Cirrhosis secondary to hepatitis C virus (HCV) infection

Transcription

1 The Use of Hepatitis C Viral RNA Levels in Liver Tissue to Distinguish Rejection From Recurrent Hepatitis C Michelle J. Gottschlich, * Kay L. Aardema, Eileen M. Burd, Raouf E. Nakhleh, Kimberly A. Brown, * Marwan S. Abouljoud, Kathryn Hirst, and Dilip K. Moonka * Persistence of hepatitis C virus (HCV) after orthotopic liver transplantation is almost universal in HCV-infected patients. Histological examination of liver biopsy specimens can be variable in distinguishing between recurrent hepatitis C and acute cellular rejection. The purpose of this study is to determine whether hepatic HCV RNA levels can be used to distinguish rejection from recurrent HCV by determining whether hepatic HCV RNA levels correlate with histological characteristics and clinical course. Seventy-two biopsy specimens were evaluated from 36 liver transplant recipients with HCV and elevated liver-related enzyme levels. Based on histological findings and clinical response to therapy, patients were defined as belonging to 1 of 5 groups: (1) definite rejection, (2) probable rejection, (3) indeterminate findings, (4) probable HCV, and (5) definite HCV. Hepatic HCV RNA was quantified using the Amplicor Monitor assay (Roche Diagnostic Systems Inc, Branchburg, NJ). There was a difference across groups in HCV RNA levels (P.046). The median HCV RNA level was 10,695 copies/mg of tissue DNA in the definite-hcv group compared with 1,024 copies/mg of tissue DNA in the definite-rejection group. Using pairwise comparisons, significant differences were found between definite HCV and definite rejection, probable HCV and definite rejection, probable HCV and probable rejection, and probable HCV and indeterminate. Our findings support the following conclusions. (1) In liver transplant recipients, hepatic HCV RNA levels are statistically greater in patients with recurrent HCV than rejection, although there is considerable overlap between groups. (2) Patients with low HCV RNA levels were unlikely to have recurrent HCV. (3) Patients with minimal and indeterminate findings on biopsy (group 3) had low HCV RNA levels. (Liver Transpl 2001;7: ) Cirrhosis secondary to hepatitis C virus (HCV) infection is the most common indication for orthotopic liver transplantation (OLT) in the United States and the world. HCV persists in almost all individuals after OLT, and recurrence of HCV hepatitis is a significant cause of morbidity and graft loss. Distinguishing recurrent HCV hepatitis from acute cellular rejection (ACR) in the transplanted liver can be difficult because the histological features of the 2 conditions can overlap. At the same time, the distinction is critical because ACR is treated with an increase in immunosuppressive medication and recurrent HCV hepatitis is treated with a decrease of the same. Serum HCV RNA levels have been shown to correlate with recurrent HCV hepatitis by some investigators, 1-4 but not others. 5-7 The current study endeavors to determine whether the use of hepatic HCV RNA levels from liver biopsy specimens after OLT can help distinguish between the 2 entities. In this study, we evaluated liver biopsy specimens and clinical course of 36 patients with hepatitis C who underwent OLT at our institution. Each biopsy specimen was classified based on histological and clinical criteria, including response to medical intervention. Biopsy specimens indicative of either recurrent HCV or ACR were compared with those characteristic of but less definitive for recurrent HCV or ACR and with those in which findings were mild or indeterminate. We examined whether differences in hepatic HCV RNA levels among these groups could help distinguish recurrent HCV and ACR. From the Divisions of *Gastroenterology, Pathology and Transplant Surgery, Henry Ford Health System, Detroit, MI; and the Biostatistics Center, George Washington University, Rockville, MD. Address reprint requests to Dilip K. Moonka, MD, Division of Gastroenterology, K-7, Henry Ford Hospital, 2799 West Grand Blvd, Detroit, MI Telephone: ; FAX: ; dmoonka1@hfhs.org Copyright 2001 by the American Association for the Study of Liver Diseases /01/ $35.00/0 doi: /jlts Materials and Methods Study Design Thirty-six patients who underwent OLT for hepatitis C related liver disease between 1990 and 1998 were analyzed retrospectively in this study. Fourteen patients contributed 1 biopsy specimen; 9 patients, 2 biopsy specimens, 12 patients, 3 biopsy specimens; and 1 patient, 4 biopsy specimens. There were 1 Asian, 8 black, and 27 white patients. Seventytwo biopsy specimens were analyzed for histological characteristics and hepatic HCV RNA levels. Percutaneous liver biopsies are performed at our institution for abnormalities in 436 Liver Transplantation, Vol 7, No 5 (May), 2001: pp

