Posttransplant plasma cell hepatitis (de novo autoimmune hepatitis) is a variant of rejection and may lead to a negative outcome in patients with hepatitis C virus

Authors


  • See Editorial on Page 750

  • Presented in part in poster form at the American Association for the Study of Liver Diseases Annual Meeting, Boston, MA, October 2006 (Hepatology 2006;44:414A).

Abstract

De novo autoimmune hepatitis has been described in both pediatric and adult liver transplantation (LT) recipients. Studies of small numbers of patients have proposed it to be an alloimmune hepatitis or form of chronic rejection. We have recently noted an increasing number of patients with post-LT recurrent hepatitis C virus (HCV) developing this, with an apparent negative impact on outcome and survival. We term this entity posttransplant plasma cell hepatitis (PCH). A search of our institution's pathology database was performed with the terms “plasma cell(s),” “lymphoplasmacytic infiltrate,” and “liver allograft.” A histological scoring system was devised to more reliably diagnose PCH in the setting of recurrent HCV. Thirty-eight patients were identified, and their clinical data were analyzed. Sixty percent had a negative outcome as defined by the development of cirrhosis, need for retransplantation, or death. Eighty-two percent had recent lowering of immunosuppression or subtherapeutic calcineurin inhibitor levels; 58% developed PCH within 2 years post-LT. Histologic resolution of PCH was associated with good outcome (P < 0.001). Patients not receiving treatment had a negative outcome (P = 0.007) as did patients receiving corticosteroids as therapy (P = 0.02). Persistence (P = 0.007) or recurrence of PCH was associated with negative outcome. In conclusion, PCH is a histologic variant of rejection. On liver biopsy, PCH can at times be difficult to diagnose, and the use of a standardized scoring system is recommended to differentiate it from other forms of allograft dysfunction. Treatment by optimization of immunosuppression without the use of corticosteroids appears effective. The development of PCH in the setting of recurrent HCV is a negative prognostic factor for patient outcome and allograft failure. Liver Transpl 14:861–871, 2008. © 2008 AASLD.

Recurrence of autoimmune hepatitis (AIH) after liver transplantation (LT) occurs frequently, and its natural history has been well defined.1–3 Idiopathic hepatitis with histological features of AIH has been described in both adult and pediatric patients undergoing LT for indications other than AIH.4–17 Typically, patients have a lymphoplasmacytic cellular infiltrate on liver biopsy.5, 18, 19 It is unknown whether this histological entity is a variant of acute cellular rejection (ACR)2, 4, 18–20 or chronic rejection7, 21 or in fact represents an alloimmune hepatitis or a true de novo AIH.9, 11, 13–17 Several investigators have reported a severe histological progression and the development of graft failure in patients with de novo AIH, sometimes occurring despite increased immunosuppressive therapy.2, 7, 10, 11, 16, 21, 22 Recently, Berardi et al.23 reported the occurrence of de novo AIH in a small number of patients undergoing interferon (IFN) treatment for recurrent hepatitis C virus (HCV) post-LT.

Despite increasing reports of small numbers of patients with de novo AIH post-LT,9, 10, 12–16, 22 the natural history of de novo AIH is not well described in the nonpediatric LT setting because of the heterogeneity of patient populations and liver histological features. There is even less known about the occurrence of de novo AIH in patients with HCV.24, 25 This is particularly important in HCV post-LT because of the potential risks of severe virological recurrence with increased immunosuppression.25–29 At our center, we have noted an increasing number of patients with post-LT recurrence of HCV developing de novo AIH with an apparent negative impact on outcome and survival. Herein we describe in detail a large series of patients with recurrent HCV post-LT developing plasma cell hepatitis (PCH) histologically similar to previous reports of de novo AIH with special emphasis on their clinical and histological features and response to treatment. A uniform histological scoring system was devised in an effort to more reliably diagnose post-LT PCH specifically in the setting of recurrent HCV and to discern whether PCH could be a histological variant of ACR.

Abbreviations

ACR, acute cellular rejection; AIH, autoimmune hepatitis; AMA, anti-mitochondrial antibody; ANA, anti-nuclear antibody; ASMA, anti-smooth muscle antibody; AST, aspartate aminotransferase; CNI, calcineurin inhibitor; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; IFN, interferon; LKM, liver-kidney-microsomal antibody; LT, liver transplantation; PCH, plasma cell hepatitis.

PATIENTS AND METHODS

A free text search of the Mount Sinai Medical Center Department of Pathology database was performed, encompassing a 13-year period (1994-2006), with the key words “liver allograft,” “lymphoplasmacytic,” and “plasma cell(s).” We believed that this would identify any patient with evidence of recurrent or de novo AIH or PCH as the presence of plasma cells typifies these diagnoses. It is conceivable that some cases of PCH may not have been identified if these combined terms were not part of the histological diagnosis and microscopic description, but at our center the liver biopsy diagnosis format includes a detailed microscopic description. We cross-referenced these data with the names of all patients undergoing LT for HCV at the Mount Sinai Medical Center from 1994 to 2006, which made up the study group. The diagnosis of HCV was confirmed serologically and with the presence of a detectable HCV viral load and upon review of the explanted native liver. This study was approved by the Mount Sinai Institutional Review Board.

Slides from the explanted native livers and all subsequent biopsies were blindly reviewed by a hepatopathologist (M.I.F.). Histological assessment was performed with at least one hematoxylin-eosin slide and a trichrome stain for the stage of fibrosis. At least 6 portal tracts were present in every biopsy reviewed. A histological scoring system adapted from both the METAVIR system30 and Knodell histological activity index31 was devised (see the appendix).

The grade of necroinflammation in chronic hepatitis is based on the degree of portal inflammation, interface hepatitis, and lobular necrosis. Portal inflammation was classified as mild to severe on the basis of the degree of lymphoplasmacytic infiltrates, interface hepatitis was based on the extent of the erosion of the limiting plate by inflammation, and the degree of lobular activity was based on the presence of parenchymal necrosis (see the appendix). The severity of plasma cell infiltrate was based on the approximate ratio of plasma cells to the rest of the inflammatory infiltrate as follows: score of 0 = absent, score of 1 = occasional plasma cells, score of 2 = 10%-30% plasma cells, and score of 3 = >30% plasma cell infiltrate often occurring in sheets (Fig. 1). In addition, the presence of centrilobular necrosis and dropout accompanied by plasma cell infiltrate was assessed as present or absent (Fig. 2). Upon initial review, an overall impression (initial assessment) of the biopsy diagnosis was given as follows: other diagnosis (neither HCV nor AIH) = 0, histologic features consistent with recurrent HCV without AIH features = 1, a histological picture of HCV with AIH features = 2, and biopsy typical of AIH = 3. Histological features of AIH are centrilobular necrosis with or without bridging necrosis and a prominent plasma cell infiltrate. Plasma cells are present in all portal tracts although not to the same degree. Fibrosis was scored with the Ishak system,32 and the degree of steatosis was adopted from the Nonalcoholic Steatohepatitis Clinical Research Network.33 ACR and chronic ductopenic rejection were classified according to the respective Banff criteria.34, 35 Other concomitant histological features such as biliary tract abnormalities and granulomas were also recorded. The first biopsy with a plasma cell score of 3, not showing chronic rejection or the typical mixed cellular infiltrate of ACR, was deemed the index biopsy. We used these criteria because large numbers of plasma cells are the most characteristic feature of AIH and are not typically seen in HCV recurrence post-LT to this degree.36–38 Portal and lobular inflammation and interface hepatitis may be seen in any form of chronic hepatitis and are thus less discriminating features.31

Figure 1.

Representative photographs depicting the different plasma cell scores for plasma cell hepatitis. (A) Plasma cell score = 1. In the center is a portal tract that shows a mild mononuclear infiltrate. Occasional plasma cells (arrowheads, inset) are identified. (B) Plasma cell score = 2. Plasma cells (arrowheads) constitute approximately 20% of the entire inflammatory infiltrate. (C) Plasma cell score = 3. This area shows a dense mononuclear inflammatory infiltrate involving an area of parenchymal collapse. The predominant component of the infiltrate is plasma cells that occur in groups and sheets (arrowheads, inset). Also seen is severe interface hepatitis with spillage of plasma cells into the lobules. (D) Plasma cell score = 3. The overall inflammation seen in this portal tract is mild, but >30% of the inflammatory infiltrate is composed of plasma cells. This denotes that an appreciable plasma cell infiltrate can be seen even with a mild degree of inflammation. Note that in the examples shown, there is absence of endothelialitis, significant eosinophilic infiltrate, or immature lymphoid cells, which would suggest acute cellular rejection. Also, well-formed lymphoid follicles (seen in recurrent HCV) are not present.

Figure 2.

Centrilobular hepatocyte necrosis accompanied by plasma cells (arrowheads). There is no central venulitis present.

At our institution, biopsies in patients with HCV are performed in a protocol fashion every 6 months when patients are started on IFN treatment and for 1 year after its cessation. This practice started in early 2004; prior to this time, biopsies were performed only when clinically indicated in the setting of abnormal liver chemistry tests. Mild ACR in patients with HCV is not treated; moderate ACR is treated with optimization of calcineurin inhibitor (CNI) levels; and with severe ACR, patients receive 1 g of intravenous corticosteroid. Steroid-resistant ACR is treated with OKT3. Treatment of HCV is initiated at the discretion of the hepatologist once histologic evidence of recurrence is confirmed by biopsy. In 2004, our institution started following a strict protocol using pegylated IFN-α2b and ribavirin; prior to this period, patients received different treatment regimens of IFN. Once a diagnosis of PCH was made on biopsy, treatment was left to the discretion of the physician.

A chart review of all patients fitting inclusion criteria was undertaken, and the following clinical information was recorded: recipient age and gender, donor age and gender, type of immunosuppression and any changes made within a 3-month period prior to the diagnosis of PCH, cyclosporine or tacrolimus trough levels, HCV genotype, history of ACR, presence of autoantibodies at the time of diagnosis of PCH when serologically assessed (anti-nuclear antibody, anti-smooth muscle antibody, liver-kidney-microsomal antibody, and anti-mitochondrial antibody), and liver chemistry tests (aspartate aminotransferase, alanine aminotransferase, and bilirubin) at the time of diagnosis of PCH. Autoantibody titers ≥ 1:40 were considered positive. Tacrolimus trough levels > 7.0 ng/mL and cyclosporine trough levels > 90 ng/mL were considered therapeutic. HLA typing of recipients is not routinely performed at our institution. The administration, timing, and HCV virological response to IFN therapy and the treatment, if any, of PCH were also noted. Special emphasis was placed on biopsies performed preceding and subsequent to the diagnosis of PCH. Poor outcome was defined as death, the histological development of cirrhosis and/or clinical portal hypertension, or the need for re-LT.

Statistical Analysis

Data were analyzed with the statistical software package SPSS for Windows, version 15.0 (SPSS, Inc., Chicago, IL). Results are reported as median values and ranges. Characteristics before and during treatment were compared with the Mann-Whitney exact test for continuous data and the exact chi-square test for categorical data. Changes in clinical and histological features over time were assessed with Wilcoxon exact tests for continuous data and the exact chi-square test. We excluded 2 patients from the analysis with respect to histological resolution of PCH because their last biopsy was within days to weeks of the index biopsy. P values lower than 0.05 were considered statistically significant.

RESULTS

A total of 66 patients with HCV were identified from the pathology database; a total of 354 liver biopsies and explants were examined. Upon review of the histological and clinical data, 28 (42%) patients were excluded. Reasons for exclusion were concomitant ACR or chronic rejection, mechanical bile duct obstruction, hepatic artery thrombosis/stenosis, cytomegalovirus or other nonhepatotropic viral infection, and medication-induced hepatotoxicity. In an effort to obtain a group of patients as homogeneous as possible, patients with peak plasma cell scores of 2 were not included because plasma cells may be seen in recurrent viral hepatitis and in rejection.35, 38 Thirty-eight patients met inclusion criteria; native explants showed cirrhosis and evidence of HCV in all patients. A total of 214 biopsies and explants were histologically assessed and scored; there was an average of 4 biopsies per patient post-LT (range 2-11). The clinical characteristics of these 38 patients are shown in Table 1.

Table 1. Clinical Characteristics of Patients with Plasma Cell Hepatitis
  1. NOTE: All numbers are presented as median (range).

  2. Abbreviations: ACR, acute cellular rejection; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; LT, liver transplantation.

n38
Gender23 male, 15 female
Age53 (34–70)
 <50 years old10 (26%)
 >50 years old28 (74%)
Era of liver transplantation 
 1994–19956 (16%)
 1996–200011 (29%)
 2001–200621 (55%)
ImmunosuppressionTacrolimus = 27 (71%)
 Cyclosporine = 11 (29%)
HCV genotype 
 123 (61%)
 22 (5%)
 32 (5%)
 42 (5%)
 Not assessed9 (24%)
Presence of HCC11 (29%)
Episodes of ACR 
 Yes21 (55%)
 No17 (45%)
Donor age40 (11–76)
 <259 (24%)
 25–5017 (45%)
 >50 years12 (31%)
Gender mismatch16 (42%)
Number of liver biopsies post-LT4 (2–11)
Outcome 
 Positive (alive)15 (40%)
 Negative23 (60%)
 Cirrhosis/portal hypertension10
 Death10
 Retransplantation3

Clinical Analysis

A majority of the patients were male (60%), were >50 years old, and had HCV genotype 1; these features were typical of our LT population. All patients had a detectable HCV viral load post-LT. Seventy-one percent of the patients received tacrolimus-based immunosuppression, and 42% underwent a gender-mismatched LT. Only 23/38 (61%) had determination of autoantibodies, 14/23 (61%) with a positive titer ≥ 1:40. A majority of the patients (32/38, 84%) underwent LT after 1996. Interestingly, no cases of PCH were diagnosed before 1999. A total of 23/38 (60%) with PCH died or developed allograft failure (see Table 1); 10/23 died, and 13/23 developed graft failure and portal hypertension. Overall, death occurred at a mean of 19 (2-54) months after PCH was diagnosed; 4 patients died within 5 months of PCH because of sepsis or allograft failure.

Clinical data of the 38 patients at the time of index biopsy demonstrating PCH are shown in Table 2. PCH was diagnosed in 34% of patients within the first year and in 58% within the first 2 years post-LT. At the time of PCH, almost all patients had aminotransferase levels above the normal range, with approximately one-third having aspartate aminotransferase or alanine aminotransferase > 250 IU/L; a majority of patients had a normal bilirubin level. However, a bilirubin level above 1.3 mg/dL was significantly associated with negative outcome (P = 0.02). Eighty-two percent (31/38) of patients had either recent lowering of their immunosuppression or subtherapeutic CNI trough levels; 2 of the 7 remaining patients were receiving IFN. Fifty-five percent of patients had a history of histological ACR; however, with the specific histological grading system of the appendix, only 7 (18%) cases met strict criteria for ACR (see Table 3). The index biopsy showing PCH was initially diagnosed as ACR in 6 patients (4/6 in 2003 or earlier). All of the patients transplanted > 5 years (7/38) prior to the development of PCH had a recent lowering of their immunosuppression.

Table 2. Clinical Data at the Time of Index Liver Biopsy Demonstrating PCH
Time between liver transplantation and index biopsy (months)17 (range = 3.6–173.4)
  • NOTE: All numbers are reported as medians.

  • Abbreviations: ALT, alanine aminotransferase; AMA, anti-mitochondrial antibody; ANA, anti-nuclear antibody; ASMA, anti-smooth muscle antibody; AST, aspartate aminotransferase; LKM, liver-kidney-microsomal antibody; PCH, plasma cell hepatitis.

  • *

    Positive AMA, ANA, ASMA, or LKM in titer ≥ 1:40.

  • Several patients received 2 forms of treatment.

 <1 year13 (34%)
 1–2 years9 (24%)
 2–3 years5 (13%)
 3–54 (11%)
 >5 years7 (18%)
Year of PCH diagnosis 
 19992 (4%)
 20004 (11%)
 20016 (16%)
 20025 (13%)
 20036 (16%)
 20045 (13%)
 20054 (11%)
 20066 (16%)
Patients having liver biopsy prior to index biopsy32
 0–6 months11 (34%)
 6–12 months7 (22%)
 1–2 years6 (19%)
 >2 years8 (25%)
Lowering of immunosuppression within 3 months18 (47%)
Mean immunosuppressant trough level (ng/mL) 
 Cyclosporine93 (range = 22–420)
 Tacrolimus7 (range = 2–14)
Therapeutic immunosuppressant levels16 (42%)
 Cyclosporine5 (31%)
 Tacrolimus11 (69%)
Patients having recent immunosuppression lowering or subtherapeutic calcineurin inhibitor levels31/38 (82%)
Autoantibodies* 
 Present14 (37%)
 ASMA6
 ANA3
 ANA, ASMA2
 AMA1
 ANA, LKM1
 ASMA, ANA1
 Absent9 (24%)
 Not Assessed15 (39%)
AST (IU/L)159 (range = 38–863)
 Normal0 (0%)
 50–15018 (47%)
 151–2507 (18%)
 >25013 (35%)
ALT (IU/L)71 (range = 24–950)
 Normal1 (3%)
 50–15014 (37%)
 151–2509 (23%)
 >25014 (27%)
Bilirubin (mg/dL)1.0 (range = 0.3–20)
 Normal22 (58%)
 1.3–3.06 (16%)
 >3.010(26%)
Interferon use at index biopsy14 (37%)
Undetectable hepatitis C viral load4/14 (29%)
Treatment16/38 (42%)
 Azathioprine/6-mercaptopurine8
 Prednisone or solumedrol alone6
 Mycophenolate mofetil2
 Increase calcineurin inhibitors5
Table 3. Summary of Histological Features of Liver Biopsies Prior to, at the Time of, and Subsequent to the Index Liver Biopsy Demonstrating PCH
Histological ScoringPre-PCH (n = 33)PCH (n = 38)Initial Biopsy Post-PCH (n = 28)Last Follow-Up Biopsy (n = 18)
  • Abbreviations: AIH, autoimmune hepatitis; HCV, hepatitis C virus; PCH, plasma cell hepatitis.

