This study was supported by Helsinki University Central Hospital (through a grant), the Finnish Pediatric Research Foundation, the Finnish Transplantation Society, the Sigrid Juselius Foundation, the Päivikki and Sakari Sohlberg Foundation, and the Academy of Finland.
Low-dose steroids associated with milder histological changes after pediatric liver transplantation
Article first published online: 12 FEB 2013
Copyright © 2012 American Association for the Study of Liver Diseases
Volume 19, Issue 2, pages 145–154, February 2013
How to Cite
Kosola, S., Lampela, H., Jalanko, H., Mäkisalo, H., Lohi, J., Arola, J. and Pakarinen, M. P. (2013), Low-dose steroids associated with milder histological changes after pediatric liver transplantation. Liver Transpl, 19: 145–154. doi: 10.1002/lt.23565
- Issue published online: 12 FEB 2013
- Article first published online: 12 FEB 2013
- Accepted manuscript online: 29 OCT 2012 03:06AM EST
- Manuscript Accepted: 15 OCT 2012
- Manuscript Received: 17 JUL 2012
Controversy remains about the role of protocol liver biopsy for symptom-free recipients and about the long-term use of low-dose steroids after pediatric liver transplantation (LT). We conducted a national cross-sectional study of pediatric recipients who underwent LT between 1987 and 2007. Liver biopsy samples were taken from 54 patients (82% of survivors) after a median posttransplant follow-up of 11 years, and they were reviewed by 2 pathologists blinded to the clinical data. Biopsy samples from 18 patients (33%) showed nearly normal histology with no inflammation, fibrosis, or steatosis. Portal inflammation was detected in 14 samples (26%), showed no correlation with anti-nuclear antibodies, and was less frequent in the 35 patients whose immunosuppression included steroids (14% versus 47% of patients not using steroids, P = 0.009). Fibrosis was present in 21 biopsy samples (39%). According to the Metavir classification, 15 were stage 1, 4 were stage 2, and 2 were stage 3. The fibrosis stage correlated negatively with serum prealbumin levels (r = −0.364, P = 0.007) and positively with chronic cholestasis (cytokeratin 7 staining; r = 0.546, P < 0.001) and portal inflammation (r = 0.350, P = 0.01). Microvesicular steatosis was found in 23 biopsy samples (43% of patients in 5%-80% of hepatocytes), and it correlated with the body mass index (r = 0.517, P < 0.001) but not with steroid use. The age of the allograft (donor age plus follow-up time) correlated with higher serum gamma-glutamyltransferase (r = 0.472, P < 0.001) and conjugated bilirubin levels (r = 0.420, P = 0.002) as well as chronic cholestasis (r = 0.305, P = 0.03). The biopsy findings led to treatment changes in 10 patients (19%), whereas only 1 complication (subcapsular hematoma) was encountered. In conclusion, continuing low-dose steroids indefinitely after pediatric LT may have a positive effect on the long-term histological state of the liver graft. Allograft aging may lead to chronic cholestasis and thus contribute to the development of liver fibrosis. Liver Transpl 19:145–154, 2013. © 2012 AASLD.
body mass index
complement component 4d
homeostasis model assessment
Because good long-term survival after pediatric liver transplantation (LT) may now be expected in the majority of patients, issues related to the persistence of stable liver function and preserved histology are becoming more relevant. Although liver biopsy is widely used in the early posttransplant period, controversy remains about whether the routine follow-up of symptom-free LT recipients should include liver biopsy. Liver biopsy samples from clinically stable pediatric LT recipients with at least 3 years of follow-up after transplantation have shown surprisingly high rates of histological changes, with fibrosis in 69% to 97%, inflammatory changes in 43%, and de novo autoimmune hepatitis (AIH) in 6% to 22%. Although the diagnosis of de novo AIH is based on chronic hepatitis, associated liver dysfunction, the presence of autoantibodies, and raised serum immunoglobulin levels, complement component 4d (C4d) has been proposed to aid in the differential diagnosis.
