The Inability of B-type Natriuretic Protein to Predict Short-Term Risk of Death or Myocardial Infarction in Non-Heart-Failure Patients With Marginally Increased Troponin Levels

Abstract

We want to know whether a low B-type natriuretic peptide (BNP) level, obtained shortly after presentation and independent of information provided by other clinical findings and laboratory tests, would affect management decisions for emergency department (ED) patients with nondiagnostic troponin levels. Previous studies have generally been retrospective or inclusive of patients with heart failure. We prospectively studied patients evaluated for possible acute coronary syndromes who had nondiagnostic levels of serum troponin, nondiagnostic ECGs, and no clinical heart failure within 4 hours of presentation. BNP levels were obtained but results not provided to clinical staff. The primary outcome was the composite of acute myocardial infarction or death within 30 days. The secondary outcome was the composite of the primary outcome, percutaneous coronary intervention, or coronary artery bypass grafting. Almost half of the patients screened for but excluded from the study had known heart failure or a history of heart failure. The resulting cohort was composed of 348 patients, with a median age of 64 years and 51% women. The primary outcome occurred in 16.1% of patients; the secondary outcome, in 27.6%. At a standard cutoff of BNP greater than or equal to 80 pg/mL, the negative predictive value for the primary outcome was 80% (95% confidence interval 73% to 86%). The negative predictive value for the secondary outcome was 69% (95% confidence interval 61% to 75%). Multivariable analyses supported these findings. A single, low BNP level obtained shortly after presentation to the ED could not identify patients at low risk for 30-day acute myocardial infarction or death.