BACKGROUND: Since March 2007, priority access to liver transplantation in France has been given to patients with the highest MELD scores. OBJECTIVE: To undertake an intent-to-treat comparison of center-based vs. MELD-based liver graft allocation. METHODS: Retrospective cohort analysis (patients listed 6th March 2007 to 5th March 2009; MELD period) with a matched historical cohort (patients listed 6th March 2005 to 5th March 2007; pre-MELD period) in a single high-volume center. Analysis was on an intent-to-treat basis, i.e. starting on the day of wait listing. RESULTS: Compared to pre-MELD, fewer patients with a MELD score less or equal to 14 (P=0.002), and more patients with a MELD greater or equal to 24 (P<0.05) were transplanted during the MELD period. For HCC candidates, median waiting time increased (121 vs. 54 days, P=0.01), transplantation rate halved (35% vs. 73.5%, P<0.001) and dropouts due to tumor progression increased (16% vs. 0%, P<0.001). Moreover, postoperative course did not change significantly except for infectious complications (35% vs. 24%, P=0.02); overall patient survival was 69.8 ± 3.1% vs. 76 ± 2.9% (P=0.29) and overall graft survival was 77.6 ± 3.4% vs. 82.8 ± 2.9% (P=0.29). Transplant failures were mainly due to deaths on the waiting list in the previous system, but to dropouts related to disease progression in the new system. Cirrhotic patient survival rate did not change (78.1 ± 4.4% vs. 73.5 ± 4.5%, P=0.42), while that of HCC patients decreased (65.3 ± 5.3% vs. 86.8 ± 4.4%, P=0.01). Post-transplant survival worsened significantly according to pre-transplant MELD score (P=0.009). CONCLUSION: The MELD-based graft allocation system introduced discrimination against HCC patients, whose incidence has increased dramatically, and should be reevaluated.