To assess late results of aortic homograft and pulmonary autograft valves implanted into the left ventricular outflow tract of children, we reviewed the case histories of 146 patients 18 years of age or younger who underwent aortic valve or root replacement between November 1964 and April 1990. One hundred three patients (mean, 12 +/- 3.9 years) received an aortic homograft and 43 (mean, 14 +/- 4.1 years) had their own pulmonary valve transferred to the aortic position. There were 54 valve and 49 root replacements with homografts and 36 valve and seven root replacements with autografts. Hospital mortality rate was 15.5% (16 patients) in the homograft group and 11.6% (five patients) in the autograft group. Survivors were followed up for a total of 867 (homograft) and 297 (autograft) patient-years. The late mortality rate was 16.7% (1.9% per patient-year) for patients with homografts and 13.2% (4.4% per patient-year) for patients with autografts, whereas the incidence for reoperation per patient-year was 2.9% and 2.0%, respectively. At 15 years actuarial rates for homografts and autografts for freedom from reoperation were 54% +/- 8.1% and 68% +/- 11.1%; freedom from endocarditis, 97% +/- 2.4% and 75% +/- 10.2%; and freedom from any complication, 41% +/- 6.5% and 50% +/- 10.3%. Valve degeneration occurred in 19 homografts (2.2% per patient-year), whereas there was no definite instance of primary tissue failure among the pulmonary autografts. This experience would indicate that either the homograft or the autograft valve can be used with acceptable results in children. However, the pulmonary autograft gives better long-term performance and, if growth potential is realized, may be the ideal valve substitute in children.