Evaluation of graft histology is an essential component of the diagnostic assessment and therapeutic monitoring of the liver transplant recipient in the early posttransplant period. The threshold for performing a biopsy is low when liver test results suggest graft dysfunction during the first year following liver transplantation because this is a common time to diagnose both rejection and infectious complications. Conversely, the role of protocol or surveillance biopsies in patients with normal graft function remains controversial. Analyses of protocol liver biopsies performed during long-term follow-up of predominantly adult liver transplant recipients have revealed abnormal histology in 36% to 88%,1–3 which has been attributed frequently to viral hepatitis and disease recurrence. Little information is available regarding histological abnormalities in pediatric liver transplant recipients. A recent study4 of a cohort of pediatric liver transplant recipients who underwent protocol liver biopsies at 1, 5, and 10 years post liver transplantation reported chronic hepatitis as being the most common histological abnormality, increasing in incidence with time from transplantation. Additionally, the prevalence and severity of fibrosis associated with chronic hepatitis increased with time from transplantation.4
One possible explanation for the scarcity of long-term histological studies in pediatric patients may be that asymptomatic children are not routinely subjected to liver biopsy in most transplant centers. However, maintenance of long-term graft function is important to all liver transplant recipients and perhaps most important in recipients who may have a longer posttransplant course ahead. Because it can be assumed that graft histology is associated with graft survival and therefore with patient survival,5 it seems reasonable for children to receive the same type of histological monitoring as adult recipients.
Considering the reports of high rates of histological abnormalities in long-term follow-up, our center adopted recommendations for surveillance liver biopsy evaluation in all patients surviving more than 5 years after transplantation in early 2003. The aims of this study were to describe the graft histology in the long-term follow-up of pediatric liver allograft recipients and to analyze factors correlating with long-term histological outcome.
This analysis included a cross-sectional sample of long-term survivors of pediatric liver transplantation who received liver biopsies between November 1, 2003 and October 31, 2005 for 1 of 2 indications: (1) the patient was ≥3 years post transplant and had elevated alanine aminotransferase (ALT) and aspartate aminotransferase (AST) ± elevated gamma-glutamyl transferase (GGT), or (2) the patient was ≥5 years post transplant and had either elevated ALT and AST ± GGT or had normal ALT, AST, and GGT but agreed to a surveillance biopsy to monitor for occult graft injury. However, the patient and family were ultimately responsible for the decision to follow through with these recommendations. This study, which retrospectively evaluated the results of these clinically indicated biopsies, was approved by the Institutional Review Board of Children's Memorial Hospital.
During the study period, 222 patients followed in our institution's organ transplant tracking record database had survived ≥3 years post liver transplantation. Of these patients, 68 had liver biopsies performed, and histological slides were available for review in 63 patients. The indications for liver biopsy were as follows: evaluation of elevated ALT and AST ± GGT (n = 24), follow-up of de novo autoimmune hepatitis with fluctuating ALT and AST ± GGT (n = 8), and surveillance biopsies for ≥5-year survivors (n = 31). There were 154 patients who had survived ≥3 years but who did not have a liver biopsy during this time period. Fifty had moved out of state and/or had transferred aspects of their care to other institutions; 17 had survived between 3 and 5 years post transplant and had normal ALT, AST, and GGT; and 87 had survived more than 5 years and had normal ALT, AST, and GGT but elected not to return for surveillance biopsy.
The indications for liver transplantation in the 63 patients with slides for review were autoimmune diseases (n = 4), biliary atresia (n = 31), genetic cholestatic diseases (n = 5), metabolic diseases (n = 6), fulminant liver failure (n = 7), and miscellaneous (included malignancies, congenital hepatic fibrosis, choledochal cyst, and cystic fibrosis; n = 10). Recipient data were retrieved from the medical records and organ transplant tracking record database. Data included history and radiological evidence of late hepatic artery thrombosis, de novo autoimmune hepatitis, donor age, cold and warm ischemia times. These data were selected because they had previously been associated with fibrosis progression following liver transplantation and graft survival following liver transplantation.5–7
During this time period, our standard immunosuppressive regimen was cyclosporine, azathioprine, and prednisone, with azathioprine being discontinued at 6 months post transplant and prednisone being discontinued at 1 year post transplant. Tacrolimus was substituted for cyclosporine in our adolescent patients older than 12 years at the time of transplant and in those recipients who either had chronic rejection or had developed acute rejection while on cyclosporine. Within our study population, 28 patients were on cyclosporine, 33 were on tacrolimus, and 2 were on sirolimus. Additionally, 13 patients were also on azathioprine, 25 were on prednisone, and 6 were on mycophenolate mofetil. The reasons for receiving prednisone included a remote history of late acute cellular rejection, a history of de novo autoimmune hepatitis, and treatment of respiratory flare of cystic fibrosis. The reasons for receiving azathioprine or mycophenolate included a history of de novo autoimmune hepatitis and a remote history of late acute cellular rejection. The reason for receiving sirolimus included a history of de novo autoimmune hepatitis.
All patients underwent Doppler ultrasound evaluation of the liver, biliary tree, and hepatic vasculature prior to liver biopsy. Patients in whom Doppler ultrasound was inconclusive underwent computed tomography angiogram of the hepatic vasculature.
Liver tissue was fixed in 10% phosphate buffered formalin and stained with hematoxylin/eosin and Masson's trichrome technique and for Cytokeratin 19 (monoclonal mouse anti-human cytokeratin 19 antibody, Dako North America, Inc., Carpinteria, CA). Slides were reviewed by a pathologist (H.M.-A) blinded to the clinical history. Fibrosis was scored according to the Metavir system: F0, no fibrosis; F1, mild portal fibrosis with no septa; F2, portal fibrosis with few septa; F3, portal fibrosis with numerous septa without cirrhosis; and F4, cirrhosis.8 Chronic allograft rejection was defined according to the Banff criteria9 (late chronic rejection: ductopenia exceeding 50%; early chronic rejection: ductopenia exceeding 20% but less than 50%). Inflammation was graded as follows: 0, no inflammation; 1, mild portal inflammation, no interface hepatitis; 2, mild/moderate portal inflammation, interface hepatitis in some but not all portal tracts; 3, moderate portal inflammation, interface hepatitis in the majority of portal tracts; and 4, severe portal inflammation with bridging necrosis.10
Viral Studies and Autoantibodies
All patients whose biopsies showed portal inflammation with interface hepatitis had hepatitis C polymerase chain reaction performed and antinuclear, anti-smooth muscle, and anti–liver-kidney-microsomal antibodies measured by indirect fluorescent assay. The normal ranges for the autoantibodies are as follows: <1:80, <1:80, and <1:40 respectively. We defined de novo autoimmune hepatitis as the presence of graft dysfunction, portal inflammation with limiting plate disruption and lobular hepatitis, and positive autoantibodies in the absence of evidence of hepatitis C infection.11
Means and standard deviations were described for continuous variables, and frequencies were described for categorical variables. Univariate analysis of factors associated with fibrosis included donor age, biopsy indication, cold and warm ischemia times, interval from transplant to biopsy, and history of de novo autoimmune hepatitis. The interval from transplant to biopsy was dichotomized at 6 years on the basis of the sample distribution. Chi-square tests were used to determine the association between categorical predictors and ≥stage 3 fibrosis, except when cell sizes were less than 5; then, Fisher's exact test was used. The Wilcoxon test was used to test the association of continuous predictors with ≥stage 3 fibrosis. A post hoc power analysis (alpha = 0.05, power = 80%) was performed to determine the number of patients with normal and elevated ALT, AST, and GGT needed to detect a difference in significant fibrosis (≥stage 3) between the 2 groups. Statistical analyses were conducted with SAS version 9.2 statistical software, and conclusions were made at a 0.05 level of significance.
As shown in Table 1, there were no significant differences in gender, pretransplant diagnosis, and age at transplantation between the 63 patients included in the study sample and the 154 who did not receive a biopsy. The patients that did not have a biopsy were significantly older and had a longer interval from transplant.
Table 1. Baseline Characteristics of Biopsied and Nonbiopsied Patients
Patients with Biopsy and Available Slides [n (%)]
Patients with No Biopsy [n (%)]
Abbreviation: LT, liver transplantation.
Number of patients
Mean age at LT (years)
3.28 ± 4.32
3.64 ± 4.39
Male gender (n)
Mean interval from LT (years)
7.99 ± 3.40
9.22 ± 4.07
Mean age at biopsy (years)
11.28 ± 4.30
13.13 ± 5.83
Indications for transplant
Fulminant hepatic failure
Genetic cholestatic disease
The histological stage and grade are shown in Tables 2 and 3. Fibrosis was observed in 61 of 63 patients (97%; Figs. 1 and 2); 14 (22%) patients scored F3, 17 (27%) patients scored F2, and 30 (48%) patients scored F1, with only 2 (3%) patients having no fibrosis. None of the biopsies contained regenerative nodules or other evidence of cirrhosis. The mean donor age and ischemia times are displayed in Table 4. As shown in Table 5, grade 2 or greater inflammatory changes were significantly associated with ≥stage 3 fibrosis (P = 0.004). There were 6 (9.5%) patients with chronic rejection: of these, 3 had early chronic rejection with bile duct loss > 20%, and the other 3 had late chronic rejection with ductopenia exceeding 50%. In the 2 years prior to the index biopsy, only 1 patient had a biopsy suggestive of chronic rejection. All 6 patients with chronic rejection had elevated ALT, AST, and GGT.
Table 2. Percentage of Patients with Fibrosis
Table 3. Percentage of Patients with Inflammatory Changes on Biopsy
Table 4. Characteristics of 63 Transplant Recipients
Mean ± SD
Median (Interquartile Range)
Abbreviation: SD, standard deviation.
Data were available from a retrospective chart review on a subset of the entire sample.
Table 5. Association Between Advanced Fibrosis (≥Stage 3) and ≥Grade 2 Inflammation in Allograft Biopsies in 63 Patients
NOTE: P = 0.004.
Of the patients on cyclosporine, 7 had advanced fibrosis (≥stage 3), and 21 had <stage 3 fibrosis; of the patients on tacrolimus, 7 had advanced fibrosis, and 26 had <stage 3 fibrosis; the 2 patients on sirolimus had <stage 3 fibrosis; the 6 patients on mycophenolate had <stage 3 fibrosis; of the patients on steroids, 7 had advanced fibrosis, and 18 had <stage 3 fibrosis; and of the patients on azathioprine, 4 had advanced fibrosis, and 9 had <stage 3 fibrosis.
Late hepatic artery thrombosis was present in only 3 patients, and chronic rejection was present in 6 patients, as detailed previously. The small number of patients with these outcomes precluded any meaningful statistical analysis of the influence of these factors on fibrosis and long-term graft histology. The interval from transplantation to biopsy of ≥6 years was significantly associated with the presence of ≥stage 3 fibrosis (P = 0.003; Table 6). We found no significant difference between donor age (P = 0.17), cold ischemia time (P = 0.17), warm ischemia time (P = 0.17), and advanced fibrosis stage. Of the 63 patients, 14 patients had a diagnosis of de novo autoimmune hepatitis. A history of de novo autoimmune hepatitis was also not associated with advanced fibrosis (P = 0.08). Hepatitis C polymerase chain reaction was negative in all patients for whom it was performed.
Table 6. Association Between Advanced Fibrosis (≥Stage 3) in Allograft Biopsies and Interval from Transplantation to Biopsy in 63 Patients
Interval from Transplantation to Biopsy
NOTE: P = 0.003.
Liver tests (ALT, AST, and GGT) and tests of liver synthetic function sampled within 48 hours of the liver biopsy were available for all 63 patients. All patients had normal liver synthetic function. As shown in Table 7, 24 had elevated liver tests (2 of the 24 patients had normal GGT), and 39 had normal liver tests at the time of biopsy. For those with elevated liver tests, the median ALT was 125 IU/L (range, 46–615), AST was 111 IU/L (range, 55–110), and GGT was 145 IU/L (range, 15–1724). (Our laboratory normals are as follows: ALT, 2–30 IU/L; AST, 16–52 IU/L; and GGT, 8–55 IU/L.) Eight patients with normal ALT, AST, and GGT and 6 patients with elevated ALT and AST ± GGT had ≥stage 3 fibrosis. The relationship between advanced fibrosis and ALT, AST, and GGT elevation was not statistically significant (P = 0.68). In order to determine if our study had sufficient statistical power to have detected a difference in advanced fibrosis between subjects with normal and elevated ALT, AST, and GGT, we performed a post hoc power analysis. Our study had 99.2% statistical power (β = 0.20) to have detected a difference in advanced fibrosis between those patients with normal ALT, AST, and GGT and those with elevated ALT, AST, and GGT.
Table 7. Association Between Advanced Fibrosis (≥Stage 3) in Allograft Biopsies and Liver Tests (Alanine Aminotransferase, Aspartate Aminotransferase, and Gamma-Glutamyl Transferase) at the Time of Biopsy in 63 Patients
Normal Liver Enzymes
Increased Liver Enzymes
NOTE: The 8 patients biopsied for follow-up of de novo autoimmune hepatitis had normal liver tests at the time of biopsy. P = 0.68.
The usefulness of long-term surveillance liver biopsies in liver transplant recipients is controversial. Costs and potential morbidity and mortality are the main drawbacks associated with percutaneous liver biopsies. However, there are several studies that show the lack of sensitivity and specificity of liver function tests in the diagnosis and grading of severity of graft dysfunction.3, 12–14 This cross-sectional study demonstrated that liver biopsies from a significant proportion of long-term survivors of pediatric liver transplantation display histological abnormalities, with less than 5% of liver grafts having no fibrosis in this study group. Sebagh et al.15 in their series of 143 patients (of which 11 were <18 years of age) reported only 20% of liver grafts with normal histology at 5 and 10 years post liver transplantation. Evans et al.4 in their series of 158 children biopsied at 5 (n = 135) and 10 (n = 64) years post liver transplant reported only 7.4% of liver grafts with normal histology at 5 years post liver transplant, this number dropping to 1.6% at 10 years post transplant. In our series, normal liver histology was rare in long-term follow-up, even though the majority of patients had normal ALT, AST, and GGT.
A large proportion of the long-term survivors in our program elected not to proceed with liver biopsy during the study period. Inconvenience and concerns regarding complications related to the procedure were frequently cited reasons. The patients without biopsy data did not appear to differ from the sample population in any systematic fashion, with the exception that they were slightly older. Because the mean age at transplantation was the same for both groups, the patients without biopsy data also had a longer interval from transplant. The differences in mean age and interval from transplantation were both less than 2 years, which in the long-term phase of care may not be a clinically relevant period of time. Likewise, although we cannot assess progression of fibrosis over time in this cross-sectional sample, considering that an interval from transplantation > 6 years was associated with a higher rate of fibrosis, we might expect that the percentage of patients with fibrosis would have increased if we had included these patients. We therefore believe that this potential ascertainment bias did not lead to an overestimation of the prevalence of fibrosis.
We explored a correlation between ≥stage 3 fibrosis and multiple variables and found only 2 significant relationships. In this series, the only factors associated with advanced fibrosis (≥stage 3) were surviving more than 6 years following liver transplantation and having ≥grade 2 inflammation. Interestingly, donor age was not associated with advanced fibrosis in this series. This patient sample may not have been adequate to test this predictor because the mean donor age in our series was only 16.1 years. Animal data do suggest that the frequency of fibrosis and bile duct lesions increases with donor age, but graft survival does not appear to be influenced by donor age.16 A larger patient sample with a wider range of donor ages would be necessary to assess this impact in children.
In our study, duration from transplant correlated with the degree of fibrosis, with patients who were more than 6 years post liver transplant being more likely to have ≥stage 3 fibrosis. The association between fibrosis and inflammation suggests that fibrosis may be due to ongoing allograft injury. The normal life expectancy of a child receiving liver transplantation is as yet unknown. However, we hope that children might live 50 to 60 years or longer with the same graft. Progressive fibrosis with a gradual decline in liver function threatens that potential and is an outcome that must be carefully studied. Serial liver biopsy assessment in the second and third decades after liver transplant will be essential to understanding the natural history of allograft survival.
Using the 2000 update of the Banff classification of liver allograft rejection,9 of the 63 patients, a histological diagnosis of chronic rejection was made in roughly 10%. None of these patients had findings of acute rejection. We sought a correlation between histological evidence of rejection and biologic changes; all 6 patients with chronic rejection had elevated ALT, AST, and GGT levels. Thus, in our experience, liver enzyme elevation was a reasonable screening method for histological changes associated with rejection.
In the present study, we describe the long-term histological outcome of a group of 63 pediatric patients with a mean histological follow-up of 7.9 years post liver transplant. Our results can be summarized as follows: (1) a significant proportion of pediatric liver transplant recipients had fibrosis in long-term follow-up; (2) normal ALT, AST, and GGT values in the late posttransplant period did not exclude allograft fibrosis and/or injury; (3) duration of transplantation of >6 years was associated with significant fibrosis; and (4) ≥grade 2 inflammation was associated with significant fibrosis.
This descriptive study is a hypothesis-generating study. It is important to note that although this population has a high prevalence of fibrosis, we cannot conclude anything about the etiology of fibrosis, and we do not know if these changes are progressive or static. Our institution has adopted a policy of surveillance biopsies at 5 and 10 years post liver transplantation to monitor for fibrosis progression. The care of the patients described in this study was individualized: if the patient met the definition of de novo autoimmune hepatitis, they were treated with steroids and an antimetabolite; patients with nonspecific portal triaditis had no change made in their immunosuppression; and the presence of fibrosis did not alter our management. Further studies to elucidate the natural progression and cause of liver fibrosis in these recipients are of vital importance.
We conclude that surveillance liver biopsies in pediatric liver transplant recipients are useful in assessing allograft status in long-term follow-up.