Clinical Parameters Are More Predictive of Mortality in Alcoholic Hepatitis than Histopathologic Severity

BACKGROUND: Alcoholic hepatitis (AH) is primarily diagnosed by clinical parameters, but is often misdiagnosed due to nonspecific symptoms, leading to high mortality rates. While histology aids definitive diagnosis, the role of the liver biopsy in its workup is still controversial. Currently, there is no widely accepted grading histologic grading system for AH. The relationship between biopsy findings and clinical course is also yet unknown. The Alcoholic Hepatitis Histologic Score (AHHS) was recently developed to define patient prognosis by histologic criteria. OBJECTIVE: The purpose of this study was to compare histologic severity defined by the AHHS with clinical severity of AH, as seen with symptoms, laboratory markers, and patient survival. METHODS: We conducted a retrospective case series of 56 patients with biopsy-proven AH from two hospitals in Dallas, TX, USA. Clinical and demographic data were collected from electronic medical records. Two trained pathologists blinded to patients' outcomes graded liver biopsies using the AHHS criteria. Relationships between clinical symptoms and complications, laboratory investigations, patient outcomes, individual histologic features, and the AHHS were analyzed. RESULTS: No hematologic or biochemical laboratory markers significantly correlated with the AHHS. Higher AST correlated with a greater degree of steatosis on biopsy (p<0.0019). Severe neutrophil infiltration on biopsy correlated with higher serum bilirubin, INR, MELD, and DF (p=0.034). Survival analysis by Kaplan-Meier curves and log-rank tests showed no significant correlation between AHHS and 90-day survival (p=0.09), while multiple clinical scoring systems accurately stratified prognosis (p<0.018 for all). Severe neutrophil infiltration on biopsy also correlated significantly with death (p=0.0001). CONCLUSION: Retrospective analysis in a diverse U.S. urban cohort did not confirm the validity of AHHS to predict survival in AH. In contrast, clinical parameters were better predictors of survival. Our results suggest that clinical deterioration, rather than histopathologic severity, is more informative in determining prognosis in AH. The relationship between neutrophil infiltration and mortality deserves further study.