Histopathology of De Novo Autoimmune Hepatitis

Transcription

1 LIVER TRANSPLANTATION 18: , 2012 ORIGINAL ARTICLE Histopathology of De Novo Autoimmune Hepatitis Ananya Pongpaibul, 1 Robert S. Venick, 2 Sue V. McDiarmid, 3 and Charles R. Lassman 4 1 Department of Pathology, Faculty of Medicine, Siriraj Hospital, Mahidol University, Nakhon Pathom, Thailand; and 2 Pediatrics, 3 Pediatrics and Surgery, and 4 Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California Los Angeles, Los Angeles, CA De novo autoimmune hepatitis (DAIH) is a well-recognized complication of pediatric liver transplantation (LT). The diagnosis is largely based on elevated liver function test results and the development of autoimmune antibodies. The histology of DAIH was first described in We present detailed histological data from the largest series to date of pretreatment and posttreatment biopsy samples from pediatric LT patients with DAIH. The histological evaluation included first an assessment of the predominant pattern of injury (hepatitis, rejection, or bile duct obstruction). Then, the necroinflammatory activity (interface, lobular, and perivenular), plasma cell density, rejection activity index, and fibrosis were scored. Seventy of 685 pediatric patients (10.2%) who underwent LT developed DAIH according to clinical and biopsy findings. Fifty-one pretreatment biopsy samples and 38 posttreatment biopsy samples were available for a retrospective review. The predominant pattern of injury (hepatitis, rejection, or bile duct obstruction) was determined, and biopsy samples were scored for the necroinflammatory activity (interface, lobular, and perivenular), plasma cell density, rejection activity index, and fibrosis. The most common pattern of injury was lobular hepatitis, which was frequently unaccompanied by interface necroinflammatory activity or prominent plasma cell infiltrates. Seven of the 51 cases had features strongly suggestive of acute rejection. Posttreatment biopsy samples showed a reduction in the degree of necroinflammatory activity and plasma cell infiltrates. In most patients, the degree of fibrosis was stable or had regressed. Because the histological features of DAIH are variable and nonspecific, a high index of suspicion and correlation with autoimmune antibodies are necessary to establish the diagnosis. In the majority of patients with DAIH, treatment appears to yield good clinical outcomes and histological improvements. Liver Transpl 18: , VC 2012 AASLD. Received September 1, 2011; accepted February 17, De novo autoimmune hepatitis (DAIH) is a well-recognized cause of late graft dysfunction in pediatric recipients of liver transplantation (LT). The prevalence of DAIH varies from 2.1% to 11%. 1-9 This wide range is likely due to diagnostic difficulties and the lack of standard diagnostic criteria. At our institution, the diagnostic criteria incorporate clinical and histological features, including histological evidence of hepatitis, an absence of previous autoimmune liver disease, late graft dysfunction [as characterized by elevated liver function test (LFT) results], the development of autoimmune antibodies, and the exclusion of other causes of graft dysfunction. 8 The development of autoimmune antibodies after pediatric LT is common, with the prevalence ranging from 26% to 74%. 10 The histopathology of DAIH was first described in 7 patients by Kerkar et al. in The described histological changes included chronic hepatitis with portal lymphoplasmacytic infiltrates, interface hepatitis, bridging collapse, and zone 3 necrosis without changes typical of acute or chronic rejection. 1 Later studies described similar histopathology. 2-7 A large and detailed study with a description of the histology at the time of the diagnosis of DAIH and after treatment has not yet been published. In this study, we present detailed pretreatment and posttreatment histological Abbreviations: DAIH, de novo autoimmune hepatitis; LFT, liver function tests; LT, liver transplantation. Address reprint requests to Charles R. Lassman, M.D., PhD., Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California Los Angeles, Le Conte Avenue, CHS , Los Angeles CA Telephone: ; FAX: ; classman@mednet.ucla.edu DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2012 American Association for the Study of Liver Diseases.

2 812 PONGPAIBUL ET AL. LIVER TRANSPLANTATION, July 2012 findings and clinical data, including outcome data, for 51 pediatric patients with a diagnosis of DAIH. PATIENTS AND METHODS Between 1984 and August 2008, 685 pediatric patients underwent 883 isolated LT procedures at a single center. Seventy recipients (10.2%) were determined to have DAIH according to our institutional diagnostic criteria. 8 After approval by the institutional review board, the medical records were reviewed, and the following parameters were recorded: age, sex, indication for LT, type and level of immunosuppression, treatment for DAIH, LFT at the time of diagnosis and at intervals after the treatment, and patient and graft survival. Two sets of biopsy samples were reviewed. The first set consisted of 51 biopsy samples taken from patients at or near the time of the diagnosis of DAIH (pretreatment samples) that were available for retrospective review. The histological evaluation included assessment of the following: interface and lobular necroinflammatory activity in accordance with Batts and Ludwig 11 ; perivenular hepatitis (perivenular inflammation and hepatocyte dropout or necrosis with or without central vein endotheliitis), which was scored as 0 (none), 1 (minimal), 2 (mild), 3 (moderate), or 4 (severe); the density of plasma cells as a percentage of the inflammatory cells in portal, lobular, and perivenular areas, which was scored as 0 (none), 1 (<25%), 2 (25%-<50%), 3 (50%-75%), or 4 (>75%); the rejection activity index according to the Banff criteria 12 ; and fibrosis in accordance with Batts and Ludwig. The predominant pattern of injury in each biopsy sample was summarized according to clinical practice. If the findings were strongly suggestive of a diagnosis other than hepatitis (eg, acute or chronic rejection or bile duct obstruction), they were summarized as such. In all biopsy samples, the degree of necroinflammatory activity (mild, moderate, or severe) and the location of the injury (interface, lobular, or perivenular) were recorded. In the minority of cases that demonstrated only minimal lymphocytic inflammation in portal areas and no necroinflammatory activity, the samples were categorized as having minimal nonspecific infiltrates. The second set of biopsy samples that were available for review consisted of 38 samples from 25 patients. Protocol biopsies are not routinely performed at our institution; therefore, these posttreatment biopsy samples were obtained only from patients who experienced a significant increase in LFT or a worsening of their clinical status after treatment. Sixteen patients underwent one posttreatment biopsy, 9 patients underwent 2 4 posttreatment biopsies. The timing of posttreatment biopsy ranged from 5 days to 5 years after treatment. For the statistical analysis, the biopsy samples were grouped into 4 categories according to the amount of time between treatment and rebiopsy: immediate, or less than 4 weeks (group 1); early, or 4 weeks to 3 months (group 2); intermediate, or more than 3 months to 2 years (group 3); and late, or more than 2 years (group 4). Patients who underwent biopsy more than once were categorized into these groups according to the timing of the first biopsy. Histologic evaluation was performed as previously described. All features were scored according to a consensus opinion by 2 pathologists (A.P. and C.R.L.) who reviewed slides together with a doubleheaded scope. The histological findings and the clinical results are presented as frequencies and percentages (for categorical variables) or as means and medians (for continuous variables). RESULTS Clinical Findings The clinical findings are described in Table 1. The mean age at the time of LT was years. There were 19 male patients and 32 female patients. The leading indications for LT were primary biliary atresia (53%), fulminant hepatic failure (18%), and metabolic liver disease (18%). The mean time from LT to the diagnosis of DAIH was years. At the time of diagnosis, the median LFT results were as follows: alanine aminotransferase, 230 U/L; aspartate aminotransferase, 172 U/L; total bilirubin, 0.7 mg/dl; direct bilirubin, 0.2 mg/dl; and alkaline phosphatase, 295 U/L. The immunosuppressant levels were within the therapeutic range; the median tacrolimus trough level was 5.8 ng/ml, and the median cyclosporine trough level was 96 ng/ml. The majority of the patients had strongly positive autoimmune antibody titers: 61% were positive for anti-nuclear antibodies (median titer ¼ 1:640), 59% were positive for double-stranded DNA antibodies by an enzyme immunoassay (median titer ¼ 545 IU/ml), 27% were positive for smooth muscle actin antibodies (median titer ¼ 1:40), and 2% were positive for liver-kidney microsomal antibodies (median titer ¼ 1:80). Nearly half (43%) were positive for more than 1 type of autoimmune antibody. The majority of the patients (76%) showed resolution of abnormal results after treatment. The median LFT results 24 months after treatment were as follows: alanine aminotransferase, 35 U/L; aspartate aminotransferase, 35 U/L; total bilirubin, 0.6 mg/dl; direct bilirubin, 0.1 mg/dl; and alkaline phosphatase, 216 U/L. The mean follow-up time after the diagnosis of DAIH was 6.7 years. The 1-, 5-, and 10-year survival rates after the diagnosis of DAIH were 96%, 89%, and 76%, respectively, for grafts and 98%, 94%, and 81%, respectively, for patients. Histological Features of the Pretreatment Biopsy Samples Overall Pattern of Injury A summary of the pathological findings is shown in Table 2. Not unexpectedly, the predominant pattern of

3 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 PONGPAIBUL ET AL. 813 TABLE 1. Clinical Parameters of the 51 Patients With DAIH Parameter Value Age at transplantation (years)* Sex: male/female (n/n) 19/32 Indications for LT [n (%)] Biliary atresia 27 (53) Fulminant hepatic failure 9 (18) Metabolic liver disease 9 (18) Other 6 (12) Mean time from LT to DAIH (years)* After Treatment Parameter Diagnosis 3 Months 6 Months 12 Months 24 Months Median LFT results Alanine aminotransferase (U/L) Aspartate aminotransferase (U/L) Total bilirubin (mg/dl) Direct bilirubin (mg/dl) Alkaline phosphatase (U/L) Median immunosuppressant levels (ng/ml) Tacrolimus trough level (n ¼ 40) Cyclosporine trough level (n ¼ 11) Parameter Value Positive findings for autoantibodies [n (%)] Anti-nuclear antibodies (median titer ¼ 1:640) 31 (61) Double-stranded DNA antibodies according to an enzyme immunoassay (median titer ¼ 545 IU/ml) 30 (59) Smooth muscle actin antibodies (median titer ¼ 1:40) 14 (27) Liver-kidney microsomal antibodies (median titer ¼ 1:80) 1 (2) Multiple positive titers 22 (43) Treatment [n (%)] Continuation of calcineurin inhibitor and addition of mycophenolate mofetil and prednisone 27 (53) Continuation of calcineurin inhibitor and increase in or addition of prednisone 10 (20) Continuation of calcineurin inhibitor and addition of mycophenolate mofetil 9 (18) Other 5 (10) Follow-up time (years)* Survival after DAIH diagnosis at 1, 5, and 10 years (%) Graft 96, 89, and 76 Patient 98, 94, and 81 *The data are presented as means and standard deviations. injury was hepatitis in a majority of the biopsy samples (36/51 or 71%; Fig. 1). There was variable necroinflammatory activity: many cases (21/51) demonstrated mild activity (grade 2), 8 demonstrated moderate activity (grade 3), and 7 demonstrated severe activity (grade 4). Interestingly, the predominant pattern of injury in 7 of the 51 samples (14%) was acute rejection (the rejection activity indices were 3, 6, 6, 7, 8, 8, and 8), and in 2 samples (4%), the features were suggestive of chronic rejection. One biopsy sample had features suggestive of bile duct obstruction. All biopsy samples with a predominant pattern of acute rejection, chronic rejection, or bile duct obstruction also showed mild to moderate necroinflammatory activity (hepatitis). In 5 of the 51 samples (10%), only minimal nonspecific inflammatory infiltrates without necroinflammatory activity were present. When the predominant pattern of injury was hepatitis (n ¼ 36) and the location of the necroinflammatory activity (interface, lobular, or perivenular) was considered, lobular hepatitis was found to be the most common form of injury; it was present in 31 of the 36 biopsy samples (86%). Interestingly, lobular activity was unaccompanied by either interface or perivenular necroinflammatory activity in 7 of the 36 samples (19%). Lobular necroinflammatory activity was accompanied by interface activity in 7 samples (19%) and by interface and perivenular activity in 13 samples (36%). Lobular necroinflammatory was accompanied only by perivenular activity (no interface activity) in 4 samples (11%). Interface activity without

4 814 PONGPAIBUL ET AL. LIVER TRANSPLANTATION, July 2012 TABLE 2. Summary of the Patterns of Injury in Pretreatment Liver Biopsy Samples from Patients with DAIH Pattern of Injury Mild Moderate Severe Total Value Hepatitis 36 (71) Pure lobular (19) Pure interface (6) Pure perivenular (3) Lobular/interface (19) Lobular/perivenular (11) Interface/perivenular (6) Lobular/interface/perivenular (36) Total 21 (58) 8 (22) 7 (19) 36 Acute rejection (plus mild to moderate hepatitis) 7 (14) Chronic rejection (plus moderate hepatitis) 2 (4) Bile duct obstruction (plus mild hepatitis) 1 (2) Minimal nonspecific infiltrates (portal, lobular, and perivenular) 5 (10) Total 51 NOTE: The data are presented as numbers or as numbers and percentages. Subgroup Value lobular necroinflammatory activity was seen in only 2 samples (6%). In 1 biopsy sample, hepatitis manifested only as perivenular necroinflammatory activity. Finally and perhaps most interestingly, in 12 of the 36 samples (33%) with hepatitis as the predominant feature, there was no appreciable interface necroinflammatory activity. Plasma Cell Rich Infiltrates Plasma cell rich infiltrates (plasma cell density > 25% of inflammatory cells) were found in only 16 of the 51 biopsy samples (31%). In 13 of the biopsy samples with plasma cell rich infiltrates, the predominant pattern of injury was hepatitis; however, in 2 cases, the predominant pattern of injury was acute rejection. In 1 case, the predominant pattern was chronic rejection. When plasma cell rich infiltrates were present, they were almost always present in the portal areas (14/16; Fig. 1C); half of these were exclusively portal, and half were portal, lobular, and perivenular. Interestingly, in 2 cases, plasma cell rich infiltrates were exclusively perivenular. Central Vein Endotheliitis Central vein endotheliitis was an uncommon finding and was present in only 6 biopsy samples. In 3 samples, other features typical of acute rejection, such as mixed portal tract inflammation, portal vein endotheliitis, and bile duct injury, accompanied central vein endotheliitis. In these 3 samples, there was no significant perivenulitis, and a diagnosis of acute rejection was easily rendered. The other 3 biopsy samples with central vein endotheliitis, however, lacked these features of acute rejection and instead demonstrated moderate to severe necroinflammatory activity. Bile Duct Inflammation Not surprisingly, bile duct inflammation was prominent in 5 of the biopsy samples with other features of acute and chronic rejection or bile duct obstruction. However, mild duct inflammation without features of rejection or obstruction was also seen in 7 of the 36(19%) samples in which the predominant pattern of injury was hepatitis. Fibrosis The staging results were as follows: no fibrosis (stage 0) in 18 of the 51 biopsy samples (35%), minimal portal fibrosis (stage 1) in 6 samples (12%), mild portal and periportal fibrosis (stage 2) in 18 samples (35%), and bridging fibrosis (stage 3) in 9 samples (18%). None of the pretreatment biopsy samples showed cirrhosis. Histological Features of the First Posttreatment Biopsy Samples Overall Pattern of Injury A summary of the histological features of the first posttreatment biopsy samples is shown in Table 3. Again, the predominant pattern of injury in most samples was hepatitis. There were 16 patients whose first posttreatment biopsy was not performed until more than 3 months after the pretreatment biopsy (groups 3 and 4). Most members of this group (9/16) demonstrated mild or insignificant activity: 6 demonstrated mild necroinflammatory activity, and 3 demonstrated only minimal nonspecific infiltrates without necroinflammatory activity. In 2 of the 16 cases, there was moderate necroinflammatory activity, and in another 2 cases, there was severe necroinflammatory activity. In 1 of the 16 cases, features of acute rejection were

5 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 PONGPAIBUL ET AL. 815 Figure 1. Necroinflammatory activity. (A) Lobular inflammation and necroinflammatory activity. There were numerous mononuclear cells within the lobule along with scattered apoptotic and ballooning cells (hematoxylin and eosin, 200). (B) Portal inflammation with interface necroinflammatory activity. The interface between the portal area and the parenchyma was obscured by inflammatory cells extending into the parenchyma (hematoxylin and eosin, 400). (C) Clusters of plasma cells at the edge of a portal area (hematoxylin and eosin, 600). present, and not surprisingly, in 2 cases features of chronic rejection were present. In contrast, in one third (3/9) of posttreatment biopsies taken from patients whose first posttreatment biopsy was performed within 3 months of the pretreatment biopsy (groups 1 and 2), there was moderate to severe necroinflammatory activity. Again, not surprisingly, 4 of these 7 patients with moderate to severe activity underwent at least 1 more posttreatment biopsy. In 3 of the 4 samples, necroinflammatory activity was still present on the follow-up biopsy samples. Eight of the 15 posttreatment biopsy samples (53%) with features of hepatitis showed mild activity, 4 (27%) showed moderate activity, and 3 (20%) showed severe activity. In contrast, 36% of all pretreatment biopsy samples with features of hepatitis showed mild activity, 16% showed moderate activity, and 20% showed severe activity. Finally, 4 of the 25 patients (16%) who underwent posttreatment biopsy had minimal nonspecific infiltrates. Only 1 member of this group (4%) had minimal nonspecific infiltrates in the pretreatment biopsy sample. Plasma Cell Rich Infiltrates The frequency of plasma cell rich infiltrates (plasma cell density > 25% of inflammatory cells) was lower in the posttreatment biopsy samples 5 of 25 (20%) versus the pretreatment biopsy samples (31%). Plasma

6 816 PONGPAIBUL ET AL. LIVER TRANSPLANTATION, July 2012 TABLE 3. Comparison of Histological Features of Pretreatment Biopsy Samples and First Posttreatment Biopsy Samples From 25 Patients Who Underwent 1 or More Posttreatment Biopsy Procedures First Posttreatment Biopsy Pretreatment Biopsy (n ¼ 25) Total (n ¼ 25) Immediate (n ¼ 4) Early (n ¼ 5) Intermediate (n ¼ 9) Late (n ¼ 7) Hepatitis Mild 9 (50) 8 (53) Moderate 4 (22) 4 (27) Severe 5 (28) 3 (20) Total 18 (72) 15 (60) Acute rejection 5* (20) 2 (8) 0 1* 1* 0 Chronic rejection 1* (4) 3 (12) * Bile duct obstruction 0 1 (4) 0 1* 0 0 Minimal nonspecific infiltrate 1 (4) 4 (16) Fibrosis stage 0 8 (32) 8 (32) Fibrosis stage 1 2 (8) 5 (20) Fibrosis stage 2 8 (32) 6 (24) Fibrosis stage 3 7 (28) 4 (16) Fibrosis stage (8) NOTE: The data are presented as numbers or as numbers and percentages. *Plus moderate hepatitis. Plus mild hepatitis. Evaluated with the last biopsy. cell rich infiltrates were present only in posttreatment biopsy samples that showed at least moderate (grade 3) necroinflammatory activity, and they were mostly present in samples taken more than 3 months after treatment (3/4). Plasma cell infiltrates were seen in all locations (portal, lobular, and perivenular; data not shown). Fibrosis In Table 3, the frequencies of fibrosis stages 0 to 4 in pretreatment biopsy samples and first posttreatment biopsy samples are compared. When we considered the data in this manner, we found that the percentage of patients with low-stage fibrosis (0-2) was slightly greater in the posttreatment group (76%) versus the pretreatment group (72%). Likewise the percentage of patients with higher stage fibrosis (3 or 4) was slightly higher in the pretreatment group (28%) versus the posttreatment group (24%). These results may suggest that treatment results in a reduction of the stage. The results are, however, not statistically significant. In Table 4, the stages of pretreatment and posttreatment biopsy samples are directly compared for each of the 25 patients with a posttreatment biopsy sample. For the 9 patients with more than 1 posttreatment biopsy sample, the comparison is made between the pretreatment sample and the last posttreatment sample. In 5 of the 25 patients (20%), there was no change in the degree of fibrosis. In 10 of the 25 patients (40%), there was an increase in the stage (from stage 0 to stage 1 in 2, from stage 0 to stage 2 in 3, from stage 0 to stage 3 in 1, from stage 1 to stage 3 in 1, from stage 2 to stage 3 in 1, from stage 2 to 4 in 1, and from stage 3 to 4 in 1; Fig. 2). In another 10 patients (40%), there was a reduction in TABLE 4. Fibrosis Stages for Each Patient Who Underwent Posttreatment Biopsy According to Pretreatment and Posttreatment Biopsy Samples Decrease in stage (n 10) No change (n ¼ 5) Increase in stage (n ¼ 10) Pretreatment Biopsy Fibrosis Stage Last Posttreatment Biopsy NOTE: For patients with more than 1 posttreatment biopsy sample, data are provided for only the last one.

7 LIVER TRANSPLANTATION, Vol. 18, No. 7, 2012 PONGPAIBUL ET AL. 817 Figure 2. Cirrhosis in a posttreatment biopsy sample (Masson trichrome, 100). A pretreatment biopsy sample taken 9 days earlier showed bridging fibrosis. the stage (from stage 1 to stage 0 in 1, from stage 2 to stage 0 in 2, from stage 2 to stage 0 in 2, from stage 3 to stage 0 in 3, from stage 3 to stage 2 in 1, and from stage 3 to stage 1 in 1). Therefore, according to the results of the last posttreatment biopsy, 40% of the patients demonstrated an increase in the stage, and 60% experienced either no change or a reduction in the stage. DISCUSSION In this study, we have demonstrated that the most common pattern of injury in liver biopsy samples taken for late graft dysfunction associated with the development of autoimmune antibodies in pediatric LT patients was, not surprisingly, hepatitis. However, the degree of activity was often mild, and the lobule was more frequently involved than the portal/parenchymal interface or perivenular parenchyma. In fact, 19% of the cases demonstrated only lobular necroinflammatory activity. A number of biopsy samples (18%) had sufficient histological features to suggest acute or chronic rejection along with hepatitis-like necroinflammatory activity. We believe that this is an important point because necroinflammatory activity in biopsy samples with features of acute rejection is often taken to indicate severe acute rejection 11 and may not trigger a search for autoimmune antibodies. Therefore, necroinflammatory activity in biopsy samples with features of acute rejection may indicate autoimmune disease and merit the measurement of autoimmune antibodies. Prominent bile duct inflammation was not a common feature except in biopsy samples that demonstrated other features of acute rejection. Interestingly, only one-third of the samples with a predominant pattern of hepatitis (no acute cellular rejection, chronic rejection, or bile duct obstruction) had plasma cell rich infiltrates. Importantly, in one-third (33%) of the pretreatment biopsy samples in which the predominant pattern was hepatitis, there was no interface necroinflammatory activity. Furthermore, 10% of the biopsy samples obtained at the time of diagnosis had minimal nonspecific infiltrates without interface or lobular necroinflammatory activity. One may speculate that these patients did not have DAIH and developed autoimmune antibodies of uncertain significance. However, one may also ponder whether these patients, the patients without plasma cell infiltrates, and the patients with no interface activity had partially treated DAIH because they were already receiving immunosuppression. Indeed, perhaps it is unreasonable to expect the histology of DAIH to closely match the histology of autoimmune hepatitis because patients with DAIH are in effect already at least partially treated for autoimmune disease. The biopsy findings of DAIH are truly variable and not specific. A hepatitic pattern of injury has been described for acute rejection in the pediatric population, 13 for late-onset acute rejection, 14 and in protocol biopsy samples after pediatric LT. 15 In other words, serological tests for autoimmune antibodies should not await plasma cell rich infiltrates, interface hepatitis, and the classic features of autoimmune hepatitis in nontransplant patients. In our opinion, serological tests may be warranted if features of acute or chronic rejection predominate but are accompanied by interface and or lobular hepatitis. The majority of the patients in this study were treated for DAIH with mycophenolate mofetil and/or low-dose prednisone. Clinically, three-quarters of the patients showed significant improvements in their LFTs. Furthermore, the patient and graft survival rates for patients with DAIH do not appear to be different than the rates for all other pediatric LT recipients. 16 Although it is difficult to completely elaborate the histological changes occurring after treatment, posttreatment biopsy samples in this series demonstrated a reduction in the degree of hepatitis and plasma cell rich infiltrates. In some patients, there may have been a decrease in fibrosis. There was no progression of fibrosis in more than half of the patients. There were only 2 patients whose posttreatment biopsy samples demonstrated cirrhosis. Both patients underwent posttreatment biopsy only once (9 days and 8.5 months after treatment). One patient had stage 3 fibrosis on to the pretreatment biopsy sample, and the other had stage 2 fibrosis. Both patients had severe necroinflammatory activity on their pretreatment biopsy samples; importantly, they continued to have severe activity on their posttreatment biopsy samples. Furthermore, these patients were diagnosed with DAIH nearly 10 and 11 years after LT. It is quite possible that these 2 exceptional cases can be explained by unrecognized and untreated DAIH, which resulted in severe hepatitis with advanced fibrosis that was poorly responsive to

8 818 PONGPAIBUL ET AL. LIVER TRANSPLANTATION, July 2012 therapy. It is also conceivable that cirrhosis was present at the time of the first biopsy but could not be appreciated because of sampling. The same, of course, may be said for all staging scores. These findings are similar to those of a previous report by Miyagawa-Hayashino et al. 5 In general, the outcome of patients with unrecognized or untreated DAIH is very poor. 17 In conclusion, it is not possible to confidently render a diagnosis of DAIH by histology alone; perhaps more importantly, it is not possible to exclude DAIH by histology alone. Not all biopsy samples show plasma cell rich infiltrates, many have no interface necroinflammatory activity, some show features of acute rejection, and in approximately 10%, there are only minimal nonspecific infiltrates with no necroinflammatory activity. Of course, other causes of hepatitis and late graft dysfunction must be ruled out, and histological findings must be used in conjunction with clinical findings and serological studies in order to make a diagnosis of DAIH. Pathologists should be aware of the existence of DAIH and raise the possibility of this diagnosis, and they should recommend a correlation with autoimmune antibodies when there is even a mild degree of hepatitic injury. The treatment of DAIH appears to yield good clinical and histological results with respect to inflammatory activity and fibrosis in most patients. The development of less invasive tools for monitoring disease activity and identifying patients at risk for developing DAIH would be helpful. REFERENCES 1. Kerkar N, Hadzić N, Davies ET, Portmann B, Donaldson PT, Rela M, et al. De-novo autoimmune hepatitis after liver transplantation. Lancet 1998;351: Hernandez HM, Kovarik P, Whitington PF, Alonso EM. Autoimmune hepatitis as a late complication of liver transplantation. J Pediatr Gastroenterol Nutr 2001;32: Andries S, Casamayou L, Sempoux C, Burlet M, Reding R, Bernard Otte J, et al. Posttransplant immune hepatitis in pediatric liver transplant recipients: incidence and maintenance therapy with azathioprine. Transplantation 2001;72: Petz W, Sonzogni A, Bertani A, Spada M, Lucianetti A, Colledan M, Gridelli B. A cause of late graft dysfunction after pediatric liver transplantation: de novo autoimmune hepatitis. Transplant Proc 2002;34: Miyagawa-Hayashino A, Haga H, Egawa H, Hayashino Y, Sakurai T, Minamiguchi S, et al. Outcome and risk factors of de novo autoimmune hepatitis in living-donor liver transplantation. Transplantation 2004;78: Spada M, Bertani A, Sonzogni A, Petz W, Riva S, Torre G, et al. A cause of late graft dysfunction after liver transplantation in children: de-novo autoimmune hepatitis. Transplant Proc 2001;33: Gupta P, Hart J, Millis JM, Cronin D, Brady L. De novo hepatitis with autoimmune antibodies and atypical histology: a rare cause of late graft dysfunction after pediatric liver transplantation. Transplantation 2001;71(5): Venick RS, McDiarmid SV, Farmer DG, Gornbein J, Martin MG, Vargas JH, et al. Rejection and steroid dependence: unique risk factors in the development of pediatric posttransplant de novo autoimmune hepatitis. Am J Transplant 2007;7: Hadzic N, Srinivasan R, Rela M, Heaton N, Vergani D, Mieli-Vergani G. De novo autoimmune hepatitis after liver transplantation in children: a single centre experience [abstract]. Hepatology 2003;38(suppl 4):202A. 10. Richter A, Grabhorn E, Helmke K, Manns MP, Ganschow R, Burdelski M. Clinical relevance of autoantibodies after pediatric liver transplantation. Clin Transplant 2007;21: Batts KP, Ludwig J. Chronic hepatitis. An update on terminology and reporting. Am J Surg Pathol 1995;19: Banff schema for grading liver allograft rejection: an international consensus document. Hepatology 1997;25: Abraham SC, Freese DK, Ishitani MB, Krasinskas AM, Wu TT. Significance of central perivenulitis in pediatric liver transplantation. Am J Surg Pathol 2008;32: Banff Working Group, Demetris AJ, Adeyi O, Bellamy CO, Clouston A, Charlotte F, et al. Liver biopsy interpretation for causes of late liver allograft dysfunction. Hepatology 2006;44: Evans HM, Kelly DA, McKiernan PJ, Hübscher S. Progressive histological damage in liver allografts following pediatric liver transplantation. Hepatology 2006;43: Farmer DG, Venick RS, McDiarmid SV, Ghobrial RM, Gordon SA, Yersiz H, et al. Predictors of outcomes after pediatric liver transplantation: an analysis of more than 800 cases performed at a single institution. J Am Coll Surg 2007;204: Salcedo M, Vaquero J, Ba~nares R, Rodríguez-Mahou M, Alvarez E, Vicario JL, et al. Response to steroids in de novo autoimmune hepatitis after liver transplantation. Hepatology 2002;35: