Prognostic information provided by serial measurements of brain natriuretic peptide in heart failure

Background: Brain natriuretic peptide (BNP) levels predict prognosis in heart failure patients. We aimed to evaluate if serial measurements of BNP can give additional prognostic information. Methods: Eighty-four patients with systolic dysfunction had two measurements of BNP with an interval of 8 to 12 months and were followed in order to register the occurrence of death. The study was observational and prospectively designed. During follow-up, patients were treated according to state of the art. Physicians were kept blind to BNP levels. Results: The median follow-up was 1190 days. The median initial BNP level was 260.4 pg/ml and decreased to 123 pg/ml in the second measurement (P=0.001). The decrease in BNP was significantly associated with ACE-i dosage and with the use of a beta-blocker. All-cause mortality was 20.2%. Patients whose initial BNP level was above the median had a significantly higher hazard of dying (HR 2.96, 95% CI 1.06-8.26). The same was observed for those whose BNP increased between the first and the second measurement (HR 2.64, 95% CI 1.00-7.00). In multivariable analysis, baseline BNP above the median and increasing BNP were associated with shorter survival. Conclusions: Higher baseline BNP and the increasing levels during follow-up were independently associated with mortality. The decrease in BNP levels was proportional to ACE-i dosage and larger among patients on beta-blockers. These results confirm the prognostic information provided by BNP determination and suggest that serial measurements give additional prognostic information. (C) 2003 Elsevier Ireland Ltd. All rights reserved.