BACKGROUND Survival after liver transplantation has improved significantly over the last decade with pediatric recipients faring better than adults. The 20-year experience of pediatric liver transplantation at Children's Hospital of Pittsburgh is reported in terms of patient survival; graft survival in relation to age, gender, and immunosuppressive protocols; causes of death; and indications for retransplantation. METHODS From March 1981 to April 1998, 808 children received liver transplants at Children's Hospital of Pittsburgh. All patients were followed until March 2001, with a mean follow-up of 12.2+/-3.9 years (median=12.6; range=2.9-20). There were 405 female (50.2%) and 403 male (49.8%) pediatric recipients. Mean age at transplant was 5.3+/-4.9 years (mean=3.3; range 0.04-17.95), with 285 children (25.3%) being less than 2 years of age at transplant. Cyclosporine (CsA)-based immunosuppression was used before November 1989 in 482 children (50.7%), and the subsequent 326 recipients (40.3%) were treated with tacrolimus-based immunosuppression. Actuarial survival was calculated using the Kaplan-Meier statistical method. Differences in survival were calculated by log-rank analysis. RESULTS Overall patient survival at 1, 5, 10, 15, and 20 years was 77.1%, 72.6%, 69.4%, 65.8%, and 64.4%, respectively. There was no difference in survival for male or female patients at any time point. At up to 10 years posttransplant, the survival for children greater than 2 years of age (79.5%, 75.7%, and 71.6% at 1, 5, and 10 years, respectively) was slightly higher than those at less than 2 years of age (72.6%, 66.9%, and 65.3% at 1, 5, and 10 years, respectively). However, at 15 and 20 years posttransplant, survival rates were similar (>2 years=67.3% and 65.8%; <2 years=64.1% and 64.1%). A significant difference in survival was seen in CsA-based immunosuppression (71.2%, 68.1%, 65.4%, and 61%) versus tacrolimus-based immunosuppression (85.8%, 84.7%, 83.3%, and 82.9%) at 1, 3, 5, and 10 years, respectively (P=0.0001). The maximum difference in survival was noted in the first 3 months between CsA and tacrolimus; thus, indicating there may have been other factors (nonimmunological factors) involved in terms of donor and recipient selection and technical issues. The mean annual death rate beyond 2 years posttransplant was 0.47%, with the mean annual death rate for patients who received tacrolimus-based immunosuppression being significantly lower than those who received CsA-based immunosuppression (0.14% vs. 0.8%; P=0.001). The most common etiologies of graft loss were hepatic artery thrombosis (33.4%), acute or chronic rejection (26.6%), and primary nonfunction (16.7%). Of note, retransplantation for graft loss because of acute or chronic rejection occurred only in those patients who received CsA-based immuno-suppression. CONCLUSIONS The overall 20-year actuarial survival for pediatric liver transplantation is 64%. Survival has increased by 20% in the last 12 years with tacrolimus-based immunosuppression. Although this improvement may be the result of several factors, retransplantation as a result of acute or chronic rejection has been completely eliminated in patients treated with tacrolimus.