Repeat orthotopic liver transplantation (ReOLT) is controversial because of limited donor organ availability and increasing health care costs. The purpose of this study is to analyse and compare the outcome of reOLT in the 1990s and the 1980s. Prospective data of 1077 adult OLT from the Liver Unit database were used for the study. The log-rank test was used for statistical analysis. Between January 1982 and December 1996, a total of 1077 adult OLTs were performed including 107 reOLTs. The proportion of retransplants decreased from 13% in the 1980s to 9% in the 1990s. There was a significant improvement in outcome; the overall 1-year graft and patient survival for reOLT was 60% and 74% in the 1990s compared to 29% (P < 0.0001) and 51% (P < 0.0001) in the 1980s. In the second half of the study between January 1990 and December 1996, 732 adult OLTs were undertaken including 70 (9%) reOLTs which consisted of 62 second, 7 third and 1 fourth grafts. The main indications for retransplantation were chronic rejection (31%), hepatic artery thrombosis (30%), primary non-function (16%), ischaemic injury (11%), recurrent disease (6%) and biliary complications (6%). During this period, the 1-year graft survival for all reOLTs was significantly lower than for primary OLTs (67% vs 78%, P < 0.001). The timing of reOLT was found to be associated with graft survival; 1-year graft survival for early reOLT (< 30 days) was 50% compared to 73% for late reOLT (P < 0.001). The worse outcome associated with early reOLT is explained by the poor preoperative medical condition of patients who were retransplanted from intensive care. Subgroup analysis of indications for reOLT revealed 1-year graft survival of 81% for late vascular complications, 75% for early vascular complications, 69% for chronic rejection and 30% for primary non-function. One-year graft survival rates for third and fourth grafts were 42% and 0%, respectively. Graft survival and resource utilisation in patients who received a late regraft for the first time is now comparable to that for primary OLT. The favourable overall results should not preclude this group of patients from consideration for reOLT.