A wig et all have highlighted a histological lesion

Transcription

1 Centrilobular Necrosis After Orthotopic Liver Transplantation: A Longitudinal Clinicopathologic Study in 71 Patients Bmno Turlin, * Gabriella I. Slapah, * Karen M. Hayllar, * Nigel Heaton, f Roger Williams, * and Bernard Portmann * Centrilobular necrosis (CLN) is a histological finding often encountered after orthotopic liver transplantation, but its pathogenesis is still unknown. In this study, the significance of CLN was assessed in a series of 227 consecutive liver transplantations performed between January 1989, and December Seventy-one patients (30.9%) showed CLN on at least one biopsy result, which were obtained because of an increase of aspartate aminotransferase activity. Their liver specimens were reviewed, and 19 histological features were recorded with particular attention given to lobular changes in acinar zone 3, to features commonly attributed to cellular and ductopenic rejection, and to changes suggestive of ischemia. CLN could first be observed either soon (within 4 days) or late (up to 3 years) after transplantation. Only 23 (32.4%) specimens had centrilobular necrosis affecting more than 75% of acinar zones 3. In 60 cases (84.5%) the lesion was limited to acinar zone 3. An important associated feature was sinusoidal congestion in 73.2% of cases. Fifty-one of 71 patients (71.8%) had histological features of cellular rejection before or at the time of CLN, and 13 of these progressed to ductopenic rejection versus 3 of the 156 patients without CLN (P c.0001). Nine patients had a recurrence of CLN, of whom 2 progressed to ductopenic rejection, a recurrence rate of 16.7% in this series. The survival of patients with CLN is worsened by associated ductopenic portal tracts compared with those without ductopenia (P =.0189-Mantel- Cox). This histological combination, irrespective of the serum bilirubin level, may warrant an early conversion to FK506-based immunosuppression. Copyright o 1995 by the American Association for the Study of Liver Diseases fter orthotopic liver transplantation (OLT), Lud- A wig et all have highlighted a histological lesion characterized by cell loss in acinar zone 3, which they named centrilobular necrosis (CLN). The occurence of CLN after OLT had been mentioned in earlier W O ~ ~ and S ~ variably. ~ interpreted as part of reperfusion damage, cellular rejection, or vascular compromise. The change was used to up-grade the overall severity of rejection in hover s grading ~ystem.~ CLN has been identified in up to 23% of hepatic allografts in Ludwig s study, either in isolation or associated with other allograft pathology such as rejection, ischemia, or toxic injury. In that study persistent CLN appeared to be a harbinger of a poor prognosis, whereas a recent work emphasizes its re~ersibility~ indicating that not only the pathogenesis but the prognostic significance of CLN remain uncertain. The aims of the present study are to establish the prevalence of CLN in a large series of biopsy specimens from consecutive liver transplantations and to investigate its relationship with other graft pathology and its impact on graft survival. Patients and Methods Patients and Biopsy Specimens Seven hundred twenty liver biopsy specimens that were obtained from 227 patients who received an orthotopic liver graft between January 1, 1989, and December 31, 1991, were reviewed. There was no protocol biopsy policy in our center beiore September 1990, and all biopsy specimens were obtained on clinical indication. After September 1990, 76 patients entered a randomized trial of FK506-T and cyclosporin-based immunosuppressive regimens that involved a protocol biopsy on day 7. Overall, 151 patients?the immunosuppresant drug FW06 has been licensed for prescription under the name Tacrolimus. From the *Institute of Liver Studies and tliver Transplant Surgical Sewice, King s College School of Medicine and Dentistiy, London, United Kingdom. Address reprint requests to: Bernard Portmann, MD, Institute of Liver Studies, King s College Hospital, Denmark Hill, London SE5 9RS, United Kingdom. Copyright by the American Association fur the Study of Liver Diseases /95/ $3.00/0 Liver Transplantation and Surgeiy, Vof 1, No 5 (Septemh], 1995: pp

2 ~ ~~ ~ ~~ 286 Turlin et a1 received triple therapy with cyclosporin, azathioprine, and prednisolone, and 37 had an FK506-based regimen. The 71 patients included in the study, 13 on FK506-based immunosuppression, underwent at least one follow-up biopsy that showed CLN. The 371 biopsy specimens available from these 71 patients were divided into four groups according to the following characteristics (Table 1): group 1, reperfusion biopsy specimens; group 2, biopsy specimens obtained before the first occurrence of CLN; group 3, biopsy specimens first showing CLN; group 4, biopsy specimens taken subsequent to the first occurrence of CLN and up to the end of the follow-up period, December 31, All biopsy specimens with CLN, including 14 protocol specimens taken on day 7, had been preceded by an increase in aspartate aminotransferase two or more times above the upper limit of the normal level (normal, 5 to 35 U/L). CLN has not been an incidental finding in the present study. Histological Methods All liver specimens were formalin-futed and routinely paraffinembedded; 3-pm thick sections were stained with hematoxylineosin, periodic acid-schiff after diastase, orcein, reticulin, and Perk methods. The following histological features were assessed semiquantitatively: percentage of acinar zones 3 with CLN; extent of CLN-only acinar zone 3; acinar zone or acinar zone 3-2 with bridging; lobular inflammation-?, 1 +, 2+ according to increasing number of inflammatory cells. Other lobular features assessed as present (+) or absent (-1 were as follows: sinusoidal congestion, endotheliitis of the termi- Table 1. Age, Gender, Original Disease, and Biopsy Specimens of 71 Patients With CLN in at Least One Biopsy Specimen Age 37.1 f 18 Gender F 47 (66.2) M 24 (33.8) Original disease PBC 24 (33.8) CIR 17 (23.9) ALF 15 (21.1) PSC 9 (12.7) EHBA 5 (7.0) HCC 1 (1.4) Biopsy specimens Reperfusion 22 (31.O) Before CLN (Group 2) 33 (46.5) First occurrence of CLN (Group 3) 71 (100.0) After CLN (Group 4) 59 (83.1) NOTE. Results expressed as percentage or means? SD. Abbreviations: PBC, primary biliary cirrhosis: CIR, cirrhosis; ALF, acute liver failure: PSC, primary sclerosing cholangitis; EHBA, extrahepatic biliary atresia; HCC, hepatocellular carcinoma. nal hepatic venule, canalicular cholestasis, feathery degeneration, fibrosis in acinar zone 3, mitosis, apoptosis, anisocaryosis, hepatocytes ballooning, coagulative necrosis, lobular hepatitis (viruslike). Associated changes recorded were cellular and ductopenic rejection according to accepted criteria (the severity of the latter was estimated by the percentage of portal tracts without an identifiable bile duct), cholangiolar cholestasis (sepsislike), feature of large duct obstruction (portal edema, ductular proliferation, and cholangiolitic inflammation), and minimal changes. In biopsy specimens of group 4, the persistence of CLN was recorded as (+) its absence as (-1. The degree of ductopenia was estimated in all group 3 biopsy specimens, including 16 hepatectomy specimens obtained at retransplantation. Evidence of ductopenic rejection and presence of CLN were assessed in subsequent allograft I1 biopsy specimens in view of the known risk of recurrence of ductopenic rejection. Statistical Methods The results are expressed as mean 2 SD (minimum-maximum) number for quantitative characters and as number (percentage) for qualitative characters. The statistical tests used were nonparametric tests (Mann Withney [MW], x2 test and Fisher s exact test [FET]). The percentage rate of survival graft (or patient) at 5 years was assessed by life table method and significance by the Mantel-Cox s test. P <.05 was considered significant. Results None of the reperfusion specimens showed significant CLN, but this was found in at least one biopsy specimen in 71 of 227 consecutive grafts (30.9%). CLN was observed in the first allograft in all but 2 patients, in whom it was first observed in a second allograft. Fifty-one (71.8%) patients had features of cellular rejection in at least one biopsy specimen before CLN (22 of 33) or at the time of CLN (39 of 71), and this was mild in 10, moderate in 31, and severe in the remaining 10. Thirteen of 33 pre-cln (group 2) specimens showed perivenular hepatocyte ballooning, but none exhibited significant ductopenia. Table 2 shows the timing of group 3 biopsies and the main histological features of and associated with CLN. The lesion had a patchy distribution, because only 23 (32.4%) of the grafts had more than 75% of their acinar zone 3 affected. It was rarely extensive; in 84.5% of the cases, CLN was limited to acinar zone 3 (Fig 1). More than 50% of the grafts with CLN showed sinusoidal congestion and/or cholestasis, whereas multinucleated giant hepatocytes and granulomas were incidentally observed in two and three cases only, respectively. Except for endotheliitis, the various lobular features observed with CLN had a similar distribution whether associated with rejection, hepatitis, or ischemia. As ex-

3 Centrilobular Necrosis in Liver Allograft 287 Table 2. Timing of (means 2 SD) and Main Lobular Changes (%) in 71 Specimens With CLN (Group 3) Days after OLT 1 st week 2nd week After % of lobules with CLN < 25% 25 to 50% 50 to 75% > 75% Extent of CLN Acinar zone 3 Acinar zone Bridging Inflammatory cells * Endotheliitis Sinusoidal congestion Cholestasis Feathery degeneration Eosinophilic necrosis Fibrosis Anisocaryosis Mitoses Granulomas Multinucleated giant hepatocytes 84.3? 173 (4-1,100) 13 (18.3) 10 (14.1) 48 (67.6) 11 (1 5.5) 21 (29.6) 16 (22.5) 23 (32.4) 60 (84.5) 5 (7.0) 27 (38.0) 33 (46.5) 11 (15.5) 9 (12.7) 52 (73.2) 44 (62.0) 11 (15.5) 30 (42.3) 12 (16.9) 3 (4.2) 2 (2.8) pected, endotheliitis was more frequently observed with cellular rejection (25.8%) than in absence of rejection (3.2%) (P =,0352; FET). In a single case endotheliitis was associated with the changes of severe lobular hepatitis. The incidence of histological features was similar in the different etiological groups. CLN persisted in at least one subsequent biopsy specimen in 51 (86.4%) of 59 patients who underwent a follow-up biopsy. Thirtythree patients (47.1%) had lost their graft after 229 days k 308 (15 to 1219) through death17 or at retransplantation.16 These patients had more cholestasis (78.8% v 47.4% - P =,0078; FET) and a higher number of portal tracts without bile duct (10 k 19.5~ 2.6 k 8.5; P =.0023; MW) in their biopsy specimens than patients with CLN who had kept their graft. Of 45 allografts with CLN but no ductopenia, 15 (33.3%) progressed to graft failurev 18 (69.2%) of 26 grafts with ductopenic portal tracts (P =,0061; FET). Thirteen of 71 patients (18.3%) with CLN in their liver had evidence of or progressed to ductopenic rejection, but only 3 out of the 156 (1.9%) patients without CLN (P <.0001; FET). There was no statistical difference in the rate of ductopenic rejection between patients with persistent CLN on biopsy and those in whom CLN had resolved (23.5%~ 12.5%; ~ 2). A second OLT was performed in 16 patients for ductopenic rejection in 12 (75%), biliary sepsis in 2, hepatic artery thrombosis in 1, and recurrence of primary sclerosing cholangitis in 1. Two patients died shortly after the second graft. Of 14 patients who had a biopsy performed on their second graft, recurrence of CLN was observed in 9 (64.3%), of whom 2 progressed to ductopenic rejection. A third OLT was needed in 3 patients (2 in whom ductopenic rejection developed and one for ischemia). No patients without CLN recurrence in their graft progressed to ductopenic rejection or needed a third OLT. The recurrence rate of ductopenic rejection was 2 of 12 (16.7%). Of 9 patients with CLN recurrence, 3 have died compared with one of 5 without CLN recurrence in the graft. The percentage of survival at the end of follow-up periods of 2 to 5 years was not statistically different between grafts in which CLN had either persisted or resolved (36 of 51 v 7 of 8). The rate and duration of survival were similar in patients in whom CLN was first observed early (1 to 2 weeks) or late after OLT. The Mantel-Cox test showed that ductopenic portal tracts associated with CLN had a worse effect on graft (29% v 63%; P =.0036) and patient (49% v 75%; P =.0189) survival at 5 years. Graft survival was also impaired when cholestasis was associated with CLN (46% v 69%; P =,0097). Discussion In the present study, the percentage of CLN in 71 of 227 consecutive grafts (30.9%), is not statistically different from the 18% found by Ludwig et all (P =,0644-FET) but significantly higher than the 2.4% found by Lagenaux et a15 (P <,0001-FET). The average time of appearance of CLN was 84.3 days, but the change could be first observed either soon (within 4 days) or late (up to 3 years) after transplantation. Unlike Markin et a1,6 we have not observed CLN in any of the reperfusion specimens. Foci of hemopoiesis were not associated with CLN in this series, which is at variance with Ludwig's findings.' Recently Dhillon et a17 have described stenosis and fibrosis of the central vein associated with CLN in nine allografts. We have found some degree of perivenular fibrosis in 16.9% of the grafts

4 288 Turlin et al Figure 1. Liver histology on the 18th day after OLT. Characteristic CLN with confluent hepatocyte drop-out and a light mixed inflammatory cell infiltrate affecting most of acinar zone 3 in a case that subsequently developed ductopenic rejection. (Hematoxylin and eosin; original magnification x 160.) with CLN, but stenosis of the central vein was not a feature in any of our cases. Shibayama et a1* described centrilobular necrosis of the liver of nontransplant patients in which the necrosis followed liver congestion. Similarly, in the liver graft, the high frequency of associated congestion suggests that some degree of venous outflow block might be the primary event. However, congestive heart failure or interference with venous outflow at the hepatic vein or at the caval anastomosis levels were not shown in the present patients who underwent transplantation. Alternatively, the lesion may follow damage to the hepatic venular wall and sinusoidal outlets because of endotheliitis, as 51 (71.8%) of patients had one (or more) biopsy specimen(s> showing features of acute cellular rejection, before or at the time CLN was observed. Difference in the timing of the biopsy would explain differences in morphological appearances, namely persistent endotheliitis in some, predominant cell drop out and sinusoidal congestion in other, whereas late changes might include perivenular fibrosis. The progression to ductopenic rejection was more frequent in patients with CLN (18.3%) than without CLN (1.9%), a difference that is highly significant (P <,0001; FET). This association between CLN and subsequent development of ductopenic rejection agrees with both Ludwig et all results and those of a large recently published series, the latter one showing an even stronger association, 53.8% of grafts with CLN having developed chronic rejection against 13.1% of the control^.^ It differs from the results of Lagenaux et al, who report that only 1 of their 12 patients (8.3%) with CLN on biopsy has progressed to ductopenic reje~tion.~ However, the latter study, published in abstract form, does not indicate the prevalence of ductopenic rejection in that particular series. In our cases ductopenic rejection was invariably associated with foam cell obliterative arteriopathy, as shown in the grafts removed at retransplantation. It is logical to envisage that ischemia may have been partially responsible for CLN, similar to the role it may play in the progressive disappearance of the intrahepatic bile ducts. However, ischemia is unlikely to be the only factor. There was no direct correlation between the severity of CLN and degree of arterial occlusion and, in some instances, CLN resolved in presence of extensive arterial occlusion or

5 Centiiilobular Necrosis in Liver Allograjt 289 progressed despite a minimal degree of arteriopathy. A direct lymphocytotoxic'* or antibody-mediated13 damage to the perivenular parenchyma may additionally be involved, similar to that implicated in bile duct loss.'4 CLN recurred in the second allograft in only one of the patients reported by Ludwig et al,' but only two underwent a second OLT. We report here nine patients in whom CLN recurred after retransplantation (63.4%). The only two cases with recurrence of ductopenic rejection were preceded by CLN recurrence. The 16.6% recurrence rate of ductopenic rejection in second allograft found in this series is markedly lower than the 70% reported by van Hoek et a1,15 and more in keeping with the 27% quoted by Adams and Neuberger.16 The better prognosis of resolutive CLN compared with persistent CLN found by Ludwig et all and Gomez et a19 was not confirmed in our series. CLN associated with either bile duct loss or cholestasis had a worse effect on graft survival than CLN alone. Only the association with bile duct loss, not with cholestasis, affected patients survival. The impairment in survival is likely because of deleterious effect of ductopenic rejection, not CLN alone. Indeed, the majority of patients with both bile duct loss and CLN progressed to irreversible ductopenic rejection. Because CLN with incipient ductopenia is strongly associated with a progression to ductopenic rejection, this histological picture may be an early marker of evolutive ductopenic rejection; at a time a conversion from cyclosporin to FK506 has the most chance to be beneficial. References Ludwig J, Gross JB, Perkins JD, Moore SB. Persistent centrilobular necrosis in hepatic allografts. Hum Pathol 1990;21: Grond J, Gouw ASH, Poppema S, Slooff MJH, Gips CH. Chronic rejection in liver transplants: a histopathologic analysis of failed grafts and antecedent serial biopsies. Transplant Proc 1986;18: Demetris AJ, Jaffe R, Starzl TE. A review of adult and pediatric post transplant liver pathology. Pathol Annu 1987;22: Snover DC, Freese DK, Sharp HL, Bloomer JR, Najarian JS, Ascher NL. Liver allograft rejection. An analysis of the use of biopsy in determining the outcome of rejection. Am J Surg Pathol 1987;ll:l Lagenaux G, Garbar C, Rahier J. Centrilobular hepatic drop out after liver transplantation: an analysis of 4500 biopsies [abstract]. Path Res Pract 1993;189: Markin RS, Wood RP, Stratta RJ, Langhas AN, Pillen TJ, Davidson I, et al. Predictive value of intraoperative liver biopsies of donor organs in patients undergoing orthotopic liver transplantation. Transplant Proc 1990;22: Dhillon AP, Burroughs AK, Hudson M, Shah N, Rolles K, Scheuer PJ. Hepatic venular stenosis after orthotopic liver transplantation. Hepatology 1994;19: Shibayama Y, Urano T, Asaka S, Hashimoto K, Nakata K. Pathogenesis of centrilobular necrosis following congestion of the liver. J Gastroenterol Hepatol 1993;8: Gornez R, Colina F, Moreno E, Gonzalez I, Loinaz C, Garcia I, et al. Etiopathogenesis and prognosis of centrilobular necrosis in hepatic grafts. J Hepatol 1994; 21 : Oguma S, Zerbe T, Banner B, Belle S, Starzl TE, Demetris AJ. Chronic liver allograft rejection and obliterative arteriopathy: possible pathogenic mechanisms. Transplant Proc 1989;21: Matsumoto Y, McCaughan GW, Painter DM, Bishop GA. Evidence that portal tract microvasculature destruction precedes bile duct loss in human liver allograft rejection. Transplantation 1993;56: Wiesner RH, Ludwig J, van Hoek B, Krom RAF. Current concepts in cell-mediated hepatic allograft rejection leading to ductopenia and liver failure. Hepatology 1991 ;14: Demetris AJ, Murase N, Nakamura K, lwaki Y, Yagihashi A, Valdivia L. lmmunopathology of antibodies as effectors of orthotopic liver allograft rejection. Sem Liver Dis 1992;12: Oguma S, Belle S, Starzl TE, Demetris AJ. A histometric analysis of chronically rejected human liver allgrafts. Insight into the mechanisms of bile duct loss: direct immunologic or ischaemic factors. Hepatology 1989;9: Van Hoek B, Wiesner RH, Ludwig, J, Paya C. Recurrence of ductopenic rejection in liver allografts after retransplantation for vanishing bile duct syndrome. Transplant Proc 1991 ;23: Adams DH, Neuberger JM. Pattern of graft rejection following liver transplantation. J Hepatol 1990;10: Wong P, Devlin J, Gane E, Ramage J, Portmann B, Williams R. FK506 rescue therapy for intractable liver allograft rejection. Transpl Int 1994;7(suppl 1):70-76.