Does BNP Augment Acute Decompensated CHF ED Management?

November 2009

Does BNP Augment Acute Decompensated CHF ED Management?

Search Strategy: Feeling lucky, you first turn to Best Bets entering the search term “BNP” and identify two incomplete reviews on this topic – the major problem with Best Bets. Next you turn to the always reliable PUBMED Clinical Queries selecting diagnostics (broad, sensitive search) for “congestive heart failure” and then combine that search with a search for “natriuretic peptid” yielding 1419 citations which you don’t have time to review right now. Therefore, you apply several limitations to these search results (humans, randomized controlled trial) and obtain 65 citations which you are able to quickly peruse to identify three ED-based RCT’s. Reviewing the bibliography for each, you also note the ACEP Clinical Policy for acute heart failure management which you decide to include in your analysis.

Dyspnea and geriatrics. Two of your favorite clinical topics. On this cold November night with flu season’s early arrival, these two issues seem to be the preordained topic of the night. Your last five patients have all had some combination of the intertwined complaints so you fully expect your next case to be the same. You won’t be disappointed.

The pleasant 85-year old doesn’t look his age so you decide to “card” him. Smiling, he shows you his driver’s license as he begins to describe his ailment. For the last 2-days he has noted progressively increased dyspnea associated with a non-productive cough. He has no fever or viral symptoms and notes no orthopnea or edema. In addition to a 20-year history of systolic hypertension, he notes a myocardial infarction “about 25-years ago”. He appears comfortable with the following vital signs (160/90, 64, 14, 36.8, 95% on room air) without any rales, rhonchi, wheezing, gallop, murmur, or edema, but from JAMA’s Rational Clinical Exam series you are aware that diagnostic findings to distinguish CHF can be unreliable. His ECG shows non-specific T-wave flattening in the high-lateral leads (aVL, V6), unchanged compared with a 2004 tracing you find on CLINDESK. While the chest x-ray is pending you contemplate your order set. Recalling that two “opinion leaders” (Bozo and Cletus) had debated BNP-testing to exclude CHF in one of the throw-away journals last year, you decide to quickly evaluate the BNP diagnostic literature yourself.

PICO Question

Population: Emergency Department patients with suspected acute decompensated CHF

Intervention: ED physician aware of BNP result + clinical gestalt

Comparison: ED physician clinical gestalt

Outcome: CHF diagnostic accuracy, admission rates, ED length-of-stay, hospital length-of-stay, ED & re-admission recidivism, mortality


First years: Clinical policy: critical issues in the evaluation and management of adult patients presenting to the emergency department with acute heart failure syndromes, Ann Emerg Med 2007; 49: 627-669. ( [FOCUS ON CRITICAL QUESTION #1 page 629-631

Second years: Improving the diagnosis of acute heart failure using a validated prediction model, J Am Coll Cardiology 2009; 54: 1515-1521. (

Third years: N-terminal pro-brain natriuretic peptide testing in the emergency department: Beneficial effects on hospitalization ,costs, and outcome, Am Heart J 2008; 156: 71-77. (

Fourth years: B-type natriuretic peptid testing, clinical outcome, and health services use in emergency department patients with dyspnea, Ann Intern Med 2009; 150: 365-371. (


Article 1: Clinical Policy: Critical Issues in the Evaluation and Management of Adult Patients Presenting to the Emergency Department with Acute Heart Failure Syndromes, Ann Emerg Med 2007: 49: 627-669

Article 2: Improving the Diagnosis of Acute Heart Failure Using a Validated Prediction Model, Journal of the American College of Cardiology 2009; 54: 1515-1521

Article 3: N-terminal pro–brain natriuretic peptide testing in the emergency department: Beneficial effects on hospitalization, costs, and outcome, American Heart Journal 2008; 156: 71-77

Article 4: B-Type Natriuretic Peptide Testing, Clinical Outcomes, and Health Services Use in Emergency Department Patients with Dyspnea, Annals of Internal Medicine 2009; 150:365-371

BNP Presentation_Nov 2009 Journal Club

CHF Decision Model.xls

Bottom Line

The (Absolutely Not So) Succinct Answer

The diagnosis of congestive heart failure is the second leading reason for hospital length of stay with a 33% re-admission rate and costs $27.9 billion annually as the leading discharge diagnosis in American’s older adults. Even the commonly ordered, highly valued chest x-ray (20% are unremarkable) may not reliably differentiate acute decompensated CHF from non-CHF etiologies of dyspnea. Unfortunately, among undifferentiated ED patients with dyspnea findings from history and physical exam (in isolation) are generally not clinically useful and are often unreliable, (JAMA Rational Clinical Exam):

History of CHF5.80.45
History of MI3.10.69
Edema (history)2.10.64
Fatigue & Weight Gain1.00.99
Pitting Edema (exam)2.30.64
BP > 150 mm Hg1.00.99
Initial clinical judgment4.40.45

Remember that a clinically useful positive-Likelihood Ratio (to rule-in) is > 10, while a clinically useful negative Likelihood Ratio (to rule-out) is < 0.1. Note that by that criteria NONE of these measures rule out CHF. Unfortunately, among the more promising physical exam findings to rule-in CHF, S3 has recently demonstrated less impressive diagnostic accuracy (Collins 2009).

BNP and amino-terminal pro-BNP (NT-proBNP) have therefore been widely touted as a valuable tool to improve the rapid recognition of CHF. However, the few randomized controlled trials assessing the impact of clinician knowledge of BNP (BASEL, IMPROVE-CHF, Rutten 2008, Schneider 2009) on outcomes (diagnostic accuracy, ED length-of-stay, hospital length-of-stay, ancillary testing, mortality) have yielded conflicting results. As a result, clinical researchers continue to debate the usefulness of BNP (Yealy vs. Diercks; Klauer vs. Carpenter) leaving most of us confused and dismayed!

ACEP guidelines recommend that very low (BNP <100 pg/dL, NT-proBNP <300 pg/dL) or very high (BNP >500 pg/dL, NT-proBNP >1,000 pg/dL) levels may be more useful than unaided clinical gestalt to rule out or rule in an acute CHF presentation in ED patients. However, more research is needed to ascertain the external validity of these recommendations to assess the ability and utility of probability based BNP-testing to augment patient-important outcomes above current standard of care. One complicated algebraic model derived in the multi-center Canadian IMPROVE-CHF cohort and validated retrospectively in the United States-based PRIDE cohort improves the diagnostic accuracy for AHF in ED patients with undifferentiated dyspnea. Future studies should validate and/or refine this model while assessing mechanisms for clinical uptake, cost, and impact on patient important outcomes (length-of-stay, morbidity, mortality). For now though, this algebraic equation (computed for you on the attached Excel sheet by simply plugging in AGE, pre-test CHF PROBABILITY, and NT-proBNP level measured) is probably the most clinically useful application of BNP for now.

The RCT’s are inconclusive. A single-center Dutch study suggests that NT- proBNP is cost-effective without increasing patient mortality, although future cost analyses will need to analyze truly undifferentiated dyspnea patients (not those patients sent by their PCP with suspected CHF) in conjunction with clinically useful BNP/NT-proBNP cut points and accepted CHF criterion standards. On the other hand, an Australian investigation indicates that BNP should not be ordered indiscriminately on ED dyspnea patients since it does not alter length of stay, admission rates, or mortality in those with severe dyspnea.

Bottom Line? I intend to order BNP on patients with an uncertain etiology of dyspnea AFTER I obtain HPI, imaging, and assess therapeutic response. When I do order a BNP, I’ll use the attached Excel sheet to yield an estimate of post-test probability so that I know how to interpret my BNP result.