Article Text


Is histology required for the diagnosis of alcoholic hepatitis? a review of published randomised controlled trials
  1. R Hamid *1,
  2. E H Forrest2
  1. 1Department of Gastroenterology, Royal Alexandra Hospital, Paisley, UK
  2. 2Department of Gastroenterology, Glasgow Royal Infirmary, Glasgow, UK


Introduction The need for liver biopsy to diagnose alcoholic hepatitis (AH) in the clinical setting remains controversial. Coagulopathy and/or ascites may preclude percutaneous biopsy and facilities for transjugular biopsy are not universally available. Despite this, some experts still advocate that biopsy confirmation of AH is a prerequisite for determining diagnosis, prognosis and treatment strategy for this condition.

Aim The authors aimed to review the already published literature in the form of randomised controlled trials (RCTs) of therapy for AH. This was primarily to determine if these data could determine the need for histology in the selection of patients for study.

Methods Studies published from 1970 to 2010 were identified through a MEDLINE computer search and manual search of the bibliographies of identified articles. Inclusion and exclusion criteria for each study were determined. The number of biopsies performed and the patients excluded as a consequence of incompatible histology were recorded.

Results The authors identified 43 published RCTs trials, with adequate data available in 39.

11 of the trials required ‘biopsy proven’ AH prior to randomisation of patients, however the histological criteria for the diagnosis of AH varied widely between studies. Where figures were provided, it appears 1668 liver biopsies were undertaken in these trials with 1409 showing evidence of AH (84.5%). However if studies defined a minimum value of bilirubin (usually >80–85 μmol/l) as an entry criterion, or the minimum value of bilirubin described was >80 μmol/l, 96% of biopsies confirmed AH (405 of 422 biopsies: p<0.0001 (95% CI 12.2% to 18.2% compared with studies without such stipulations).

Where co-existent cirrhosis was mentioned, 839 of 1187 (70.7%) of biopsies had evidence of this. There was no correlation between the proportion of patients with cirrhosis and the mortality in the placebo treated patients (p=0.81; 95% CI −0.54 to 0.44).

Conclusion Overall the diagnosis of AH is confirmed in over 80% of cases, but this clinical diagnostic accuracy increases to over 95% with a minimum value of bilirubin. Co-existent cirrhosis does not influence short-term prognosis. These data indicate that in most cases histology is not necessary for the diagnosis and management of AH.

  • alcoholic hepatitis
  • bilirubin
  • liver biopsy
  • randomised controlled trial.

Statistics from


  • Competing interests None.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.