Liver biopsy interpretation for causes of late liver allograft dysfunction


  • Potential conflict of interest: Nothing to report.

  • See end of article text for complete list of authors and affiliations.


Evaluation of needle biopsies and extensive clinicopathological correlation play an important role in the determination of liver allograft dysfunction occurring more than 1 year after transplantation. Interpretation of these biopsies can be quite difficult because of the high incidence of recurrent diseases that show histopathological, clinical, and serological features that overlap with each other and with rejection. Also, more than one insult can contribute to allograft injury. In an attempt to enable centers to compare and pool results, improve therapy, and better understand pathophysiological disease mechanisms, the Banff Working Group on Liver Allograft Pathology herein proposes a set of consensus criteria for the most common and problematic causes of late liver allograft dysfunction, including late-onset acute and chronic rejection, recurrent and new-onset viral and autoimmune hepatitis, biliary strictures, and recurrent primary biliary cirrhosis and primary sclerosing cholangitis. A discussion of differential diagnosis is also presented. (HEPATOLOGY 2006;44:489–501.)

Distinguishing among potential causes of late liver allograft dysfunction can be difficult because of overlapping clinical, serological, and histopathological features. Most problematic biopsies are obtained more than 1 year after transplantation. Currently, diagnoses are made using center-specific criteria, but a standardized set of criteria has not been generally agreed upon. Availability of standardized criteria1, 2 would enable centers to compare and pool results, improve therapy, and better understand pathophysiological disease mechanisms.

Native disease recurrence is a significant problem and can be categorized as follows: (1) infectious (viral hepatitis A, B, C, D.), (2) dysregulated immunity (autoimmune hepatitis [AIH], primary biliary cirrhosis [PBC], primary sclerosing cholangitis [PSC], and sarcoidosis),3 (3) malignancies, (4) toxic (e.g., alcohol, adverse drug reactions.), (5) metabolic disorders, including nonalcoholic steatohepatitis, and (6) other diseases, such as idiopathic granulomatous hepatitis,4 postinfantile giant cell hepatitis,5 and Budd-Chiari syndrome,6 that are of uncertain etiology or multifactorial in origin. Recurrent infectious and dysregulated immunity diseases pose the most difficult diagnostic challenges and are addressed herein. Some diseases in the remaining categories can also recur, but because they do not usually present diagnostic challenges they are not discussed further.

Immunological Considerations

Immune recognition of differences in major histocompatibility complex antigens triggers a characteristically robust inflammatory response in the first few months after transplantation referred to as early acute rejection.2 Like all other immune responses, acute and especially chronic rejection reactions7, 8 evolve over time and diversify via “epitope spreading.”9 Tissue damage during the initial phase releases cryptic antigens that activate endogenous danger signals. Recipient dendritic cell antigen uptake and self-reactive T and/or B lymphocyte priming10 triggers “autoantibody” production and immunity directed against non–major histocompatibility complex determinants. Some non–major histocompatibility complex cytoplasmic, nuclear, and matrix protein antigens11–14 (reviewed in Graft15–18) are shared by the donor and recipient, whereas others may be donor-specific.


AIH, autoimmune hepatitis; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; HCV, hepatitis C virus; LAR, late-onset acute rejection.

Long-Term Protocol Biopsies

Most programs obtain biopsies when changes in liver tests represent a significant deviation from baseline values. Obtaining protocol allograft biopsies in asymptomatic long-term survivors with normal or near-normal liver tests is controversial. Considerations such as potential morbidity and mortality, cost, inconvenience, use of resources, and potential impact of unexplained histopathological findings should be weighed against potential individual and/or societal benefits.4, 19–24 These include (1) early detection of clinically inapparent disease,19, 24 (2) recognition of nonalcoholic steatohepatitis as a significant cause of cryptogenic cirrhosis in the United States25 but not in England,26 (3) identification of recipients that might be successfully weaned from immunosuppression,27 (4) recognition of late-onset rapid hepatitis C virus (HCV) progression,21 and (5) impact of alcohol use.20

Approximately 75% of biopsies from long-surviving recipients with abnormal liver tests or symptoms show significant histopathological abnormalities.4, 19–23 These abnormalities are usually attributable to recurrent disease or biliary tract strictures, some of which occur as a late complication of preservation injury.4, 19–23 The incidence and significance of histopathological abnormalities in long-surviving recipients without abnormal liver tests or symptoms is dependent on the original disease: up to 25% show significant abnormalities when obtained from recipients with original diseases that commonly recur (e.g., HCV, PBC, AIH).4, 19–23

Even in the absence of recurrent disease, minor histopathological abnormalities appear in approximately two thirds of biopsies obtained from long-surviving asymptomatic recipients with normal liver tests.4, 19–23 These include nodular regenerative hyperplasia changes and thickening/hyalinization of small hepatic artery branches4, 28 (probably side effects of immunosuppression) and “nonspecific” portal and lobular inflammation.4, 22–24 The pathogenesis, significance, and long-term consequences of nonspecific inflammation (e.g., idiopathic posttransplantation hepatitis), portal venopathy, and nodular regenerative hyperplasia are in need of further study.

Recurrent HCV disease progression is significantly more rapid than HCV in native livers. Disease progression rates for recurrent hepatitis B virus, PBC, PSC, nonalcoholic steatohepatitis, and AIH are difficult to study because of the small number of long-term survivors with biopsies and chronic immunosuppression, as well as introduction of new medical therapies. Regardless, nearly all recurrent diseases can potentially cause allograft cirrhosis.

Practical Problems and Approach to Biopsy Interpretation

Most late causes of liver allograft injury are first detected because of abnormalities in routinely monitored liver tests; clinical signs and symptoms are much less common. When signs or symptoms do occur, they are similar to those seen in the general population with the same causes of liver injury. Examples include fever and upper right quadrant pain in ascending cholangitis; fatigue, nausea, vomiting, and jaundice in viral hepatitis; relapsing bacteria in hepatic infarcts, etc.

Many late posttransplantation biopsies show portal-based chronic inflammation with variable interface activity. Subtle histopathological differences relied upon to distinguish among several possible specific causes of dysfunction are not always present or reliable. Occasionally, rendering a definitive diagnosis may not be possible in the early stages of a disorder. A caveat of “features suggestive of early” emphasizes a tentative diagnosis.

Laboratory tests used to establish a diagnosis before transplantation may not have the same significance after transplantation. Antimitochondrial antibodies and antinuclear antibodies often persist after transplantation in patients with PBC or AIH, albeit at lower titers, even without histopathological evidence of recurrent disease. Patients without AIH before transplantation can develop autoantibodies either as a complication of otherwise typical rejection11, 12, 29 or in association with new-onset AIH.30–36 “Non–organ-specific” autoantibodies have been detected in up to 71% of patients after liver transplantation,37 emphasizing the need for clinicopathological correlation.

More than 1 insult can contribute to late posttransplantation dysfunction. Biopsy analysis can help to determine the main component of injury, but careful clinicopathological correlation is needed. Levels of immunosuppression can influence biopsy findings and the severity of recurrent viral hepatitis, AIH, and rejection. For example, late-onset acute rejection (LAR) is often precipitated by inadequate immunosuppression and recipients with AIH and other autoimmune disorders usually require more immunosuppression to prevent rejection and disease recurrence. Too much immunosuppression can trigger cholestatic HCV hepatitis. Lymphoid depletion followed by rapid withdrawal of immunosuppression can precipitate aggressive HCV recurrence.38

Biopsy interpretation should include an assessment of adequacy, systematic examination, and thorough clinicopathological correlation. Adequacy is ultimately the subjective opinion of the pathologist, but in general, at least 6 small portal tracts should be sampled. The findings should then be correlated with the original disease, immunosuppression, liver tests, viral serology, and immunology and radiology findings.

Generalized Criteria

Criteria used to distinguish rejection from AIH can be melded into generalized criteria applicable to other causes of late liver allograft dysfunction,39–56 including: (1) histopathological evidence of liver injury showing a pattern compatible with the diagnosis (liver tests are usually elevated in a pattern consistent with the diagnosis); (2) positive serological, molecular biological, immunological, or radiographic evidence of pathogen or possible cause of injury; and (3) other causes of similar histopathological changes and elevated liver tests, if present, have been reasonably excluded.

Table 1 shows approximate incidences, risk factors, and clinical, immunological, and radiological observations for common causes of late dysfunction. Specific diagnoses can be rendered when these observations are combined with histopathological findings (Table 2), timing and pattern of liver test elevations, and important exclusionary criteria (Table 3). A discussion of histological findings in late posttransplant biopsies and their differential diagnosis follows.

Table 1. Incidence, Risk Factors, and Clinical Observations
DiagnosisIncidence at 5 Years of Recurrent DiseaseRisk Factors for Disease Recurrence and/or Severe Recurrent DiseaseClinical/Immunological/Radiological Observations
  1. Abbreviations: AIH, autoimmune hepatitis; HBV, hepatitis C virus; HCV, hepatitis C virus; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis.

Recurrent AIH∼30%Suboptimal immunosuppression; type I > type II disease; severe inflammation in native liver before transplantation; longer duration of follow-up HLA DR3 or DR4 recipient status may reflect more severe diseaseUsually need higher baseline immunosuprression (see text) HLA DR3 and/or DR4 genotype often present
De novo AIH<5%May be more common in children, but this assumption has been questioned recentlySame as above
Recurrent HBV100% if HBV DNA is positive; less frequent if HBV DNA is negativeAnti-HBc–positive donor Inadequate anti-HBV treatment HBV mutantsRecurrent HBV disease not usually a significant problem because of treatment with effective antiviral drugs
Recurrent HCVNearly universal in those with HCV replication before transplantationHCV RNA in blood helpful in differential diagnosis (>30,000,000 IU/L); increased risk of cholestatic hepatitis Significant acute or chronic rejection usually occurs only in association with relatively low HCV RNA levels (<5,000,000 IU/L)Greater viral burden and more rapid progression of fibrosis than in general population Severity of hepatitis often worse with genotype 1 viruses Variable disease progression Subset of recipients with late-onset rapid progression
Recurrent PBC20%-30%; increases with timeTacrolimus as baseline immunosuppression; living-related donor; steroid and other immunosuppression withdrawal May recur as AIHInitial diagnosis often made via biopsy in asymptomatic recipient with or without increased liver tests
Recurrent PSC20%-30%; increases with timeMale sex; donor–recipient sex mismatch Intact colon at time of transplantation Patients at increased risk of rejectionCholangiographically confirmed biliary strictures occurring >90 days after liver transplantation Mural irregularity, diverticulum-like outpouchings, and an overall appearance resembling PSC Patient and allograft survival not adversely affected up to 5 years; later outcome uncertain
Acute rejectionVariable; <30% of causes of late dysfunctionInadequate immunosuppression Treatment with immune-activating drugs (e.g., interferon) History of autoimmune liver diseaseMuch less common than early after transplantation May be more difficult to treat, perhaps related to delay in diagnosis.
Chronic rejection∼3 %Inadequate immunosuppression Treatment with immune-activating drugs (e.g., interferon) Refractory acute rejection Chronic rejection in a previous failed allograftImportant cause of late dysfunction Most cases occur within first year Does not appear to increase with time after transplantation, but more follow-up is needed.
Idiopathic posttransplantation hepatitis5%-60%; wide variation 5%-15% of patients followed for a minimum of 10 years will develop progressive fibrosis resulting in established cirrhosis Incidence varies widely among centers
Table 2. Histopathologic Features Most Commonly Detected With Various Causes of Late Liver Allograft Dysfunction
Histopathological FeaturesAutoimmune Hepatitis*Acute RejectionChronic RejectionChronic Viral Hepatitis Types B and CPrimary Biliary CirrhosisPSC/BD Strictures
  • NOTE. The histopathological findings in this table should be combined with clinical, serological, radiographic, and important exclusionary criteria listed in Table 2 to arrive at a final diagnosis.

  • Abbreviation: PSC/BD, primary sclerosing cholangitis/bile duct.

  • *

    The same findings apply to recurrent and de novo autoimmune hepatitis.

Distribution, severity, and composition of portal inflammationUsually diffuse; predominantly mononuclear of varying intensity; often prominent plasma cell componentUsually diffuse; variable intensity; mixed “rejection-type” (see text) infiltratePatchy; usually minimal or mild lymphoplasmacyticPatchy; variable intensity; predominantly mononuclear; nodular aggregatesNoticeably patchy and variable intensity; predominantly mononuclear; nodular aggregates and granulomasUsually patchy to diffuse depending on stage; mild neutrophilic, eosinophilic, or occasionally mononuclear predominant
Presence and type of interface activityProminent and defining feature is usually necroinflammatory-type; often plasma cell–richFocally present and mild necroinflammatory typeMinimal to absentVariable; usually not prominent: necroinflammatory- and ductular-typeImportant feature later in disease development: ductular and necroinflammatory-type with copper depositionProminent and defining feature: ductular-type with portal and periportal edema
Bile duct inflammation and damageVariable; if present, involves a minority of bile ductsPresent and usually involves a majority of bile ductsFocal ongoing lymphocytic bile duct damage; inflammation wanes with duct lossVariable; if present, involves a minority of bile ductsGranulomatous or focally severe lymphocytic cholangitis is diagnostic in proper settingPeriductal lamellar edema; “fibrous cholangitis”; acute cholangitis; multiple intra-portal ductal profiles
Biliary epithelial senescence changes and small bile lossAbsent or involves only a minority of ducts/portal tracts, but may be focally severeAbsent or involves only a minority of ductsSenescence/atrophy/ atypia involve a majority of remaining ducts (see text)Absent or involves only a minority of ductsSmall bile duct loss associated with ductular reactionSmall bile duct loss associated with ductular reaction
Perivenular mononuclear inflammation and/or hepatocyte dropoutVariable; can involve a majority of perivenular regions, similar to rejection (see text); may be plasma cell–rich.Variable, if defining feature should involve a majority of perivenular regions; may also show subendothelial inflammation of vein (see text)Usually present, but variableVariable but generally mild; if present, involves a minority of perivenular regionsVariable but generally mild; if present, involves a minority of perivenular regionsAbsent
Lobular findings and necroinflammatory activityVariable severity; rosettes may be present and/or prominentVariable; if present, concentrated in perivenular regionsVariable; if present, concentrated in perivenular regionsDisarray variable; variable severity; necroinflammatory activityMild disarray; parenchymal granulomas; periportal copper deposition and cholatestasis are late featuresDisarray unusual; neutrophil clusters; ± cholestasis
Pattern of fibrosis during progression toward cirrhosisUsually macronodular; posthepatitic patternRareUncommon, if present usually a venocentric pattern; may evolve to biliary pattern over timeUsually macronodular, hepatitic pattern; may be micronodular (see text)Biliary patternBiliary pattern
Table 3. Inclusionary and Exclusionary Criteria for the Diagnosis of Recurrent and New-Onset Chronic Necroinflammatory Diseases After Liver Transplantation and Timing of First Onset and Pattern of Liver Test Elevation
DiagnosisOriginal DiseaseSerology/Molecular Testing*Timing and Liver Injury Test ProfileImportant Exclusionary Criteria
  • NOTE. See Table 1 for compatible histopathological findings.

