Patients with aortic stenosis (AS) may remain asymptomatic with good prognoses for many years but have poor prognoses once they develop symptoms. Because the presence of symptoms is subjective, B-type natriuretic peptide (BNP) may provide a more objective indication of the prognoses of patients with AS. We evaluated 124 patients with AS (valve area <1.2 cm(2)) with clinical evaluation, Doppler echocardiography, and BNP assessment and obtained up to 2 years of follow-up without valve replacement. Patients with syncope, angina, and/or heart failure were considered to have symptoms. The 24 patients without symptoms had lower BNP levels (187 +/- 193 pg/ml) than the 100 patients with symptoms (930 +/- 928 pg/ml, p <0.001). BNP indicated symptom status, with an area under the receiver-operating characteristic curve of 0.87 (p <0.001). The optimal discrimination of symptoms occurred with BNP >190 pg/ml. Survival was significantly influenced by the presence of symptoms (relative risk [RR] 7.5, p <0.01) and BNP tertile (RR 2.9, p <0.001). The 1-year mortality rate without surgery was 6% for BNP <296 pg/ml, 34% for BNP 296 to 819 pg/ml, and 60% for BNP >819 pg/ml. No patients with BNP <100 pg/ml died. The combination of BNP and symptoms provided a better prediction of survival than symptoms alone (chi-square 13.6, p <0.001). BNP significantly (RR 2.8, p <0.01) influenced survival after correction for other univariate predictors (coronary artery disease, symptoms, functional class, ejection fraction, and aortic valve area). In conclusion, elevated BNP indicates progressively worse survival in patients with AS treated medically. Thus, the measurement of BNP supplements the evaluation of symptoms in determining the prognoses of patients with AS.