2 Hepatic HCV RNA to Distinguish Rejection From HCV 437 liver-related enzyme levels. Protocol biopsies were not performed. Fifty-four OLTs were performed in HCV-infected patients during this period. Selection was determined primarily by the availability of liver biopsy specimens for HCV RNA analysis. Each biopsy specimen was assigned to 1 of 5 groups depending on histological findings and response to clinical intervention: (1) definite rejection, (2) probable rejection, (3) indeterminate, (4) probable recurrent HCV, and (5) definite recurrent HCV. Hepatic HCV RNA levels were determined from the same biopsy specimens. Histological and Clinical Evaluation The 5 groups were defined in the following manner. For both ACR and recurrent HCV, we defined 3 histological criteria for each diagnosis that were generally believed to be consistent with that diagnosis. Histological features believed to be consistent with recurrent HCV included acidophilic bodies, predominant portal lymphocytic infiltrate, and lobular hepatitis. Features consistent with ACR included mixed cellular infiltrate in the majority of portal tracts, endothelialitis, and bile duct damage. We defined the definite groups (groups 1 and 5) as meeting at least 2 of the 3 histological criteria for ACR or recurrent HCV, along with an appropriate clinical course for the diagnosis. We defined the probable groups (groups 2 and 4) as meeting at least 1 of the 3 histological criteria for either recurrent HCV or ACR, together with an appropriate clinical course. The indeterminate group was defined as having mixed and relatively mild histological findings. Patients were believed to have a clinical course consistent with recurrent HCV if their liver-related enzyme levels improved over 2 months with a reduction in immunosuppression. A clinical course consistent with ACR included a persistent improvement in liver-related enzyme levels over 2 months after an increase in immunosuppression. Histological findings of each liver biopsy specimen were reviewed in a blinded fashion. Clinical records of each patient were also reviewed, and the following data were collected at the time of biopsy: serum alanine aminotransferase (ALT) level, serum total bilirubin level, time from OLT, type of primary immunosuppressant, steroid dose, recent treatment for rejection, and race. Quantification of Liver HCV RNA by Reverse- Transcriptase Polymerase Chain Reaction The amount of HCV RNA in each biopsy specimen was quantified by reverse-transcriptase polymerase chain reaction (RT-PCR) using the Amplicor Monitor assay (Roche Diagnostic Systems Inc, Branchburg, NJ). This system uses a quantitation standard (QS) to quantitate the amount of HCV RNA present. The QS is a synthetic RNA molecule with PCR primer binding sites identical to the target HCV RNA but with a unique internal sequence. Our modification of the technique for analysis of liver tissue is described in detail elsewhere. 8 Briefly, liver biopsy specimens were deparaffinized in xylene, minced, and digested with proteinase K. The QS was added at this time. RNA was extracted using guanidine thiocyanate, precipitated with alcohol, and then subjected to RT-PCR. Results were expressed as micrograms of DNA in tissue because we have no accurate method to weigh tissue that has been paraffin embedded. Statistical Analysis Repeated-measures ANOVA (SAS PROC MIXED; SAS Institute Inc, Cary, NC) was used to compare HCV RNA levels and other clinical measures across the 5 clinical outcome groups. The analysis tests the null hypothesis that there is equality across groups, appropriately adjusting for multiple observations per patient. If the null hypothesis was rejected, pairwise comparisons were examined. HCV RNA levels were transformed to log 10 values for all analyses because of extreme values and skewed distribution on the original scale. Log 10 values were normally distributed. Fisher s exact test was used to test tabulated data. All tests are 2-sided. Statistical significance is P less than.05, unless noted. Results Seventy-two biopsy specimens were available from 36 patients. Table 1 lists characteristics for the different groups at the time of biopsy. When the sample was restricted to first biopsy specimen per patient, the breakdown by race and age was not significant across groups. Serum ALT levels were not different among groups. Bilirubin level differed by group: the indeterminate group had significantly lower values than the definite-hcv group (P.005), definite-rejection group (P.031), and probable-rejection group (P.048). The difference among groups in time from OLT was marginally significant (P.078). Figure 1 shows HCV RNA levels in biopsy specimens plotted against time. Symbols are used to show whether the patient had recurrent HCV or rejection based on the histological and clinical criteria described in Materials and Methods. Rejection typically occurs within the first 3 to 4 months after OLT, whereas recurrent HCV is seen throughout the post-olt period. Hepatic HCV RNA Levels by Group Table 1 lists descriptive statistics for hepatic HCV RNA levels across the 5 histological and clinical outcome groups. Figure 2 shows the distribution on the original scale of both original data points and box and whisker plots. Again, HCV RNA levels showed a wide range (the probable-rejection group in particular) and assumed a normal distribution only when expressed in logs. Differences across groups were significant (P.046). The groups, ranked from highest to lowest,

3 438 Gottschlich et al Table 1. Patient Demographics Definite Rejection Probable Rejection Indeterminate Probable HCV Definite HCV P No. of biopsies Age (yr) ALT (IU/L) Bilirubin (mg/dl) Time from OLT (d) Steroid dose (mg/drug) TAC/CSA 5/3 5/2 12/15 9/13 2/5.453 Mean tissue HCV RNA (log10 copies/ g DNA) Median tissue HCV RNA as natural value ( g DNA) 1,024 2, ,883 10,695 Minimum and maximum values 0-59, , , ,941 2, ,762 NOTE. Mean HCV RNA levels are expressed in log values and all statistical analysis was performed in log values because only log values are normally distributed. Repeated-measures ANOVA was used to compare values among groups. Abbreviations: TAC, tacrolimus; CSA, cyclosporine. are definite HCV, probable HCV, probable rejection, definite rejection, and indeterminate. Pairwise analyses yielded the following significant differences: definite HCV greater than definite rejection (P.030), probable HCV greater than definite rejection (P.014), probable HCV greater than probable rejection (P.040), and probable HCV greater than indeterminate (P.014). The difference between definite HCV and probable rejection is marginally significant (P.052). In a separate analysis, we combined the 2 rejection groups (groups 1 and 2) and the 2 recurrent-hcv groups (groups 4 and 5). We then repeated the statistical evaluation using 3 groups: rejection (n 15), indeterminate (n 28), and recurrent HCV (n 29). The mean in log 10 values for the rejection group is , and for the recurrent HCV group, Differences across groups were now significant with P of.007. In pairwise analysis, differences were seen between the recurrent-hcv and indeterminate groups (P.002) and the recurrent-hcv and rejection groups (P.003). Figure 1. Variations in hepatic HCV RNA levels from the time of OLT. Each point represents a single biopsy specimen with the HCV RNA titer charted (using a log scale) on the y-axis against the time from OLT to biopsy (in days) on the x-axis. All 72 biopsy specimens are represented. In addition, different symbols signify which of the 5 histological/clinical groups the patient was most consistent with. Rejection typically occurred within the first few months after OLT, whereas recurrent HCV occurred at all time intervals. No trend was noted in hepatic HCV RNA levels over time.

