The current policy for determining priority for organ allocation is based on the model for end stage liver disease (MELD). We hypothesize that severity of graft dysfunction assessed by either the MELD score or the Child‐Turcotte‐Pugh (CTP) score correlates with mortality after liver retransplantation (re‐OLT). To test this hypothesis, we analyzed the outcome of 40 consecutive patients who received re‐OLT more than 90 days after primary orthotopic liver transplantation (OLT). The Kaplan‐Meier 1‐year and 5‐year survival rates after re‐OLT were 69% and 62%, respectively. The area under the curve (AUC) values generated by the receiver operating characteristics (ROC) curves were 0.82 (CI 0.70‐0.94) and 0.68 (CI 0.49‐0.86), respectively (P = .11), for the CTP and MELD models in predicting 1‐year mortality after re‐OLT. The 1‐year and 5‐year survival rates for patients with CTP scores less than 10 were 100% versus 50% and 40%, respectively, for CTP scores of at least 10 (P = .0006). Patients with MELD scores less than or equal to 25 had 1‐year and 5‐year survival rates of 89% and 79%, respectively, versus 53% and 47%, respectively, for MELD scores greater than 25 (P = .038). Other mortality predictors include hepatic encephalopathy, intensive care unit (ICU) stay, recurrent hepatitis C virus (HCV) infection, and creatinine level of 2 mg/dL or higher. Analysis of an independent cohort of 49 patients showed a trend for a correlation between CTP and MELD scores with 1‐year mortality, with AUC of 0.59 and 0.57, in respective ROC curves. In conclusion, our results suggest that severity of graft failure based on CTP and MELD scores may be associated with worse outcome after re‐OLT and provide a cautionary note for the “sickest first” policy of organ allocation. (HEPATOLOGY 2004;39:230–238.)