To assess the efficacy of liver transplantation in patients with alcoholic cirrhosis, we compared 2-year survival of 169 liver transplantation patients in 12 French centres with survival of two control groups treated conservatively. The matched group was 169 patients of similar age, cirrhosis severity, and bleeding history; the simulated group was 169 patients whose theoretical survival was determined in a cohort of 797 patients with alcoholic cirrhosis. The probability of survival to 2 years in transplanted patients was 73 (95% confidence interval 67-79%) versus 67% (59-75) in the matched and 67% (63-71) in simulated controls. When prognostic factors were taken into account, transplantation was associated with survival (r = 0.527; p = 0.069). Patients with severe liver disease (high-risk group) benefited most for 2-year survival: 64% (42-86) vs 41% (23-59) in the matched and 23% (19-27) in the simulated control groups (p < 0.01). There was no difference for patients at low and at medium risk. Liver transplantation increases the 2-year survival of patients with severe alcoholic cirrhosis. In patients with less severe disease, further studies should be done by non-randomised controlled studies with longer follow-up or by randomised trials.