Although immunosuppression (IS) regimens at different centers vary greatly, with some centers recommending steroid and azathioprine (AZA) withdrawal to enhance patient adherence and others using the more conventional triple-medication regimen, the long-term effects of IS on liver histology have rarely been considered. Because the alarmingly high rates of fibrosis reported elsewhere require some explanation, cytokeratin 7 (CK7) as a marker of liver cell injury and subtle cholestasis seems to be an interesting addition to the more conventional stains. However, we found no reports of CK7 use in a pediatric population after LT. Since the 1990s, poorer outcomes have been associated with greater donor age both in adults[10, 11] and in children. Although several studies of children have included donor age as a variable, we found no reports on the effects of actual allograft aging in the pediatric population. We hypothesized that both IS agents and the age of the allograft may influence liver histology in long-term, symptom-free survivors of pediatric LT.
PATIENTS AND METHODS
We conducted a population-based cross-sectional study of all 99 pediatric recipients (age < 18 years) who underwent deceased donor LT in Finland between 1987 and 2007. All transplants were performed at Helsinki University Central Hospital, to which the LT recipients also returned for their follow-up visits. Fifty-four of the 66 survivors (82%) underwent liver biopsy after a median posttransplant follow-up of 11 years (range = 3-22 years) at a median age of 18 years (range = 4-36 years). The age at LT and other demographic data are presented in Table 1. Most patients (36/54 or 67%) received a reduced size graft (most often segments 2 and 3), and in 39 cases, the graft bile duct was attached to a Roux-en-Y loop of jejunum. Three patients underwent re-LT 1 month to 2 years after primary LT. The causes were hepatic artery thrombosis, primary nonfunction, and chronic rejection. Six patients were treated for bile duct strictures (2 reoperations, 3 stents, and 1 percutaneous transhepatic catheter), and 5 had 1 or 2 cholangitis episodes each. Three patients had arterial complications: 1 patient underwent arterial reconstruction for hepatic artery thrombosis 1 day after LT, 1 patient underwent angioplasty and stenting for hepatic artery stenosis 1 month after LT, and 1 patient developed portal thrombosis 6 years after LT. At the time of the study, all participating patients were symptom-free, and laboratory tests showed stable graft function (Table 1). Twelve patients (22%) used an antihypertensive medication. None of the patients had diabetes, but 11 (20%) had insulin resistance [homeostasis model assessment (HOMA) score > 2.5]. Forty-two patients (78%) were cytomegalovirus (CMV)-positive, but none of the patients had viral hepatitis.
|Age (years)||18 (4-36)|
|Sex: male/female (n/n)||25/29|
|Liver disease diagnosis [n (%)]|
|Biliary atresia||19 (35)|
|Metabolic diseasea||9 (17)|
|Hepatic malignancy||6 (11)|
|Reduced graft (%)||67|
|Cold ischemia time (minutes)||457 (244-831)|
|Anhepatic time (minutes)||57 (20-115)|
|Age at LT (years)||7 (0.4-18)|
|Donor age (years)||21 (1-62)|
|Follow-up after LT (years)||11 (3-22)|
|Graft age (years)c||35 (11-72)|
|BMI group [n (%)]d|
|Underweight (<19 kg/m2)||5 (9)|
|Normal weight (19-25 kg/m2)||34 (63)|
|Overweight (25.1-30 kg/m2)||6 (11)|
|Obese (>30 kg/m2)||9 (17)|
|Total serum bilirubin (μmol/L)||10 (4-34)|
|AST (U/L)||29 (20-68)|
|ALT (U/L)||23 (9-72)|
|GGT (U/L)||19 (8-99)|
|Serum prealbumin (mg/L)||202 (102-519)|
|Fasting plasma glucose (mmol/L)||5.1 (4.2-6.1)|
|HOMA score||1.9 (0.6-8.7)|
Two patients were excluded from this study because they had undergone retransplantation less than a year before the study. Three of the 10 nonparticipants were young adults with suspected nonadherence because they had missed several appointments, and 7 did not consent to liver biopsy. No differences were found in age, diagnoses of liver disease, or length of follow-up between participants and nonparticipants.
IS remained basically constant during the 20-year period. Maintenance IS was based on a triple-medication regimen, which included cyclosporine A (CSA), AZA, and methylprednisolone (MP). On an individual basis, CSA was switched to tacrolimus and AZA was switched to mycophenolate mofetil if this was clinically indicated. MP was tapered to 0.25 mg/kg/day at 2 weeks, switched to alternate-day dosing at 6 months (0.1 mg/kg/day), and often discontinued after the patient reached adulthood. Fifty-two of the patients (96%) used calcineurin inhibitors at the time of the study; 37 (69%) used CSA, and 15 (28%) used tacrolimus. Half of the patients (27/54) were on a triple-IS regimen, which included a calcineurin inhibitor, AZA or mycophenolate mofetil, and low-dose MP. The 35 patients (65%) still using low-dose MP as part of their IS at the time of the study had been using MP for 6 years (range = 1-22 years), whereas the patients no longer using MP had used it for 9 years (range = 1-17 years, P = 0.88) and had been off MP for 4 years (range = 1-18 years). All patients still on MP used it on an alternate-day basis and received ≤0.1 mg/kg/day. The CSA doses were lower for patients using MP versus patients not using MP (120 ± 35 versus 160 ± 40 mg/day, P = 0.02), but no difference was found in tacrolimus doses. No significant differences in diagnoses were found between MP users and nonusers. Half of the patients (27/54) had been treated for rejection (2-4 episodes each for 8 patients), whereas half had no rejection episodes in their history.
Measurements and Biochemical Analyses
The blood levels of total and conjugated bilirubin, aspartate transaminase (AST), alanine transaminase (ALT), gamma-glutamyltransferase (GGT), prealbumin, glucose, and insulin were taken during the same hospital visit as the liver biopsy sample. The body mass index (BMI) was calculated, and the patients were classified as underweight, normal weight, overweight, or obese according to the recently determined BMI-for-age groups of Finnish children. Insulin resistance was evaluated with the HOMA score according to the following formula:
The cutoff for the diagnosis of insulin resistance was set at 2.5. Anti-nuclear, anti-mitochondrial, anti–smooth muscle, and anti–liver-kidney microsomal antibodies were measured with an accredited indirect immunofluorescence assay, and titers greater than the following values were considered positive: 80 for anti-nuclear antibodies, 50 for anti-mitochondrial antibodies and anti–smooth muscle antibodies, and 10 for anti–liver-kidney microsomal antibodies.
All liver biopsy samples were percutaneous core needle biopsy samples taken by an experienced radiologist under ultrasound guidance and, for the youngest children, under general anesthesia. Biopsy samples were fixed in formalin, embedded in paraffin, cut, and stained with conventional stains. Immunostaining for CK7 and C4d was also performed. Two pathologists (J.A. and J.L.) who had expertise in liver histopathology and were blinded to the clinical data reviewed the biopsy samples to reach a consensus. Inflammation was graded according to the Banff criteria. Portal, lobular, interface, endothelial, and bile duct inflammation was evaluated separately and graded semiquantitatively on a scale of 0 to 3 [(0) none, (1) mild, (2) moderate, and (3) severe]. When inflammation was present, the distribution of inflammatory cell types was recorded. Fibrosis was staged according to the Metavir system on a semiquantitative scale of 0 to 4 [(0) none, (1) mild portal fibrosis, (2) moderate scarring extending from portal areas, (3) moderate fibrosis with bridging, and (4) extensive fibrosis or cirrhosis]. Macrovesicular steatosis and microvesicular steatosis were recorded as the percentages of hepatocytes affected. For statistical analyses, a composite variable of steatosis (yes/no) also was used. The CK7 immunopositivity of periportal hepatocytes was graded semiquantitatively on a scale of 0 to 3 [(0) none, (1) mild, (2) moderate, and (3) marked]. CK7 samples were also used to measure bile duct proliferation, which was reported on a scale of 0 to 2 [(0) none, (1) mild, and (2) marked]. C4d immunopositivity was recorded as present or absent.
This study was approved by the ethics committee of the Helsinki and Uusimaa Hospital District. All participating patients and controls (and parents in the case of minors) signed an informed consent form.
Statistical analyses were performed with SPSS 19.0 statistics software (IBM, Somers, NY). Descriptive statistics included frequencies, percentages, means and standard deviations, and, in the case of skewed distributions, medians and ranges. For continuous variables, the Mann-Whitney U test, the chi-square test, and Fisher's exact test were used as appropriate. For dichotomous variables, Fisher's exact test was used. Correlations were calculated with the Spearman rho 2-tailed test. A P value < 0.05 was considered statistically significant.
Biopsy samples were considered representative because the median sample length was 18 mm (range = 10-31 mm), and the samples included 14 portal tracts (range = 10-44). Biopsy samples from 18 patients (33%) showed nearly normal histology with no inflammation, fibrosis, or steatosis. Portal inflammation was detected in 14 samples (26%; Table 2 and Fig. 1A); 4 of these samples also demonstrated interface inflammation, and 2 showed additional lobular inflammation. Lymphocytes were the predominant cell type in inflammatory infiltrates (70%-100%), and they were followed by plasma cells (5%-20%), neutrophils (5%-10%), macrophages (5%-10%), and eosinophils (5%). Four biopsy samples showed C4d immunopositivity. Fibrosis was present in 21 biopsy samples (39%; Table 2). According to the Metavir classification, 15 were stage 1, 4 were stage 2, and 2 were stage 3. Microvesicular steatosis was found in 23 biopsy samples (43% of patients in 5%-80% of hepatocytes; Fig. 1B); 5 of these samples also showed macrovesicular steatosis (in 5%-30% of hepatocytes). None of the biopsy samples had visible cholestasis or periportal copper deposits. Thirteen biopsy samples (24%) showed CK7-immunopositive periportal hepatocytes (Fig. 1C), whereas a ductular reaction was found in 23 biopsy samples (43%; Fig. 1D), and bile duct loss was found in 4 samples (7%).
|Histological Finding||n (%)|
|0 (none)||41 (76)|
|1 (mild)||9 (17)|
|2 (moderate)||3 (6)|
|3 (marked)||1 (2)|
|0 (none)||31 (57)|
|1 (mild)||20 (37)|
|2 (marked)||3 (6)|
|Loss of bile ducts|
Associations of Histology With Laboratory and Clinical Markers
Although prealbumin levels were higher in the patient group with nearly normal histology versus the patients with histological changes (253 ± 89 versus 190 ± 43 mg/L, P = 0.007), no significant differences were found in patient or donor age, cold ischemia time, length of follow-up, incidence of vascular or biliary complications, or ALT, AST, GGT, or bilirubin levels.
Portal inflammation was less frequent in the 35 patients whose IS still included low-dose MP (14% versus 47% of patients not using steroids, P = 0.009; Fig. 2). Portal inflammation correlated positively with fibrosis staging (r = 0.350, P = 0.01), but no correlations were found with liver biochemistry (Table 3). Among the autoantibodies measured, anti-nuclear antibodies were present in 14 patients, with titers > 80 in 11 (20%). However, no correlation was found between anti-nuclear antibodies and portal inflammation or between portal inflammation and C4d immunopositivity, a pre-LT diagnosis of hepatitis, the CMV status, the follow-up length, or the allograft age. No other autoantibodies were found in any patients. Because none of the patients with positive autoantibodies had elevated transaminase levels or interface hepatitis, no cases of de novo AIH were diagnosed.
|Present (n = 14)||Absent (n = 40)|
|Total serum bilirubin (μmol/L)||12 (4-25)||9 (4-34)||0.33|
|AST (U/L)||30 (20-68)||29 (20-55)||0.99|
|ALT (U/L)||23 (12-72)||23 (9-61)||0.50|
|GGT (U/L)||30 (12-99)||16 (8-99)||0.22|
|Serum prealbumin (mg/L)||178 (102-289)||204 (114-519)||0.19|
|Reduced graft (%)||50||73||0.19|
|Cold ischemia time (minutes)||432 (245-779)||471 (244-831)||0.64|
|Anhepatic time (minutes)||60 (32-115)||54 (20-113)||0.99|
|CMV-positive status (%)||85||78||0.63|
|MP use (%)||36||75||0.008|
|Age at LT (years)||7.2 (0.9-17)||2.6 (0.4-17)||0.08|
|Donor age (years)||26 (1.0-62)||19 (1.0-59)||0.45|
|Follow-up after LT (years)||15 (3.0-22)||10 (2.8-22)||0.08|
|Graft age (years)a||37 (20-68)||33 (11-72)||0.16|
|BMI for age (kg/m2)b||22 (17-39)||22 (16-40)||0.95|
The fibrosis stage correlated negatively with serum prealbumin levels (r = −0.364, P = 0.007) and positively with CK7 immunopositivity (r = 0.546, P < 0.001; Table 4). All 4 patients with interface inflammation had fibrosis. The 2 patients with a Metavir score of 3 had no signs of inflammation, whereas all 4 patients with a Metavir score of 2 also had mild portal inflammation. Vascular or biliary complications showed no correlation with fibrosis, nor did the length of follow-up.
|Present (n = 21)||Absent (n = 33)|
|Total serum bilirubin (μmol/L)||9 (4-43)||10 (4-33)||0.86|
|AST (U/L)||35 (20-116)||28 (20-55)||0.11|
|ALT (U/L)||21 (10-118)||24 (9-61)||0.46|
|GGT (U/L)||18 (10-99)||25 (8-103)||0.84|
|Serum prealbumin (mg/L)||175 (102-254)||220 (135-519)||0.01|
|Reduced graft (%)||71||64||0.55|
|Cold ischemia time (minutes)||458 (245-595)||463 (244-831)||0.57|
|Anhepatic time (minutes)||58 (21-115)||54 (20-113)||0.83|
|CMV-positive status (%)||75||83||0.47|
|MP use (%)||62||67||0.72|
|Age at LT (years)||4.9 (0.7-17)||4.4 (0.4-17)||0.80|
|Donor age (years)||19 (1-52)||25 (1-62)||0.64|
|Follow-up after LT (years)||11 (3-22)||11 (3-22)||0.60|
|Graft age (years)a||33 (11-68)||35 (11-72)||0.65|
|BMI for age (kg/m2)b||23 (17-40)||22 (16-40)||0.45|
|Positive CK7 stain [n (%)]||10 (48)||3 (9)||<0.001|
Microvesicular steatosis (r = 0.517, P < 0.001) and macrovesicular steatosis (r = 0.429, P = 0.001) correlated with the BMI, and macrovesicular steatosis also correlated with HOMA (r = 0.407, P = 0.02). Patients using an antihypertensive medication had a higher BMI than patients not using an antihypertensive medication (26.1 ± 6.6 versus 20.5 ± 4.7 kg/m2, P = 0.004) and more often had steatosis (33% versus 2%, P = 0.001). No connection between steatosis and MP use, inflammation, fibrosis, or donor age was found (Table 5).
|Present (n = 23)||Absent (n = 31)|
|Total serum bilirubin (μmol/L)||11 (5-33)||9 (4-34)||0.09|
|AST (U/L)||29 (20-68)||30 (20-68)||0.61|
|ALT (U/L)||24 (9-72)||23 (10-72)||0.79|
|GGT (U/L)||19 (10-99)||20 (8-99)||0.55|
|Serum prealbumin (mg/L)||188 (114-289)||208 (102-519)||0.18|
|Reduced graft (%)||65||68||0.85|
|Cold ischemia time (minutes)||462 (244-779)||455 (258-831)||0.85|
|Anhepatic time (minutes)||56 (32-85)||59 (20-115)||0.61|
|CMV-positive status (%)||85||77||0.47|
|MP use (%)||65||65||>0.99|
|Age at LT (years)||5.1 (0.7-17)||4.4 (0.4-17)||0.56|
|Donor age (years)||32 (1-59)||18 (1-62)||0.40|
|Follow-up after LT (years)||11 (3-22)||12 (3-22)||0.83|
|Graft age (years)a||38 (18-72)||30 (11-68)||0.33|
|BMI for age (kg/m2)b||24 (16-40)||21 (16-28)||0.001|
|HOMA score||2.6 (1.0-8.7)||1.8 (0.6-3.5)||0.09|
The CK7 immunopositivity of periportal hepatocytes correlated positively with the serum GGT levels (r = 0.495, P < 0.001; Fig. 3) and conjugated bilirubin levels (r = 0.402, P = 0.003) but not with any other liver function tests or portal inflammation (Table 6). In patients with CK7 immunopositivity, the GGT levels (59 versus 15 U/L, P < 0.001) and the conjugated bilirubin levels (5 versus 2 μmol/L, P = 0.004) were significantly higher. Five of the 6 patients treated for bile duct strictures showed CK7-immunopositive periportal hepatocytes, and 4 showed bile duct proliferation. One of the 5 patients treated for cholangitis had CK7-immunopositive periportal hepatocytes. Half of the biopsy samples with CK7-immunopositive hepatocytes were from patients with no clinically diagnosed biliary complications.
|Positive (n = 12)||Negative (n = 42)|
|Total serum bilirubin (μmol/L)||13 (5-34)||9 (4-24)||0.06|
|AST (U/L)||34 (20-68)||28 (20-51)||0.13|
|ALT (U/L)||29 (12-72)||23 (9-50)||0.09|
|GGT (U/L)||59 (16-99)||15 (8-68)||<0.001|
|Serum prealbumin (mg/L)||175 (114-331)||206 (102-519)||0.15|
|Reduced graft (%)||77||63||0.34|
|Cold ischemia time (minutes)||466 (245-514)||460 (244-831)||0.84|
|Anhepatic time (minutes)||57 (35-115)||59 (20-113)||0.56|
|CMV-positive status (%)||92||78||0.30|
|MP use (%)||62||68||0.69|
|Age at LT (years)||5.1 (0.9-17)||3.9 (0.4-17)||0.62|
|Donor age (years)||38 (11-51)||18 (1-62)||0.11|
|Follow-up after LT (years)||16 (3-22)||10 (3-22)||0.17|
|Graft age (years)a||43 (20-68)||29 (11-72)||0.02|
|BMI for age (kg/m2)b||23 (18-34)||22 (16-40)||0.47|
The graft type (whole liver/reduced), the type of biliary anastomosis (end-to-end/Roux-en-Y), the cold ischemia time, and the number of rejection episodes showed no significant correlations with histological findings.
The age of the allograft (donor age plus follow-up time) correlated positively with serum GGT levels (r = 0.472, P < 0.001; Fig. 3) and conjugated bilirubin levels (r = 0.420, P = 0.002) as well as CK7 immunopositivity of periportal hepatocytes (r = 0.305, P = 0.03). No correlations were found between the allograft age and inflammation, fibrosis, or steatosis. As for the 30 patients with 10 or more years of follow-up after LT (median = 17 years), inflammation was present in 10 biopsy samples (33%), fibrosis was present in 13 (43%), and steatosis was present in 13 (43%). Fourteen of these patients (47%) still used low-dose MP more than 10 years after LT, and only 3 of these 14 patients (21%) had portal inflammation, whereas 9 of the 16 patients (56%) not using MP did (P = 0.07). The fibrosis stages were slightly more advanced in comparison with the whole patient population (stage 1 for 8, stage 2 for 4, and stage 3 for 1).
The biopsy findings led to treatment changes in 10 patients (19%). IS was increased in 4 patients and was decreased in 4. One patient was scheduled to undergo magnetic resonance imaging because of marked venous congestion, and 1 patient's control visits were scheduled more frequently because of a precirrhotic condition. Only 1 complication (subcapsular hematoma) was encountered, and it was treated conservatively.
Biopsy samples from long-term, symptom-free survivors of pediatric LT showed inflammation in 26%, fibrosis in 39%, microvesicular steatosis in 43%, and CK7-immunopositive periportal hepatocytes in 24%. The severity of fibrosis correlated with CK7 immunopositivity and inflammation. The use of low-dose MP seemed to protect against inflammatory and fibrotic changes to the liver graft, whereas MP use was unassociated with insulin resistance or liver graft steatosis. No cases of de novo AIH were diagnosed. Allograft aging correlated with serum GGT and conjugated bilirubin levels as well as CK7 immunopositivity.
Our study is limited by the fact that the biopsy samples were acquired cross-sectionally instead of at a certain follow-up interval after LT. None of our patients were completely steroid-free, but the patients not using MP at the time of the study had been off steroids for years. These limitations are countered by the population-based, unselected nature of the study as well as the high participation rate (82%), the long follow-up time (median = 11 years), the consensus of 2 experienced pathologists (which increased the validity of the histological grading), and the meticulous collection of valid follow-up data from hospital records. Although a needle biopsy sample includes only 1/50,000 of the liver volume, it is still the gold standard for assessing liver histology. Also, 29 patients (54%) had already successfully transitioned to adult services; this is a time period potentially hazardous to both the LT recipient and the liver graft.[21, 22] Thus, the histological findings of our study seem even more promising.
The rates of both inflammation and fibrosis were clearly lower in our study population versus several recent reports. Despite many confounding factors between studies from different centers, one possible explanation is our indefinitely continued use of low-dose steroids. Steroid withdrawal is mainly based on the adverse effects of steroids on growth, bone health, and metabolic changes, which are dose-dependent and may be overestimated in patients using a low-dose medication on alternate days.[23, 24] Up to 22% of pediatric LT recipients may later be diagnosed with de novo AIH, which may actually be a part of a continuum of immune-mediated injury. Because its treatment consists of steroids with or without AZA, our post-LT IS protocol (a long-term triple-medication regimen including AZA and low-dose MP) may in fact protect against this entity.[4, 26] C4d provided no additional information because we found no differences in the 4 patients with C4d-positive biopsy samples in comparison with the other patients. Already in 2000, Peeters et al. noted that portal fibrosis was less frequent in patients who had suffered at least 1 acute rejection episode, and they concluded that the steroids used to treat rejection may play a role. Chronic inflammatory changes often lead to fibrosis, especially in children,[5, 28] and this finding persuaded our colleagues in Birmingham to continue low-dose steroids indefinitely. Because of continued MP use, the rate and grades of inflammation among our patients were low, and this in turn led to a lower rate and lower grades of fibrosis.
Besides inflammation, fibrosis may result from concealed cholestasis due to biliary complications and allograft aging. Biliary complications are clearly associated with fibrosis 1 year after LT, but this effect may be diluted with longer follow-up. Although none of our patients had histological cholestasis or significantly elevated bilirubin levels, 13 biopsy samples (24%) showed CK7-immunopositive periportal hepatocytes, and a ductular reaction was present in 23 biopsy samples (43%). Mature hepatocytes are CK7-negative but may undergo dedifferentiation into CK7-positive intermediate hepatobiliary cells under persistent cholestasis. This cholestasis is clinically subtle but prolonged because of incomplete bile duct obstruction and is possibly exacerbated by hypoxia. Mild persistent cholestasis may also lead to a ductular reaction as the liver tries to compensate for the relative bile acid overload. These reactive changes also induce mild inflammation and progressive fibrosis[9, 29] and bind inflammation, fibrosis, and allograft aging even closer together.
Greater donor age is known to influence outcomes and graft function.[10, 11, 13, 14] To the best of our knowledge, this is the first time that the actual graft age has been associated with clinical markers in children. Although the actual allograft ages in our study varied from 11 to 72 years, we were still able to show correlations of allograft aging with higher serum GGT and conjugated bilirubin levels as well as CK7 positivity of periportal hepatocytes. Despite their magnificent regeneration potential, hepatocytes also grow old, and this makes pulling a second lifetime out of a donor liver inconceivable. Children in need of LT should, therefore, be entitled to the livers of the youngest donors. This should also minimize the chances of acquiring a liver that is already fatty at the time of LT.
Steatosis has not been mentioned in any reports on long-term follow-up biopsy samples from pediatric LT recipients from other centers, although it is a well-recognized risk in adults after LT.[32, 33] We find it highly unlikely, however, that no steatosis would be present elsewhere because obesity is a growing global problem, and LT recipients are at risk of acquiring metabolic syndrome. Our patients also illustrated a clear correlation between antihypertensive medication use, insulin resistance, BMI, and liver steatosis. Microvesicular steatosis may represent a milder form of fat accumulation because smaller droplets may fuse into larger ones. Microvesicular steatosis is, however, even harder to grade definitively than the macrovesicular form of fat accumulation. The patterns of fat accumulation may also be affected by patient age, but because of the scarcity of reports on liver steatosis in children, this suspicion has been left unconfirmed. Because none of the biopsy samples in our study showed lobular inflammation or hepatocyte ballooning and only 5 had macrovesicular steatosis (in the most severe case, 30% of the hepatocytes were affected), which led to a maximum nonalcoholic fatty liver disease activity score of 1, nonalcoholic fatty liver disease was mild, and no cases of steatohepatitis were diagnosed. We found no association between steroid use and steatosis. Besides the positive effects that low-dose steroids may have on long-term liver histology, low-dose steroids and lower doses of calcineurin inhibitors are also associated with a more favorable cholesterol metabolism profile after LT. Because LT recipients are at greater risk for cardiovascular disease than the general population, further research involving all associated factors is needed.
Because only 1 minor complication was encountered (2%), whereas treatment changes were made for 10 patients (19%) and histological changes showed very limited correlations with biochemical markers, our study found liver biopsy to be both useful and justifiable in the long-term follow-up of pediatric LT recipients. The follow-up protocol at our center has been updated to include liver biopsy at 5-year intervals.
In conclusion, our results suggest that continuing low-dose steroids indefinitely after pediatric LT may have a positive effect on the long-term histological state of the liver graft. Allograft aging may lead to chronic cholestasis and thus contribute to the development of liver fibrosis, and children should, therefore, receive grafts from the youngest donors.
- 15New Finnish growth references for children and adolescents aged 0 to 20 years: length/height-for-age, weight-for-length/height, and body mass index-for-age. Ann Med 2011;43:235-248., , , , , .
- 19Intraobserver and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. The French METAVIR Cooperative Study Group. Hepatology 1994;20( pt 1):15-20.
- 25What is the long-term outcome of the liver allograft? J Hepatol 2011;55:702-717..
- 29Ductal plates in hepatic ductular reactions. Hypothesis and implications. I. Types of ductular reaction reconsidered. Virchows Arch 2011;458:251-259..
- 30Ductal plates in hepatic ductular reactions. Hypothesis and implications. II. Ontogenic liver growth in childhood. Virchows Arch 2011;458:261-270..
- 33Non-alcoholic fatty liver disease in liver transplant recipients: another story of “seed and soil”. Am J Gastroenterol 2010;105:613-620., , , , , , et al.