Résumé : Background & Aims: Patients with alcoholic hepatitis and a modified Maddrey's discriminant function (mDF) <32 have a low risk of short-term mortality. However, few data exist concerning long-term outcomes. The aims of this study were to evaluate 5-year survival rates and to identify predictive factors for long-term prognosis in this patient population. Methods: We studied patients from 2 centers who were admitted for hepatic decompensation (ascites, hepatic encephalopathy, or jaundice) and who had histological findings of steatohepatitis and an mDF <32. Clinical and biological parameters were recorded at the time of liver biopsy and alcohol consumption was recorded during follow-up. We performed Cox proportional hazard survival analysis to identify factors associated with 5-year survival. Results: One hundred and twenty-one patients were included (male: 64%, mean age: 51.5 ± 10.3 years, presence of cirrhosis: 84%). The median model for end-stage liver disease and mDF scores were 14 (IQR 11.7–16.1) and 19 (IQR 11.1–24), respectively. During follow-up, 30% of the patients remained abstinent. Survival rates at 1, 6, 12, 24, and 60 months were 96.7 ± 1.6%, 90.1 ± 2.7%, 80.8 ± 3.6%, 69.9 ± 4.3%, and 50.7 ± 4.9%, respectively. The majority of deaths (80%) were liver related. In multivariable analysis, encephalopathy at baseline and alcohol abstinence were predictive of 5-year survival. The 5-year survival rates of patients without and with encephalopathy at baseline were 60.5 ± 5.8% and 29.7 ± 8.0%, respectively, and the 5-year survival rates of abstinent and non-abstinent patients were 74.0 ± 8.0% and 40.9 ± 5.8%, respectively. Conclusions: The mortality rate of patients with alcoholic hepatitis and an mDF <32 is around 50% at 5 years. Hepatic encephalopathy at baseline and lack of alcohol abstinence impair long-term prognosis. New treatment strategies, including measures to ensure abstinence, are required. Lay summary: Patients with alcoholic hepatitis that is of intermediate severity have a low risk of short-term mortality but not much is known regarding long-term outcomes for these patients. This study clearly indicates that patients with intermediate disease characteristics have poor long-term outcomes. The presence of hepatic encephalopathy at the time of diagnosis and the absence of alcohol abstinence during follow-up are factors that predict poor long-term mortality.