Liver allograft pathology: approach to interpretation of needle biopsies with clinicopathological correlation

Laboratory Mediine Program, University Health Network/ University of Toronto, Toronto, Ontario, Canada Correspondene to: Maha Guindi, Laboratory Mediine Program, University Health Network/University of Toronto, 200 Elizabeth Street, Eaton 11-444, Toronto, Ontario, Canada M5G 2C4; maha. guindi@uhn.on.a Aepted 1 Otober 2009 Published Online First 9 November 2009 Liver allograft pathology: approah to interpretation of needle biopsies with liniopathologial orrelation O Adeyi, S E Fisher, M Guindi ABSTRACT The spetrum of diseases enountered in post-transplant liver pathology biopsies is broad. In this review, these have been divided as belonging to one of three ategories: (1) new-onset/de novo post-transplant abnormalities (early and late), (2) rejetion, and (3) reurrene of original disease. The linial and pathologial features of the entities making up eah ategory, with the relevant differential diagnosis and overlaps between and within these groups, are disussed and illustrated. Reurrent or de novo neoplasms make up a fourth ategory not inluded in this review. Early newonset onditions are mostly related to surgial ompliations, donor fators and ishaemia to the graft. These inlude reperfusion/preservation injury, lipopeliosis, smallfor-size-syndrome, biliary sludge syndrome and hepati artery thrombosis. The various forms of rejetion (ellular, hroni, antibody-mediated, and late atypial rejetion) are detailed. Most hroni liver diseases an and do reur in the graft. They may display features that overlap with de novo onditions (eg, primary slerosing holangitis versus hroni rejetion). As with most ases of allograft biopsy interpretation, aurate diagnosis rests with areful orrelation of histologial features with linial, imaging and laboratory findings, and often omparison with previous sequential and follow-up biopsies. Lateonset new diseases inlude biliary stritures, idiopathi hroni hepatitis and de novo autoimmune hepatitis, among others. This review provides a pratial approah to the interpretation of these hallenging biopsies. Seleted diffiult senarios or onundrums are identified and disussed in the relevant setions. The spetrum of pathologial findings and proesses seen in the liver allograft is vast and transends non-transplant-related and post-transplant-related pathology. Suh pathology may be enountered in post-transplant liver biopsies and in the exised failed liver allografts. Traditionally post-transplant pathology has been divided anatomially or as to early or late ourrene in the posttransplant period. These approahes, while very useful, tend to generate long lists that may be diffiult to remember. We have attempted to distil the questions posed to the pathologist interpreting suh biopsies into thee broad ategories: (1) is there rejetion (2) is there reurrene of the underlying liver disease for whih transplantation was needed, and (3) is there a new proess ourring post transplant apart from rejetion? Needless to say, these are ompliated patients and two or more liniopathologial proesses may oexist. The pathologist plays an important role in My approah defining these proesses, espeially sine the patterns of liver enzyme abnormalities and other linial parameters leading to a liver biopsy are not always lear-ut in differentiating between diverse onditions potentially affeting the allograft, and whih not infrequently require diametrially opposite interventions. Most diagnostially important pathologial studies an be ompleted on routinely proessed formalin-fixed, paraffin-embedded tissue setions. Routinely two H&E-stained slides are prepared from eah biopsy, eah of whih ontains three or four setions. Trihrome, iron, periodi aid Shiff with diastase (PASD), and any other histohemial or immunohistohemial stains are ordered after reviewing the H&E findings. Immunofluoresene staining, to exlude antibody-mediated rejetion, optimally requires fresh frozen tissue, but C4d staining on formalin-fixed, paraffin-embedded tissues an add important anillary information in ases of suspeted antibody-mediated rejetion. 1 In the post-transplant setting, immunohistohemistry is often performed in liver setions for loalisation of viral antigens, identifiation of bile dut epithelium and post-transplant lymphoproliferative disease (PTLD) immunophenotyping. INDICATIONS FOR LIVER ALLOGRAFT BIOPSY Follow-up of transplanted solid organs involves monitoring for evidene of graft injury. Liver enzymes hanges are quite sensitive to hepatoellular (transaminases) or biliary (alkaline phosphatase) injuries. In some ases, abnormalities of liver funtion (bilirubin, albumin and oagulation parameters) are present, either from failure to normalise post-transplant or as a result of severe or advaned-stage post-transplant injuries. In most ases, liver allograft biopsies are performed in response to hanges in liver enzyme levels, abnormality in one or more liver funtion parameters, imaging abnormalities or funtional abnormalities, to follow-up an earlier biopsy, or as part of a protool that requires time-speifi biopsies. The pathologist s first role is to understand the linial question(s) the biopsy is meant to answer. Speifi indiations for liver allograft biopsy in an individual patient typially depend on the age of the graft (ie, time from transplant grafting) and they an be divided into early and late periods. Early graft dysfuntion refers to hanges ourring within the first 3 months of transplantation, while late hanges refer to those ourring after 6 months. 2 3 The period of 3 6 months represents J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 47

an intermediate time, when early and late hanges overlap. Summaries of the indiations for allograft biopsy in early and late graft periods are given in boxes 1 and 2. The entities to onsider in relation to eah of these indiations are listed, to be followed later by the histopathologial harateristis of some of these the entities. EARLY NEW ONSET DISEASES/INJURIES IN THE IN LIVER ALLOGRAFT Preservation and reperfusion injury At the time of organ harvesting, preservation injury may our. This refers to tissue damage ausing graft dysfuntion immediately after transplantation. Fators that ontribute to preservation injury inlude donor and reipient hypotension and other auses of warm ishaemia, old ishaemia during organ preservation, and reperfusion injury. Warm ishaemia ours when the organ is maintained at body temperature but is inadequately perfused with blood. It Box 1: Early indiations for liver allograft biopsy and onsiderations Worsening or failure of liver funtion or enzymes to normalise post-transplant (primary or seondary non-funtion): Tehnial problems (anastomoti: dut or vasular; nonanastomoti vasular fators: eg, hepati artery thrombosis) Immunologial (ellular rejetion, ABO inompatibility/ antibody-mediated rejetion) Donor fators ( marginal grafts inluding fatty liver, long warm and/or old ishaemi period; small-for-size syndrome in live donor grafts) Extreme preservation/reperfusion injury Rise in liver enzymes after initial fall (or unsatisfatory nadir): Immunologial fators (rejetion) Infetion (new or reativated) Delayed manifestation of anastomoti problems Adverse drug reation Reurrene of primary disease Donor fators Post-transplant lymphoproliferative disease Less than expeted normalisation of liver enzymes following a treated event: Wrong initial linial and/or pathologial diagnosis Corret initial diagnosis, but no response to treatment Corret initial diagnosis, but missed or unmasked other pathology Adverse reation to mediation Patient non-ompliane Follow-up to a prior biopsy: Compare response to prior intervention, progression and QA prior biopsy Other fators dependent on indiation for follow-up biopsy Abnormalities of post-transplant imaging: Poor flow (ishaemi parenhymal injury, vasular thrombi, bile dut nerosis/ishaemi holangitis, outflow obstrution, sinusoidal obstrution) Colletions (haematoma, absess, infart, neoplasm) Protool biopsy (time defined): Compare with prior biopsies if available Doument any pathology or absene of any Fibrosis staging preferentially damages hepatoytes. However, if restrited to less than 120 min in duration, it is not usually problemati. 4 5 Warm ishaemia ours in livers harvested from ardia death donors as well. Cold ishaemia ours during storage of the liver in preservation fluid and ie bath immersion. It preferentially damages sinusoidal endothelial ells. 6 7 Cold ishaemia auses loss of mitohondrial respiration and, onsequently, adenosine triphosphate depletion. 6 8 10 Eventually there is lifting of the sinusoidal endothelial ells away from the underlying matrix with loss of sinusoidal mirovasular integrity and funtion. The pathophysiology of preservation/reperfusion injury has been reviewed elsewhere. 1 Muh of the injury that ours with liver preservation is attributable to the reperfusion phase. 6 8 10 A asade of proesses is triggered that leads to an imbalane of vasoonstritors over vasodilators an additional important fator that ontributes to miroirulatory failure. 6 8 10 Reperfusion injury thus manifests in the biliary tree and in the hepatoyti parenhyma. Bile dut ells are diretly suseptible to preservation and reperfusion injury. The biliary sludge syndrome is a aused by the pathophysiologial mehanisms relevant to preservation/reperfusion injury and wound healing in the biliary tree. 11 12 Biopsy appearanes of this will be disussed later in the setion on biliary ompliations. With regard to the hepati parenhyma, biopsy samples obtained within hours of omplete revasularisation are also referred to as post-perfusion biopsies. In severe preservation/ reperfusion injury, 7 there is zonal or onfluent oagulative nerosis, sometimes with periportal or bridging nerosis, and severe neutrophili exudation may be seen. The subapsular parenhyma is espeially suseptible to damage and drying. Histologially, this area may show a more severe pathologial proess than the deeper parenhyma and might not be representative. Biopsies taken a few days after transplantation may show mild hepatoellular injury, suh as mirovesiular steatosis, rounding-up of hepatoyte ytoplasm with detahment from adjaent hepatoytes, and hepatoellular swelling. 7 13 Repair responses in suh ases are in the form of inreased hepatoellular mitosis, thikening of the ell plates and nulear enlargement. Mild zone 3 hepatoellular swelling and analiular holestasis may be present. Portal inflammation and dutular reation at the portal/periportal interfae are usually absent in mild injury. In more severe injury, if hepatoellular nerosis was mainly in zone 3, entrilobular hepatoyte dropout is seen. The adjaent viable zone 2 hepatoytes proliferate to restore the liver parenhyma, and mitoses are seen. If periportal nerosis and bridging nerosis are present, the parenhymal ollapse triggers dutular reation 7 13 that an link adjaent portal trats and distort the arhiteture. More severe injury is also usually aompanied by entrilobular hepatoellular swelling, and analiular and holangiolar holestasis. 7 13 The pathologial differential diagnosis inludes sepsis, biliary obstrution, antibody-mediated rejetion and holestati hepatitis. Correlation is needed in suh ases with the linial history (eg, donor age, donation after ardia death liver, details of old and warm ishaemi times, operative note and mirobiologial studies). Distinguishing between preservation injury and obstrution/ holangitis (table 1) requires areful examination of the bile duts loated within the portal trat onnetive tissue and omparing them with the dutules loated at the interfae zone. 48 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

Box 2: Late indiations for liver allograft biopsy and onsiderations New-onset abnormality in liver funtion/rise in liver enzymes from baseline: Reurrent disease Infetion (new or reativated) Immunologial (ellular rejetion, dutopeni rejetion) De novo post-transplant neoplasm (post-transplant lymphoproliferative disease, other) Reurrent neoplasm (usually hepatoellular arinoma) Adverse drug reation Newly aquired liver disease (eg. de novo hepatitis or any other form of liver disease seen in native livers) Late anastomotial ompliations (eg, biliary striture) Vasulopathies (hroni rejetion related, sinusoidal obstrution syndrome, irrhosis) Metastati neoplasm Liver involvement by another systemi disease Less than expeted liver enzymes normalisation following a treated event: Wrong initial linial and/or pathologial diagnosis Corret initial diagnosis, but no response to treatment Corret initial diagnosis, but missed or unmasked other pathology Adverse reation to mediation Patient non-ompliane Follow-up to a prior biopsy Imaging abnormalities: Neoplasm (primary, reurrent, metastati) Non-neoplasti mass lesions Protool (time defined): Compare with prior biopsies if available Doument any pathology or important negatives Fibrosis staging In obstrution or holangitis, there is onentri peridutal lamellar oedema, aompanied by neutrophils within the lumen or infiltrating between biliary epithelial ells. These bile dut hanges are not seen in preservation injury. There is aute periholangiolitis in preservation injury. Both disorders an show marked zone 3 hepatoanaliular and/or holangiolar holestasis and intralobular neutrophil lusters. 1 Lipopeliosis Lipopeliosis is a lesion that ours in the early post-transplant period. It ours when the engrafted donor liver is fatty, and is seen in approximately 5% of transplants. 14 15 The liver sinusoids have the appearane of being engorged with fat. The mehanism by whih the lesion develops is that hepatoyte nerosis ours in a steatoti graft after transplantation due to ishaemia or preservation injury. The fat globules are then released from the injured hepatoytes and are sequestered in the sinusoids and/or the spae of Disse. 16 The linial outome of lipopeliosis an vary greatly and probably depends on the severity of the underlying hepatoellular nerosis that aused the release of fat droplets from hepatoytes in the first plae. 15 Lipopeliosis may be assoiated with prolonged post-transplant holestasis, 15 whih, as disussed earlier, is an important indiation for liver biopsy in the early post-transplant period. The linial and pathologial ourse of the reported ases suggests that Table 1 Differential diagnosis of biliary striture Histologial Preservation feature Biliary striture injury Aute rejetion Portal inflammation Bile dut epithelium Perivenular mononulear inflammation Dutular reation Peridutal oedema Neutrophils and bile duts Periportal arhitetural ollapse Parenhyma Predominantly neutrophili Relatively normal nuleus-toytoplasm ratio Mild non-speifi inflammation Lymphoytes, plasma ells, and eosinophils (whih may predominate when patients are treated with ortiosteroid-sparing immunosuppressive regimens) Reative hanges; inreased nuleus-to-ytoplasm ratio Absent Absent Present Usually present Usually present Intraepithelial and intraluminal neutrophils may be present in interlobular duts Absent Centrilobular holestasis in hepatoytes and analiuli; small lusters of neutrophils in lobules may be seen May be prominent if biliary sludge syndrome present Neutrophili periholangitis (if severe) May be present in severe injury Zonal onfluent nerosis early. Hepatoellular, swelling, rounding up, entrilobular holestasis in hepatoytes and analiuli Usually absent Absent No; dut injury by lymphoytes seen infiltrating biliary epithelium Usually absent lipopeliosis by itself is reversible and not toxi to the liver but 14 15 is indiative of a more severe form of preservation injury. In our experiene, the lesion is easily detetable when florid, but an be very subtle when mild, or when the biopsy is done later on in the ourse of the lesion, whih by then may have started to resolve. Thus the pathologist needs to remember to think of the lesion and searh for it. Trihrome stains help to make the extruded fat droplets stand out in ontrast against the darker-staining surrounding hepatoytes (fig 1). Fator-VIIIrelated antigen and type IV ollagen immunoperoxidase stains help to delineate the ontours of dilated sinusoids, or may show that fat droplets are present just outside the sinusoid, in the spae of Disse, and are ompressing the sinusoid. CD68 immunoperoxidase stain demonstrates the ytoplasm of marophages surrounding the empty spaes that represent the 15 16 extruded fat droplets, indiating that the fat is no longer within the hepatoyti ytoplasm. Small-for-size graft syndrome The portal hyperperfusion (PHP) or small-for-size graft syndrome (SFSS) is a ompliation that ours primarily in living donor or redued-size liver allografts. This ompliation ours when the transplanted donor segment is less than 30% of the standard or expeted liver volume of the reipient, or less than 0.8% reipient body weight, 17 20 that is when a transplanted liver is not large enough to aommodate the markedly inreased portal vein blood flow. Patients with irrhosis oming to liver transplantation have markedly inreased portal blood flow. 21 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 49

