Centrilobular necrosis (CN) has been reported in

Transcription

1 RAPID COMMUNICATION Centrilobular Necrosis After Orthotopic Liver Transplantation: Association With Acute Cellular Rejection and Impact on Outcome Ziad Hassoun, 1 Vijay Shah, 1 Christine M. Lohse, 2 V. Shane Pankratz, 2 and Lydia M. Petrovic 3 Several studies have linked centrilobular necrosis (CN) to acute cellular rejection (ACR) following liver transplantation. However, it may be difficult to establish the diagnosis of ACR when the classic portal features are absent. The aim of the present study was to identify specific features that would help to recognize ACR in biopsies with CN. One hundred and forty liver biopsies with CN were identified from 97 patients who underwent liver transplantation. The following histopathologic features were assessed: CN, steatosis, lobular inflammation, cholestasis, endothelialitis, and fibrosis. CN was graded semiquantitatively. A number of clinical and biochemical parameters were also recorded. Biopsies with CN were assessed for the presence or absence of ACR and divided into two groups accordingly. The associations of the biochemical, pathologic, and clinical features with ACR were assessed using a multivariate logistic regression model. The outcomes of patients with and without rejection were compared using the Cox proportional hazards regression model. Seventy-four biopsies (52.9%) showed evidence of ACR, and 52 patients (53.6%) had evidence of ACR at the first biopsy with CN. The multivariate analysis showed the presence of cholestasis, lobular inflammation, the ALT level, and time since liver transplantation to be independent predictors of the presence of ACR in biopsies with CN. Patients with ACR on their first biopsy with CN were significantly more likely to experience graft loss compared with patients without ACR. In conclusion, the presence of cholestasis and lobular inflammation on biopsies with CN appeared helpful in predicting its association with ACR. (Liver Transpl 2004;10: ) Abbreviations: CN, centrilobular necrosis; OLT, orthotopic liver transplantation; ACR, acute cellular rejection; CR, chronic rejection; ALT, alanine aminotransferase; ALP, alkaline phosphatase; ACR, acute cellular rejection. From the 1 Advanced Liver Disease Study Group, 2 Division of Biostatistics, and 3 Department of Pathology and Laboratory Medicine, Mayo Clinic, Rochester, MN. Dr. Petrovic is currently affiliated with CHUM Research Center, Montreal, Quebec, Canada. Address reprint requests to Ziad Hassoun, MD, CHUM Research Center, Edouard-Asselin Bldg., 3rd Fl., 264 E. Rene-Levesque Blvd., Montreal, Quebec H2X 1P1, Canada. Telephone: , ext ; FAX: ; ziad.hassoun@umontreal.ca Copyright 2004 by the American Association for the Study of Liver Diseases Published online in Wiley InterScience ( DOI /lt Centrilobular necrosis (CN) has been reported in up to 30% of hepatic allografts following orthotopic liver transplantation (OLT). 1 Various insults have been incriminated in its etiopathogenesis, including ischemia-reperfusion injury, 2 vascular inflow or outflow obstruction, viral and autoimmune hepatitis, and drug toxicity. 3 In particular, CN has been attributed to tacrolimus toxicity in some cases, although definitive evidence to support this concept is lacking. 3,4 In contrast, a large body of evidence supports a frequent association of CN with acute cellular rejection (ACR). 1,5 10 Traditionally, CN has been associated with ACR only if the classic portal findings of the latter are present. 11,12 However, it is has been suggested that ACR can manifest as isolated perivenular and subendothelial inflammation of the terminal hepatic venules without conspicuous portal-tract inflammation or bile-duct damage, and that this form of ACR is often associated with CN. 13 CN has been associated with an increased incidence of chronic rejection (CR) 1,5,7 10 and with lower patient and graft survival rates, 1,5,7 9 although this is controversial. Specifically, antirejection treatment of ACR with centrilobular localization appears to reverse the process in the majority of cases, and may thus prevent the evolution toward CR and graft loss. 10,14,15 Thus, it appears that CN is frequently associated with ACR and could be so even in the absence of the portal features of ACR. In the latter situation, the distinction between ACR and other causes of CN can be difficult. This distinction is important because the therapeutic approach and the prognosis are radically different depending on the etiology. Therefore, the main objective of the present study was to investigate the association of CN with ACR, and more specifically to identify specific clinical, biochemical, or histopathologic features that would help to recognize ACR in biopsies with CN. A secondary objective was to determine whether the association of CN with ACR impacted on patient and graft survival. 480 Liver Transplantation, Vol 10, No 4 (April), 2004: pp

2 Centrilobular Necrosis After OLT 481 Materials and Methods Patients and Biopsy Specimens This was a retrospective study using archived liver biopsy specimens. Needle liver biopsies were obtained prospectively, according to a fixed protocol that was standard practice at the Mayo Clinic: 7 days, 4 months, 1 and 5 years, then every 5 years following liver transplantation. In addition, biopsies were performed when clinically indicated. A follow-up biopsy was always performed after treatment of a rejection episode. A 15-gauge Jamshidi needle (Bard, Murray Hill, NJ) was used under ultrasound guidance. Biopsies were routinely fixed in formalin, embedded in paraffin, and stained with hematoxylin & eosin. All biopsy specimens were reviewed by an experienced hepatopathologist (LMP) blinded to the clinical context of the biopsy and laboratory parameters. The study was approved by the Mayo Clinic Institutional Review Board. Posttransplant liver biopsies with CN were identified through a keyword search of the electronic archive of the Division of Anatomic Pathology for the period from January 1994 to September The following histopathologic features were assessed: CN, steatosis, lobular inflammation, cholestasis, endothelialitis, and fibrosis. CN was diagnosed when the following features were present: centrilobular hepatocyte necrosis associated with sinusoidal congestion and/or hemorrhage, with or without a lobular inflammatory infiltrate. CN was graded semiquantitatively on a scale of grade 1 to grade 3, based on the extent of lobular involvement (Fig. 1) 16 (1: scattered hepatocyte necrosis limited to the immediate perivenular area; 2: small foci of necrosis; 3: confluent foci of necrosis with more extensive involvement of the lobule). When hemorrhagic necrosis was present, the extent of the hemorrhage was also taken into account in the grading of CN. Early posttransplant biopsies in which cholestasis and ballooning degeneration predominated were interpreted as preservation-reperfusion injury 16 and were excluded from the analysis. Next, liver biopsies with CN were assessed for the presence or absence of ACR and divided into two groups accordingly. ACR was diagnosed and graded according to international consensus criteria. 11,12 Clinical and biochemical parameters were then recorded from patients chart review and computerized records. Results of ultrasound with Doppler vascular examination, angiograms, or cholangiograms (when indicated by abnormal biochemistries or ultrasound) were available for all patients. Immunosuppression At the initiation of the liver transplant program, a doubledrug immunosuppressive regimen of cyclosporine and prednisone was used. Azathioprine was added to this regimen in Cyclosporine was started orally at a dose of 8 mg/kg/ day divided in two doses or as a continuous intravenous infusion at a rate of 2 mg/kg/day. The dose was titrated for a trough blood level of ng/ml during the first three months following OLT, and ng/ml thereafter. Azathioprine was administered at a dose of 2 mg/kg/day and Figure 1. Photomicrographs showing the grades of CN. (A) Grade 1 (mild CN), consisting predominantly of hepatocellular ballooning with occasional focal liver cell necrosis, involving zone 3 of the acinus. (B) Grade 2 (moderate CN), with focal liver cell necrosis, congestion, and mild inflammatory infiltrate, involving predominantly zone 3 of the acinus. (C) Grade 3 (severe CN), consisting of confluent lobular necrosis, congestion, and hemorrhage, with more extensive involvement of the lobule (zones 2 and 3).

3 482 Hassoun et al. was discontinued 4 months after OLT. Prednisone was initiated intravenously at a dose of 100 mg/day and tapered gradually to 50 mg/day over the first week and 10 mg/day by 6 months after OLT. This regimen was altered progressively. Tacrolimus was introduced in 1990 and was used routinely as replacement for cyclosporine since 1997 (0.1 mg/kg/day divided into two oral doses, titrated for a trough blood level of ng/ml during the first two months following OLT and 5 10 ng/ml thereafter). In 1999, mycophenolate mofetil (2 g/day divided into two doses) was substituted for azathioprine and was also discontinued between 2 and 4 months following OLT. Finally, the doses of prednisone were reduced to 20 mg/day over the first week and gradually tapered, and prednisone is now discontinued 4 months after OLT. An initial episode of ACR or a recurrence more than 30 days after the initial episode was treated with intravenous steroid pulse therapy (1 g methylprednisolone IV every other day, for a total of 3 doses). Refractory rejection or a recurrence within 30 days after the initial episode was treated with OKT3 (2.5 5 mg/day for 10 days). Statistical Methods Clinical, biochemical, and pathologic features were compared among biopsies with and without ACR, using chi-square, Fisher exact, and Wilcoxon rank sum tests. All tests were two-sided. The relationships of these features with the presence of ACR were also evaluated in a multivariate logistic regression model. Since there were many features to evaluate, only those that were statistically significant on univariate analysis were entered in the multivariate model. This model was developed using both stepwise and backward selection procedures. Natural logarithmic transformations for the continuous variables were explored to improve model fit. Time since OLT (in months) was also entered in the final model. The rate of graft loss, defined as death or retransplant, was calculated using the Aalen estimator. 17 The outcomes of patients with and without ACR in their first liver biopsy with CN were compared using Cox proportional hazards regression model with a time-dependent covariate accounting for the time elapsed between OLT and the first liver biopsy with CN. P values less than.05 were considered statistically significant. Results Patients One hundred and forty liver biopsies with CN were identified from 97 patients (50 males and 47 females) who underwent OLT between January 1986 and August The baseline characteristics recorded for these patients included age at transplant, weight at time of biopsy, donor and recipient gender, cytomegalovirus immune status of donor and recipient, indication for liver transplantation, type of immunosuppression used at the time of the biopsy (mono-, double or triple therapy), and whether the patient was receiving azathioprine, mycophenolate mofetil, cyclosporine, or tacrolimus at the time of the biopsy. We recorded also if patients were found to have hepatic vascular inflow or outflow obstruction. For patients in whom CN was observed in their first liver allograft (85.6% of cases, n 83), the indications for liver transplantation were fulminant hepatic failure (n 4), alcoholic liver disease (n 5), chronic viral hepatitis (n 23), cholestatic liver disease (n 29), autoimmune hepatitis (n 7), and other indications (n 15). The median age at the time of transplantation was 50 years (range, 7 mo to 69 yr). The median time from transplantation to the first biopsy with CN was 3.6 months (range, 1 day to 12.2 yr). Sixty-eight patients were treated with tacrolimus at the time of the first biopsy with CN, 23 were on cyclosporine, and 6 did not receive any calcineurin inhibitor. A diagnosis of hepatic artery thrombosis or venous outflow obstruction was made in 10 patients. Eighty-five (87.6%) patients with CN had at least one follow-up biopsy. CN had resolved on the follow-up biopsy of 39 patients and was found in more than one biopsy in 46 patients. In 2 patients, CN was transiently absent, but it reappeared in later biopsies. In the 46 patients with persistent CN, its median duration was 6.9 months (range, 5 days to 10 yr). Fifty-two (53.6%) patients had evidence of ACR at the first posttransplant biopsy with CN. Baseline clinical characteristics of patients with CN, with or without the presence of ACR, are summarized in Table 1. There was no significant difference between the two groups. Biopsy Results The median number of portal tracts per biopsy was 7 (range, 1 18). Seventy-four of the 140 biopsies with CN (52.9%) showed concomitant evidence of ACR. ACR was mild in 57 (40.7%) biopsies and moderate in 17 (12.1%). The clinical and biochemical features that were evaluated included time since liver transplant, cyclosporine or tacrolimus levels, alanine aminotransferase (ALT), alkaline phosphatase (ALP), bilirubin level, and international normalized ratio (Table 2). On univariate analysis, ALT and ALP were significantly higher in the group with ACR. Among the histopathologic features, lobular inflammation, cholestasis, and endothelialitis occurred significantly more frequently in this group (Fig. 2A). In addition, more patients in the ACR group had grade 2 CN, while patients without ACR more frequently had grade 1 CN (Fig. 2B). These features were not meaningfully different among biop-

4 Centrilobular Necrosis After OLT 483 Table 1. Baseline Clinical Characteristics of 97 Patients With CN on Liver Biopsy According to the Presence or Absence of ACR on the First Biopsy With CN Without ACR n 45 With ACR n 52 P value* Age at transplant (yr) 49.8 ( ) 51.0 ( ).323 Weight at first biopsy (kg) 74.7 ( ) 65.1 ( ).388 Gender of recipient (F / M) 24/21 (53.3/46.7) 23/29 (44.2/55.8).835 Gender of donor (F / M) 15/26 (36.6/63.4) 20/30 (40.0/60.0).091 CMV mismatch 5 (11.6) 7 (13.5).223 Indication for OLT.457 Chronic viral hepatitis 11 (24.4) 12 (23.1) AIH / PBC / PSC 15 (33.3) 21 (40.4) Other 19 (42.2) 19 (36.5) Immunosuppression.769 Double therapy 21 (48.8) 25 (51.0) Triple therapy 22 (51.2) 24 (49.0) Antimetabolites.665 Azathioprine 21 (47.7) 25 (49.0) MMF 6 (13.6) 5 (9.8) None 17 (38.6) 21 (41.2) Calcineurin inhibitors.120 Cyclosporine 11 (26.2) 12 (24.5) Tacrolimus 31 (73.8) 37 (75.5) Vascular obstruction 5 (11.4) 5 (9.6).155 Abbreviations: CN, centrilobular necrosis; ACR, acute cellular rejection; F, female; M, male; CMV, cytomegalovirus; OLT, orthotopic liver transplantation; AIH, autoimmune hepatitis; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; MMF, mycophenolate mofetil. NOTE. Results are presented as n (%) or median (range). *P values were based on Cox proportional hazards models for continuously scaled variables and on log-rank tests for discrete variables. sies with mild or moderate ACR. The features significantly associated with ACR in liver biopsies with CN on multivariate analysis included cholestasis, lobular inflammation, serum ALT, and time since OLT (Table 3). The odds of ACR were 16 times higher among specimens with cholestasis relative to specimens without this feature (P.001). The odds of ACR were 4 times higher among specimens with lobular inflammation relative to those without this feature (P.010). The odds ratios for the associations between ALT and months since OLT with ACR (with both features transformed to the natural logarithmic scale to improve Table 2. Clinical and Biochemical Features Associated With 140 Liver Biopsies With CN Without ACR n 66 With ACR n 74 P value* Months since OLT 4.0 ( ) 2.5 ( ).105 Cyclosporine level (ng/ml) 176 (66 433), n (25 272), n Tacrolimus level (ng/ml) 8.8 ( ),n ( ),n ALT (U/L) 59 ( ) 208 ( ).001 ALP (U/L) ( ) 611 ( ).004 Bilirubin (mg/dl) 1.5 ( ) 3.0 ( ).093 INR 1.1 ( ) 1.1 ( ).752 Abbreviations: CN, centrilobular necrosis; ACR, acute cellular rejection; OLT, orthotopic liver transplantation; ALT, alanine aminotransferase; ALP, alkaline phosphatase; INR, international normalized ratio. NOTE. Results are presented as n (%) or median (range). *P values were based on Wilcoxon rank sum tests.

5 484 Hassoun et al. Table 4. Reason for Graft Loss in 25 Patients With CN Without ACR n 11 With ACR n 14 Chronic rejection 2 4 Vascular obstruction 3 3 Recurrent hepatitis C 1 2 Malignancy 1 1 Cardiovascular event 1 1 Primary nonfunction 1 1 Other 2 2 Abbreviations: CN, centrilobular necrosis; ACR, acute cellular rejection. model fit) were 1.8 and 1.4, respectively, (P.015 and.021, respectively), indicating that as the ALT level and time since OLT increased, the likelihood of having ACR in a biopsy with CN also increased. In addition to analyzing time since OLT as a continuously scaled variable, an indicator variable for biopsies taken more than 6 months after OLT was added to the multivariate model. Biopsies with CN taken more than 6 months after OLT were 3.6 times more likely to have ACR compared with biopsies taken less than 6 months after OLT, after adjusting for cholestasis, lobular inflammation, and ALT level (P.038). Figure 2. (A) Among histopathologic features of 140 liver biopsies with CN, lobular inflammation, cholestasis, and endothelialitis differed significantly on univariate analysis between biopsies with and without concomitant ACR (P <.001 for all). (B) The grade of CN also differed significantly between biopsies with and without concomitant ACR (P <.001). Table 3. Features That Were Significantly Associated With ACR in Liver Biopsies With CN on Multivariate Analysis Odds Ratio [95% CI] P value Cholestasis [4.74,54.74].001 Lobular inflammation 3.92 [1.39,11.04].010 ALT* 1.80 [1.12,2.88].015 Months since OLT* 1.43 [1.06,1.93].021 Abbreviations: ACR, acute cellular rejection; CN, centrilobular necrosis; ALT, alanine aminotransferase; OLT, orthotopic liver transplantation. *ALT and months since OLT were transformed to the natural logarithmic scale in order to improve model fit. Outcome Seven patients from this cohort of patients with CN ultimately developed CR. Two of them did not have ACR in the first posttransplant biopsy with CN, and 5 did (P.636). Twenty-five patients from this cohort of patients with CN experienced graft loss following their initial transplant. The median time from OLT to death or retransplant was 1.5 years, ranging from 7 days to 12.3 years. The median duration of follow-up of the survivors was 4 years, ranging from 23 days to 13.8 years. The hazard ratio for the association between graft loss and the presence of ACR in the first posttransplant biopsy with CN was 3.10 (95% CI: ), indicating that patients with ACR in the first posttransplant biopsy with CN were 3 times more likely to experience graft loss compared with patients without ACR in the presence of CN on the biopsy. This difference was statistically significant (P.007). The causes of graft loss are shown in Table 4. There was no meaningful difference between the two groups. Discussion In the present study, CN was associated with ACR (defined by its classic portal features) in liver allografts in slightly more than one half of the total cases of CN. This figure is comparable with what has been reported previously in the literature. 1,6 10 It has been suggested that CN can be associated with ACR even in the

6 Centrilobular Necrosis After OLT 485 absence of the classic portal findings of the latter. 13 Two studies have examined this particular entity (termed isolated central venulitis ) in adult 14 and pediatric 15 liver transplant recipients and have compared it to cases where central venulitis is associated with portal features of ACR. Both have found that the clinicopathologic features of isolated central venulitis and of central venulitis associated with portal features of ACR are very similar. In particular, the rates of response to treatment of ACR are identical in both groups and close to 100% in one study. 14 Thus, these studies support the concept that CN can be associated with ACR even when the portal features of the latter are absent. This raises the problem of the diagnosis of ACR in this setting. We have shown that cholestasis, lobular inflammation, elevated ALT, and time since OLT were independent predictors of the presence of ACR (defined by its classic portal features) in biopsies with CN. These findings may thus prove useful in predicting the presence of ACR in posttransplant liver biopsies with CN, when the portal features of ACR are absent or nondiagnostic. This should be evaluated prospectively in a separate study. Lobular inflammation has been reported in association with CN in posttransplant liver biopsies, 1,5,6,10,14 16,18 21 and its coexistence with ACR has been noted by some. 5,6,10,14,15,20 However, the etiopathogenetic relationship of this finding with ACR has remained obscure. A strong correlation between CN and lobular inflammation has been described recently, and the same authors have reported a correlation between lobular inflammation and endothelialitis of the central vein. 16,21 In addition, lobular inflammation and CN have been found to respond to antirejection therapy in a number of cases. 14,15,18 A similar lobular necroinflammatory picture without prominent portal inflammation has also been reported after the withdrawal of immunosuppression in stable long-term liver recipients, in whom reinstitution of immunosuppression has resulted in normalization of the biochemical parameters. 22 Taken together, these observations suggest that in the setting of CN lobular inflammation is a manifestation of ACR. Our results support this concept by showing that lobular inflammation is independently related to ACR in biopsies with CN. Cholestasis has also been described in association with CN. 1,10,16,21 However, its diagnostic value has not been examined thoroughly, even though its frequent association with ACR has been observed in this setting. 10 In two studies from the same authors, the severity of cholestasis did not correlate with that of any other histopathologic finding, including those of ACR. 16,21 The first of these studies 16 examined day 7 posttransplant biopsies with preservation-reperfusion injury, and the second examined biopsies of 28 patients who underwent retransplantation for CR. 21 In the present study, early posttransplant biopsies in which cholestasis and ballooning degeneration predominated were interpreted as preservation-reperfusion injury and were excluded from the analysis. Only 11 biopsies were taken within one week of OLT. The presence of cholestasis in biopsies with CN was the strongest independent predictor of the simultaneous presence of ACR. The role of biochemical liver tests, in particular transaminases, in the diagnosis of ACR in liver allografts is controversial. Several studies have demonstrated their lack of sensitivity and specificity. 23,24 It is nevertheless common to observe an increase in transaminase values when ACR develops 25 and a decrease after successful treatment. An elevation of the transaminases has been reported in association with CN. 9,10,14,15,18 Similar to our findings, Nakazawa et al. have reported higher values of ALT when CN is associated with ACR. 10 We found that ALT elevation independently predicted the presence of ACR in posttransplant liver biopsies with CN. CN has been associated with an increased incidence of CR. 1,5,7 10 Indeed, both ACR and CN have recently been shown to be constant features of the evolution toward CR. 19,21 When biopsies with ACR that proceeds to CR are compared with biopsies with ACR that eventually resolves, the former group is found to have more severe CN, while the portal features of ACR appear significantly worse in the group with resolving ACR. 21 Both the incidence and severity of central-vein endothelialitis are higher in the group that ultimately develops CR. 19 Lobular inflammation also appears to play a role in this setting. 19,21 In an earlier study, a perivenular necroinflammatory process had been identified in all patients who developed CR, and ACR coexisted in many. 20 Other authors have observed that lobular inflammation is more frequent in patients who develop CR. 26 These observations suggest that ACR with centrilobular predilection, which is frequently associated with CN, may represent a subset with an increased propensity to evolve into CR. In the present study, the incidence of CR in the group without ACR appeared comparable to the overall rate of 3% observed in liver transplant recipients at the Mayo Clinic. 27 It was more than twice as high in the ACR group, but this difference was not statistically significant. This may be due to insufficient statistical power. Alternatively, the relatively aggressive policy of treatment of ACR in our

7 486 Hassoun et al. institution could account for a lower incidence of CR in the group of patients with ACR and CN. Finally, since ACR was defined by its classic portal features, some of the biopsies without ACR might actually represent ACR with centrilobular predilection, possibly including the two patients in this group who developed CR. CN has also been associated with lower patient and graft survival rates. 1,5,7 9 However, all authors have not noted such a poor prognosis. In particular, in cases of ACR with centrilobular localization (whether associated with portal features of ACR or not), antirejection treatment appears to reverse the process in the majority of cases, and may thus prevent the evolution toward CR and graft loss. 2,10,14,15 A beneficial response to an increase in immunosuppression has also been noted in children with CN after combined liver and small-bowel transplantation. 18 In our study, the 5-year survival free of retransplantation of patients with CN was 69%, somewhat lower than the overall graft survival rate observed in our institution. 27 On the other hand, when the time between OLT and the first biopsy with CN was taken into account in a Cox regression model, patients with CN and ACR appeared approximately 3 times more likely to experience graft loss compared with patients with CN alone. The reason why the association of ACR and CN carries a worse prognosis in our patients is not clear, as the causes of graft loss were similar in the two groups (Table 4). In summary, we found that the presence of cholestasis and lobular inflammation in biopsies with CN may be helpful in the identification of associated ACR. A prospective study is needed to determine whether this association remains valid when the classic portal features of ACR are absent. Additionally, the outcome of CN may be worse when it is associated with rejection. References 1. Turlin B, Slapak GI, Hayllar KM, Heaton N, Williams R, Portmann B. Centrilobular necrosis after orthotopic liver transplantation: A longitudinal clinicopathologic study in 71 patients. Liver Transpl Surg 1995;1: Khettry U, Backer A, Ayata G, Lewis WD, Jenkins RL, Gordon FD. Centrilobular histopathologic changes in liver transplant biopsies. Hum Pathol 2002;33: Demetris AJ. Central venulitis in liver allografts: considerations of differential diagnosis. Hepatology 2001;33: Hassoun Z, Shah V, Petrovic LM. The choice of calcineurin inhibitor does not influence centrilobular necrosis after orthotopic liver transplantation. Transplant Proc 2002;34: Ludwig J, Gross JB Jr., Perkins JD, Moore SB. Persistent centrilobular necroses in hepatic allografts. Hum Pathol 1990;21: Dhillon AP, Burroughs AK, Hudson M, Shah N, Rolles K, Scheuer PJ. Hepatic venular stenosis after orthotopic liver transplantation. Hepatology 1994;19: Gomez 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: Allen KJ, Rand EB, Hart J, Whitington PF. Prognostic implications of centrilobular necrosis in pediatric liver transplant recipients. Transplantation 1998;65: Sebagh M, Debette M, Samuel D, Emile JF, Falissard B, Cailliez V, et al. Silent presentation of veno-occlusive disease after liver transplantation as part of the process of cellular rejection with endothelial predilection. Hepatology 1999;30: Nakazawa Y, Walker NI, Kerlin P, Steadman C, Lynch SV, Strong RW, et al. Clinicopathological analysis of liver allograft biopsies with late centrilobular necrosis: A comparative study in 54 patients. Transplantation 2000;69: Terminology for hepatic allograft rejection. International Working Party. Hepatology 1995;22: Banff schema for grading liver allograft rejection: An international consensus document. Hepatology 1997;25: Demetris A, Adams D, Bellamy C, Blakolmer K, Clouston A, Dhillon AP, et al. Update of the International Banff Schema for Liver Allograft Rejection: Working recommendations for the histopathologic staging and reporting of chronic rejection. An International Panel. Hepatology 2000;31: Tsamandas AC, Jain AB, Felekouras ES, Fung JJ, Demetris AJ, Lee RG. Central venulitis in the allograft liver: a clinicopathologic study. Transplantation 1997;64: Krasinskas AM, Ruchelli ED, Rand EB, Chittams JL, Furth EE. Central venulitis in pediatric liver allografts. Hepatology 2001; 33: Neil DA, Hubscher SG. Are parenchymal changes in early posttransplant biopsies related to preservation-reperfusion injury or rejection? Transplantation 2001;71: Tsiatis AA. A large sample study of Cox s regression model. Annals of Statistics 1981;9: Lacaille F, Canioni D, Fournet JC, Revillon Y, Cezard JP, Goulet O. Centrilobular necrosis in children after combined liver and small bowel transplantation. Transplantation 2002;73: Gouw AS, van den Heuvel MC, van den Berg AP, Slooff MJ, de Jong KP, Poppema S. The significance of parenchymal changes of acute cellular rejection in predicting chronic liver graft rejection. Transplantation 2002;73: Quaglia AF, Del Vecchio Blanco G, Greaves R, Burroughs AK, Dhillon AP. Development of ductopaenic liver allograft rejection includes a hepatitic phase prior to duct loss. J Hepatol 2000;33: Neil DA, Hubscher SG. Histologic and biochemical changes during the evolution of chronic rejection of liver allografts. Hepatology 2002;35: Devlin J, Doherty D, Thomson L, Wong T, Donaldson P, Portmann B, et al. Defining the outcome of immunosuppression withdrawal after liver transplantation. Hepatology 1998; 27: Abraham SC, Furth EE. Receiver operating characteristic analysis of serum chemical parameters as tests of liver transplant rejection and correlation with histology. Transplantation 1995;59: Neuberger J, Wilson P, Adams D. Protocol liver biopsies: The case in favour. Transplant Proc 1998;30:

8 Centrilobular Necrosis After OLT Figueiredo FAF, Menon KVN, Wiesner RH. Diagnosis and treatment of hepatic allograft dysfunction. In: Norman DJ, Turka LA, eds. Primer on transplantation. Mt. Laurel, NJ: American Society of Transplant Physicians;2001: Sebagh M, Blakolmer K, Falissard B, Roche B, Emile JF, Bismuth H, et al. Accuracy of bile duct changes for the diagnosis of chronic liver allograft rejection: Reliability of the 1999 Banff schema. Hepatology 2002;35: Wiesner RH, Krom RAF. Twelve years of liver transplantation at the Mayo Clinic. In: Cecka JM, Terasaki PI, eds. Clinical Transplantation. Los Angeles: UCLA Tissue Typing Laboratory;1997: