Silent Presentation of Veno-occlusive Disease After Liver Transplantation as Part of the Process of Cellular Rejection With Endothelial Predilection

Transcription

1 Silent Presentation of Veno-occlusive Disease After Liver Transplantation as Part of the Process of Cellular Rejection With Endothelial Predilection MYLÈNE SEBAGH, 1 MARYLINE DEBETTE, 2 DIDIER SAMUEL, 2 JEAN-FRANCOIS EMILE, 1 BRUNO FALISSARD, 3 VALERIE CAILLIEZ, 2 DANIEL SHOUVAL, 2 HENRI BISMUTH, 2 AND MICHEL REYNÈS 1 Hemorrhagic centrilobular necrosis and fibrous stenosis of hepatic venules, suggesting veno-occlusive disease (VOD) have rarely been observed after orthotopic liver transplantation (OLT). The aim of this study was to determine the prevalence of this syndrome after OLT in relation to the course with particular reference to acute rejection and to azathioprine administration. VOD was identified in 19 of 1,023 patients transplanted over a 9-year period. VOD occurred at a median of 30 days posttransplantation, without clear cut clinical evidence for hepatic vein outlet obstruction. Seventeen of the 19 patients had an episode of acute rejection before or at the time of VOD. These episodes were compared with that of patients without VOD. In patients with VOD, portal inflammation and endothelialitis were enhanced (P.014 and P.048) and endothelialitis was also higher than bile duct damage (P.03). The incidence of a centrilobular endothelialitis for both groups was not different although an increased trend was observed in the study group (64% vs. 46%; P.18). The incidence of persistent rejection was similar between both groups (47% vs. 41%). The incidence of chronic rejection was higher in the study group (29% vs. 10%; P.04). All patients with VOD received azathioprine as part of immunosuppressive regimen. Despite azathioprine withdrawal, zone 3 changes persisted in 57% of patients. In conclusion, the incidence of VOD was 1.9% after OLT. The association of prominent endothelial involvement and VOD with acute rejection in most cases suggests an immunological phenomenon. (HEPA- TOLOGY 1999;30: ) Veno-occlusive disease (VOD) of the liver was first described by Bras et al. 1 The clinical diagnosis of VOD was based on the triad of jaundice, painful hepatomegaly, and ascites/weight gain. It was confirmed by histological findings including fibrous obliteration of small hepatic veins by Abbreviations: VOD, veno-occlusive disease; OLT, orthotopic liver transplantation; RAI, rejection activity index; VBDS, vanishing bile duct syndrome. From 1 Service d Anatomie Pathologique, 2 Centre Hépato-biliaire, and 3 Unité de Santé Publique, Hôpital Paul Brousse, Villejuif, UPRES virus hépatotropes et cancers, Université Paris-Sud, France. Received February 25, 1999; accepted August 13, Address reprint requests to: Mylène Sebagh, M.D., Laboratoire d Anatomie Pathologique, Hôpital Paul Brousse, 14 avenue Paul Vaillant Couturier, Villejuif Cedex, France. mylene.sebagh@pbr.ap-hop-paris.fr; fax: (33) Copyright 1999 by the American Association for the Study of Liver Diseases /99/ $3.00/ connective tissue and centrilobular hemorrhagic necrosis. In the transplantation setting, VOD is responsible for liver dysfunction after bone marrow 2-4 and kidney transplantation VOD after bone marrow transplantation is thought to be the consequence of hepatoxicity, arising mainly from pretransplantation conditioning chemoirradiation therapy given for hematologic and other malignancies. On the other hand, VOD after kidney transplantation is usually thought to be related to azathioprine therapy. Surprisingly, there have been few reports to date of this condition developing in liver allografts, although these recipients often receive azathioprine as part of an immunosuppressive regimen. The present retrospective study was undertaken to assess the incidence of the VOD syndrome after orthotopic liver transplantation (OLT) at our institution in relation to the posttransplantation course with particular reference to acute rejection and to azathioprine therapy. PATIENTS AND METHODS The files of the Liver Transplantation Database of the Department of Pathology were searched to identify all patients with VOD between February 1988 and November During that time, a total of 19 patients was identified among 1,023 consecutive patients that underwent 1,145 liver transplantations at the Hepatobiliary Center of Paul Brousse Hospital, Villejuif, France. Immunosuppression protocols varied over the time period. Most patients were treated with triple-drug immunoprophylaxis including cyclosporine, corticosteroids and azathioprine. The other patients were treated with Tacrolimus in a dual- or triple-drug regimen. Rejection episodes were confirmed by histological assessment before treatment. They were treated with 1 to 3 pulses of methylprednisolone. A 10-day course of Orthoklone OKT3 (Ortho, Raritan, NJ) and/or Tacrolimus was administered in patients with steroid-resistant rejection. Clinical and biochemical parameters were obtained by review of records and included presence of jaundice, hepatomegaly, ascites, serum values of total bilirubin, alkaline phosphatase, gamma glutamyl transpeptidase and transaminases, and changes in immunosuppressive regimen. In all 19 patients, ultrasound Doppler examination excluded hepatic vein obstruction (Budd-Chiari syndrome), which may produce histological changes similar to those of VOD. The liver tissue investigated in this study was obtained because of a clinical indication for routine histology. Ethical considerations were thus satisfied. Protocol liver biopsies were routinely performed at 1, 2, 3, 5, and 10 years after OLT and when clinical and biochemical abnormalities required it. Liver specimens stained with hematoxylin-eosin safran and picrosirius were reviewed by 2 pathologists (M.R. and M.S.) blinded to clinical findings. This report was concerned with all posttransplantation liver specimens of the 19

2 HEPATOLOGY Vol. 30, No. 5, 1999 SEBAGH ET AL patients with histological proven VOD. VOD was histologically diagnosed if there was total or subtotal fibrous obliteration of hepatic veins by connective tissue and centrilobular hemorrhagic necrosis. Both criteria were required for the diagnosis. Other vascular lesions, including sinusoidal dilatation, sinusoidal congestion, peliosis hepatis, and nodular regenerative hyperplasia, were recorded. Special attention had turned to the presence of rejection. The diagnosis of acute rejection was based on the classical histological triad including portal inflammation, bile duct inflammation/ damage, and inflammation of portal or terminal hepatic venules. Acute rejection was considered persistent when it was present in successive biopsies performed at intervals of more than a week and in association with impaired liver function tests. Chronic rejection was defined as the presence of a vanishing bile duct syndrome (VBDS) (disappearance of interlobular bile ducts) in greater than 50% of the portal tracts and/or of a foamy cell arteriopathy seen in the graft at the time of retransplantation. Grading of Rejection. At the time of histological diagnosis, acute rejection had been graded according to an international scheme recently published. 15 In summary, portal inflammation, bile duct inflammation/damage, and inflammation of portal or terminal hepatic venules in that order are each graded semiquantitatively on a scale of 0 (absent) to 3 (severe). The individual scores are then added together to arrive at a final rejection activity index (RAI). For the purpose of the present study, modifications of this system were made to provide further information regarding the relative severity of portal and hepatic venular endothelial inflammation. Severity of endothelialitis was semiquantitatively classified as mild, moderate, or severe on a scale of 0 to 3, by estimating the proportion of veins that were inflamed and the severity of the inflammatory infiltrate. This was followed by reporting of 2 endothelial scores, one for portal veins and the other for hepatic veins. In contrast with the Banff system, severe endothelialitis scored 3 did not account for the presence of perivenular hepatocyte necrosis. Control Group. Our study patients that experienced an acute rejection before or at the time of VOD were compared with a control group of patients with acute rejection, but without evolution towards VOD. The number of control cases was determined at 4 times the number of study cases (a higher number of control cases does not increase the power of the comparison). For inclusion in the control group, patients experienced an episode of acute rejection occurring during a period similar to our study patients. They were followed up for at least 1 year and underwent at least 1 liver biopsy after the diagnosis of acute rejection. The RAI had to be superior or equivalent to 3 (i.e., borderline or consistent with the Banff system). We then compared data from the review of liver biopsies, particularly regarding details of the grading of the acute rejection (individual scores and RAI in the Banff system, 2 additional endothelial scores) and of its outcome (persistence, VBDS, and chronic rejection) Statistical Methods. Differences were analyzed by means of Mann- Whitney test