Utility of brain natriuritic peptide as a diagnostic tool for congestive heart failure in the elderly.

We studied the performance characteristics of B-type natriuretic peptide (BNP) as a diagnostic test for congestive heart failure in the elderly dyspneic population. In a retrospective chart review study, dyspneic patients who had a BNP level drawn were included. To diagnose congestive heart failure (CHF), the Framingham Criteria were used. To diagnose pneumonia or a lower respiratory tract infection, the consensus development conference Criteria for Pneumonia/Lower Respiratory Tract Infection were used. Based on the criteria satisfied, the patients were categorized into one of 4 groups: group 1, pneumonia/lower respiratory tract infection; group 2, CHF; group 3, both; group 4, neither. Sensitivity, specificity, and positive and negative likelihood ratios were calculated for various BNP reference limits from 100 pg/ml upwards in increments of 100. A total of 70 patients (mean age 76.5) presenting with dyspnea were included in the study. Forty-six were females. The mean (+/-SD) BNP level for group 1 (n = 13) was 273 (+/-360) pg/ml, for group 2 (n = 30) it was 1394 (+/-934) pg/ml, for group 3 (n = 17) it was 1138 (+/-842) pg/ml, and for group 4 (n = 10) it was 403 (+/-362) pg/ml. Forty-seven patients (groups 2 and 3) met the Framingham criteria (CHF+). The other 23 (groups 1 and 4) did not (CHF-). The sensitivity and specificity of BNP for CHF at a cutoff of 100 pg/ml was 96% and 26%, respectively. The sensitivity (87%) and specificity (74%) were optimal at a cutoff of 400 pg/ml. Our study indicates that the specificity of a BNP level of > or = 100 pg/ml for diagnosing CHF in the elderly is poor. Our data suggest an optimal BNP value of > or = 400 pg/ml. Elderly patients frequently have multiple etiologies contributing to dyspnea. In our study, one-fourth of the patients satisfied the criteria for a dual diagnosis of CHF and pneumonia.