Late graft dysfunction and autoantibodies after liver transplantation in children: Preliminary results of an italian experience



Late graft dysfunction (GD) associated with the development of autoantibodies is a common event after pediatric liver transplantation (OLTx) and can present in 2 clinicohistological subsets: de novo autoimmune hepatitis (DNAH) and early chronic rejection (ECR). Sixty out of 247 children developed autoantibodies after OLTx. GD was demonstrated in 22 (37%); based on histology, patients were divided in a DNAH and an ECR group. Portal/periportal inflammatory infiltrate with interface/lobular hepatitis was suggestive for DNAH. Pericentral hepatocytes confluent dropout with a variable degree of central vein endothelitis, but not with ductopenia (loss of >50% of interlobular bile ducts), was diagnosed as ECR. Nine patients had DNAH and 13 ECR. Five out of 9 in the DNAH group were on cyclosporin (CsA) and 4/9 were on tacrolimus (Tac). In the ECR group, 11 children were treated with CsA and 2 with Tac. All DNAH patients had normal liver function tests on steroids and azathioprine (AZA). Five patients with ECR recovered by increasing calcineurin inhibitors (CNIs) dosage, but in 8/13, including 7 switched from CsA to Tac, AZA and steroids were added to obtain remission of disease. Two patients developed late chronic rejection. DNAH and ECR associated with autoantibodies are forms of late GD after OLTx. DNAH improves after standard treatment of autoimmune hepatitis. ECR has a good response to increased doses of CNIs, although ductopenic chronic rejection may occur. In conclusion, the early differential diagnosis of these conditions and an appropriate treatment seem to allow good overall results reflected by a graft survival of more than 90%. Liver Transpl 12:573–577, 2006. © 2006 AASLD.

Recent reports suggest a close relationship between late graft dysfunction (GD), a progressive posttransplant disease associated with both chronic hepatitis and severe GD, and high immunoglobulin G (IgG) and serum autoantibodies. This situation, known as de novo autoimmune hepatitis (DNAH), may occur a few months up to years after surgery and may be characterized by bridging fibrosis or cirrhosis, leading to graft failure. Usually it does not respond to the conventional posttransplant immunosuppression schedule.1–3

Another late post-pediatric orthotopic liver transplantation (OLTx) complication is a GD characterized by confluent centrilobular necrosis and central vein endothelitis. It is an early form of chronic rejection, and in some reports it was associated with autoantibodies.4–6

In this study we analyzed the clinical course, the different features, and the response to treatment of these two conditions.


OLTx, orthotopic liver transplantation; DNAH, de novo autoimmune hepatitis; ECR, early chronic rejection; GD, graft dysfunction; CNIs, calcineurin inhibitors; CsA, cyclosporin; AZA, azathioprine; Tac, tacrolimus; LFT, liver function tests.


From October 1997 to October 2003, 227 OLTx were performed in 206 children at the Ospedali Riuniti di Bergamo; 41 additional children transplanted elsewhere were followed and included in studied population. Immunosuppression consisted of cyclosporin (CsA) and steroids until 1998. Prednisone was tapered and stopped 3 months after the transplant. From 1999 to 2001, according to a controlled trial, basiline immunosuppression consisted in CsA, AZA, and steroid versus Tac and steroid. AZA and prednisone were tapered and stopped 6 and 3 months respectively after OLT.

Tac was used as rescue therapy, and as primary treatment, in conjunction with steroids, since 2001. Acute rejection was treated with 10 mg/kg/day of intravenous methyl prednisolone pulses for 3 days.

Long-term follow-up after hospital discharge was scheduled as follows: twice weekly from the third week to the end of the first month, twice monthly until the sixth month, then monthly until the end of the first year; then every 3-6 months. Unexpected signs or symptoms prompted a complete work-up. The work-up consisted of complete clinical evaluation, liver function tests (LFT), and other laboratory tests (serum proteins, blotting tests, kidney function tests, blood cell count, venous blood pH, electrolytes), EBV [Epstein-Barr virus]-DNA polymerase chain reaction, cytomegalovirus pp65 antigen, hepatitis B surface antigen, HCVab, CsA (or Tac) trough blood level, abdominal ultrasounds, and Doppler ultrasound of liver blood vessels. In all patients the level of circulating EBV-DNA copies was routinely assessed. The onset of EBV-related clinical manifestations (mononucleosis-like picture, posttransplant lymphoproliferative disease, unexplained alteration of LFT, mainly elevation of alanine aminotransferase) associated with an increase in the blood EBV-DNA level was promptly treated by reduction of immunosuppression, until EBV-DNA levels were under 500 copies/100,000 mononuclear cells. Since 1999 anti-CD20 monoclonal antibody was also used in case of oligoclonal or monoclonal posttransplant lymphoproliferative disease. Gammaglobulins and autoimmune markers—antinuclear factor (ANA), anti–smooth muscle antibodies (SMA), anti–liver-kidney microsomal antibodies (LKM), and anti-neutrophil cytoplasmic antibodies (ANCA)—were systematically tested since January 2000. Autoantibodies level was performed by indirect immunofluorescence, using commercially available Kits (INOVA, San Diego, California); ANA were tested on HEP 2 cells; SMA and LKM on kidney, stomach and liver mouse tissues, and ANCA on ethanol-fixed and formalin-fixed granulocytes. All tests were performed at the initial dilution of 1:20 and titers ≥1:40 were regarded as positive.

In all cases of persistent GD, biliary, vascular, and infectious complications were ruled, and liver biopsy was performed in all cases. DNAH was treated by prednisone, 2 mg/kg/day, progressively tapered to a minimal dose (2.5-5 mg/day) and azathioprine (AZA), 1-2 mg/kg/day. Calcineurin inhibitors (CNIs) were decreased to reduce trough levels of 20%.

Patients diagnosed with early chronic rejection (ECR) were treated by increasing dosage of CNIs to obtain trough levels of 30% higher than the standard level (or switching from CsA to Tac). If normalization of LFT was not obtained, AZA and a steroid taper were added, and levels of CNIs were restored to standard. The same schedule was used as a first-line treatment in the case of a recent or concomitant EBV infection.


We studied 247 children who underwent OLTx. Sixty out of 247 children developed autoantibodies. Thirty-eight out of 60 patients developed autoantibodies without GD. Liver biopsy, obtained in 3 of them showed only mild aspecific histological changes (1 case) and nearly normal liver (2 cases). Nine out of 247 children who underwent OLTx (3.6%) fulfilled the criteria for DNAH and 13 (5.2 %) for ECR. Median follow-up was 39.3 and 35.8 months, respectively. Median age at transplantation was 2.7 years (range, 0.4-19.1 years). The native liver disease was biliary atresia (18), Alagille syndrome (2), progressive familial intrahepatic cholestasis type 1 (1), and alpha 1 antitrypsin deficiency (1). None were affected by primary autoimmune liver disease.

De Novo Autoimmune Hepatitis Group

In this group, both CsA (5 patients) and Tac (4 patients) levels were within the standard range. Average onset of graft dysfunction was at 34.8 months after OLTx (±20.8); median alanine aminotransferase level was 173 UI/L (range, 95-779), median value 44 UI/L (range, 21-100), and median IgG level was 1,970 mg/dL (range, 1,020-3,010).

LKM antibodies were present in 2 patients, ANCA in 2, and SMA in none. In 3 patients ANA antibodies were detected, and in 2 patients, ANA was associated with SMA and ANCA. Global data are summarized in Table 1. None of the patients had documented CMV or EBV infection in a 1-year period before GD onset. Previous biliary obstruction occurred in 1 patient, while 1 had a vascular complication in the past. Previous acute rejection occurred in 5 patients.

Table 1. Serological Data on Patients with DNAH
PatientALT (UI/L)IgG (mL/dL)Auto Ab Titers
  1. Abbreviations: ALT, alanine aminotransferase; Auto Ab, autoantibody; ANA, antinuclear antibody; LKM, anti-liver-kidney microscomal antibody; ANCA, anti-neutrophil cytoplasmic antibody.

11071,970ANA 1:40, SMA 1:80
22582,450pANCA 1:40
37791,300LKM 1:40
41201,600ANA 1:640
51883,010ANA 1:1280
61732,570LKM 1:1280
7951,200ANA 1:320, pANCA 1:80
81001,020ANA 1:1280
91822,000cANCA 1:40

All the patients had normalization of LFT on steroids and AZA at a median of 2.5 months (range, 1-14.1) after the onset of treatment. Therapy initially consisted of steroids in only 3 patients; this regimen determined a normalization in 1, while the other 2 patients required AZA shortly after. One patient initially thought acute rejection was unsuccessfully treated with pulse intravenous steroids and switched from CsA to Tac. Another patient was switched from CsA to Tac. Both of them had normalized LFT only when the treatment of autoimmune hepatitis with steroids and AZA was started.

After treatment 4 patients were negative for immune markers (1 LKM, 1 ANA/SMA and 2 ANCA), 5 had stable or decreasing titers. IgG levels were back to normal values at a median time of 4.9 months (range, 1-26). One patient experienced relapse of disease after withdrawal of the specific treatment: He normalized again when AZA and steroids were restarted.

Early Chronic Rejection Group

Eleven out of 13 patients experiencing ECR were on CsA and 2 on Tac. Immunosuppressive drug levels were adequate in 8 cases, while 5 patients had a low immunosuppressive trough level due to a concurrent or recent EBV infection.

GD occurred at an average of 45.7 months after liver transplantation (±39.3); median alanine aminotransferase and levels at diagnosis were respectively 177 UI/mL (range, 60-560) and 34 UI/L (range, 10-150); median IgG value was 1,380 mg/dL (range, 603-2120). ANA and SMA positivity was documented in 5 and 3 patients respectively; ANA positivity was associated to SMA positivity in 4 patients and to ANCA positivity in 1 patient, while LKM antibodies were absent in all the patients. Laboratory findings are summarized in Table 2.

Table 2. Serological Data on Patients With ECR
pt dataALT (U/L)IgG (mg/dL)Auto Ab Titers (UI/L)
  1. Abbreviations: ALT, alanine aminotransferase; Auto Ab, autoantibody; ANA, antinuclear antibody; SMA, smooth muscle antibody.

1931,370SMA 1:160
2560603SMA 1:80
32421,380ANA 1:160, SMA 1:80
41201,800ANA 1:80, SMA 1:40
51611,750ANA 1:640, SMA 1:80
61301,310ANA 1:80
71771,140ANA 1:160, SMA 1:80
83701,870ANA 1:320
9601,450ANA 1:40, pANCA 1:40
103161,410ANA 1:1280
114181,350ANA 1:640
124911,230ANA 1:640
131052,120SMA 1:40

Recent or previous EBV infection was documented in 5 patients, while none had cytomegalovirus infection. Previous biliary and portal vein thrombosis occurred in 2 and 3 children, respectively. Three patients had previous biopsy proven acute rejection.

Eight patients, including 7 children switched from CsA to Tac, initially responded to an increase of CNIs only, but 3/8 required AZA and oral steroids to obtain normal LFT without overdosage of CNIs (overdosage was considered >30% of the standard trough level). Five children (4 on CsA and 1 on Tac) were treated by steroids and AZA because of a diagnosis of ECR due to concurrent severe EBV infection or posttransplant lymphoproliferative disease. In 3 cases autoantibodies disappeared. Normalization of LFT was obtained in 9 patients at a median time of 7.3 months (range, 0.7-27.6) after the beginning of treatment; in 4 cases we observed that LFT improved without normalization. Two out of these 4 patients developed ductopenic.

The type of CNIs (CsA/Tac) treatment, LFT values and IgG levels were not significantly different between patients with DNAH and ECR (P > 0.05).

ANA and/or SMA positivity occurred significantly more frequently in the ECR group (P = 0.01). The incidence of EBV infection was significantly higher in the ECR group compared to the DNAH group (P = 0.03). Cumulative data concerning treatment and outcome of each group are reported in Table 3.

Table 3. Patient Treatment and Outcome Summary
 DNAH (n = 9 patients)ECR (n = 13 patients)
  1. Abbreviations:PTLD, post transplant lymphoproliferative disease; DNAH, de novo autoimmune hepatitis; ECR, early chronic rejection; EBV, Epstein-Barr virus; AZA, azathioprine; CsA, cyclosporine; Tac, tacrolimus.

Treatment↓CsA/Tac + steroid/AZA (9/9)↑CNIs, including 7 switch CsA>Tac (8/13)
  Introduction of steroid/AZA (3/8)
  CNIs + steroid/AZA (EBV, PTLD) (5/13)
OutcomeBiochemical recovery (9/9)Biochemical recovery (9/13)
  Biochemical improvement (4/13)
  Ductopenic chronic rejection (2/13)


There is increasing evidence that late GD may occur in forms other than ductopenic rejection. One of these patterns is DNAH, characterized by lobular/interface hepatitis, high levels of transaminases and IgGs, autoantibodies, and good response to steroids.1–3 A variety of potential mechanisms can lead to de novo autoimmunity after liver transplantation. In addition to the immune thymus downregulation due to long-term immunosuppression, a possible mechanism is molecular mimicry, determining the release of cross-reactive autoantibodies. Again, development of auto(allo)immunity may be a consequence of allogenic transplantation, and a close relationship between rejection and autoimmunity has been demonstrated by animal studies.7 However the primum movens of autoreactive autoantibodies in patients whose native liver disease was not autoimmune is still unclear.

On the other hand, GD may present with a mainly centrilobular localized process with centrilobular necrotic damage in conjunction with central vein endothelitis, a feature known as early chronic rejection.4 An association between early chronic rejection and autoantibodies was previously reported.6, 8

Both DNAH and ECR seem to be a pattern of late GD associated with autoantibodies after pediatric OLTx. In our series, 8.9% of all transplanted children had a late GD and autoantibodies positivity at the same time. Patients were divided into 2 groups (DNAH and ECR) on the basis of the histological findings: 9 DNAH and 13 ECR.

We did not find any difference in LFT pattern between DNAH and ECR. There was no difference in CNIs treatment (CsA or Tac). The IgG level was higher in the DNAH group than in the ECR. ANA and/or SMA positivity was more frequent in the ECR group. LKM were present only in the DNAH group. The incidence of recent biliary and vascular complications was similar in 2 groups. EBV infection (recent or ongoing) was higher in the ECR group.

It must be pointed out that, by performing statistical analysis of the parameters in the study, we could not show any significant difference between the two groups, except for autoantibodies profile and EBV infection. This is possibly related to the short number of patients of the population.

As previously reported,9 our patients with DNAH also did not respond to pulse intravenous steroids (1 patient) nor to switch from CsA to Tac (1 patient) but require the standard treatment for classic autoimmune hepatitis (steroids and azathioprine) with excellent results in all cases. In 3 patients the treatment initially consisted only of steroids; in all but 1 patient LFT normalization was obtained only when AZA was added. Alanine aminotransferase levels were back to normal at a median time of 2.5 months (range, 1-14.1 months) after starting treatment. Relapse occurred in 1 case when therapy was withdrawn; this child's LFT normalized after therapy was restarted.

Overall, 9/13 ECR patients (69%) had a complete response of LTF to an increased immunosuppression, 4/13 had an improvement of LFT without normalization, and 2/13 experienced ductopenic rejection. In 5 cases the mere elevation of CNIs was effective.

In 6 cases, biochemical recovery was achieved with an association of CNIs and AZA/steroid instituted at the time of diagnosis in 3 cases (EBV infection/posttransplant lymphoproliferative disease) and started after obtaining no satisfactory response to mere CNIs upload in the other 3 cases.

Several authors reported a poor outcome of late graft dysfunction despite treatment; in our series only 2 out of 13 ECR patients experienced ductopenic chronic rejection.

Based on our own experience, the best therapy for DNAH seems to be prompt treatment with AZA/steroid and without any increase of CNIs. A sustained immunosuppression by an increase of CNIs is mandatory for ECR. In case of an EBV infection or posttransplant lymphoproliferative disease (which we consider as a contraindication for CNIs' high blood trough levels) or in case of poor response, association with AZA/steroid has proved to be effective in improving or normalizing LFT.

It remains still unclear if ductopenic rejection includes a hepatic-like phase before a centrilobular phase, both associated with autoantibodies, preceding duct loss, or if the 2 are completely independent clinicopathological manifestations.

The good outcome after DNHA treatment suggests that this condition could represent a very early, and possibly reversible, stage of chronic rejection. ECR might be a second and less reversible step of damage characterized by a higher immunologic severity.

We should be aware that these provisional indications arise from a reduced cohort of liver transplanted children, but it is difficult to find any larger groups in which patients are routinely tested for autoantibodies after surgery. We are now increasing the number of followed patients and more extensive clinical and morphological findings will soon be available, in particular concerning cohort of children with new onset of antibodies without GD.

If further data can confirm our findings or a possible modification of liver damage in DNAH and ECR after treatment, an early differential diagnosis would become crucial to realizing a tailored treatment and to avoid graft loss. 1, 2

Figure 1.

Chronic hepatitis with piecemeal necrosis and normal-looking central vein.

Figure 2.

Central vein with hepatocyte confluent dropout and focal venulitis.