for semiquantitative variables and 2 test for qualitative values. Intervals between the following events were recorded: (1) liver transplantation to diagnosis of histological VOD; (2) diagnosis of an eventual preceding acute rejection to diagnosis of histological VOD. RESULTS Demographic and laboratory data are summarized in Table 1. The 19 patients (8 men and 11 women) underwent liver transplantation for fulminant liver failure related to virus, drug, or unidentified etiology (n 6); alcohol-induced (n 4); related to virus infection (n 5); TABLE 1. Demographic and Laboratory Data of the 19 Patients With VOD After Liver Transplantation Liver Function Tests at the Time of VOD Patient Age at OLT/Sex Indication for OLT Initial Therapy Symptoms at the Time of VOD Bilirubin N F 17 mol/l AST N F 35 ALT N F 43 PT % ALP N F 180 GGT N F /F Alcohol-induced C Ascites, jaundice /F Amyloidosis C Ascites /M Alcohol-induced C /F HBV-FH C /M FH T , /F HBV- C 364 4,764 2, /F FH C Pleural effusion 200 2,106 1, /F HBV-FH C ,200 1, /F PBC C /M HBV- C Triad* /F HBV-FH C-chemotherapy /M Alcohol-induced T /F PBC C 600 1, /M HBV- C /F FH C-Rifadin (drug pyrilen) 16 49/M HCV-/HCC C-chemotherapy 1, , /F PBC C Ascites /M Alcohol-induced C /M HCV- C Abbreviations: AST, aspartate aminotransferase; ALT, alanine aminotransferase; PT, prothrombin time; ALP, alkaline phosphatase; GGT, gamma glutamyl transpeptidase; HBV, hepatitis B virus; FH, fulminant hepatitis; HCV, hepatitis C virus; HCC, hepatocellular carcinoma; PBC, primary biliary ; C, cyclosporine azathioprine steroids; T, Tacrolimus azathioprine steroids. *Ascites, painful hepatomegaly, jaundice.

3 1146 SEBAGH ET AL. HEPATOLOGY November 1999 primary biliary (n 3); and amyloidotic neuropathy (n 1). The median age was 44 years (range, years) at the time of transplantation. All patients received an ABO-identical or -compatible graft. One of these grafts showed evidence of severe macrovesicular steatosis at the time of transplantation for a fulminant liver failure (patient 4). Patients were treated with azathioprine, steroids, and either cyclosporine in 17 patients or Tacrolimus in 2 patients (patients 5 and 12). Azathioprine was administered at a daily dose of 1 to 2 mg/kg. At the time of histological diagnosis, VOD was never suggested on the basis of clinical grounds in any of the 19 patients, although 1 (patient 10) had a typical clinical presentation (ascites, painful hepatomegaly, and jaundice) obscured by other complications (development of a gut lymphoproliferative disorder 6 years after the transplantation followed by a severe episode of acute rejection 10 days before the onset of VOD). Four patients had nonspecific signs independently or in association, including ascites, pleural effusion, and jaundice. Fourteen patients were completely asymptomatic. In these asymptomatic patients, the liver biopsy was indicated by investigation for abnormal liver function tests. Liver function tests were abnormal in all patients. The mean value was 386 µmol/l (range, 14-1,068 µmol/l) for serum bilirubin, 312 (range, 81-1,200 ) for alkaline phosphatase, 649 (range, 17-4,764 ) for aspartate transaminase, and 471 (range, 26-2,304 ) for alanine transaminase. Specimens with VOD were obtained at a median of 30 days after transplantation (range, 10 to 3,972 days) (Table 2). VOD was first noted in liver biopsy specimens from 14 patients and in failed allografts from 5 patients at the time of retransplantation (patients 6, 11, 16, 17, and 19). Damaged veins were present frequently at the same location as the centrilobular area necrosis. The severity of lesions (number of veins and centrilobular areas involved, degree of venous occlusion, and extent of hemorrhagic necrosis) was variable (Fig. 1). In addition to the above findings, nonspecific changes commonly attributed to azathioprine-liver disease were present in 3 patients (16%): peliosis hepatis in 1 patient (patient 12) and nodular transformation in 2 (patients 11 and 14). The review of the 42 earlier biopsy results from 15 of the 19 patients (average, 2.8; 1 to 8 per patient) showed centrilobular congestion in 3 (patients 2, 9, and 18), sinusoidal dilatation in 1 (patient 9), and nodular transformation in 1 (patient 11). A common finding was the presence of acute rejection (Table 2; Fig. 2). Rejection was diagnosed in 13 of 15 patients before the development of VOD (87%) and in 11 of the 19 patients at the time of VOD (58%). Only 2 patients (patients 4 and 11) had no history of acute rejection at any time. The first rejection episode (previous or concomitant to VOD) occurred at a median of 15 days after