Late Protocol Liver Biopsies in the Liver Allograft: A Neglected Investigation?

Transcription

1 LIVER TRANSPLANTATION 15: , 2009 ORIGINAL ARTICLE Late Protocol Liver Biopsies in the Liver Allograft: A Neglected Investigation? George Mells, 1 Caroline Mann, 1 Stefan Hubscher, 2 and James Neuberger 1 1 Liver Unit, Queen Elizabeth Hospital, University Hospital Birmingham NHS Foundation Trust, Birmingham, United Kingdom; and 2 Department of Pathology, University of Birmingham, Birmingham, United Kingdom As outcomes from liver transplantation have improved, attention has focused on long-term outcomes: patient and graft survival is affected by many factors, including the consequences of both overimmunosuppression (eg, renal failure and cancer) and underimmunosuppression (eg, rejection). The use of protocol (rather than event-driven) biopsies of the liver allograft, except for those grafted for HCV infection, has been largely abandoned. The aim of this study was to determine if protocol biopsies can improve the management of liver allograft recipients. A retrospective analysis of liver allograft recipients who had undergone protocol liver biopsies between 2000 and 2006 was performed. One hundred seventy-eight patients with normal liver tests (alcoholic liver disease, 49; autoimmune hepatitis, 20; and primary biliary cirrhosis, 107) who had undergone 235 protocol biopsies were identified. No significant complication from the biopsy was recorded. Liver histology was reported as normal or nearly normal in only 57 (24%). Chronic hepatitis (not obviously related to disease recurrence) was present in 78 (33%). Interpreted in the light of the calculated creatinine clearance, the biopsy findings indicated that overall immunosuppression (IMS) should be maintained or increased with standard calcineurin inhibitor (CNI) based IMS in 25% of cases, that overall IMS should be reduced in 15% of cases, and that overall IMS should be maintained or increased by the substitution of non-nephrotoxic agents for CNIs in 9% of cases. The histological findings led to a documented change in IMS in 76 (32%) (increased IMS, 11; decreased IMS, 58; and switch from CNI, 7). In conclusion, protocol liver biopsy provides important histological information about graft function that is not available from standard liver tests and safely allows modification of IMS to ensure that long-term side effects of drug therapy (eg, renal failure) are minimized while graft function is sustained. Liver Transpl 15: , AASLD. Received September 28, 2008; accepted January 26, Liver transplantation (LT) is well established as an excellent treatment for selected patients with acute or chronic liver failure. Because of surgical, anesthetic, and medical advances, early mortality after LT has decreased substantially since the procedure was introduced. However, late mortality rates are essentially unchanged. 1 Some of the major causes of late mortality, such as renal impairment, vascular disease, and de novo malignancy, are clearly related to immunosuppression (IMS). 2-4 Thus, late mortality may be reduced by minimization of IMS to avoid exposure to unnecessary side effects while still protecting the graft. Determining the minimum IMS required to protect the graft is difficult. Standard liver function tests (LFTs) are unreliable. Several studies have recognized the low sensitivity and specificity of LFTs. 5-7 LFTs may fail to detect important histological abnormalities, including chronic hepatitis (CH). In our experience, unexplained (idiopathic) CH, not obviously related to recurrent or de novo disease, is a common finding in late allograft biopsies, occurring in 30% to 70% of biopsies taken after 12 months However, the extent to which CH has been observed in other studies of late posttransplant biopsies is variable. 12,13 CH is clinically important because it may be associated with progressive fibrosis. 9,10 Many cases are associated with autoantibodies or features of rejection, which suggest an immune-mediated mechanism. 9,14 Limited Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; CH, chronic hepatitis; CI, confidence interval; Cic, cyclosporine; CNI, calcineurin inhibitor; ecrcl, estimated creatinine clearance; HCV, hepatitis C virus; IMS, immunosuppression; LFT, liver function test; LT, liver transplantation; NS, not significant; PBC, primary biliary cirrhosis; Tac, tacrolimus. Address reprint requests to James Neuberger, Liver Unit, Queen Elizabeth Hospital, Birmingham B15 2TH, United Kingdom. Telephone: ; FAX: ; j.m.neuberger@bham.ac.uk DOI /lt Published online in Wiley InterScience ( American Association for the Study of Liver Diseases.

2 932 MELLS ET AL. data suggest that progressive fibrosis associated with CH may be prevented by an increase in IMS. 9,10 CH may therefore be an indication to increase overall IMS in order to prolong graft survival. Conversely, the absence of CH may allow for a safe reduction in IMS with long-term benefits for the patient. Protocol liver allograft biopsies are liver biopsies undertaken as part of the routine management of the transplant recipient rather than to investigate changes in the clinical state or liver tests. In the early years of LT, protocol biopsies were part of standard practice in most LT centers. This is no longer the case. We recently conducted an informal survey of transplant centers on 3 different continents and confirmed that protocol biopsies, except for those with hepatitis C virus (HCV) infection, are undertaken by few centers. 15 Protocol biopsies remain standard practice at the Birmingham Liver Transplant Unit because we believe that knowledge of liver histology will help us optimize IMS and understand further the natural history of the allograft structure and function. We therefore undertook a retrospective review of the liver allograft histology of late protocol allograft biopsies in patients with normal LFTs. We selected patients who had undergone LT for alcoholic liver disease (ALD), autoimmune hepatitis (AIH), or primary biliary cirrhosis (PBC). The main aim was to determine whether information obtained from a protocol liver allograft biopsy might allow more appropriate IMS. We also evaluated the prevalence of CH and other histological abnormalities in these patients. PATIENTS AND METHODS Protocol liver biopsies were initially done at day 7 posttransplant, at 1 year, and annually thereafter. Six years ago, after an internal review, the policy was changed to protocol biopsy at 1 year and, for those grafted for indications other than HCV infection and ALD, at 3 yearly intervals thereafter. Those grafted for HCV and ALD continue to undergo annual protocol biopsy. The IMS regimen has been published elsewhere. 10 In brief, until 1995, first-line IMS consisted of dual therapy with cyclosporine and azathioprine; since 1995, first-line IMS has been tacrolimus and azathioprine. Most LT recipients who received a graft before 1995 have continued to receive cyclosporine. Long-term corticosteroids are routinely used only in those grafted for AIH or for those with late or recurrent rejection. For cyclosporine, the target trough level is 100 to 150 ng/ml (whole blood measured by radio immunoassay); for tacrolimus, it is 4 to 8 ng/ml (whole blood measured by radio immunoassay). The calcineurin inhibitor (CNI) dose is adjusted in light of clinical, serological, and histological findings. The Birmingham Liver Unit maintains a transplant database that contains clinical and demographic information about LT recipients under follow-up in Birmingham. Information for the database is collected prospectively. The transplant database was searched for all patients who had undergone LT for ALD, AIH, or PBC and who had undergone protocol liver biopsy between the years 2000 and We chose these disease categories to determine if the frequency of CH depends on whether the indication for LT is an immune liver disease (AIH and PBC) or a nonimmune liver disease (ALD). We chose ALD as a single, homogeneous cohort in preference to a heterogeneous cohort containing patients with LT for multiple, uncommon conditions. Viral hepatitis was avoided because of the difficulty in distinguishing idiopathic CH from viral allograft hepatitis. LFTs at the time of the protocol biopsy were reviewed, and the biopsy was excluded if the contemporaneous LFTs were abnormal. In Birmingham, routine LFTs consist of total bilirubin, aspartate aminotransferase (AST), alkaline phosphatase (ALP), and albumin. Abnormal LFTs were defined as total bilirubin, AST, or ALP above the upper reference level (ie, total bilirubin 17 mol/l, AST 50 IU/L, and ALP ). The relevant histology reports were reviewed by GM and SH. Reports were categorized according to the histopathological diagnosis. Thus, 6 categories were identified: Normal or nearly normal (minor abnormalities not consistent with a specific histological diagnosis). Unexplained CH. Fatty liver disease (including steatosis and steatohepatitis). Recurrent disease (AIH or PBC). De novo AIH. Other abnormalities (including siderosis and nodular changes). Biopsies with abnormalities consistent with more than 1 pathological diagnosis were categorized according to the predominant abnormalities. The histological diagnoses of CH and recurrent disease were made according to published criteria. 9,11,13,16 The histological diagnosis of recurrent PBC was based on the presence of inflammatory bile duct lesions associated with ductopenia and secondary features of chronic cholestasis (ductular reaction and deposition of copper-associated protein) in periportal hepatocytes. 13 Biopsies were classified as showing no evidence of PBC, or possible, probable, or definite recurrent PBC, as previously described. 16 The histological diagnosis of AIH (de novo and recurrent) was based on the presence of a mononuclear inflammatory infiltrate involving the portal tracts, interface hepatitis, and numerous plasma cells. 13,17 For the purpose of this analysis, biopsies were classified as possible AIH if portal inflammation was present with mild interface hepatitis and probable if inflammation was more severe, including plasma cells. Idiopathic CH was diagnosed on the basis of a predominantly mononuclear portal and/or lobular infiltrate lacking typical features, biliary or vascular features of acute or chronic rejection, or features suggestive of recurrent disease (PBC or AIH). 9,11 For patients with a histological diagnosis of CH, inflammatory activity was graded as mild, moderate, or severe, and fibrosis was staged as none, mild (fibrous portal expansion without bridging), moderate

3 PROTOCOL LIVER ALLOGRAFT BIOPSIES 933 TABLE 1. Case Characteristics of Liver Allograft Recipients with Normal Liver Tests Undergoing Liver Biopsy ALD % AIH % PBC % P Number of patients Number of males Number of females Age at LT (years) Age at biopsy (years) Time from LT to biopsy (years) NS Number of biopsies First biopsy NS Second biopsy Third biopsy ecrcl 50 ml/minute Number on Tac-based IMS NS Number on Cic-based IMS NS Number on non-cni IMS NS Unclear Mean trough Tac (ng/ml) Mean trough Cic (ng/ml) NS Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; Cic, cyclosporine; CNI, calcineurin inhibitor; ecrcl, estimated creatinine clearance; IMS, immunosuppression; LT, liver transplantation; NS, not significant; PBC, primary biliary cirrhosis; Tac, tacrolimus. (bridging fibrosis), or severe (cirrhosis), as previously described. 9,11 Renal tests at the time of biopsy were also recorded, and the estimated creatinine clearance (ecrcl) was estimated with the Cockcroft-Gault formula. Significant renal impairment was defined as ecrcl 50 ml/min. The immunosuppressive regimen at the time of biopsy and subsequent clinical visits was recorded. It was assumed that a change in IMS implemented at a subsequent clinic appointment was prompted by the biopsy unless CNI concentrations were in the toxic range (tacrolimus 15 mg/ml or cyclosporine 300 ng/ ml). Statistical Analysis All proportions are reported as percentages with 95% confidence intervals. The chi-square test or Fisher s exact test was used to look for statistically significant differences in the prevalence of given histological diagnoses between the LT recipient cohorts reviewed in this study. RESULTS Cases and Patient Characteristics (Table 1) Between 2000 and 2006, 258 patients had undergone LT for ALD (n 71), AIH (n 21), or PBC (n 166) with a combined total of 338 biopsies (ALD, 88; AIH, 31; and PBC, 219). Each biopsy was treated as a separate case. Contemporaneous liver tests were abnormal in 101 cases (ALD, 28; AIH, 3; and PBC, 70) and were excluded from further analysis. The information available for 2 PBC cases was incomplete; these cases were also excluded. Thus, 235 cases with normal contemporaneous LFTs were identified (ALD, 60; AIH, 28; and PBC, 147). Patient characteristics are summarized in Table 1. In brief, there were 49 patients with LT for ALD, 20 patients with LT for AIH, and 107 patients with LT for PBC; the median time from LT to biopsy was 2.1, 3, and 3.3 years, respectively (nonsignificant). Significant differences between the groups included the proportion of female LT recipients, the age at LT and biopsy, the proportion with significant renal impairment, and the mean trough tacrolimus level within each group. Thus, a greater proportion of ALD cases were male (90%) in comparison with AIH (21%) and PBC (12%) cases (P 0.005). PBC patients were older than AIH patients at LT (P 0.005) and protocol biopsy (P 0.02). Renal impairment (defined as ecrcl 50 ml/minute) was present in 34% of those grafted for PBC versus 15% and 7% of those grafted for ALD and AIH, respectively (P 0.02). The mean trough tacrolimus level was higher in AIH patients receiving tacrolimus-based IMS in comparison with PBC patients (P 0.02). Although a greater proportion of AIH patients were taking tacrolimus-based IMS (71%) in comparison with those grafted for ALD (59%) and PBC (56%), this difference was not statistically significant. Similarly, although the median cyclosporine concentration was higher in those grafted for ALD (92.9 ng/ml) compared with those grafted for AIH and PBC (83 and 89 ng/ml, respectively), this difference was not statistically significant. No major complications from the biopsy (defined as death, additional intervention, or prolongation of inhospital stay) were recorded. Histological Findings (Table 2) Liver histology was reported as normal or nearly normal in only 30%, 29%, and 24% of ALD, AIH, and PBC

4 934 MELLS ET AL. TABLE 2. Histology in Liver Allograft Recipients with Normal Liver Tests Undergoing Protocol Liver Biopsies Histology ALD % (95% CI) AIH % (95% CI) PBC % (95% CI) Total % (95% CI) P Value Normal/nearly normal (10 30) 8 29 (15 47) (19 33) (19 30) NS Fatty liver (30 54) 4 14 (6 32) 4 3 (1 7) (10 19) Idiopathic CH (19 41) 5 18 (8 36) (30 46) (27 39) NS Recurrent disease (AIH/PBC) NA NA (24 58) (22 37) (18 29) De novo AIH 1 2 (0 9) 0 0 (0 12) 4 3 (1 7) 5 2 (1 5) NS Other abnormalities 6 10 (5 20) 0 0 (0 12) 2 1 (0 5) 8 3 (2 7) Total biopsies Recurrent PBC Possible (22 50) Probable (22 50) Definite (19 45) Total recurrent PBC Recurrent AIH Possible 7 64 (35 85) Probable 4 36 (15 65) Total recurrent AIH CH: inflammatory activity Mild CH (81 100) (54 100) (85 98) (89 99) NS Moderate CH 0 0 (0 19) 0 0 (0 46) 3 5 (2 15) 3 4 (1 11) Severe CH 0 0 (0 19) 0 0 (0 46) 0 0 (0 6) 0 0 (0 5) Total CH CH-associated fibrosis: Severity of fibrosis Mild fibrosis 9 53 (31 74) 2 40 (12 78) (41 66) (41 63) NS Moderate fibrosis 2 12 (4 35) 1 20 (4 64) 5 9 (4 19) 8 10 (5 19) Severe fibrosis 0 0 (0 19) 0 0 (0 46) 0 0 (0 6) 0 0 (0 5) Total fibrosis (41 83) 3 60 (22 88) (50 74) (50 73) Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; CH, idiopathic chronic hepatitis; CI, confidence interval; NS, not significant; PBC, primary biliary cirrhosis. cases, respectively. Histological abnormalities observed in protocol biopsies included fatty liver disease (steatosis and steatohepatitis), idiopathic CH, recurrent PBC, and recurrent and de novo AIH, as shown in Table 2. Of the 33 biopsies with fatty liver disease, 29 biopsies showed steatosis, and 4 had additional features of steatohepatitis, the latter all in the ALD group. The remaining 8 biopsies with other abnormalities included 3 with siderosis (all in the ALD group) and 5 with nodular changes (3 transplanted for ALD and 3 for PBC). A greater proportion of those with ALD had fatty liver disease (42%) in comparison with those with AIH (14%) or PBC (2%; P 0.005). Steatosis in patients with LT for ALD is likely to reflect recurrent alcohol consumption, but this was not confirmed by the questioning of the recipients. Steatohepatitis (n 4) and siderosis (n 3) were also more common in the ALD group, but the small numbers preclude valid statistical analysis. Recurrent PBC (possible, probable, or definite) was observed in 30% of PBC cases. Recurrent AIH (possible or probable) was observed in 40% of AIH cases. De novo AIH was observed in 1 ALD case (1.7%) and 4 PBC cases (2.7%) Table 2 also shows the frequency of idiopathic CH and the severity of inflammatory activity and fibrosis. Overall, idiopathic CH was observed in 33% of protocol biopsies (ALD, 28%; AIH, 18%; and PBC 34%; the difference was not significant). Idiopathic CH was associated with mild inflammatory activity in 96% of cases and moderate activity in 4% of cases. Fibrosis was present in 65% of ALD cases with idiopathic CH (18% of all ALD cases), in 60% of AIH cases with idiopathic CH (11% of all AIH cases), and in 63% of PBC cases with idiopathic CH (24% of all PBC cases). There was no significant difference in the proportion of ALD, AIH, or PBC cases with idiopathic CH-associated fibrosis. Overall, fibrosis was mild in severity in 84% of all idiopathic CH cases and moderate in 16% of cases. CH and Renal Impairment in Patients with Normal Contemporaneous LFTs (Table 3) The ecrcl could not be calculated in 2 ALD cases. With the exclusion of these cases, 57 of 233 cases (24.5%) had significant renal impairment with ecrcl 50 ml/ minute. Of these, 23 had no evidence of CH, and 12 had recurrent PBC. Thus, in 35 cases (15%), idiopathic CH was absent, but a significant side effect of IMS was present. These patients might have derived benefit from reduced exposure to CNI. Of those with ecrcl 50 ml/minute, 13 cases (6%) had significant idiopathic CH (classified as moderate or severe idiopathic CH or idiopathic CH associated with any degree of fibrosis). In these cases, additional IMS

5 PROTOCOL LIVER ALLOGRAFT BIOPSIES 935 TABLE 3. Idiopathic Chronic Hepatitis and Fibrosis and Calculated Creatinine Clearance in All Cases with Normal Liver Function Tests ecrcl 50 (ml/min) % ecrcl 50 (ml/min) % Any ecrcl % No inflammation (7 14) (26 38) (35 48) Recurrent/de novo AIH 2 1 (0 3) 14 6 (4 10) 16 7 (4 11) Recurrent PBC 12 5 (3 9) (10 18) (14 24) Mild CH/no fibrosis 7 3 (2 6) 21 9 (6 13) (8 17) Moderate-to-severe CH/any fibrosis 13 6 (3 9) (12 21) (17 27) Total (19 30) (69 81) Abbreviations: AIH, autoimmune hepatitis; CH, idiopathic chronic hepatitis; ecrcl, estimated creatinine clearance; PBC, primary biliary cirrhosis. TABLE 4. Changes in IMS Following Protocol Liver Biopsy ALD (60) % AIH (28) % PBC (147) % Total (235) % Increased overall IMS 2 3 (1 11) 3 11 (4 27) 6 4 (2 9) 11 5 (3 8) Reduced overall IMS (19 41) 7 25 (13 44) (17 31) (20 31) Switch to CNI-sparing regimen 1 2 (0 9) 1 4 (1 18) 5 3 (2 8) 7 3 (1 6) Total change (23 46) (24 58) (24 38) (27 39) Abbreviations: AIH, autoimmune hepatitis; ALD, alcoholic liver disease; CNI, calcineurin inhibitor; IMS, immunosuppression; PBC, primary biliary cirrhosis. was probably appropriate, but because of significant renal impairment, increased IMS should have been achieved with non-nephrotoxic agents. Of those with ecrcl 50 ml/minute, 7 cases (3%) had mild idiopathic CH and no fibrosis. These patients might have derived benefit from overall IMS being maintained. However, the presence of significant renal impairment suggests that IMS should have been maintained with non-nephrotoxic agents. A total of 176 cases had ecrcl 50 ml/minute (74%). Of these, 58 cases (25%) had idiopathic CH (mild, moderate, or severe). These patients might also have derived benefit from overall IMS being increased or maintained. The absence of significant renal impairment suggests that conversion to non-nephrotoxic agents would not have been necessary. Of those with ecrcl 50 ml/minute, 73 cases (31%) had no inflammation on liver biopsy. In these cases, there was no immediate reason to alter IMS. However, a reduction in overall IMS to avoid future complications from IMS could have been considered. There were 16 cases with recurrent/de novo AIH. In the Birmingham Liver Unit, recurrent or de novo AIH is treated with the addition of a low-dose corticosteroid to the IMS regimen. The IMS regimen may also be altered according to the presence of side effects. Thus, in this study, there were 2 cases with recurrent/de novo AIH and ecrcl 50 ml/minute who may have benefited from the use of non-nephrotoxic agents (in addition to a low-dose corticosteroid) to maintain overall IMS. Figure 1. Change in the grade of chronic hepatitis or the stage of fibrosis in sequential biopsies. Documented Changes in IMS Following Protocol Liver Biopsy (Table 4) In total, protocol biopsy was followed by a change in IMS in 30% of cases (33% of biopsies in ALD cases, 39% of biopsies in AIH cases, and 28% of biopsies in PBC cases). Sequential Biopsies (Fig. 1) A total of 102 patients had undergone 2 or more biopsies between 2000 and 2006 (ALD, 27; AIH, 6; and PBC, 69). In 42 cases, idiopathic CH was evident in the first biopsy (ALD, 9; AIH, 4; and PBC, 29). The subsequent

6 936 MELLS ET AL. biopsy revealed progression of the inflammatory grade in 12 cases (ALD, 3; AIH, 1; and PBC, 8) and progression of fibrosis in 13 cases (ALD, 3; AIH, 0; and PBC, 10). No change in the inflammatory grade was observed in 18 cases (ALD, 4; AIH, 0; and PBC, 14), and there was no change in fibrosis in 5 cases (ALD, 2; AIH, 0; and PBC, 3). Regression of inflammatory grade was seen in 13 cases (ALD, 2; AIH, 3; and PBC, 8), and regression of fibrosis was seen in 11 cases (ALD, 2; AIH, 2; and PBC, 7). In the AIH group, the subsequent biopsy revealed recurrent AIH in 2 cases. In the PBC cohort, the subsequent biopsy revealed recurrent PBC in 11 cases. We sought a relationship between changes in CH and/or fibrosis and changes in IMS. Unfortunately, partly as a result of retrospective data collection, it was impossible to determine how overall IMS had changed between biopsies. DISCUSSION HISTOLOGICAL ABNORMALITIES ARE FREQUENT IN LT RECIPIENTS WITH NORMAL LFTS In this group of liver allograft recipients with normal liver tests, graft histology was normal or nearly normal in less than one-third of the cohort. This is consistent with previous studies in which abnormal allograft histology has been reported in 36% to 90% of liver allograft recipients with normal LFTs. 6,18-20 THE MOST FREQUENT HISTOLOGICAL ABNORMALITY DETECTED WAS CH In this study, the most common histological abnormality was idiopathic CH, which was reported in one-third of all cases. The reported prevalence of idiopathic CH is variable. Berenguer et al. 5 reported CH in almost 15% of patients grafted for non HCV-related cirrhosis, whereas Heneghan et al. 21 reported nonspecific inflammation in 7%, 13%, and 17% of patients grafted for HCV, ALD, and cryptogenic cirrhosis, respectively. Maor-Kendler et al. 22 reported inflammation (hepatitis activity index 2) in 29% of 1-year biopsies from LT recipients grafted for cryptogenic cirrhosis, and Rosenthal et al. 23 reported unexplained inflammation in nearly half of annual protocol liver biopsy specimens from healthy pediatric LT recipients. Other studies have reported CH in up to 40% of liver allograft biopsies. 8 Conversely, Burra et al. 24 reported no cases of idiopathic CH in 118 biopsies from 34 LT recipients grafted for ALD, and Abraham et al. 7 reported no cases of idiopathic CH in a series of 394 protocol biopsies. Reasons for the variable prevalence of CH are not clear. It may reflect greater use of protocol biopsies in some centers, differences in the definitions of CH employed in different centers, differences in routine IMS regimens, or some other as yet unidentified center-specific effect. The etiology and significance of CH are uncertain, and it may represent the histological features of 1 or more of a number of different processes, including recurrent disease, rejection, de novo AIH, viral infection (including hepatitis E), and drug toxicity. 25 Regardless of its etiology, CH is clinically important because it may be associated with progressive fibrosis in both adult and pediatric recipients, as shown in previous studies from this unit. 9-11,26 More recently, Herzog et al. 14 reported that interface hepatitis was found in nearly onequarter of 119 liver allograft recipients with abnormal liver tests and was associated with a high probability of fibrosis progression. In the present study, we found that approximately one-third of LT recipients with normal LFTs have CH and, therefore, are at risk of progressive fibrosis, cirrhosis, and graft loss. CH Was Similarly Prevalent in All Groups, Regardless of the Indication for Transplantation It has been suggested that CH may represent a hepatitic form of late rejection. 9 Idiopathic CH might thus be more prevalent in LT recipients grafted for an immunemediated liver disorder such as PBC or AIH because rejection has been reported to occur more frequently in this setting. However, no significant difference in the prevalence of CH between those grafted for immunemediated disease (PBC and AIH) and other groups (ALD) was found. This observation does not elucidate the etiology of idiopathic CH but does suggest that all LT recipients may benefit from protocol allograft biopsies, regardless of the indication for transplantation. In Some Cases, CH Is an Early Manifestation of Recurrent Disease Of those who had undergone more than 1 biopsy during the study period, CH was reported in the first biopsy but recurrent PBC was reported in the second in 11 cases. In 2 cases, idiopathic CH was reported in the first biopsy, and recurrent AIH was reported in the second. This suggests that, in some cases, CH is an early manifestation of disease recurrence. The presence of nonspecific CH preceding diagnostic abnormalities of recurrent disease has been described previously for both PBC and AIH. 27,28 However, in many cases in this study, features of idiopathic CH persisted in subsequent biopsies, and this suggests that CH is not necessarily an early manifestation of disease recurrence but may be an independent pathological entity. In our unit, we take the pragmatic approach that, in the absence of other histological changes, CH is not a manifestation of recurrent disease, and consideration is given to increasing IMS (discussed later). We acknowledge that this approach will sometimes result in unnecessary exposure to excessive IMS. Phenotypic characterization of the infiltrates and use of more histological markers may help in the distinction between CH and other forms of allograft hepatitis.

7 PROTOCOL LIVER ALLOGRAFT BIOPSIES 937 CH May Become More Severe or Less Severe or Remain Static Between Biopsies In this study, we have shown that the severity of inflammatory activity and fibrosis may progress, remain static, or regress. We were interested in determining whether changes in the severity of inflammation and fibrosis were related to changes in IMS. Unfortunately, it was not feasible to investigate this relationship because, for several reasons, we were unable to quantify how overall IMS had changed from 1 biopsy to the next. For example, it was difficult to gauge the change in overall IMS if the CNI dose had been reduced but the non-cni dose had been increased, if tacrolimus had been converted to cyclosporine, or if azathioprine had been converted to mycophenolate. Hence, apart from previous data suggesting that IMS may be important, the factors associated with progressive fibrosis remain unclear. It is likely that a carefully designed prospective study will be required to elucidate these factors. Protocol Allograft Biopsies Guide IMS on the Basis of the Presence or Absence of CH and Side Effects from IMS On the basis of other studies, CH may respond to increased IMS. 10 In the study by Evans et al., 9 the presence of CH was viewed as an indication to add a steroid to the IMS regimen. Unfortunately, in this study, it was not possible to look for a relationship between overall IMS and changes in the inflammatory grade or fibrosis stage. However, if we accept the previous evidence that at least some cases of CH respond to increased IMS, approximately one-third of patients in this study may benefit from overall IMS being increased or maintained. Without a protocol biopsy, these patients could not have been identified. Immunosuppressive agents have side effects that contribute to long-term morbidity and mortality posttransplant. In this study, 98% of cases were receiving CNI, and approximately 20% had renal impairment with ecrcl 50 ml/minute. There are many causes of renal impairment in the liver allograft recipient, but the most likely factor that either causes or contributes to renal impairment is the long-term use of a CNI. 29 To prevent CNI-dependent renal impairment and other long-term morbidity, the dose of the immunosuppressant should be the minimum required to prevent immune-mediated graft damage. Interpreted in the context of side effects from IMS, protocol biopsies can help identify cases in which IMS can be reduced or modified. In this study, the median follow-up was relatively short ( years), and there were relatively few cases with sequential biopsies. Longer follow-up and additional patients with sequential allograft biopsies may provide further information about the time course of CH. Nevertheless, on the basis of the studies by Syn et al., 10 Evans, 9 and Seyam et al., 26 we believe that overall IMS should be maintained or increased when there is ongoing inflammatory activity (as shown by CH), but it may be judiciously reduced if CH is absent. We recognize that prospective randomized controlled trials are required to provide definitive evidence to support this approach, although other centers might consider reviewing their use of protocol biopsies as an aid to optimizing and even weaning from IMS. In this study, the absence of idiopathic CH meant that a reduction in overall IMS could have been considered in 60% of cases, with the possibility of long-term benefit for the patient. One quarter of all cases had significant renal impairment (ecrcl of 50 ml/ minute), and this suggests that the CNI dose should have been reduced. However, one-third of these patients had CH, and this indicates that overall IMS should have been maintained or increased. These patients would have benefited from switching from CNIdependent IMS to non CNI-dependent IMS. Because liver tests were normal in all cases, the decisions to reduce or alter IMS could not have been made without a protocol biopsy. In theory, therefore, the use of protocol biopsies does provide benefit to the patient because the amount and type of IMS can be tailored to the individual. It appears that protocol biopsies are also useful in practice because, in this study, the biopsy was followed by a change in management in approximately one-third of all cases. With a more structured approach, it is likely that this proportion could be increased. ACKNOWLEDGMENT The authors are grateful to Dr. Desley Neil and Dr. Philippe Taniere for their assistance with the histopathological reports, to Ms. Bridget Gunson and Ms. Sue Paris for their help with accessing and maintaining the Liver Unit database, and to the surgical, medical, nursing, and other staff for the clinical care. REFERENCES 1. Adam R, McMaster P, O Grady JG, Castaing D, Klempnauer JL, Jamieson N, et al., and the European Liver Transplant Association. Evolution of liver transplantation in Europe: report of the European Liver Transplant Registry. Liver Transpl 2003;9: Mells G, Neuberger J. Reducing the risks of cardiovascular disease in liver allograft recipients. Transplantation 2007; 83: Ojo AO, Held PJ, Port FK, Wolfe RA, Leichtman AB, Young EW, et al. Chronic renal failure after transplantation of a nonrenal organ. N Engl J Med 2003;349: Galve ML, Cuervas-Mons V, Figueras J, Herrero I, Mata M, Clemente G, et al. Incidence and outcome of de novo malignancies after liver transplantation. Transplant Proc 1999;31: Berenguer M, Rayón JM, Prieto M, Aguilera V, Nicolás D, Ortiz V, et al. Are post-transplantation protocol liver biopsies useful in the long term? Liver Transpl 2001;7: Slapak GI, Saxena R, Portmann B, Gane E, Devlin J, Calne RY, Williams R. Graft and systemic disease in long-term survivors of liver transplantation. Hepatology 1997;25: Abraham SC, Poterucha JJ, Rosen CB, Demetris AJ, Krasinskas AM. Histologic abnormalities are common in pro-

8 938 MELLS ET AL. tocol liver allograft biopsies from patients with normal liver function tests. Am J Surg Pathol 2008;32: Hubscher SG. Chronic hepatitis in liver allografts. Hepatology 1990;12: Evans HM, Kelly DA, McKiernan PJ, Hubscher S. Progressive histological damage in liver allografts following paediatric liver transplantation. Hepatology 2006;43: Syn W-K, Nightingale P, Gunson B, Hubscher SG, Neuberger JM. Natural history of unexplained chronic hepatitis after liver transplantation. Liver Transpl 2007;13: Mohamed R, Hubscher SG, Mirza DF, Gunson BK, Mutimer DJ. Post-transplantation chronic hepatitis in fulminant hepatic failure. Hepatology 1997;25: Shaikh OS, Demetris AJ. Idiopathic posttransplantation hepatitis? Liver Transpl 2007;13: Banff Working Group. Liver biopsy interpretation for causes of late allograft dysfunction. Hepatology 2006;44: Herzog D, Soglio DB, Fournet JC, Martin S, Marleau D, Alvarez F. Interface hepatitis is associated with a high incidence of late graft fibrosis in a group of tightly monitored pediatric orthotopic liver transplantation patients. Liver Transpl 2008;14: Mells G, Neuberger J. Protocol liver allograft biopsies. Transplantation 2008;85: Hubscher SG, Elias E, Buckels JA, Mayer AD, McMaster P, Neuberger JM. Primary biliary cirrhosis. Histological evidence of disease recurrence after liver transplantation. J Hepatol 1993;18: Hubscher SG. Recurrent autoimmune hepatitis after liver transplantation: diagnostic criteria, risk factors, and outcome. Liver Transpl 2001;7: Neuberger J, Wilson P, Adams D. Protocol liver biopsies: the case in favour. Transplant Proc 1998;30: Pappo O, Ramos H, Starzl TE, Fung JJ, Demetris AJ. Structural integrity and identification of causes of liver allograft dysfunction occurring more than 5 years after transplantation. Am J Surg Pathol 1995;19: Sebagh M, Rifai K, Feray C, Yilmaz F, Falissard B, Roche B, et al. All liver recipients benefit from the protocol 10- year liver biopsy. Hepatology 2003;37: Heneghan MA, Zolfino T, Muisan P, Portmann BC, Rela M, Heaton ND, O Grady JG. An evaluation of long-term outcomes after liver transplantation for cryptogenic cirrhosis. Liver Transpl 2003;9: Maor-Kendler Y, Batts KP, Burgart LJ, Wiesner RH, Krom RA, Rosen CB, Charlton MR. Comparative allograft histology after liver transplantation for cryptogenic cirrhosis, alcohol, hepatitis C, and cholestatic liver diseases. Transplantation 2000;70: Rosenthal P, Emond JC, Heyman MB, Snyder J, Roberts J, Ascher N, Ferrell L. Pathological changes in yearly protocol liver biopsy specimens from healthy pediatric liver recipients. Liver Transpl Surg 1997;3: Burra P, Miono D, Cecchetto A, Cillo U, Zanus G, Fagiuoli S, et al. Histological features after liver transplantation in alcoholic cirrhotics. J Hepatol 2001;34: Kamar N, Selves J, Mansuy JM, Ouezzani L, Péron JM, Guitard J, et al. Hepatitis E virus and chronic hepatitis in organ-transplant recipients. N Engl J Med 2008;358: Seyam M, Neuberger JM, Gunson BK, Hubscher SG. Cirrhosis after orthotopic liver transplantation in the absence of primary disease recurrence. Liver Transpl 2007;13: Sebagh M, Farges O, Dubel L, Samuel D, Bismuth H, Reynes M. Histological features predictive of recurrence of primary biliary cirrhosis after liver transplantation. Transplantation 1998;65: Duclos-Vallee JC, Sebagh M, Rifai K, Johanet C, Ballot E, Guettier C, et al. A 10 year follow up study of patients transplanted for autoimmune hepatitis: histological recurrence precedes clinical and biochemical recurrence. Gut 2003;52: Ojo AO. Renal disease in recipients of non-renal solid organ transplantation. Semin Nephrol 2007;27: