By Van Selby, MD

Assistant Professor of Medicine, University of California, San Francisco Cardiology Division, Advanced Heart Failure Section

Dr. Selby reports no financial relationships relevant to this field of study.

SYNOPSIS: Among patients ≥ 80 years of age presenting with acute dyspnea, B-type natriuretic peptide level was not useful for differentiating cardiac vs. respiratory etiologies when added to a model of clinical predictors.

SOURCE: Plichart M, Orvoën G, Jourdain P, et al. Brain natriuretic peptide usefulness in very elderly dyspnoeic patients: The BED study. Eur J Heart Fail 2017;19:540-548.

B-type natriuretic peptide (BNP) is used frequently to identify cardiac vs. respiratory etiologies in patients presenting with dyspnea. However, many factors influence BNP level, limiting its usefulness in certain populations. The diagnostic accuracy of BNP concentration in the assessment of dyspnea in very elderly (> 80 years of age) patients has not been studied adequately. The authors of the BNP Usefulness in Elderly Dyspneic Patients (BED) study enrolled 383 patients ≥ 80 years of age who were evaluated for acute dyspnea. All patients had BNP levels measured in addition to other clinical testing, including echocardiography. Independent cardiologists blinded to the BNP result evaluated each case according to standard guidelines to determine whether the cause of dyspnea was cardiac vs. respiratory.

Sixty-two percent of patients had cardiac dyspnea, and 38% had respiratory dyspnea. BNP levels were significantly higher among patients with cardiac vs. respiratory etiologies (median level 385.5 vs. 172.0 ng/L; P < 0.001). However, BNP was not a good test for discriminating cardiac vs. respiratory etiologies (area under the curve [AUC] = 0.68). The authors created a multivariate model of clinical predictors that discriminated cardiac vs. respiratory dyspnea with high accuracy (AUC = 0.915). When added to this model, BNP was independently associated with cardiac etiology but did not improve the AUC significantly (P = 0.16). No single BNP cutoff value was found that diagnosed or excluded cardiac etiologies with adequate reliability. Clinical predictors associated with a cardiac etiology included higher body mass index, history of heart failure (HF), X-ray findings consistent with pulmonary edema, and lower ejection fraction. History of chronic respiratory disease, rhonchi, and higher white blood cell count all were associated with a respiratory etiology. The authors concluded that BNP is not a useful diagnostic tool among very elderly patients with acute dyspnea, but noted that it may be of interest for prognosis in heart failure.


Both American and European guidelines give a class IA recommendation for the use of natriuretic peptide biomarkers such as BNP to support or exclude heart failure in patients presenting with dyspnea. A cutoff level of 100 ng/L is recommended often to rule out cardiac dyspnea, regardless of age. However, guidelines also acknowledge that comorbidities can influence BNP levels and recommend they be taken into account when interpreting a given patient’s BNP. Elderly patients are more likely to present with comorbidities, and age alone can influence BNP level. Therefore, it is important to understand how this changes the utility of BNP as a diagnostic test.

The Breathing Not Properly Multinational Study was one of the first and largest to evaluate the utility of BNP measurement in patients presenting with dyspnea. A post-hoc subanalysis of this study found that BNP was a weaker predictor in subjects > 70 years of age. BNP levels tended to be higher in elderly subjects, and this decreased the specificity for any given cutpoint. Several smaller studies subsequently found that BNP remained useful in elderly patients, though higher cut-points were needed.

Plichard et al now add the largest study to date specifically evaluating the utility of BNP in elderly patients. Their primary finding is that a clinical model consisting of age, body mass index, gender, and other covariates discriminated cardiac vs. respiratory etiologies with high accuracy. Although BNP levels were higher in patients with cardiac dyspnea, adding BNP levels to their clinical model did not improve the discriminative ability of their model significantly.

Although the findings of this study weaken enthusiasm for BNP use in elderly patients, there are a few important aspects to keep in mind. The baseline multivariable model already had an impressive ability to discriminate cardiac vs. respiratory dyspnea, and it would be difficult for the addition of BNP to improve on this model significantly. The model these authors used involved many variables and was not practical for routine use in clinical settings. BNP levels were clearly higher in patients with cardiac etiologies, and perhaps if a simpler, more realistic baseline model were used, then adding BNP would improve the diagnostic accuracy. Another limitation is the lack of a “gold standard” for the diagnosis of cardiac (as opposed to respiratory) dyspnea. Instead of abandoning BNP testing in elderly patients, it may be better to continue using it to differentiate cardiac vs. respiratory causes of dyspnea while taking into consideration its limitations. Elderly patients have higher BNP levels than younger patients, and a higher BNP level cannot “rule in” cardiac dyspnea with the same accuracy as it can in younger patients. BNP levels should be used along with all other available clinical data when determining the etiology of acute dyspnea in elderly patients.