We read with interest the excellent article by Feld et al. 1 in which they quoted our work. A recent study by Verma et al. also found that patients with cirrhosis at presentation were more likely to die or develop complications of their liver disease during follow-up. 2 In fact, in our study individuals with cirrhosis at presentation had similar outcomes (development of cirrhosis complications/liver failure, need for a transplant, or death) to those without cirrhosis at onset/follow-up (15% vs. 5%, respectively). However, the cohort that developed cirrhosis during follow-up was significantly more likely to have an adverse outcome (43%). 2 About 20% to 40% of patients with autoimmune hepatitis (AIH) develop cirrhosis during a follow-up period ranging from 5 to 10 years despite undergoing therapy. 2,3 This group is more likely to take a longer time to achieve remission, the elevated aminotransferases reflecting ongoing necroinflammation, which in the long term translates to hepatic fibrosis and cirrhosis. 2,4 Therefore, the poorer outcome in this cohort is not surprising. However, those with cirrhosis at presentation appear to have similar outcomes to those without cirrhosis (5-and 10-year survival rates 97% vs. 94% and 89% vs. 95%, respectively). 2,5 Therefore, the observation by Feld et al. of significantly poorer outcome (38.1% vs. 4.8%) and survival (5-and 10-year survival rates 78.7% vs. 96.7% and 67.2% vs. 94%, respectively) in those with cirrhosis at presentation versus those without cirrhosis is in variance with previously reported data. There may be a number of reasons for this. First, in Feld et al.'s study, of the 16 patients with cirrhosis who developed an endpoint, at least 5 had "burned-out AIH." In such a cohort it is incredibly difficult to make a diagnosis of AIH with certainty, as histology invariably reveals inactive cirrhosis. The positive autoantibodies could be coincidental and the elevated globulins may just reflect the underlying cirrhosis. Was nonalcoholic fatty liver disease (NAFLD) excluded in these 5 cases? Up to 25% of patients with NAFLD may have positive autoantibodies, and it is not unusual to see minimal or no fat in presence of well-established NAFLD cirrhosis. 6 Even if those with "burned-out AIH" did indeed have AIH, their long-term course could be different from those without "burned-out" cirrhosis. Second, of the total cohort of 125, 2 were hepatitis C virus (HCV) RNA positive, 6 had not had HCV serology, 5 were antimitochondrial antibody (AMA) positive, and 5 had type 2 AIH. If any of these 18 patients were among the 16 with cirrhosis at presentation with a poor outcome, the results might have been further biased. Finally, the authors did not clarify what proportion of their patients developed cirrhosis during follow-up. Did any of the 4 patients without cirrhosis at presentation with an endpoint eventually develop cirrhosis during follow-up?
Therefore, before any firm conclusions can be drawn, it would be prudent if the authors could clarify how many of the 16 patients with cirrhosis and a poor outcome definitely had type 1 AIH. It might also be useful to further analyze the data by stratifying the cohort into cirrhosis at presentation, cirrhosis during follow-up, and no cirrhosis.