Central perivenulitis: A common and potentially important finding in late posttransplant liver biopsies

Authors

  • Stefan G. Hübscher

    Corresponding author
    1. Department of Pathology, University of Birmingham and University Hospital Birmingham National Health Service Foundation Trust, Birmingham, United Kingdom
    • Department of Pathology, University of Birmingham, Birmingham B15 2TT, United Kingdom
    Search for more papers by this author
    • Telephone: 00-44-(0)121-414-4005; FAX: 00-44-(0)121-414-4019


  • See Article on Page 625

The term central perivenulitis (CP) emerged from discussions held by the Banff Working Party at the 8th Banff Conference on Allograft Pathology in Edmonton, Canada, in July 2005.1 It is used to describe a spectrum of inflammatory lesions involving perivenular regions of the liver parenchyma that are in most cases thought to be a manifestation of liver allograft rejection. Other related terms that have been used previously include centrilobular necrosis,2–7central venulitis,8, 9hepatitic phase of rejection,10centrilobular inflammation,11centrilobular necroinflammation,12centrilobular changes,13centrilobular alterations,14 and centrilobular injury.15

At one end of the histological spectrum of CP, usually occurring within the first few weeks of liver transplantation, perivenular inflammation is typically seen in association with characteristic portal tract changes of acute rejection and often also involves the endothelium of hepatic veins. In such cases, a diagnosis of rejection is straightforward, and rejection is usually graded as severe according to the Banff criteria.7, 16 More frequently, features of CP are seen in patients presenting with graft dysfunction several months or more following transplantation. In such cases, hepatic venous endothelial inflammation is rarely conspicuous. Portal tract inflammation is also variable in severity and composition and often lacks features typically seen in early acute cellular rejection (ACR). In cases where portal inflammation is absent or lacks typical features of ACR, the term isolated central perivenulitis (ICP) may be used, as was done in the study by Krasinskas et al. that appears in this issue of Liver Transplantation.17 Although there is increasing evidence to suggest that late CP is usually related to rejection,1, 3, 4, 6, 8, 9, 13, 15 other causes of centrilobular necroinflammation should also be considered in the differential diagnosis (discussed later). Furthermore, even in those cases where late acute rejection is the most likely diagnosis, grading is difficult according to the standard Banff criteria, which are largely based on portal tract changes, and alternative approaches are thus required.1

Early studies of liver allograft rejection were focused mainly on inflammatory changes occurring in portal tracts but also recognized the presence of inflammation involving hepatic venous endothelium and the surrounding liver parenchyma.18–22 The susceptibility of perivenular regions to immune-mediated injury probably is related to these areas containing potent donor antigen-presenting cells, which can trigger an alloimmune response.23 Subsequently, several studies have attempted to assess the etiology, clinicopathological features, and natural history of centrilobular changes in the liver allograft.2–15, 24–30 Most studies have investigated centrilobular changes occurring as a component of rejection or as a consequence of other recognized graft complications, but a few have identified centrilobular lesions presenting as an isolated phenomenon.8, 9, 15 Direct comparisons between previously published studies are difficult to make because of differences in patient demographics and selection criteria, times at which biopsies were obtained, and the range of centrilobular lesions studied. Nevertheless, a number of general themes have emerged. First, cases with features of CP tend to present later than those with purely portal-based rejection changes. For example, in a study of pediatric liver allograft recipients, the mean times to the development of acute portal rejection alone, acute portal rejection and central venulitis, and isolated central venulitis were 13, 34, and 65 days, respectively.9 Second, CP is often associated with a rise in aminotransferase levels,4, 7, 9, 12 contrasting with the predominantly cholestatic biochemical profile that is seen in association with bile duct inflammation in acute rejection or with bile duct loss in chronic rejection. Third, the presence of CP indicates a more severe form of acute rejection that is less likely to respond to conventional immunosuppression and more likely to progress to chronic rejection. Adverse outcome may be partly related to the presence of more severe concomitant portal tract changes than are seen in cases without centrilobular lesions14 but also appears to apply to cases where centrilobular changes are present as an isolated finding9, 15 Fourth, in studies of serial biopsies, CP typically precedes bile duct loss and thus appears to represent a transitional phase between acute (reversible) rejection and chronic (irreversible) rejection.12, 27, 31 Recognition of this fact and prompt treatment with appropriate immunosuppressive therapy are likely to be important reasons for the declining prevalence of graft failure from chronic rejection in recent years.

The mechanisms involved in the pathogenesis of centrilobular injury in liver allograft rejection are complex and can result in a range of overlapping histological appearances. In cases where there is a combination of perivenular inflammation and hepatocyte dropout (CP), it is likely that inflammatory cell–mediated hepatocyte apoptosis is an important mechanism.32, 33 Although inflammatory lesions typically have a perivenular distribution, diffuse spotty lobular inflammation has also been noted during the transitional phase from acute to chronic rejection.10, 29 Ischemia due to occlusive lesions in large-sized and medium-sized arteries has been postulated as a cause for persistent centrilobular necrosis, often occurring without conspicuous inflammation, in cases of chronic rejection.26, 34 However, other studies have failed to demonstrate a clear association between the presence/severity of centrilobular necrosis and occlusive arteriopathy in chronic rejection.2, 12 Rejection-related endothelial inflammation in hepatic venules may be important in the pathogenesis of congestive changes resembling those seen in hepatic veno-occlusive disease.29, 35, 36 Perivenular congestion is also commonly present in cases of rejection-related centrilobular necroinflammation without obvious occlusive lesions in hepatic veins.9, 12, 15 Antibody-mediated damage to sinusoidal endothelial cells has also been suggested in the pathogenesis of hepatocyte loss in liver allograft rejection.33, 37 Other centrilobular changes that can be seen as part of the spectrum of rejection-related CP are cholestasis, hepatocyte ballooning, and fibrosis. Fibrosis is typically seen in a perivenular distribution but in some cases appears to result in thickening of the hepatic vein wall itself.27 Fibrous obliteration of the hepatic vein lumen can also be seen as part of the spectrum of rejection-related veno-occlusive disease35, 36 or in the absence of prominent congestive changes.38 In some cases, there is progression to bridging fibrosis or cirrhosis.38, 39

There are several other graft insults that can result in perivenular hepatocyte injury.1, 40–42 A list of possible causes is summarized in Table 1. Ischemic causes of centrilobular necrosis tend to present in the early posttransplant period and are rarely associated with prominent inflammation. Distinction from rejection-related CP is thus rarely a problem. Inflammation is also not a prominent finding in the centrilobular injury occurring with azathioprine toxicity, which instead is typically associated with sinusoidal dilatation and congestion and features of nodular regenerative hyperplasia.43, 44 Centrilobular necrosis has also been attributed to tacrolimus toxicity,45 but this observation has not been verified by other studies.40 Viral hepatitis behaves more aggressively in the setting of immunosuppression and may be associated with severe necroinflammatory changes including foci of confluent or bridging necrosis, features not seen with hepatitis C virus (HCV) infection and very rarely seen with hepatitis B virus infection in the immunocompetent individual. Clinicopathological correlation, including the identification of recent changes in immunosuppression and viral nucleic acid levels, will usually help to identify cases where viral infection is a likely cause for centrilobular necroinflammatory lesions. Some cases of recurrent HCV infection have atypical features resembling those seen in autoimmune hepatitis (AIH)—these have a higher prevalence of CP lesions than cases with typical recurrent HCV.46 The most important conditions to consider in the differential diagnosis of CP are recurrent and de novo AIH.1, 47, 48 In most cases of AIH (recurrent or de novo), centrilobular necroinflammatory lesions are associated with typical portal tract changes including a plasma-cell rich inflammatory infiltrate associated with interface hepatitis. However, some cases may present with isolated or predominant features of CP before subsequently developing typical portal tract changes.49–52 Areas of overlap exist between de novo AIH and rejection,53 and it has been suggested that so called de novo AIH could represent a form of late cellular rejection. Finally, features of CP can be seen as part of the spectrum of “idiopathic” posttransplant chronic hepatitis (IPCH).1, 54, 55 Most cases occur in association with a portal tract inflammatory infiltrate that lacks typical features of acute or chronic rejection, but in some instances centrilobular necroinflammation presents as an isolated finding. Many patients with IPCH also have autoantibodies but lack the biochemical changes that are required to make a diagnosis of de novo AIH.54 By a process of exclusion, it has been suggested that many of these cases are best regarded as a form of late cellular rejection. There is increasing evidence from studies carried out in the Birmingham Liver Unit to suggest that IPCH is associated with the development of progressive fibrosis, which can lead to bridging fibrosis or cirrhosis in up to 50% of patients at 10 years,54, 56, 57 but important clinical problems have not been reported to the same extent in other studies.55

Table 1. Possible Causes of Perivenular Necrosis in the Liver Allograft
Ischemia• Preservation/reperfusion injury
 • Vascular occlusion (hepatic artery, portal vein, hepatic vein)
Rejection• Acute
 • Chronic
Viral hepatitis (recurrent or acquired)• Hepatitis B
 • Hepatitis C
Autoimmune hepatitis (recurrent or acquired) 
Drugs• Azathioprine
 • Tacrolimus (FK 506)
Other• “Idiopathic” chronic hepatitis

The study by Krasinskas et al.17 analyzes histological findings in a consecutive series of 100 adult patients undergoing orthotopic liver transplantation (OLT). Importantly, the study population excluded patients transplanted for HCV infection and AIH, both of which have the potential to recur with features of CP. How do the findings compare with other published studies?

PREVALENCE OF ICP

The overall prevalence of ICP (defined as CP occurring in the absence of concomitant or recent portal-based features of ACR) was 28%. This compares with a frequency of 9%-31% in 5 previous studies.2–5, 9 Reasons for these apparent differences are not clear, but they could simply reflect different study protocols, as discussed earlier.

Abbreviations

ACR, acute cellular rejection; AIH, autoimmune hepatitis; CP, central perivenulitis; HCV, hepatitis C virus; ICP, isolated central perivenulitis; IPCH, “idiopathic” posttransplant chronic hepatitis; OLT, orthotopic liver transplantation.

TIME OF PRESENTATION

In line with previous studies, ICP occurred later than portal-based acute rejection. The great majority of cases presented >3 months post-transplant, and the mean time to the first biopsy was 422 days. Furthermore, in biopsies obtained >12 months post-OLT, ICP represented the commonest pattern of graft inflammation, accounting for 19/21 cases (90%) where inflammation was present.

BIOCHEMICAL CHANGES

Cases of late CP (>3 months post-OLT) were typically associated with relatively mild elevations in liver function tests. This is in line with other studies describing potentially important histological abnormalities in patients who appear to be clinically well with normal or near-normal liver biochemistry54, 57 and supports the case for using protocol biopsies in the assessment of late graft function.

ETIOLOGY OF CP

With the exclusion of recurrent viral hepatitis and AIH as potential causes of CP, it seems reasonable to assume that many if not all cases of ICP are a manifestation of late acute rejection. In support of this observation, 28/40 (70%) patients with CP also had portal-based ACR (12 concomitantly, 16 as prior episodes) versus only 12/60 (20%) patients without CP. Three out of 40 (7.5%) patients with CP subsequently developed features compatible with de novo AIH, again emphasizing potential areas of overlap between rejection and de novo AIH.

NATURAL HISTORY AND OUTCOME

A mean of 5.7 biopsies per patient were obtained, with follow-up periods up to 7 years. A diagnosis of ICP was strongly associated with the development of “late inflammation,” which was not seen in cases with purely portal-based ACR. In 94% of cases, the pattern of late inflammation could also be described as a form of CP, suggesting differences in the targets of immune-mediated damage in early and late posttransplant studies. “Adverse outcomes” (defined as zone 3 fibrosis, ductopenia, and de novo AIH) were commonly seen in patients with CP, presenting either as an isolated lesion (10/28 cases, 36%) or in association with portal-based ACR (3/12, 25%), and they were not seen in any of the 12 patients with pure portal ACR or in the 48 patients without ACR. Ten patients developed zone 3 fibrosis (6 of which progressed to bridging fibrosis), and 3 had duct loss; both are recognized features of chronic rejection. However, only 1 of the 13 patients with these adverse findings progressed to graft failure from chronic rejection during the study period. Although the histologic diagnosis of ICP was not an indication for therapy according to the original study design, 9 of the 13 patients with adverse outcomes were treated with increased or altered immunosuppression upon diagnosis of the adverse finding, and the natural history of disease progression may thereby have been modified. A favorable response to immunosuppressive therapy, which has also been reported in other studies,8, 15, 30 further supports the concept of CP being an immune-mediated lesion.

GRADING OF CP

The centrilobular necroinflammatory changes occurring in CP constitute a broad spectrum in terms of severity. Mild cases are characterized by mild, often patchy perivenular inflammation with little or no hepatocyte loss. More severe cases are associated with diffuse inflammatory lesions accompanied by areas of confluent or bridging necrosis.1 Previous studies have made little attempt to grade the severity of zone 3 changes occurring in the absence of typical portal ACR. The present study is the first to use the grading scheme recently proposed by the Banff Working Group.1 Encouragingly, cases classified as severe CP according to the Banff criteria had a significantly greater frequency of adverse outcomes than those with mild CP.

In conclusion, the study by Krasinskas and colleagues17 demonstrates that ICP is a common and potentially important finding in late posttransplant biopsies. Most cases are probably related to rejection and are associated with a worse outcome than cases of purely portal-based rejection. In an era in which noninvasive methods for investigating liver dysfunction are becoming increasingly popular, the findings in the present study emphasize the clinically important information that can be obtained from protocol biopsies in the liver transplant patient. Future studies should be directed toward defining robust criteria for determining the etiology of ICP (rejection-related versus other causes), validating the observations made in the present study regarding the potential value of grading the severity of CP and establishing treatment protocols that specifically target the distinctive immunological events that are likely to be occurring in the microenvironment of the small hepatic vein branches and surrounding liver parenchyma.

Ancillary