4 Hepatic HCV RNA to Distinguish Rejection From HCV 439 doses were greater in patients in the probable-rejection group than the probable-hcv group (P.025). Finally, there was no overall relationship between hepatic HCV RNA levels and treatment of rejection within the preceding 90 days with either OKT3, antithymocyte antibody, or bolus steroid treatment. Figure 2. Hepatic HCV RNA levels expressed by group using dot plots and box and whisker figures. In the box and whisker figures, whiskers represent minimum and maximum values and the box represents values with the 25th to 75th percentiles. The solid bar represents median values. The greatest values are seen in the definite-hcv group (category 5) and the lowest values, in the indeterminate group (category 3). Immunosuppression At our center between 1990 and 1997, there was a shift from the use of cyclosporine to tacrolimus. Starting in 1997, tacrolimus was used as the primary calcineurin inhibitor. Patients believed to experience adverse effects attributed to this medication were converted to cyclosporine according to a specific protocol. Antithymocyte globulin and OKT3 were used for steroid-resistant rejection. There is no relationship between primary immunosuppressant used (cyclosporine versus tacrolimus) and likelihood of being in the rejection, recurrent-hcv, or indeterminate groups (Table 1; P.453). Patients on tacrolimus therapy were no more prone to develop recurrent HCV (or rejection) than patients on cyclosporine therapy. A separate analysis was performed to determine whether there was a difference in hepatic HCV RNA levels within groups depending on the use of cyclosporine or tacrolimus. For this analysis, 3 groups were analyzed: the 2 rejection groups were pooled, as were the 2 HCV groups (Fig. 3). For biopsy specimens in the recurrent-hcv groups, hepatic HCV RNA levels were significantly greater if patients were treated with tacrolimus than if cyclosporine was the primary immunosuppressant (P.024). There were no differences in HCV RNA levels between treatment with tacrolimus and cyclosporine within the indeterminate (P.923) or rejection (P.610) groups. However, steroid doses differed by group. Patients with definite rejection were receiving greater steroid doses than patients in both the definite-hcv (P.045) and probable-hcv (P.014) groups, and Discussion The management of patients who underwent OLT for hepatitis C is complicated by the difficulty distinguishing recurrent HCV from ACR after OLT. The distinction between the 2 entities may be difficult because some histological features of recurrent HCV and ACR overlap, and the 2 entities may coexist in some settings. At the same time, correctly distinguishing between the 2 entities is critical because they are treated in an opposed manner. ACR is treated with an increase in immunosuppressive medications, and recurrent HCV is treated with a decrease. The current study examines whether hepatic HCV RNA levels can help distinguish between the 2 entities. The current study uses a design that retrospectively tries to use histological criteria, as well as response to clinical intervention, to place each biopsy result into 1 of 5 groups: definite rejection, probable rejection, indeterminate or mild findings, probable HCV, and definite Figure 3. Hepatic HCV RNA levels within the rejection, indeterminate, and recurrent-hcv biopsy specimens as a function of the primary immunosuppression used. Within the recurrent-hcv group only, patients administered tacrolimus (TAC) had significantly greater viral RNA levels than those administered cyclosporine (CSA). The total number of biopsy specimens is 70 (instead of 72) because 2 patients were not administered tacrolimus or cyclosporine at the time of biopsy.

5 440 Gottschlich et al HCV. An effort then is made to determine whether there is a pattern of hepatic HCV RNA levels across the different groups and whether these levels can provide information to distinguish recurrent HCV and ACR. Our data, shown in Fig. 2, find a statistically significant difference across groups and in pairwise comparisons between groups. Greatest levels of viral RNA were found in the recurrent-hcv groups. Lowest levels were found in the indeterminate group, followed by the definite-rejection and probable-rejection groups. Most telling is the observation that biopsy specimens from the recurrent-hcv groups are unlikely to have low levels of hepatic HCV RNA, and finding low hepatic HCV RNA levels argues against recurrent HCV. The findings are tempered by the observation that although HCV RNA levels in the rejection group are lower than in the recurrent-hcv groups, there is a wide range of levels in the rejection groups, and some biopsy specimens contained high levels of viral RNA. There are several potential explanations for this observation. One explanation is that in some of these biopsy specimens, recurrent HCV coexists with ACR. Another explanation is that in some patients with rejection, the resulting inflammatory milieu provides an environment conducive to viral replication. A final explanation is that in some of these individuals, the predominant process is recurrent HCV and not rejection. At this time, we are not in a position to distinguish between these and other possibilities. There are other observations of note. There were no differences among groups in age, race, or serum ALT levels. In particular, patients in the indeterminate group, who overall had mild findings on biopsy and low hepatic HCV RNA levels, had serum ALT values similar to the other groups. This differed from serum bilirubin levels, for which the indeterminate group had lower levels than either the rejection or recurrent-hcv groups. There was some difference in groups in time from OLT. Consistent with clinical experience, the longer the time from OLT, the less likely a biopsy specimen was to be consistent with rejection (Fig. 1). Recurrent HCV was seen at almost all time points after OLT. We did not find either an overall increase or decrease in hepatic HCV RNA levels over time. Some differences were noted in immunosuppression among groups. Steroid doses were significantly greater in the rejection groups than in the recurrent-hcv groups because patients were administered greater steroid doses soon after OLT when rejection and low hepatic HCV RNA levels were more common. We did not find differences among groups in the primary immunosuppressant used. However, in patients with recurrent HCV, those on tacrolimus therapy had significantly greater levels of hepatic HCV RNA than those on cyclosporine therapy. This was not seen in biopsy specimens from the indeterminate or rejection groups and suggests that the use of tacrolimus may result in enhanced viral replication only in patients who develop recurrent HCV (Fig. 3). Previous reports examined the relationship between serum HCV RNA levels and recurrent HCV after OLT. Some investigators found a correlation between serum HCV RNA levels and the presence and severity of recurrent HCV hepatitis, 1-4 but an almost equal number found no correlation. 5-7 In some reports in which a positive correlation was noted, the investigators noted that recurrent HCV was typically associated with high viral RNA levels, but not all individuals with high viral RNA levels had recurrent HCV. 1,4 This observation is similar to our conclusions about hepatic HCV RNA levels. Several studies examined hepatic HCV RNA levels after OLT. Di Martino et al 9 were the first to do so. They analyzed 84 biopsy specimens from 33 patients and showed that patients with HCV who developed lobular hepatitis after OLT had significantly greater levels of hepatic HCV RNA than patients with normal histological characteristics. A report from Miami also measured hepatic HCV RNA levels in patients who developed recurrent HCV hepatitis after OLT. 10 The investigators examined 23 biopsy specimens and found that patients with recurrent hepatitis C had significantly greater hepatic HCV RNA levels than those who did not. It is critical to note that many of these studies were primarily designed to determine the relationship between HCV RNA levels and the natural history of hepatitis C after OLT. They were not designed to determine whether hepatic HCV RNA levels can distinguish recurrent HCV or rejection at a single time point. Di Martino et al 9 specifically excluded patients with features of rejection from their analysis. The current study has important limitations. We used formalin-fixed paraffin-embedded liver biopsy specimens. Previous reports have specifically evaluated the stability of HCV RNA in liver tissue in this setting. A report from the Mayo Clinic found no loss of HCV RNA detection in tissue fixed in formalin for less than 24 hours, and in this analysis, long-term paraffin-block storage had no effect on HCV RNA detection. 11 There was significant loss of HCV RNA in tissue fixed for more than 24 hours. In our study, specimen fixation was consistently less than 8 hours. Moreover, because our fixation method is uniform, if degradation occurred, it would have been equal in both the recurrent-