transplantation (range, 6 to 3,972 days). Only in 4 cases, it developed later at 117, 392, 2,232, and 3,972 days after transplantation. For the 3 former, rejection occurred after discontinuation of immunosuppressive regimen in 2 patients (because of self decision in patient 1 and of complication of chemotherapy for prevention of recurrent hepatocellular carcinoma in patient 16) and a reduction in immunosuppression associated with chemother- TABLE 2. Relationship Between Acute Rejection, Azathioprine Duration, and VOD After Liver Transplantation Patient Onset Score (Banff grade*) CL Acute Rejection Endothelial Scores VP/CL Persistence VBDS Onset VOD Persistence Interval Acute Rejection/ VOD Aza Duration /0 No 423 Yes /2 No 30 Yes /2 No 18 No No /0 Yes 11 Yes /1 Yes 253 NA /0 No 25 No /2 Yes 125 Yes /1 Yes 78 No , /3 No 2,242 Yes ,052 NA 1, /0 No 201 No /3 Yes 18 Yes , /2 No 3,972 Yes 0 3, /0 No 14 No /0 Yes 145 NA /2 Yes 12 NA /1 Yes 10 Yes /2 NA 15 NA 0 15 Median Mean Abbreviations: CL, centrilobular endothelialitis; VP, portal endothelialitis; Aza, azathioprine; NA, not applicable. *In this system, the first score refers to portal infiltrate, the second score to bile duct damage, and the third score to venous endothelial inflammation. The individual scores were then added to produce a final RAI. As assessed in the Patients and Methods section, we expanded the Banff system by reporting 2 semiquantitative endothelial scores. The first one (VP) refers to the portal venous endothelial inflammation, the second (CL) refers to hepatic venous endothelial inflammation. Patient in whom VOD was first detected at retransplantation. Not evaluable because the diagnosis of VOD and/or acute rejection was made on the graft at the time of retransplantation.

4 HEPATOLOGY Vol. 30, No. 5, 1999 SEBAGH ET AL FIG. 1. VOD of the liver occurring at 18 days after transplantation (patient 13). The centrilobular vein shows an intimal thickening by red cells. Its lumen is still free. (A) Perivenular hepatocyte necrosis is present (hematoxylin-eosin-safran). (B) Another section of a centrilobular vein is totally occluded by fibrillar material within the same liver biopsy specimen (picrosirius). apy in 1 patient because of lymphoproliferative disorder (patient 10). For the remaining patient (patient 14), no cause was found. The median interval between the first rejection episode and histological VOD was 8 days (range, 0 to 238 days). There were 4 patients (patients 6, 8, 9, and 12) in whom the interval between the 2 conditions was greater than 60 days. In at least 3 of them, there was a history of successive rejection episodes and VOD was not identified through a prospective protocol but through a retrospective protocol. This may in part explain the variation in observation. For the remaining patient (patient 12), there was a single rejection episode occurring in the early posttransplantation period and rapidly responsive to corticoids, suggesting that the 2 conditions were independent. To validate our observations, we chose to compare these episodes of acute rejection (17 cases) with those of a control group without evolution towards VOD (75 cases) in terms of histological signs, severity of histological score, and outcome: (1) Histological signs that conducted to the diagnosis of acute rejection were identical in both groups. (2) In the Banff system, the portal inflammation (first score) and the venous endothelial inflammation (third score) were significantly greater in the study group (P.014 and P.048, respectively), whereas the bile duct inflammation (second score) and the RAI were comparable (P.559 and P.114, respectively). The severity of the venous endothelial inflammation was significantly greater than that of the bile duct inflammation in the study group (P.03). The incidence of a centrilobular endothelialitis for the 2 groups was not significantly different although the trend was increased in the study group (64% vs. 46%; P.18). There was no statistical difference in the severity of portal endothelial inflammation (P.075), in the severity of hepatic venular endothelial FIG. 2. Acute rejection occurring at 10 days after transplantation, 8 days before the histological diagnosis of VOD in the same patient as seen in Fig. 1. (A) Acute rejection involves a portal tract on the left and a centrilobular vein on the right (hematoxylin-eosin-safran). (B) At higher magnification, centrilobular endothelialitis is severe (graded 3). This centrilobular vein is nearly occluded by the rejection infiltrate. However, it differs from VOD where the lumen is occluded by connective tissue without inflammation.

