Liver allograft biopsy is vital to the posttransplant management of patients treated with this option for end-stage chronic liver disease, acute liver failure, or hepatocellular carcinoma. Graft injury can often be detected via the monitoring of markers of parenchymal and biliary injury (transaminases and alkaline phosphatase, respectively), but these changes are far from specific, so there is always the need to specify a cause for liver enzyme abnormalities via liver biopsy and/or other means. Also, the sensitivity of enzyme changes as markers of injury is uncertain in the context of transplantation; it is possible to detect significant inflammation in protocol biopsy samples from patients with enzyme levels within normal ranges.1, 2 Then, there is the added factor of how to define normal enzymes in a patient whose best posttransplant baseline levels never fall within the usual normal range. Therefore, whether the pathologist is looking at a liver biopsy sample for a cause or as part of a protocol, he or she is faced with answering specific questions posed by the hepatologist, which could include any or all of the following: determining the cause and/or severity of graft injury; resolving a dilemma in which the cause is thought to be known but the response to the intervention is not as predicted; and assessing the progression of a known injury for the purpose of planning when an intervention should be performed, how an intervention should be performed, or which intervention is necessary. There are several reviews and guidelines addressing approaches to defining causes and severity and other ways of evaluating liver allograft biopsy samples.3 Not infrequently, however, the indication for liver biopsy is a situation in which a known condition has been adequately treated but the response is suboptimal or the effect is even the opposite of what was expected.4, 5, 6 Resolving this dilemma requires making special accommodations for changes due to the preceding therapy, which has invariably changed the classic appearance of the original pattern of injury and/or introduced new features. Viral hepatitis C remains for now the commonest cause of end-stage liver disease leading to transplantation. One common scenario that highlights this problem, therefore, is persistent liver enzyme abnormalities despite negative findings for viral RNA in hepatitis C transplant patients treated with an interferon-based regimen. Probably the commonest explanation for these patients is the development of a new immunological injury (or the aggravation of a preexisting immunological injury) after interferon exposure.3, 6, 7, 8, 9 Such injuries unfortunately do not present with predictable features of acute cellular rejection (ACR) or chronic rejection (CR). This review, therefore, describes the typical features of recurrent viral hepatitis C (rHCV), ACR, and CR in the first section. The second section discusses the various patterns of immunological injuries attributable to late-onset ACR with an emphasis on interferon-exposed individuals, and it highlights how these differ from typical changes. The final section discusses other, nonimmunological causes of persistent liver enzyme abnormalities despite negative findings for viral RNA. Table 1 summarizes the teaching points of this review.
Recurrent Viral Hepatitis and Alloimmune Causes of Graft Injury in Hepatitis C–Positive Patients
rHCV After Transplantation
Hepatitis recurs in virtually every liver allograft in patients undergoing transplantation for that reason, but the progression of recurrence varies from patient to patient and ranges from chronic smoldering to rapid and aggressive progression to fibrosis and graft loss. It is, therefore, important to be able to recognize rHCV and differentiate it from other causes of graft injury as part of graft management. Liver biopsy, whether performed as part of a protocol or indicated by abnormal liver enzymes, shows features of chronic hepatitis in the majority of patients similar to those in nontransplant patients (Fig. 1A,B). Characteristic histopathological features include portal, interface, and lobular inflammation with a tendency toward small lymphocyte predominance. Portal inflammation without significant hepatocellular injury (necrosis and/or apoptosis) should be cautiously interpreted because the features of ACR (a primarily portal-based injury) can overlap with rHCV features.10 A minority of patients who are characterized by very high viral RNA levels and clearly other factors that remain poorly understood experience a more aggressive fibrosing cholestatic pattern (Fig. 1C,D). This pattern is characterized by less prominent inflammation; cholestasis; hepatocellular swelling and disarray; and rapidly advancing portal, periportal, and subsinusoidal fibrosis.
Unlike rHCV, which benefits from a reduction in immunosuppression, ACR is managed by increased immunosuppression either on a short-term regimen or for an extended period (depending on the severity and other factors). Therefore, this should be clearly differentiated from rHCV, although the two can coexist. Most episodes of ACR occur within days to weeks of transplantation, whereas rHCV becomes histologically recognizable after 2 to 3 months. Therefore, knowledge of the timing, the pattern of enzyme elevation, and the levels of immunosuppression should assist in the overall interpretation. Classically, ACR shows portal- and/or perivenular-centered mixed inflammation with little lobular inflammation (between the portal tract and the central vein). The infiltrates can also be lymphocyte-predominant, but unlike in rHCV, these usually include activated types (e.g., immunoblasts and centroblasts) and are often mixed with eosinophils, neutrophils, and other cell types. As shown in Fig. 2A-C, the inflammation of ACR clusters in portal and/or perivenular regions because antigens residing on bile duct epithelium and endothelial cells of the portal and hepatic veins are the primary targets. As discussed later, however, a variation from this classic pattern of ACR is one of the problems encountered in post-interferon and other late occurring cell-mediated immunological graft injuries.10
Unlike in almost all other solid organs, ACR and CR are not time-defined. ACR can occur several years after transplantation, and CR can occur within a few weeks. The distinguishing features are histological in the context of supporting clinical and biochemical profiles. CR is a cholestatic disease explained histologically by atrophy and/or a loss of terminal bile ducts (Fig. 2D,E). There is typically very little or no active inflammation. Unlike bile duct injuries in other scenarios, CR occurs without any meaningful ductular reaction or portal edema, and it is unassociated with copper retention in periportal areas, at least in the early stages. The features that one sees are cholestasis (sometimes from the onset), little or no inflammation, and a loss or senescence/atrophy of small bile ducts. According to the Banff criteria, early CR is defined as the atrophying of a majority of sampled small bile ducts, and this implies that the biopsy sample has to be adequate (with more than 11 portal tracts). Late CR is the loss of more than 50% of sampled small bile ducts. For this reason, the term chronic ductopenic rejection is often applied to late CR. Intermediate-size and larger bile ducts, even in the most severe cases, could remain intact.
Persistent Liver Enzyme Abnormalities After the Successful Treatment of rHCV
The posttransplant treatment of a recurrent infection (rHCV) with an interferon-based regimen leads to a virological response in up to half of treated patients.11 However, a normalization of liver enzymes (transaminases and/or alkaline phosphatase) does not always occur, and this triggers liver biopsy for resolution. Table 2 lists some of the common entities that one has to consider in these instances. The approach to the resulting liver biopsy sample begins with an understanding of the clinical question. The pathologist should confirm that the patient remains negative for viral RNA. This is because the initial virological response is not invariably sustained, and a patient in whom viral RNA was previously undetectable could again turn positive. Next is the study of the pattern of enzyme changes. Sudden changes in numbers at any time are unlikely to be related to rHCV in a patient who has hitherto established a baseline, and the presence of a newly introduced source of injury or a flare of a latent injury should be entertained. Reviewing the morphological pattern of injury in the biopsy sample then follows, and the subsequent interpretation should be made in this context.
|De novo AIH (also known as plasma cell hepatitis)|
|Steatosis with or without steatohepatitis|
|Idiopathic posttransplant hepatitis|
|Chronic ductopenic rejection|
|Typical ACR (with portal-based features)|
|Infection (e.g., an unsampled abscess)a|
|Other focal lesions (e.g., neoplasms and infarcts)b|
|• Liver biopsy is an essential tool for resolving various posttransplant clinical dilemmas.|
|• rHCV, ACR, and CR are common causes of graft injury.|
|• The resolution of persistently abnormal liver enzymes despite negative findings for hepatitis C virus RNA should be pursued systematically, and one should keep in mind the scope of possible explanations (Table 1).|
|• De novo AIH, in particular, should be recognized and communicated clearly.|
|• Hepatitis without prominence of plasma cells has been called idiopathic, but it is probably also a form of cellular rejection.|
|• NRH can easily be missed without a high level of suspicion and good reticulin staining.|
|• Steatosis with or without steatohepatitis, when it is present, is less of a challenge to diagnose than other causes.|
|• One cause of abnormal liver enzymes could be a focal lesion, which may not be represented in a needle core sample.|
De Novo Autoimmune Hepatitis (AIH; Also Known as Plasma Cell Hepatitis or Atypical ACR)
ACR with the classic features described earlier can develop in this context. However, it is more likely that the rejection will have atypical features, and it will tend to be more hepatitic than the classic form. There is a form that is plasma cell−rich (discussed now) and another with a less pronounced plasma cell population (discussed next). If there is sufficient necroinflammation for a diagnosis of hepatitis, a high proportion of plasma cells with or without accentuation around zone 3 areas should suggest de novo AIH. This pattern of injury (Fig. 3) is the most important factor in these cases because the usual AIH serum markers (total immunoglobulin G and autoantibodies) are present in only 50% to 70% of cases,12 probably because they are suppressed by concurrent immunosuppression in others. These serum markers are, therefore, not reliable diagnostic parameters of de novo posttransplant AIH. Prior or ongoing interferon treatment is a recognized risk for de novo AIH, and there is also the likelihood of steroid resistance. Unlike the usual short-term rejection treatment, treatment is often required for an extended period (as for nontransplant AIH).5, 6, 12 Lastly, when the aforementioned de novo AIH features are seen in hepatitis C virus RNA−positive individuals, the question of whether it is AIH or (plasma cell−rich) rHCV needs to be resolved. This is hardly possible just with morphology, and should prompt a clear discussion in the final report. Even when such patients are treated as having rHCV, close monitoring with more frequent follow-up biopsies may be necessary.
Idiopathic Posttransplant Hepatitis
Hepatitis not meeting the aforementioned criteria (i.e., few or no plasma cells) should warrant a search for an exposure to new (but sometimes not so new) medications, alcohol, herbs, or other things. In the absence of any of these, the term idiopathic posttransplant hepatitis has been used, especially in the pediatric population.8 Because some of these cases also progress to fibrosis, I and other authors have come to believe that they, like de novo AIH, represent a (hepatitic) variant of cellular rejection12 that cannot and should not be graded with the usual Banff criteria. This discussion should be part of the report, and many transplant hepatologists will consider additional immunosuppressive therapy.
Nonhepatitic and Probably Nonimmunological Causes of Abnormal Liver Enzymes
Among the nonhepatitic injuries listed in Table 1, steatosis/steatohepatitis (Fig. 4) and chronic ductopenic rejection (discussed earlier) should be documented when they are present, and one should bear in mind that like de novo AIH, CR can also complicate interferon therapy.7 Steatohepatitis can be seen with alcohol; in nonalcoholic states such as morbid obesity, insulin resistance, malnutrition, and hyperlipidemia; and with exposure to some environmental toxins and adverse reactions to certain medications (including prednisone, a common drug in this patient population).
Nodular Regenerative Hyperplasia (NRH)
However, what happens when the biopsy sample is seemingly normal? NRH is increasingly recognized as a cause of liver enzyme abnormalities occurring late or in older grafts.13 Reticulin staining is the best way (if not the only way) to document NRH, but it is not always part of the routine panel of stains at many centers. Therefore, when an obvious biopsy abnormality is not initially apparent, reticulin staining should always be employed. NRH (Fig. 5) is an entity characterized by diffuse micronodularity to the liver parenchyma, but it is unassociated with fibrosis or other features of cirrhosis. Reticulin staining is used to demonstrate these nodules. Its other features include a loss of small portal veins and clinical portal hypertension in some patients. Also, liver enzymes can be normal in some patients and especially in those undergoing protocol biopsy, but when they are elevated, this tends to occur preferentially in a cholestatic pattern with alkaline phosphatase more than transaminases.
Truly Normal Liver Biopsy in a Patient With Abnormal Liver Enzymes
When this occurs, the pathologist should suggest follow-up imaging to exclude a focal lesion such as an infarct, an infection, or an expanding recurrent or de novo neoplasm, which may not have been included in the sampling.