Rejection and Steroid Dependence: Unique Risk Factors in the Development of Pediatric Posttransplant De Novo Autoimmune Hepatitis


*Corresponding author: Robert S. Venick,


Posttransplant de novo autoimmune hepatitis (d-AIH) is increasingly described as a long-term complication after pediatric liver transplantation (LT). d-AIH is characterized by graft dysfunction, the development of autoimmune antibodies and histologic evidence of hepatitis in liver transplant recipients without previous history of autoimmune liver disease. This study is a matched case-control, univariate analysis aimed at identifying risk factors for the development of d-AIH and evaluating response to treatment. From 1984 to 2003, 619 children received 788 LTs at a single center. Forty-one patients developed d-AIH and were matched with controls for year of LT, age at time of LT and diagnosis. The following variables were insignificant in the development of d-AIH: age, gender, race, initial diagnosis, ischemia time, graft type, Epstein-Barr virus and cytomegalovirus status, HLA typing and primary immunosuppression. Compared to controls, d-AIH patients were less likely to be on monotherapy immunosuppression or weaned off prednisone at the time of diagnosis. The d-AIH group relative to the controls had statistically significant greater numbers of rejection episodes. d-AIH was treated with prednisone and/or MMF in 39 of 41 patients and lead to significant improvements in liver function tests. Thirty-nine patients are alive at a mean of 4.0 years follow-up after diagnosis. Three have required retransplantation.


Autoimmune hepatitis is a progressive inflammatory liver disorder characterized serologically by the presence of autoantibodies and hypergammaglobulinemia, and histologically by a dense mononuclear cell infiltrate of the portal tracts with interface hepatitis (1,2). Posttransplant de novo autoimmune hepatitis (d-AIH) was initially described in 1998 in liver transplant (LT) recipients without a prior history of autoimmune liver disease (3). The diagnostic criteria for d-AIH are shown in Figure 1. This disorder affects between 2.1% and 5.2% of pediatric LT patients (3–12), and between 0.4% and 3.4% of adult recipients (13–16). The specific long-term treatment and outcome of d-AIH are not well established (4,15).

Figure 1.

Criteria for the diagnosis of de novo autoimmune hepatitis.

Posttransplant d-AIH differs from recurrence of autoimmune hepatitis in that d-AIH occurs in patients who were transplanted for reasons other than autoimmune hepatitis. The presence of autoantibodies with or without evidence of liver dysfunction has been consistently reported following LT in up to 15–41% of pediatric patients (3,9,17–19), and in 11–71% of adult recipients (20–22). Therefore, the diagnosis of d-AIH is made by fulfilling the criteria in Figure 1 and excluding other causes of late graft dysfunction. However, the overlap in histologic features of atypical acute rejection, chronic rejection and nonspecific hepatitis, frequently noted in biopsies from patients with late graft dysfunction, can make the pathologic diagnosis challenging for experienced pathologists.

Previous reports on d-AIH have been limited to small case series so that the important clinical characteristics associated with d-AIH or the outcomes of this entity are not yet well understood. The aim of this Institutional Review Board approved, retrospective matched case-control study was to determine significant risk factors for the development of d-AIH, and to evaluate the response to treatment.


Between February 1984 and May 2003, 788 pediatric orthotopic liver transplants were performed in 619 children less than 18 years of age at the University of California Los Angeles Medical Center. From 1984 to 1987 induction and maintenance therapy consisted of cyclosporine (CsA) and prednisone. Azathioprine (AzA) was also added to this induction regimen between 1987 and 1994 in an attempt to lower steroid and CsA doses. Specific maintenance requirements during this era included CsA dosed according to therapeutic trough levels (goal of 200–250 ng/mL level measured by high-performance liquid chromatography of whole blood for the first 3 months following LT, and 100–150 ng/mL by year one post-LT in patients without episodes of rejection), AzA (1 mg/kg/day) and prednisone (0.3 mg/kg/day by postoperative day [POD] 6, gradually tapered to 0.15–0.2 mg/kg/day by 1 year post-LT) (23). For children transplanted after 1994, primary immunosuppression consisted of tacrolimus (Tac) and prednisone. Tac was administered orally at an initial dose 0.05–0.1 mg/kg/day and was titrated according to trough concentrations (goal of 8–12 ng/mL whole blood microparticle enzyme immunoassay during the first month post-LT and 5–8 ng/mL thereafter). After 1995, prednisone was tapered to 0.3 mg/kg/day by POD 6, and was weaned off as tolerated within 1 year post-LT. Long-term management including the rate of steroid weaning and the dosing of calcineurin inhibitor was individualized based on clinical circumstances. Acute rejection episodes during all time periods were treated with high-dose intravenous methylprednisolone boluses tapered over 7 days. OKT3 was reserved for steroid resistant rejection, and after 1994 conversion from CsA to Tac was also an option in such cases. Mycophenolate mofetil (MMF) was also used in cases of steroid resistant rejection and retransplantation.

