Efficacy of MELD score in predicting survival after liver retransplantation.

OBJECTIVE: We retrospectively investigated the efficacy of the MELD score to predict the outcome of liver retransplantation and serve as selection criteria. MATERIALS AND METHODS: From 1987 to 2003, the 765 liver transplantations included 87 patients (11.4%) who received a second graft. In addition to graft and patient survivals, ROC curves were used to establish the best MELD score to select cases with poor outcomes. RESULTS: Indications for retransplantation were: 38 (43.7%) surgical complications; 12 (13.8%) chronic rejections; 15 (17.2%) disease recurrences; and 22 (15.3%) primary graft nonfunction. Overall patient survivals at 1, 3, and 5 years were 62.4%, 50.7%, and 49.1%, respectively. A MELD score of 25, calculated by ROC curves, significantly predicted graft and patient survival (44.2% vs 22.5%, P < .05 and 58.6% vs 27.8%, P < .005). During the first 30 postoperative days, patients with a MELD higher than 25 lost the second graft in 48% of cases compared to 16% in the other group (P < .005). Patients retransplanted for primary graft nonfunction showed significant lower 5-year survival rates than those for other indications (28.6% vs 54.5%, P < .05) and higher mean MELD score (30.7 vs 21.9, P < .05). CONCLUSION: A MELD score of 25 is a valid cut-off to predict the outcome of retransplantations, it may be useful to select patients among those who require a second graft. Cases with primary graft nonfunction displayed lower survival, because of their compromised clinical status as evidenced by their high MELD scores.