Natural history of unexplained chronic hepatitis after liver transplantation

Authors


Abstract

Unexplained chronic hepatitis (CH) in the adult liver allograft recipient is not uncommon, but its natural history and clinical significance is unknown. A retrospective study was undertaken of adult liver allograft recipients to determine the frequency and natural history of unexplained CH. We evaluated only those patients who had undergone ≥2 liver biopsies after 6 months and were grafted for indications where recurrent disease could be confidently excluded (alcoholic liver disease [ALD] in those who remained abstinent and fulminant hepatic failure [FHF] from drug reactions). Of 288 patients who were transplanted for ALD or FHF, 30 fulfilled the above criteria. CH was first diagnosed at a median of 15.25 months after transplantation. A total of 24 patients showed mild necroinflammatory changes, and 12 had mild or moderate fibrosis. Liver tests did not reflect the presence or degree of inflammation or fibrosis. After a median of 4 years, necroinflammatory scores were increased in 5; new or progressive fibrosis was noted in 13%; 3 had developed cirrhosis; and 5 developed clinical evidence of portal hypertension. Progressive fibrosis was associated with a high titer of anti-nuclear antibodies (>1:1600) and a portal tract plasma cell infiltrate. There was a trend for correlation between necroinflammatory activity and fibrosis stage, but this did not reach statistical significance (P = 0.06). Serum alkaline phosphatase (P = 0.012) and female gender of the donor (P = 0.033) were associated with progressive fibrosis. Unexplained CH is not uncommon in the liver allograft and may progress to established cirrhosis in a subgroup of patients transplanted for ALD or FHF. Standard liver tests do not reflect the extent of these changes, so protocol liver biopsies may be required to detect these changes. We recommend careful history and follow-up in these patients. Liver Transpl, 2007. © 2007 AASLD.

Liver histology in the allograft is rarely normal.1, 2 The interpretation of the histological findings must be done in conjunction with clinical and serological findings, but there remain many patients for whom no cause for the abnormal histology can be identified.3

Chronic hepatitis (CH) is the most common histological finding in late posttransplantation biopsy samples and occurs in up to 40% of biopsy samples taken after 12 months.4–8 Moreover, there is a poor correlation between the histological findings and both clinical and serologic findings. CH is characterized by the presence of mononuclear portal and lobular infiltrate with or without fibrosis, in the absence of typical features of acute or chronic rejection.9 CH may be the result of identifiable factors such as viral infections (recurrent or acquired), recurrent autoimmune diseases (autoimmune hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis), de novo autoimmune hepatitis, and drug toxicity.1–3, 10, 11 Features of CH are also described in fatty liver disease (both alcohol and nonalcohol related).12–14 However, many patients with CH still do not have any readily identifiable cause of graft damage.1–3 In this subgroup of patients, the natural history, implication for treatment, and the effect on long-term graft function remain unclear.15 To examine the frequency and natural history of CH, we undertook a retrospective clinicopathological study of unexplained CH occurring in liver allograft recipients.

Abbreviations

CH, chronic hepatitis; ALD, alcoholic liver disease; FHF, fulminant hepatic failure; AST, aspartate transaminase; ALP, alkaline phosphatase; BMI, body mass index; ANA, anti-nuclear antibody.

MATERIALS AND METHODS

Patients

Criteria for inclusion in this retrospective analysis were survival for at least 6 months, an index liver biopsy finding showing features of CH, at least one follow-up biopsy, and transplantation for a disease where recurrence could be confidently excluded. We selected patients transplanted for alcoholic liver disease (ALD) where there was no clinical or biochemical evidence of a return to possibly high alcohol consumption (defined as more than 5 units a week). The alcohol consumption was determined by direct questioning of the patient, interviews with family members, and measurement of random blood alcohol levels. We also selected patients transplanted for fulminant hepatic failure (FHF) due to deliberate or accidental poisoning from acetaminophen or idiosyncratic drug reactions. Patients transplanted for viral hepatitis (hepatitis B and C), autoimmune hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, seronegative hepatitis, cryptogenic cirrhosis, or nonalcoholic steatohepatitis-associated cirrhosis were excluded because all of these diseases have the potential to recur and may be associated with features of CH.

Immunosuppression

The immunosuppression protocol was, in general, based on triple immunosuppression with corticosteroids, azathioprine, and a calcineurin inhibitor. Cyclosporine was used until 2000; thereafter, tacrolimus was used. Hydrocortisone 200 mg/day was used immediately after transplant, then switched to prednisolone 20 mg/day. This was gradually reduced and withdrawn at 3 months except in those grafted for autoimmune hepatitis, who were maintained on long-term low-dose prednisolone. The target whole blood trough levels for cyclosporine were 150–200 ng/mL for the first 3 months and 100–150 ng/mL thereafter; for tacrolimus, target levels were 10–15 ng/mL for the first 3 months and 5–10 ng/mL thereafter, with the dose adjusted according to liver and renal function. Mycophenolate (2 g/day) was used if azathioprine could not be tolerated and was also used in combination with calcineurin inhibitors in patients with renal impairment. Since 2004, sirolimus has also been used in those who develop recurrent acute rejections, chronic ductopenic rejection, or renal impairment.

Liver Biopsy and Histology

In the absence of contraindications, liver biopsies were performed either when clinically indicated or as part of the review protocol. Protocol liver biopsies, defined as a biopsy performed irrespective of the results of liver tests, were performed annually after transplantation from 1987 until 1996; after 1996, they were done at year 1 and thereafter at intervals of every 3 years. Liver sections were routinely stained with hematoxylin and eosin, hematoxylin van Gieson, reticulin, orcein, Perl, and periodic acid–Schiff with and without diastase pretreatment. CH was characterized by a predominantly portal-based mononuclear inflammatory infiltrate with interface hepatitis and without biliary or vascular changes characteristic of acute or chronic rejection. Index and subsequent liver biopsy samples were assessed by a system devised when the protocol for obtaining annual review biopsies samples commenced in 1987. Details are as previously reported.4, 6, 16 Briefly, the inflammatory grade was assigned as mild, moderate, or severe, depending on the severity of periportal and/or lobular activity; the fibrosis stage was classified as mild (periportal fibrosis without bridging), moderate (bridging fibrosis), and severe (cirrhosis).

Liver Biochemistry, Serology, and Immunology

All patients had routine blood tests for liver biochemistry (serum bilirubin, aspartate transaminase [AST], alkaline phosphatase [ALP]) taken on the day of liver biopsies. Normal ranges are as follows: bilirubin, 1–22 μmol/L; AST, 5–43 U/L; ALP, 70–320 U/L. Immunological tests (anti-nuclear antibodies [ANAs], anti-mitochondrial antibodies, and anti–smooth muscle antibodies) for autoimmune diseases and serological tests for hepatitis B and C viruses (hepatitis B surface antigen, anti–hepatitis C antibodies) were performed in patients with CH.

Statistical Analysis

Data were analyzed with SPSS for Windows 10.0.7 (SPSS, Chicago, IL). Variables studied included donor and recipient age and gender, body mass index (BMI), AST, bilirubin, and ALP. The Kendall tau-b coefficient was used to measure the association between grading, fibrosis, and donor/recipient gender. The Spearman correlation coefficients were used to measure the associations between other variables.

RESULTS

Patients

Between 1982 and June 2005, a total of 1,968 adults underwent transplantation in this unit. Of these, 288 were transplanted for either ALD or FHF associated with adverse drug reactions. A total of 199 had survived >6 months, and 143 had undergone 1 liver biopsy. CH was found in 46 biopsy samples. Thus, the prevalence of CH in this population was 46 (32%) of 143. Of these, 30 had undergone >1 liver biopsy and so were included in the study. Twenty-one patients were grafted for ALD and 9 for drug-induced FHF. There were 18 men and 12 women; age ranged 19 to 62 years (median, 45 years), and their median BMI was 24 kg/m2 (range, 15–34 kg/m2). Twenty-five were initially treated with corticosteroids, azathioprine, and ciclosporin; 4 with corticosteroids, azathioprine, and tacrolimus; and 1 with tacrolimus and mycophenolate. Immunosuppression at the time of initial biopsy that revealed CH was azathioprine and ciclosporin in 14; azathioprine and tacrolimus in 4; ciclosporin and mycophenolate in 1; prednisolone, ciclosporin, and mycophenolate in 1; prednisolone, azathioprine, and tacrolimus in 1; ciclosporin monotherapy in 4; and tacrolimus monotherapy in 3. In 2 patients, the immunosuppressive regimen was not recorded.

Donor Characteristics

Of the donors, 16 were men, with a median age of 43 years (range, 19–67 years) and a median BMI of 24.1 kg/m2 (range, 20–29.3 kg/m2).

Initial Clinical and Histological Features

CH was diagnosed at a median of 15.25 months after transplantation (range, 6.0–72 months) by protocol biopsy. Of these, 9 were first diagnosed within the first year, 13 (43%) between the first and second year, and the remaining 8 after the second year.

At the time of diagnosis, the median values of routine liver biochemistry were as follows: bilirubin 16 μmol/L (range, 8–68 μmol/L), AST 26 U/L (range, 12–235 U/L) and ALP 279 U/L (range, 105–1279 U/L). No patient was infected with hepatitis B or C virus. ANA results were available for 13 patients. High ANA titers (>1:1600) were detected in 5 patients. Two others had low ANA titers (1:40 to 1:25). Immunoglobulin G levels were measured in 20 patients; 4 showed increased levels (median, 23 g/L; range, 21.5–25.5 g/L). Two of these patients with increased immunoglobulins also had high ANA titers.

Overall, the initial inflammatory activity was classified as mild in 80% (24 of 30 cases) and moderate to severe in the remaining 20%. Fibrosis was noted in 12 of 30 cases but was mild or moderate in severity. Plasma cells were noted to be prominent in 9 cases.

Changes in Immunosuppression

After receiving a diagnosis of CH, 20 patients received changes in their immunosuppression. The majority (13 cases) had the calcineurin inhibitor dose reduced (from a median of 250 mg to 100 mg per day for ciclosporin and 6 mg to 4 mg per day for tacrolimus) because of renal dysfunction; 2 other patients were converted to mycophenolate therapy. One patient developed neurological side effects and was switched from ciclosporin to tacrolimus.

Four patients were treated with additional immunosuppression. All received prednisolone at a median dose of 20 mg (range, 5–20 mg). Two of these patients had plasma cells in the allograft biopsy sample, while a third patient had increased ANA titers in addition to plasma cells at biopsy. The last patient had increased ANA titers in isolation.

Follow-up Clinical and Histological Features

Repeat liver biopsies were performed between 0.58 and 9.4 years (median, 3.9 years) after the index biopsy findings showed CH. In cases where there were more than one follow-up biopsy, data are presented for changes relating to the last biopsy. Three were event-driven biopsies (precipitated by abnormal liver biochemistry or changes in immunosuppression); the others were protocol directed. Follow-up median liver biochemistry values were as follows: bilirubin 19 μmol/L (range, 3–635 μmol/L), AST 27 U/L (range, 15–157 U/L), and ALP 290 U/L (range, 100–1932 U/L).

The follow-up biopsy samples showed worsening inflammatory changes in 5 patients. Results of analysis of previous biopsy samples for all 5 had showed prominent plasma cell infiltrates. Fifteen of 30 patients showed either complete resolution of or a reduction in inflammatory activity, including 4 patients treated by increased immunosuppression (Table 1).

Table 1. Necroinflammatory Activity and Fibrosis Staging at Index and Follow-up Biopsies
(n = 30)Follow-up biopsy
NoneMildModerateSevere
Necroinflammatory index biopsy    
 None0000
 Mild101031
 Moderate1301
 Severe0010
Fibrosis staging index biopsy    
 None10521
 Mild2431
 Moderate0101
 Severe0000

New or progressive fibrosis occurred in 13 of 30 cases at a median of 4 years (range, 0.58–8.4 years) after initial diagnosis. Among the remaining 17, the fibrosis stage remained the same in 14 and improved in the other 3. All 5 patients with high ANA titers (>1:1600) had progressive fibrosis. The greatest progression was seen in patients who had both increased ANA titers and plasma cells at their index biopsies (Table 2). Two of these patients fulfilled the criteria of de novo autoimmune hepatitis,17–21 with increased aminotransferase levels (to at least 1.5 times the upper limit of normal), autoantibody positivity, and increased immunoglobulins. Three patients with progressive fibrosis were noted to have increased severity of steatosis (from initial biopsy) or developed features of steatohepatitis. However, there was no statistically significant association between BMI and changes in inflammation (tau-b = −0.329; P = 0.091) or fibrosis (tau-b = −0.028; P = 0.902). Moreover, there was no significant correlation between the grades of necroinflammatory activity and progression of fibrosis (tau-b = 0.306; P = 0.06). Two patients developed progressive fibrosis despite receiving corticosteroid treatment and achieving a reduction in necroinflammatory scores.

Table 2. Histological and Immunological Associations in Patients With Varying Degrees of Fibrosis Progression
Increase in fibrosis stageHistological and immunological markers (n)
1Steatosis (1)
 Steatohepatitis (1)
 Plasma cell (2)
 High ANA titers (1)
 High ANA titers + plasma cell (2)
 No obvious (2)
2Steatosis (1)
 Plasma cell (1)
 High ANA titers + plasma cell (1)
3High ANA titers + plasma cell (1)

Factors found to be correlated with progressive fibrosis were the levels of ALP (rho = 0.476; P = 0.012) (Fig. 1 and Table 3) and the female donor gender (tau-b = 0.365; P = 0.033).

Figure 1.

Graph showing serum alkaline phosphatase (ALP) and degree of fibrosis changes in 30 patients at time of follow-up posttransplantation biopsy.

Table 3. Association Between Routine Parameters of Liver Biochemistry and Inflammation/Fibrosis
CharacteristicBilirubinASTALT
RhoSigRhoSigRhoSig
  • Abbreviations: AST, aspartate aminotransferase; ALP, alkaline phosphatase; rho, Spearman correlation coefficient; sig, significance (2 tailed).

  • *

    Illustrated in Figure 1.

Degree of inflammation0.1120.5780.3190.1050.150.454
Degree of fibrosis−0.0530.794−0.3030.125−0.2570.196
Change in degree of inflammation−0.0730.719−0.2130.287−0.1070.597
Change in degree of fibrosis−0.1220.5430.3740.0540.4760.012*

Overall, 3 of 30 patients with CH had progressed to established cirrhosis at the time of the latest biopsy (1.7, 6.2, and 8.4 years, respectively, after diagnosis of CH).

Retransplantation and Mortality

Of 13 patients who had progressive fibrosis, 5 developed marked graft dysfunction (portal hypertension, as shown by the presence of ascites and/or oesophageal varices or liver failure) between 0.75 and 8.4 years (median, 2 years) after the diagnosis of CH. One patient underwent a successful regraft for decompensated liver disease, and 2 others died from sepsis and multiorgan failure (mean, 73 months after liver transplantation). Another is alive; the fifth patient has been lost to follow-up.

There were 3 other (non-liver-related) deaths, occurring at median of 62 months after liver transplantation. Two of these were cancer related (oropharyngeal and lung).

DISCUSSION

CH after liver transplantation is common.1, 3, 7 We found that just over 30% of the patients transplanted for ALD or FHF associated with adverse drug reactions developed CH between 6 months and 6 years after liver transplantation. In a recent study within the pediatric population, a population in which allograft recipients may be at low risk of developing recurrent diseases, >60% of recipients developed CH 10 years after liver transplantation.16 At presentation, most patients had mild inflammatory activity and mild or moderate fibrosis. Progressive fibrosis occurred in >40% of patients. An important association with progressive fibrosis was the presence of high ANA titers (>1:1600). Two of these patients had increased immunoglobulins and increased AST, features compatible with de novo autoimmune hepatitis.18–21 Anti–smooth muscle antibodies were present in one patient. The greatest increase in fibrosis was seen in the patient found at biopsy to have both high ANA titers and plasma cells. Previous studies did not assess fibrosis progression and had shorter follow-up.11, 22 Although the role of these plasma cells is unclear, Novobrantseva and colleagues23 have recently shown, by use of carbon tetrachloride–induced fibrosis in a mouse model, that B lymphocytes could be involved in liver fibrosis. Alternatively, the presence of plasma cells and increased ANA titers may simply reflect an activated state of the adaptive immune system that is acting in concert with cytotoxic T lymphocytes to drive liver fibrosis.

The degree of inflammation increased in 5 patients with prominent plasma cell infiltrates. Nakhleh and colleagues11 noted similar findings, when none of their patients with prominent plasma cells experienced resolution in the degree of inflammation on follow-up biopsies (none received treatment). On the other hand, 4 others in our cohort with either high ANA titers or plasma cells at biopsy showed improvement in necroinflammatory scores after corticosteroid therapy was initiated.

Late cellular rejection histologically may resemble autoimmune and viral hepatitis, with more prominent interface hepatitis, less bile duct inflammation, and less subendothelial venular inflammation than typically seen in early acute rejection.1, 24 It has been shown that antibodies are found in association with rejection episodes.25, 26 Nearly half of our patients had their immunosuppression reduced for renal dysfunction, and this could have led to rejection from inadequate immunosuppression. Some cases of CH may therefore be a hepatitic form of late cellular rejection.2, 10 We found that female gender of the donor was a risk factor for progressive fibrosis. This could be related to donor minor histocompatibility antigens mismatches.27

Many centers have abandoned the use of protocol biopsies.28 However, we found that most patients with CH had normal liver biochemistry at initial and subsequent biopsies and were asymptomatic. In this study, all initial diagnoses of CH were made by protocol biopsies. Furthermore, only 3 of 13 cases of new or progressive fibrosis were detected through event-driven biopsies. Although serum ALP levels were found to be far higher in those with progressive disease, individual values of ALP could fall within normal limits. Three of our patients with progressive fibrosis had increased steatosis or developed features of steatohepatitis on subsequent allograft biopsies. In some cases, steatosis and steatohepatitis may be a risk factor for the progression of otherwise unexplained CH. Because there is such poor correlation between symptoms, liver tests, and graft histology, this has strengthened the need for performing protocol liver biopsies because management may be affected by the histological findings. For example, the finding of CH may represent late cellular rejection from suboptimal immunosuppression and hence trigger the need for a thorough history review and closer follow-up of individual patients. We recruited only patients grafted for ALD and drug-induced FHF, where recurrent disease can be confidently excluded. This is in contrast to previous studies in the adult population, where CH may well be attributed to recurrent disease. Despite our small study population, we have shown that CH may lead to progressive liver damage and found that routine biochemical tests cannot be relied on in monitoring the progress of CH, results broadly similar to those reported in a recent study of CH in pediatric liver allograft recipients.16

In conclusion, CH is common and may lead to graft cirrhosis and failure in a subgroup of patients. The risks of disease progression appear to be correlated with the presence of autoantibodies and portal tract plasma cells at liver biopsy. Routine liver biochemistry is not helpful in identifying these patients, and we recommend that protocol biopsies be performed to assess development of allograft abnormalities. The possibility of de novo autoimmune hepatitis or nonalcoholic steatohepatitis should be borne in mind and managed appropriately after usual causes of CH have been excluded. Treatment of unexplained CH remains undefined, but the possibility that this represents a form of late cellular rejection may indicate a need for increased immunosuppression. The role of corticosteroids in these patients remains unproven, and prospective studies will be needed to examine their effectiveness.

Acknowledgements

We thank Claire Thorne for assistance in the retrieval of clinical notes.

Ancillary