5 1148 SEBAGH ET AL. HEPATOLOGY November 1999 inflammation (P.075), and in their relative severity (P.626) between the 2 groups. (3) In terms of outcome, there was no statistical difference in the incidence of persistent rejection (47% vs. 41%; P.66) and of a VBDS (29% vs. 26%; P.81) between the 2 groups. On the contrary, the incidence of chronic rejection (VBDS 50%) was significantly greater in the study group (29% vs. 10%; P.04). Follow-up liver specimens showed a variable outcome in regard to azathioprine administration (Table 2). All 19 patients initially received azathioprine after transplantation. Azathioprine was withdrawn from 5 patients (patients 6, 8, 9, 10, and 14) at the time of conversion to Tacrolimus (as rescue therapy for ductopenic or intractable acute rejection) and was systematically withdrawn from all the other patients at the time of histological diagnosis of VOD. Furthermore, 1 of the patients (patient 14) underwent a transjugular intrahepatic portosystemic shunt. It should be emphasized that the histological outcome is not available for the 5 patients in whom the diagnosis of VOD was made in the explant at the time of retransplantation. For the 14 others, the histological outcome (available in subsequent liver biopsy specimens, autopsy material, and explants at the time of retransplantation) displayed improvement in only 6 patients (patients 3, 4, 7, 9, 12, and 15) and persistence/aggravation in the 8 others. In patients with persistence/aggravation, 6 had a hemorrhagic lesion initially confined to centrilobular areas that progressed to a pan-acinar distribution with time. Two patients (patients 10 and 14) developed progressive zone 3 fibrosis. The outcome of the patients with histological VOD is shown in Table 3. At the time of reporting, 12 patients died, 4 without retransplantation and 8 soon after; only 7 patients are alive, 4 without retransplantation and 3 after. Considering death, VOD was fully implicated in 4 patients (those dying without retransplantation [patients 1, 5, 10, and 13]) in whom hepatic failure was the primary cause of death. In contrast, VOD was not directly implicated in those dying after retransplantation (patients 2, 6, 8, 9, and 16-19) in whom death was from multifactorial origin. Of those 11 patients who were retransplanted, graft failure was mainly related to chronic rejection in 3 (patients 8, 9, and 14), herpetic hepatitis in 1 (patient 2), massive hemorrhagic necrosis in 4 (patients 16-19), primary dysfunction (likely caused by a severe macrovesicular steatosis) in 1 (patient 4), VOD in 1 (patient 6), and nodular regenerative hyperplasia in 1 (patient 11). Like that, considering retransplantation, VOD may be fully implicated in 1 retransplanted patient (patient 6) and partially in 8 retransplanted patients when VOD was present within their explant but seemed to be a secondary phenomenon. In contrast, VOD was not implicated in 2 retransplanted patients in whom VOD disappeared (patients 4 and 9). At last follow-up, no patient had developed evidence of recurrence of VOD. DISCUSSION The aims of this study were to assess the incidence of VOD after liver transplantation and the clinical outcome with regard to acute rejection and azathioprine therapy (as they may relate to the pathogenesis of VOD). We have shown that the incidence of VOD after OLT is low, occurring only in 1.9% of about 1,000 consecutive patients over a 9-year period. However, the outcome of our patients with VOD was poor. We recognize a strong association with acute rejection, while the role of azathioprine is more questionable. The first result of the present study suggests that despite the rare and anecdotal reports of VOD after liver transplantation in the literature, where only 16 patients are described (Table 4), this complication does occur in almost 2% of our patients. VOD has been noted relatively early between 5 and 133 days after transplantation. VOD is also recognized as a complication of renal transplantation, 5-10 often described in TABLE 3. Outcome of Patients With VOD and Main Pathological Findings in the Last Follow-up Material Main Pathological Findings Nonretransplanted Patients Retransplanted Patients Patient Outcome Cause of Death Postmortem Material Last Follow-up Liver Biopsy Explant 1 D (444) Liver failure VOD, CR 2 retx (38) then D (11 post retx) Sepsis Herpetic hepatitis, VOD 3 A N 4 retx (32) Severe steatosis 5 D (137) Liver failure VOD 6* retx (253) then D (1 post retx) Sepsis VOD 7 A N 8 retx (214) then D (90 post retx) Sepsis CR, VOD 9 retx (122) then D (3 post retx) Brain hemorrhage CR 10 D (2270) Liver failure VOD, CR 11* retx (1052) NRH, VOD 12 A N 13 D (20) Liver failure VOD 14 retx (4272) CR, VOD 15 A N 16* retx (145) then D (17 post retx) Sepsis MHN, VOD 17* retx (12) then D (81 post retx) Sepsis MHN, VOD 18 retx (13) then D (18 post retx) Brain hemorrhage MHN, VOD 19* retx (15) then D (9 post retx) Sepsis MHN, VOD Abbreviations: D, dead; retx, retransplanted; A, alive; CR, chronic rejection; N, (sub)normal histology; NRH, nodular regenerative hyperplasia; MHN, massive hemorrhagic necrosis. *Patient in whom VOD was first detected at retransplantation.

6 HEPATOLOGY Vol. 30, No. 5, 1999 SEBAGH ET AL TABLE 4. Characteristics of the Patients With VOD After Liver Transplantation in the Literature VOD Reference Patient Age/Sex Indication for OLT Symptoms Acute Rejection Onset Onset Persistence Aza Duration Sterneck et al. 1 58/F HCV- Jaundice /M HCV- Jaundice, ascites Mion et al. 1 47/M HBV-, HCV- Ascites, hepatomegaly (3) Dhillon et al. 1 18/M PSC NA (5) /M Biliary atresia NA (5) 5 Throughout 3 55/F HCC NA (6) 6 Throughout 4 23/M PSC NA (7) 7 Throughout 5 61/M HBV-/HCC NA (19) 19 Throughout 6 30/F HCV- NA 72 Throughout 7 48/M HBV- NA 133 Throughout 8 43/F HCV-/HCC NA /M HBV-, HDV- NA Hübscher et al. 1 59/F Primary biliary NA (6) 10 NA NA 2 40/M Cryptogenic NA 5 NA NA 3 11/F Cryptogenic NA (6) 20 NA NA 4 53/F Primary biliary NA (6) 15 NA NA Abbreviations: Aza, Azathioprine; HCV, hepatitis C virus; HBV, hepatitis B virus; PSC, primary sclerosing cholangitis; HCC, hepatocellular carcinoma; HDV, hepatitis delta virus; NA, not available. association with nodular regenerative hyperplasia and peliosis hepatis and occurring relatively late at 1 or 2 years after transplantation. On the other hand, VOD is a more common complication after bone marrow transplantation. 2-4 Its frequency varies greatly, likely because of differences in the intensity of conditioning regimens, individual pharmakokinetics and metabolism of drugs used, type of marrow donor, and presence or absence of risk factors. 3 It usually occurs early in the first weeks after transplantation. It is of interest that in the present series the diagnosis of VOD was made on histological grounds and was never suspected on clinical grounds. Five of our patients (26%) were symptomatic with only one of them presenting the typical clinical triad with evidence of hepatic vein outlet obstruction. All patients had abnormal liver function tests with sometimes an unusual pattern of liver injury. Because the features used to diagnose VOD, namely ascites, painful hepatomegaly, jaundice, and hyperbilirubinemia are not specific in the OLT setting, we believe that the diagnosis of VOD after OLT requires a histological assessment. In contrast, those features developing in the first weeks after bone marrow transplantation are so highly suggestive of the diagnosis that, among bone marrow transplantation centers, there is controversy about whether histological confirmation is required for an accurate diagnosis of VOD (after the other causes of liver disease are ruled out). 4 The histological diagnosis of VOD may at times be difficult and hazardous. In our study, we required the presence of the 2 lesions (i.e., obliterative venous lesion and centrilobular hemorrhagic necrosis). These may occur within the same specimen but may often be in separate foci. Obliterative venous lesion, although considered the main feature of VOD, is most difficult to recognize. Liver biopsy may have a large false-negative sampling error because terminal hepatic veins do not show uniform involvement with VOD. The number of veins involved is often low. The venous lumen may be masked by the hemorrhagic necrosis. Thus, problematic cases must have additional sections and reticulin stain. In contrast, centrilobular necrosis is a familiar feature after liver transplantation and may be caused by other conditions than VOD. 16 Our definition is somewhat more strict than that used in bone marrow transplantation centers in which centrilobular necrosis and/or centrilobular hemorrhage alone is sufficient to make the diagnosis in the presence of abnormal liver function tests and/or typical clinical signs. Sterneck et al. 11 decided that only centrilobular hemorrhagic necrosis would be used as a diagnostic feature because they recognized that early cases may not show obliterative venous lesion. The most significant finding of our study is that VOD is strongly associated with acute rejection. It raises questions to its pathogenesis, via endothelialitis-induced damage to the centrilobular wall ,16-18 We believe that this association is not an incidental finding for several reasons: First, the association rate was particularly high in our study (89%) and through the literature (56%, 9 of 16 cases) Second, the interval between the 2 events was short (median, 8 days). Third, the comparison of these episodes with control cases (acute rejection without evolution towards VOD) in terms of severity of histological scores (in the Banff system) showed that the former had an endothelial predilection: The venous endothelial inflammation was significantly enhanced compared with the bile duct inflammation in the study group. The presence of centrilobular endothelialitis tended to be more frequent in the study group. Thus, the presence of centrilobular endothelialitis should alert the pathologist to consider irreversible rejection or other serious complication such as VOD. Although severe endothelial score, as defined in the recent Banff consensus scheme, 15 requires the presence of centrilobular endothelialitis in association with a centrilobular necrosis, we wonder whether the sole presence of centrilobular endothelialitis favors a diagnosis of severe acute rejection that requires additional immunosuppression. Because of a combination of VOD and massive hemorrhagic necrosis, the possibility of immune mediated-humoral mechanisms as a cause for these lesions may be envisaged in 4 of our patients (patients 16-19). However, they all received an ABO-compatible and cross-match negative graft. Three of them had acute rejection. One of them had a long delay

7 1150 SEBAGH ET AL. HEPATOLOGY November 1999 between transplantation and diagnosis of massive hemorrhagic necrosis (145 days). For these reasons, we feel that the likelihood of humoral rejection in these cases is very low. A more likely explanation is that veno-occlusive lesions are a secondary event occurring as a result of perivenular ischemic injury. 14 Other factors may be incriminated. Chemotherapy prescribed in 2 patients (for treatment of a gut lymphoproliferative disorder in one and for prevention of a recurrent hepatocellular carcinoma in the other) may be a risk factor, but its role is unlikely because of the small number of courses before the onset of VOD (4 and 2 courses, respectively) and because of the absence of radiation therapy. Azathioprine administration as part of the immunosuppressive regimen in all patients may be another cofactor involved in development of VOD. However, the duration of azathioprine exposure was short (median, 25 days). Histological features commonly attributed to azathioprine hepatotoxicity (i.e., sinusoidal congestion and dilatation, peliosis hepatis, and nodular regenerative hyperplasia) were seen in only 25% of our patients. Furthermore, histological features of VOD were reversed after azathioprine discontinuation in only 42% of cases (6 of 14). The nonreversibility of the hepatic injury after the drug discontinuation is thought to reflect a fully developed VOD supported by the presence of fibrosis, which occludes the venous lumen (by definition in our study), in contrast to the cases of VOD reported by Sterneck et al. 11 characterized by the absence of fibrosis. In no patient of our study was the relationship between azathioprine administration and the hepatic injury proven by drug withdrawal and repeat challenge. However, of those who received another liver transplant and are alive (3 patients), it is interesting to note that none had recurrence of VOD despite the restitution of azathioprine. It would be useful to examine the incidence of VOD in centers not using azathioprine as part of the immunosuppressive regimen. In our series, patients with histological VOD seem to have a poor outcome, as assessed by the number of deaths and retransplantations. However, only 4 patients (those dying without retransplantation) died of VOD. Because of its association with other pathological processes in most of them and because of the good outcome of patients with VOD previously reported, no correlation may be established between the presence of VOD and the outcome. In conclusion, VOD may occur in hepatic allografts but may escape detection unless the patients undergo a liver biopsy. The pathogenesis remains unknown but the strong association with acute rejection suggests that endothelialitisinduced damage to the centrilobular venous wall may play an important role. We were unable to determine with certainty the role of azathioprine toxicity. Thus, the policy toward cessation of azathioprine may warrant further prospective investigation. REFERENCES 1. Bras G, Jellife DB, Stuart KL. Veno-occlusive disease of liver with non portal type of, occurring in Jamaica. Arch Pathol 1957;57: Bearman SC. The syndrome of veno-occlusive disease after marrow transplantation. Blood 1995;85: Sculman HM, Fisher LB, Schoch HG, Henne KW, McDonald GB. Veno-occlusive disease of the liver following marrow transplantation: histological correlates of clinical signs and symptoms. HEPATOLOGY 1994;19: Jones RJ, Lee KSK, Beschorner WE, Vogel VG, Grochow LB, Braine HG, Vogelsang GB, et al. Veno-occlusive disease of the liver following bone marrow transplantation. Transplantation 1987;44: Weitz H, Gokel JM, Loeschke K, Possinger K, Eder M. Veno-occlusive disease of the liver in patients receiving immunosuppressive therapy. Virchows Arch Pathol Anat 1982;1: Marubbio AT, Danielson B. Hepatic veno-occlusive disease in a renal transplant patient receiving azathioprine. Gastroenterology 1975;69: Eisenhauer T, Hartman H, Rumpf KW, Helchen U, Scheler F, Creutzfeldt W. Favourable outcome of hepatic veno-occlusive disease in a renal transplant patient receiving azathioprine, treated by portocaval shunt. Digestion 1984;30: Katzka DA, Saul SH, Jorkasky D, Sigal H, Reynolds JC, Soloway RD. Azathioprine and hepatic veno-occlusive disease in renal transplant patients. Gastroenterology 1986;40: Read AE, Wiesner RH, La Brecque DR, Tifft JG, Mullen KD, Sheer RL, Petrelle M, et al. Hepatic veno-occlusive disease associated with renal transplantation and azathioprine therapy. Ann Intern Med 1986;104: Liano F, Moreno A, Matesanz R, Teruel JL, Redondo C, Garcia-Martin F, Orte L, et al. Veno-occlusive disease of the liver in renal transplantation: is azathioprine the cause? Nephron 1989;51: Sterneck M, Wiesner R, Ascher N, Roberts J, Ferrel L, Ludwig J, Lake. Azathioprine hepatoxicity after liver transplantation. HEPATOLOGY 1991; 14: Mion F, Cloix P, Boillot O, Gille D, Bouvier R, Paliard P, Berger F. Maladie veino-occlusive après transplantation hépatique. Association d un rejet aigu et de la toxicité de l azathioprine. Gastroenterol Clin Biol 1993;17: Dhillon AP, Burroughs AK, Hudson M, Shah N, Rolles K, Scheuer PJ. Hepatic venular stenosis after orthotopic liver transplantation. HEPATOL- OGY 1994;19: Hubscher SG, Adams DH, Buckels JAC, McMaster P, Neuberger J, Elias E. Massive haemorrhagic necrosis of the liver after liver transplantation. J Clin Pathol 1989;42: Demetris AJ, Batts KP, Dhillon AP, Ferrell L, Fung J, Geller SA, Hart J, et al. Banff schema for grading liver allograft rejection: an international consensus document. HEPATOLOGY 1997;25: Ludwig J, Gross JB, Perkins JD, Moore SB. Persistent centrilobular necrosis in hepatic allografts. Hum Pathol 1990;21: Turlin B, Slapak GI, Hayllar KM, Heaton N, Williams R, Portmann B. Centrilobular necrosis after orthotopic liver transplantation: a longitudinal clinico-pathologic study in 71 patients. Liver Transpl Surg 1995;1: Allen KJ, Rand EB, Hart J, Whitington PF. Prognostic implications of centrilobular necrosis in pediatric liver transplant recipients. Transplantation 1998;65: