In 2003, the first International Liver Transplantation Society expert panel consensus conference recognized that there are 2 distinct forms of severe hepatitis C virus (HCV) recurrence, one more common and easily diagnosed (chronic HCV disease) and the other less frequent but more aggressive (cholestatic HCV disease). It was then recognized that these forms of recurrence differ substantially in pathogenesis, clinical presentation, prognosis, and outcome.
A definition of cholestatic HCV disease was then proposed. It was based on several criteria that included the following:
- More than 1 month after transplantation (usually <6 months).
- Serum bilirubin level greater than 6 mg/dL.
- Serum alkaline phosphatase and gamma-glutamyl transferase levels greater than 5 times the upper limits of normal.
- Characteristic histological state with ballooning of hepatocytes predominantly in the perivenular zone (not necrosis or fallout), a paucity of inflammation, and variable degrees of cholangiolar proliferation without bile duct loss.
- Very high serum HCV RNA levels.
- Absence of surgical biliary complications (normal cholangiogram) and absence of evidence for hepatic artery thrombosis.
Studies of posttransplant HCV have reported mostly on the commonest form of recurrent disease, and not many studies have specifically dealt with this patient population.[1-4] In most publications, patients with cholestatic hepatitis have been grouped with patients with progressive fibrosis and poor outcomes. Publications that have specifically dealt with this entity include those reporting on the pathophysiological mechanisms involved,[5, 6] others describing this complication in specific populations (particularly human immunodeficiency virus–coinfected patients), and some examining antiviral responses to emphasize the need to maintain therapy over the long term.[1, 3, 8]
In fact, most studies focusing on cholestatic hepatitis typically have been case reports or case series, and these have varied in their defining criteria or in the types of clinical information provided. Indeed, in a recent systematic review, the authors found that only 3 of the 12 studies published after 2003 used the definition of the consensus conference. In that systematic review, although 28 of 30 articles provided a definition of posttransplant cholestatic hepatitis, the definitions and the diagnostic criteria that were applied varied among the articles. In particular, histological features were not uniform, with the most common feature being cholestasis with portal/periportal fibrosis together with structural hepatocyte changes marked by ballooning degeneration. In addition, the types of data collected lacked uniformity. In particular, only 13 articles reported on serum HCV RNA levels at the time of diagnosis. In order to better understand this disease, the authors of this systematic review proposed that studies focusing on cholestatic hepatitis should provide some minimum information: serum HCV RNA levels at the time of diagnosis; HCV genotype; time from liver transplantation to the diagnosis of cholestatic hepatitis; type of immunosuppression; drug level goals; total and direct bilirubin, alanine and aspartate aminotransferase, alkaline phosphatase, and gamma-glutamyl transpeptidase levels at the time of diagnosis; histological findings; treatment data (including the dosing, duration, and biochemical and virological responses); and outcomes (including autopsy findings if they are available).
Under the assumption that the consensus conference definition was vague and hence was rarely applied in clinical practice and that “little is known about the pathophysiology” of this entity, Verna et al. tried a different approach to establish the incidence of cholestatic hepatitis and potentially modify the diagnostic criteria. They selected a group of cholestatic hepatitis recipients with a combination of outcome and histopathological criteria, and they compared the results of these patients to the results of patients with either minimal fibrosis or advanced fibrosis who did not meet specific cholestatic criteria. Biopsy samples taken 1 year after transplantation (range = 8-16 months) and 19 biopsy samples not taken during this time period but read initially as cholestatic hepatitis were re-read by 1 pathologist. Specific cholestatic hepatitis parameters (a prominent ductular reaction in most portal tracts, cholestasis, prominent hepatocyte ballooning with lobular disarray, and any degree of sinusoidal/pericellular fibrosis) were given a score, and each score was then assessed for its relationship with posttransplant survival in order to select cholestatic hepatitis cases on the basis of their inferior outcomes independently of the stage of fibrosis. Patients with 3 or more of the 4 cholestatic HCV criteria had significantly worse survival regardless of the fibrosis stage, and they constituted the cholestatic hepatitis group (n = 27 or 15%), which was then compared to the group of patients with minimal fibrosis (≤F1; n = 99 or 55%) and the group of patients with advanced fibrosis (F2-F4; n = 53 or 30%) who did not meet the criteria for cholestatic HCV. Independent predictors of this novel cholestatic HCV criteria group were donor age, previous rejection, and total bilirubin. The HCV viral load did not predict this entity. Posttransplant survival was predicted by the pathological group (cholestatic hepatitis), significant fibrosis, donor age, and cold ischemia time. On the basis of their findings, the authors have proposed a new set of diagnostic criteria for cholestatic hepatitis: (1) the presence of at least 3 of 4 histopathological lesions on a biopsy sample taken at least 1 month after transplantation, (2) supportive criteria [total bilirubin level > 2 mg/dL, gamma-glutamyl transpeptidase level > 150 IU/L, aspartate aminotransferase level > 70 IU/L, older donor age, and recent biopsy-proven acute rejection (Banff rejection activity index ≥ 5)], and (3) the absence of exclusionary criteria (biliary obstruction or hepatic artery thrombosis).
How much does this new definition differ from prior recommendations? The specific criteria (the presence of at least 3 typical histopathological lesions) do not add anything new to what is already known.[1, 3, 9, 11] Regretfully, the authors did not provide data on earlier histological findings that in clinical practice could possibly be much more helpful in managing this disease by predicting cholestatic hepatitis before it is fully developed and hence is less amenable to responding to therapy. In one study, fibrosis (especially sinusoidal fibrosis) and cholestasis were found to be earlier findings that should raise the suspicion of cholestatic hepatitis. The exclusionary criteria are the same. The most important differences from the criteria proposed by the consensus conference lie in the supportive criteria and particularly in the laboratory tests. As for classical laboratory parameters, the new criteria modify the cutoffs, particularly for bilirubin levels. The authors use a significantly lower cutoff than that proposed by the consensus conference (2 instead of 6 mg/dL) because at that cutoff, the total bilirubin level is 100% sensitive and 60% specific for a diagnosis of cholestatic hepatitis. This means that a bilirubin level ≤ 2 mg/dL excludes cholestatic hepatitis. Such a suggestion is so obvious that it does not withstand scrutiny. Furthermore, the new cutoffs for gamma-glutamyl transpeptidase and aspartate aminotransferase are set at 150 and 70 IU/L, respectively, because at these cutoffs they are 96% sensitive and 27% to 42% specific for cholestatic hepatitis. The original biochemical cutoffs were designed to combine specificity with pathology findings. They were not designed to withstand individual sensitivity parameters. If such biochemical criteria were accepted, then cases of simple cholestasis with mild biochemical abnormalities would have the potential to be overdiagnosed and wrongly classified as cholestatic HCV. This would set the field back considerably.
Additionally, the viral load is no longer considered a predictive factor of cholestatic hepatitis, and this finding is at odds with major prior publications as well as the underlying pathogenesis associated with this entity.[1, 3-6, 13, 14] It should be recognized that the biopsy samples used to diagnose cholestatic HCV in this study were taken 281 days after transplantation. The viral loads seem to have been analyzed only in a proportion of patients at this time point. No difference was seen in the viral loads at this time point versus 356 (significant fibrosis) and 381 days (mild fibrosis). The major issue here is that serial peak viral loads are key to recognizing the potential to develop cholestatic HCV. It has been well established that viral loads peak at approximately 3 months and afterward slowly decline no matter what the diagnosis is.[15-17] Furthermore, the consensus conference definition suggests that most cases are originally diagnosed within 6 months of transplantation, whereas the viral loads of the cases in this study were estimated much later (281 days). In another study with a similar design, 17 cholestatic hepatitis cases that were initially identified on the basis of pathology were further characterized clinically to better assess specific clinical features as well as outcomes. The major findings of that study are that these cases tend to occur early after liver transplantation at a mean of 7 months and that they are associated with hyperbilirubinemia (bilirubin > 5 mg/dL in 88% of cases and > 10 mg/dL in 59%), extremely high HCV RNA levels in comparison with levels in patients not developing cholestatic hepatitis, a donor age greater than 50 years, and extremely poor survival; these conclusions are more in line with the original consensus definitions. Thus, we agree that cross-sectional viral loads may not separate cholestatic HCV from noncholestatic HCV, particularly if the viral load is estimated late in the illness. However, we would claim that serial viral loads from the time of transplantation would certainly do so. Thus, the viral load should remain a key to the diagnosis. It is also key to recognizing cholestatic HCV as a different pathogenic entity. Thus, we would reject the proposal that the viral load be eliminated from the definition of cholestatic HCV. To be generous, one could claim that if a clinician did not have access to serial viral load measurements, then other criteria could be applied.
We conclude that the article by Verna et al. does not really clarify the study of cholestatic HCV. It is crucial for us to understand this entity better. Increasingly, when it occurs, it will be the target of new non–interferon-based antiviral therapies. The key is clear pathological criteria, very high viral loads, and known exclusion criteria. The current article challenges the International Liver Transplantation Society consensus criteria, but the challenge falls short.