To identify predictors of long-term outcome after balloon aortic valvuloplasty, we analyzed data on 674 adults (mean age, 78 +/- 9 years; 56% were women) undergoing this procedure at 24 clinical centers who had a mean initial increase in aortic valve area of 0.3 cm2. Baseline data included clinical, echocardiographic, and catheterization variables. Follow-up data included mortality, cause of death, rehospitalization, 6-month echocardiography, and functional status. Kaplan-Meier curves and log-rank tests were used to evaluate survival in subgroups. Multivariate Cox regression models were used to identify independent predictors of survival. Overall survival was 55% at 1 year, 35% at 2 years, and 23% at 3 years, with the majority of deaths (70%) classified as cardiac by an independent review committee. Rehospitalization was common (64%), although 61% of survivors at 2 years reported improved symptoms. Echocardiography at 6 months (n = 115) showed restenosis from the postprocedural valve area of 0.78 +/- 0.31 cm2 to 0.65 +/- 0.25 cm2 (P < .0001). With stepwise multivariate analysis, sequentially adding clinical, echocardiographic, and catheterization variables, the overall model identified independent predictors of survival as baseline functional status, baseline cardiac output, renal function, cachexia, female gender, left ventricular systolic function, and mitral regurgitation. Baseline and postprocedural variables were examined to identify which subgroup of patients has the best outcome after aortic valvuloplasty. A "lower-risk" subgroup (28% of the study population), defined by normal left ventricular systolic function and mild clinical functional limitation, had a 3-year survival of 36% compared with 17% in the remainder of the study group. Long-term survival after balloon aortic valvuloplasty is poor with 1- and 3-year survival rates of 55% and 23%, respectively. Although survivors report fewer symptoms, early restenosis and recurrent hospitalization are common.