The combined utility of an S3 heart sound and B-type natriuretic peptide levels in emergency department patients with dyspnea.

BACKGROUND: Emergency department (ED) patients with undifferentiated dyspnea are a diagnostic dilemma. We hypothesized that electronic detection of an S3 would be more accurate in determining decompensated heart failure than physician auscultation, and that combining electronic heart sounds with B-type natriuretic peptide (BNP) would provide additional decision making information to the emergency physician, especially in the BNP indeterminate range (100-500 pg/mL). METHODS AND RESULTS: We collected demographic, clinical, and laboratory data in a convenience sample of ED patients presenting with signs or symptoms of acute decompensated heart failure between September 2003 and June 2004. The electronic presence of an S3 or S4 was determined using the Audicor system, a validated device that algorithmically detects S3 and S4 heart sounds. Two independent reviewers determined the presence or absence of acute decompensated heart failure (primary HF) based on chart review, while blinded to BNP and Audicor results. Test characteristics were determined with 95% confidence intervals. Of 422 enrolled patients, 343 had complete data and were included in the final analysis. Median age was 61 years, 54% were female, and 48% were white. The sensitivity, specificity, positive and negative predictive value, and diagnostic accuracy of an electronic S3 for primary HF were 34% (26% to 43%), 93% (89% to 96%), 66% (57% to 74%), 7% (4% to 11%), and 70% (65% to 75%) and for physician auscultation were 16% (11% to 24%), 97% (93% to 99%), 84% (76% to 89%), 3% (2% to 7%), and 66% (61% to 71%). The addition of an Audicor S3 to intermediate BNP levels improved the positive LR from 1.3 to 2.9; the positive predictive value from 53% to 80%. CONCLUSION: An S3 is highly specific for primary HF and it is ideally suited for use in combination with BNP to improve diagnostic accuracy in ED patients with dyspnea of unclear etiology.

Duke Scholars

Author Christopher Lindsell Biostatistics & Bioinformatics, Division of Biostatistics