proceed to developing a BNP- or proBNP-guided strategy for managing HF.
An alternative interpretation of our observations is that disease progression in HF may occur in parallel to the development of load-independent mechanisms of cardiac (or extracardiac) BNP generation. This may explain why patients with similar clinical and echocardiographic severity of HF may have vastly different plasma BNP values. Earlier stages or more reversible forms of disease may be associated with less cardiac (or extracardiac) activation of the “fetal gene program,” leading to relatively lower baseline plasma BNP levels in the compensated state.8 The pathogenesis of HF as well as unknown polymorphisms may also regulate the synthesis, release, and degradation of BNP. Indeed, recent data from the Framingham Study suggested a potential genetic contribution to interindividual variability of plasma BNP levels in the general population.9
This study was limited by its retrospective and cross-sectional nature. The multivariable analysis was based on variables easily available in the clinical setting. Not all patients had available information regarding duration and history of treatment and long-term follow-up, and none had invasive hemodynamic assessment. Echocardiography was not performed simultaneously with BNP testing. Nevertheless, our findings clearly illustrate the importance of interpreting plasma BNP levels within the specific clinical context. Better understanding of the test characteristics is needed before we can effectively use this valuable test to guide therapeutic strategies.
In the ambulatory care setting, both symptomatic and asymptomatic patients with chronic, stable systolic HF may present with a wide range of plasma BNP levels. In a subset of symptomatic patients (21% in our cohort), plasma BNP levels are below what would be considered “diagnostic” (<100 pg/mL). This intraindividual heterogeneity of plasma BNP values may affect our ability to use a single “cutoff” value as a target to guide therapy. Further studies are needed to understand the mechanisms of BNP generation, which may affect the interpretation of plasma BNP levels in specific clinical settings.
We thank Mark Zakrajsek (Biosite Corp) for his assistance.
Dr Francis serves as a consultant to Biosite Diagnostics and Roche Diagnostics. Dr Young has received a research grant from Biosite Diagnostics.
Albert NM, Young JB. Heart failure disease management: a team approach. Cleve Clin J Med. 2001; 68: 53–62.
Lowes BD, Minobe W, Abraham WT, et al. Changes in gene expression in the intact human heart: downregulation of alpha-myosin heavy chain in hypertrophied, failing ventricular myocardium. J Clin Invest. 2001; 100: 2315–2324.
Wang TJ, Larson MG, Levy D, et al. Heritability and genetic linkage of plasma natriuretic peptide levels. Circulation. 2003; 108: 13–16.