The purpose of the study is to determine if a series of BNP blood tests performed on patients who present to the Emergency Department with congestive heart failure (CHF) can predict which patients may have adverse outcomes. If the BNP is shown to be predictive of bad outcomes in certain patients, those patients might receive more intensive therapy early to prevent such outcomes. This was a prospective trial enrolling patients who presented to the ED and were diagnosed with heart failure. Subjects had a blood test for BNP, which is elevated in the presence of heart failure, collected twelve hours after their initial clinical BNP was obtained in the ED. Demographics, history, length of hospital stay, and other approved data were collected. At 30 days and 6 months after discharge, a follow up call was made to determine if the subject had required additional emergency care, had been admitted to a hospital, or had died during that period of time.

BNP is a 32-amino acid peptide secreted almost exclusively by the ventricles under conditions of ventricular wall strain.Multiple studies have examined the utility of BNP as a marker of decompensated heart failure, both in the inpatient and outpatient settings. Several studies have examined the use of a single BNP measure to predict adverse outcomes in CHF patients. Typical study designs have utilized CHF specialty clinics or patients admitted for congestive heart failure (CHF) exacerbation as the study population. Very few studies have utilized BNP in the acute care setting; these studies have also used BNP as a single measure. As the serum half-life of BNP is approximately 22 minutes,BNP levels will respond to measures decreasing ventricular wall strain, such that a decreasing BNP level should be diagnostic of improved CHF. It is our hypothesis that Emergency Department patients with acute CHF exacerbation and a decreasing serum BNP level after 12 hours of therapy will demonstrate an improved outcome when compared to patients whose BNP does not decrease.

Our goal is to establish the utility of serial BNP levels during the intensive initial management of acute CHF. One and six month follow-up patient outcomes will be collected and compared against the serial BNP measurements while the patient was in the acute phase of treatment. It is our hypothesis that serial brain natriuretic peptide (BNP) levels compared at 0 and 12 hours for research purposes will be predictive of a combined endpoint of death, readmission, or repeat Emergency Department visit for recurrent CHF in patients presenting to the Emergency Department with an acute exacerbation of CHF.

To date, no data has been published regarding the utility of serial BNP measurements in the observation unit management of CHF. Our study proposes to compare outcomes in patients managed in the Clinical Decision Unit (CDU), a six bed observation unit in our Emergency Department, against patients admitted for CHF exacerbation. The primary outcome of the study will be the combined endpoint of death, readmission, or repeat Emergency Department visit for recurrent CHF at 30 days and 6 months. For inpatients, and those patients admitted from the CDU to the inpatient service, additional outcome measures will be adverse outcomes (intubation, balloon pump, death, etc) during the index hospitalization. As the clinicians managing the patient will not have access to the 12 hour BNP level, the measurement of BNP itself will not be a source of bias. Our secondary hypothesis is that the occurrence of the combined endpoint will be independent of treatment location (CDU vs inpatient) as opposed to responder vs. non-responder status.

Inclusion Criteria:

• age18 or older
• primary diagnosis of CHF in the Emergency Department
• admission to the hospital or transfer to the Observation Unit

Exclusion Criteria:

• minors, prisoners,pregnant women,unable to provide consent
• unstable angina or acute myocardial infarction in the ED
• dialysis dependent patients
• use of nesiritide as a therapy in the ED