Corticosteroids have been shown to significantly decrease short-term mortality in patients with severe alcoholic hepatitis. However, independent factors associated with a favorable outcome and long-term survival are unknown. The aim of this study was to examine prognostic factors and long-term survival in patients with biopsy-proven severe alcoholic hepatitis.

Of 183 patients studied, 61 had been randomized in a previous trial; 32 of them were treated with prednisolone (group I) and 29 were not treated (group II); 61 were treated from the end of this randomized trial (group III); and 61 were simulated (group IV).

At 1 year, survival in group I (69%; confidence interval [CI], 57%-81%) and group III (71%; CI, 55%- 87%) was better than in the nontreated groups (group II, 41%; CI, 23%- 59%; P = 0.01) (group IV, 50%; CI, 37%-63%; P = 0.05). At 2 years, survival was not significantly different. Treated patients with marked liver polymorphonuclear infiltrate had better 1-year survival (76%; CI, 64%-88%) than the others (53%; CI, 35%-71%; P = 0.05). Treated patients with polymorphonuclear counts of > 5500/mm3 had better 1-year survival (77%; CI, 65%-89%) than the others (40%; CI, 14%-66%; P = 0.003). In the 93 treated patients, liver polymorphonuclear infiltrate (P < 0.03) and polymorphonuclear count (P < 0.001) were independently correlated with 1-year survival.

Prednisolone reduced mortality by at least 1 year. Liver polymorphonuclear infiltrate and polymorphonuclear count were independent prognostic factors. (Gastroenterology 1996 Jun;110(6):1847-53)