  • *

    Reported as median scores.

Initial assessment*1321.5
 Normal/other (0)2 (6%)02 (7%)2 (11%)
 HCV without AIH features (1)21 (64%)07 (26%)7 (39%)
 HCV with AIH features (2)9 (27%)06 (21%)4 (22%)
 Typical of AIH (3)1 (3%)38 (100%)13 (46%)5 (28%)
Total score*711109
 <716 (48%)06 (21%)6 (33%)
 7-914 (42%)2 (5%)6 (21%)6 (33%)
 10-123 (10%)36 (95%)16 (58%)6 (34%)
Portal inflammation*2332.5
 Normal (0)1 (3%)001 (5%)
 Mild (1)12 (36%)3 (8%)3 (10%)3 (17%)
 Moderate (2)17 (52%)12 (31%)10 (36%)5 (28%)
 Severe (3)3 (9%)23 (61%)15 (54%)9 (50%)
Interface hepatitis*1322
 Normal (0)5 (15%)001 (5%)
 Mild (1)13 (39%)2 (5%)4 (14%)3 (17%)
 Moderate (2)12 (36%)9 (24%)12 (43%)9 (50%)
 Severe (3)3 (10%)27 (71%)12 (43%)5 (28%)
Lobular activity*2332
 Normal (0)001 (4%)1 (5%)
 Mild (1)13 (39%)02 (7%)3 (17%)
 Moderate (2)15 (46)4 (10%)5 (18%)5 (28%)
 Severe (3)5 (15%)34 (90%)20 (71%)9 (50%)
Plasma cells*1322
 None (0)6 (18%)05 (18%)4 (22%)
 Occasional (1)15 (46%)03 (11%)4 (22%)
 Moderate (2)10 (30%)08 (28%)5 (28%)
 Severe (3)2 (6%)38 (100%)12 (43%)5 (28%)
Centrilobular necrosis with plasma cells    
 Yes14 (42%)38 (100%)20 (71%)9 (50%)
 No19 (58%)08 (29%)9 (50%)
Fibrosis (Knodell/Ishak)*2344
 06 (18%)001 (6%)
 16 (18%)2 (5%)1 (3%)2 (11%)
 210 (30%)8 (22%)3 (11%)1 (6%)
 37 (19%)10 (26%)8 (29%)4 (21%)
 43 (9%)6 (16%)4 (14%)2 (12%)
 51 (2%)10 (26%)9 (32%)6 (33%)
 602 (5%)3 (11%)2 (11%)
Acute cellular rejection    
 Yes5 (15%)Not1 (4%)1 (6%)
 No28 (85%)applicable27 (96%)17 (94%)

Fifty-six percent of patients treated for PCH had a good outcome, whereas 27% of patients not receiving treatment for PCH had a good outcome (P = 0.07). There was no difference in outcome between patients treated with steroids and those not receiving any treatment at all (P = 0.68), but after exclusion of patients treated with steroids, a significant difference in outcome was found between patients treated without the use of steroids and those not treated at all (P = 0.02).

Eleven of 14 patients had liver biopsy performed when they were receiving IFN because their liver chemistry tests increased. Three patients had undetectable HCV viral load at the time of PCH.

Statistically, IFN use was not associated with the development of PCH. Twelve of 14 (86%) patients receiving IFN had a recent lowering of immunosuppression or subtherapeutic CNI levels. None of these patients had their IFN discontinued after PCH was diagnosed. Fifty-four percent (12/22) received IFN after PCH was diagnosed; 8/12 (67%) had a virological response. IFN use at the time of or subsequent to PCH did not significantly influence patient outcome.

Using logistic regression analysis, we found no clinical variables such as gender, age, era of LT, type of CNI, HCV genotype, presence of autoantibodies, donor age, treatment of PCH, or gender mismatch and no histologic variables to be independent predictors of resolution of PCH.

Histological Evaluation

Table 3 summarizes the histological features of each patient's index biopsy, pre-PCH biopsy, and all post-PCH liver biopsies. Of the 33 patients having a pre-PCH biopsy, 21 had histological features of recurrent HCV; very few patients had severe portal inflammation, interface hepatitis, or lobular activity or a total score above 9 that would be consistent with a severe HCV recurrence.38 We did not examine the significance of portal inflammation, interface hepatitis, or lobular activity on clinical outcome or histological resolution because we could not reliably differentiate PCH from recurrent HCV solely on the basis of these features. However, 25 patients (76%) had some degree of plasma cell infiltration pre-PCH, with over half (14/25) already manifesting focal centrilobular necrosis (severe plasma cell infiltrate was found in 2 other patients, but they also had ACR.).

At the time of the index biopsy, 95% of patients had a total score ≥ 10, with a large number having severe portal inflammation, interface hepatitis, or lobular activity; 75% of these patients had a liver biopsy performed within 2 years prior to the index biopsy. Seventy-three percent of patients had ≥stage 3 fibrosis at index biopsy, 31% with stage 5 or 6. The stage of fibrosis, however, did not impact on the resolution of PCH or clinical outcome.

Twenty-eight patients had liver biopsies performed at a mean of 18 months (range 0.2-62.4) after diagnosis of PCH (see Table 2). Nine patients no longer showed evidence of PCH after receiving treatment; however, 16/28 (57%) had a total score ≥ 10 with significant plasma cell infiltrate consistent with PCH persistence. More than 50% of patients continued to have severe portal inflammation, interface hepatitis, or lobular activity and severe plasma cell infiltrate. Eighteen patients had progression of their fibrosis from the index biopsy, 11/18 (61%) progressing 1 stage. Seven patients had no change in fibrosis, but 5/7 (71%) were stage ≥ 5. Three patients regressed 2 stages of fibrosis. Seven of 28 patients (25%) had a follow-up liver biopsy within 1 month of PCH diagnosis; 1/7 (14%) had fibrosis progression, and 3/7 (43%) had histologic resolution of plasma cell infiltrate, all patients having been treated for PCH.

Eighteen patients had longer term histological follow-up (mean 27 months, range 0.5-59). Table 3 shows that the initial assessment changed overall from 3 to 1.5, and this was consistent with improvement or resolution of PCH. Overall, there was improvement in portal inflammation, interface hepatitis, and lobular activity scores, but it was not possible to distinguish how much of the residual inflammation was due to persistent PCH or HCV. There was a decrease in the amount of plasma cell infiltration over time, with there being a greater decrease in patients receiving treatment for PCH versus those not treated. Patients with persistent plasma cell infiltrate on last follow-up biopsy all had a negative outcome (P = 0.001). A good outcome was more likely to be present in patients not having severe inflammation on the index biopsy (score 4-6; P = 0.04) or in those with histological resolution of PCH (P < 0.001). Six patients continued to have an initial assessment score of 3 (persistent features of PCH), 4/6 with fibrosis progression. Five of these patients had a negative outcome (P = 0.007). Four patients had an eventual recurrence of PCH after resolution, all having a negative outcome.

There were no pre-PCH or index biopsy histological features that impacted on histological resolution of PCH or clinical outcome. Patients with a persistently high total score on their initial postindex biopsy were less likely to resolve PCH (P = 0.03) and trended toward a negative clinical outcome (P = 0.09). Patients with fibrosis stage 3 or greater on their last biopsy post-PCH were more likely to have a negative outcome (P = 0.04). Two patients had an initial assessment score of 0 both before and after the index biopsy of PCH.

DISCUSSION

De novo AIH has been mainly described in the pediatric LT population6, 7, 8, 14, 15, 17 but has been increasingly reported in adults transplanted for various liver diseases.9–13, 16 Initially described in adults transplanted for primary biliary cirrhosis,12de novo AIH was felt to be a form of allograft dysfunction within the spectrum of autoimmune liver disease. Evans et al.7 found that this de novo AIH frequently resulted in graft failure in children and proposed it to be a form of chronic rejection. More recently, Khettry et al.39 found 9/92 patients transplanted for HCV to have an AIH-type histology with a more rapidly progressive fibrosis. They suggested that this AIH histology was a variant form of recurrent HCV.

In the current study, we have identified a group of 38 patients undergoing LT for HCV that developed a dense plasma cell infiltrate that may be seen in de novo AIH, which we have termed posttransplant PCH. Eighty-two percent of patients developed PCH after a recent lowering of their immunosuppression or subtherapeutic CNI levels, and this suggests that PCH was a variant of allograft rejection. Fifty-five percent of patients had a previous episode of ACR, and this suggests some predisposition to developing rejection. Typical ACR rates at most LT centers vary from <10% to 20%; an initial episode of ACR is a risk factor for developing further episodes of ACR.40 Upon retrospective review, some of these cases of ACR in fact were found to be PCH, and this demonstrated the potential histological similarity between the 2 entities. Centrilobular necrosis can be seen in severe ACR,35 and this may have led to the diagnosis of ACR instead of PCH in those patients who concurrently had a plasma cell infiltrate. Endothelialitis, bile duct damage, and a mixed cellular infiltrate occur in typical ACR35 but were not present in these cases; this suggests more of a hepatocellular rather than endothelial cell immunologic insult. Patients had a mean follow-up of 18 months, with none developing chronic rejection. This differs from the findings of Evans et al.7 in their pediatric population.

Several features of our study suggest that PCH is not a true AIH. AIH rarely recurs within the first 2 years post-LT1–3; 58% of our patients were diagnosed with PCH within 2 years of LT, 34% within 12 months. Of 7 patients having a follow-up biopsy within 1 month of PCH, 3/7 had a fairly rapid resolution of plasma cell infiltrate with treatment; this rapid histological recovery is not typical for AIH.1–3 Fourteen of 23 patients assessed had positive autoantibody titers; however, low titer autoantibodies have been reported to occur in the setting of both recurrent HCV41, 42 and ACR.43 Patients with titers less than 1:160 are less likely to have a true autoimmune process20; almost all of the patients in the current study had low titer autoantibodies.

HCV patients may not have as high a risk of ACR as patients who have immunologic forms of liver disease such as PBC and PSC44, 45 and who may potentially have a predisposition to developing de novo AIH. Concurrence of HCV and AIH is infrequently described.46 There are reports of patients developing AIH and other autoimmune phenomena with IFN therapy.47, 48 Berardi et al.23 reported 9/44 patients receiving pegylated IFN and ribavirin who developed a de novo AIH post-LT. They concluded that IFN may have induced this AIH, which resulted in graft failure in 4 patients. Their patients did not all have histologic evidence of AIH, some being diagnosed on the grounds of autoantibodies and the International Autoimmune Hepatitis Group score.20 Some of the patients in fact had histologic evidence for ductopenia; the occurrence of chronic ductopenic rejection has now been increasingly reported in patients receiving pegylated IFN and ribavirin49 and may have been the cause of the allograft failure in Berardi et al.'s patients rather than de novo AIH. In our study, IFN did not appear to play a role in the development of PCH, its histological resolution, or the clinical outcome. PCH was diagnosed in most patients receiving IFN therapy because an increase in liver chemistry tests triggered the performance of a liver biopsy. This may have occurred secondary to a lowering of immunosuppression or subtherapeutic immunosuppression. PCH should be included in the differential diagnosis of abnormal liver chemistry tests occurring in patients receiving IFN therapy, especially if the HCV viral load is undetectable.

The strength of our study is the use of a uniform histological scoring system to diagnose PCH and the review of a large number of biopsies and patients. The scoring system was based on several standardized and well-accepted scoring schemes30–33, 35; the only nonstandardized components of our scoring system were “initial impression” and “plasma cell score,” both of which are easily assessed. We have tried to better evaluate the nature of PCH by studying a homogeneous group of patients with nonimmunologic liver disease. A few plasma cells can be seen in the presence of HCV recurrence and may also be seen in ACR,35, 38 making the diagnosis of PCH sometimes problematic. This is why we limited the inclusion criteria only to patients having >30% plasma cells. Future studies should examine the significance, progression/regression, and treatment of lesser degrees of plasma cell infiltration occurring in post-LT HCV recurrence.

Treatment of PCH with increased immunosuppression itself showed a trend toward better clinical outcome; persistence of PCH on subsequent biopsies was strongly associated with negative outcome (P = 0.007), as was the presence of advanced fibrosis (P = 0.03). Although only a small number of PCH patients were treated, it appears that those treated without steroids (that is, optimizing CNI dose) did better than those treated with steroids and those not receiving treatment at all. This implies that steroids may have improved PCH but ultimately led to a more severe recurrence of HCV. It may thus be reasonable to recommend optimizing CNI or adding an agent such as azathioprine instead of increasing corticosteroids in this group of patients because of the potential risks of steroid treatment.26–28 Recurrence of PCH after histological resolution was similarly associated with negative outcome as was persistence of PCH, as all 4 patients with PCH recurrence died or needed re-LT. The significance of these results, however, must be tempered by the possibility of a type 2 statistical error due to the small sample size.

It is unlikely that we missed post-LT patients having >30% plasma cells by using this stringent database search as we tried to be as inclusive as possible by using broad search terms. Interestingly, no cases of PCH were identified before 1999. It is the impression of our liver pathologists that in fact cases of PCH were not seen in the earlier years of our LT program. One explanation for this may be the more aggressive approach adopted over the last 5 to 7 years26–28 at weaning immunosuppression in patients transplanted for HCV and greater awareness about and recognition of “de novo AIH.” As the aim of our study was to better characterize in a standardized fashion the entity of PCH, specifically in HCV patients, and not to compare its outcome and so forth to other post-LT HCV patients, we did not feel a comparative HCV control population was necessary to validate our findings. As numerous studies of heterogeneous groups of patients have shown that increasing immunosuppression in de novo AIH may or may not affect outcome,6–8, 10, 11, 13, 16, 17 we did not feel that comparing our cohort with a control group of non-HCV LT patients developing de novo AIH would be useful. In this study, we did not consider patients with lesser degrees of plasma cell infiltration as having PCH in an effort to avoid overdiagnosing PCH in patients frequently having one or more concurrent disease entities that could have plasma cells, that is, recurrent HCV with mild ACR.

The current study demonstrates that a large number of HCV patients having PCH go on to develop allograft failure. Although we cannot be sure that HCV by itself did not contribute significantly to this, the majority of patients did not have severe necroinflammatory activity or advanced fibrosis prior to PCH, and close to half of patients returned to an initial assessment score of 1 (histological HCV) on post-PCH biopsy. Thus, PCH does not appear to be a variant form of recurrent HCV, as suggested by Khettry et al.39 Our patients with persistent or recurrent PCH had a significantly negative outcome, and this implies that the occurrence of PCH in the setting of recurrent HCV may be a risk factor for developing graft failure. Liver biopsy showing persistent or recurrent PCH with a fibrosis score of 3 or greater (P = 0.03) may identify patients at risk for negative outcome. Thus, performing a follow-up liver biopsy after PCH is diagnosed may be prudent.

Limitations to our study include the lack of comparative autoimmune serological markers before and after the diagnosis of PCH, the lack of protocol liver biopsies, and the fact that the histological scoring system that we used has not been previously used and thus is not standardized. We studied only 38 patients, and although this is in fact a large number of patients having the entity of PCH, studies of larger numbers of patients are essential. Future studies assessing the significance of lesser degrees of plasma cell infiltration in patients post-LT are also warranted. Even though in the current study IFN use did not appear to play a role in the development of PCH, because IFN has been associated with other immunologic phenomena, further evaluation of post-LT HCV patients receiving IFN should be undertaken to define its role in the outcome of PCH

In conclusion, in this group of patients with recurrent HCV post-LT, the development of a dense PCH is a negative prognostic factor for allograft failure and survival. This rich PCH appears to be more likely a histologic variant of rejection rather than a true de novo AIH. It often occurs in the setting of recently lowered or subtherapeutic immunosuppression; in contrast to the findings of previous investigators, we have found no correlation with the use of IFN or the development of autoantibodies.23 The identification of a dense plasma cell infiltrate may be key to differentiating PCH from ACR or recurrent HCV. Our results suggest that treating PCH by optimizing immunosuppression but not using corticosteroids is warranted because of the potential negative effects of corticosteroids on the natural history of post-LT HCV.

APPENDIX:

Table  . HISTOLOGICAL SCORING SYSTEM UTILIZED TO ASSESS PLASMA CELL HEPATITIS IN THE STUDY PATIENTS
  1. NOTE: The grading of portal inflammation, interface hepatitis, and lobular activity is adapted from METAVIR30 and the Knodell Histological Activity Index.31 Fibrosis was staged with the Ishak scoring system.32 Steatosis was graded with the Nonalcoholic Steatohepatitis Clinical Research Network Histologic Scoring System.33 The rejection activity index was graded with the Banff criteria.35

  2. Abbreviations: AIH, autoimmune hepatitis; HCV, hepatitis C virus.

Initial assessmentScore
 Other0
 HCV histological features without AIH features1
 HCV with AIH features2
 Typical of AIH3
A. Portal inflammationScore
 Normal0
 Mild: sprinkling of inflammatory cells in occasional portal tracts1
 Moderate: dense inflammatory infiltrates in some portal tracts2
 Severe: dense inflammatory infiltrates in most portal tracts3
B. Interface hepatitisScore
 None0
 Mild: focal, only seen in a few portal tracts1
 Moderate: affecting <50% of the limiting plate in most portal tracts or continuous in few portal tracts2
 Severe: continuous in most portal tracts3
C. Lobular activityScore
 None0
 Mild: scattered foci of parenchymal necrosis (<5 foci/10× field)1
 Moderate: numerous foci of parenchymal necrosis (>10 foci/10× field)2
 Severe: moderate as above plus bridging necrosis and/or collapse3
D. Plasma cellsScore
 Absent0
 Occasional1
 Moderate (<30%)2
 Severe (>30% or in sheets or clusters)3
Total score (A, B, C, and D)
Centrilobular necrosisScore
 Mixed inflammatory infiltrate (including plasma cells)Yes
 No
Fibrosis (0–6) 
Steatosis (0–3) 
Rejection activity index (0–9) 

Ancillary