Chronic allograft dysfunction: Diagnosis and management. Is it always progressive?



Key Points

  • 1There are many causes of graft dysfunction post–liver transplant, but recurrent disease remains the most common cause.
  • 2Viral hepatitis, nonalcoholic and alcoholic steatohepatitis, and autoimmune diseases are the most common causes of recurrent disease.
  • 3Graft hepatitis occurs frequently and in many cases will not progress.
  • 4Cirrhosis in the absence of any identifiable cause develops in a minority.
  • 5Treatment is of the underlying cause but some, such as recurrent and de novo autoimmune hepatitis and recurrent primary sclerosing cholangitis may not respond well, and regraft may be required. (Liver Transpl 2005;11:63–68.)


There are many causes of graft dysfunction after liver transplantation. In the early days after liver transplantation, ischemic and reperfusion injury predominate. Acute cellular rejection is seen in the first 3 months and thereafter becomes less common. Thereafter, causes of graft dysfunction are variable and are shown in Table 1. As can be seen from Figure 1, graft loss from recurrent disease is a major source of graft loss. In this review, the causes of graft dysfunction after the first 6 months will be discussed, with the exception of recurrent and de novo viral infection.

Table 1. Causes of Liver Allograft Dysfunction
  1. Abbreviations: HBV, hepatitis B virus; CMV, cytomegalovirus; EBV, Epstein-Barr virus; I/R ischemia reperfusion.

ImmunologicalAcute, cellular rejection
 Chronic ductopenic rejection
 Chronic hepatitis
 De novo autoimmune disease
 Recurrent autoimmune disease
ViralRecurrent viral hepatitis (HBV, HCV)
 De novo viral hepatitis (CMV, EBV, HCV, HBV and others)
ToxicDrugs and other toxins
IschemiaLate effects of I/R injury
Recurrent diseaseAutoimmune
Figure 1.

Graft loss from recurrent disease. Diagnoses codes: dark blue line, alcoholic liver disease (ALD); green line, biliary atresia; tan line, chronic active hepatitis (CAH); dark purple line, cryptogenic; yellow line, fulminant; red line, hepatitis B virus; aqua line, HCV; gray line, malignancy; light blue line, metabolic; dark green line, other; orange line, PBC; light purple line, PSC; dark blue cross, ALD censored; green cross; biliary atresia censored; tan cross, CAH censored; dark purple cross, cryptogenic censored; yellow cross, fulminant censored; red cross, hepatitis B virus censored; aqua cross, HCV censored. (Birmingham data; prepared by Rebecca Russell, medical student, and Kerry Webb, CPN.)


AIH, autoimmune hepatitis; HCV, hepatitis C virus; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; CC, cryptogenic cirrhosis.

In medium-term (>5 years) survivors, there are many causes for graft dysfunction. Thus Pappo and colleagues1 in Pittsburgh analyzed 51 symptomatic and 14 asymptomatic patients. In the symptomatic patients, causes for dysfunction included viral hepatitis in 33% (of these, two thirds (64%) were recurrent viral hepatitis and the other third de novo viral hepatitis), 22% had rejection, 14% had recurrent nonviral disease, and 6% had obstructive cholangiopathy. There was a poor correlation between the symptoms, laboratory findings, and histological changes. Broadly similar observations were made by Slapak and colleagues2 in King's College Hospital in London, who studied the biochemical and histologic features of 116 liver allograft recipients who had survived for 5 or more years (up to 19 years). Although 41 (35%) had normal liver tests, 19 (46%) had histological abnormalities. Of the remaining 75 patients with abnormal liver tests, 65 (87%) had abnormal histology; these changes were classified as shown in Table 2. As can be seen, the majority had either structural abnormalities or chronic hepatitis. Of those with chronic hepatitis, about half (14 of 27) were attributed to recurrent B and C viral disease and recurrent autoimmune hepatitis (AIH); however, in 13, no cause could be identified. Of the 26 with structural abnormalities, various types of occlusive vasculopathy were found. In 14 cases, the recipient was maintained on azathioprine and was considered to be the only etiological cause, and 5 showed some improvement when the azathioprine was withdrawn.

Table 2. Causes of Graft Dysfunction in Long-Term Liver Allograft Survivors
  1. NOTE: Data from Slapak et al., 1997.

Chronic hepatitis (C and autoimmune)27
Structural abnormalities26
Primary biliary cirrhosis–like changes8
Mild acute rejection1

Chronic Hepatitis

Chronic hepatitis can be defined as the presence of a portal and lobular mononuclear infiltrate in the absence of evidence of rejection (such as bile duct damage, endothelialitis, and arterial foam cells) and the absence of other identifiable causes of graft damage such as viral infection (Fig. 2). Chronic hepatitis is found relatively commonly in the allograft; indeed, it is very rare to get a report from the histopathologist stating that the graft is completely normal. However, neither the significance nor the prognosis of graft hepatitis is known.

Figure 2.

Chronic hepatitis post–liver transplant: (A) high and (B) low power. (Histopathology courtesy of Dr. Desley Neil, Department of Pathology, University Hospital, Birmingham.)

Birmingham Experience of graft Hepatitis

It has been the practice in our unit in Birmingham to undertake protocol liver biopsies–that is, to take liver histology irrespective of liver tests. Initially, we took liver biopsies annually, but after review in 1996, this was amended to taking liver biopsies at year 1 and then at 3 yearly intervals (with some exceptions, such as those grafted for hepatitis C viral infection). To determine if chronic hepatitis in the allograft was progressive, 50 consecutive recipients were selected who had paired liver biopsies with no specific etiologic features; those with hepatitis C virus (HCV) infection were not included in the analysis. Indications for transplantation in this series included primary biliary cirrhosis (PBC) (14), alcoholic liver disease (10), seronegative hepatitis (8), sclerosing cholangitis (6), alpha 1 antitrypsin deficiency (2), cryptogenic cirrhosis (2), cystic fibrosis (2), drug toxicity (2), Wilson's disease (1), and secondary biliary cirrhosis (1). The median time to first biopsy was 15 months (range, 9 to 71 months) and the median interval between biopsies was 35 months (range, 7-63 months). Liver tests were not greatly deranged, with 13 having elevated serum aminotransferase level and 19 elevated alkaline phosphatase. The degree of fibrosis was assessed on a semi-quantitative scale of 0 to 3. Fibrosis was present in 29 patients. Comparing the first and second biopsy, the degree of fibrosis was unchanged in the 21, improved in 15, and worse in 14.

The increase of chronic hepatitis in those grafted for seronegative hepatitis may affect graft survival.3

Other Experiences of Graft Chronic Hepatitis

A Medline search has produced very little data on the natural history of chronic hepatitis. Nakhleh and colleagues4 in Jacksonville conducted a retrospective review of 704 patients and identified 282 who were at low risk of recurrence. Graft hepatitis was identified in 5% of the series and in 11% of those at low risk of recurrence. Indications for those at low risk included cryptogenic cirrhosis (13), steatohepatitis (12), α1-antitrypsin deficiency (3), acetaminophen toxicity (1), neuroendocrine tumor(1), and epithelioid hemangioendothelioma (1). Hepatitis developed between 2 weeks and 6 years after liver transplantation, but in 22 cases, it was identified within the first 8 months. Follow-up biopsies were available in 25: in 10 the hepatitis had resolved, and in the 11 where the hepatitis persisted, 4 showed reduced infiltration, but there was no significant change in the degree of fibrosis.

Does Chronic Hepatitis Lead to Cirrhosis?

The development of cirrhosis in the graft is uncommon and is usually related to recurrent disease, especially viral hepatitis, recurrent steatohepatitis, and autoimmune liver disease.5, 6 In an attempt to determine the causes and factors leading to graft cirrhosis that is not due to recurrent disease (viral hepatitis, recurrent autoimmune diseases such as PBC, primary sclerosing cholangitis (PSC), AIH, recurrent or de novo viral hepatitis B or C, alcoholic liver disease).7 We examined our database. We identified 48 of 1,287 (3.7%) adult liver allograft recipients who had survived more than 12 months and who had developed a cirrhosis in the graft. In 29, the cirrhosis was attributed to disease recurrence. In 9, a cause for the cirrhosis could be identified: de novo AIH (4), a biliary-type cirrhosis following biliary complications (4), and de novo–acquired hepatitis B viral infection.1 In 10 patients, however, no cause for the cirrhosis was identified. In these cases, the cirrhosis was diagnosed at a median of 84 months post–liver transplantation (14 to 166 months). Of these 10 patients, the transplant was done for PBC (3), PSC (1), and seronegative fulminant hepatitis (6). The degree of inflammation recorded scored a median of 2 (of 5) and fibrosis a median of 2.

Some cases of graft hepatitis may be related to recurrence of disease. In particular, inflammation in the graft may be commoner in those grafted for cryptogenic cirrhosis suggesting that, in some cases, there is recurrence of the as yet unidentified causal agent.

Summary of Chronic Hepatitis

The causes and natural history of graft hepatitis is poorly characterised. Studies of our own and others suggest that graft hepatitis is relatively common. The significance is unclear and that progression is variable. It is uncertain whether graft hepatitis reflects a form of modified rejection, which requires additional immunosuppression or other causes of graft damage, such as drug toxicity. In some cases, it seems that graft hepatitis may lead to a cirrhosis.

Centrilobular Fibrosis

Centrilobular fibrosis has been described in less than 10% allograft recipients where no other cause (viral hepatitis, azathioprine toxicity, or rejection) could be identified.8 This appears to be more common in children. Centrilobular fibrosis is associated with portal fibrosis. Switching from cyclosporine to tacrolimus was associated with some improvement, but stopping azathioprine had little effect. Progression in the short term was not observed. These findings may relate to cold ischemia and biliary complications; of interest, acute rejection appeared to protect from portal fibrosis.9 Whether this progresses is not certain.

De Novo AIH

The initial report of de novo AIH from Kerkar and colleagues10 was of 7 children (4% of their series) who developed a characteristic form of graft damage at a median period of 2 years post–liver transplant. Since then, there have been many reports and series of recurrent AIH occurring in the allograft.11 This condition is characterized by:

  • a biochemical hepatitis

  • circulating autoantibodies

  • elevated immunoglobulins

  • inflammatory infiltration in the graft with interface hepatitis

  • indication for transplantation other than AIH

In most cases, the autoantibodies are anti-nuclear, anti–smooth muscle, or anti–liver/kidney microsomal. But in a recent study, Inui and colleagues have shown antibodies against cytokeratin 8/18 in a patient with this syndrome occurring in a living-donor recipient.12

Most studies suggest that the condition is relatively uncommon, occurring in less than 1% of liver allograft recipients. Children are more commonly affected than adults. There is often a good response to treatment with additional immunosuppression with corticosteroids, but in some cases there is relentless progression to cirrhosis and graft failure.13

Although the term “de novo AIH” has been used, the term has been debated. It has been argued that de novo AIH is appropriate, since the condition arises de novo and fulfills the criteria for AIH. Others have argued that the condition is not necessarily autoimmune. Reasons for this have come largely from work by Aguilera and colleagues.14 They have shown that some patients with de novo AIH have circulating antibodies to a protein, glutathione S-transferase T1, that is present in the graft but not in the donor. There is a good analogy in renal transplantation: Some patients undergoing a renal transplant for Alport's syndrome will develop, posttransplant, antiglomerular basement antibodies, similar to those seen in Goodpasture's syndrome.15 These antibodies may lead to graft damage. Thes antibodies react with the alpha 3 and alpha 5 chains of type IV collagen. The alpha 3 chain is the Goodpasture antigen, and both alpha 3 and alpha 5 are undetectable in the glomerular basement membrane in Alport's syndrome.

Thus, if the graft-damaging immune response seen in de novo AIH is directed against graft rather than host antigen, then this would surely fit the criteria for rejection rather than autoimmune disease. Until these issues are generally resolved, then the phrase suggested by Heneghan—“graft dysfunction mimicking AIH following liver transplantation”—while less easy to use is more appropriate.

Recurrent Disease

It is now clear that many metabolic and immunologic diseases recur and may affect graft survival. These are summarized in Table 3. In most cases, the diagnosis of recurrent disease can be made on clinical, serological, and histological grounds. The main exception is with PSCwhere differentiation between recurrent primary and secondary sclerosing cholangitis may be difficult. Nonalcoholic steatohepatitis may recur in the graft if the underlying trigger has not been corrected.16

Table 3. Recurrent Autoimmune and Metabolic Diseases After Liver Transplantation
DiseaseDiagnosis5-Year IncidenceEffect on Graft SurvivalRisk FactorsTreatmentResponse to Treatment
  1. Abbreviations: AMA, antimitochondrial antibodies; UDCA, ursodeoxycholic acid; Igs, immunoglobulins; HLA, human leukocyte antigen; NASH, nonalcoholic steatohepatitis.

Primary biliary cirrhosisAMA, histology40%MinimalTacrolimusUDCAUnknown
Autoimmune hepatitisAutoantibodiesRaised Igs Histology30%Moderate?HLA absence of steroidsSteroidsVariable
Primary sclerosing cholangitisNonanastomotic strictures exclusion30%Significant?Intact colonUDCAPoor
AlcoholHistoryBlood tests20% graft damageSmallAlcoholism and substance abuseAbstinenceGood
NASHClinicalUltrasound ModerateObesity Drugs, DiabetesSmall bowel surgeryWeight lossCorrection of bowel anatomy?GlitazonesGood
Cryptogenic40%SmallUncertainNil specific 

Treatment of recurrent PSC and PBC is based not on evidence based but on extrapolation of experience of the use of ursodeoxycholic acid in the native liver. A further argument for the use of ursodeoxycholic acid in this setting is the finding that use of ursodeoxycholic acid in both PBC and in PSC is associated with a reduction in the risk of colon cancer development. The risk of colon cancer is increased in liver allograft recipients and more so in those with ulcerative colitis.

The response to increased steroids in those with recurrent AIH is variable, with a significant proportion developing graft failure and needing a regraft.

Recurrence of Cryptogenic Cirrhosis (CC)

For obvious reasons, it is not possible to define the features of recurrent CC. Nonetheless, several authors have reported that those grafted for CC may have a higher incidence of graft hepatitis. Thus, Charlton17 found 6 of 27 patients grafted for CC had persistent hepatitis without an obvious cause; there was no evidence of recurrent cirrhosis. Heneghan18 compared the incidence of graft hepatitis in those grafted for CC, alcoholic liver disease, and HCV infection. Biopsies were all done for investigation of abnormal liver tests. Inflammation was seen in 22% of those with CC, 25% of those with alcoholic liver disease, and 68% of those grafted for HCV; cirrhosis was seen in 0%, 2%, and 16%, respectively. Others too have shown that of those grafted for CC, there is a high incidence of steatohepatitis that may lead to fibrosis and cirrhosis, suggesting that some, at least, of those labeled as cryptogenic may have end-stage steatohepatitis.19, 20

Other Causes of Graft Dysfunction

Drugs are a rare cause of graft dysfunction; both the calcineurin inhibitors are associated with liver dysfunction; changes are usually mild and nonspecific.21 Azathioprine may also result in graft damage and include hepatitis, cholestasis, nodular regenerative hyperplasia, and veno-occlusive disease.22, 23 The presence of sinusoidal congestion and centrizonal necrosis on a liver allograft recipient on azathioprine should lead to a review and probable alteration of immunosuppression.

Investigation of Graft Dysfunction

In most cases, graft dysfunction will be detected on the basis of routine liver tests. However, liver tests are not specific for the cause and severity of the graft damage. A strategy for the investigation in shown in Table 4.

Table 4. Investigation of Chronic Graft Dysfunction
  1. Abbreviations: BMI,; MRC, magnetic resonance cholangiography; ERC, endoscopic retrograde cholangiography; PTC, percutaneous cholangiography.

ClinicalIndication for transplant
 Drug and alcohol history
SerologicalStandard liver tests
 Autoantibodies and immunoglobulins
 Viral studies (especially hepatitis B and C)
ImagingUltrasound with Doppler
 Imaging the biliary tree (MRC, ERC, PTC)
HistologyLiver Biopsy

Protocol Biopsies

The debate as to whether protocol biopsies should be done in stable, well, liver allograft recipients has remained unsettled. To justify liver biopsy in a person who is well and asymptomatic, the following points should be considered:

  • Will the biopsy show abnormalities in the presence of normal liver tests?

  • Will these abnormalities lead to a change in management?

  • Will the findings give any useful information about management and prognosis?

Our initial evaluation24 in 655 protocol biopsies taken >1 year after transplant showed that normal histology was reported in 33 of 367 with normal histology and in 15 of 291 with abnormal liver tests; however, in only a minority did the findings result in a change of treatment (usually amending the immunosuppression). Sebagh25 suggested that protocol biopsies were justified although not on an annual basis. In contrast, Rosenthal26 found frequent abnormalities in children with normal liver tests (portal and parenchymal infiltrations in 48% and 25% of 158 annual biopsies from 54 children, minimal and mild ductular proliferation in 15% and 9%, mild fibrosis (8%), and focal pericholangitis (6%), but annual biopsies did not in alter management. This view is broadly supported by others (outside the context of HCV infection).27