Background & aims: In severe (Maddrey score >or=32) alcoholic hepatitis (AH), infection is classically viewed as a contraindication for corticosteroids, although specific data are lacking. This study's aims were (1) to evaluate the incidence of infection in patients with severe AH before and after corticosteroid treatment; (2) to determine whether infection contraindicates corticosteroids; and (3) to focus on predictive factors of development of infection.
Methods: At admission, systematic screening of infection consisted of chest x-ray and blood, ascites, and urinary cultures. All patients were treated with prednisolone. Response to steroids was defined using the Lille model.
Results: Two hundred forty-six patients with severe AH were prospectively included. Infections at admission were as follows: 63 infections (25.6%) were diagnosed: 28 (44.4%) spontaneous bacterial peritonitis or bacteremia, 8 (12.7%) pulmonary infections, 20 (31.7%) urinary tract infections, and 7 (11.2%) other infections. Patients infected before using corticosteroids had 2-month survival similar to that of others: 70.9% +/- 6.1% vs 71.6% +/- 3.4%, respectively, P = .99. Development of infection after steroids: 57 patients (23.7%) developed infection: 16 (28.1%) spontaneous bacterial peritonitis or bacteremia, 23 (40.3%) pulmonary, 10 (17.5%) urinary tract, and 8 (14.1%) other infections. Infection occurred more frequently in nonresponders than in responders: 42.5% vs 11.1%, respectively, P < .000001. In multivariate analysis, only the Lille model (P = .0002) independently predicted infection upon steroids use. The Lille model (P = .000001) and Model for End-Stage Liver Disease score (P = .006) were independently associated with survival, whereas infection was not (P = .52).
Conclusions: Severe AH is associated with high risk of infection. Infection screening is warranted but should not contraindicate steroids. In terms of mechanisms, nonresponse to steroids is the key factor in development of infection and prediction of survival.