BACKGROUND/AIMS The alcoholic hepatitis histologic score (AHHS) is a recently developed clinical model for predicting short-term mortality in Caucasian patients with alcoholic hepatitis (AH). The AHHS has not been extensively validated in other ethnic populations. This study validated the AHHS in a Korean patient cohort.

METHODS We conducted a prospective cohort study of hospitalized Korean patients with AH between January 2010 and August 2017. Histopathological findings were assessed to determine the AHHS in all study subjects. Histopathological risk factors were examined by Cox regression analysis to predict overall survival (OS). Kaplan-Meier curves were plotted to assess the diagnostic performance of the AHHS.

RESULTS We recruited a total of 107 patients with biopsy-proven AH. None of the individual AHHS components were associated with 3-month mortality. However, the bilirubinostasis type and fibrosis severity were significantly associated with AH mortality beyond 6 months (all p<0.05, except fibrosis severity for 6-month mortality) and OS (all p<0.05). The modified AHHS classification as a binary variable (<5 vs ≥5) was also associated with OS (hazard ratio, 2.88; 95% confidence interval [CI], 1.50 to 5.56; p=0.002), and had higher predictive performance for OS (concordance index [C-index], 0.634; 95% CI, 0.561 to 0.707) than the original AHHS classification (mild vs moderate vs severe: C-index, 0.577; 95% CI, 0.498 to 0.656). This difference was statistically significant (p=0.045).

CONCLUSIONS In this prospective Korean AH cohort, the modified AHHS was significantly associated with OS. Therefore, the AHHS might be a useful histological prognosticator for long-term prognosis in patients with nonsevere AH.