As interest increase in the Ross procedure performed as a therapeutic option for children with congenital aortic valvar stenosis, it becomes increasinly important to know the late results of aortic valvotomy in this population. We have therefore examined retrospectively the medical records of 121 consecutives survivors undergoing aortic valvotomy before 10 years of age between 1974 and 1992. The mean age at the first valvotomy was 29 months, with a range from 3 days to 10 years. The mean duration of follow up was 9.4 years, with a range from 1.6 to 22 years. Fifteen patients (12.3%; 70% CL: 10-16) died: 9 following reoperation, and 6 late after surgery. Death was related to the hearts in 86% of cases. The actuarial survival rate was 79% (70% CL: 72/84) at 10 years. Young age at the first valvotomy, and the number of procedures, emerged as risk factors of secondary mortality. Reoperations on the aortic valve, 73 in all, were required in 56 patients. The second procedure was done after a mean interval of 6 years, with a range from 1 day to 18 years. This was for restenosis in three-quarters of the cases. The aortic valve was replaced in 30 patients, at a mean of 9 years, and with a range from 9 months to 18 years, after the first procedure. The survival without replacement at 20 years was 29% (70% CL: 15-49). No factor was identified with a relationship either to reoperation or valvar replacement. Long term results after aortic valvotomy, therefore, show a high late mortality, frequent reinterventions, and an almost inescapable eventual need for valvar replacement. The ongoing use of the Ross operation is justified, even if longterm studies in children are still needed to validate its use.