BACKGROUND: The prognostic value of BNP in acute coronary syndrome (ACS) has been repeatedly assessed, but not completely well established. Literature data for establishing the best time for assessing BNP, be it on hospital admission or after coronary intervention, are controversial. OBJECTIVE: To analyze BNP in non-ST segment elevation ACS (NSTE-ACS) in the long term, and to assess the association between BNP (pg/ml), death, coronary anatomy, and TIMI Risk Score. METHODS: Forty patients with NSTE-ACS and troponin>0.50 ng/ml had their BNP levels measured on admission and 96 hours after, and were followed up for four years. The difference between the two measures was assessed by use of Wilcoxon test (p<0.05). The ROC curve was used to evaluate 96-hour BNP accuracy as a death predictor, and logistic regression was used to assess a possible confounding factor among 96-hour BNP, age, and outcome. RESULTS: There was an increase in the 96-hour BNP (148 on admission vs. 267 after 96 hours; p=0.04). Thirteen patients died. For the 300 pg/ml cutoff, 96-hour BNP was a death predictor (sensitivity, 92.30%; specificity, 77.80%; positive predictive value, 66.70%; negative predictive value, 95.50%). The area under the ROC curve was 0.92. A 7.4-time increase in the relative risk of death in four years was observed with a 96-hour BNP> 300 pg/ml (95% CI 1.90 a 29.30 p<0.01). An association between 96-hour BNP and TIMI Risk Score was observed (p<0.01). An association was observed between the increase in 96-hour BNP and multivessel disease (p=0.02). CONCLUSION: In NSTE-ACS with positive troponin, 96-hour BNP can be a tool for risk stratification. (Arq Bras Cardiol. 2012; [online].ahead print, PP.0-0).