Following LT, patients were regularly evaluated every 1–3 months in clinic with routine laboratory studies including alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase, total and conjugated bilirubin. Patients who were found to have graft dysfunction, manifest by elevated liver function tests (LFTs), underwent percutaneous liver biopsies which were interpreted by an experienced transplant hepatopathologist. In the event that the liver biopsy had an appearance of a portal-tract infiltrate and periportal hepatitis involving lymphocytes and plasma cells, we then pursued an extensive laboratory workup. This included antidouble stranded DNA, antinuclear antibody, liver-kidney microsomal antibody, smooth muscle antibody, perinuclear antineutrophil cytoplasmic antibody, antimitochondrial antibody, hepatitis A antibody, hepatitis B surface antigen, surface antibody and core antibody, hepatitis C antibody or hepatitis C quantitative polymerase chain reaction (PCR, once it became available at our center in the late 1990s), Epstein-Barr virus (EBV) PCR and cytomegalovirus (CMV) DNA quantitation. Vascular insufficiency and biliary strictures were excluded via ultrasound, CT scan, MRI/MRA and cholangiogram when clinically indicated.

Patients identified with d-AIH were matched with controls for the following variables: age at transplant ± 12 months, year of transplant ± 12 months and primary diagnosis (cholestatic liver disease, fulminant hepatic failure or metabolic liver disease). In instances in which more than one potential match was found, the control was chosen randomly so that the authors were blinded with regard to patient selection. In 39 of 41 cases, controls were found which met all three of the above criteria. In two instances of children with metabolic liver disease, appropriate age and year of transplant matches were found, but a perfect match could not be found for primary diagnosis. Therefore, one child was matched with a control with cholestatic liver disease, and the other with a child with fulminant hepatic failure. The medical records of the case and control patients were reviewed retrospectively, and the following variables were recorded: age, gender, race, ischemia time, graft type, EBV and CMV status of donors and recipients, primary diagnosis, primary immunosuppressant, immunosuppressant levels, time at which steroids were weaned, HLA-typing of the donor and recipient, as well as number and timing of rejection episodes. For the case patients the following were also collected: treatment of d-AIH, LFTs (at diagnosis of d-AIH, 3, 6, 12 and 24 months after treatment), patient and graft survival.

Statistical analysis was performed using JMP (SAS Corp, Cary, NC) and StatXact (Cytel Inc., Cambridge, MA). Proportions of categorical covariates for the case and control groups were compared using exact chi-square methods (Fisher's exact test), while mean and median values for continuous variables were compared using 2-tailed paired Student's t-tests or the nonparametric Wilcoxon rank sum test, respectively, when the distribution was not unimodel symmetric. For the case patients, repeated measure analysis of variance (ANOVA) was used to analyze the geometric means of LFTs and immunosuppression levels over time. While multivariate logistic regression and classification tree (CART) analyses were attempted in order to evaluate all the factors simultaneously, the relatively small number of patients compared to the number of covariates made the results inconclusive. Therefore multivariate results are not reported.


Of the 619 children who received liver transplants during this time frame, 50 patients were identified who were transplanted for a reason other than a primary autoimmune disease, and who developed posttransplant graft dysfunction as well as an elevation of one of the autoimmune markers listed above. The biopsy reports of these 50 patients were reviewed retrospectively. Nine patients were excluded because their biopsies or laboratories revealed other causes of graft dysfunction or did not support the diagnosis of d-AIH. Forty-one of 619 patients were ultimately identified as having d-AIH (incidence = 6.6%). Thirty-three of these 41 patients (80%) developed d-AIH after 1999. The clinical characteristics and indications for LT of the cases and controls are shown in Table 1. No statistically significant demographic, underlying diagnostic, surgical or virological differences were noted between the two groups.

Table 1.  Baseline clinical characteristics of case and control patients
Variable Gender FemaleCase n = 26Control n = 23p-Value 0.651
  1. 1Fisher's exact = 0.46, df = 1.

  2. 2Two-tailed Student's paired t-test.

  3. 3Chi-square = 6.51, df = 4.

  4. 4Chi-square = 0.56, df = 2.

  5. 5Fisher's exact = 2.33, df = 1.

  6. 6Fisher's exact = 0.12, df = 1.

Mean age at transplant (years ± 95% CI)3.6 ± 0.93.5 ± 1.00.612
 African Americann = 9n = 5 
 Asiann = 2n = 3 
 Caucasiann = 7n = 16 
 Latinon = 22n = 17 
 Native Americann = 1n = 00.173
Primary diagnosis
 Cholestatic liver diseasen = 23n = 240.754
 Fulminant hepatic failuren = 13n = 14 
 Metabolic liver diseasen = 5n = 3 
Surgical variables 0.262
 Mean cold ischemia time (h)8.66 ± 1.447.66 ± 1.280.102
 Mean warm ischemia time (h)0.83 ± 0.100.69 ± 0.13 
Graft type 0.342
 Wholen = 26n = 30 
EBV and CMV status
 EBV high risk (D+/R− or recipient < 1 y/o at LT)n = 19n = 280.205
 CMV high risk (D+/R− or recipient < 1 y/o at LT)n = 10n = 80.666

The mean time from LT to the development of d-AIH was 7.0 ± 1.2 years. At the time of diagnosis there was a notable increase from baseline in the case patients' LFTs, particularly in the transaminases (Table 2). The autoimmune markers found to be positive at the time of diagnosis in the d-AIH patients are shown in Figure 2. According to our laboratory's criteria many patients had strongly positive markers, and 18 of 41 patients (44%) had multiple positive autoantibodies.

Table 2.  Characteristics of cases and control patients at diagnosis of d-AIH or corresponding time interval
VariableMean ± 95% confidence intervals of casesMean ± 95% confidence intervals of controlsp-Values of two-tailed t-tests
  1. N/A = not applicable.

Time from transplant to development of d-AIH7.0 ± 1.2 yearsN/AN/A
LFTs at diagnosis of d-AIH (cases) or corresponding time interval from LT (controls)ALT 317 ± 119ALT 32 ± 11p = 0.00004,
AST 291 ± 136AST 40 ± 12p = 0.0009,
T Bilirubin 1.5 ± 0.7T Bilirubin 0.8 ± 0.4p = 0.09,
D Bilirubin 0.7 ± 0.4D Bilirubin 0.3 ± 0.3p = 0.18,
Alk Phos 318 ± 46Alk Phos 239 ± 33p = 0.007,
Figure 2.

Autoimmune markers of d-AIH patients at diagnosis.

A statistically significant difference was not noted in the type of initial primary immunosuppression used in the case and control groups (p = 0.40, Table 3). However, within the first 2 years post-LT only 8% of the patients with d-AIH, versus 29% of the controls could be weaned off of prednisone (Table 4, p = 0.0094). Significant differences were also noted in the immunosuppressants used at the time of diagnosis of d-AIH (cases) and corresponding matched time interval (control) (p = 0.030, Table 5). Specific differences at the time of diagnosis revealed the following: fewer of the cases (46%) versus controls (81%) were off of prednisone (chi-square = 10.3, df = 1 p = 0.0013), fewer of the cases (29%) versus controls (63%) were on monotherapy CsA or Tac (chi-square = 9.6, df = 1, p = 0.0019), and a trend was noted that most of the cases (39%) versus controls (24%) required MMF or AzA, both medications used in the treatment of autoimmune hepatitis, as maintenance medications (chi-square = 2.0, df = 1, p = 0.15). Additionally, there was a statistically significant difference in the mean Tac levels at diagnosis of the d-AIH patients (7.1 ± 1.2 ng/mL) relative to the controls (5.3 ± 1.0 ng/mL) (Table 6, p = 0.03).

Table 3.  Initial primary immunosuppressant following liver transplant
1° ImmunosuppressantCaseControlp-Value
  1. 1Chi-square = 2.90, df = 3.

  2. AzA = azathioprine; CsA = cyclosporine; MMF = mycophenolate mofetil; Tac = tacrolimus.

CsA and prednisone53 
CsA, prednisone and AzA17170.401
Tac and prednisone1721 
Tac, prednisone and MMF20 
Table 4.  Prednisone requirements 2 years post liver transplant
  1. 1Chi-square = 6.74, df = 1.

Percent of patients off of prednisone by 2 years post-LT8%29%0.00941
Table 5.  Immunosuppressant at diagnosis
Immunosuppressant at the time of d-AIH or time interval from LTCasesControlsp-Value of chi-square test
CsA and AzA76 
CsA, AzA and prednisone72 
Tac and MMF01p = 0.03
Tac and prednisone115 
Tac, prednisone and sirolimus20 
Tac, prednisone and MMF21 
Table 6.  Immunosuppressant levels at diagnosis
VariableMean ± 95% CI of casesMean ± 95% CI of controlsp-Values of 2-tailed t-tests
Immunosuppressant levels at diagnosis of d-AIH or corresponding time interval from LTTac Lvl (n = 26) 7.1 ± 1.2Tac Lvl (n = 25) 5.3 ± 1.0p = 0.03,
CsA Lvl (n = 15)102 ± 41CsA Lvl (n = 16) 73 ± 16p = 0.22

Additional variables which were analyzed and not found to have significant differences between cases and controls were the HLA typing and mismatching of donors and recipients (data not shown).

The number of total (cases: 2.8 ± 0.7 vs. controls: 1.0 ± 0.4, p = 0.0002), acute (2.0 ± 0.5 vs. 0.7 ± 0.2, p = 0.0001), chronic (0.6 ± 0.3 vs. 0.3 ± 0.2, p = 0.05) and steroid resistant rejection (0.2 ± 0.1 vs. 0.02 ± 0.05, p = 0.05) episodes were significantly different between the case and control groups (Table 7).

Table 7.  Rejection post-LT in the case and control groups
  1. 1p-Values are based on 2-tailed Student's paired t-test.

Mean number of total biopsy proven rejection episodes2.8 ± 0.71.0 ± 0.40.0002
Mean number of acute steroid-sensitive rejection episodes2.0 ± 0.50.7 ± 0.20.0001
Mean number of chronic rejection episodes0.6 ± 0.30.3 ± 0.20.05
Mean number of steroid resistant rejection episodes0.2 ± 0.10.02 ± 0.050.05

In terms of the treatment of d-AIH, 21 patients were treated with a combination of MMF (25 mg/kg/day) and prednisone (1–2 mg/kg/day with a maximum of 40 mg, gradually tapered to 5–10 mg/day over 6–8 weeks), 8 with increased doses of prednisone alone, 7 with MMF alone, 3 with MMF and a change of CsA to Tac and 2 who were already on triple therapy with no change. Of note CsA and Tac dosage and levels stayed the same after the diagnosis of d-AIH as can be seen in Figure 3A. The mean LFTs at diagnosis compared by repeated measures ANOVA to those at 3, 6, 12 and 24 months after the diagnosis of d-AIH significantly decreased with treatment (Figure 3B).

Figure 3.

Figure 3.

(A) FK and CsA Levels at diagnosis of d-AIH and during treatment. The geometric means were compared using repeated measures ANOVA methods on the log scale. For Tac Levels: p-values of 0 versus 3 months = 0.51, 0 versus 6 months = 0.85, 0 versus 12 months = 0.73, 0 versus 24 months = 0.92. For CsA Levels: p-values of 0 versus 3 months = 0.14, 0 versus 6 months = 0.25, 0 versus 12 months = 0.70, 0 versus 24 months = 0.46. (B) Response to treatment after diagnosis of d-AIH.

Figure 3.

Figure 3.

(A) FK and CsA Levels at diagnosis of d-AIH and during treatment. The geometric means were compared using repeated measures ANOVA methods on the log scale. For Tac Levels: p-values of 0 versus 3 months = 0.51, 0 versus 6 months = 0.85, 0 versus 12 months = 0.73, 0 versus 24 months = 0.92. For CsA Levels: p-values of 0 versus 3 months = 0.14, 0 versus 6 months = 0.25, 0 versus 12 months = 0.70, 0 versus 24 months = 0.46. (B) Response to treatment after diagnosis of d-AIH.

Thirty-nine of 41 patients (95%) are currently alive at a mean of 4.0 ± 0.4 years after the diagnosis of d-AIH. One patient died of complications related to end-stage liver disease at 1.1 years after d-AIH diagnosis, the other of sepsis at 2.5 years after diagnosis with normal graft function. Additionally three patients have been retransplanted at 0.1, 1.8 and 2.4 years after diagnosis of d-AIH due to end-stage liver disease associated with d-AIH. Twenty-three of 39 surviving patients are on low-dose daily (n = 16) or every other day prednisone (n = 7) with a mean dose of 8.0 ± 2.7 mg/day or 0.2 ± 0.09 mg/kg/day. There were no cases of posttransplant lymphoproliferative disease attributable to the increase in immunosuppression among the case patients, nor was there a statistically significant difference between the case and control patients in the overall long-term patient (log-rank test, chi-square = 0.31, p = 0.58) or graft survival (log-rank test, chi-square = 0.45, p = 0.50).


This study is the largest and only case-control study to date on d-AIH. The handful of previous pediatric reports listed in Tables 8 and 9 have lacked sufficient size and study design to uncover important associations with the development of d-AIH (3–21). Our intent was to determine risk factors for the development of d-AIH. We found that prior to the diagnosis of d-AIH patients had more episodes of rejection, an increased dependence on steroids and overall greater immunosuppression requirements than their matched controls. This study also offers a long-term view of the clinical course of d-AIH and confirms that the majority of children can be treated with MMF and low-dose prednisone.

Table 8.  Previous pediatric reports of d-AIH
Author/location/yearNo. of patients w/d-AIH/No. of patients followed% Patients with d-AIHMedian age Dx (years)Median time to d-AIH (years)Female/Male ratio
Table 9.  Previous pediatric reports of d-AIH
AuthorTreatment of d-AIHMedian Follow-upLFTs trendFollow-up pathologyOutcomes
  1. 1Typical prednisone regimen = 1–2 mg/kg/day gradually tapered to low dose; typical AzA regimen = 1–2 mg/kg/day.

  2. 2M-H regimen = solumedrol (10–20 mg/kg/day) × 3 days followed by prednisolone 1 mg/kg/day gradually tapered.

  3. AzA = azathioprine; CsA = cyclosporine; Tac = tacrolimus.

KerkarPrednisolone + AzA1283 daysNormalized (median 32 days)2 patients w/improved necroinflamatory activityN/A
AndriesPrednisolone + AzA1N/AAll normalized w/in 3 months9/10 improved w/ treatmentN/A
HernandezPrednisone + AzA1+ CsA dose ↓ed or discontinuedN/AAll normalized w/in 3 months1 normalized, 2 improved, 2 acute rjxnN/A
GuptaPrednisolone + AzA1N/AImproved4/6 developed bridging fibrosis2 re-LT; 1 death
SpadaPrednisolone + AzA1N/AInitially improved, increased during taperN/AN/A
Miyagawa-HayashinoSteroids2+ increased Tac dosing3.5 yearsImprovedHistological remission in 1/11 patients with follow-up3 re-LT

Our overall incidence of d-AIH of 6.6%, was similar to the range previously reported at other pediatric LT centers. The median duration between transplant and the development of d-AIH of 6.7 years illustrates the important message that this process should be included in the differential diagnosis of children who develop late graft dysfunction. Only 12% of patients developed d-AIH in their first 2 years post-LT; therefore, clinicians should have the highest index of suspicion for d-AIH in the late post-LT period. While a female preponderance for developing d-AIH has been reported in other series and in our study (female/male ratio of 1.7:1), this prevalence is not as strong as in autoimmune hepatitis in which up to 75% of children are female (2). No racial predisposition was seen in our study for patients with d-AIH versus controls. Given that the indication for LT was a variable for which we controlled the study was not well equipped to identify specific primary diseases which may be predisposed to developing d-AIH. Nonetheless, the primary diseases and their frequencies in the patients who developed d-AIH in our series were relatively representative of the indications for LT at our center, other studies on pediatric d-AIH (24) and of world-wide indications for pediatric LT (25). The only caveat was the percentage of cases (34%) in our series who were transplanted for fulminant hepatic failure was higher than the percentage of all children transplanted for fulminant hepatic failure, and warrants further investigation. It is of interest that patients transplanted for fulminant hepatic failure have been described as having a higher incidence of acute rejection than their controls (84% vs. 56%) (26), because certainly rejection in this study is a risk factor associated with the development of d-AIH.

Nearly two-thirds of the patients in our study had positive ANA and anti-dsDNA titers at the time of diagnosis, and in many the titers were strongly positive. Given the previously reported high prevalence of autoantibodies in patients post-LT (3,9,17–22,27,28), it is important that the diagnosis of d-AIH is strictly reserved for those patients with positive titers, biopsy evidence of autoimmune hepatitis with exclusion of other potential causes of graft dysfunction. Nine patients were excluded from our study for precisely these reasons. Ideally in a prospective trial, autoimmune antibodies would be routinely monitored on all patients pre- and post-LT to aid in the screening and early diagnosis of d-AIH.

While hypergammaglobulinemia has been used by other centers as a criterion in diagnosing d-AIH, we have not routinely measured immunoglobulin levels in our pediatric LT recipients. In some series patients with liver disease have been described as having higher IgG levels secondary to portosystemic shunting (29,30). Alternatively, others have described up to 26% of post-LT recipients as being hypogammaglobulinemic (31). Our experience with immunoglobulin levels post-LT has mirrored this widely reported variation, and has not given us significant additional information than the autoantibodies have provided.

Our study offered a large sample size to study the relationship between HLA typing of donors and recipients and the development of d-AIH. However, partly due to missing data, especially among the early transplant patients, we were unable to demonstrate a meaningful association. It is clear that type 1 autoimmune hepatitis, displays an immunogenetic association with the HLA DR3 (DRB1*0301) and DR4 (DRB1*0401) haplotypes (32,33). Whether or not the same holds true for d-AIH remains unclear (3,8,14,15).

Interestingly, 80% of our patients have developed d-AIH after 1999. It is not clear if the incidence of d-AIH is truly increasing, or rather if this reflects an increased awareness, screening and ability to diagnose d-AIH. Initially we suspected that this trend might be associated with changes in our immunosuppression protocol including changing from CsA to Tac in 1994, more aggressive steroid weaning and stopping the routine use of AzA. However, in univariate analysis there was not a significant difference in the development of d-AIH and primary immunosuppressant (Table 3, p = 0.40). Previous studies have shown variability in the type of primary immunosuppression and the development of d-AIH (3–8). One of the weaknesses of a retrospective study such as ours is that while it may demonstrate associations, it does not lend itself well toward evaluating the role that one medication versus another may play in the development of d-AIH. To get at the heart of this question, and to better determine causality of d-AIH would require randomization and prospective monitoring.

A history of steroid dependence, and response to steroids and AzA have been mentioned as characteristic features of d-AIH patients (4). Our study confirms that steroid dependence is an important defining characteristic of patients who develop d-AIH (Table 4). Furthermore, case patients were more difficult to maintain on monotherapy Tac or CsA, required higher levels of Tac to maintain normal graft function and tended to be on AzA or MMF at the time of diagnosis. While these findings may be related to the increased incidence of rejection in the case patients, there also may be an underlying up-regulation of the immune system of d-AIH patients which makes them dependent on additional immunosuppression in order to maintain normal graft function.

Another hypothesis which we had prior to our retrospective review was that this increasing phenomenon of d-AIH may have been related to the emphasis which was placed on the weaning of steroids in pediatrics in the late 1990s (23). However, as mentioned above, the role which steroids and their wean play in d-AIH development is difficult to isolate given that steroid use is confounded by rejection and perhaps by underlying characteristics of individual patients' immune systems. In the series by Gupta et al. all patients were on low-dose monotherapy calcineurin inhibitors at the time of d-AIH diagnosis, which the authors speculate may have triggered d-AIH (6). On the contrary, Andries et al. reported that prior to the diagnosis of d-AIH 10 of 11 patients required daily prednisone to keep their LFTs within normal limits (4). Such variability highlights differences in management of immunosuppression from center to center. Future studies may be well-served by trying to examine cumulative doses of steroids prior to the development of d-AIH. Additionally, larger series would permit meaningful multivariable analysis which could separate the role which steroid use and rejection may play in the pathogenesis of d-AIH.

The small sample size and retrospective nature of previous studies in this field make the role which rejection may play in the development of d-AIH somewhat controversial. Hernandez et al. only noted one out of their five patients with d-AIH to have experienced a previous episode of rejection (5). On the contrary, the Kyoto group found acute rejection to be a significant predictor in the development of d-AIH (8). Additional studies have shown a strong association between chronic rejection and the development of antitissue antibodies (33). Our study revealed a statistically significant difference in the number of total, acute, chronic and steroid resistant rejection episodes among the case and control groups (Table 7). Again the retrospective nature of our study only allows us to conclude that a higher incidence of rejection is associated with a predisposition to developing d-AIH, yet the causality which rejection may or may not play remains unclear and will require prospective monitoring.

Standard treatment for pediatric autoimmune hepatitis and posttransplant d-AIH, typically entails a prednisone taper followed by maintenance with either MMF or AzA (3–7,9,11). Thirty-nine of 41 patients in our study were treated with a combination of MMF and/or prednisone. CsA and Tac doses of our patients were not adjusted once d-AIH was diagnosed, as there does not seem to be convincing evidence that this will alter clinical outcome. Prior authors have advocated decreasing CsA or Tac doses once d-AIH was diagnosed; however, they have reported subsequent episodes of rejection, histologic relapse of d-AIH and the need for retransplantation (5,9). On the contrary, Miyagawa-Hayashino increased the dose of Tac in addition to instituting steroids, yet 3 of 13 of their patients required re-LT and only one patient in their series showed histologic evidence of remission. Other potential treatments have been used as adjunctive therapy in autoimmune hepatitis including ursodeoxycholic acid, budesonide and methotrexate (6,34); these medications have not yet been studied in d-AIH.

Awareness that successful treatment of d-AIH is possible with steroids plus AzA or MMF has led to excellent graft and patient survival. Once therapy for d-AIH was started, LFTs decreased rapidly in all of our patients, and remained significantly reduced at 3 months follow-up. This normalization is well described in other pediatric studies and typically occurs within weeks of treatment (Table 9). While few of the pediatric d-AIH studies provide long-term data, our patients sustained their improvement in LFTs at 6, 12 and 24 months follow-up. At a mean of 4.0 ± 0.4 years after the diagnosis of d-AIH, 41% of our patients were off of prednisone, with the remainder requiring low-dose daily or every other day steroids in order to maintain normal LFTs. We advocate maintenance on calcinuerin inhibitors and MMF with low-dose steroids used only as needed to maintain LFTs in an acceptable range. On this regimen, the majority of d-AIH patients do not require any closer long-term follow-up than their age-matched controls. A weakness of our series is that we do not routinely perform protocol biopsies on our patients when their LFTs are within normal range. Therefore, the interval and long-term histopathologic course of d-AIH among our patients is not well understood. A review of the literature shows significant variability among centers (Table 9). Prospective histopathologic monitoring of patients undergoing treatment for d-AIH will be important in future studies.

The mechanism leading to d-AIH remains unclear. Prior hypotheses have included immunosuppression-induced autoimmunity (35–38), molecular mimicry (17,39), autoantibodies directed against glutathione-S-transferase T1 (16,40) and a preponderance of acute cellular rejection to lead to the development of d-AIH (17,41–43). Many of these theories have only been examined in animal models, and clearly additional work is needed to understand mechanisms involved in d-AIH.


Posttransplant d-AIH should be included in the differential diagnosis of pediatric LT patients without previous autoimmune liver disease who develop late graft dysfunction. In this center's experience the incidence of d-AIH is 6.6%. While age, gender, initial diagnosis, HLA typing, surgical variables, viral status and primary immunosuppression do not appear to be significant risk factors for the development of d-AIH, the number of episodes of rejection and steroid dependence are strongly associated with d-AIH. Fortunately early diagnosis and treatment of d-AIH with low-dose steroids and MMF allows for good patient and graft survival. The pathogenesis of d-AIH remains poorly understood, and requires prospective monitoring.


Dr. Venick's work on this study was supported by a grant from the Glaser Pediatric Research Network, a nonprofit organization dedicated to pediatric clinical research and children's health, and by the National Institute of Health-Loan Repayment Program.