The long-term follow-up of 35 patients with mean age of 74.3 +/- 8 years (64-86) who underwent percutaneous aortic valvuloplasty (PAV) is presented. The mean duration of the follow-up study was 33 months (range 1-48). Global mortality was 42% (14 patients) and was related to post-PAV aortic valvular area (47% mortality in less than 0.7 cm2 area group vs 28% in greater than or equal to 0.7 cm2 group) and left ventricular ejection fraction (67% in EF less than 35% group vs 35% in EF greater than or equal to 35% group). The actuarial probability of remaining alive and free of symptoms and of cardiac surgery or new PAV was 57, 41, 33, and 20% at 1, 2, 3 and 4 years. After PAV clinical improvement was obtained in most of the patients (68%), but only 42% of those with successful dilatation remain asymptomatic after a 24 months period (85 +/- 28 mmHg, p = NS). Transvalvular gradient studied by Doppler decreased immediately after PAV (92.8 +/- 26 mmHg vs 51 +/- 16 mmHg, p less than 0.001), but returned to baseline values after 12 months. These results show that PAV yields a clinical improvement in most of the patients, but this improvement is transitory, does not modify the natural history of the disease and has a high degree of restenosis. Thus, its use must be limited to a reduced group of patients who are not candidates for cardiac surgery.