Secondary Logo

Journal Logo

Myths in Emergency Medicine: BNP's Use Based on Superstition, Not Sound Reasoning

Spiegel, Rory, MD

doi: 10.1097/01.EEM.0000542262.32745.46
Myths in Emergency Medicine

Dr. Spiegelis a clinical instructor in emergency medicine and a critical care fellow in the division of pulmonary and critical care medicine at the University of Maryland Medical Center. Visit his blog at, follow him on Twitter @emnerd_, and read his past articles at



Superstitious beliefs are forged in a caldron of fear, uncertainty, and the inability to control one's environment, so it should come as no surprise that we as emergency physicians have adopted a few superstitious mannerisms, bearing them as protective talismans in the hope they will ward off evil.

No chief complaint is fraught with more uncertainty than a patient presenting after a syncopal episode. A generalized decrease in blood supply to one's brain long enough to cause a transient loss of consciousness should concern us. Longitudinal data suggest that despite a negative initial diagnostic workup, a clinically significant portion of these patients go on to have a serious adverse event in the first few days following their ED presentation. (N Engl J Med 1983;309[4]:197.)

We have proven ourselves incapable of identifying which portion of these patients will have negative outcomes despite decades of research. (Ann Emerg Med 2006;47[5]:448; Ann Emerg Med 2008;52[2]:151.) It was only a matter of time before we turned toward biomarkers, such as natriuretic peptides, hoping they would provide us the comfort in the workup of these patients that up until now has eluded us.

Reed, et al., conducted the ROSE (Risk Stratification of Syncope in the Emergency Department) study to evaluate the performance of BNP alone and in cooperation with a decision tool. (J Am Coll Cardiol 2010;55[8]:713.) They enrolled adult patients (16 years or older) presenting to a single ED with syncope. Research assistants collected 32 predetermined variables during the initial phase, and examined their ability to predict clinical outcomes (the composite of all-cause mortality, acute myocardial infarction, life-threatening arrhythmia, the decision to implant a pacemaker, a cardiac or pulmonary embolus, cerebrovascular accident, intracranial hemorrhage, subarachnoid hemorrhage, an acute surgical procedure, and endoscopic intervention).

The authors assembled a clinical decision rule using a 529-patient derivation cohort that included seven variables (BNP level ≥300 pg/ml, bradycardia ≤50 bpm, rectal examination showing fecal occult blood, hemoglobin ≤90 g/l, chest pain associated with syncope, ECG showing Q wave, and a saturation ≤94% on room air). They then enrolled an additional 538 patients in the hope of validating their original findings.

Back to Top | Article Outline

Perfection Improbable

The ROSE rule performed admirably in its derivation cohort, boasting a sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios of 92.5%, 73.8%, 22.4% and 99.2%, and 3.5 and 0.1, respectively. The area under the receiver-operator characteristic (ROC) curve for the ROSE rule in the derivation cohort was 0.83.

The authors reported slightly less ideal numbers when they validated them in the same ED by the same study group. The sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios were 87.2%, 65.5%, 16.5% and 98.5%, and 2.5 and 0.2, respectively. The area under the ROC curve fell to 0.76 (95% CI: 0.70 to 0.83).

The BNP fared even worse when examined in isolation. At a cutoff of 300, the natriuretic peptide identified only 41 percent of the patients who experienced a composite event at 30 days.

Despite the authors' optimistic interpretation, the data in no way support the use of the ROSE clinical decision rule or BNP in isolation when determining which patients are safe for discharge from the ED. Even if you believe the numbers presented by the authors, this rule is nowhere near sensitive enough to use in clinical practice and in fact performs worse than the San Francisco Syncope Rule or unstructured clinical judgment, equally maligned for their lack of sensitivity. (Am J Med 2014;127[11]:1126.e13.) But, more importantly, this is a small single-center derivation and validation cohort. It is highly likely its diagnostic performance would be even worse when applied to an external cohort.

The likelihood of developing a strategy that achieves diagnostic perfection is improbable, given the multidimensional complexity of a presenting complaint such as syncope. Our current approach of admitting the vast majority of these patients is not sustainable, but simply offloading the diagnostic burden to a poorly performing biomarker is not the answer. The data supporting the use of BNP and the ROSE criteria are incomplete at best, but we can confidently say that they are incapable of safely differentiating the patients who will inevitably have a bad outcome following a syncopal event even with this limited perspective.

We also have no idea what effect the implementation of such a strategy would have when tossed haphazardly in the clinical arena. A quantifiable promise of certitude in the face of the uncertainty that we confront daily is understandably alluring. Their continued use is based more on superstition than sound scientific reasoning when those assurances fail to predict the outcomes they promise.

Share this article on Twitter and Facebook.

Access the links in EMN by reading this on our website or in our free iPad app, both available at

Comments? Write to us at

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.