Interface hepatitis is associated with a high incidence of late graft fibrosis in a group of tightly monitored pediatric orthotopic liver transplantation patients


  • Denise Herzog,

    Corresponding author
    1. Department of Pediatrics, Centre Hospitalier Universitaire Sainte-Justine, Hôpital St-Luc, Université de Montréal, Montreal, Québec, Canada
    • Division of Gastroenterology, Hepatology, and Nutrition, Hôpital Sainte-Justine, 3175, Côte Sainte-Catherine, Montreal, Quebec, Canada H3T 1C5
    Search for more papers by this author
    • Telephone: 514-345-4626; FAX: 514-345-4999

  • Dorothée Bouron-Dal Soglio,

    1. Department of Pathology, Centre Hospitalier Universitaire Sainte-Justine, Hôpital St-Luc, Université de Montréal, Montreal, Québec, Canada
    Search for more papers by this author
  • Jean-Christophe Fournet,

    1. Department of Pathology, Centre Hospitalier Universitaire Sainte-Justine, Hôpital St-Luc, Université de Montréal, Montreal, Québec, Canada
    Search for more papers by this author
  • Steven Martin,

    1. Department of Pediatrics, Centre Hospitalier Universitaire Sainte-Justine, Hôpital St-Luc, Université de Montréal, Montreal, Québec, Canada
    Search for more papers by this author
  • Denis Marleau,

    1. Department of Liver Unit, Centre Hospitalier de l'Université de Montréal, Hôpital St-Luc, Université de Montréal, Montreal, Québec, Canada
    Search for more papers by this author
  • Fernando Alvarez

    1. Department of Pediatrics, Centre Hospitalier Universitaire Sainte-Justine, Hôpital St-Luc, Université de Montréal, Montreal, Québec, Canada
    Search for more papers by this author


Chronic graft dysfunction, manifesting with elevated liver enzymes and histological features of interface hepatitis (IH), is being increasingly recognized as a long-term problem after liver transplantation. The aim of this study was to characterize our group of post–orthotopic liver transplantation (OLT) patients with respect to clinical, laboratory, and histological signs of IH. A retrospective study of charts and liver biopsy specimens from patients transplanted between 1986 and 1999 was used. Histological features of IH were found in 29/119 patients at a median interval of 23.9 months (95% confidence interval −28.2 to 52.6) after OLT. All patients with IH had risk factors for chronic rejection, such as steroid-resistant rejection, acute rejection later than 3 months post-OLT, female receiver of male graft, or pretransplant cytomegalovirus (CMV)-positive serology with a CMV-negative donor liver. None of the 29 had features favoring a diagnosis of de novo autoimmune hepatitis, but 4 had isolated hypergammaglobulinemia, and 4 had non–organ-specific autoantibodies without hyperimmunoglobulin G. Sixteen of 29 patients also had features of chronic rejection, such as foam cell arteriopathy, loss of bile ducts, or pericentral fibrosis. After abnormal biopsy, all but 1 patient were switched to tacrolimus. During a median follow-up of 12 years, death occurred in 5, retransplantation occurred in 7, and definite cirrhosis occurred in 4. In conclusion, IH was detected in 24.4% of our patients and was associated with a high degree of fibrosis development. Most likely, IH represents a form of chronic rejection directed against periportal hepatocytes. Liver Transpl 14:946–955, 2008. © 2008 AASLD.

Liver transplantation has become a routine, life-saving therapy for children with end-stage liver disease. Current long-term survival rates are over 80%.1, 2 However, chronic graft injuries, manifesting with elevated serum liver enzymes and histological abnormalities, are being recognized more frequently as a long-term problem.3–5 Chronic rejection usually presents with foam cell arteriopathy, loss of bile ducts in at least 50% of portal tracts, significant initial elevation of serum aspartate aminotransferase (AST), and slowly increasing bilirubin throughout its evolution.6 A less frequent form of chronic rejection is directed against pericentral hepatocytes, manifesting as a chronic inflammatory infiltrate, cellular collapse, and ultimate fibrosis.7 Both humoral immunity and cellular immunity are believed to be responsible for this process, and endothelial and biliary epithelial cells are the target. Recent studies have revealed a special type of graft dysfunction, de novo autoimmune hepatitis (AIH).8–10 This term refers to a progressive posttransplant disease associated with chronic active hepatitis at liver biopsy, elevated serum liver enzymes, increased serum immunoglobulin G (IgG) concentrations, and the presence of non–organ-specific autoantibodies. Its incidence in pediatric patients is between 2.3% and 5.2%, and its incidence is even lower in adult orthotopic liver transplantation (OLT) recipients.11 The etiology and pathogenesis of this chronic disease are not known, but molecular mimicry and autoimmunity have been hypothesized.12 Others have tried to classify it as a form of late graft dysfunction in the context of chronic rejection.4, 8, 13, 14 The aim of the present study in our liver-transplanted pediatric population was to assess (1) the frequency of this form of chronic graft dysfunction manifesting with biochemical characteristics also found in patients with AIH, (2) its association with serological parameters of autoimmunity or viral infection and type of immunosuppression, and (3) its outcome in a group of liver transplant recipients followed without protocol biopsies.


