Only patients suffering from a severe form of alcoholic hepatitis (Maddrey's discriminant function greater than 32) require medical treatment. Oral prednisolone for 28 days is the only treatment which has been proven to improve short-term survival over placebo in patients with severe alcoholic hepatitis. However, prednisolone alone cannot be regarded as an ideal treatment because some patients still have a bad outcome despite being treated with corticosteroids. Response to treatment can be predicted by the Lille score, a simple tool that is calculated after 7 days of prednisolone course. The ideal binary cut-off of the Lille is 0.45, responders having a Lille score < 0.45 and non-responders having a Lille score ≥0.45. In terms of treatment management, approximately 30% of patients with severe alcoholic hepatitis do not take benefit from prednisolone and are classified as null responders by a Lille score greater than 0.56. In them, there is a consensus for stopping prednisolone after a 7-day course of treatment (Lille score is calculated after 7 days) while patients with a Lille score <0.56 continue treatment for a total of 30 days.

Numerous trials have attempted to test the impact of other strategies in association with prednisolone, but none of them has shown an improvement in survival (primary endpoint) as compared to prednisolone alone. These strategies include for instance pentoxifylline, amoxicillin-clavulanic acid and enteral nutrition.

Because oxidative stress is a major driver of liver injury during alcohol-related liver disease, antioxidants, especially N-acetylcysteine, have been tested for many years to treat alcoholic hepatitis. N-acetylcysteine alone does not seem to bring a survival benefit over placebo while it may improve outcome when combined to prednisolone.

Historically in severe alcoholic hepatitis, treatment is only given for one month. However, a significant proportion of patients still disclose impaired hepatic function after treatment has been stopped (e.g. 50% of patients still have a MELD score ≥17 after 60 days in). It is thus tempting to hypothesize that a proportion of patients will recover slowly and may take benefit from a prolonged treatment. Such strategy has been proposed in some old studies with relatively limited sample size but never tested with a rigorous approach.

In the present study, for the first time in alcoholic hepatitis, we will take into account the recent recommendations of international experts by choosing an innovative primary endpoint that does not only include mortality and evaluate this endpoint at the preferred timepoint of 90 days.

After more than 30 years of negative trials in severe alcoholic hepatitis, the present study is aimed to evaluate two important new strategies to decrease both mortality and liver impairment.

Inclusion Criteria:

• Patients aged 18-75
• Alcohol consumption of more than 40g/day (women) and 50g/day (men)
• Recent onset of jaundice (<3 months)
• Biopsy proven alcoholic hepatitis (transjugular liver biopsy)
• Maddrey's discriminant function ≥ 32, defining severe alcoholic hepatitis
• MELD score ≥ 17
• Patients covered with social insurance
• Patients having provided written informed consent to participate

Exclusion Criteria:

• Hepatocellular carcinoma
• Uncontrolled gastrointestinal bleeding
• Previous severe allergy or hypersensitivity to N-acetylcysteine (anaphylactic shock, Quincke edema, severe urticaria)
• Hypersensitivity to any component of the medication
• MELD score <17
• Type 1 hepatorenal syndrome before the initiation of treatment
• Severe extrahepatic disease, with life expectancy < 6 months
• Any malignant tumor < 2 years (except skin carcinomas)
• Ongoing viral or parasitic infection
• Untreated bacterial infection
• Tuberculosis < 5 years
• Positive blood PCR in patients with positive antibodies against HCV
• Patient carrying HBV or HIV
• Treatment with corticosteroids, immunosuppression therapy or budesonide within 6 months before the study