We investigated whether N-terminal pro-brain natriuretic peptide (NT-pro-BNP) that was measured on admission in patients with acute myocardial infarction (AMI) predicts the efficacy of reperfusion or whether NT-pro-BNP provides prognostic information independent of infarct size as estimated by single-photon emission computed tomographic scintigraphy. The study included 174 patients with ST-segment elevation AMI who were admitted within 24 hours of pain onset. NT-pro-BNP level was measured on admission. Paired scintigraphic studies (before and 7 to 14 days after reperfusion) were performed to assess infarct size and define myocardial salvage. One-year clinical follow-up was assessed. Patients were categorized into the high NT-pro-BNP group (57 patients in the upper tertile of NT-pro-BNP) and low NT-pro-BNP group (117 patients in the middle and lower tertiles of NT-pro-BNP). Initial median perfusion defect was 35.0% (interquartile rage 20.0 to 53.0%) of the left ventricle in the high NT-pro-BNP group versus 19.0% (interquartile range 10.0 to 32.2) of the left ventricle in the low NT-pro-BNP group (p <0.001). Median salvage index was 0.36 (interquartile range 0.16 to 0.86) in the high NT-pro-BNP group versus 0.53 (interquartile range 0.31 to 0.75) in the low NT-pro-BNP group (p = 0.22). After adjustment in Cox's proportional hazards model, NT-pro-BNP remained an independent correlate of 1-year mortality (adjusted hazard ratio 2.31, 95% confidence interval 1.09 to 4.89, p = 0.03, high vs low NT-pro-BNP group). In conclusion, NT-pro-BNP measured on admission in patients with AMI correlates with scintigraphic area at risk and predicts prognosis but does not predict the efficacy of mechanical reperfusion by stenting or angioplasty.