I read with great interest the article by Björnsson et al.,1 which describes the clinical, histological, and prognostic features of drug-induced autoimmune hepatitis (DIAIH) (n = 24 patients) compared to classical AIH (n = 237 patients). The clinical and histological scores were similar in both groups, but the prognosis was more favorable in DIAIH cases. Corticosteroid responsiveness was similar in both groups, while discontinuation of immunosuppression was tried and successful in 14 DIAIH cases, with no relapses (0%), whereas 65% of the patients with classical AIH had a relapse after discontinuation of immunosuppression (P < 0.0001). I would make some comments about their findings.
Although they authors did not specifically mention, it seems that the DIAIH cases were all acute in presentation (whether purely acute or an acute flare of chronic liver disease). There was stage 0 fibrosis histologically in all cases, presence of centrilobular (65%) and confluent necrosis (30.4%) in a substantial proportion of cases, and a median increase in aminotransferases > 10 times of the upper limit of the normal range (tables 1-3 of Björnsson et al.1). Moreover, those figures were much more prominent for nitrofurantoin-induced DIAIH (table 4 of Björnsson et al.1). However, how can we explain the appearance of the liver as “cirrhotic” in a majority of cases with nitrofurantoin-induced DIAIH? The radiologic appearance of confluent necrosis, fibrosis, or massive fibrotic bands could not be confirmed with histological analysis. Sampling variability of liver biopsy or radiologic mimics of cirrhosis, such as severe or fulminant hepatitis,2 may explain this discordance. I think that the latter scenario is more consistent based on the abovementioned data. So, I would like to assume that nitrofurantoin-induced DIAIH cases in this series were acute and severe (although not fulminant) in clinical presentation. If this assumption is true, two further comments arise.
First, the findings of Björnsson et al. represent new clues about the potential of liver fibrosis reversibility. Fibrotic deposition related to recent disease and characterized by the presence of thin reticulin fibers, often in the presence of a diffuse inflammatory infiltrate, is likely to be fully reversible, whereas long-standing fibrosis, branded by extensive collagen cross-linking by tissue transglutaminase, presence of elastin, dense acellular/paucicellular extracellular matrix, and decreased expression and/or activity of specific metalloproteinases, is not.3, 4 So, the successful and sustained remission in DIAIH cases supports this pathophysiological basis.
Second, in addition to centrilobular or confluent necrosis, seronegativity of all markers was proposed as distinctive features of acute-onset classical AIH.5 However, this was not the case in the present series, whereas antinuclear antigen and/or alpha-smooth muscle actin was positive in 23 of 24 DIAIH cases. So, seronegativity does not seem to be a feature of acute-onset DIAIH.
Finally, I applaud the efforts of Björnsson et al.1 in that we will be more comfortable starting steroids in a patient with DIAIH even with radiologic features of “cirrhosis”.