The arterial buffer response regulates a balaned portal vein and hepati artery inflow. 21 23 There is reiproal regulation between portal vein and hepati arterial inflow. Inreased portal venous flow diminishes hepati artery flow, whereas dereased portal flow inreases hepati artery flow. A onstant release of adenosine, a vasodilator substane, among the hepati arterioles and portal venules maintains balaned inflows. Inreased portal flow dereases loal adenosine onentrations resulting in hepati artery branh onstrition and a redution in arterial flow. This is observed in PHP/SFSS liver allografts. Conversely, dereased portal flow results in dereased adenosine and hepati artery vasodilatation. 21 22 However PHP/SFSS also ours following transplantation of whole adaveri livers and partial allografts that are greater than 0.8% body weight, thus fators other than graft size may be at play. The findings of Demetris et al suggest that portal hyperperfusion, venous pathology and the arterial buffer response ontribute to early and late linial and histopathologial manifestations of the SFSS. 24 The features of SFSS may be divided into early and late. Early findings inlude: portal hyperperfusion resulting in portal vein and periportal sinusoidal endothelial denudation and foal haemorrhage into the portal trat onnetive tissue. This dissets into the periportal hepati parenhyma when severe; and poor hepati arterial flow and vasospasm. In severe ases, this an result in funtional dearterialisation, ishaemi holangitis and parenhymal infarts. Late sequelae seen in exised grafts that survive the earlier insults are small portal vein branh thrombosis with oasional luminal obliteration or reanalisation, nodular regenerative hyperplasia and biliary stritures. 24 Thus, SFSS results in hanges that are present in the peripheral and entral liver. Sine ore biopsies sample the peripheral liver parenhyma, it follows that not all features of SFSS will be aptured in a ore biopsy. In peripheral ore needle biopsies, affeted grafts most ommonly show the following triad: entrilobular hepatoanaliular holestasis, entrilobular hepatoyte mirovesiular steatosis, and a dutular reation at the interfae zone. 24 Figure 1 (a) Low-power view showing empty spaes representing extruded fat droplets that appear to be engorging sinusoids (Masson trihrome stain 6100). (b) Higher magnifiation of liver parenhyma showing findings similar to those in (a). Note that the appearanes of nulei abutting the extruded fat droplets are not those of hepatoyte nulei but rather of marophage nulei (H&E 6200). () CD68 immunostain showing immunoreativity (brown olour) in the ytoplasm abutting and surrounding the extruded fat droplets, onfirming their loation within marophages (6400). (d) Type IV ollagen immunostain showing linear staining of sinusoidal wall (arrow), with adjaent fat droplet (*) loated outside the sinusoid in the spae of Disse (6400). However venous findings are unommon in peripheral ore needle biopsies. 24 Detetion in the liver ore biopsy of the above mentioned in the proper setting suh as early post transplant merits a reommendation to the liniian to investigate this possibility. It should be noted that the zone 3 hanges and dutular reation are not speifi for SFSS. The differential diagnosis would also inlude suboptimal arterial flow beause of hepati artery thrombosis or bile dut strituring not related to the SFSS, and systemi auses suh as sepsis with or without systemi hypotension. Hepati artery thrombosis The liver allograft, ontrary to the native liver, is devoid of a ollateral arterial irulation, espeially early post transplantation, thus suseptibility to ishaemi injury is inreased. Although with improving surgial tehniques the inidene of hepati artery thrombosis post-transplant has dramatially improved, it remains the most frequent ause of vasular ompliations after liver transplantation. 25 27 Due to their predominant or exlusive dependene on arterial supply, the strutures most ommonly affeted inlude are the hilum and large bile duts, whih are not routinely sampled in the liver biopsy. Peripheral ore needle biopsies thus may show variable hanges, but are not always reliable for establishing a diagnosis of hepati artery thrombosis. 13 Findings an range from ompletely normal to marked entrilobular hepatoyte swelling (later entrilobular hepatoellular atrophy and sinusoidal widening), dutular reation, with or without bile plugs, and aute holangiolitis or frank oagulative nerosis. In some ases, spotty aidophili nerosis of hepatoytes, so-alled ishaemi hepatitis, an mimi aute viral hepatitis, while dut ishaemia often leads to biliary trat injury and strituring (disussed in Biliary ompliations). 1 REJECTION As in other solid organ transplants, liver allografts are prone to immunologially mediated rejetion, but the roles played by the 50 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

major histoompatibility omplex (MHC) antigens are not as well defined, 28 and many programmes do not perform donor reipient human leuoyte antigen omparison or ross-mathes prior to transplant. Types of rejetion in liver allografts inlude aute ellular (ell-mediated) rejetion, atypial features assoiated with late ellular rejetion, hroni rejetion, and antibody-mediated rejetion. Antibody-mediated rejetion an be hyperaute (very rare outside of ABO mismath) or it may our days to weeks post-transplant. The parameters for reognising this entity, as disussed later, are largely dependent on the demonstration of relevant irulating donor-speifi antibodies. Aute ellular rejetion Aute ell-mediated rejetion (ACR) remains the ommonest ause of early graft dysfuntion, with inidene ranging from 24 80%, with a mean of 49.8%. The reported inidene often inludes ACR diagnosed linially, with or without onfirmatory biopsies. The Banff 1997 doument defining riteria for soring ACR on a sale of 0 9 is widely used and the total sore of all the rejetion features present in a given biopsy is known as the Rejetion Ativity Index (RAI) (table 2). At the University Health Network Laboratory Mediine Program, Toronto, Ontario, Canada, only 8.11% of 887 liver allograft biopsies performed between July 2007 and June 2009 showed histopathologial features of rejetion, 94.4% of whih were either mild (RAI 3 4) or moderate (RAI 5 6) rejetion. Severe rejetion (RAI >7) was less ommonly seen, representing only 5.6% of biopsies showing rejetion (unpublished data). Fators determining the inidene of ACR inlude site of transplant, type of immunosuppression, perioperative fators (ishaemia, infetions), type of post-transplant surveillane, donor harateristis (inluding age, adaveri versus living, et), and other fators. 29 34 The definition of ACR is not based on time of ourrene from transplant, but rather on harateristi morphologial hanges. Although most ACR our in the early post-transplant period, late-ourring ACR, up to several years post-transplant, is not unommon. As disussed earlier, late ACRs are those ourring 6 months post-transplant or later, and have been estimated to our in 6 10% of adult patients, and are more likely to show atypial histopathologial features, be resistant to treatment, require resue therapy, or progress to dutopeni 3 31 35 37 rejetion. The atypial features in ACR are disussed below. Histopathologial features of ACR The 1997 Banff soring system for liver allografts is the most widely employed soring system for liver allograft injuries, inluding ACR, 38 and will be employed throughout this disussion. ACR is an immunologially mediated injury direted at speifi targets, (ie, bile dut epithelium and vasular endothelial ells). 39 41 This understanding undersores the resulting morphology. Therefore inflammatory infiltrates of ellular rejetion are to be sought and found around these targets, namely portal trats and in perivenular areas of zone 3. Notably the lobular regions between the portal trat and zone 3 venules show no signifiant involvement by the immune effetor ells; this is a potentially helpful fator in differentiating typial ACR from hepatitis, where lobular inflammation with evidene of hepatoellular injury and death/apoptosis would be expeted with absent or not-so-prominent endothelial and bile dut injury (table 2). In ACR, sine the portal vein Table 2 Summary of the Banff 1997 riteria for soring aute ellular rejetion in liver allografts Parameter sored Criteria* RAI sore Portal inflammation Bile dut injury/ inflammation Venous phlebitis Inflammation in minority of portal trats, not 1 expanding and mostly lymphoyti Inflammation in and expanding majority or all portal 2 trats, mixed lymphoyti inluding oasional blast/ativated lymphoytes and neutrophils, eosinophils Inflammation in and expanding majority or all portal 3 trats, mixed lymphoyti inluding numerous blast/ ativated forms and neutrophils, eosinophils, with spillover to interfae/periportal hepatoytes Inflammation affeting dut epithelium in minority of 1 portal trats with mild evidene of epithelial injury suh as inreased nuleoytoplasmi ratio, or irregular spaing in epithelium Inflammation affeting dut epithelium in most portal 2 trats with marked evidene of epithelial injury in few duts, suh as inreased nuleoytoplasmi ratio, ytoplasmi vauolisation, disordered polarity and/or irregular spaing in epithelium Inflammation affeting dut epithelium in most portal 3 trats with marked evidene of epithelial injury in most duts, suh as inreased nuleoytoplasmi ratio, ytoplasmi vauolisation, disordered polarity and/or irregular spaing in epithelium; outright dut nerosis an be seen in some duts Subendothelial lymphoytes in some but not the 1 majority of portal and/or hepati venules Subendothelial lymphoytes in most portal and/or 2 hepati venules The table is modified from the Banff shema for grading liver allograft rejetion: an international onsensus doument. 38 *Key features differentiating one sore in eah parameter from the next higher sore are underlined. RAI, Rejetion Ativity Index. endothelium and bile duts are targets, the infiltrates, exept in the more severe forms, tend to luster around these targets with little to no spillover to the lobule through the interfae hepatoytes. When the terminal hepati venules are involved, the infiltrates are seen under the endothelial ells (endotheliitis, phlebitis), but in the more severe forms of rejetion these infiltrates involve the perivenular parenhyma, with or without hepatoellular nerosis. The three histologial parameters underlying ACR are eah sored on a sale of 0 3 to give a total RAI on a sale of 0 9, using the Banff 1997 riteria. 38 These parameters are portal inflammation, bile dut injury, and portal and/or terminal hepati venule endothelial injury; the soring parameters are harted in table 2, as defined by the Banff 1997 riteria. 38 Portal inflammation Portal infiltrates in ACR are usually but not always mixed, and may inlude ativated lymphoytes (inluding blast forms), eosinophils and neutrophils. These infiltrates range from mild to severe depending on the density, and an involve few to all sampled portal trats. As seen in table 2, the density and extent of portal trat involvement both fator into apportioning a sore for this aspet of ACR, examples of whih are illustrated in fig 2. Bile dut injury Bile dut injury in ellular rejetion is haraterised by the presene of inflammatory ells within dut epithelial ells, but assoiated with evidene of epithelial injury, suh as high N:C J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 51

ratio, variation in nulear size, ytoplasmi vauolisation and disruption of lumen. Outright nerosis of duts an be present in the most severe injuries (fig 2). Endothelial injury Injury to the portal and/or terminal hepati vein endothelium omprises inflammatory infiltrates beneath the endothelium, variably referred to as endotheliitis or phlebitis. The authors prefer the term phlebitis as this seems to more aptly desribe the proess, and is used in the remaining portion of this artile. Phlebitis an be seen in non-rejetion proesses inluding reurrent hroni hepatitis C virus (HCV); however, the rejetion-related phlebitis is usually assoiated with morphologial evidene of endothelial injury; this ranges from mere lifting of the endothelium that onsequently assumes a more plump shape, to embolisation into the vasular lumen and/or nulear atypia. In the more severe forms, phlebitis is aompanied by perivenular extension (usually around terminal hepati venules) of inflammation into the lobule produing nerosis of surrounding hepatoytes (fig 2B D) Other helpful findings Histologial features of ACR an also be seen at the hepati hilum, although this portion of the liver is rarely inluded in diagnosti liver allograft biopsies. However, when present, inflammation of hilar nerve twigs and/or large hepati arterial intimal inflammation with endothelial injury are helpful findings, although not typially sored in the Banff shema (fig 3). Evaluation of liver allografts for ellular rejetion involves the reognition of diagnosti histopathologial features as well as exlusion of non-rejetion differentials and o-existing independent pathologial proesses. Sine ACR is an inflammatory proess, the most diffiult and important differential diagnoses are other inflammatory proesses, espeially those due to viral infetions, de novo non-viral and non-infetive hepatitis, and Figure 2 (A) Typial low-power view of aute ellular rejetion (ACR) showing portal but no lobular inflammation (H&E 650). (B) Higher magnifiation of the portal trat within the box in (A) showing bile dut injury (blue arrows), portal phlebitis and mixed infiltration that inludes eosinophils and ativated lymphoytes. Subendothelial inflammation with lifting of portal vein endothelial ells (green arrow) exemplifies endothelial injury (H&E 6200). (C) Hepati vein phlebitis is another feature of aute ellular rejetion, as shown in this illustration; perivenular inflammation is also present (trihrome stain 6200). (D) Inflammatory infiltrates from two portal trats in a ase with severe aute ellular rejetion illustrate the typial mixed nature of ells that inlude ativated/blastoid lymphoytes as well as eosinophils and neutrophils, but largely limited to portal trats with some interfae ativity but otherwise quiesent hepati lobule; note in addition prominent portal phlebitis in the right panel (top left panel and right panel, H&E 6100; bottom left panel, trihrome stain 6100). Figure 3 When the hepati hilum is inluded, inflammation of nerves (top left) and/or hepati artery intima with endothelia injury (blue arrow) an be seen in aute ell-mediated rejetion (H&E 6100). lymphoproliferative diseases. The infetive agents that pose the most problems are reurrent (or less ommonly de novo) viral hepatitis B or C. However other viral infetions (Epstein Barr virus (EBV), ytomegalovirus (CMV) and others) should always be onsidered in this ontext of immunosuppression. Drugindued hepatitis and de novo autoimmune hepatitis are diagnoses of exlusion when the inflammatory proess shows a prominent hepatiti profile, and should be arefully orrelated with linial and serologial markers. The history of temporal assoiation of liver injury markers with exposure to potentially hepatotoxi drugs, pattern, rate and peak of enzyme hanges, and presene or absene of relevant autoantibodies, should be employed in making this orrelation. Table 3 highlights helpful histopathologial (and linial) features to onsider in making 52 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

the distintion between ACR and HCV; this represents the most ommonly enountered diagnosti dilemma. Atypial features in late ellular rejetion ACR ours late in transplant, but not as often as in early graft period. When late ACR ours, the pattern and nature of inflammatory infiltrates sometimes differ from early ACR, in having more hepatiti features (interfae ativity, entral perivenulitis and lobular inflammation), tendeny towards monotypi/less mixed portal infiltrates, and less prominent dut injury. 42 43 Late ACRs are those ourring after 6 months, but these atypial features are more likely to be observed in the muh older grafts. Poor ompliane to mediation, biliary ompliations, and fored redution in immunosuppressive regimen due to infetions (suh as tuberulosis, CMV, reurrent HCV) or PTLD, are some of the doumented reasons for late ACR, although in many patients, no identifiable reasons are 2 37 44 apparent. Of the features assoiated with late ACR, entral perivenulitis (CPV) (fig 4) warrants more disussion. CPV an be mild, moderate or severe, and ours with or without perivenular hepatoellular nerosis. 42 Allograft CPV an be isolated or assoiated with hepatiti lobular inflammation, and/or portalbased features of ACR, inluding dut injury. 45 47 When it ours in assoiation with portal features of ACR, the diagnosis is usually easy to make, and the treatment approah often follows established anti-rejetion protools. Isolated CPV is a more diffiult feature to haraterise as it an present with normal or only minimally elevated liver enzymes. 46 When CPV is present in the biopsy with parenhymal or portal hanges other than those of ACR, the aetiology of CPV in suh a setting beomes the subjet of debate. For example, when there are biopsy Table 3 Histopathologial and linial differenes between aute ellular rejetion reurrent viral hepatitis C Aute ellular rejetion Viral hepatitis Portal inflammation + +/2 Nature of portal inflammation Mixed inludes ativated lymphoytes, eosinophils +/2, neutrophils +/2 Tends to be more monotypi; lymphoytes predominate Interfae ativity * +* Lobular inflammation * + Lobular apoptoses +/ Zone 3 aentuation of inflammation * /+ without phlebitis Zone 3 phlebitis with or without + /+ perivenular dropout Portal vein phlebitis + /+ Bile dut injury + /+ HCV, HBV or other viral serology Not helpful + Rate and timing of liver enzyme hanges Predominant enzyme pattern Reent hange from baseline and may be assoiated with suboptimum immunosuppression ALT/AST, ALP or mixed Usually smouldering, rarely steep exept in FCH ALT/AST *Lobular inflammation, interfae ativity or zone 3 aentuation without phlebitis an be seen in late aute ell-mediated rejetion (also known as atypial aute ellmediated rejetion). 2, Absent; +/2, usually present but an be absent; 2/+, usually absent but an be present oasionally; +, should be present. ALP, alkaline phosphatase; ALT, transaminase; EBV, Epstein Barr virus; FCH, fibrosing holestati hepatitis; HBV, hepatitis B virus; HCV, hepatitis C virus; PTLD, posttransplant lymphoproliferative disease. hanges to suggest reurrent HCV, or when the infiltrate is plasma-ell rih, suh as may be seen in reurrent or de novo autoimmune hepatitis (AIH), the issue beomes whether oexistent CPV in the biopsy is part of eah of these proesses (ie, one diagnosis in the biopsy), or whether CPV ontinues to represent ACR that is oexistent with other disease. This issue is further disussed below. Chroni rejetion The 2000 Banff doument desribes the riteria for grading, reognising and soring hroni rejetion (CR) in liver allografts. 48 CR ours in less than 3% of liver transplant patients at 5 years, and the inidene does not appear to inrease with inreasing years post-transplant; this supports the view that fators determining its ourrene are determined early in 43 49 50 transplant. These fators inlude repeated ACR, CMV infetion, high donor age, long old ishaemi period, and 49 51 52 inadequate/suboptimum immunosuppression. CR is an immunologial injury direted at the vasular endothelium of the hepati artery and peribiliary plexus, as well as the bile dut epithelium. The resulting dut injury is haraterised by epithelial senesene and ultimately loss of small bile duts (dutopenia), with hanges in the intima of hepati artery that inlude thikening and aumulation of foamy marophages. Most CR is diagnosed several months post-transplant, with a mean of 25.1 months in one large series. 53 However, aelerated hroni rejetion an also our within few weeks posttransplant, although suh events are fortunately rare, and usually seen in highly sensitised or suboptimally immunosuppressed patients. Clinial history ould inlude prior (multiple) or ongoing ellular rejetion episodes, problems attaining satisfatory serum levels of immunosuppression, and rising alkaline phosphatase; bilirubin elevation is usually late. Needle biopsy should first be evaluated for adequay in all ases of allograft biopsies, but more importantly in evaluating suspeted CR, where diagnosis hangs on review of an adequate number of portal trats (a minimum of seven fully sampled portal trats is required). Histologial features of CR are listed in table 4, and inlude senesene affeting the majority of interlobular bile duts with less than 50% dutopenia in early CR, or dut loss in more than 50% in late ACR. Senesene or atrophy of bile duts is haraterised by epithelial disruption with irregular spaing, ytoplasmi eosinophilia, high nuleoytoplasmi ratio, nulear hyperhromasia and luminal narrowing. There is none to minimally expanded portal trats, and dutular proliferation or opper retention is not seen; inflammation is minimal (typially lymphoplasmayti when present) (fig 5). Other helpful findings inlude zone 3 perivenular fibrosis, holestasis that is typially hepatoellular with or without aentuation in zone 3, sinusoidal foam ells, and foam ell arteriopathy (if hilar strutures are inluded). 49 Although dutular reation is typially absent in CR, this an be seen, as in the response reovery phase following inrease in or swith of immunosuppressive agent. 49 Early dut senesene and/or loss an be subtle on routine stains and thus should always be searhed for in patients with unexplained elevation of holestati liver enzymes. The use of ytokeratin 7 immunostain (or other biliary epithelial markers) for additional evaluation should be employed (fig 5C). CR is an evolving proess and several biopsies may be needed before a definite histologial diagnosis is feasible. Reurrent primary slerosing holangitis (PSC) is diffiult to distinguish from CR. This subjet is disussed below in the J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 53

Figure 4 Late-ourring ellular rejetion tends to be more hepatiti than typial aute ell-mediated rejetion in the early post-transplant period; the inflammation involves portal, entral perivenular (CV) and mid-lobular areas (upper panel, H&E 6200). As shown in the lower panels, the infiltrates in portal (left lower panel, H&E 6400) and zone 3 areas (right lower panel, H&E 6400) inlude mild dut injury (blue arrow) and rih perivenular plasma ell population (H&E prominent). Figure 5 (A) Chroni rejetion is haraterised by none to minimal lymphoplasmayti inflammation and holestasis in the more severe ases (left panel, H&E 6100; right panel, H&E 6200). (B) Higher magnifiation shows a portal trat with senesent (atrophi) bile duts (arrows) in portal trats with sparse inflammation. Note the absene of signifiant proliferation. Senesene is haraterised by irregular bile dut epithelial size with non-uniform spaing, ytoplasmi eosinophilia and high nuleoytoplasmi ratio. Note the absene of signifiant dutular reation. Senesene in a majority of sampled bile duts or dut loss in more than 50% is required for the diagnosis of hroni rejetion (H&E 6200). (C) Portal trats are seen with absent duts (inset: onfirmed by negative ytokeratin 7 stain 650), minimal inflammation and early periportal fibrosis (trihrome stain 650). (D) Other features of hroni rejetion are highlighted. Left panel, perivenular fibrosis around the entral vein (CV) probably refleting healed foi of severe hepati phlebitis from prior aute ell-mediated rejetion (trihrome stain 6100). Right panels, hepati artery from the hilum of an explanted allograft showing intimal thikening with aggregation of foamy histioytes (inset), and luminal narrowing, alled hroni graft arteriopathy (upper panel, H&E 625; lower panel, elasti trihrome stain 625; inset, H&E 6100). 54 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

Table 4 Histopathologial features of early and late phases of hroni rejetion Struture Early hroni rejetion Late hroni rejetion Interlobular bile duts,,60 mm Hepati venules/perivenular zone 3 hepatoytes Senesene/atrophy affeting the majority of small duts, but,50% dut loss Mild perivenular fibrosis, intimal or luminal inflammation setion on reurrent PSC, and helpful distinguishing features are listed in table 5. Dutopenia in the liver allograft alls to mind CR, but other auses of the so alled vanishing dut syndrome should always be onsidered, explored, and reasonably exluded, espeially those emanating from drugs suh as antibiotis like SeptraH 54 and ACE-inhibitors, to whih many transplant patients are invariably exposed (see later setions). Antibody-mediated rejetion There is no doubt that antibody-mediated rejetion (AMR) ours in liver allografts, either rapidly in the immediate perioperative period (hyperaute), usually but not only due to ABO inompatibility, or later but in the first few days following transplant. 55 58 The inidene of these events is diffiult to determine as they are both unommon and under-reognised. Compared with other solid organs, the liver is relatively proteted from hyperaute AMR beause the Kupffer ells of the liver are able to mop up deleterious antibodies, in addition to the helpful presene of dual irulation providing immense vasular reserve. Also many other immediate post-transplant ompliations, inluding sepsis and hepati artery thrombosis, superimposed on prolonged preservation ishaemia, usually provide alternative explanations for primary non-funtion in liver allografts. However, most agree that the protetion of the liver from antibody-mediated injury is not absolute, and that Dut loss in >50% of portal trats and senesene of remaining duts Marked perivenular fibrosis; hepati vein remodelling, inflammation Hepati artery branhes in interlobular portal trats Missing interlobular artery in,25% of portal trats Missing interlobular artery in.25% of portal trats Hilar (large) hepati artery Foam ell arteriopathy with no signifiant narrowing Foam ell arteriopathy with luminal narrowing/fibrointimal proliferation Others Transition hepatitis with spotty/lyti hepatoellular nerosis Cholestasis, sinusoidal foam ell aumulation Table modified from that published by Demetris et al (2000). 48 Table 5 Comparison of the features of reurrent primary slerosing holangitis and hroni rejetion Reurrent primary slerosing holangitis Chroni rejetion Clinial features Cholangiogram Liver biopsy: Portal hanges Interfae hanges Original disease PSC, years after transplantation, seletive rise of ALP/ GGT Mural irregularity, divertiulum-like outpouhings, beading and pruning of bile duts Uneven, portal expansion by mixed infiltrate, peridutal lamellar oedema, and periholangitis; foal biliary epithelial degenerative hanges Oedema, holate stasis, dutular reation, opper deposits in periportal hepatoytes Typially within 1st year post-transplant, inadequate immunosuppression, unresolved ACR, or after multiple episodes of ACR Pruning of peripheral bile duts No signifiant expansion, biliary epithelial degenerative hanges in most portal trats, dut loss Usually not signifiant CPV Usually not signifiant Often present Cholestasis Variable Usually present Fibrosis Slow progression in time, biliary type Perivenular with or without bridging septa, if present. ACR, aute ell-mediated rejetion; ALP, alkaline phosphatase; CPV, entral perivenulitis; GGT, glutamyl transpeptidase; PSC, primary slerosing holangitis. AMR does our. When it does our, it is usually within the first week post-transplant. 57 The more aggressive (hyperaute) variant ours immediately post-perfusion, usually but not only in the ontext of ABO isoagglutinins. The linial features inlude hypotension, oagulopathy, progressive hyperbilirubinaemia, renal failure and refratory thromboytopenia (ie, rapidly evolving and worsening hepati failure). Imaging studies an reveal portal vein thrombosis and parenhymal nerosis. Serologial demonstration of donor reipient inompatibility at the ABO, MHC or other levels is present, as well as relevant donor-speifi antibodies. The pathogenesis stems from transplantation into a sensitised host with preformed antidonor antibodies. The titre, lass and speifiity of these preformed antibodies determine the degree of injury, and the higher titres are more likely to result in 59 60 more severe injury. Beause of the assoiated oagulation problems, as well as the rapidity of progression, these ases hardly ome to the pathologist as diagnosti needle biopsies, but are more likely to be seen as explanted failed graft or at autopsy. Pathologially the liver is enlarged up to twie the preengraftment weight, 57 and ongested/mottled on the apsular and ut surfaes. Pathy foi of haemorrhage and nerosis an be grossly reognised. Mirosopially the liver shows haemorrhagi and oagulative areas of nerosis and, in the less affeted areas, features mimiking preservation/reperfusion injuries (zone 3 hepatoellular loosening and holestasis) an be identified. Ishaemi nerosis of bile duts of all alibres is also present, as well as loss of small bile duts, despite the short duration post-grafting. Findings in the vessels (small, intermediate and sometimes large) inlude thromboses and vasulitis, with evidene of neutrophili exudation and fibrin deposits in and around vasular walls. Tissue demonstration of antibody ativity by the presene of C4d, C1q or immunoglobulins (almost always IgG, IgM also) in vasular and sinusoidal walls an be seen, but is hardly neessary. Less aggressive development of AMR as a ause of graft dysfuntion an present in the setting of slow normalisation of liver enzymes post-transplant and/or inreasing liver enzymes in the first week of transplant. The diagnosis of AMR in this ontext should inlude a ombination of linial, serologial (demonstration of donor reipient mismath and presene of donor-speifi antibody). The histopathologial features of AMR are not as well defined as in ACR. AMR shows features that overlap with preservation/reperfusion injury, but are ommoner and more pronouned in patients with preformed irulating donor antibodies. 60 These features inlude zone-3- aentuated holestasis, portal expansion with oedema and dutular proliferation (exlude obstrution/strituring), as well as C4d deposition in the walls of portal apillaries and veins and hepati venule; sinusoidal C4d an be seen, but its speifiity is 57 60 yet to be determined. If perihilar tissue is inluded in the J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 55

biopsy, arterial injury and thrombi are present, as evidened by endothelial hypertrophy and myoyte nerosis, vauolisation and thikening, while nerosis of large bile duts and ongestion 60 61 of peribiliary plexus omplete the piture. DISEASE RECURRENCE Long-term survival after liver transplantation is now expeted due to improvements in surgial tehniques, patient seletion and immunosuppression regimens. Approximately 85 90% of liver transplant patients are alive after 1 year and 75% after 5 years. 1 Disease reurrene represents the major ause of graft dysfuntion and loss after 6 months. 29 The frequeny of disease reurrene after liver transplantation is highly influened by the aetiology of the primary liver disease 29 62 64 ; for example hepatitis B and C almost universally reur. Furthermore, a different set of human leuoyte antigen (HLA) moleules present in the allograft may alter the reognition and response of the body to viral partiles. Other fators influening the prevalene and/ or severity of disease reurrene inlude the immunosuppressive environment, whih learly failitates viral repliation and an underlie aggressive behaviour when the disease reurs. 1 On a more positive note, the advent of effetive prophylati and therapeuti measures in the last 15 20 years has dramatially improved the post-transplantation outome of patients transplanted for hepatitis B virus (HBV) infetion. 65 70 Primary biliary irrhosis (PBC), PSC and AIH may all reur after liver transplantation and diagnosti riteria for reurrene may differ from the ones used for similar native liver disease. 71 75 Reurrene of aloholi liver disease, non-aloholi fatty liver disease and non-aloholi steatohepatitis have also been reported. 1 30 76 78 Liver transplantation for primary hepati malignanies (hepatoellular arinoma or holangioarinoma) is based on stage of disease, and the likelihood of reurrene is dependent on several risk fators, suh as mirosopi vasular invasion, multiple tumours and tumour burden, for hepatoellular arinoma. Cholangioarinoma has a poor prog- 1 70 79 84 nosis after liver transplantation and its reurrene influened largely by disease stage. 79 81 Liver disease is ured by transplantation in ases of a 1 -antitrypsin disease, Wilson disease and ysti fibrosis, and most metaboli liver diseases, do not reur after transplantation. Post-transplant outome for patients with haemohromatosis is yet unertain, but there seems to be little impat up to 5 years after transplantation. 1 85 Reurrent HBV infetion In the past, liver transplantation for HBV-related liver disease was often followed by aggressive reurrene and redued 62 63 65 68 70 86 89 survival. Therefore, in the 1980s, HBV infetion was onsidered an absolute or relative ontraindiation for orthotopi liver transplantation (OLT). Currently, ombination therapy by oral antiviral agents (eg, lamivudine, adefovir) and hepatitis B immunoglobulin in the pretransplant and posttransplant setting ahieves nearly 100% of protetion against 65 67 70 86 87 90 adverse outomes from graft reinfetion. The ourse of post-transplant HBV infetion is well known. 88 Similar to that whih ours in non-transplanted patients with HBV infetion, reinfetion also ours in three phases, with an inubation period (approximately 3 months post-transplant), followed by aute infetion (up to 6 months) and hroni infetion (more than 6 months). 91 HBV infetion soures inlude the patient s irulation and extrahepati repliating sites. 88 Disease ativity and progression tend to be more severe ompared with non-transplanted livers, but may be attenuated by oexistent HCV or hepatitis D virus (HDV) infetion due to dereased viral repliation. 1 80 Nevertheless, fulminant reurrent HBV and HDV have been reported in patients with HDV infetion and ative HBV repliation. 92 The most ommon linial feature of aute hepatitis B in the allograft is mild elevation of liver enzymes. Nausea, vomiting, jaundie and hepati failure signal severe reurrent disease. Within the first month after transplantation, an atypial pattern of reurrent HBV infetion, known as fibrosing holestati hepatitis (FCH), may rapidly lead to graft failure. Patients with hroni HBV infetion may be without symptoms or may omplain of fatigue or other non-speifi gastrointestinal symptoms. 92 Pathogenesis of reurrene As in non-immunosuppressed patients, liver damage in allograft reipients is also seondary to virally direted immunologial injury. Reognition of viral antigens by memory T helper ells leads to expansion and ativation of antigen-speifi TH1-type CD4+ lymphoytes. 1 93 This results in marophage ativation and pro-inflammatory ytokine prodution. Interferon (IFN) and tumour nerosis fator (TNF) a ause damage by reruiting and ativating non-speifi inflammatory ells, upregulating TNF-reeptor expression, exerting a diret ytotoxi effet on the hepatitis B surfae antigen (HBsAg)-expressing hepatoytes, and induing loal mediators of tissue injury suh as nitri oxide. 1 It has been suggested that despite immunosuppression, HLA-lass-I-independent immune mehanisms have a signifiant pathogeni role in liver damage assoiated with HBV reurrene after liver transplantation. 93 In the setting of fibrosing holestati hepatitis, whih usually affets overimmunosuppressed patients and is assoiated with massive viral repliation, liver injury may be attributable to a diret HBV-mediated ytopathi effet. 1 Histologial findings The pathologial features of HBV infetion in liver allografts are similar to those seen in non-allograft livers. However, in liver transplant patients, antiviral therapy an derease viral repliation and therefore alter disease histology. 1 81 Inadequate antiviral treatment or the development of resistant viral strains may result in progression from aute to hroni hepatitis and irrhosis. Resolution of disease an sometimes our after a bout of aute hepatitis. Sattered nulear and ytoplasmi staining for hepatitis B ore antigen (HBAg) an be demonstrated by immunohistohemistry 2 5 weeks after transplantation in biopsies reinfeted by HBV. Subsequently, surfae antigen is expressed, but ground glass ells are not easily found during the aute phase. 92 This generally orresponds to the onset of neroinflammatory ativity, Kupffer ell hypertrophy, lobular disarray and portal 62 69 92 94 inflammation. Clinially, these hanges are assoiated with graft dysfuntion. Individual hepatoytes undergoing eosinophili or ballooning degeneration (spotty nerosis) an be found sporadially in the lobules. Patients with low immunosuppression may develop bridging or submassive nerosis, whih is a refletion of more robust immune response against the virus. Reurrent hroni HBV infetion has a more aggressive ourse with more rapid progression of fibrosis and more severe ativity, and it is most likely to be due to enhaned viral repliation and attenuated host response. 92 Liver biopsies show lymphoplasmayti portal inflammation and fibrosis, interfae hepatitis, dutular reation, lobular disarray, Kupffer ell hypertrophy 56 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

and variable lobular neroinflammatory ativity. There may be ground glass ells or sanded-appearing nulei orresponding to HBV surfae antigen or ore antigen expression, respetively. These an be demonstrated by immunostaining for HBsAg and HBAg. A omplete absene of stainable HBAg should raise the possibility of other auses of graft dysfuntion, inluding oinfetion with HCV or HDV. FCH is an atypial pattern of liver injury assoiated with HBV infetion, and it an also affet renal transplant reipients. 92 Histologially, FCH is haraterised by diffuse hepatoellular swelling, holestasis, prominent dutular reation and perisinusoidal fibrosis, but lak of signifiant inflammatory infiltrate (fig 6). The swollen hepatoytes often show ytoplasmi and nulear immunoreativity for HBAg as indiative of viral repliation. 92 Some other atypial forms of reurrent HBV infetion assoiated with heavy viral load have been desribed as fibrosing ytolyti hepatitis, fibroviral 69 92 hepatitis B and steatoviral hepatitis B. In HDV/HBV oinfetion, hepatitis D antigen an be identified in the nulei of infeted hepatoytes by immunohistohemistry. 92 Reurrent HDV infetion an also be onfirmed by HDV RNA detetion by moleular studies or the presene of anti-hdv IgM. 92 HDV hepatitis assoiated with non-repliative HBV infetion an result in hepatiti lesions similar to fibrosing holestati, fibrosing ytolyti or steatoviral hepatitis, but without HBAg expression. In ontrast, in the presene of ative HBV repliation, ombined HBV/HDV hepatitis in allografts is histologially similar to that in non-allograft livers. 1 81 The differential diagnosis of reurrent HBV infetion inludes HCV and non-hepatotropi viruses, drug-indued liver injury and immune-mediated hepatitis. There may be some overlapping features between reurrent HBV infetion and aute or hroni rejetion. Preferential lobular involvement and serologial data are helpful in distinguishing HBV infetion from rejetion. 94 Figure 6 Fibrosing holestati hepatitis B virus infetion in the liver allograft. (A) Diffuse hepatoellular swelling, without aompanying inflammatory reation (H&E 6100). (B) Lobular disarray with several eosinophili apoptoti bodies (H&E 6200). (C) Perisinusoidal fibrosis (Masson trihrome stain 6200). (D) Staining of hepatitis B surfae antigen (orein 6200). Reurrent HCV infetion HCV infetion is the most ommon indiation for OLT, and disease reurrene in the allograft is among the leading auses of graft loss and the need for re-transplantation. 62 63 88 91 95 116 Viral reurrene is universal and graft injury ours routinely. Reinfetion ours during allograft reperfusion, and pretransplant viral titres are reahed in about 72 h. 95 Histologial reurrene with hepatitis due to HCV ours in up to 90% of individuals by 5 years after transplant. 95 96 116 However, progression of HCV infetion is variable: some patients will present an indolent ourse, whereas others rapidly progress to irrhosis and graft failure. Overall, progression of HCV infetion is aelerated after liver transplantation as ompared with patients 95 101 102 without transplants. The linial presentation of allograft reipients with reurrent HCV hepatitis is similar to that of non-allograft patients with primary infetion. Liver enzymes inrease in parallel with histologial evidene of hepatitis, usually within 3 6 weeks after transplantation. Severe reurrent HCV an ause fibrosing holestati hepatitis. This is usually assoiated with overimmunosuppression and is linially manifested by fatigue, jaundie and a marked inrease of serum bilirubin, alkaline phosphatase (ALP) and glutamyl transpeptidase (GGT). The presene of markedly elevated HCV RNA levels is helpful to establish a orret diagnosis. Cirrhosis will develop in 5 20% of 62 97 98 patients beause of reurrent HCV hepatitis. Pathogenesis of reurrene Several fators related to the virus (ie, genotype 1b, viral genomi heterogeneity), the host, the environment and the 101 102 104 107 donor are impliated in the outome. The immune status is likely to be the more important fator influening disease severity: more intense immunosuppression leads to 95 99 worse outomes. Some other preditors of HCV infetion after liver transplantation inlude donor age at time of transplant, donor steatosis, length of old ishaemi time at J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 57

transplant, host immunogeneti bakground (ie, HLA mathing), and timing of reurrene and early histologial findings. Furthermore, rapid tapering doses of steroids and 95 102 steroid-free immunosuppression, with or without indution antibodies, have been thought to redue the likelihood of severe 95 102 117 118 reurrent HCV infetion. Pretransplant viral eradiation by antiviral therapy prevents disease progression and improves survival, whereas post-transplant treatment before or after histologial reurrene has shown variable outomes. 95 102 107 The presene of oexistent CMV infetion after transplant and a history of aute allograft rejetion are also assoiated with inreased severity of HCV reurrene. 1 Obesity and alohol influenes are likely to be similar to those in nontransplanted patients. 95 As mentioned earlier, the initial biohemial and histologial hepatitis usually ours between 1 and 3 months after transplant. The aute phase is marked by a peak of HCV repliation and indution of hepatoyte apoptosis and proliferation, CD8/NKT ellular infiltrate in the graft, and speifi anti- HCV CD4 response. 119 Persistent HCV infetion in the allograft most often evolves to hroni hepatitis (6 12 months). At this point, the inreased viral load seems to overome the inhibitory effets of immunosuppressive therapy on the immune system with subsequent events haraterised by: (1) enhaned inflammatory response and upregulation of IFN-induible genes, (2) indution of antiviral IFNa-induible genes that are not assoiated with redued viral repliation, and (3) a HCV-driven enhaned proliferation, apoptosis and fibrosis response in the allograft. 103 Infiltrating inflammatory ells often lak a speifi HCV-direted antigen response. 1 Less than 10% of patients may develop severe liver injury (ie, fibrosing holestati hepatitis C). In this partiular senario, a redued immune response with undetetable HCV-speifi CD4 response and stable quasispeies takes plae as a result of overimmunosuppression. 103 These allografts typially show a non-speifi TH2 ytokine response, with high levels of interleukin (IL) 10 and/or IL4. Together, these events are believed to allow rapid HCV repliation resulting in extremely high viral burdens (HCV RNA levels in serum are typially more than 30 million IU/ml), and ytopathi allograft injury. 1 92 Early hepati stellate ativation has been shown to our in patients at greater risk to develop progressive fibrosis and more 90 108 114 aggressive reurrent HCV infetion in the allograft. Histologial findings The pathologial features of HCV in liver allografts are similar to those of primary infetion in non-allograft livers (fig 7A, B). Histologial reurrene may be evident within 3 6 weeks after transplant, or sometimes as early as 10 14 days. Liver biopsies performed during the aute phase of reurrent HCV infetion may show lobular disarray, Kupffer ell hypertrophy, hepatoyte apoptosis, mild sinusoidal lymphoytosis and mild mononulear portal inflammation. Periportal and mid-zonal large droplet steatosis is often seen. 1 Mild bile dut injury may be present in the form of intraepithelial lymphoytes and sattered biliary epithelial reative hanges. Suh mild dut injury needs to be interpreted with aution so as not to overall ACR in this setting. As disease progresses into a hroni phase, usually beginning at 4 12 months after transplant, the portal inflammation inreases, often with lymphoid aggregates, and interfae hepatitis of variable severity, lobular disarray, and mild neroinflammatory ativity. If present, inflammatory bile injury is, most ommonly, mild and foal. Bile dut loss is not a feature of reurrent HCV infetion. Perivenular (zone 3) inflammation an be present but it typially involves a minority of hepati veins. 42 Prominent interfae ativity an our in aggressive onventional reurrent HCV infetion. Fibrosing holestati HCV This is an aggressive variant of reurrent HCV that ours in oasional patients with rapid deterioration. This is haraterised histologially by extensive dense portal fibrosis with immature periellular/sinusoidal fibrous bands, extensive hepatoyte swelling and degeneration, dutular reation, marked analiular and ellular bilirubinostasis, and moderate mononulear inflammation (fig 7C, D). 96 It is very important to reognise FCH in order to ensure proper therapy. It has been shown that a small subset of patients with reurrent HCV infetion present linial and morphologial features that overlap with AIH (ie, post-liver transplant AIHlike hepatitis). 97 In these partiular ases, the allograft biopsies show a prominent portal, periportal and lobular plasma-ell-rih infiltrate, and perivenular (zone 3) nerosis. Some of these patients also have positive autoimmune serology with inreased serum globulins, presene of anti-nulear antibody and antismooth musle antibody. Its reognition is linially important beause of inreased fibrosis progression; however, this may be diffiult to separate from de novo AIH and atypial aute rejetion. Differential diagnosis Distintion between ACR and reurrent HCV infetion is very important beause treatment for ACR with ortiosteroids and OKT3 is assoiated with inreased risk of allograft irrhosis and mortality. On the other hand, if left untreated, ACR may progress to hroni rejetion, espeially in IFN-treated patients. 42 Not only do these two onditions often share similar linial and histologial features, but they may also oexist in the liver allograft. A areful review of the post-transplant linial ourse, inluding liver enzymes results and HCV RNA levels when available, should be onsidered in parallel to interpretation of the biopsy findings. Mononulear portal inflammation and lymphoyti holangitis are ommon features of reurrent HCV infetion and ACR. However, in ACR, inflammatory bile dut injury tends to involve a majority of bile duts. Perivenular (zone 3) inflammation involving a majority 36 91 116 of hepati veins also favours rejetion. Lobular neroinflammatory ativity and interfae hepatitis with dutular reation tend to be more prominent in reurrent HCV infetion (table 3). If ACR and reurrent HCV oexist, the predominant proess should be identified. In doubtful ases where low-grade ACR annot be reliably exluded, patients should be monitored losely with re-biopsy reommended if liver enzymes ontinue 36 120 to rise. Chroni rejetion with or without oexistent reurrent HCV infetion is identified by biliary epithelial degenerative hanges or small dut loss or perivenular (zone 3) inflammation and fibrosis involving a majority of bile duts/ terminal hepati veins. 36 Distinguishing reurrent HCV from other viral hepatitides, de novo AIH, drug-indued hepatitis, PBC and PSC is based primarily on a ombination of linial, biohemial, serologial and histopathologial findings. AIH usually shows more prominent plasma ell inflammation and less steatosis ompared with reurrent HCV infetion. FCH due to HCV needs to be distinguished from large dut obstrution and hepati artery thrombosis (ishaemi holangitis). Bile dut obstrution is identified by portal oedema and 58 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

Figure 7 Reurrent hepatitis C virus infetion in the liver allograft. (A) Portal inflammation with interfae ativity and mild bile dut injury (H&E 6200). (B) Foi of neroinflammation in the lobule (H&E 6100). (C) Fibrosing holestati hepatitis C infetion: marked hepatoellular ballooning with minimal inflammation, and dutular reation (H&E 6100). (D) Portal expansion and perisinusoidal fibrosis (Masson trihrome stain 650). dutular reation with or without aute holangitis. 121 A far more hallenging senario is when reurrent HCV (ie, non-fch) oexists with a biliary problem or another ause of holestasis (eg, adverse drug reation or sepsis). In these situations, the histologial features may be very similar to FCH. A areful review of the linial ourse to rule out an infetious proess, or history of new mediations, and imaging of the biliary tree may be helpful to exlude or onfirm a seond problem. Still, in some instanes, espeially when the atual viral load annot be determined by reliable quantitative methods, FCH annot be exluded with ertainty, and the deision whether or not to treat HCV infetion will depend on weighing the risks and benefits. Anti-HCV therapy may be given as the only feasible intervention and, on oasion, response to anti-hcv therapy or lak thereof provides the answer in hindsight. Reurrent autoimmune hepatitis Autoimmune hepatitis is a relatively unommon indiation of liver transplant. Outomes are good with 1-year and 5-year patient survival rates of about 87% and 80 90%, respetively. Graft survival rates at 1 year and 5 years are 84% and 74 76%, 88 122 126 respetively. The reported reurrene rate for AIH in 71 72 74 82 most studies is in the range of 17 42% at 5 years. In general, reurrent AIH appears at variable time periods after transplantation, and progression seems to be slow. 122 Reurrent AIH responds well to inreases in immunosuppression or addition of ortiosteroids. 72 The pathogenesis of AIH is unknown. Whether or not an autoimmune response will perpetuate seems to be influened by a geneti suseptibility to present self or ross-reating antigens, a sensibility to aetiologial triggers (ie, viruses or toxins), and the omposition 74 123 ytokine environment. Aberrant exposure of HLA-II antigens and enhaned presentation of normal onstituents on hepatoytes with subsequent ativation and proliferation of ytotoxi T lymphoytes may take plae. Hepatoellular damage seems to be seondary to proinflammatory ytokines 71 127 released by sensitised T ells. There seems to be no onsistent risk fators for reurrene, but reurrent AIH has been shown to be more ommon in transplant reipients who were HLA-DR3 positive or HLA-DR4 positive in one study. 74 Furthermore, suboptimal immunosuppression, the presene of type I autoimmune disease, and severe inflammation in the native liver before transplantation, may also be assoiated with a greater inidene of reurrent 122 124 disease. Diagnosti riteria for reurrent AIH are similar to those used in the non-transplanted liver and inlude biohemial, serologial and histologial abnormalities, and steroid dependeny (see box 3). 128 However, these riteria are more diffiult to apply in the allograft liver for a number of reasons inluding biohemial and histologial overlap with ACR, the immunosuppressive environment, and the possibility of alloimmune disease direted against allograft antigens. 124 Beause of the lak of reliable disease markers, a liver biopsy is often the main or sole diagnosti tool for identifying reurrent AIH in the allograft. Histologial findings The histologial hanges attributed to reurrent AIH are not speifi and need to be distinguished from other auses of hroni hepatitis, ACR, hroni rejetion, adverse drug reations and reurrent PBC and PSC. Early hanges inlude lobular hepatitis with hepatoyte rossetting. 42 74 124 The hroni phase is usually marked by portal infiltrate omposed of lymphoytes and plasma ells with prominent interfae ativity. A plasma-ell-rih infiltrate direts attention to the possibility of AIH, but it is not a requirement for diagnosis. Lobular neroinflammatory ativity is variable, and onfluent and bridging nerosis are not unommon. Perivenular (zone 3) inflammation an be present in a majority of hepati venules similar to rejetion. If present, bile 42 124 dut inflammatory damage involves a minority of duts. J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 59

Box 3: Criteria for the diagnosis of reurrent autoimmune hepatitis, primary slerosing holangitis and primary biliary irrhosis Reurrent autoimmune hepatitis Liver transplant for autoimmune hepatitis Autoantibodies in signifiant titre (.1:40) Sustained rise in serum aminotransferase ativity ( more than two times normal) Elevated serum immunoglobulins Diagnosti or ompatible liver histology Cortiosteroid dependeny Exlusion of other auses of graft dysfuntion (eg, HCV infetion, rejetion) Reurrent primary slerosing holangitis Liver transplant for primary slerosing holangitis Multiple nonanastomoti biliary stritures Exlusion of other auses (ie, rejetion, infetion, ishaemia) Diagnosti or ompatible liver histology Reurrent primary biliary irrhosis Liver transplant for primary biliary irrhosis Persistene of antimitohondrial antibodies Elevated immunoglobulins Diagnosti or ompatible liver histology Exlusion of other auses of graft damage Reurrent PBC PBC is onsidered to be a disease of disordered immune regulation haraterised by progressive loss of interlobular and septal bile duts leading to holestasis and irrhosis. Antimitohondrial antibodies are present in 95% of patients. Liver transplantation is indiated for advaned PBC, with exellent overall patient and graft outomes. The 5-year survival rate after deeased donor liver transplantation is approximately 80%. Reurrent PBC after transplantation is ontroversial, but has now beome aepted. Reurrent PBC is seen in 17% of patients at a mean of 36 months, and 30% at 10 years. The reported median time reurrene is between 3.7 and 5 years. 71 75 129 132 The role of ursodeoxyholi aid in the treatment and prevention of reurrent PBC is ontroversial. 73 It is not lear whether donor and reipient age, old and warm ishaemia time, and type of immunosuppression used, may influene disease reurrene. However, it seems to be more ommon after living-related liver transplantation and after ortiosteroid withdrawal. 130 The diagnosti riteria for reurrent PBC are summarised in box 3. A diagnosis of reurrent PBC is made in the setting of harateristi histology and absene of other auses of graft damage. Elevated serum immunoglobulins and persisting antimitohondrial antibodies are not suffiient for the diagnosis of disease reurrene. Histologial hanges may be present in the allograft, even in the absene of biohemial abnormalities. 73 Histologial findings Changes of reurrent PBC are similar to those present in native livers. Liver biopsies with reurrent PBC may show one or more of the following features: variable portal inflammation with mononulear (or mixed) infiltrate, lymphoid aggregates with germinal entres, lymphoyti holangitis with biliary epithelial eosinophilia, and peridutal epithelioid non-nerotising granulomatous reation. The diagnosti lesions (ie, epithelioid 121 132 granulomas and florid dut lesions) are often foal, and therefore may not be present in needle biopsies in the early stages (fig 8). As disease progresses, there is development of lymphoplasmayti interfae ativity resembling AIH, and biliary interfae ativity with holate stasis. Additional features inlude dutular reation, portal and periportal fibrosis, small bile dut loss and periportal oedema (halo sign). The parenhyma may show spotty nerosis or even sattered foi of lyti nerosis, and deposition of opper and opper-assoiated proteins at the portal/parenhymal interfae provides supportive evidene of reurrent PBC if other auses of biliary trat disease have been exluded. Bile dut injury or loss due to reurrent PBC needs to be distinguished from ACR, hroni rejetion, adverse drug reation, CMV and HCV infetion, reurrent PSC, ishaemi holangitis, reurrent or de novo AIH, and graft versus host disease. Usually, the linial senario, serologial investigations and imaging results are very important to make the diagnosis lear. A diagnosis of reurrent PBC an be definitive when granulomatous bile dut destrution and/or florid bile dut lesions are present in the proper linial ontext. In the absene Figure 8 Reurrent primary biliary irrhosis. (A) Predominantly portalbased hanges with marked inflammatory infiltrate and interfae ativity (H&E 650). (B) Portal trat with a granulomatous bile dut lesion (H&E 6100). 60 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

of these features, the presene of a prominent but foal lymphoyti holangitis, aompanied by portal-based lymphoid aggregates with germinal entres and bile dutular reation, are highly suggestive, although not diagnosti, of reurrent PBC. 121 Sometimes the time frame of rise of ALP is a lue: a sudden rise in ALP is unlikely to be due to reurrent PBC. Reurrent PSC PSC is a progressive holestati disease of unknown aetiology that usually involves both the extrahepati biliary tree and the intrahepati biliary tree, and has a lose assoiation with inflammatory bowel disease. The hallmark linial lesion of PSC is an abnormal holangiogram. Endosopi retrograde holangiopanreatography and magneti resonane holangiopanreatography typially show irregular stritures, beading, divertiular outpouhing, and pruning of bile duts. Liver transplantation is indiated for patients with end-stage disease. The long-term outome after transplantation is very good, with survival rates of 86% at 5 years, and 70% at 10 years. 62 A higher inidene of aute and hroni and steroid-resistant rejetion in PSC patients has been reported, espeially in the presene of oexistent inflammatory bowel disease. 49 62 Reurrene of PSC 62 63 70 71 74 133 139 after transplantation ranges from 9% to 47%. Risk fators for disease reurrene inlude donor reipient gender mismath, male gender, and intat olon at the time of transplantation. 49 140 The presene of hilar holangioarinoma before transplantation signifiantly dereases survival after transplantation. 135 As a matter of fat, the presene of holangioarinoma is onsidered an absolute ontraindiation to transplantation at most entres due to the high risk of reurrene in the graft. 136 There is yet no effetive treatment to delay the presentation or progression of reurrent PSC in the allograft. 136 Seletive elevation of ALP and GGT due to PSC reurrene usually manifests 1 year after transplantation. It is very diffiult to separate reurrent PSC from other auses of biliary stritures (eg, holedohojejunal anastomoti striture, hepati artery thrombosis, preservation injury, hroni dutopeni rejetion, ABO blood group inompatibility, viral/baterial biliary trat infetion, SFSS in living donors, non-heartbeating donors). The diagnosti riteria for reurrent PSC 49 136 are summarised in box 3. Non-anastomoti intrahepati stritures that develop within 90 days after transplantation are not onsidered reurrent disease. The diagnosis of reurrent PSC requires holangiographi and histologial evaluation. Histologial findings The histologial features of reurrent PSC are idential to those seen in the native livers with PSC. Early hanges in the peripheral liver inlude mild non-speifi aute and hroni 42 72 periholangitis and mild dutular reation. As disease progresses, there is peridutal lamellar oedema with inreased dutules and mixed portal inflammation with eosinophils and neutrophils, periportal oedema, dutular reation and sattered small dut loss. 121 Later stages are featured by biliary irrhosis, holestasis, marked opper deposition, and Mallory bodies in paraseptal hepatoytes. Peridutal onentri fibrosis and dut loss involve small and medium-sized bile duts (fig 9A E). These so-alled fibro-obliterative dut lesions an also be seen in patients with ishaemi holangitis (hepati artery thrombosis) and other post-transplant auses of seondary slerosing holangitis. The large intrahepati and extrahepati bile duts may show uleration, biliary sludge and marked peridutal lymphoplasmayti infiltrate. 49 The distintion between reurrent PSC and hroni rejetion may be hallenging, as both ause a holestati pattern of liver enzyme elevation and dut loss (fig 9F). The linial history, evaluation of serial biopsies and histopathologial findings are useful to separate these two onditions (table 5). 121 Reurrent aloholi liver disease, non-aloholi fatty liver disease and non-aloholi steatohepatitis Aloholi liver disease represents a leading ause indiation for liver transplantation with short-term survival rates omparable to those for patients who undergo liver transplantation for other onditions. 76 78 The rate of alohol relapse is onsidered low, and resumption of alohol seems to begin within the first year after transplantation. 141 142 There is no signifiant evidene supporting a detrimental effet on graft or patient survival assoiated with reidivism. Fatty liver and steatohepatitis are the main histologial features of alohol relapse. 142 More severe reidivism an lead to frank aloholi hepatitis with Mallory s hyaline, foamy degeneration of hepatoytes and perivenular fibrosis. 121 Aurate data on the perentage of liver transplants performed for non-aloholi steatohepatitis (NASH)-related irrhosis are not available, in part beause many ases identified as ryptogeni irrhosis may in fat represent burnt out NASH. 143 Steatosis has been reported to our within 6 12 months and irrhosis within 2 years of transplantation in 143 144 patients undergoing liver transplantation for NASH. Reurrent NASH seems to our at later times than fatty liver alone, with inreasing inidene over time during follow-up. Reurrent metaboli diseases In disorders suh as type 1 tyrosinaemia, a 1 -antitrypsin defiieny, Wilson disease, neonatal haemohromatosis, and glyogen storage disease types 1, 3 and 4, the liver is replaed by a genetially normal one that is not suseptible to reurrent disease. 85 The risk of reurrene is higher in patients with metaboli defets involving extrahepati sites, and the effets on the liver are largely seondary in those that are at highest risk of reurrene (eg, Niemann Pik disease, Gauher s disease, ystinosis and erythropoieti protoporphyria). 85 LATER NEW-ONSET DISEASES/INJURIES IN THE IN LIVER ALLOGRAFT Biliary ompliations At the time of transplantation, reonstrution of the biliary trat ours in the form of a dut-to-dut anastomosis or holedohojejunal anastomosis. Muosal and/mural damage may our in the proess and lead to biliary trat ompliations, suh as bile leaks, and anastomoti or intrahepati stritures. The proess of biliary wound healing ours and 145 146 may or may not be ineffetual. The general onepts of wound healing as they apply to the biliary trat, interleukin 6/gp130 signalling pathways, and ineffetive wound healing, are disussed in detail elsewhere. 12 This an affet the small extrahepati biliary tree and/or the large extrahepati biliary tree. Briefly, in the extrahepati large bile duts, biliary healing may lead to sarring and striture formation. In the small extrahepati bile duts, impaired proliferation of the bile dut epithelium or exuberant responses an ontribute to liver injury. 12 Radiologial tests suh as MRI and/or allograft biopsies may be performed in the ourse of investigation of biliary ompliation post-transplant. Biliary ompliations or their J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 61

Figure 9 Reurrent primary slerosing holangitis in a failed liver allograft. (A) Portal trat with mild hanges, inluding peridutal oedema and biliary epithelial degenerative hanges (H&E 6100). (B) Charateristi peridutal lamellar oedema (Masson trihrome stain 6100). (C) Fibro-obliterative dut lesions ( fibrous knots ) (Masson trihrome stain 6100). (D) Marked portal expansion dutular reation and fibrosis (Masson trihrome stain 6100). (E) Periportal hepatoyte opper aumulation (rhodamine stain 6200). (F) Features of early hroni rejetion are present in the same liver allograft: marked biliary epithelial senesent hanges (H&E 6100). This patient also had multiple episodes of aute ellular rejetion during the early and late post-transplant period. sequela may also be seen in biopsies performed for other reasons (eg, protool surveillane biopsies for HCV in a patient with biliary striture). Biliary ompliations our early and late in the post-transplant ourse. All are disussed here for onveniene. Biliary sludge syndrome Cold ishaemi-preservation injury depletes energy stores in mirovasular endothelial ells and bile dut epithelium. As a result, metalloproteinases are ativated. Biliary epithelium and endothelium are detahed from underlying matrix. In the mirovasulature, detahment of endothelium predisposes to thrombosis after reperfusion. 8 10 Reperfusion brings leuoytes that beome ativated by tissue damage, and they release effetor moleules, ausing more tissue damage and further promote thrombogenesis. 8 10 Several fators, inluding inreased sensitivity of bile dut ells to reperfusion injury, poor funtional reovery after ATP depletion, invasion of polymorphonulear leuoytes into bile duts, and hydrophobi bile salts, appear to ontribute to preservation-related injury of bile duts. 8 10 147 148 Damaged biliary epithelial ells are sloughed into the bile. Exposure of the underlying stroma to bile appears to serve as a nidus for rystallisation of biliary sludge. 12 Injury of bile duts is assoiated with hyperbilirubinaemia and underlies the longlasting phase of reperfusion graft injury. 8 10 Morphologial hanges produed by bile sludge syndrome are present in the extrahepati large bile duts and intrahepati large bile duts, as well as in the small intrahepati duts. Sine biopsies usually sample the peripheral liver, it is the latter that are enountered in biopsies. There is prominent dutular reation onsisting of biliary ells and peridutal myofibroblasts. This ours beause of inreased pressure in the biliary tree distal to the point of luminal obliteration. The proliferating dutules and myofibroblasts form a wedge of tissue that arises from the portal trat and distorts the liver arhiteture. Morphologial hanges of the large bile duts are usually seen at the time of re-transplantation in the exised failed graft. There is biliary sludge, muosal ulers and inflamed granulation tissue and myofibroblast proliferation in the wall of extrahepati bile duts and large intrahepati bile duts. 11 Exposure of the underlying stroma serves as a stimulus for inflammation and ativation of myofibroblasts. This leads to wound ontration and fibrosis, and eventually to stritures in large-alibre duts. Complete fibrous obliteration of the bile dut lumens by onentri rings of fibrous tissue ours. 62 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

Bile dut stritures To appreiate the pathophysiology of bile dut striture formation, the blood supply to bile duts is summarised as follows. Blood is supplied to intrahepati bile duts and extrahepati bile duts exlusively through hepati arteries (HAs). 149 Over 50% of the blood onveyed by HAs is primarily destined to the bile duts. 150 Intrahepati arteries ourse in lose proximity to the bile duts. They drain into the peribiliary plexus, whih is a rih mirovasular network surrounding bile duts. 150 Blood supplying the bile duts drains into the sinusoids via the portal system. Ishaemia-indued bile dut lesions have been olletively labelled as ishaemi holangitis. 151 The biliary epithelium is dependent on arterial blood flow, unlike the hepati parenhyma with its dual arterial and portal venous blood supply. As suh, the biliary epithelium is suseptible to injury when arterial blood flow is ompromised. Ishaemi holangitis manifests as segmental stritures with resultant mehanial impairment of bile flow and, oasionally, seondary infetion of the biliary system. Biliary stritures in liver transplant reipients are related to a ombination of large HA olusion related to surgial reonstrution or sepsis, and/or damage to small-sized arteries and peribiliary plexus due to preservation, reperfusion, rejetion, ABO inompatibility or CMV infetion. 152 Biliary stritures may be anastomoti or nonanastomoti. Anastomoti biliary stritures are thought to result from tehnial surgial problems or loal ishaemia; the inidene has been reported to be,10% for deeased donor liver transplantation, and to be higher, up to 30%, for living donor liver transplantation. 153 159 Most anastomoti stritures appear within the first several months after transplantation, but they may also develop, less frequently, many years after transplantation. Intrahepati biliary stritures are inluded in nonanastomoti stritures; they result primarily from hepati artery thrombosis. 152 Intrahepati biliary stritures, without hepati artery olusion, are related to hroni dutopeni rejetion, ABO inompatibility, ishaemia reperfusion injury, or reurrene of primary disease, suh as PSC or AIH. 146 147 159 162 Nonanastomoti stritures tend to our later after transplantation, are less amenable to treatment, are generally progressive, and adversely impat graft and patient survival. The biopsy features of dut obstrution in allografts are the same as those enountered in native livers. Most ompliations show predominantly neutrophili portal inflammation, peridutal oedema, and intraepithelial and intraluminal neutrophils within true portal bile duts. Mild dutular proliferation, entrilobular hepatoanaliular holestasis and small lusters of neutrophils throughout the lobules are also ommonly seen. 163 Chroni biliary trat stritures may be assoiated with hroni portal inflammation and biliary epithelial ell senesene. The differential diagnosis of early dut obstrution (within 6 months post OLT) is preservation injury with biliary sludge syndrome and ACR; the differentiating histologial features are ompared in table 1. After 6 months post transplantation, biliary obstrution/ striture mimis a broader spetrum of proesses that inlude aute and hroni rejetion, viral hepatitis and reurrent autoimmune disorders. Chroni intermittent biliary obstrution or holangitis, as in patients with the biliary sludge syndrome, an be assoiated with a mixed, or a predominantly mononulear, portal infiltrate and biliary epithelial ell senesene hanges. Portal fibrosis with mild dut proliferation, mild portal neutrophili or eosinophili inflammation, and mild entrilobular holestasis are features that suggest obstrutive holangiopathy 1 INFECTIOUS COMPLICATIONS Apart from reurrene original viral infetions, usually viral hepatitis B and C (desribed above), other viral infetions an affet liver graft, inluding EBV and CMV. Also newly aquired hepatitis B or hepatitis C infetion should be borne in mind in patients transplanted for other reasons other than viral hepatitis. The sope of infetions in liver allograft patients is broad, and also inludes several non-viral agents that either onstitute new or re-ativated (mostly opportunisti) infetions. The histologial features in most ases are not different from those seen in the other organ systems, and therefore they are not disussed further in this review. However, the syndromes assoiated with EBV and CMV warrant further disussion in that they are not only well studied in liver patients, but they also represent important differential diagnoses to other more ommonly enountered histopathologial problems, and are therefore disussed in the following setions. Post-transplant EBV infetion Post-transplant liver biopsy in the setting of EBV disease may be enountered in the ontext of linial hepatitis, or in the setting of suspeted or known PTLDs, for example: (a) post-transplant EBV infetion inluding hepatitis, (b) PTLD in the liver, () spetrum of allograft onditions assoiated with extrahepati PTLD. The inidene of EBV infetion post-transplant varies depending on whether the patient was exposed to EBV prior to transplantation. EBV infetion post-transplant ours in 63 80% of patients who are seronegative at the time of transplantation. Reativation infetions our in 20 22% of 164 165 patients exposed to the virus before transplantation. It appears that many patients are asymptomati. In a paediatri series, only 15% of patients with primary post-transplant EBV hepatitis were symptomati. 165 A range of histologial patterns an be seen in the allograft biopsies in EBV hepatitis. 166 168 These range from the harateristi piture desribed in the immunoompetent host in infetious mononuleosis to a distintive onstellation of features. 166 At one end of the spetrum is the typial EBV hepatitis pattern that usually shows mild portal and sinusoidal mononulear infiltrates (fig 10). These infiltrates are omposed of small and mildly atypial lymphoytes. The lymphoytes arrange themselves in a single file pattern within the sinusoids, and this pattern should suggest an EBV-related disorder. In situ hybridisation for EBV-enoded RNA (EBER) is onfirmatory. Other lobular hanges inlude foal hepatoellular swelling, aidophili nerosis of hepatoytes and mild lobular disarray. Granulomas may our. 169 Another pattern is that of nonspeifi reative hepatitis. The liver sinusoids show mild lymphoytosis. Portal mononulear inflammatory infiltrates may be of variable severity. Portal trats with sparse infiltrates oexist with dense portal lymphoyti infiltrates elsewhere in the same biopsy. At the other end of the spetrum is a harateristi onstellation of features desribed by Randhawa et al that onsists of mixed mononulear portal and sinusoidal infiltrates ontaining atypial large non-leaved mononulear ells and immunoblasts, assoiated hepatiti lobular ativity, and relatively mild dut damage disproportionate to the severity of the infiltrate. 166 Any of these patterns should arouse suspiion for EBV infetion, and this should be relayed to the treating physiian so that serologial onfirmation an be obtained. The patterns at the ends of the histologial spetrum desribed above are suffiiently harateristi to raise suspiion for EBV. The pattern of non-speifi reative hepatitis arries a wider J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 63

Figure 10 Epstein Barr virus hepatitis. Low-power view of parenhyma showing sinusoidal lymphoid infiltrate in a single file pattern (H&E 6100). Insert: high-power view of portal trat showing moderately dense portal infiltrate inluding small and mildly atypial lymphoytes (H&E 6400). differential diagnosis that should inlude EBV, espeially if serology for usual hepatotropi viruses (HBV, HCV) is negative. The differential diagnosis of a mononulear infiltrate in the post-transplant setting is ellular rejetion. In favour of EBV is the monomorphous omposition of the portal infiltrate in EBV hepatitis, the lak of eosinophils and neutrophils and presene of atypial large leaved lymphoid ells, and assoiated hepatiti hanges. While EBV-related bile dut damage and phlebitis may our, 170 the tempo of these is more brisk when the portal infiltrate is mixed in the setting of ellular rejetion. It should be remembered that ellular rejetion and EBV hepatitis may Figure 11 (a) Low-power view of needle biopsy showing no obvious parenhymal inflammation (H&E 650). (b) Higher magnifiation of the box in (a) showing a typial ytomegalovirus (CMV) intranulear inlusion in an enlarged hepatoyte liver surrounded by few neutrophils. This was the only inlusion in the ore biopsy and was seen on one level of one slide, emphasising the need for metiulous srutiny of the biopsy in suh ases (H&E 6630). () CMV immunostain showing immunoreativity (brown olour) in the intranulear CMV inlusions from another ase, onfirming their nature (6400). oexist. This is espeially more prone to our in biopsies taken as follow-up after treatment for EBV hepatitis that inludes redution of immunosuppression When there is a pre-existing diagnosis of PTLD and there is linial ause to perform a liver biopsy, the most important onsideration is whether aute rejetion is present. Possible findings in suh biopsies inlude aute rejetion, EBV hepatitis and reurrent disease. In one series, aute rejetion was found in 54% of biopsies, and features indeterminate for rejetion in 29%. EBV hepatitis was the primary diagnosis in 8% of biopsies, and seondary in an additional 25%. Non-speifi reative hepatitis was seen in approximately 16% of biopsies, and 20% showed reurrene of primary liver disease. 171 Immunoperoxidase staining showed primarily T ells. EBER was deteted in 58.3% of biopsies, with and without aute rejetion. 171 In the rejetion ases, the presene of oasional EBV-infeted ells was thought to reflet an inreased irulating viral burden in these patients. 171 Positive in situ hybridisation for EBER in liver biopsies is only seen in the ontext of a high viral load in patients with PTLD or those at high risk for developing this ompliation. 168 Post-transplant CMV infetion In liver transplant reipients, the overall inidene of CMV disease has been desribed as being up to 29%. 172 In transplant patients, the greatest risk of CMV hepatitis is transplantation of an organ from a CMV-serology-positive donor to a serologynegative patient (D+/R2). In liver transplant reipients, in the risk group of D+/R2 the risk may be as high as 44 65% if no prophylaxis is given. 173 174 CMV hepatitis is a signifiant ompliation of CMV infetion after liver transplantation, with 172 175 176 an inidene of 2 17%. CMV infetion of the liver transplant is haraterised by graft dysfuntion. Thus the biopsy indiation may be one of elevated serum transaminases. The diagnosis, however, should be based 64 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

on liver biopsy. 177 178 The histologial features of CMV hepatitis are variable. In high-risk patients (ie, heavily immunosuppressed or naive reipients), CMV ytopathi effets may our in any ell type of the liver. The harateristi CMV inlusions are large eosinophili and intranulear, surrounded by a lear halo (fig 11). Oasional small basophili or amphophili ytoplasmi inlusions may be present. 179 CMV hepatitis is haraterised by spotty lobular nerosis, and mild lobular disarray. There is mononulear, or mixed, portal inflammation, foal bile dut damage and aggregates of marophages (mirogranulomas) sattered throughout the parenhyma. Hepatoytes ontaining CMV inlusions may be assoiated with small lusters of neutrophils (miroabsesses). Dilemmas arise when, as is ommonly the ase, none of the harateristi features are apparent, or when features of CMV overlap or oexist with those of ellular rejetion. Absene of harateristi positive signs The histologial alterations assoiated with intrahepati CMV infetion may ause differential diagnosti problems, as the harateristi CMV inlusions are rarely seen. Neutrophili absesses are not speifi or sensitive, and again are not always present in the biopsy. Lamps et al found miroabsesses in only 17% of post OLT biopsies done for graft dysfuntion; of the 17% who had miroabsesses, only 19% were proven to have CMV. 180 Miroabsesses were seen in other onditions, inluding infetious onditions (baterial, viral and fungal), graft ishaemia and biliary obstrution, and they were of undetermined aetiology in 29% of ases. Also, although CMV hepatitis tends to exhibit milder lobular disarray and hepatoyte swelling as ompared with HCV or HBV hepatitis, CMV hepatitis an sometimes be diffiult to distinguish from early reurrent HCV or HBV. CMV hepatitis may also posses overlapping features with EBV hepatitis: mild lymphoplasmayti portal and lobular inflammation. If present, miroabsesses and mirogranulomas are helpful, as they are not generally assoiated with HBV or HCV infetion, but immunoperoxidase staining for viral antigens (CMV, HSV and HBV) or in situ hybridisation for EBV should be performed, as either is usually diagnosti in these diffiult ases. Overlapping features with, or assoiated rejetion Other features of CMV hepatitis (eg, portal lymphoyti infiltration, some degree of endotheliitis and holestasis) may lead to the misinterpretation of mild aute rejetion. Follow-up biopsies after CMV treatment may be diffiult to interpret as to the presene of residual CMV versus evolving aute or hroni rejetion possibly related to lowering of immunosuppression in the fae of the initial diagnosis of CMV hepatitis. In pratie, the demonstration of CMV inlusions or its antigens by immunostaining takes preedene, resulting in redution in immunosuppression and initiation of ganilovir therapy. Correlation of time lines of enzyme improvement or worsening with immunosuppression lowering and/or initiation of ganilovir treatment an be helpful in resolving the differential diagnosis. There is reported assoiation between CMV and hroni 181 182 rejetion. It appears that persistene of CMV DNA, and not aute CMV hepatitis, is assoiated with hroni rejetion. Suessful antiviral treatment of CMV infetion does not exlude the persistene of the virus and the risk of hroni rejetion. DE NOVO HEPATITIS Idiopathi hroni hepatitis Unexplained hroni hepatitis (idiopathi hroni hepatitis (ICH)) in the adult liver allograft reipient is not unommon, although its reported inidene is quite variable between different entres. 183 The inidene ranges from,10% and up 43 144 184 189 to approximately 50% in different series. ICH is haraterised by a predominantly portal mononulear inflammatory infiltrate assoiated with interfae hepatitis. Lobular inflammation is of variable degrees, and hepatoyte nerosis or apoptosis is frequently present. By definition, the immunemediated pattern of inflammation designated as ICH also requires that bile dut injury or vasular lesions harateristi of aute or hroni rejetion are minimal or absent. 190 Thus, the features are largely similar to those seen in hroni hepatitis in the non-transplant setting. 191 There are several potential reasons for the variable inidene of ICH among transplant entres. These inlude fators that ould affet long-term outome or late biopsy findings. 183 To name some of these fators: the use of extended-riteria organs with influene from donor fators and the early postoperative ourse, severity of early aute rejetion, low-grade rejetion going unreognised in entres where protool biopsies are not routinely performed, and variation in terminology and histologial threshold for diagnosis of ICH ( portal and lobular mononulear inflammation at one entre may be diagnosed as ICH, whereas a similar infiltrate may be termed as nonspeifi inflammation in another entre). The differential diagnosis inludes a number of immunemediated entities that inlude ellular rejetion, infetions and de novo post OLT AIH. A hroni hepatiti form of rejetion is probably the leading onsideration. 190 Centrilobular-based aute rejetion is another possibility in those ases of ICH where entral perivenulitis involving a majority of hepati veins is present. 42 Reurrent or newly aquired HCV or HBV should always be onsidered. These are disussed in detail in the setion on reurrent disease. Other potential viral hepatitides inlude EBV (see setion on post-transplant infetions) and hepatitis E virus (HEV). Chroni HEV infetion an develop in immunosuppressed patients, who may then serve as long-term arriers of the virus. It has been hypothesised that HEV may be a ause of hroni hepatitis in liver transplant reipients. Haagsma et al diagnosed hroni HEV infetion in two liver transplant reipients. 192 Fourteen ases of aute HEV infetion developed in patients reeiving solid organ transplants: three liver transplants, nine kidney transplants, and two kidney and panreas transplants. All patients were positive for serum HEV RNA. Histologial features of hroni hepatitis were present. 193 Unidentified hepatotropi viruses have been suggested as a possible ause of hroni hepatitis, but HEV 183 194 may be one of these. New-onset AIH diagnosis is based on a ombination of the histopathologial pattern of injury, autoimmune serology, and 190 194 195 inreased serum gammaglobulins. This was mentioned above as a onsideration in the diagnosis of late-ourring ellular rejetion, and is further disussed below as a linial entity. The histologial features of ICH on routine stains are nonspeifi and thus, often, the aetiology annot be determined on histologial grounds alone. Correlation of histologial findings with linial events and time lines of enzyme elevation, if any, is required (eg, reent redution in immunosuppression, reent illness interfering with absorption of mediations, introdution J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 65

of new drugs that may ause drug-indued liver injury or affet levels of immunosuppressants). Autoimmune and viral (inluding hepatotropi and non-hepatotropi viruses, eg, EBV) serology and immunoglobulin levels, partiularly IgG, are an integral part of evaluation to omplement histology. A areful searh for ground glass hepatoytes to suggest infetion with HBV is needed, and immunoperoxidase stains for HBsAg an further onfirm the nature of suh ells. This an be helpful in situations where a risk fator for HBV infetion is not readily known. Immunostains for other viruses, suh as herpes simplex and CMV, may be performed. In situ hybridisation for EBV performed on the liver biopsy should also be onsidered. In most series of long-term biopsy follow-up studies it appears that ICH has a benign rather than a signifiantly 43 144 184 189 adverse linial outome, although there are a few studies that show evidene to the ontrary. Among the latter, in their retrospetive study, the Birmingham group reported mild inflammatory ativity and mild to moderate fibrosis in the initial biopsies of 12 of 30 reipients with ICH. Approximately 41% of these 12 reipients subsequently developed marked graft dysfuntion, new or progressive fibrosis was noted in approximately 50%, and three patients developed irrhosis. 194 Seyam et al found histologially proven graft irrhosis in 48 (representing 3.7%) of patients in this series. 196 Ten patients at a median of 7 years after transplantation had no apparent aetiology for irrhosis; in all 10 patients, previous biopsies had shown features of hroni hepatitis of unertain aetiology. The authors onluded that hroni hepatitis was the most frequent underlying pathologial proess in ases where the ause of irrhosis remains unertain, as they found in 20% of patients with graft irrhosis. The prevalene of ryptogeni posttransplant irrhosis was signifiantly higher in patients initially transplanted for fulminant seronegative hepatitis (6%) than in those transplanted for other diseases (0.3%). 196 Therefore, until there is a better understanding of the true impat of ICH, the histologial pattern of hroni hepatitis should be reported and Figure 12 (a) De novo post-transplant autoimmune hepatitis: low-power view of needle biopsy showing entral perivenulitis with zone 3 hepatoyte drop out and portal inflammation. There is milder inflammation of the parenhyma outside zone 3. The edge of the zone 3 areas with dropped-out hepatoytes is marked by arrows (H&E 650). HV, hepati vein; PT, portal trat. (b) De novo post-transplant autoimmune hepatitis: higher magnifiation of zone 3 showing onentri inflammation and parenhymal nerosis around the hepati vein (ie, entral perivenulitis) (H&E 6400). () De novo post-transplant autoimmune hepatitis: higher magnifiation of the portal trat, showing plasma-ell-rih infiltrate (H&E 6400). should trigger pathologial and histologial investigation as muh as possible to exlude an underlying ause. De novo AIH In native liver, AIH does not have a pathognomoni feature, and its laboratory, serologial and histologial manifestations are found in aute and hroni liver disease of diverse auses. Thus, the diagnosis of autoimmune hepatitis requires onfident exlusion of other ausative fators. In 1993, the International Autoimmune Hepatitis Group 128 197 199 proposed diagnosti riteria, whih were revised in 1999. Due to the omplexity of these riteria and their insuffiient validation, the International Autoimmune Hepatitis Group devised a simplified soring system for wider appliability in routine linial pratie. 198 Histologially, mononulear infiltrates that are plasma-ell predominant and interfae ativity are reasonably good markers of AIH in native livers. 199 200 Thus, de novo AIH in the liver allograft is diagnosed by a plasma-ellrih infiltrate showing signifiant neroinflammatory interfae and perivenular ativity (fig 12). However, in the posttransplant setting, a new diagnosis of AIH is ompliated by the need to distinguish this entity from reurrent HCV and rejetion. HCV infetion appears to indue a geneti suseptibility to autoimmune proesses, inluding in the liver. 201 203 When differentiating HCV from de novo AIH post-transplant, the diffiulty stems from the fat that HCV infetion by itself in the non-transplant setting an be assoiated with multiple immunemediated extrahepati manifestations, and hroni HCV liver disease an be assoiated with AIH-like features in native 199 204 liver. Czaja et al reported higher serum levels of gammaglobulin and immunoglobulin G, higher frequeny of irrhosis, a higher mean Knodell sore, a higher frequeny of HLA-DR3, and a high titre of smooth musle antibodies assoiated with the AIH-like pattern of HCV-indued liver injury in the general population. 199 66 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254

Figure 13 Nodular regenerative hyperplasia: retiulin stain showing the nodular parenhymal arhiteture. There are nodular hyperplasti areas of parenhyma (*) entred on portal trats and not bound by fibrosis alternating with atrophi areas (650). There is evidene in the literature to suggest that patients transplanted for HCV liver disease, like non-transplant patients with HCV, an develop AIH-like features in the graft assoiated with reurrent HCV or, after suessful HCV therapy, with HCV RNA learane. Khettry et al reported AIH-like HCV based on histologial hepatitis with moderate to severe portal, periportal and lobular neroinflammation with prominent plasma ells in 10% (nine patients) of HCV reipients in their programme. There was a signifiantly higher inidene of CPV. 97 Serologial evidene of autoimmunity was present in six of nine (66%) of the AIH-like HCV patients, supporting the hypothesis that prominent Figure 14 (a) First biopsy from patient treated with previous episodes of ellular rejetion and azathioprine therapy: hepati vein with lumen obstruted by fibrous tissue (*) from markedly thikened intima (Masson trihrome stain 6400). (b) Higher magnifiation of the biopsy in (a) showing severe sinusoidal ongestion in zone 3 around the hepati vein (Masson trihrome stain 6400). () Seond biopsy from the same patient 4 years later: low-power view showing organisation of ollapsed parenhyma by fibrosis that is severe (Masson trihrome stain 625). (d) Seond biopsy, as in (), showing healed hepati veins. There are subtle remnants of strutural ollagen from the distorted wall of a small hepati vein (arrows) (Masson trihrome stain 6400). plasma ells are a marker of autoimmunity in liver allograft biopsies with hroni hepatitis. 97 Berardi et al reported a series of 9 of 44 HCV liver allograft reipients who developed what they termed de novo autoimmune hepatitis after at least 6 months of anti-hcv therapy for HCV reurrene. 205 Although HCV RNA learane was ahieved in all but one ase, these patients developed signifiant graft dysfuntion and hepatitis. Extensive investigations exluded other auses inluding infetious ones, and led to a diagnosis of de novo AIH aording to liniopathologial riteria defined by the International Autoimmune Hepatitis soring system. 128 Prednisone treatment and essation of antiviral therapy resulted in varying outomes: five remissions, and four graft failures with two deaths. That the liver allograft damage in this series was related to autoimmunity was reasonably onvining given HCV learane. In a study of highly seleted patients, Fiel et al use the phrase plasma ell hepatitis (de novo autoimmune hepatitis) to desribe what they eventually onlude to be a variant of aute rejetion. 206 The authors onluded that plasma ell hepatitis (PCH) represents a form of aute rejetion beause PCH in these patients frequently developed in assoiation with suboptimal immunosuppression; their ohort had a high inidene of aute rejetion prior to developing PCH (whih was felt to indiate a tendeny toward rejetion), and had a better outome when treated with inreased immunosuppression. The onundrum: CPV in reurrent HCV and de novo AIH rejetion or what is it? CPV (zone 3 inflammation surrounding the hepati vein) with or without assoiated zone 3 nerosis is usually immunologially mediated, and mostly represents rejetion in the ontext of liver transplantation. 207 211 Besides AIH, other auses of perivenular nerosis and inflammation are muh less likely. 212 In a series of 100 patients in a study published by the University of Pittsburgh Medial Center, 46 40 patients had CPV, J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254 67

Figure 15 (a) Low-power view from a post-transplant liver biopsy of a patient with reurrent hepatitis C virus and reent elevated transaminases (H&E 625). Note mild portal inflammation aentuating towards one end of the ore (arrowhead). Insert: high-power view of the portal trat with aentuated inflammation. Note that the inflammation here is mixed and there is bile dut injury by lymphoytes (arrow); these are features that mimi ellular rejetion (H&E 6400). (b) Repeat biopsy after the absess was searhed for by imaging: low-power view of organising absess with portal trat at the edge (PT) (H&E 650). () Higher-power view of portal trat at periphery of absess showing bile dut injury as a bystander next to an inflammatory lesion biopsy from the same patient 4 years later: low-power view showing organisation of ollapsed parenhyma by fibrosis that is severe (H&E 6200). (d) Higher-power view of absess showing inflammatory ells inluding neutrophils (H&E 6200). of whih 12 were lassified as ACR beause of assoiated portalbased findings. Follow-up revealed that five of the remaining 28 patients subsequently developed portal-based ACR on separate biopsies, two patients developed de novo AIH, and two developed hroni dutopeni rejetion. In general, and until new insight emerges, CPV aompanied by portal-based features of ACR with or without lobular hepatitis is best regarded and treated as late ACR. When CPV is isolated, serologial evidene for autoimmunity should be sought and, if present, (for now) is best regarded as AIH in an allograft, and lassified as de novo or reurrent as the ase may be. The remaining ases probably represent a variety of onditions not easily lumped together, and as these patients are followed-up a better understanding of the ontributing disease proess(es) should beome lear, but should nevertheless be reported as idiopathi hepatitis, disussed with the treating liniians and, if possible, followed up losely with early re-biopsy, the goal being to detet fibrosis progression and/or development of dutopenia. It remains unertain whether plasma-ell-rih entrilobular lesions (fig 12b) in HCV liver allograft reipients represent rejetion, as onluded by Fiel at al, 206 or an altered immune variant of HCV, or de novo AIH. The arguments for and against have been reently disussed extensively by Demetris et al, 203 but their onlusion was that perivenular inflammation and nerosis involving a majority of entral veins, plasma-ell-rih or not, are immune-mediated injuries (ellular rejetion and AIH), and responsive to inreased immunosuppression, regardless of the HCV status of the patient, and irrespetive of whether or not the patient was post-transplant. In reurrent HCV, prominent lobular inflammation with zone 3 nerosis in HCV transplant patients should prompt the onsideration as to whether these hanges represent a severe or aggressive form of HCV reurrene, or HCV with superimposed ellular rejetion with CPV, or de novo AIH. It would appear that lobular hanges in de novo post-transplant AIH are more prominent than in the native liver AIH. 213 214 The features that favour post-transplant AIH are nevertheless similar to those onsidered in the native liver and inlude: portal inflammation with numerous plasma ells, prominent interfae hepatitis, and lobular inflammation (plasma ell rih) with zone 3 nerosis. Perivenular inflammation, and nerosis involving a majority of entral veins, are not typial features of reurrent HCV in allografts; when they are found, an alternative explanation that inludes an aompanying immune-mediated injury/ellular rejetion should always be onsidered and further investigated linially. The deision to treat rejetion or autoimmune phenomena in HCV-positive patients requires lose liniopathologial orrelation and an informed disussion with the liniian. In our experiene at a large transplant entre, and similar to the opinion of Demetris et al, 203 this deision depends on the severity of the perivenular damage and whether the short-term gain of less liver damage offsets a diminished ability to eventually lear the HCV infetion. If left untreated, signifiant (moderate or severe) perivenular damage, whether plasma-ell-rih or not, promotes the ourrene of bridging fibrosis (portal-to-entral 42 97 205 206 and entral-to-entral). DRUG-INDUCED LIVER INJURY Drug-indued liver injury an mimi many patterns of transplant-related and non-transplant-related liver pathology. A detailed desription of drug-indued liver injury is beyond the sope of this review and is given elsewhere. 215 Brief omments will however be made on some of the drugs ommonly used in the post-transplant setting: immunosuppressants and sulfamethoxazole trimethoprim (Septra). Cilosporin (CyA) and tarolimus (KF506) hepatotoxiity were reported in liver allograft reipients in the early 1990s. However, newer insights indiate that many of the features thought to represent CyA and/or FK506 toxiities, suh as perivenular nerosis/fibrosis, bile dut epithelial hanges and 48 216 218 sinusoidal foam ells, are in fat due to hroni rejetion. In pratial terms, these alineurin inhibitors are more likely to 68 J Clin Pathol 2010;63:47 74. doi:10.1136/jp.2009.068254