Mitral valve replacement (MVR) is associated with higher mortality and morbidity rates in children than in adults, and the use of heterograft valves has been encouraged. The results of MVR in 56 consecutive patients, aged 2 to 12 years, presenting between January, 1972, and January 1979, were reviewed to test these beliefs. The etiology of mitral valve disease was rheumatic in 46, congenital in eight, and acute bacterial endocarditis in two. All children were seriously disabled (NYHA Classes III and IV). Cardiac catheterization in 36 patients revealed mixed valve disease in 26, pure mitral regurgitation in seven, and pure mitral stenosis in three. Seventeen Starr-Edwards (SE), five Lillehei (L), Björk-Shiley (BS), eight Hancock (H), and 25 Carpentier-Edwards (CE) mitral prosthesis were inserted. Operative mortality was 2% (1 BS) and late mortality was 10% (three SE, one L, two CE). Serious late complications occurred in 30% of survivors, including 11 instances of calcific valve stenosis (five H, six CE), one case of valve thrombosis (1 L), and two embolic episodes (1 SER, 1 H). Survival curves were similar for patients with heterograft and mechanical valves (92% and 77% at 5 years). Event-free curves showed heterograft valves to have a far higher complication rate than mechanical valves (10% complication free at 4 years compared to 84% free at 5 years). Early operative results in children are excellent, and the overall mortality (10%) compares favorably with figures for MVR in adults. However the long-term durability and choice of prostheses remain problematical, since less than 10% of heterograft valves survive beyond 5 years.