Between 1984 and 1997, a total of 444 children (< 15 years) received an orthotopic liver transplant (OLT) at Saint-Luc University Clinics. Biliary atresia constituted the indication for OLT in 304 cases (68%). Median age (range) at OLT was 2.1 years (0.3-14.5). 177 children (40%) received a whole liver graft, 184 (41%) a reduced-size graft, 26 (6%) a split liver graft, whereas a living-related donor graft was used in 57 children (13%). Most grafts were ABO-identical or -compatible, in 395 cases (89%) and 40 (9%), respectively. Overall actuarial patient survivals at one and five years were 85% and 81%. The overall retransplantation rate was 19%. The results of uni- and multivariate statistical analyses showed a significant impact of year of transplantation (learning curve effect), with a 15% improvement of patient survival between the 1984-7 and the 1995-7 periods (p < 0.002). Results differed according to the indications for OLT, the best survivals being recorded for familial cholestasis, the worst for liver tumor (p = 0.004). Five year patient survival was significantly better after elective OLT (82%), when compared to highly urgent OLT (63%) (p < 0.001). Patient survivals were comparable in the children receiving primary cyclosporin-A microemulsion or tacrolimus immunosuppression, which were significantly higher than in the historical group treated with cyclosporin-A. No impact of the age at OLT, type of graft and ischemic time could be reported. In conclusion, this series illustrates the progressive improvements introduced in our pediatric liver transplant program between 1984 and 1997, including the technical variants allowing pediatric OLT using adult donors as well as the introduction of new immunosuppressive strategies.