AIH, autoimmune hepatitis; ALT, alanine aminotransferase; AMA, anti-mitochondrial antibody; ANA, anti-nuclear antibody; ANCA, anti-neutrophil cytoplasmic antibody; AR, acute rejection; AST, aspartate aminotransferase; BA, biliary atresia; BR, biliary leak and reanastomosis; CHF, congenital hepatic fibrosis; CI, confidence interval; CMV, cytomegalovirus; COACH, cerebellar vermis hypoplasia, ocular coloboma, ataxia congenital, congenital hepatic fibrosis; CyA, cyclosporin A; DC, degenerative changes; EBV, Epstein-Barr virus; F, fibrosis; FO, foamy obliteration; HAT, hepatic artery thrombosis; HCV, hepatitis C virus; Ig, immunoglobulin; IH, interface hepatitis; LKM1, anti-liver kidney microsome antibody 1; MF, media fibrosis; Mmf, mycophenolate mofetil; NAIC, North American Indian childhood cirrhosis; ns, not significant; OLT, orthotopic liver transplantation; PFIC, progressive familial intrahepatic cholestasis; PMN, piecemeal necrosis; PV, portal vein; PVT, portal vein thrombosis; R/D, receiver/donor; SD, standard deviation; SMA, anti-smooth muscle antibody; Tyr, tyrosinemia.


Data on 119 children undergoing OLT at Hôpital Sainte-Justine between 1986 and 1999 and surviving >1 year were analyzed retrospectively. Patient data considered as potential risk factors for graft dysfunction were collected: age, gender, basic disease at OLT, preexisting immune dysfunction (autoimmune diseases and immune deficit), donor and receiver cytomegalovirus (CMV) and Epstein-Barr virus (EBV) status prior to OLT, retransplantation, number of acute rejection episodes, post-OLT viral infections, and type of immunosuppression. Furthermore, data were collected from laboratory tests repeated every 4 to 6 weeks: blood levels of cyclosporin A (CyA) and tacrolimus, serum alanine aminotransferase (ALT), AST, bilirubin, and gamma-glutamyl transferase, complete blood cell count, kidney function tests, electrolytes, proteins, and weight and height. The following tests were conducted on a yearly basis: Ig and autoantibody levels [anti-nuclear antibody (ANA), anti-smooth muscle antibody (SMA), anti-liver kidney microsome antibody 1 (LKM1), anti-liver cytosol antibody 1 (LC1), anti-neutrophil cytoplasmic antibody (ANCA), and anti-mitochondrial antibody (AMA)]. Autoantibody titers were quantified with commercially available kits for indirect immunofluorescence (INOVA, San Diego, CA). ANA was tested on HEP 2 cells, SMA and AMA were tested on stomach and liver mouse tissue, and ANCA was tested on ethanol-fixed and formalin-fixed granulocytes. LKM1 and LC1 were measured by enzyme-linked immunosorbent assay with cytochrome P450 2D6 and human formiminotransferase cyclodeaminase as an antigen and by Western blotting on human liver microsomal and cytosolic proteins. Patients were not tested for genetic susceptibility to AIH. Viral infections were investigated by testing for the presence of EBV-DNA (polymerase chain reaction) or CMV-DNA (nucleic acid sequence–based amplification) and by serology testing (EBV nuclear antigen IgG immunofluorescence, viral capsid antigen IgM and IgG, and early antigen and CMV IgG by enzyme-linked immunosorbent assay). Abdominal ultrasound and complete physical examination were carried out every 3 to 6 months during the first 3 years and then annually. Hepatitis C virus (HCV), hepatitis B virus, hepatitis A virus, parvovirus, herpes virus, EBV, and CMV infections were excluded routinely in the case of elevated aminotransferases, and echographic control was carried out to exclude morphological changes.

Postoperative immunosuppression was as follows: After initial intravenous CyA, oral CyA was introduced when intestinal motility resumed, between 3 and 4 days after OLT, and was divided into 3 daily doses. The CyA targets are listed in Table 1. Azathioprine was administered at a dose level of 1 to 2 mg/kg/day according to the leukocyte count and was continued long-term. Prednisone was started at 2 mg/kg/day and was tapered to 0.3 mg/kg/day by the end of the first month. In patients with normal aminotransferases, an alternate-day prednisone regimen was achieved by 6 months, and corticosteroids were discontinued at 1 year. In some patients, CyA levels were reduced further to 50 to 80 ng/mL after 1 year. Alternatively, tacrolimus was given in combination with prednisone at the target levels listed in Table 1. Table 1 also includes data on the changes in immunosuppressive therapy and in the method of CyA blood level analysis over the years. Patients diagnosed as being in acute rejection were treated with high-dose corticosteroids (methyl prednisolone, 600 mg/m2 of body surface/day for 3 days, tapering at 3-day intervals by 50% of the dose). Patients in chronic rejection or with acute steroid-resistant rejection were changed from CyA to tacrolimus and, before the availability of tacrolimus, were treated with high doses of steroids (600 mg/m2 of body surface), anti-CD3 monoclonal antibody, or polyclonal anti-lymphocyte globulin. Liver biopsies were performed according to clinical indications and/or abnormal laboratory features for more than 2 months. Histological examination was undertaken on needle biopsy specimens routinely processed into paraffin blocks. The following stains were applied routinely: hematoxylin-phloxin-safran, periodic acid Schiff with and without diastase pretreatment, and trichrome collagen staining. Acute and chronic rejections were defined according to Banff criteria,13, 14 and chronic rejection manifesting with foam cell arteriopathy and bile duct loss was recorded separately from that with centrilobular hepatocyte loss.7 Interface hepatitis (IH) was diagnosed according to the International Autoimmune Hepatitis Group consensus.15–17

Table 1. Use of Calcineurin Inhibitors in the Hôpital Sainte-Justine Liver Transplant Program
Target Trough Levels 1986–2004
Time Post-OLTCyA (μg/L)*Tacrolimus (ng/L)
  • Abbreviations: CyA, cyclosporin A; OLT, orthotopic liver transplantation.

  • *

    1986–1998: whole blood monoclonal radioimmunoassay; 1998–2004: fluorescence-polarized immunoassay (AxSYM, Abbott Laboratories, Chicago, IL). Trough target levels declined by 20%.

  • Microparticle enzyme immunoassay (Abbott Laboratories).

0–2 weeks350–40012–15
3–4 weeks250–30012–15
5–8 weeks200–25010–12
9–12 weeks150–20010–12
4–12 months100–1505–10 up to 6 months
>12 months80–1004–5
Changes in the Immunosuppression Regimen
YearsImmunosuppressive Agent
1986–1995CyA Sandimmune
1996–2001Tacrolimus (Prograf) as rescue therapy or for cosmetic problems in stable patients
2001–2004Tacrolimus (Prograf) as primary immunosuppression

AIH-like syndrome was defined as the presence of IH, the elevation of liver enzymes, IgG, and autoantibodies such as SMA, ANA, LKM1, and AMA, and the response to treatment of AIH (2 mg/kg of body weight/day prednisolone and 1.5–2 mg/kg azathioprine).17 Fibrosis was classified according to the METAVIR Study Group with a 5-point score: F0 = no fibrosis, F1 = portal fibrosis without septa, F2 = few septa, F3 = numerous septa without cirrhosis, and F4 = cirrhosis.18, 19 Patients showing histological features compatible with IH7, 15, 20 were selected for further analysis. IH was classified according to the Ishak modification of the histologic activity index grading with a 4-point score: PMN0 = no piecemeal necrosis (PMN), PMN1 = mild (focal, few portal areas), PMN2 = mild/moderate (focal, most portal areas), PMN3 = moderate (continuous around, <50% of portal tracts or septa), and PMN4 = severe (continuous around, >50% of portal tracts or septa).21 Viral infections with an increase of blood EBV-DNA or CMV-DNA levels were treated by reduction of immunosuppression and intravenous ganciclovir until the disappearance of measurable numbers of viral copies.

The patients were divided into 2 groups: group 1 had IH without signs of acute rejection, and group 2 had no IH. Pathology specimens were reviewed in a blinded fashion by 2 independent pathologists (D.B.-D.S. and J.-C.F.). Group 1 patients were further subdivided into those with early IH appearing within 1 year of OLT or with precipitating factors and those with late IH appearing beyond 1 year after OLT. This subdivision appeared necessary because patients with early IH had short follow-up periods ending in retransplantation or death.


All testing was based on determining statistical significance at a 2-sided alpha level of 0.05. The Kolmogorov-Smirnov test was used to flag normal distribution. The 2-sample t test served to compare frequency distributions between subgroups, with the Mann-Whitney U test for comparison of groups without normal distribution. The study was approved by the Research Ethics Committee of Hôpital Sainte-Justine.


Of 119 patients, 29 (24.4%, 15 girls, 14 boys) were found to have IH without simultaneous signs of acute rejection in 1 or more biopsy specimens (group 1). The remaining 90 patients (45 girls) had no IH (group 2).

Patients with IH (Group 1)

IH was detected in 29 patients at a median interval of 23.9 months (95% confidence interval −28.2 to 52.6). The same patients had a median follow-up time of 12 years [range 1–19, standard deviation (SD) 4.8, P = 0.018, Mann-Whitney U], an overall median of 8 clinically indicated biopsies (range 3–15, SD 3.6), and a median of 3 (range 1–3) biopsies taken after the diagnosis of IH. The repeated biopsies taken before the diagnosis of IH showed exclusively acute rejection or normal tissue in 15/29 patients. In 14/29 patients, 1 of the biopsies showed no rejection but steatosis (5 patients), focal lobular hepatitis (3 patients), cholestasis (5 patients), or cholangitis (3 patients). All 29 patients presented 1 or more risk factors for chronic rejection (see Table 2), 22/29 experienced at least 1 episode of acute rejection later than 3 months post-OLT (P = 0.004, Fisher's exact test), and 16/29 displayed features of chronic rejection against endothelial and/or bile duct epithelial cells or centrilobular hepatocytes in 1 or more biopsy specimens. All patients had been investigated for infections, mechanical or ischemic bile duct problems, drug toxicity, or drug abuse prior to biopsy.

Table 2. Primary diseases, year of transplantation, risk factors for chronic rejection and features common of autoimmune hepatitis
inline image


Immunosuppression began with CyA in 28/29 patients and changed to tacrolimus at the diagnosis of IH. If tacrolimus was given alone and targeted levels had not induced improvement, they were increased by 30% (12 patients), then azathioprine was added (7 patients), and lastly it was replaced by mycophenolate mofetil (Mmf; 9 patients; Fig. 1).

Figure 1.

Post-OLT follow-up of fibrosis in relation to immunosuppression in pediatric patients with post-OLT late IH. Only biopsies taken after the diagnosis of IH are listed. The fibrosis score (METAVIR classification), the immunosuppression type, and the grade of IH (Ishak score) are indicated at each biopsy, showing IH without simultaneous acute rejection. Abbreviations: CyA, cyclosporin A; F0–F4, fibrosis grades; IH, interface hepatitis; mmf, mycophenolate mofetil; OLT, orthotopic liver transplantation; P1–P4, piecemeal necrosis grades; siro, sirolimus; tacro, tacrolimus.

Initial Diseases

A significant number of group 1 patients had required OLT for diseases clustered in the province of Quebec: 9 for tyrosinemia and 4 for North American Indian childhood cirrhosis (NAIC; 12/29 = 44.8%). Additionally, Alagille syndrome was represented more frequently than expected in this group. NAIC patients were 67, 146, 63, and 185 months old at OLT, and all were male. IH appeared as early as 13.1 to 24.1 months after OLT. A common risk factor for chronic rejection other than non-Caucasian origin22, 23 in this group was their highly unreliable compliance to medication.

Patients with Alagille syndrome were 54, 56, 131, and 19 months old at OLT (3 males, 1 female), and IH appeared 2.1 to 108.2 months after OLT. The 9 patients with tyrosinemia were younger at OLT (12–99 months, median 30 months, SD 29.7), and IH appeared later (60–108.2 months) than in the 2 former groups. The common risk factor for chronic rejection in this group was young age at OLT.22, 24 For additional individual risk factors for chronic rejection in patients from the 3 groups, see Table 2.

Early IH or IH with Precipitating Factors

In 6/29 patients (20.7%), IH was noted within 1 year of OLT. No hyper-IgG was seen, and autoantibodies (ANA 1/20) were seen only on 1 occasion in 1 patient. One of the 6 died from infectious causes while waiting for retransplantation for chronic rejection, 3 were retransplanted within 2 years for chronic rejection, 1 acquired HCV infection, and 1 developed compensated cirrhosis.

Late IH

In 23/29 patients, IH appeared beyond 1 year post-OLT. One died of postoperative complications after retransplantation in year 3, 1 acquired HCV infection, and a patient with Alagille syndrome developed dialysis-dependent renal insufficiency with compensated liver cirrhosis. Grade of IH, progression of fibrosis, and immunosuppression of the remaining 20 patients are shown in Fig. 1, portal and centrilobular lesions other than IH and fibrosis are shown in Table 3, and a photograph of typical IH is shown in Fig. 2.

Table 3. Histological Changes of Portal Fields and Centrilobular Regions in Patients with IH
PatientPortal FieldsCentrilobular ZoneInterval Between Portal Field/Centrilobular Changes and Diagnosis of IH
Bile DuctsArteriesFibrosisInfiltrateNecrosis/BallooningCholestasisFibrosis
  1. Abbreviations: +, extension of fibrosis or inflammatory infiltrate into zone 3; DC, degenerative changes; F, fibrosis; FO, foamy obliteration; IH, interface hepatitis; MF, media fibrosis; mt, months; n, normal; N, no; w, weeks; y, years; Y, yes.

 Patients with late interface hepatitis
2DCnF3–F4Y+YYN+3 y
4nnF3–F4Y+YNY++1 w
6nnF3Y+NNY+−2 y
7DCnF3–F4Y+NNY++8 y
8nnF0Y+YNY++15 mt
9nnF2Y+NNY++11 mt
10nnF2NNNN−5 y
12DuctopenianF3–F4Y+NNY++6 y
15ProliferationnF3Y+NNY+−2 y
16DCnF3Y+YNY+19 mt
17nnF3NNNY+4 y
19nnF2–F3Y+NNY+2 y
20nnF2–F3Y+NNY+−2 y
23DCnF2–F3NYNY+−2 mt
 Patients with early interface hepatitis
1nnF2NYYY−2 y
2DCnF2YNNY+−6 y
3nMFF3–F4NYYY+6 y
4DuctopeniaMFF2NYYN3 w
5DuctopeniaFOF2NYYN11 mt
6DuctopenianF2Y+YNY+3 d
Figure 2.

Histologic section showing an inflammatory interface lymphoplasmocytic infiltrate. A small bile duct branch (top, center) shows focal lymphocytic infiltration and epithelial irregularity, changes that are suggestive of early chronic rejection (hematoxylin-phloxin-safran, original amplification ×20).

Patients with Autoantibodies and/or Elevated IgG Titers (Table 2)

Autoantibodies and IgG were monitored at yearly intervals from 1998 on. Four patients had persistently elevated IgG titers. The first had additionally a single episode of positive ANA (1/20) and negative titers for the remaining follow-up; the second had autoantibodies tested only once (ANA 1/1280), was transferred to adult care before 1998, required retransplantation for chronic rejection 18 years after OLT for NAIC, and had serious problems with compliance to immunosuppression. Patients 3 and 4 never had elevated autoantibody titers. Patient 3 had been transplanted for AIH type I and had problems with compliance to medication. Four more patients with late IH developed autoantibodies and fibrosis grade F3, while preserving normal IgG or gamma-globulin levels. The first and only ANA elevation (ANA 1/40) of patient 1 was found when the biopsy showed fibrosis grade 3–4. Patient 2 also had autoantibodies elevated (ANA 1/640) only when cirrhosis was established and centrilobular fibrosis, infiltrate, and hepatocellular ballooning were present. Compliance to medication had always been a problem. Patients 3 and 4 had persistent (ANA 1/80) or fluctuating (ANA 1/40 to negative) autoantibody titers, the first with portal fibrosis grade 3 and the second with grade 2 only. Viral infections (EBV, CMV, herpes simplex, parvovirus, or HCV), mechanical or ischemic bile duct problems, or bone marrow disease was repeatedly excluded. The IH was noted at a median interval of 53.4 (range 10.1–117.6) months after OLT. It is worth mentioning that 3 of the patients with elevated IgG titers (including the patient transplanted for AIH I) and both patients with persistently elevated autoantibodies had severe problems with compliance to immunosuppression.

Patients Without IH (Group 2)

Their median follow-up was 7 years (range 1–18). Forty of the 90 (44%) patients experienced an episode of acute rejection and required liver biopsy later than 3 months post-OLT, 23/90 (19.3%) additionally required liver biopsy more than 1 year post-OLT, and 17/90 required liver biopsy later than 2 years post-OLT. ANAs were never tested in 17, tested once and not elevated in 21, tested repeatedly and never elevated in 33, were found to fluctuate between negative and 1/40 in 16, to peak at 1/160 and stabilize at 1/40 upon the introduction of tacrolimus (tyrosinemia, CMV hepatitis post-OLT), to fluctuate between 1/40 and 1/80 (non-A, non-B, non-C fulminant hepatic failure, post-OLT varicella infection), or between 1/40 and 1/320 despite the introduction of tacrolimus (AIH II) in one each. None of these patients had elevated IgG levels or IH, even though liver biopsy was performed later than 3 months post-OLT in the patient with CMV hepatitis and in the patient transplanted for AIH II. Outside of rejection episodes, serum ALT, serum AST, and serum bilirubin levels remained normal in all group 2 patients.

The diseases leading to OLT in the 17 patients requiring liver biopsy > 2 years post-OLT were biliary atresia in 7, tyrosinemia in 5, and Wilson's disease, primary sclerosing cholangitis, Byler's disease, fulminant hepatic failure, and short bowel syndrome in 1 each; postoperative complications were portal vein thrombosis or stenosis in 2 and thrombosis of a hepatic artery in 1. The median follow-up was 10 years [4–7 years, P = not significant (ns) compared with the remaining group 2 and group 1 patients].

EBV and CMV Status at OLT (Groups 1 and 2)

Group 1

Pre-OLT CMV status was negative in 24/29 receivers and positive in 5/29 (17.2%) recipients. Seven CMV-negative patients received a CMV-positive liver, and 5 CMV-positive patients (17.2%) received a CMV-negative liver. Pre-OLT EBV status was negative in 18, positive in 10 (34.5%), and not available in 1/29 receivers, and 5 (17.2%) EBV-negative patients obtained an EBV-positive liver.

Group 2

Pre-OLT CMV status was negative in 61, positive in 22 (24%), and not available in 7/90 receivers (P = ns); 23 CMV-negative patients received a CMV-positive liver (25.5%), and 12 CMV-positive recipients received a CMV-negative liver (P = ns). Pre-OLT EBV status was negative in 53, positive in 30 (33%), and not available in 9/90 receivers (P = ns), and 10 EBV-negative patients obtained an EBV-positive liver (11%, P = ns).

Blood Groups, Donor Gender, Anastomoses, and Diseases Leading to OLT (Groups 1 and 2)

In the recipient blood groups, 11/29 (37.9%) were type O, 15/29 (51.7%) were type A, 2/29 (6.9%) were type B, and 1/29 (3.5%) were type AB in group 1, whereas 36/90 (40.0%) were type O, 34/90 (37.8%) were type A, 11/90 (12.2%) were type B, and 3/90 (3.3%) were type AB in group 2; the blood group was not determined in 4 (P = 0.056). The distribution of blood groups in Quebec patients22 was as follows: O = 46%, A = 42%, B = 9%, and AB = 3% (compared with group 2, P = ns). Of the 12/90 patients with signs of chronic rejection or portal or pericentral fibrosis, only 2 had blood group A. ABO-compatible nonidentical grafts were used in 10/29 (34.5%) patients of group 1 and in 17/90 (18.9%) of group 2 (P = ns). Four (14.3%) boys in group 1 and 17 (18.9%) in group 2 obtained a female liver (P = ns).

Choledochojejunal anastomosis was required in 23/29 patients of group 1 and in 65/90 of group 2 (P = ns). All 29 patients had undergone Doppler ultrasound studies that showed normal flow across the hepatic arteries, hepatic veins, and portal vein. For the basic diseases and frequency distribution, see Table 4.

Table 4. Pediatric Patients with OLT with Interface Hepatitis (Group 1) and without Interface Hepatitis (Group 2) and Preoperative Diagnoses
Primary DiagnosisGroup 1Group 2TotalP*
  • Abbreviations: CI, confidence interval; COACH, cerebellar vermis hypoplasia, ocular coloboma, ataxia congenital, congenital hepatic fibrosis; OLT, orthotopic liver transplantation.

  • *

    Fisher's exact t test.

Cholestatic liver disease    
 Biliary atresia8 (27.5%)32 (35.2%)40 (33.6%) 
 Progressive familial intrahepatic cholestasis2 (6.9%)2 (2.2%)4 (3.4%) 
 Alagille syndrome4 (13.8%)3 (3.3%)7 (5.9%) 
Autoimmune diseases    
 Autoimmune hepatitis1 (3.4%)2 (2.2%)3 (2.4%) 
 Primary sclerosing cholangitis 1 (1.1%)1 (0.8%) 
Fulminant hepatitis 7 (7.8%)7 (5.9%) 
 Hepatoblastoma 2 (2.2%)2 (1.6%) 
 Congenital hepatic fibrosis1 (3.4%)2 (2.2%)3 (2.4%) 
 Short bowel syndrome 2 (2.2%)2 (1.6%) 
 Cystic fibrosis 5 (5.6%)5 (4.1%) 
 COACH syndrome 1 (1.1%)1 (0.8%) 
 North American Indian childhood cirrhosis4 (13.8%)4 (4.4%)8 (6.7%) 
 Tyrosinemia9 (31.0%)18 (20%)27 (22.7%) 
 Others1 (3.4%)7 (7.7%)8 (6.7%) 
 Alpha 1 antitrypsin deficiency 1 (1.1%)1 (0.8%) 
 Wilson's disease 1 (1.1%)1 (0.8%) 
Total29 (24.4%)90 (75.6%)1190.28
Age at OLT [months, median (95% CI)]46 (44.2–81.7)100 (66.8–124.9) 0.5
Interval between OLT and the study year 2006 [months, median (95% CI)]149.5 (135.5–169.6)147 (131.4–170.7) 0.17


Group 1

The evolution of fibrosis under immunosuppression is described in Fig. 1. Only 6/20 (30%) patients with late IH had fibrosis scores ≤ 2 at the time of the study. Among these were 2 patients with short follow-up periods, 1 with tacrolimus alone and 1 with Mmf added. Of the 4 with ≥7 years of follow-up, 2 had tacrolimus alone, 1 had CyA alone, 1 had tacrolimus and Mmf, and 1 had developed nonorgan autoantibodies. In all, 7/20 (35%) had established cirrhosis (F4; including 2 deaths and 1 retransplantation), and 7/20 had bridging fibrosis (35%) despite the subsequent addition of azathioprine and Mmf. Patients more recently diagnosed with IH were directly put on tacrolimus and Mmf because the combination of tacrolimus with azathioprine had not stopped the evolution of fibrosis in prior patients.

Group 2

The diagnosis of the most recent liver biopsy was chronic rejection and F3 in 1, bridging fibrosis (F3) in 3, portal fibrosis (F2) in 2, and normal architecture and no features of chronic inflammation in 11 (64.7%). Initial immunosuppression was CyA in all; 7 had been changed to tacrolimus on the occasion of an episode of acute rejection.


Late liver graft dysfunction is an increasing long-term problem after successful OLT.3–5 Several risk factors have been recognized, and the most important are repeated acute rejections and/or severe steroid-resistant rejection.23, 25–28 In historic CyA-treated cohorts, numerous additional conditions have been considered risk factors for the development of chronic graft dysfunction, such as male donors and female receivers,27 viral infections after OLT especially during childhood with immature thymus function,24 blood group–compatible nonidentical donors,29 AIH leading to OLT,30 mechanical or ischemic graft lesions,31 and past CMV infections32 or human leukocyte antigen DRB*0301 or DRB*0401 positivity in the donor or the receiver.10 IH is one of the morphological features that reveal ongoing graft dysfunction. It is at the same time a histological hallmark of the newly determined type of graft dysfunction known as de novo AIH, a denomination obviously chosen to emphasize the resemblance to AIH, even though the dysfunction occurs on a transplanted organ and represents an “alloimmune” phenomenon. Patients with de novo AIH have been analyzed separately and specific risk factors, such as lymphopenia,33 autoimmune liver disease before OLT, seronegative fulminant hepatic failure, and post-OLT viral infections,24 have been recognized. IH was a frequent finding in our patient cohort. It occurred in 24.4% (29/119) of the patients, and their evolution was characterized by low-level fluctuation of ALT. More importantly, after a median follow-up of 12 years, 23 (79%) of these patients had fibrosis scores > F2 versus only 11% of those without IH. Furthermore, IH appearing within 1 year post-OLT was associated with retransplantation in 50% and death before retransplantation, cirrhosis, and HCV infection in 16.7% each, but not with the elevation of IgG levels or the presence of autoantibodies. Patients with IH appearing > 1 year post-OLT, including those with elevated IgG levels or the presence of non–organ-specific autoantibodies, had a less austere evolution: established cirrhosis, retransplantation, or death occurred in 43.5% in all. None of our patients showed all the features of de novo AIH as defined by the International Autoimmune Hepatitis Group.15–17 This finding is in contrast to the incidences of definite de novo AIH of 2.3 to 5.2% reported in pediatric liver transplant patients from other study groups.10 Instead, 16/29 manifested additional features of chronic rejection, such as foam cell arteriopathy and ductopenia or centrilobular fibrosis. The incidences of IH reported here are related to a population under protocol immunosuppression with frequent laboratory testing and dose adjustments, liver biopsies when clinically indicated, and immunosuppression changed to tacrolimus with azathioprine or Mmf at the appearance of IH.

Our type of follow-up program makes it difficult to compare graft evolution with that of the Birmingham Study Group,3 in which patients underwent 1-, 5-, and 10-year protocol biopsies. Furthermore, diseases leading to OLT have a typical Quebec distribution, with more NAIC and tyrosinemia than in other groups. Nevertheless, the periods of the liver transplantation programs and the immunosuppression regimes are the same in both institutions, and therefore a certain degree of comparison is possible. Effectively, the results are similar: Evans et al.3 reported bridging fibrosis in over 50% and cirrhosis in 15% of their patients with IH at 10-year biopsy, and we found 60% with bridging fibrosis and 14% with cirrhosis after a median follow-up of 12 years. However, only 24% of our patients showed circulating non–organ-specific autoantibodies and at very low titers, in contrast to patients of the Birmingham Study Group, in which over 72% and 80% displayed autoantibodies after 5 and 10 years, respectively. Additionally, no specific association between the number of episodes of late acute rejection and the appearance of autoantibodies could be seen according to other groups.23, 34 A comparison of the patients of the 2 study groups justifies the following questions: (1) is there a relationship between close controls with readily strengthened immunosuppression and the lack of fully developed de novo AIH and (2) are features of de novo AIH and unreliable compliance related as in our group of transplanted NAIC patients. Furthermore, the similarity of outcome in terms of late fibrotic transformation of the graft with or without follow-up protocol biopsies calls for a critical review of the indication for protocol biopsies during late follow-up.

Because various etiologies of de novo AIH are currently being discussed, they should be analyzed with respect to our patient group. Viral infections with known or unknown agents have been proposed as trigger factors,23, 24, 35, 36 especially in children with incomplete T-cell maturation. However, only 17.2% (n = 5) of our patients had a documented viral infection during the year before the diagnosis of IH. Unexpectedly, our patients with IH had blood group A more frequently than patients of group 2, and the difference was nearly significant. In the context of viral infections possibly causing chronic active post-OLT hepatitis,11 a histoblood group antigen recognizing infectious agents, such as norovirus, was a suggestive candidate. However, viruses of this group have been detected only in association with renal transplant rejection37 and not with chronic liver graft dysfunction.38, 39 Furthermore, disseminated norovirus infection, including hepatitis, has been found only in immunodeficient mice40 and not in humans. Lastly, even though episodes of diarrhea are frequent in children, norovirus has not been reported to be the causative agent in our patients until now. In fact, our hypothesis for the etiology of IH is chronic rejection targeting hepatocytes, and it is supported by the risk factors for chronic rejection present in all patients, by the progression of fibrosis despite immunosuppression with azathioprine and prednisone, and by the large proportion of patients displaying simultaneous signs of chronic rejection against other cell types. We further hypothesize that IH is the expression of chronic rejection specifically targeting periportal hepatocytes and that the formation of autoantibodies is an attempt to overcome rejection.41, 42 Effectively, chronic rejection has been described as targeting only pericentral hepatocytes,7 and postulating chronic rejection specifically against periportal hepatocytes implies that we are recognizing a new feature of the chronic allograft rejection process. AIH can serve as a model for the differential immunologic attack of hepatocytes. Indeed, 95% of patients with AIH have perishing periportal hepatocytes, but in 5%, the lesions are exclusively pericentral,43–45 and the antigens targeted are not known. Chronic rejection of hepatocytes seems to follow the same pattern: IH is frequently seen, and chronic pericentral rejection is recognized as a rare, isolated form of chronic rejection.7

In conclusion, 29/119 patients with liver transplantation developed IH in association with an increasing degree of graft fibrosis. However, the presence of all the criteria for de novo AIH, such as heightened ALT and AST, autoantibodies, and elevated IgG levels, was unusual. The lack of such manifestations may be related to the relatively high immunosuppression levels targeted as well as the frequent blood level controls and dose adjustments. The significant number of risk factors for chronic rejection in each patient and the high number of patients with simultaneous signs of chronic rejection against other cell types conjure up an attractive interpretation of IH as an additional sign of chronic rejection.