  • Abbreviations: AIH, autoimmune hepatitis; ANA, antinuclear antibodies; ASMA, anti–smooth muscle antibodies; ALKM, anti-liver–kidney microsomal antibodies; HBV, hepatitis B virus; HCV, hepatitis C virus; PCR, polymerase chain reaction; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis.

  • *

    Timing = usual timing of first onset.

  • Sustained elevation for more than 1 month. Hepatocellular = alanine aminotransferase and/or aspartate aminotransferase > alkaline phosphatase and/or γ-glutamyltranspeptidase. Cholestatic = alkaline phosphatase and/or γ-glutamyltranspeptidase > aspartate aminotransferase and/or alanine aminotransferase.

Recurrent AIHAIHAutoantibodies (ANA, ASMA, ALKM) usually in high titers (>1:160); elevated serum immunoglobulin G>6 months hepatocellularAcute and chronic rejection, HBV, HCV infection, as determined via third-generation ELISA and/or serum PCR
De novo AIHOther than AIHSame as above>6 months hepatocellularSame as above
Recurrent HBV or HCVHBV- or HCV-induced cirrhosisHBV or HCV infection using standard, third-generation serological criteria and/or positive molecular testing for HBV or HCV nucleic acidsUsually 6-8 weeks, but as early as 10 days Usually hepatocellular but may be cholestaticAcute and chronic rejection AIH
Recurrent PBCPBCPositive AMA, but little additional benefit because AMA remains elevated in the majority of patients after transplantation>1 yr CholestaticBiliary tract obstruction/strictures
Recurrent PSCPSCNAUsually >1 yr CholestaticHA thrombosis/stenosis, chronic (ductopenic) rejection, abnormal surgical anatomy, anastomotic strictures alone, nonanastomotic strictures occurring <90 d after liver transplantation, and ABO incompatibility
Acute rejectionNA (see text for risk factors)NAAny time Usually hepatocellular; may be mixed if superimposed on chronic rejectionInadequate immunosuppression usually, but not always present (see text) Important exclusions: biliary tract obstruction/strictures, HBV, HCV, AIH
Chronic rejectionNA (see text for risk factors)NAAny time, but usually <1 yr Cholestatic; rarely hepatocellular in veno-occlusive variant (see text)Inadequate immunosuppression usually, but not always present (see text) Important exclusions: biliary tract obstruction/strictures, HBV, HCV, AIH
Idiopathic posttransplantation hepatitisNonviral and non-autoimmune hepatitisNegative testing for HBV and HCV infection and autoantibodies>1 yr Usually hepatocellularAcute and chronic rejection, all other causes of chronic hepatitis, and biliary tract obstruction/strictures reasonably excluded; all attempts should be made to determine a cause

Late-Onset Acute Rejection.

LAR, which occurs more than several months after transplantation, may show slightly different features than typical acute rejection seen early after transplantation (Fig. 1). Fewer blastic lymphocytes, slightly greater interface activity, less venous subendothelial inflammation, and slightly more lobular activity cause biopsies with LAR to resemble chronic hepatitis.4, 57 LAR can also present as isolated perivenular inflammation and hepatocyte dropout (so-called “central perivenulitis”)58–60 and evolve into typical chronic rejection with ductopenia.61 Subendothelial inflammation of portal or central veins is not a required finding in such cases. LAR, however, is still most commonly characterized by: (1) predominantly mononuclear portal inflammation containing lymphocytes, neutrophils, and eosinophils; (2) venous subendothelial inflammation of portal or central veins or perivenular inflammation; and (3) inflammatory bile duct damage. Previously proposed criteria2 should be used for grading unless LAR presents as isolated central perivenulitis. For these cases, the following descriptors are recommended:

Figure 1.

Composite of late-onset acute rejection occurring more than 1 year after transplantation in a patient with low levels of baseline immunosuppression. (A) Low-magnification view (×20) shows prominent portal inflammation distributed evenly throughout the portal tracts, as well as perivenular mononuclear inflammation. Note also the irregular interface zone around the inflamed portal tracts. (B) Higher magnification (×200) of the portal tract outlined by the rectangle in panel A better illustrates the irregular interface zone and fewer blastic lymphocytes, which causes the biopsy to resemble chronic hepatitis. However, the prevalence and severity of lymphocytic cholangitis (arrows) are much worse than would be expected in chronic hepatitis and point toward acute rejection as the correct diagnosis. (C) Higher magnification (×200) of the central vein designated by the arrow in panel A better illustrates the perivenular mononuclear inflammation, or “central perivenulitis.” Abbreviations: PT, portal tract; CV, central vein.

  • minimal/indeterminate: perivenular inflammation involving a minority of central veins with patchy perivenular hepatocyte loss without confluent perivenular necrosis

  • mild: as above, but involving a majority of central veins

  • moderate: as above, with at least focal confluent perivenular hepatocyte dropout and mild moderate inflammation, but without bridging necrosis

  • severe: as above, with confluent perivenular hepatocyte dropout and inflammation involving a majority of hepatic venules with central-to-central bridging necrosis.

“Minimal” and “mild” cases, as described above, may resolve spontaneously.60 More severe perivenular changes probably warrant more aggressive treatment, but studies of long-term outcome according to therapy are needed to validate such an approach.

Chronic Rejection.

Portal tracts and perivenular regions are primarily affected in chronic rejection, and changes are divided into “early” and “late” stages.1 In a biopsy specimen, the minimum diagnostic criteria are: (1) biliary epithelial senescence changes affecting a majority of the bile ducts with or without bile duct loss; or (2) foam cell obliterative arteriopathy; or (3) bile duct loss affecting >50% of the portal tracts.1

Biliary epithelial senescence changes include cell and nuclear enlargement, multinucleation, uneven nuclear spacing, and cytoplasmic eosinophilia.62 Some small bile ducts may be only partially lined by biliary epithelial cells. Perivenular hepatocyte dropout and central perivenulitis are typical of early chronic rejection.63 Variable perivenular fibrosis occasionally progressing to veno-centric cirrhosis is typical of late chronic rejection.64 Chronic rejection rarely results in a “posthepatitic” pattern of cirrhosis. If this pattern is present, other insults should be reasonably excluded.

The safest approach to a chronic rejection diagnosis in any setting is to review prior biopsies and correlate the histopathological findings closely with the clinical course. The typical scenario usually includes persistent/unresponsive acute rejection and/or inadequate immunosuppression.

Recurrent Diseases and New-Onset Diseases

Hepatitis C Virus.

The predominant features of HCV include mononuclear portal inflammation, often arranged into nodular aggregates, necroinflammatory and ductular-type interface activity, and mild macrovesicular steatosis. Except for an association between steatosis and HCV genotype 3,65 no histopathological features reliably distinguish among different viral genotypes. Lymphocytic cholangitis, if present, involves a minority of bile ducts without ductopenia. Lobular disarray and necroinflammatory activity are usually mild. Confluent or bridging necrosis with recurrent HCV alone is unusual. Central perivenulitis, if present, involves a minority of central veins.

There are two histopathological patterns of severe chronic HCV: (1) aggressive conventional hepatitis with prominent interface activity and (2) fibrosing cholestatic hepatitis. Features of fibrosing cholestatic hepatitis include centrilobular hepatocyte swelling and degeneration; cholestasis, hepatocyte apoptosis, and portal expansion because of a ductular reaction; fibrosis; and a mild mixed portal inflammation.66 Fibrosing cholestatic hepatitis is associated with massive HCV replication (e.g., >40-50 million IU/mL67, 68).

Recurrent and New-Onset or De Novo Autoimmune Hepatitis.

AIH is difficult to distinguish, histologically and conceptually, from rejection. Immune responses against self-antigens constitute an autoimmune response, whereas those against foreign antigens constitute rejection. Donor livers undoubtedly contain non–major histocompatibility complex antigens not expressed in the native liver, and theoretically all forms of AIH after transplantation could be classified as rejection.34, 42 Serological and histological findings used to distinguish AIH from rejection may reflect the nature, density, and location of antigenic targets. There are no conventional clinical tests that differentiate an autoimmune response from rejection, and distinctions based on clinical and histopathological findings may not reflect the true pathogenesis. Some new-onset AIH cases might be attributable to polymorphic expression of glutathione S-transferase T169; transplantation of a mismatched graft into a nonexpressing recipient could trigger rejection that closely resembles AIH.

The International Autoimmune Hepatitis Group70 scoring system and criteria for the diagnosis of AIH in native livers have not been tested in allografts; however, they do provide useful guidelines. AIH is established through a combination of serological, molecular biological, and histopathological findings. Non–organ-specific autoantibodies, a requisite for diagnosis, typically include anti–smooth muscle antibodies and antinuclear antibodies, as well as antibodies to liver kidney microsome type 1.71 Their occurrence implies activation of immune mechanisms possibly involved primarily in disease pathogenesis or collateral responses to liver cell destruction and nonselective antigen release. Autoantibodies after liver transplantation do not establish the diagnosis of AIH, nor are they accurate parameters of inflammatory activity. Their principal value is to direct attention to the possibility of AIH.

The minimum diagnostic criteria for recurrent or de novo AIH in an allograft are: (1) interface hepatitis with portal lymphocytic infiltrates; (2) significant titers (≥1:160) of antinuclear antibodies, smooth muscle antibodies, or antibodies to liver kidney microsome type 1; (3) hyper-gammaglobulinemia; and (4) exclusion of virus-induced or drug-related hepatitis and late acute or chronic rejection. Titers ≥1:160 are unlikely to be nonspecific background reactivities and therefore compel a thorough evaluation for AIH.70

Initial manifestations include lobular hepatitis with hepatocyte rosetting40 that usually evolves into the chronic phase characterized by lymphoplasmacytic portal inflammation with prominent interface activity. Plasmacytic infiltrates characterize AIH, but are not diagnostic requisites. Confluent and bridging necrosis are not uncommon, particularly in patients on suboptimal immunosuppression. Lymphocytic cholangitis, if present, involves a minority of ducts.

Central perivenulitis can occur in acute onset AIH in native livers72–74 and in otherwise typical LAR. In native livers, perivenular hepatocyte injury associated with AIH usually wanes as interface hepatitis appears,75 but the evolution of changes has not been studied in allografts. Panacinar hepatitis is also within the spectrum of histological findings in AIH,70 but a cholestatic form is not recognized.

Idiopathic Posttransplantation Hepatitis.

Idiopathic posttransplantation hepatitis is defined as chronic hepatitis that cannot be ascribed to a particular cause. By definition, bile duct damage and venous endothelial inflammation are not conspicuous. In adults, the prevalence is difficult to determine, because most native diseases have the potential to recur with features of chronic hepatitis. In some centers, up to 40% of adult patients subjected to biopsy more than 12 months after transplantation have unexplained chronic hepatitis.76 A similar prevalence has been observed in the pediatric population, in which recurrent native disease is less of a problem; the frequency of “idiopathic” chronic hepatitis was 20% at 1 year of age, rising to 60% at 10 years of age.77

Cases presenting as central perivenulitis probably represent centrilobular-based acute rejection, or AIH if autoantibodies are also present,57 because allograft dysfunction usually responds to increased immunosuppression.59–61, 78 Some idiopathic posttransplantation hepatitis cases may represent rejection with chronic hepatitic features.79 However, a diagnosis of idiopathic posttransplantation hepatitis does not usually trigger treatment with increased immunosuppression. In some series, as many as 50% of such cases may develop bridging fibrosis or cirrhosis over a period of 10 years.77 This observation supports the need for protocol biopsies and clarification of management policies in those with significant activity.77

Primary Biliary Cirrhosis.

Recurrent PBC findings are nearly identical to those seen in native livers.80, 81 The pathognomonic lesion is noninfectious granulomatous cholangitis in the proper setting, which includes presence of antimitochondrial antibodies and absence of other causes such as infections and biliary strictures. Diagnostic lesions are not always present. Patchy but easily recognizable and severe lymphocytic cholangitis accompanied by biliary epithelial cell eosinophilia, portal lymphoid nodules containing germinal centers, and development of a “biliary gestalt” can also be diagnostic of recurrent PBC in the proper setting. The biliary gestalt includes a ductular reaction at the interface zone combined with portal and periportal fibrosis, small bile duct loss, periportal edema (halo sign), and lysosomal pigment and copper/protein deposition in periportal hepatocytes. Plasma cell–rich periportal hepatitis may be an early marker predictive of later PBC recurrence.82 Nonspecific lobular findings include mild spotty hepatocyte apoptosis, slight sinusoidal lymphocytosis, mild nodular regenerative hyperplasia, and Kupffer cell granulomas.

Primary Sclerosing Cholangitis.

Findings are identical to those described for native livers with PSC and to other causes of biliary strictures. Subtle histopathological clues that suggest low-grade biliary strictures include mild portal edema; mild nonspecific acute and chronic “pericholangitis” often accompanied by a very mild type I ductular reaction; sinusoidal clusters of neutrophils; and centrilobular hepatocanalicular cholestasis. More significant strictures usually cause lamellar periductal edema, increased portal tract ductal profiles, and/or concentric periductal fibrosis.83 Later-stage findings include the biliary gestalt. “Fibro-obliterative duct lesions” are not diagnostic of recurrent PSC, because they can also develop in patients with ischemic cholangitis and reflux cholangiopathy.

Differential Diagnosis

Rejection Versus Chronic Hepatitis.

This commonly encountered and difficult problem has important therapeutic implications.67 Unnecessary augmentation of immunosuppression can accelerate fibrogenesis in chronic HCV or trigger cholestatic hepatitis. Untreated acute rejection can progress to chronic rejection, particularly in interferon-treated recipients.

Mononuclear portal inflammation and lymphocytic cholangitis are features of chronic hepatitis and most cases of LAR. In LAR, however, the portal infiltrate tends to be more diffusely distributed throughout the portal tracts and throughout the biopsy rather than aggregated into nodules in occasional portal tracts, as in chronic hepatitis. In LAR and chronic rejection, lymphocytic cholangitis and/or biliary epithelial senescence changes, respectively, should involve a majority of bile ducts.67 Central perivenulitis involving a majority of central veins also favors rejection. Damage limited to a minority of bile ducts favors acute or chronic hepatitis. Key features of acute and chronic hepatitis are lobular necroinflammatory activity and necroinflammatory and ductular-type interface zone activity, respectively, which are more prevalent and severe than in acute rejection.

Because acute and/or chronic rejection and chronic hepatitis can coexist, the predominant process should be identified. Key features of acute rejection in the context of recurrent HCV are prevalence and severity of mononuclear inflammatory bile duct damage and central perivenulitis. If either feature involves a majority of bile ducts or central veins, then acute rejection is present. However, coexistent acute rejection should be listed as the primary process only when rejection-related changes are obvious. Most such cases are graded as “moderate” according to the Banff schema.67 Chronic rejection in the context of recurrent HCV is recognized by the same features as in allografts without recurrent HCV: small bile duct loss or biliary epithelial senescence or perivenular inflammation and fibrosis involving a majority of bile ducts or hepatic venules, respectively.

Chronic Rejection.

Small bile duct damage and loss and perivenular fibrosis are relied upon for the diagnosis of chronic rejection because arteries with pathognomonic changes are rarely present in needle biopsy specimens.1 Bile duct injury and ductopenia, however, can also be caused by biliary strictures, hepatic artery pathology, adverse drug reactions, and cytomegalovirus. Isolated ductopenia involving less than 50% of portal tracts can be seen occasionally without significant elevations of liver tests.80 Whether these uncommon cases are an early phase or subclinical chronic rejection is uncertain. Angiography showing pruning of intrahepatic arteries with poor peripheral filling and segmental narrowing also supports a chronic rejection diagnosis.84, 85

Perivenular fibrosis can also be caused by suboptimal hepatic venous drainage, adverse drug reactions,86 and the various causes of veno-occlusive disease and Budd-Chiari syndrome in native livers.87 In cases of chronic rejection identified by biliary epithelial senescence, bile duct loss, or perivenular fibrosis alone, non–rejection-related causes of ductal injury and loss or perivenular fibrosis should be reasonably excluded, particularly if the clinical scenario is not typical (Table 1).

Biliary Strictures Versus Acute and Chronic Rejection.

Significant biliary strictures are usually recognized by the biliary gestalt and are reinforced by preferential elevation of γ-glutamyltranspeptidase and alkaline phosphatase. However, a thorough clinicopathological correlation is needed to distinguish among many underlying causes, such as recurrent PSC, ischemic cholangitis due to injury from prolonged preservation or non–heart-beating donors, imperfect biliary anastomoses, inadequate arterial flow, and antibody-mediated rejection.4, 88–92 Periportal hepatocyte copper deposition signals chronic bile flow impediments.

Mononuclear portal inflammation usually favors acute rejection, whereas neutrophilic or eosinophilic portal inflammation, late after transplantation, favors biliary stricturing. However, chronic low-grade biliary strictures can occasionally cause predominantly mononuclear portal inflammation. Ductopenia in some portal tracts accompanied by a ductular reaction should raise the suspicion of biliary strictures. Cholangiography and/or angiography may be required to distinguish between chronic rejection and biliary strictures. Acute rejection occurring more than 6 months after transplantation is unusual in adequately immunosuppressed recipients. Therefore, checking baseline immunosuppressive drug levels and the liver test profile often point to the need for cholangiography before increased immunosuppression.

Acute and Chronic Rejection Versus Primary Biliary Cirrhosis.

In acute rejection, portal inflammation and lymphocytic cholangitis are usually more diffusely distributed throughout the portal tracts and the biopsy and typically involve small bile ducts (<20 μm). Portal inflammation and lymphocytic cholangitis in recurrent PBC are typically patchy and involve medium-sized bile ducts (>40-50 μm). In the absence of a pathognomonic lesion, recurrent PBC is most commonly recognized by the biliary gestalt occurring in the absence of mechanical biliary strictures. This gestalt is unusual in rejection. Central perivenulitis is not a feature of PBC.

Central Perivenulitis.

LAR can manifest primarily as central perivenulitis.59–61, 63, 93–96 Because of its association with severe acute rejection2 and transition to early chronic rejection,63 central perivenulitis is sometimes portrayed as a poor prognosis lesion, but this is not necessarily correct.59, 60 As in native livers, central perivenulitis in allografts has several causes (Fig. 2), including various forms of rejection (pure perivenular rejection and early chronic rejection), early autoimmune hepatitis,72, 74, 97 compromised afferent or efferent blood flow,73, 87, 98 and adverse drug reactions. Perivenular rejection can be missed clinically and present later as ascites because of a Budd-Chiari syndrome or veno-occlusive disease.63, 64, 93, 96, 99

Figure 2.

Approach to biopsies showing posttransplantation central perivenulitis. In cases with no or minor portal inflammation, the differential diagnosis includes acute rejection, chronic rejection, and prediagnostic autoimmune hepatitis. If none of these changes is present, and vascular imaging is normal, the lesion is likely to represent a form of acute rejection. Cases with a more extensive portal inflammatory infiltrate have a similar differential diagnosis. It remains unclear whether idiopathic posttransplantation hepatitis is a form of rejection, and how it is related to pure central perivenulitis. Follow-up biopsies also frequently provide important diagnostic findings. Abbreviations: AR, acute rejection; CR, chronic rejection; AIH, autoimmune hepatitis.

An acute rejection diagnosis is obvious when central perivenulitis occurs in association with other portal-based changes typical of acute rejection; the severity is graded according to standard criteria.2 Acute rejection is also the most likely diagnosis when central perivenulitis involves a majority of central veins with minimal or absent portal inflammation, except if the original disease was AIH. In this situation, isolated central perivenulitis may represent early recurrent AIH34, 42, 74, 75 or new-onset AIH. In native livers presenting with acute AIH central perivenulitis, chronic portal inflammation and interface activity usually develop over time.72, 74, 97 Therefore, in allografts, re-examination of the native liver histopathology, serological studies for autoantibodies, and close follow-up for the development of changes more typical of chronic hepatitis75 are warranted. Because increased immunosuppression effectively treats either rejection or AIH, any differences in assigned diagnoses may be semantic. Hepatic vein outflow obstruction and ischemia can also cause centrilobular necrosis, but any associated lymphocytic inflammation is usually minimal.

Mild focal central perivenulitis can coexist with other causes of late dysfunction. In such cases, central perivenulitis probably represents a focal alloreaction, because similar changes are rarely—if ever—seen with the same disorders in native livers. Therefore, we recommend mentioning its presence or suggesting a diagnosis of “indeterminate for rejection,” unless a majority of central veins are involved.

Distinguishing Among the Various Causes of Chronic Hepatitis.

Determining a specific cause of chronic hepatitis is not always possible, but subtle differences can suggest a specific etiology. Plasma cell and aggressive interface activity and confluent perivenular or bridging necrosis are suggestive of AIH, macrovesicular steatosis is suggestive of HCV, and viral inclusions are seen only in hepatitis B virus. Because potentially distinguishing features are inconsistently present and not entirely reliable, determining the underlying cause of acute and/or chronic hepatitis should be based on a complete clinicopathological evaluation (Tables 2 and 3). Steatohepatitis can coexist with other causes of injury.

Cholestatic or Biliary Disease Versus Chronic Hepatitis.

A single granulomatous duct destructive lesion is diagnostic of PBC in the proper setting. Infectious causes of granulomatous cholangitis should be excluded, but they are uncommon. Portal granulomas without granulomatous cholangitis have been reported in native livers with HCV.100 In the absence of pathognomonic lesions, recurrent PBC or PSC is most commonly distinguished from chronic hepatitis by a biliary gestalt.

Cholestatic viral hepatitis can be difficult to distinguish from biliary strictures with or without hepatic artery thrombosis. Portal edema and portal—rather than periportal—neutrophilia are common in biliary strictures. Cholangiolar proliferation and acute cholangiolitis without portal edema is more characteristic of cholestatic hepatitis. Lobular disarray and hepatocellular swelling and apoptosis are more usual for cholestatic viral hepatitis.

Banff Working Group

Anthony J. Demetris, University of Pittsburgh Medical Center, Pittsburgh, PA (Chairman)

Oyedele Adeyi, Princess Margaret Hospital, University of Toronto, Toronto, Canada

Chris O. C. Bellamy, University of Edinburgh, Edinburgh, Scotland, UK

Andrew Clouston, University of Queensland, Brisbane, Australia (Co-Chairman)

Frederic Charlotte, Pitié-salpetriere Hospital, Paris, France

Albert Czaja, Mayo Clinic, Rochester, MN

Ierachmiel Daskal, Albert Einstein Healthcare Network, Jefferson Health System, Philadelphia, PA

Magda S. El-Monayeri, Wadi El Neel Hospital, Ain Shams Faculty of Medicine, Cairo, Egypt

Paulo Fontes, University of Pittsburgh Medical Center, Pittsburgh, PA

John Fung, Cleveland Clinic Foundation, Cleveland, OH

Bruno Gridelli, ISMETT, Palermo, Italy

Maria Guido, Istituto di Anatomia Patologica, Padova, Italy

Hironori Haga, Kyoto University Hospital, Kyoto, Japan

John Hart, University of Chicago, Chicago, IL

Eva Honsova, Institute for Clinical and Experimental Medicine, Prague, Czech Republic

Stefan Hubscher, University of Birmingham, Birmingham, UK (Co-Chairman)

Tomoo Itoh, Hokkaido University Hospital, Sapporo, Japan

Nirag Jhala, University of Alabama, UAB Hospital, Birmingham, AL

Patricia Jungmann, University of Pernambuco, Recife, Brazil

Urmila Khettry, Lahey Clinic Medical Center/Tufts University School of Medicine, Burlington, MA

Charles Lassman, David Geffen School of Medicine at the University of California–Los Angeles, Los Angeles, CA

Saverio Ligato, Hartford Hospital, Hartford, CT

John G. Lunz III, University of Pittsburgh Medical Center, Pittsburgh, PA

Amadeo Marcos, University of Pittsburgh Medical Center, Pittsburgh, PA

Marta Ida Minervini, ISMETT, Palermo, Italy

Johan Mölne, Sahlgrenska University Hospital, Göteberg, Sweden

Mike Nalesnik, University of Pittsburgh Medical Center, Pittsburgh, PA

Imad Nasser, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA

Desley Neil, Queen Elizabeth Hospital, University of Birmingham, UK

Erin Ochoa, University of Pittsburgh Medical Center, Pittsburgh, PA

Orit Pappo, Hadassah University Hospital, Jerusalem, Israel

Parmjeet Randhawa, University of Pittsburgh Medical Center, Pittsburgh, PA

Finn P. Reinholt, Rikshospitalet University Hospital, Oslo, Norway

Phil Ruiz, University of Miami Medical Center, Miami, FL

Mylène Sebagh, Universite Paris XI, Hopital Paul Brousse, Villejuif, France

Marco Spada, SMETT (Istituto Mediterraneo per Trapianti e Terapie Alta Specializzazione), Palermo, Sicily

Aurelio Sonzogni, Ospedali Riuniti, Bergamo, Italy

Athanassios C. Tsamandas, University of Patras School of Medicine, Patras, Greece

Annika Wernerson, Karolinska University Hospital, Stockholm, Sweden

Tong Wu, University of Pittsburgh Medical Center, Pittsburgh, PA

Funda Yilmaz, Ege University School of Medicine, Bornova, Izmir, Turkey