Effects of interferon treatment on liver histology and allograft rejection in patients with recurrent hepatitis C following liver transplantation



Recurrent hepatitis C after liver transplantation remains a significant cause of graft loss and retransplantation. Although treatment of recurrent hepatitis C with interferon-based regimens has become widely accepted as safe and can lead to sustained virologic clearance of hepatitis C virus (HCV) RNA, long-term histologic improvement and the risk of precipitating graft rejection remain controversial. The present study is a retrospective evaluation of the clinical and histological consequences of treating recurrent hepatitis C with interferon-based therapy in a selected group of liver transplant recipients. Twenty-three liver transplant recipients with recurrent hepatitis C and histologic evidence of progressive fibrosis completed at least 6 months of interferon, 83% of whom received pegylated-interferon α-2b; only 4 tolerated ribavirin. Overall, 11 patients (48%) had undetectable HCV RNA at the end of 6 months of treatment. Of these patients, 3 remained HCV RNA–negative on maintenance interferon monotherapy for 33 months, and the other 8 (35%) completed treatment and remained HCV RNA–undetectable 24 weeks after discontinuation of interferon. Overall necroinflammatory activity in liver biopsies obtained 2 years after HCV RNA became undetectable decreased significantly (7.73 ± 2.37 vs. 5.64 ± 2.94 units before and after treatment, respectively; P = .016). However, 5 of these 11 patients had no histologic improvement in follow-up liver histology. Liver biopsies in the 12 nonresponders demonstrated disease progression. Of the 23 patients treated with interferon, 8 (35%) had evidence of acute or chronic rejection on posttreatment liver biopsy, most of whom had no previous history of rejection (P < .01 for comparison of pretreatment and posttreatment prevalence of histologic rejection), and 2 experienced graft loss from chronic rejection, requiring retransplantation. In conclusion, interferon treatment of recurrent hepatitis C does not consistently improve histologic disease after virologic response, and it may increase the risk of allograft rejection. (Liver Transpl 2004;10:850–858.)

Graft reinfection with hepatitis C virus (HCV) occurs nearly universally after liver transplantation for cirrhosis due to chronic hepatitis C. The progression of fibrosis in recurrent hepatitis C varies widely among individuals,1–3 but is often accelerated compared to progression in nontransplant patients: up to 28% develop cirrhosis within five years of transplantation.4 Although earlier series suggested that both graft and patient survival after liver transplantation for hepatitis C were similar to those observed in patients with other liver diseases, more recent data demonstrate that post–liver transplant survival is reduced in patients with chronic hepatitis C.5 Inevitably, the problem of recurrent hepatitis C will escalate in the next two decades as the epidemic of hepatitis C fully matures and more patients require liver transplantation,6 and since the progression of recurrent hepatitis C may itself be accelerating.7

As with nontransplant patients, interferon (IFN) based antiviral therapy has been used to treat recurrent hepatitis C among the transplant population but with scant knowledge of the long-term safety and efficacy of such treatment. Most published reports have concluded that sustained virological response (SVR) is lower in the transplant population than in the nontransplant population (0%-33%),8–18 but these disappointing results have not included the use of more effective pegylated IFN (PEG-IFN). Moreover, histologic improvement after antiviral therapy has been observed inconsistently, although only one series15 has provided histologic follow-up beyond the completion of treatment. In many grafts, necroinflammatory activity persisted, and fibrosis either remained unchanged or it progressed, even after achieving SVR.8–12, 15

Liver allograft rejection after IFN treatment for recurrent hepatitis C has not been a frequently recognized complication of antiviral therapy in the most recent series.12–14, 17, 18 However, an “acute vanishing bile duct syndrome,” chronic rejection, and the need for retransplantation have been reported in up to 35% of patients treated for recurrent hepatitis C with IFN in earlier series.10, 19 The present retrospective series explores the histologic effects of IFN therapy in all liver transplant recipients treated for recurrent hepatitis C at our center between 1998 and 2002, and whether such treatment was associated with acute or chronic allograft rejection.


HCV, hepatitis C virus; IFN, interferon; PEG-IFN, pegylated-interferon; RVN, ribavirin; SVR, sustained virologic response; ALT, alanine aminotransferase; NS, not significant.

Patients and Methods


Two hundred patients with cirrhosis due to chronic hepatitis C underwent liver transplantation at Virginia Commonwealth University between 1993 and 2002. All patients with recurrent viral hepatitis C after transplantation (HCV RNA–positive) underwent protocol liver biopsy at an interval of every 6–24 months. Patients were selected to undergo IFN-based therapy between 1998 and 2002 if they were at least 12 months post–liver transplant (mean, 43 months; range, 12-101 months) and had histologic evidence of ongoing necrosis and inflammation as well as evidence of progressive fibrosis (Knodell score ≥ 6). Patients were excluded from treatment for severe neutropenia (ANC <1,500/mm3), thrombocytopenia (platelet count <50,000/mm3), had rejection within the past 12 months, or had evidence of concurrent viral infections (hepatitis B virus, human immunodeficiency virus, cytomegalovirus). The present report describes the histologic findings of 23 patients who were selected on the basis of completion of at least 24 weeks of IFN-based antiviral therapy to assess histologic changes that occurred after this therapy. The 23 patients included 6 women and 4 recipients of right liver lobes from living donors. One patient with alcoholic cirrhosis was infected with HCV by her graft. Another patient underwent retransplantation for graft failure due to hepatitis C. Immunosuppression during the period 1998-2002 remained uniform: All patients received a calcineurin inhibitor (cyclosporin A or tacrolimus) maintained in therapeutic range, and all had been tapered off corticosteroids at 6 months after liver transplantation. Informed consent was obtained from all patients.

Treatment and Follow-up

Prior to the availability of PEG-IFN, patients were treated with IFN α-2b (3 mIU SC TIW; Intron A, Schering-Plough, Kenilworth, NJ). After PEG-IFN α-2b became available in 2000, 10 patients were switched to PEG-IFN (PEG-Intron, 1.0-1.5 μg/kg/week) treatment. Only 4 patients received standard IFN throughout treatment. Ribavirin (RVN) was started in 8 patients at 600 mg/d and titrated upward to1,000 mg/d as tolerated, but it was discontinued in 4 patients for severe anemia. Two patients received erythropoietin (Procrit, 40,000 units SC/week) to treat anemia caused by RVN.

At 24 weeks of treatment, those patients infected with HCV genotypes 2 or 3 and most of those who failed to become HCV RNA–undetectable (nonresponders) stopped therapy. Treatment was continued for at least 48 weeks in patients infected with genotype 1 who were HCV RNA–undetectable at week 24. Eight patients with undetectable HCV RNA at the end of treatment were followed for at least 6 months to monitor for relapse, and all remained HCV RNA–undetectable (sustained virologic responder [SVR] rate, 35%); no patients relapsed after discontinuation of antiviral therapy. Three patients with undetectable HCV RNA at 24 weeks were maintained on IFN or PEG-IFN monotherapy for a mean of 33 months because of pretransplant nonresponse to IFN (n = 1), or pretransplant relapse after discontinuation of IFN (n = 1) or IFN/RVN (n = 1). For purposes of data analysis, these 3 patients were included with the 8 patients with SVR, because HCV RNA remained undetectable during monitoring every 3 months. Four nonresponders also continued IFN maintenance therapy for 18–24 months for potential histologic benefit.

Liver Allograft Histology

Liver biopsy was performed within 3 months prior to starting antiviral therapy in all patients. Patients who became HCV RNA–undetectable underwent follow-up liver biopsy 23 months (mean) after first undetectable HCV RNA, 10 months (mean) after discontinuation of IFN. Nonresponders underwent follow-up liver biopsy within 2 months of discontinuing treatment. Paired liver biopsies from each patient were reviewed by experienced liver transplant pathologists (A.S.M. and M.C.) blinded to IFN treatment and HCV RNA status, and the severity of recurrent hepatitis C was scored according to Knodell.20 Histologic improvement in paired biopsies was defined as ≥2 unit decrease in necroinflammatory activity without an increase in fibrosis. Biopsies were also examined for evidence of acute and chronic allograft rejection according to standard criteria.21, 22 Acute rejection was assessed by degree of bile duct injury, venous endothelial inflammation, and centrilobular necrosis, but not given a formal Banff score because of the inability to assign a relative contribution of rejection to the chronic portal inflammation of hepatitis C. Chronic rejection was defined by the presence of bile duct atrophy/pyknosis, bile duct paucity (ducts absent in >50% of portal triads examined on at least 3 sections of the biopsy specimen), and foam cell obliterative arteriopathy.


HCV genotyping was performed by the line-probe hybridization assay (Inno-LiPA assay; Innogenetics, Gent, Belgium). Three patients who began treatment in 1998 did not have genotypes determined. HCV RNA was measured by the Roche Amplicor assay (Roche Diagnostics, Indianapolis, IN)(lower limit of sensitivity 600 copies/mL; upper range reported >8.5 × 105).

Statistical Analysis

Statistical significance of differences in parametric variables before and after treatment was assessed using Student t test for paired samples. Differences between unpaired data were determined by univariate ANOVA. Differences in nonparametric variables were assessed using Pearson's chi-square test. Analyses were performed using SPSS 11.5.0 for Windows (SPSS, Cary, NC).


Predictors of Virologic Response to Antiviral Therapy and Effects on Alanine Aminotransferase (ALT) and Viral Load

After 24 weeks of IFN treatment, 11 of 23 (48%) patients had undetectable HCV RNA levels, and 12 were nonresponders. Eight of the 11 responders completed antiviral therapy, and remained HCV RNA–undetectable for 6 months after discontinuation of antiviral therapy (SVR rate, 35%). Three responders were maintained on IFN with undetectable HCV RNA documented every 3 months. Baseline patient variables prior to antiviral treatment are shown for the 23 patients in Table 1. Only pretreatment viral load and genotype non-1 predicted becoming HCV RNA–undetectable at week 24. Mean pretreatment viral load for those patients with undetectable HCV RNA and nonresponders was 5.18 × 105 and 8.12 × 105, respectively (P = .01). All 3 patients infected with genotypes 2 / 3 achieved sustained response (P = .02), as did 3 patients with unknown genotypes; all other patients were infected with genotypes 1a / 1b. Although donor age strongly correlated with pretreatment fibrosis score (mean, 1.4 for donor age >35 yr and 2.8 for donor age >35 yr, respectively; P < .001; data not shown), this parameter failed to predict virologic response. The mean age at transplant, time after transplant to initiation of antiviral therapy, and immunosuppression (calcineurin inhibitor regimen and mean serum levels) were also not significantly different in patients with undetectable HCV RNA at 24 weeks vs. nonresponders. Finally, mean ALT within 1 year prior to treatment and necroinflammatory activity and fibrosis scores on pretreatment liver biopsy also failed to predict response to antiviral therapy.

Table 1. Pretreatment Characteristics of Liver Transplant Patients With Recurrent Hepatitis C Based Upon Virologic Response to IFN ± RVN Treatment
Pretreatment VariableHCV RNA (−) N = 11HCV RNA (+) N = 12P Value
  1. Abbreviations: NS, not significant at P < .05; ND, not determined. NOTE. Data are mean ± SD unless otherwise indicated. CYA and FK506 levels represents mean values within 6 months before antiviral therapy.

Age at transplant, (yr)48.4 ± 6.948.1 ± 5.3NS
Donor age, (yr)30.6 ± 11.238.5 ± 13.0NS
Viral load pretreatment, (IU/mL)5.2 × 1058.1 × 105.01
HCV genotype 2b/3a/ND, (n)2/1/30/0/0.02
Time of treatment after transplant, (mo)35.5 ± 24.349.6 ± 29.8NS
Treatment with RVN (partial/complete), (n)3/21/2NS
Immunosuppression, (n)   
CYA level, (ng/mL)189 ± 86194 ± 73NS
FK506 level, (ng/mL)10.2 ± 1.08.6 ± 2.0NS
ALT, (U/L)127 ± 67126 ± 74NS
Liver biopsy necroinflammatory activity7.7 ± 2.47.8 ± 1.4NS
Liver biopsy fibrosis score1.9 ± 1.02.4 ± 1.0NS

Mean viral load in nonresponders before and after treatment was not significantly different (8.12 × 105 vs. 6.57 × 105 IU/mL, respectively; P = .18), although 2 nonresponders had a >2 log10 decrease in viral load. Mean ALT in those patients with undetectable HCV RNA decreased significantly from 127 ± 67 to 54 ± 54 U/L (P = .002), and 3 patients had persistently elevated ALT despite achieving SVR. These 3 patients had evidence of rejection on posttreatment biopsy (see below). In contrast, mean ALT in nonresponders was unchanged by treatment (126 ± 74 and 109 ± 60 U/L, respectively; P = not significant [NS]).

Effects of Antiviral Treatment on Liver Graft Histology

The 11 patients who were HCV RNA–undetectable at week 24 completed therapy (8 patients) or were maintained on IFN (3 patients) as described in Methods, with undetectable HCV RNA documented every 3 months. Liver biopsy was repeated a mean of 23 months after first undetectable viral load in all 11 patients. Necroinflammatory activity in these paired biopsies decreased significantly (7.73 ± 2.37 vs. 5.64 ± 2.94 in pretreatment and follow-up biopsies, respectively; P = .016; Fig. 1). However, 7 of these 11 follow-up biopsies (64%) still exhibited moderate-tosevere necroinflammatory activity (score, 5-11), and histologic response to antiviral therapy, defined as a decrease in necroinflammatory activity of at least 2 points without an increase in fibrosis, was observed in only 6 patients (55%). In contrast, mean necroinflammatory activity in nonresponders increased slightly after IFN treatment (7.75 ± 1.43 vs. 8.25 ± 2.23, P = NS; Fig. 1). Mean fibrosis scores in patients with undetectable HCV RNA decreased from 1.91 ± 1.04 to 1.55 ± 1.21 (P = NS), while fibrosis scores in nonresponders increased from 2.55 1.04 to 2.73 .90 after treatment (P = NS).

Figure 1.

Effects of IFN treatment on liver biopsy histology. Necroinflammatory activity (left panel) and fibrosis score (right panel) before and after interferon treatment for 11 patients with undetectable hepatitis C virus RNA and in 12 nonresponders. (Data expressed as mean ± SD).

Histologic Evidence of Rejection After Antiviral Therapy

Paired liver biopsies obtained before and after IFN treatment were also reviewed for histologic evidence of rejection. None of the 23 pretreatment biopsies demonstrated evidence of acute cellular or chronic rejection. In contrast, the liver biopsies of 8 patients (35%) demonstrated histologic evidence of rejection after IFN treatment (P < .01 compared to pretreatment). Posttreatment biopsies from 7 patients exhibited evidence of acute cellular rejection including bile duct injury with or without endothelialitis (Table 2). Acute rejection was mild in 4 patients, moderate in 2 patients, and severe in 1 patient, and it was treated according to severity (Table 2). Three patients underwent repeat liver biopsy after treatment for acute cellular rejection (ACR) and had resolution of bile duct injury and endothelialitis. One patient (7), a nonresponder to IFN who was 7 years post-OLT and had no previous episodes of rejection, developed severe acute cellular rejection near the end of her 6-month course of IFN (Figs. 2A 2B). Despite discontinuation of IFN and administration of steroid boluses and antithymocyte globulin, the patient developed rapidly progressive cholestasis (alkaline phosphatase, 3,030 U/mL; total bilirubin, 26.6 mg/dL) and required retransplantation 3 months after the acute rejection. Pathologic examination of the explant showed foam cell obliterative arteriopathy, bile duct atrophy, and bilirubinostasis characteristic of chronic rejection (Figs. 2C 2D). A second patient (8) 4 years post-OLT with 1 episode of ACR shortly after transplant developed progressive cholestasis (alkaline phosphatase, 956; total bilirubin, 3.8) after 18 months of IFN. Liver biopsy at the end of treatment revealed ductopenia (1 bile duct observed in 9 portal tracts), as did a follow-up liver biopsy 1 year after discontinuation of IFN. The patient subsequently developed ascites and has been relisted for liver transplantation.

Table 2. Histologic Evidence, Treatment, and Outcome of Rejection After IFN Therapy for Recurrent Hepatitis C
Pt. No.Evidence of Rejection on Post-IFN  Liver Biopsy (grade)Treatment of RejectionFollow-up Histology After  Treatment of Rejection
  1. Abbreviations: ND, not done; IV, intravenous; CYA, cyclosporine A; IFN, interferon; ATG, antithymocyte globulin.

1Bile duct injury, endothelialitis (mild)rapamycinND
2Bile duct injury, endothelialitis, centrilobular necrosis (moderate)rapamycin, prednisoneresolution of rejection
3Bile duct injury (mild)IFN withdrawalresolution of rejection
4Bile duct injury, endothelialitis (mild)IV steroids, ↑ FK506 doseresolution of rejection
5Bile duct injury, endothelialitis, centrilobular necrosis (moderate)IV steroids, ↑ CYA doseND (expired)
6Bile duct injury (mild)IFN withdrawl, ↑ FK506 doseND
7Bile duct injury, endothelialitis, centrilobular necrosis (severe)ATG, IV steroids, retransplantedbile duct pyknosis, arteriopathy (explant)
8Bile duct paucity (1 bile duct/ 9 portal tracts)listed for retransplantbile duct paucity
Figure 2.

Graft loss due to acute followed by chronic rejection in a patient treated with IFN monotherapy for severe hepatitis C. (A) Alanine aminotransferase, total bilirubin, and liver biopsy findings. Seven years after orthotopic liver transplantation, protocol liver biopsy showed severe necroinflammatory activity and portal fibrosis (Knodell score [3, 4, 3, 1]). Nearing the end of six months of interferon treatment, alanine aminotransferase and bilirubin rose abruptly. (B) Second liver biopsy to investigate increased bilirubin and alanine aminotransferase showed severe bile duct injury (small arrow) and portal venous endothelial inflammation (large arrow). Hematoxylin & eosin stain (C) Explant pathology: cross-section of the common hepatic artery 85% to 90% occluded with subintimal foamy macrophages. Verhoeff-VanGieson stain. (D) Explant pathology: bile duct atrophy/pyknosis (large arrow) and bilirubinostasis (small arrows). Hematoxylin & eosin stain.

Baseline variables evaluated as possibly contributing to rejection are shown in Table 3. Patients with evidence of rejection on the posttreatment biopsy underwent liver transplantation at an earlier age than patients without evidence of rejection (43.5 ± 4.5 vs. 50.7 ± 5.2 years, respectively; P < .01). Otherwise, there was no significant relationship between rejection and donor age, time of treatment after transplant, virologic response, length of IFN exposure, history of rejection before IFN treatment, nor calcineurin inhibitor levels. Similarly, the pretreatment ALT, necroinflammatory activity, and fibrosis scores were not significantly different in patients with and without histologic evidence of rejection after IFN treatment.

Table 3. Characteristics of Liver Transplant Patients With Recurrent Hepatitis C Treated With IFN According to Presence of Histologic Evidence of Rejection on Posttreatment Liver Biopsy
Patient VariableRejection N = 8No Rejection N = 15P Value
  • Abbreviations: NS, not significant (at P < .05); IFN, interferon; CYA, cyclosporine A; FK506, tacrolimus. NOTE. Data are expressed as mean ± SD unless otherwise indicated.

  • *

    P = .013 for difference before and after IFN.

Age at transplant, (yr)43.5 ± 4.550.7 ± 5.2.008
Donor age, (yr)37.0 ± 13.234.4 ± 13.2NS
Time of treatment after transplant, mo45.8 ± 28.641.3 ± 28.2NS
Virologic responders, n (%)3 (38)8 (53)NS
IFN treatment time, (mo)10.3 ± 7.718.4 ± 12.6NS
Previous rejection, n (%)2 (25)4 (27)NS
Immunosuppression, n (%)   
 CYA5 (63)9 (60)NS
 FK5063 (38)6 (40)NS
CYA level, (ng/mL)   
 Before IFN206 ± 71207 ± 50NS
 After/during IFN198 ± 50188 ± 50NS
FK506 level, ng/mL   
 Before IFN10.2 ± 1.18.9 ± 1.9NS
 After/during IFN6.0 ± 1.49.1 ± 2.8NS
ALT, (U/L)   
 Before IFN134 ± 84122 ± 65NS
 After IFN115 ± 8065 ± 45*.046
Necroinflammatory activity   
 Before IFN8.1 ± 2.07.5 ± 2.0NS
 After IFN7.9 ± 2.66.5 ± 2.6NS
Fibrosis score   
 Before IFN2.0 ± 1.12.4 ± 1.0NS
 After IFN2.3 ± 1.02.1 ± 1.2NS

Following IFN treatment, no significant differences were noted in necroinflammatory activity and fibrosis scores in patients with and without rejection (Table 3). Although mean ALT significantly improved after IFN treatment in patients who did not develop rejection (122 ± 65 vs. 65 ± 45 U/L before and after treatment, respectively; P = .014), ALT remained elevated in patients with histologic evidence of rejection posttreatment (134 ± 84 vs. 115 ± 80 U/L, respectively; P = NS).


In the absence of randomized trials in liver transplant recipients, recurrent hepatitis C has been treated with IFN with or without RVN based upon protocols extrapolated from the nontransplant literature. Only a small number of patients with recurrent hepatitis C have had graft histology reported after IFN-induced SVR, many of whom failed to improve necroinflammatory activity or fibrosis scores.10, 12, 15 Furthermore, it is generally accepted that IFN does not increase the risk of allograft rejection, even though early reports raised this concern.10, 16, 19 In the present series, 74% of liver transplant recipients treated with IFN for recurrent hepatitis C failed to show an improvement in posttreatment histology, including 45% of those with undetectable HCV RNA for a mean of 2 years. Furthermore, evidence of allograft rejection was observed in the posttreatment liver biopsies of 35% of these patients, including 2 patients with irreversible chronic rejection leading to graft failure. Admittedly, these retrospectively collected data are compromised by heterogeneity in immunosuppression, IFN treatment regimen, and timing of posttreatment liver biopsy. However, the availability of paired liver biopsies before and after IFN treatment allows each patient to serve as his or her own control.

The virologic clearance rate at treatment week 24 (48%) and SVR rate (35%) appear to compare favorably with previous studies of liver transplant recipients.8–17 However, our patient population was selected on the basis of completion of at least 6 months of antiviral therapy to assess liver histology after treatment. These response rates are nevertheless higher than expected based upon previous studies using nonpegylated IFN monotherapy, which have reported SVR rates of 0% to 10%.8–11 Since few of our patients tolerated RVN, prolonged treatment and the use of PEG-IFN probably contributed to these favorable results. Prolonged IFN treatment has previously been suggested to improve viral clearance rates in liver transplant recipients.23

In addition to genotype and low pretreatment viral load, which are well-recognized predictors of response to IFN in the nontransplant population, normalization of ALT on treatment predicted virologic response. Indeed, 73% of patients with undetectable HCV RNA had persistently normal ALT during and after IFN treatment. Those who cleared HCV RNA but did not normalize ALT on treatment (mean, 124 U/L) had evidence of rejection on posttreatment liver biopsy, emphasizing the importance of routine follow-up liver biopsy even in patients achieving SVR, especially in patients with modestly elevated ALT.

On average, modest but significant improvement in necroinflammatory activity (mean, 2 units) was observed in follow-up liver biopsies from patients who became HCV RNA–undetectable. However, 7 of 11 patients with undetectable HCV RNA demonstrated persistent moderate-to-severe necroinflammatory activity (≥5 units), 5 of whom had unchanged or worsened fibrosis on posttreatment histology. Although the number of patients with virologic response was small, and sampling variation in liver biopsy may be significant, these results suggest that the natural history of recurrent hepatitis C may not improve after successful viral eradication in many liver transplant recipients, in contrast to the nontransplant population in whom fibrosis regression has been repeatedly demonstrated. The absence of fibrosis regression in patients achieving SVR appears to be a universal finding in the present and previous studies, even up to 3 years after antiviral therapy.12, 24 Our results are also similar to one recent series24 in which 14 patients achieving SVR, on average, improved necroinflammatory activity on liver biopsies obtained at the end of treatment and in yearly follow-up. Although the conclusion of this study was that viral clearance led to histological improvement in transplant recipients with recurrent hepatitis C, persistent mild, moderate, and severe chronic hepatitis was observed in liver biopsies of 29%, 29%, and 6% of patients, respectively, 3 years after the end of IFN / RVN treatment.

The cause of persistent hepatitis and absence of fibrosis regression in many of our transplant recipients with undetectable HCV RNA after IFN treatment remains speculative. Persistent low-level hepatitis C infection might perpetuate necrosis and inflammation, since false-negative HCV RNA results can occur.25 However, all patients in our series were tested every 3–6 months after viral clearance and remained HCV RNA–undetectable, making this possibility unlikely. Furthermore, liver transplant recipients achieving SVR have recently been shown to clear HCV RNA from liver tissue.24 Chronic infection with other hepatotrophic viruses (hepatitis B virus, cytomegalovirus) were excluded by specific tests, but infections with cryptic viruses remain possible. Chronic hepatitis without an identifiable etiology has been observed in 20% to 40% of protocol biopsies, and the possibility of allograft rejection as an explanation for “idiopathic graft hepatitis” has been previously proposed.26 Our results support the role of rejection as a cause of persistent inflammation, since almost all patients with undetectable HCV RNA and no evidence of rejection on posttreatment liver biopsies met the definition of histologic improvement. In contrast, patients with undetectable HCV RNA without histologic improvement after IFN treatment usually had histologic evidence of rejection.

In the setting of recurrent hepatitis C, the diagnosis of allograft rejection must be made cautiously, since some of the classic histologic signs of rejection (portal inflammation, bile duct injury) can be simulated by hepatitis C. Several clinical and histologic features strongly suggest that the 8 patients described in Table 2 indeed had rejection. First, none had histologic evidence of rejection prior to IFN treatment. Second, patients achieving SVR with evidence of acute rejection on posttreatment liver biopsy had persistently elevated ALT, while those without rejection had normal ALT. Third, those patients with evidence of acute rejection after IFN who underwent follow-up liver biopsy after treatment for rejection had resolution of bile duct injury and endothelialitis. Fourth, 2 patients developed chronic rejection and allograft failure during IFN therapy. These cases were well documented clinically and histologically, one by pathologic examination of the explant, and the other by repeated liver biopsies showing duct loss (1 duct / 9 portal tracts) and progressive cholestasis. Finally, the prevalence of histologic evidence of rejection in our post-IFN treatment population (35% overall, 8.7% chronic rejection) is higher than in protocol biopsies from 159 untreated liver transplant recipients with recurrent hepatitis C from our center over the same time period (16% overall, 1.3% chronic rejection).

Several features of our treatment of recurrent hepatitis C may have contributed to the high incidence of histologic rejection after IFN therapy. Only a small proportion received full courses of RVN, which suppresses cultured immune cell proliferation27 and might have protected against rejection.14 The administration of IFN relatively late after transplant, when patients had been weaned off steroids and mycophenolate, very likely contributed to a higher than expected incidence of rejection, even though we found no strong relationship between the time to treatment after transplant and evidence of rejection. Treatment with PEG-IFN may have also increased the incidence of rejection as compared to published series using nonpegylated IFN. Finally, the high incidence of rejection may simply reflect the performance of protocol biopsies looking for evidence of subclinical rejection even in patients with normal ALT, as has been recently advocated for this purpose.28

The course of 2 of our nonresponders closely resembled case reports of 8 other liver transplant patients in the literature who received IFN for recurrent hepatitis C and developed chronic rejection.10, 16, 19 Clinically, these patients presented with progressive cholestasis usually within 1–6 months after initiating IFN, and follow-up liver biopsies revealed bilirubinostasis, bile duct atrophy or loss, and (in 4, including one of our patients) foam-cell obliteration of the hepatic artery.22 Several of these cases, including our own, evolved from acute cellular rejection, were unresponsive to discontinuation of IFN and antirejection therapy, and lead to death or retransplantation. The similarity of these reports from 4 transplant centers clearly suggests that IFN can trigger irreversible chronic rejection in some patients with recurrent hepatitis C.

We conclude that IFN treatment in liver transplant recipients with recurrent hepatitis C does not consistently improve allograft histology in patients who achieve viral clearance, even after months or years, and does not improve liver histology in nonresponders at all. Conversely, IFN therapy may increase the risk allograft rejection. Rejection following IFN may smoulder sub-clinically and be detectable only by liver biopsy, and it may be responsible for the absence of histologic improvement after viral clearance. Less commonly, IFN may precipitate chronic rejection and graft loss. The treatment of recurrent HCV with IFN requires further study in large multicenter trials with longer histologic follow-up before it can be regarded as safe and effective. Minimizing immunosuppression in patients with recurrent hepatitis C to slow its progression must also be critically reassessed before the initiation of antiviral therapy with IFN.