Refining criteria for liver biopsy in severe alcoholic hepatitis: Moving the field forward

Symptomatic alcoholic hepatitis (AH), a clinical syndrome, is associated with significant morbidity and mortality.¹,² Liver biopsy is recommended to diagnose AH when a clinical diagnosis is uncertain using the National Institute on Alcohol Abuse and Alcoholism criteria.³ Typically, modified discriminant function and model for end-stage liver disease (MELD) scores are used to stratify disease severity, with respective scores of >32 and >20 characterising severe AH.³,⁴ Over the last few years, liver biopsy in AH has gained attention in predicting the outcome of patients with AH. The pivotal study on histology in AH introduced the AH histologic score (AHHS) using the presence and severity of infiltration by polymorphonuclear (PMN) cells, bridging fibrosis and cirrhosis, bilirubinostasis and megamitochondria.⁵ PMN infiltration and megamitochondria are protective, while the other two features are associated with severe disease. With a total possible score of 9, patients a score of ≤3 are stratified as mild disease, and those with a score of >5 are classified as having severe disease. Although independently validated in a cohort of 109 patients,⁵ the role of the AHHS remains controversial given contrasting results by other studies.⁶,⁷

In this issue of United European Gastroenterology Journal, Dubois et al. attempted to ascertain the prognostic value of the AHHS for predicting outcomes within six months among 107 histologically characterised patients with severe AH recruited from two different cohorts.⁸ This study also assessed the role of the Laennec staging system, which has been evaluated for patients with cirrhosis but not among AH patients. The AHHS available in 105 patients was categorised as mild, intermediate and severe in 10, 29 and 66 patients, respectively. Survival rates in three AHHS risk categories were: 90%, 72% and 69% at 28 days (p = 0.6); 80%, 52% and 63% at three months (p = 0.3); and 70%, 41% and 58% at six months (p = 0.3). Survival rates remained similar when analysed for Laennec stage on liver biopsy. After controlling for patient age and MELD score, both the AHHS and Laennec stage were not associated with patient survival. The AHHS was also not associated with response to corticosteroids.

However, the findings of this study and analysis of individual components of the AHHS provided important and clinically relevant information. For example, PMN infiltration emerged as a protective factor and was associated with a better outcome, similar to other studies.⁵,⁹ Further, agreement between pathologists on the presence of megamitochondria was modest, which not only impacts utility of this score in clinical practice, but also implies dedicated hepatopathologists who are trained to read AH biopsies to define the utility of the AHHS in clinical practice better. As the presence and degree of fibrosis is the most important determinant of liver disease, including patients with alcohol-associated liver disease,¹⁰ fibrosis stage in the AHHS does not have much of a role in predicting disease outcome, as the presence of cirrhosis or bridging cirrhosis is allocated 3 points, while all those with fibrosis stages 0, 1 and 2 do not receive any points. Similarly, for bilirubinostasis, no points are given for patients with cholestasis in hepatocytes but without bilirubinostasis in ductules and/or canaliculi.

The current study only included patients with severe AH, with a median MELD score of 22 in one cohort and 24 in the other cohort, in contrast to a median MELD score of 18 in an earlier study by Altamirano. Higher disease severity in the current study was also evidenced by intermediate- or high-risk AHHS in the majority, 89% receiving corticosteroids, and higher three-month mortality compared to the earlier study. Seemingly, clinical phenotype in severe AH, including sterile and infection-related systemic inflammatory response syndrome, MELD and Lille scores, and acute on chronic liver failure with multi-organ failure, are more robust in predicting short-term outcome in severe AH, while the AHHS may be more useful for patients with mild to moderate AH or for asymptomatic patients with histological findings of AH.¹,²,⁴,¹¹,¹² Among symptomatic AH patients who survive the index episode, liver biopsy may also be useful in predicting long-term outcome. For example, AH on histology was the best predictor for advancement to cirrhosis in one study at subsequent biopsy in four years.¹³

Clearly, the findings of the current study make some important observations which will help move the field forwards. Future large studies are needed to examine the role of individual components of the AHHS and to examine its utility in patients with moderate and less severe disease and among AH patients without cirrhosis. Studies are also needed to examine the role of liver biopsy findings in AH patients in predicting outcome beyond six months. Needless to say, these studies should take into account patient behaviour and abstinence from alcohol, given that this is a major determinant of outcomes.³,¹⁰,¹⁴