Reply:^†

We much appreciate the effort made by Muratori et al. in validating the simplified criteria for autoimmune hepatitis (AIH) in their large, well-characterized population. We are pleased that it worked well in their hands. Preliminary reports from other centers support their conclusion and suggest that the simplified score may be useful in everyday practice.1, 2 Nonetheless, we wish to add some words of caution and raise points that may need attention in future studies.

Studies thus far have been based on patients diagnosed on the basis of the revised score of the International Autoimmune Hepatitis Group, which was not primarily validated by a prospective study.3 If this score is used as the gold standard, new criteria will by definition be of inferior positive predictive value. It would be important to evaluate patients prospectively and include a comparison of both scores. However, what could then be used as the gold standard? We believe that response to immunosuppressive therapy is such a characteristic hallmark of AIH that it could be used as an additional criterion to confirm a suspected diagnosis of AIH.

Another question might be of even higher clinical relevance: in whom should liver biopsy be performed? The simplified score requires liver biopsy as an absolute prerequisite for making the diagnosis. This is partly based on the experience of false-positive scores from the use of the old score in patients with cirrhosis due to nonalcoholic steatohepatitis.4 We believe that it will not be possible to make the diagnosis of AIH with sufficient certainty without liver biopsy being performed. However, we do not advocate performing liver biopsy on every patient with mildly elevated liver enzymes; therefore, it will be important to define criteria that make it likely to detect and unlikely to overlook significant AIH in a patient presenting with raised liver enzymes.

Muratori et al. rightly point out the problem of liver-kidney microsome (LKM)–positive hepatitis C patients. Our study unfortunately did not include a sufficient number of patients with either LKM-positive AIH or LKM-positive hepatitis C to answer this question. LKM antibodies are of high diagnostic value in children with evidence of hepatitis and no other autoantibodies. However, LKM antibodies, in contrast to soluble liver antigen/liver-pancreas (SLA/LP) antibodies, are not highly specific for AIH but are also found in patients with hepatitis C.5-7 It would indeed be helpful to study a larger patient cohort with such an antibody profile. We believe it likely that LKM antibody status can be used in the score only in the absence of hepatitis C virus infection, but this will need to tested.

Finally, we are not certain how reliable the simplified score performs in patients with acute and fulminant AIH. As these are patients in special need of a quick and reasonably reliable diagnosis, it would be of particular interest to test the score in an unselected group of patients presenting with severe hepatitis. Studies such as the one performed by Muratori et al. will help us to define the role of diagnostic criteria further and hopefully help others to recognize AIH more quickly and reliably.