6 Hepatic HCV RNA to Distinguish Rejection From HCV 441 HCV and rejection groups. Although absolute levels of hepatic HCV RNA might be affected, comparisons among groups should remain valid. Finally, we performed an internal analysis to determine whether there was a correlation between level of hepatic HCV RNA and age of the biopsy specimen. No correlation was found. There are additional limitations. In designing this study, we were faced with the difficulty that there is no gold standard for either recurrent HCV or ACR. We attempted to define a set of histological and clinical criteria that would be generally acceptable as a starting point in this type of analysis. If there had been no differences in hepatic HCV RNA levels among groups, one interpretation would be that our criteria were not valid. We did not have a sufficient number of paired serum and biopsy specimens in this study to determine whether hepatic HCV RNA levels mirror or differ from serum levels. Finally, although we found hepatic HCV RNA levels useful to help distinguish between recurrent HCV and ACR at our center, there is typically a delay of 48 to 96 hours from the time of biopsy to the time data are available. Our findings support the following conclusions. We found hepatic HCV RNA levels to be a useful adjunct to histological examination to distinguish between recurrent HCV and rejection, but the utility is limited by overlap between groups and the ability to obtain data in real time. In liver transplant recipients with HCV and elevated liver-related enzyme levels, hepatic HCV RNA levels are statistically greater in patients with recurrent HCV than rejection, although there is considerable overlap between groups. Patients with low HCV RNA levels are unlikely to have recurrent HCV. Patients with recurrent HCV tend to have high HCV RNA levels, but high levels are also seen in some patients with rejection. Patients with minimal and indeterminate findings on biopsy despite elevated liver-related enzyme levels have low HCV RNA levels. The likelihood of having rejection versus recurrent HCV was not influenced overall by immunosuppression, but patients with recurrent HCV on tacrolimus therapy had greater HCV RNA levels than patients with recurrent HCV on cyclosporine therapy. References 1. Gane EJ, Naoumov NV, Qian KP, Mondelli MU, Maertens G, Portmann BC, et al. A longitudinal analysis of hepatitis C virus replication following liver transplantation. Gastroenterology 1996;110: Feray C, Gigou M, Samuel D, Paradis V, Wilber J, David MF, et al. The course of hepatitis C virus infection after liver transplantation. Hepatology 1994;20: Papatheodoridis GV, Barton SG, Andrew D, Clewley G, Davies S, Dhillon AP, et al. Longitudinal variation in hepatitis C virus (HCV) viraemia and early course of HCV infection after liver transplantation for HCV cirrhosis: The role of different immunosuppressive regimens. Gut 1999;45: Moll C, Kahler C, Edis C, Nachbauer K, Konigsrainer A, Spechtenhauser B, et al. Early-level hepatitis C viremia after orthotopic liver transplantation is of prognostic significance. Transplant Proc 1998;30: Chazouilleres O, Kim M, Combs C, Ferrell L, Bacchetti P, Roberts J, et al. Quantitation of hepatitis C virus RNA in liver transplant recipients. Gastroenterology 1994;106: Freeman RB, Tran S, Lee YM, Rohrer RJ, Kaplan MM. Serum hepatitis C RNA titers after liver transplantation are not correlated with immunosuppression or hepatitis. Transplantation 1996;61: Zhou S, Terrault NA, Ferrell L, Hahn JA, Lau JY, Simmonds P, et al. Severity of liver disease in liver transplantation recipients with hepatitis C virus infection: Relationship to genotype and level of viremia. Hepatology 1996;24: Aardema KL, Nakhleh RE, Terry LK, Burd EM, Ma CK, Moonka DK, et al. Quantitation of hepatitis C virus RNA in liver tissue of allografts: Correlation with histologic features and liver function tests. Transplant Proc 1999;31: Di Martino V, Saurini F, Samuel D, Gigou M, Dussaix E, Reynes M, et al. Long-term longitudinal study of intrahepatic hepatitis C virus replication after liver transplantation. Hepatology 1997:26: Cirocco R, Fragulidis GP, Weppler D, Zucker K, Markou M, Zervos XA, et al. Recurrent hepatitis C infection following orthotopic liver transplantation is predicted by posttransplant serum and hepatic allograft viral titers. Transplant Proc 1997;29: Guerrero RB, Batts KP, Brandhagen DJ, Germer JJ, Perez RG, Persing DH. Effects of formalin fixation and prolonged block storage on detection of hepatitis C virus RNA in liver tissue. Diagn Mol Pathol 1997;6: