fax: (44) 121 3338251
Liver Failure and Liver Disease
Progressive histological damage in liver allografts following pediatric liver transplantation†
Article first published online: 20 APR 2006
Copyright © 2006 American Association for the Study of Liver Diseases
Volume 43, Issue 5, pages 1109–1117, May 2006
How to Cite
Evans, H. M., Kelly, D. A., McKiernan, P. J. and Hübscher, S. (2006), Progressive histological damage in liver allografts following pediatric liver transplantation. Hepatology, 43: 1109–1117. doi: 10.1002/hep.21152
Potential conflict of interest: Nothing to report.
- Issue published online: 20 APR 2006
- Article first published online: 20 APR 2006
- Manuscript Accepted: 9 FEB 2006
- Manuscript Received: 7 OCT 2005
The long-term histological outcome after pediatric liver transplantation (OLT) is not yet fully understood. De novo autoimmune hepatitis, consisting of histological chronic hepatitis associated with autoantibody formation and allograft dysfunction, is increasingly recognized as an important complication of liver transplantation, particularly in the pediatric population. In this study, 158 asymptomatic children with 5-year graft survival underwent protocol liver biopsies (113, 135, and 64 at 1, 5, and 10 years after OLT, respectively). Histological changes were correlated with clinical, biochemical, and serological findings. All patients received cyclosporine A as primary immunosuppression with withdrawal of corticosteroids at 3 months post OLT. Normal or near-normal histology was reported in 77 of 113 (68%), 61 of 135 (45%), and 20 of 64 (31%) at 1, 5, and 10 years, respectively. The commonest histological abnormality was chronic hepatitis (CH), the incidence of which increased with time [25/113 (22%), 58/135 (43%), and 41/64 (64%) at 1, 5, and 10 years, respectively) (P < .0001)]. The incidence of fibrosis associated with CH increased with time [13/25 (52%), 47/58 (81%), and 37/41 (91%) at 1, 5, and 10 years, respectively) (P < .0001)]. The severity of fibrosis associated with CH also increased with time, such that by 10 years 15% had progressed to cirrhosis. Aspartate aminotransferase (AST) levels were slightly elevated in children with CH (median levels 52 IU/L, 63 IU/L, and 48 IU/L at 1, 5, and 10 years, respectively), but this did not reach statistical significance compared with those with normal histology. On multivariate analysis, the only factor predictive of chronic hepatitis was autoantibody positivity (present in 13% and 10% of children with normal biopsies at 5 and 10 years, respectively, and 72% and 80% of those with CH at 5 and 10 years, respectively) (P < .0001). Four children with CH and autoantibodies, who also had raised immunoglobulin G (IgG) levels and AST greater than 1.5× normal fulfilled the diagnostic criteria for de novo autoimmune hepatitis (AIH). Another two were found to be hepatitis C positive. No definite cause for CH could be identified in the other cases. In conclusion, chronic hepatitis is a common finding in children after liver transplantation and is associated with a high risk of developing progressive fibrosis, leading to cirrhosis. Standard liver biochemical tests cannot be relied on either in the diagnosis or in the monitoring of progress of chronic allograft hepatitis. In contrast, the presence of autoantibodies is strongly associated with the presence of CH. The cause of chronic hepatitis in transplanted allografts is uncertain but may be immune mediated, representing a hepatitic form of chronic rejection. (HEPATOLOGY 2006;43:1109–1117.)
Liver transplantation (OLT) provides life-saving treatment for children with end-stage liver disease. Current long-term survival rates are over 80%.1, 2 However, few data are available regarding the long-term histological outcome of transplanted allografts in children. Studies in adults have revealed histological abnormalities in allografts but have focused largely on patients with liver dysfunction or late graft loss.3–6 Indications for liver transplantation in adults and children are different, adults often undergoing transplantation for conditions that may recur post-transplantation. In children, most liver diseases are non-recurrent and potentially curable by transplantation. An important question within the field of pediatric liver transplantation is whether children who have undergone successful transplantation can expect a normal life expectancy or whether there will be a gradual decline in liver function and eventual graft loss.
Recent studies have revealed a phenomenon referred to as de novo autoimmune hepatitis, post-transplantation immune hepatitis, or graft hepatitis mimicking autoimmune hepatitis.7–11 This syndrome is characterized by biochemical, serological, and histological features indistinguishable from autoimmune hepatitis (AIH) occurring in patients transplanted for conditions other than autoimmune disorders. The incidence in these studies ranged from 2.3% to 5.2% of paediatric patients. A similar syndrome has been described more rarely in adults.7, 12 The cause is not understood, but it has been postulated to be a form of low-grade chronic rejection and may be related to molecular mimicry.13 The increased incidence in children may relate to disruption of normal T cell maturation secondary to immunosuppressive therapy. In adults, the most common underlying liver diseases predisposing to de novo AIH are primary biliary cirrhosis and primary sclerosing cholangitis (PSC). In protocol biopsies after OLT for the former, recurrent disease and de novo AIH may be seen together.14 The incidence of chronic allograft hepatitis with or without associated autoantibody formation in asymptomatic children with normal liver function is not known.
Whereas some centers advocate serial liver biopsies after transplantation to assess histological changes,5, 15 others have not found protocol liver biopsies to be helpful in guiding clinical management and have abandoned their use.16 The aim of this study was to prospectively assess histological changes of allografts over the first 10 post-transplantation years using protocol liver biopsies at 1, 5, and 10 years post OLT. The study is unique in reporting data from a large cohort of asymptomatic children who underwent protocol liver biopsy.
Patients and Methods
Between 1983 and 1996, 210 children (96 boys, 114 girls) underwent 247 liver transplantations at our Unit. Of these, 158 (80 boys, 78 girls) were alive 5 years after OLT with graft survival of 5 years with their first graft. The mean age at transplantation was 4.7 years (range, 0.0-15.9 years; median, 2.9 years). Indications for transplantation are shown in Table 1. Of the 158 children, 69 (44%) received whole liver allografts and 89 (56%) received reduced size organs—81 as the result of cut-down and eight from split procedures.
|Indication for Transplantation||Number||%|
|Extrahepatic biliary atresia||77||48.4|
|Fulminant hepatic failure||23||15.1|
|Hepatitis A infection||3||1.9|
|Non-A, non-B, non-C acute hepatitis||15||9.5|
|Tyrosinemia type 1||5||3.1|
|Cystic fibrosis liver disease||6||3.8|
Since the inception of the liver transplant program, we performed protocol reviews of all patients at approximately 1, 5, and 10 years after OLT. The review process comprised clinical review, biochemical and serological screening, abdominal ultrasonography, and liver biopsy.
Standard liver function tests (LFTs) were performed at each review along with immunoglobulins G, A, and M and autoantibodies at 5 and 10 years. Autoantibodies included anti-nuclear antibody (ANA), anti-smooth muscle antibody (SMA), anti-mitochondrial antibody, and anti–liver-kidney microsomal antibody (LKM). All were reported as positive if present at titers of 1 in 25 or higher.
Serology for cytomegalovirus and Epstein-Barr virus was performed at all time points and for hepatitis B, C, and G viruses if biopsies showed features of chronic hepatitis.
At 1, 5, and 10 years post OLT, all children underwent percutaneous liver biopsy, and biopsies were assessed by the same histopathologist (S.G.H.), using a system previously reported.17, 18 Diagnostic categories included normal, mild non-specific changes, chronic hepatitis (CH), acute rejection, chronic rejection, biliary obstruction, recurrent disease, or other histological abnormality. Chronic hepatitis was characterized by predominantly portal-based mononuclear inflammation associated with interface hepatitis. Varying degrees of lobular inflammation with hepatocyte necrosis or apoptosis were also frequently present. A diagnosis of chronic hepatitis also required that bile duct lesions or vascular changes characteristic of acute or chronic rejection were minimal or absent. The severity of interface hepatitis and lobular necro-inflammation were each graded semi-quantitatively on a scale of 0 to 3 (0 = none, 1 = mild, 2 = moderate, 3 = severe). Chronic hepatitis cases were then assigned an overall inflammatory grade of mild, moderate, or severe based on the severity of periportal or lobular inflammatory activity.17 Representative examples of 2 cases classified as chronic hepatitis with mild and moderate activity are illustrated in Fig. 1A -C. Fibrosis staging was carried out on a 3-point scale: 0 = no fibrosis, 1 = periportal fibrosis without bridging, 2 = bridging fibrosis, 3 = cirrhosis. Examples of cases showing mild fibrosis (stage 1) and cirrhosis (stage 3) are shown in Fig. 1D-E. A diagnosis of mild non-specific changes was made when there was mild portal or lobular inflammation without interface activity or hepatocyte necrosis or mild periportal fibrosis without bridging (stage 1).
During this time period, the standard immunosuppressive regimen was cyclosporine A (Sandimmun until 1997, then Neoral thereafter, both Novartis Pharmaceuticals, Basel, Switzerland) with azathioprine and prednisone. Cyclosporine was given to achieve trough levels of 60 to 90 μg/L after the first post-transplantation year. Azathioprine was given at a dose of 1.5 mg/kg/day and was discontinued at 12 months post OLT. Prednisone was commenced initially at 2 mg/kg/day, gradually reduced, and then discontinued at 3 months post OLT. Steroids were continued for longer in patients with acute rejection, with the aim of discontinuing by 6 months post OLT.
Statistical analysis was performed using the chi-squared test to demonstrate differences in frequencies of observations between the three time points. Logistic regression was used to investigate factors associated with histological abnormalities at 5 years post-transplantation. This time point was chosen as more biopsies were performed than at 1 and 10 years. The demographic variables studied included age at transplantation, underlying liver disease, donor age, donor and recipient sex, donor and recipient blood group, donor and recipient cytomegalovirus status, allograft cold ischemic time, type of allograft (whole, reduced size, or split allograft). Dynamic variables included alanine aminotransferase (ALT), aspartate aminotransferase (AST), immunoglobulin levels, and autoantibody positivity at the time of liver biopsy.
One hundred thirteen of 158 children (72%) had biopsies at a mean of 14.1 months (range, 9.5-17.4 months; median, 13.5 months). Twenty-six children did not have biopsies at this stage because they were transplanted between 1995 and 1996, when the procedure had been discontinued. Nineteen individuals declined biopsies because they were asymptomatic with normal LFTs at their 1-year review.
There were no graft or patient losses between 1 and 5 years post-transplantation. One hundred thirty-five children (85%) underwent liver biopsy at a mean of 62.3 months post OLT (range, 49.1-86.5 months; median, 63.5 months). Ten patients declined a liver biopsy at this stage, seven had been transferred to adult care, three were living overseas, and in three biopsy was contraindicated (previous hemobilia after liver biopsy in one, renal failure requiring dialysis in one, and neurological problems following a cerebrovascular event in one).
Between 5 and 10 years post OLT, there were 11 graft losses. Seven patients died (bacterial sepsis in two, subarachnoid hemorrhage in one, and recurrent disease without re-transplantation in four), and four underwent retransplantation (chronic rejection in one, biliary obstruction in two, and de novo autoimmune hepatitis in one). Of the remaining 147 patients, 81 (55%) underwent transplantation 10 or more years ago and 64 (79%) underwent biopsy at a mean of 116.6 months (range, 101.9-131.5 months; median, 117.1 months). Of those who did not undergo a biopsy, 12 had been transferred to adult care, one lived overseas, three declined a biopsy, and one biopsy was again contraindicated due to previous hemobilia. The configuration of patients at the 3 points was statistically similar and did not suggest a drop-out bias between 1 and 10 years.
A few complications occurred after liver biopsy in these patients (hemobilia in 1, pneumothorax in 1, and abdominal pain after biopsy in 3 patients, all of whom required an overnight stay in hospital but no specific treatment).
The biopsy findings at each of the three periods are shown in Table 2. A decrease occurred between 1 and 5 years and again between 5 and 10 years in the proportion of biopsies considered to be normal or near-normal (66% at 1 year, 41% at 5 years, and 27% at 10 years; chi-squared = 31.4, df = 2; P < .0001). However, even in these “normal” biopsies, frequently mild inflammatory changes, mainly portal, were seen, which were considered insufficient to make a diagnosis of chronic hepatitis (60% at 1 year, 34% at 5 years, and 23% at 10 years).
|1 Year (n = 113)||5 Years (n = 135)||10 Years (n = 64)|
|Histology||Number (%)||Number (%)||Number (%)|
|Normal/near normal||77 (68.2)||61 (45.2)||20 (31.3)|
|Normal||7 (6.2)||10 (7.4)||1 (1.6)|
|Mild nonspecific changes||68 (60.2)||46 (34.1)||15 (23.4)|
|Mild fibrosis||2 (1.8)||5 (3.7)||4 (6.3)|
|Chronic hepatitis||25 (22.1)||58 (43.0)||41 (64.0)|
|Rejection||3 (2.7)||3 (2.2)||0|
|Acute||1 (0.9)||1 (0.7)||0|
|Chronic||2 (1.8)||2 (1.3)||0|
|Biliary obstruction||7 (6.2)||10 (7.4)||1 (1.6)|
|Recurrent disease||1* (0.9)||1* (0.7)||1* (1.6)|
|Other||0||2† (1.3)||1‡ (1.6)|
The commonest histological abnormality at all times was chronic allograft hepatitis (Fig. 1). This increased sequentially over time: 22% at 1 year, 43% at 5 years, and 64% at 10 years. This increase was statistically significant (chi-squared = 36.21, df = 2; P < .0001). The grade of interface hepatitis, lobular necro-inflammation, and the overall severity of chronic hepatitis are summarized in Table 3. Portal inflammation with interface hepatitis, usually mild, was present in 117 of 124 cases (94%). Lobular inflammatory changes, again usually mild, were present in 63 of 124 cases (51%). In seven cases (6%), the diagnosis of CH was based on lobular necro-inflammatory changes occurring in the absence of significant periportal inflammatory activity. Overall inflammatory activity was graded as mild in 59% of cases, moderate in 33%, and severe in 8%. Although not statistically significant, there was a tendency for the severity of inflammation to increase with time (Fig. 2).
|Interface hepatitis||7 (6%)||94 (76%)||20 (16%)||3 (2%)|
|Lobular inflammation||61 (49%)||41 (33%)||18 (15%)||4(3%)|
|Overall grade||0||73 (59%)||41 (33%)||10 (8%)|
Other histological diagnoses included biliary obstruction, rejection, and one case of recurrent sclerosing cholangitis. These were uncommon, accounting for no more than 11% of cases at any time, and did not increase significantly over time.
Forty-eight children had biopsies at all three times. Of these, only one child had chronic hepatitis which appeared to resolve. The child had mild chronic hepatitis without fibrosis at 1 year but at both 5 and 10 years had mild non-specific changes only with mild (grade 1) fibrosis, which could have resulted from sampling error.
Periportal fibrosis was occasionally observed as an isolated finding—2 of 77 (3%), 5 of 66 (8%), and 4 of 20 (20%) cases at 1, 5, and 10 years, respectively—and in all cases was mild (stage1) (Fig. 3). Of these, fibrosis was persistent, with the cases demonstrating fibrosis at 1 year having fibrosis again on biopsy at 5 and 10 years. One child (above) had chronic hepatitis at 1 year, but biopsies at 5 and 10 years showed mild periportal fibrosis with no significant inflammation. The incidence of fibrosis was high in biopsies with chronic hepatitis and increased between 1 and 10 years—13 of 25 (52%), 47 of 58 (81%), 37 of 41 (91%) at 1, 5, and 10 years, respectively; chi-squared = 36.71, df = 2; P < .0001). Furthermore, the severity of fibrosis in individuals with chronic hepatitis also increased with time. None demonstrated stage 3 fibrosis (consistent with cirrhosis) at 1 year; however, 7% had stage 3 fibrosis at 5 years and 15% at 10 years (Fig. 4). For the 48 cases who had biopsies at all three times, no correlation was seen between the grade of necro-inflammatory activity (interface hepatitis, lobular inflammation, or overall severity) and the severity of fibrosis. Likewise, the grade of necroinflammatory activity at 1 year was not predictive for the subsequent development of fibrosis at 5 or 10 years. Not surprisingly, children with biliary obstruction had high rates of fibrosis, which appeared to progress with time (5/7 at 1 year—stage 1 = 4, stage 2 = 1; 10/10 at 5 years—stage 1 = 3, stage 2 = 4, stage 3 = 3; and 1/1 at 10 years—stage 2 =1).
Of the LFTs, AST (normal range, 10-50 IU/L) and ALT (normal range, 10-45 IU/L) were available for all patients at 5 years. Although median AST and ALT were higher in children with chronic hepatitis compared with those with normal biopsies (AST, 63 IU/L compared with 44 IU/L and ALT, 57 IU/L compared with 39 IU/L), these differences were not statistically significant. There was no significant correlation between the grade of necro-inflammatory activity (individual components or overall severity) and AST or ALT levels. However, AST and ALT were significantly higher in those with other histological abnormalities such as biliary obstruction and rejection at 1 year post-transplantation (P < .005). At later times, the number of these patients was too small to compare with the other groups (Table 4).
|Histology||1 Year||5 Years||10 Years|
|Median AST (Range)||Median AST (Range)||Median AST (Range)|
|Normal||43 (8–136)||44 (16–155)||38 (21–105)|
|Chronic hepatitis||52 (17–110)||63 (22–103)||48 (25–128)|
|Other histology||81 (24–138)||63 (28–248)||65 (16–142)|
Autoantibody data were not available for most patients at 1 year post-transplantation, because testing was not routine in the early years of the transplant program. At 5 and 10 years, autoantibody positivity was significantly higher in children with chronic hepatitis than in children with normal or near-normal biopsies [42/58 (72%) compared with 8/61 (13%) at 5 years; 33/41 (80%) compared with 2/20 (10%) at 10 years (chi-squared = 78.13, df = 2; P < .0001)]. Children with chronic hepatitis and positive autoantibodies were more likely to have high-titer ANA (>1 in 100), SMA, or LKM autoantibodies or a combination of ANA and SMA or LKM autoantibodies than children with normal biopsies. Indeed, no child with a normal biopsy had ANA of greater than 1 in 40, SMA, or LKM autoantibodies (Table 5). There was no significant correlation between the grade of necro-inflammatory activity (individual components or overall severity) and the presence of autoantibodies. Elevated immunoglobulin levels were not detected in most children with chronic allograft hepatitis.
|Histology||5 Years||10 Years|
|Autoantibody Positivity (No./Total No. of Biopsies)||Autoantibody Positivity (No./Total No. of Biopsies)|
|Normal/near normal||8/61 (13%)||2/20 (10%)|
|ANA 1 in 25 = 4||ANA 1 in 25 = 1|
|ANA 1 in 40 = 4||ANA 1 in 40 = 1|
|Chronic hepatitis||42/58 (72%)||33/41 (80%)|
|ANA 1 in 25 = 8||ANA 1 in 25 = 5|
|ANA 1 in 40 = 7||ANA 1 in 40 = 11|
|ANA 1 in 100 = 10||ANA 1 in 100 = 8|
|ANA 1 in 400 = 9||ANA 1 in 400 = 3|
|ANA 1 in 1600 = 4||ANA 1 in 1600 = 3|
|SMA = 12||SMA = 8|
|LKM = 2||LKM = 2|
|Mixed Abs = 10||Mixed Abs = 7|
A non-immune cause for chronic hepatitis was identified in only two patients, both testing positive at 1 year for hepatitis C virus. Their inflammatory changes persisted with time, although one died of an unrelated cause between 5 and 10 years. No other viral causes of chronic hepatitis were identified in the remaining children.
Only one child at 5 years and an additional three at 10 years demonstrated chronic hepatitis in association with liver dysfunction (aminotransferase levels elevated more than 1.5 times the upper limit of normal), autoantibody positivity, and raised immunoglobulins, thus fulfilling the diagnostic criteria of de novo autoimmune hepatitis. The child with de novo autoimmune hepatitis at 5 years post-transplantation had progressive allograft failure, requiring re-transplantation 4 years later. Two children with chronic hepatitis and positive autoantibodies at 10 years post-transplantation also had seronegative pauci-articular arthritis.
Using univariate logistic regression, factors associated with chronic hepatitis on liver biopsy included autoantibody positivity, sex mismatch between donor and recipient, and the levels of AST and ALT. However, using multivariate logistic regression analysis, autoantibody positivity was the only factor predictive of chronic allograft hepatitis on liver biopsy at 5 years post-transplantation.
The current study is unique in describing histological abnormalities in children undergoing protocol liver biopsies up to 10 years after liver transplantation. There was an unexpectedly high rate of histological abnormalities at all three time points (32%, 55%, and 79% at 1, 5, and 10 years, respectively). The commonest histological diagnosis was chronic hepatitis, the prevalence of which increased with time, being present in 22%, 43%, and 64% of cases at 1, 5, and 10 years, respectively. Furthermore, of biopsies classified as normal for statistical purposes, many had minor non-specific inflammatory changes, the significance of which is uncertain. The findings of biliary obstruction and recurrent disease (one case of recurrent PSC) on biopsy were predictable in that the children had current or previously abnormal LFTs, previous diagnostic biopsies, or abnormal ultrasound findings. This was not the case for those with chronic hepatitis, most of whom had normal LFTs.
A strong association was found between the presence of chronic hepatitis and the development of progressive fibrosis. The prevalence of fibrosis at 1 year in the current study (23%) is similar to that reported by Peeters et al.,19 who observed portal fibrosis in 31% of transplanted children within 1 year post-transplantation, mainly related to graft ischemia and biliary complications. In the current study, children with biliary obstruction also exhibited a high rate of fibrosis. However, the finding of increasing fibrosis in children with chronic hepatitis has not been reported before and has potentially important implications for long-term graft function. More than half of the children with chronic hepatitis at 1 year had some fibrosis, albeit mild, whereas by 10 years 50% of cases had bridging fibrosis or cirrhosis. A small number of children had mild periportal fibrosis as an isolated finding without any other histological abnormality. Fibrosis in these cases may be related to a previous pathological process such as cholangitis or chronic hepatitis, which has subsequently resolved.
Previous studies of late post-transplantation biopsies, mainly carried out in adults, suggest that most cases (70%-90%) will eventually develop histological abnormalities.3–6, 18, 20, 21 Many are seen in protocol biopsies obtained from people who are clinically well with good graft function. Many late post-transplantation biopsies have features of chronic hepatitis.3–5, 22 In most cases these can be attributed to recurrent disease, particularly recurrent viral infection (hepatitis B virus or hepatitis C virus). Other possible causes of chronic hepatitis in the liver allograft include acquired viral infection (hepatitis B or hepatitis C), recurrent autoimmune hepatitis, or biliary disease (primary biliary cirrhosis, sclerosing cholangitis), de novo autoimmune hepatitis, or drug toxicity.23, 24 In a varying proportion of cases, no obvious cause for chronic hepatitis can be identified.3–5, 22 In the study of Sebagh et al,5 this accounted for only 1 of 78 (1.3%) cases. By contrast, no obvious etiological factor could be identified in 6 of 29 (32%) chronic hepatitis cases reported by Pappo et al.3 and 13 of 27 (48%) of those documented by Slapak et al.4
In contrast to the adult population undergoing liver transplantation, where most of the common indications are potentially recurrent diseases, most pediatric liver allograft recipients are not at risk of developing recurrent disease. In the current study, only eight children were transplanted for potentially recurrent diseases (3 AIH, 5 PSC). One child developed recurrent PSC. No evidence of disease recurrence was seen in the three children transplanted for AIH. The high prevalence of chronic hepatitis seen in the current study thus cannot be attributed to recurrent disease.
Two previous studies documented features of chronic hepatitis in pediatric patients after liver transplantation.16, 25 Rosenthal et al.16 described 54 asymptomatic children with normal LFTs who underwent annual biopsy up to 3 years post-transplantation. Portal and parenchymal inflammation was common (48% and 25%, respectively) but not fibrosis, which occurred in 8% only. In a previous study from our liver unit, chronic hepatitis occurred in 24% of children after liver transplantation, a similar prevalence to that seen in adults.25 No viral or other cause for chronic hepatitis was identified. The previous study had a shorter follow-up (median, 4 years), and there was no assessment of autoantibody status or fibrosis progression.
The most important factor associated with chronic hepatitis was the presence of autoantibodies, noted in 72% and 80% of cases at 5 and 10 years, respectively, compared with 13% and 10% of cases with normal or near-normal histology. Furthermore, antibody titers in chronic hepatitis cases were often present in high titer and included smooth muscle antibodies and LKM antibodies as well as ANA. In approximately 25% of chronic hepatitis cases, two or more autoantibodies were present. However, only four children with chronic hepatitis and autoantibodies (one at 5 years; three at 10 years) had other features supporting a diagnosis of de novo AIH (AST > 1.5 × upper limit of normal, raised immunoglobulins). It will be important to follow these children to determine whether they develop the full clinical scenario of de novo autoimmune hepatitis. Nevertheless, these observations raise further questions regarding the pathogenesis, diagnostic criteria, prognosis, and clinical management of de novo AIH in the liver allograft.
Transient de novo autoantibody production after liver transplantation has been observed in association with episodes of rejection,26,27 suggesting that de novo autoimmune hepatitis may represent an atypical form of late cellular rejection. Other studies have identified acute rejection as a risk factor for subsequent development of de novo AIH.28, 29 Furthermore, late cellular rejection may be associated with histological changes that are different from those typically seen in early acute rejection and more closely resemble those of chronic viral or autoimmune hepatitis.3, 30 Some cases of otherwise unexplained chronic hepatitis in the liver allograft may represent a form of late cellular rejection.23
The calcineurin inhibitor drugs, cyclosporine and tacrolimus, interfere with normal T-cell maturation and the function of regulatory T-cells.31 The observation that de novo autoimmune hepatitis may be more common in childhood raises the possibility that thymic dysfunction arising from cyclosporine use may be a factor.32 The children in this study were transplanted before 1996, and all received cyclosporine-based primary immunosuppression. Since 2000, we have routinely used tacrolimus as the primary immunosuppressive agent. We do not yet know the incidence of chronic hepatitis in patients receiving tacrolimus from the time of transplantation. However, the choice of calcineurin inhibitor after OLT for autoimmune liver diseases may influence the development of recurrent disease, with tacrolimus being associated with more rapid onset of recurrent autoimmune disease (primary biliary cirrhosis) than cyclosporine in one study.33
Most of the children in this study received cyclosporine as monotherapy from 1 year post-transplantation, and steroids were usually withdrawn at 3 months. This protocol differs from the regimens used in some other pediatric transplant programs in two respects: steroid withdrawal was usual and occurred early, and target cyclosporine levels were in the lower range. Despite this, the incidence of acute rejection in our program34 was similar to35–37 or less than38, 39 that reported by others. Others have reported physical and metabolic benefits from steroid withdrawal, without any increase in acute rejection,40, 41 albeit undertaken at a later stage after transplantation than in our series. Chronic hepatitis may represent a form of chronic rejection related to under-immunosuppression that might have been prevented by continued steroid therapy. As a result, we have changed our protocol and now continue prednisone at low dose indefinitely. Furthermore, steroids have been recommenced in those children who had discontinued steroids and subsequently developed chronic hepatitis. The approach to immunosuppression in children with normal liver function but chronic hepatitis on biopsy with or without autoantibody formation has yet to be elucidated, and prospective studies are clearly indicated.
Non-adherence with immunosuppressive therapy is a potential problem after liver transplantation, especially in adolescents. The histological findings we have observed were documented in children of all ages, and therefore we do not believe that non-adherence is likely to have had a major etiological role.
The results of the current study indicate that important histological abnormalities can be seen in protocol biopsies obtained from children who are clinically well with good graft function and normal liver biochemistry. Of particular concern is that the median AST and ALT levels were not statistically different between those with normal histology and those with chronic hepatitis. Screening for chronic allograft hepatitis using liver biochemistry is therefore not possible and may instead require regular measurement of autoantibodies.
These reported findings are from a single center and retrospective. It is important they be reproduced in other centers, to exclude center- or era-specific phenomena. Other than early steroid discontinuation, our overall experience does not differ significantly that of from other pediatric transplant programs of similar vintage. The indications for transplantation, patient and graft survival, re-transplantation rate, incidence of acute rejection, and viral infections and use of innovative surgical approaches are similar in our center to contemporaneous reports.34, 35, 37, 39, 40, 42
The authors thank Carla Lloyd for assistance in data collection and Dr. Paul Davies for statistical advice.