There are approximately 1 million ED visits annually for CHF in US with ~74% of these patients being admitted
BNP is a peptide released from ventricles when there is increased pressure/volume in the heart
ED physicians are highly accurate in diagnosing CHF in ED without BNP, about 80%, a solid B-.
Study Question:
Does the addition of BNP measurement in ED patients with dyspnea increase diagnostic accuracy of CHF?
Design:
Conducted between April 1999-December 2000
Analyzed 1,586 patients across 7 EDs (5 in US and 2 in Europe). Study was cleverly titled “Breathing Not Properly” (BNP, get it?).
Patients were enrolled if had primary complaint of dyspnea. Patients were excluded if had advanced renal failure (CrCl <15 mL/min), acute MI, or obvious cause of dyspnea (e.g. like a knife in their chest)
After initial assessment of patient including history & physical, EKG, and CXR, ED physicians were asked to estimate clinical probability of CHF, but were blinded to the BNP.
Of note, 511 patients (33.2%) had prior history of CHF which was available to ED physician by patient’s history or available records
30 days after initial ED visit, cases were analyzed by 2 independent cardiologists who were not part of treatment teams to determine if diagnosis of CHF was achieved. They used the Framingham scoring system and NHANES scores as further objective criteria for diagnosing CHF. Patients were categorized as (1) Dyspnea due to CHF or (2) Dyspnea due to non-cardiac cause.
If there was disagreement between the 2 independent cardiologists, then the case was reviewed by the “study end-points committee.” This occurred in 164 of the 1,586 patients reviewed.
Results:
When patients were placed in the high probability (>80% likelihood) of CHF as cause of dyspnea by ED physicians: they had a diagnostic accuracy of 74%, sensitivity 49%, specificity 96%, PPV 91%, NPV 68%, +LR 11.5
In patients thought to be intermediate CHF probability by ED physician (meaning 20-80% probability of CHF):
BNP correctly classified 74% of patients as CHF or not CHF
BNP incorrectly classified 7% of patients
Discussion:
Generally, ED physicians are pretty good at determining if a patient has CHF by clinical impression and the traditional workup of EKG and chest xray alone (specificity 96%)
Despite these solid numbers, the addition of BNP in the work-up can increase the diagnostic accuracy from 74% (ED physicians alone) to 81.5% (ED physicians + BNP)
However, ED physicians are less good at ruling out CHF (sensitivity 49%) but the addition of BNP in this work-up can increase sensitivity to 94%.
Of note, CHF can be a tricky diagnosis and isn’t alway cut and dry (or wet). Even with the availability of echocardiography and response to HF medications, the diagnosis of CHF was uncertain in ~10.7% of patients by cardiologist reviewers.
There have been other studies since this 2002 article that have looked at the use of BNP in heart failure
One paper by Wang et. al (2005) looked at the utility of several clinical findings (e.g. rales, S3, LE edema, etc.) as well as imaging lab/testing (CXR) on diagnosing heart failure, and also found a high LR for a physician’s initial clinical judgment.
There have also been subsequent studies that have shown little impact of BNP on patient centered outcomes (More info at First10EM/BNP)
References:
Blecker S, Ladapo JA, Doran KM, Goldfeld KS, Katz S. Emergency department visits for heart failure and subsequent hospitalization or observation unit admission. Am Heart J. 2014;168(6):901-8.e1.
Mccullough PA, Nowak RM, Mccord J, et al. B-type natriuretic peptide and clinical judgment in emergency diagnosis of heart failure: analysis from Breathing Not Properly (BNP) Multinational Study. Circulation. 2002;106(4):416-22.
Morgenstern J. BNP in the emergency department: The evidence. First 10 EM. 2018. first10em.com/bnp.