Interferon-Based Combination Anti-Viral Therapy for Hepatitis C Virus After Liver Transplantation: A Review and Quantitative Analysis

Authors

  • C. S. Wang,

    1. Department of Medicine and Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • H. H. Ko,

    1. Department of Medicine and Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • E. M. Yoshida,

    Corresponding author
    1. Department of Medicine and Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, British Columbia, Canada
    2. The British Columbia Transplant Society, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • C. A. Marra,

    1. Centre for Clinical Epidemiology and Evaluation, Vancouver Coastal Health Research Institute, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • K. Richardson

    1. Centre for Clinical Epidemiology and Evaluation, Vancouver Coastal Health Research Institute, Vancouver, British Columbia, Canada
    Search for more papers by this author

* Corresponding author: Eric M. Yoshida, eyoshida@interchange.ubc.ca

Abstract

Recurrence of hepatitis C virus (HCV) infection after liver transplantation (LT) is universal. However, the efficacy, tolerability and safety of combination interferon and ribavirin (IFN–RIB) or peginterferon and ribavirin (PEG–RIB) anti-viral therapies post-LT are uncertain. We performed a comprehensive search of major medical databases (1980–2005) and conference proceedings (1996–2005). The main outcome measure was sustained virological response (SVR, undetectable HCV RNA) at 6 months. Summary estimates were calculated using random-effects models. Twenty-seven IFN–RIB and 21 PEG–RIB studies were included. IFN–RIB was associated with a pooled SVR rate of 24% (95% CI, 20–27%), while PEG–RIB was associated with an SVR rate of 27% (23–31%). Pooled discontinuation rates were 24% (21–27%) with IFN–RIB and 26% (20–32%) with PEG–RIB. The pooled rate of acute graft rejection was 2% (1–3%) with IFN–RIB and 5% (3–7%) with PEG–RIB. IFN–RIB and PEG–RIB therapies in HCV infection post-LT were associated with similar but overall low SVR and were poorly tolerated. The rate of acute rejection was small. The therapeutic advantage of PEG–RIB therapy observed in non-transplant chronic HCV infection appears to be attenuated post-LT. Clinical trials are needed to evaluate reasons for this post-transplant therapeutic disadvantage and to find strategies to ameliorate them.

Introduction

Hepatitis C virus (HCV) infection is an important public health problem, with more than 4 million people affected in the United States and 170 million affected worldwide (1,2). Chronic hepatitis occurs in the majority of HCV-infected individuals, and up to 20% of those with chronic disease may progress to cirrhosis after 20 years (3). Liver transplantation (LT) is the only available treatment for patients with decompensated cirrhosis, and is also indicated for early stages of hepatocellular carcinoma (4). Chronic HCV infection is now the most common indication for LT worldwide (5).

Unfortunately, re-infection of the graft with HCV post-LT is universal, and is a major concern in transplant hepatology (6,7). Current evidence suggests that the course of HCV recurrence post-LT progresses more rapidly and more aggressively than in the immunocompetent non-transplant HCV-infected population. Chronic liver disease develops in at least 70% of patients after 3 years (8), with 3–6 times faster progression of fibrosis (9), and a shortened course to cirrhosis. The median time to cirrhosis is estimated to be as early as 6–9 years after LT (10–12). Furthermore, the rate of decompensation and death after the development of cirrhosis is accelerated (13). Studies of large numbers of LT patients have demonstrated that untreated HCV recurrence leads to decreased graft and patient survival when compared to non-transplant chronic HCV-infected individuals, and also when compared to non-HCV-infected LT recipients transplanted for other indications (11,14). Approximately 10–25% of patients with recurrent disease will die or require re-transplantation within 5 years of initial transplantation (15).

The aggressive course of HCV infection post-LT has led to multiple studies investigating the use of combination anti-viral therapy [interferon plus ribavirin (IFN–RIB) or peginterferon plus ribavirin (PEG–RIB)] to prevent or slow the course of recurrent HCV infection post-LT. However, unlike in the non-transplant hepatitis literature, interpretation of these studies is limited by their small sample sizes and the lack of large, randomized controlled trials. As a result, there is not a clear understanding of the overall treatment benefits. It was with this in mind that we performed a quantitative analysis of published data to provide a concise estimate of the efficacy and tolerability of combination anti-viral therapy for the treatment of HCV infection post-LT.

Materials and Methods

Search strategy

We performed a comprehensive search of the published literature for controlled and observational studies regarding the efficacy of interferon-based combination therapy (IFN–RIB and PEG–RIB) for HCV infection post-LT. An apriori decision was made to specifically focus our review on combination therapy (and not monotherapy), as it is the optimal treatment in non-transplant chronic HCV infection, and is the first therapy of choice in our institution for HCV recurrence post-LT.

We searched the following electronic databases: MEDLINE (OVID Technologies, 1980–2005 August 9), EMBASE (1980–2005 Week 32), Cochrane Controlled Clinical Trials Registry (to 2005 3rd Quarter), Cochrane Database of Systematic Reviews (to 2005 3rd Quarter), Database of Abstracts and Reviews (to 2005 3rd Quarter) and ACP Journal Club (1991 to July/August 2005). The search strategy was limited to humans and combined the following keywords: ‘Hepatitis C or Hepatitis C Virus or HCV’, ‘Transplantation or Transplant’, ‘Liver or Hepatic’, ‘Interferon or Interferon-alfa’, ‘Pegylated interferon or Peginterferon or Peg interferon or Peginterferon-alfa’, and ‘Ribavirin’.

The computerized search was supplemented with a manual search of reference lists of review articles, original studies, and of conference abstracts from the following journals: Hepatology, Journal of Hepatology, American Journal of Transplantation, Gastroenterology and American Journal of Gastroenterology between 1995 and August 2005.

Study selection

One author (C.S.W. or H.H.K.) screened the titles and abstracts of the computerized and manual searches. All retrieved full-text articles and conference abstracts were independently reviewed by two authors (C.S.W. and H.H.K.) for relevance, inclusion and data extraction using a standardized data collection form. Disagreements between reviewers were resolved by consensus.

Given the potential for heterogeneity in the published literature, strict inclusion and exclusion criteria were developed prior to the literature search. The inclusion and exclusion criteria were designed to be broad enough to capture all relevant literature, but narrow enough so as to exclude small and poorly conducted studies, as well as to limit heterogeneity. We included studies (controlled or uncontrolled, randomized or non-randomized) published in full-text or in abstract form, of IFN–RIB or PEG–RIB for both early prophylactic treatment of HCV infection as well as treatment of established recurrent HCV infection post-LT. Combination therapy in the studies had to be planned for 6 months or more. We excluded reports which: (i) were review articles, (ii) were interim reports of ongoing studies, (iii) included patients aged less than 18 years, (iv) were small studies or case series with less than 10 patients, (v) involved non-uniform treatment during the study period, (vi) included only patients with HCV co-infected with human immunodeficiency virus (HIV) or hepatitis B virus (HBV) and/or (vii) dealt only with therapy for HCV infection in the re-transplant population.

We recognize the drawbacks of including studies published only in abstract form, particularly the wide variation in design, quality and lack of stringent peer-reviewed methodology. However, we felt it was vital to search for and include relevant abstracts, as omitting abstracts could miss a significant portion of available data. We utilized the same inclusion and exclusion criteria to identify relevant abstracts, but made sure they reported on completed work (e.g. were not merely interim or preliminary analyses) to reduce the risk of bias from having only partial results. As well, every attempt was made to identify and omit multiple publications from the same group, which were obviously in duplicate (if the same studies were published as both abstracts and full-text articles, we included only the full articles).

Data extraction and outcomes

Intention-to-treat methods were used to extract response rates for each study. The primary outcome measure was virological response (undetectable HCV RNA) 6 months after cessation of treatment (sustained virological response, SVR). Secondary outcome measures were: (i) virological response at the end of treatment (ETVR), (ii) biochemical response (normalization of serum transaminases) at the end of treatment (ETBR) and end of follow-up (EFBR), (iii) histological response (stable or improved inflammatory and/or fibrosis indices) at the end of treatment (ETHR) and end of follow-up (EFHR), (iv) rate of acute rejection (AR) and (v) compliance (completion of full duration at target drug dosages, completion of full duration but at reduced drug dosages, premature discontinuation of treatment).

Statistical methods

Pooled quantitative summary estimates of the pre-defined outcome rates were calculated across individual studies by applying the random-effects model of DerSimonian and Laird (16). The ‘meta-summary’ estimate produced by this model differs from a simple arithmetic average, in that it represents a weighted average of results from individual studies, with studies of smaller sample sizes given less weight (initial weights for each study were given by their sample sizes). We assumed that heterogeneity would be present, specifically that significant variation in treatment effects would be a consequence of inter-trial differences. Therefore, we performed all calculations using random-effects models, which adjust not only for within-study variance, but for heterogeneity between studies as well. This produces more conservative estimates than does a fixed-effects model, and lowers the risk of too optimistically narrow confidence intervals. Funnel plots of the primary outcome (SVR) against sample size were constructed to assess for publication bias (17).

Since the majority of studies in post-LT HCV infection utilized a non-controlled and non-randomized design, we performed the quantitative analysis carefully considering the biases that may result from lack of randomization (18). A recommended approach to deal with the heterogeneity is to sort the heterogeneous group of studies into subgroups, according to variables suspected of contributing to the inconsistency. We used seven stratifying variables (full-text papers vs. conference abstracts, high vs. low target drug dosages, duration of planned treatment, use of growth factors, greater vs. less than 75% genotype 1 patients, inclusion of patients already treated post-LT, interval from LT to initiation of treatment). For each subgroup, we calculated a pooled random-effects summary estimate, as well as the p-value associated with testing, assuming that the summary outcome rates from each subgroup were equal. The effect of patient and trial characteristics on the probability of experiencing the estimated intervention benefit (SVR) was also analyzed using multivariate models. The multivariate random-effects logistic regression models were constructed by first entering all variables with univariate p-values <0.2 into the model. Variables were then removed from the model in a stepwise fashion, retaining only those variables with p-values <0.2. Two-sided significance tests were used throughout the analysis with a p-value <0.05 considered statistically significant. All analyses were performed using SAS software version 8.2 (SAS, Cary, NC, USA).

Results

Search results

A total of 475 citations and 83 conference abstracts were identified in the computerized and manual literature searches. Of these, 32 full-text papers and 21 abstracts met eligibility criteria and were included in the quantitative analysis (Figure 1). Funnel plot analysis of SVR for each treatment strategy did not suggest the presence of publication bias (Figure 2).

Figure 1.

Study selection process.

Figure 2.

Funnel plots of the sample size for each study plotted against the sustained virological response (SVR) percentage for IFN–RIB in HCV recurrence (A), PEG–RIB in HCV recurrence (B), and IFN–RIB in HCV prophylaxis (C). The dashed vertical lines represent the pooled summary estimates.

Anti-viral therapy for recurrent HCV infection post-LT

For the treatment of established HCV recurrence, we identified 27 studies [20 papers (19–37) and 7 abstracts (38–44), total 689 patients] of IFN–RIB, of which 2 were controlled (21,28) (1 vs. placebo, 1 vs. IFN monotherapy) and the remainder observational (Table 1). The studies of IFN–RIB were generally small (range 12–54 patients per study). The median proportion of men was 75% (range 64–93%), and with genotype 1 infection was 82% (range 35–100%). Immunosuppressive regimes varied widely.

Table 1.  Summary of studies of IFN–RIB for treatment of recurrent HCV infection post-LT (19–44)
SourceDesignTotal n; % Male; % Genotype 1Age (mean or median)Interval from LT to treatmentTreatment protocolDuration of combination treatment (months)Duration of follow-up (months)Immunosuppression
  1. RCT = randomized controlled trial; CT = controlled trial; UCT = uncontrolled trial; M = male; G-1 = genotype 1; IF = interferon; Rib = ribavirin; MU = million units; TIW = three times per week; Cy = cyclosporine; Cs = corticosteroids; Aza = azathioprine; MMF = mycophenolate mofetil; n/a = not available or applicable.

  2. 1Published in abstract form only.

Samuel et al. (28)RCT28; 64% M; 82% G-1Mean 56 yearsMean 52 monthsIF 3 MU TIW Rib 1 000–1 200 mg/day12 months6 months46% Cy + Cs; 18% Cy + Cs + Aza; 18% FK506 + Cs; 14% FK506 + Cs + Aza; 4% Cy
Ahmad et al. (21)CT20; 90% M; 35% G-1Mean 53 yearsMean 38 monthsIF 5 MU TIW Rib 1 200 mg/day12 months6 monthsFK506-based
Targhetta et al. (34)UCT25; 72% M; n/a G-1Mean 52 yearsMean 16 monthsIF 6 MU TIW Rib 1 000 mg/day12 months0n/a
Bizollon et al. (19)UCT21; 67% M; 95% G-1Mean 56 yearsMean 9 monthsIF 3 MU TIW Rib 1 200 mg/day6 months12 months95% Cy + Cs; 5% FK506 + Cs
Yedibela et al. (36)UCT15; 80% M; 87% G-1Mean 49 yearsMean 18 monthsIF 3–6 MU TIW Rib 800–1 200 mg/day12 months6 months80% FK506 + Cs; 20% Cy + Cs + Aza
De Vera et al. (23)UCT13; 72% M; 84% G-1Mean 52 yearsn/aIF 3 MU TIW Rib 1 000 mg/day>12 monthsMedian 14 monthsn/a
Shakil et al. (24)UCT38; 74% M; 82% G-1Median 50 yearsMedian 23 monthsIF 3 MU TIW Rib 1 000 mg/day12 months6 months79% FK506; 21% Cy; 37% Cs
Narayanan Menon et al. (26)UCT26; 69% M; 50% G-1Mean 47 yearsMedian 15 monthsIF 3 MU TIW Rib 800–1 000 mg/day12 months6 monthsn/a
Firpi et al. (27)UCT54; 80% M; 78% G-1Mean 49 yearsMedian 31 monthsIF 3 MU TIW Rib 800–1 000 mg/day12 months6 months65% FK506; 35% Cy
Lavezzo et al. (25)UCT27; 77% M; 74% G-1Median 55 yearsMedian 9 monthsIF 3 MU TIW Rib 800 mg/day6 months12 months42% Cy; 5% Cy + Aza; 10% Cy + Aza + Cs; 33% FK506; 9% FK506 + Cs
Lavezzo et al. (25)UCT30; 77% M; 77% G-1Median 55 yearsMedian 9 monthsIF 3 MU TIW Rib 800 mg/day12 months12 months42% Cy; 5% Cy Aza; 10% Cy + Aza + Cs; 33% FK506; 9% FK506 + Cs
Nair et al. (29)UCT33; 85% M; n/a G-1Mean 48 yearsMedian 5 monthsIF 3 MU TIW Rib 800 mg/day12 months0FK506 + MMF
Giostra et al. (30)UCT31; 74% M; 61% G-1Median 54 yearsMedian 18 monthsIF 3 MU TIW Rib 10 mg/kg/day12 months6 months35% Cy; 35% FK506; 13% Cy + Aza; 3% Cy + MMF; 10% FK506 + Cs; 3% Cy + Aza + Cs
Alberti et al. (20)UCT18; 78% M; 72% G-1n/an/aIF 3 MU TIW Rib 600 mg/day12 months>12 months100% Aza + Cs + Cy; then continued on Cy alone or Cy + Cs
Kornberg et al. (22)UCT15; 73% M; 93% G-1Mean 50 yearsMean 10 monthsIF 3 MU TIW Rib 600 mg/day12 monthsMean 20 months100% (Cy or FK506) + Cs + (Aza or MMF)
Dumortier et al. (35)UCT24; n/a M; n/a G-1n/aMedian 7 monthsIF 3 MU TIW Rib 1 200 mg/day12–24 monthsMedian 39 monthsn/a
Mukherjee et al. (33)UCT38; 76% M; 89% G-1Mean 49 yearsMedian 4 monthsIF 3 MU TIW Rib 1 000–1 200 mg/day6 or 12 months depending on genotype6 monthsCy; FK506; Cs
Ross et al. (32)UCT15; 93% M; 53% G-1Mean 47 yearsMedian 4 monthsIF 3 MU TIW Rib 800–1 200 mg/day>6 months6 monthsCs; FK506; Aza; MMF
Toniutto et al. (37)UCT12; 83% M; 100% G-1Median 52 yearsMedian 9 monthsIF 3 MU TIW Rib 600–800 mg/day12 months6 months(Cy or FK506) + Cs
Berenguer et al. (31)UCT24; 75% M; 100% G-1Median 54 yearsMedian 14 monthsIF 1.5–3 MU TIW Rib 600–1 200 mg/day12 months6 monthsn/a
Firpi et al. (40)1UCT30; 67% M; n/a G-1Mean 48 yearsMean 6 monthsIF 3 MU TIW Rib 1 000–1 200 mg/day>6 months6 months(Cy or FK506) ± Cs
Bizollon et al. (41)1UCT20; n/a M; 95% G-1Mean 53 yearsMean 11 monthsIF 3 MU TIW Rib 1 000 mg/day6 months6 monthsn/a
Bizollon et al. (42)1UCT23; n/a M; 85% G-1Mean 51 yearsn/aIF 3 MU TIW Rib 1 000 mg/day6 monthsMean 15 monthsn/a
Willner et al. (38)1UCT16; 75% M; 81% G-1Mean 48 yearsn/aIF 3 MU TIW Rib 800–1 200 mg/day6–12 months0n/a
Firpi et al. (43)1UCT50; 71% M; 88% G-1Mean 51 yearsn/aIF 3 MU TIW Rib 800–1 000 mg/day6–18 months6 monthsCy or FK506
Gordon et al. (39)1UCT14; n/a M; 57% G-1n/aMedian 16 monthsIF 3 MU TIW Rib 800 mg/day12 months6 monthsn/a
Belli et al. (44)1UCT29; n/a M; 71% G-1n/an/aIF 3 MU TIW Rib 200–600 mg/day12 months>12 monthsn/a

We identified 21 studies [9 papers (37,45–51) and 12 abstracts (52–63), total 587 patients] of PEG–RIB, of which only 1 was controlled (49) (vs. placebo) (Table 2). The studies of PEG–RIB were also generally small (range 11–86 patients per study). The median proportion of men was 77% (range 59–100%), and with genotype 1 infection was 81.5% (range 62–100%). Many studies did not report immunosuppressive regimes, but those that did showed wide variability.

Table 2.  Summary of studies of PEG–RIB for treatment of recurrent HCV infection post-LT (37,45–63)
SourceDesignTotal n; % Male; % Genotype 1Age (mean or median)Interval from LT to treatmentTreatment protocolDuration of combination treatment (months)Duration of follow-up (months)Immunosuppression
  1. CT = controlled trial; UCT = uncontrolled trial; M = male; G-1 = genotype 1; PIF = peginterferon; Rib = ribavirin; QW = every week; Cy = cyclosporine; Cs = corticosteroids; Aza = azathioprine; Siro = sirolimus; MMF = mycophenolate mofetil; n/a = not available or applicable.

  2. 1Published in abstract form only.

Castells et al. (49)CT24; 71% M; 100% G-1Mean 61 yearsMean 4 monthsPIF 1.5 mcg/kg QW Rib 800 mg/day6–12 months6 months71% FK506 + Cs; 29% FK506 + Cs + MMF
Rodriguez-Luna et al. (47)UCT19; 74% M; 63% G-1Mean 53 yearsMean 32 monthsPIF 1.5 mcg/kg QW Rib 800–1 000 mg/day>12 months6 monthsFK506; Cs; Cy; MMF
Mukherjee et al. (45)UCT39; 85% M; 79% G-1Median 50 yearsMedian 20 monthsPIF 1.5 mcg/kg QW Rib 800 mg/day6 or 12 months depending on genotype6 months62% Cy; 38% FK506
Planas et al. (50)UCT30; 73% M; 90% G-1Mean 56 yearsMean 57 monthsPIF 1.5 mcg/kg QW Rib 10.6 mg/kg/day6 or 12 months depending on genotype6 months70% FK506; 20% Cy; 10% MMF
Babatin et al. (51)UCT13; 77% M; 62% G-1Mean 49 yearsMean 24 monthsPIF 1.5 mcg/kg QW Rib 10.6 mg/kg/day12 months6 months54% Cy; 15% FK506; 23% Siro; 8% FK506 + Siro
Dumortier et al. (48)UCT20; 65% M; 80% G-1Mean 54 yearsMedian 28 monthsPIF 1 mcg/kg QW Rib 1 000–1 200 mg/day12 months6 months85% FK506 ± (MMF or Cs); 15% Cy
Neff et al. (46)UCT29; 59% M; 97% G-1Mean 53 yearsMean 20 monthsPIF 1.5 mcg/kg QW Rib 400–600 mg/day12 months6 monthsFK506 + Cs
Neff et al. (46)UCT28; 79% M; 100% G-1Mean 51 yearsMean 27 monthsPIF 1.5 mcg/kg QW Rib 400–600 mg/day12 months6 monthsFK506 + Cs
Toniutto et al. (37)UCT12; 100% M; 100% G-1Median 57 yearsMedian 14 monthsPIF 0.5 mcg/kg QW Rib 600–800 mg/day12 months6 months(Cy or FK506) + Cs
Neff et al. (56)1UCT30; 63% M; n/a G-1Mean 54 yearsMean 29 monthsPIF 1.5 mcg/kg QW Rib 400–600 mg/day>6 months0FK506
Kontorinis et al. (57)1UCT25; n/a M; n/a G-1n/an/aPIF 180 mcg QW Rib 14 mg/kg/day6 months0n/a
Bahra et al. (52)1UCT25; n/a M; n/a G-1n/an/aPIF 1–1.5 mcg/kg QW Rib 200–800 mg/day12 months6 monthsn/a
Lavezzo et al. (53)1UCT16; 81% M; 62% G-1Mean 54 yearsMedian 10 monthsPIF 1 mcg/kg QW Rib 800 mg/day12 months6 monthsn/a
Samuel (54)1UCT22; 59% M; 77% G-1Mean 58 yearsMean 96 monthsPIF 1 mcg/kg QW Rib 7.5 mg/kg/day6 or 12 months depending on genotype6 monthsn/a
Lorenzini et al. (55)1UCT18; n/a M; 78% G-1n/an/aPIF 80 mcg QW Rib 600 mg/day6–12 months6 monthsn/a
Gordon et al. (58)1UCT13; n/a M; 69% G-1n/aMedian 44 monthsPIF 0.5–1.5 mcg/kg QW Rib 800 mg/day12 months6 months92% FK506; 8% Cy
Martinez et al. (59)1UCT11; n/a M; n/a G-1n/an/aPIF n/a Rib n/aMean 12 monthsn/an/a
Samanta et al. (60)1UCT26; 85% M; n/a G-1Mean 52 yearsn/aPIF 1.5 mcg/kg QW Rib 1 000 mg/day>6 months6 monthsn/a
Kontorinis et al. (61)1UCT40; n/a M; 90% G-1n/aMean 35 monthsPIF 180 mcg QW Rib 14 mg/kg/day12 months0n/a
Martini et al. (62)1UCT86; 79% M; 83% G-1Mean 53 yearsMedian 12 monthsPIF 1 mcg/kg QW Rib 800 mg/day12 months6 monthsn/a
Picciotto et al. (63)1UCT61; 79% M; 87% G-1Mean 47 yearsn/aPIF 1 mcg/kg QW Rib 10 mg/kg/day6 or 12 months depending on genotype6 months38% Cy; 48% FK506; 14% FK506+MMF

Interferon plus ribavirin.  Pooled results across all IFN–RIB studies are shown in Table 3. While end-of-treatment virological response (ETVR, defined as loss of detectable HCV RNA at the end of therapy) was 34% (95% CI, 30–37%), the SVR (defined as loss of detectable HCV RNA 6 months after therapy) dropped to 24% (95% CI, 20–27%). As expected, biochemical responses were higher. Despite the low virological responses, both the end-of-treatment and end-of-follow-up histological responses (defined as stable or improved inflammatory and/or fibrosis indices) were over 50%. Compliance to therapy was poor—only 33% (95% CI, 21–27%) of patients were able to achieve and maintain target drug dosages and duration of therapy. The remainder required either dose reduction (44%; 95% CI, 38–50%) or early discontinuation (24%; 95% CI, 21–27%). The rate of acute rejection observed was very low (2%; 95% CI, 1–3%).

Table 3.  Pooled outcomes of IFN–RIB and PEG–RIB for treatment of recurrent HCV infection post-LT
OutcomeIFN–RIBPEG–RIB
No. of studiesPooled proportion (95% CI)No. of studiesPooled proportion (95% CI)
  1. ETBR = end of treatment biochemical response; ETVR = end of treatment virological response; ETHR = end of treatment histological response; EFBR = end of follow-up biochemical response; SVR = sustained virological response; EFHR = end of follow-up histological response.

ETBR1957% (53–61%)1054% (48–60%)
ETVR2634% (30–37%)2042% (38–46%)
ETHR1453% (38–68%)668% (57–79%)
EFBR1537% (26–48%)232% (20–43%)
SVR2224% (20–27%)1627% (23–31%)
EFHR659% (48–70%)285% (68–100%)
Compliance—full duration, full dose1833% (28–38%)1421% (9–34%)
Compliance—full duration, reduced dose1844% (38–50%)1466% (61–70%)
Compliance—early cessation2624% (21–27%)2026% (20–32%)
Acute rejection252% (1–3%)165% (3–7%)

Subgroup analyses of SVR based on stratifying variables are shown in Table 4. Of the subgroups studied, only prior anti-viral therapy post-LT showed a significant difference in SVR in the univariate analysis, and was also the only variable that was statistically significant in the multivariate logistic regression model (OR 1.9; 95% CI, 1.0–3.4; p = 0.04). Of note, neither drug dosage, proportion with genotype 1, growth factor use, interval from LT, nor duration of therapy in the IFN–RIB trials appeared to have a significant association with SVR.

Table 4.  Univariate analysis of predictor variables for SVR for treatment of recurrent HCV infection post-LT
VariableIFN–RIBPEG–RIB
No. of studiesPooled proportion (95% CI)p-ValueNo. of studiesPooled proportion (95% CI)p-Value
  1. IFN = interferon; PEGIFN = peginterferon; Rib = ribavirin; GF = growth factors (erythropoietin and/or colony stimulating factors); G-1 = genotype 1; LT = liver transplantation; n/a = not available or applicable.

Published papers vs.1723% (19–27%)0.30929% (23–34%)0.39
Abstracts527% (20–34%) 725% (20–30%) 
IFN > 3 MU sc/week vs.226% (11–40%)0.79n/an/an/a
IFN = 3 MU sc/week2024% (20–27%) 
PEGIFN = 1.5 mcg/kg/day vs.n/an/an/a825% (19–31%)0.45
PEGIFN < 1.5 mcg/kg/day 828% (23–33%) 
Rib > 800 mg/day vs.1224% (20–29%)0.70323% (14–32%)0.41
Rib ≤ 800 mg/day1023% (18–28%) 1327% (23–32%) 
GF used vs.920% (15–26%)0.141028% (23–34%)0.40
GF not used1326% (21–30%) 625% (20–31%) 
≥75% G-1 patients vs.1322% (18–26%)0.271126% (22–30%)0.05
<75% G-1 patients726% (20–33%) 340% (27–53%) 
Previously treated vs.235% (24–45%)0.02722% (17–27%)0.02
Not previously treated2022% (19–26%) 932% (26–38%) 
6 months therapy vs.626% (20–33%)0.33728% (22–34%)0.47
≥12 months therapy1623% (19–27%) 925% (20–31%) 
Therapy started <6 months from LT vs.327% (18–35%) 133% (14–52%) 
6–24 months from LT vs.1222% (18–27%)0.37523% (16–29%)0.30
>24 months from LT325% (17–34%)0.84633% (25–40%)0.96

There has been only one randomized placebo controlled study (28). Treatment lasted for 48 weeks, with 6-month follow-up. As expected, no patients in the control group achieved SVR (vs. 21% in the treatment group). The other controlled study was non-randomized, and compared combination IFN–RIB to IFN monotherapy (21). SVR was significantly different between the two groups (2.5% with IFN vs. 20% with IFN–RIB, p = 0.03).

Peginterferon plus ribavirin.  Pooled results across all PEG–RIB studies are shown in Table 3. While ETVR was 42% (95% CI, 38–46%), the SVR dropped to 27% (95% CI, 23–31%), which was very similar to that for IFN–RIB. Biochemical responses were also similar between PEG–RIB and IFN–RIB. As with IFN–RIB, both the ETHR and EFHR were high. Compliance to therapy was poor and worse than IFN–RIB—only 21% (95% CI, 9–34%) of patients were able to achieve and maintain target drug dosages and duration of therapy. The majority required dose reduction (66%; 95% CI, 61–70%). The proportion of patients who discontinued therapy early (26%; 95% CI, 20–32%) was similar to that for IFN–RIB. Similar to IFN–RIB, PEG–RIB anti-viral therapy was associated with a low rate of acute rejection (5%; 95% CI, 3–7%).

Subgroup analyses of SVR based on stratifying variables are shown in Table 4. Of the subgroups studied, the proportion of genotype 1 patients in the studies and prior anti-viral therapy post-LT showed a significant difference in SVR in the univariate analysis. In the multivariate logistic regression model, only the proportion of genotype 1 patients in the studies had a p-value less than 0.2 (OR 0.5; 95% CI, 0.2–1.4; p = 0.16). Of note, neither drug dosage, growth factor use, interval from LT, nor duration of therapy in the PEG–RIB trials appeared to have a significant association with SVR.

In the only controlled non-randomized trial, 6 months of PEG–RIB therapy was compared to placebo (49). As expected, no patients in the control group achieved SVR (vs. 33% in the treatment group).

Pre-emptive HCV anti-viral therapy post-LT

For the prophylaxis of HCV recurrence post-LT, we identified five studies [three papers (64–66) and two abstracts (67,68), total 134 patients], of which one was controlled (67) (vs. placebo) (Table 5). The studies of IFN–RIB prophylaxis were small (range 21–36 patients per study). The median proportion of men was 77% (range 59–81%), and with genotype 1 infection was 83% (range 83–91%). We were unable to identify any studies of PEG–RIB for prophylactic therapy that met all eligibility criteria.

Table 5.  Summary of studies of IFN–RIB for prophylactic treatment of HCV infection post-LT (64–68)
SourceDesignTotal n; % Male; % Genotype 1Age (mean or median)Interval from LT to treatmentTreatment protocolDuration of combination treatment (months)Duration of follow-up (months)Immunosuppression
  1. UCT = uncontrolled trial; RCT = randomized controlled trial; M = male; G-1 = genotype 1; IF = interferon; Rib = ribavirin; MU = million units; TIW = three times per week; Cy = cyclosporine; Cs = corticosteroids; Aza = azathioprine; MMF = mycophenolate mofetil; n/a = not available or applicable.

  2. 1Published in abstract form only.

Sugawara et al. (65)UCT21; 78% M; 83% G-1Range 23–63 yearsMedian 30 daysIF 6 MU TIW Rib 600 mg/day>12 monthsMedian 26 monthsFK506 + Cs
Mazzaferro et al. (64)UCT21; 76% M; 91% G-1Median 50 yearsMedian 21 daysIF 3 MU TIW Rib 10 mg/kg/day12 monthsMedian 12 monthsCy + Aza + Cs
Mazzaferro et al. (66)UCT36; 81% M; 83% G-1Median 52 yearsMedian 18 daysIF 3 MU TIW Rib 10 mg/kg/day12 monthsMedian 52 monthsCy + Cs
Reddy et al. (67)1RCT21; 59% M; n/a G-1Mean 50 yearsRange 14–28 dIF 3 MU TIW Rib 1 000 mg/day12 months6 monthsn/a
Nair et al. (68)1UCT35; n/a M; n/a G-1n/an/aIF 3 MU TIW Rib 800 mg/day12 months0n/a

Interferon plus ribavirin.  Pooled results across all IFN–RIB prophylactic studies are shown in Table 6. ETVR (30%; 95% CI, 22–38%) and SVR (31%; 95% CI, 21–42%) were similar. Biochemical and histological responses were based on only a few studies. It is likely that the lower histological responses seen in the prophylactic setting are due to better baseline histology and thus less margin for improvement with treatment. As with treatment of HCV recurrence, compliance to therapy was poor—only 50% (95% CI, 39–61%) of patients were able to achieve and maintain target drug dosages and duration of therapy, while 21% (95% CI, 1–40%) withdrew early from treatment. The observed rate of acute rejection was low but slightly higher (6%; 95% CI, 2–11%) than that seen in the treatment of established HCV recurrence. However, this may not be a direct consequence of anti-viral therapy, as treatment was started early after transplantation when the rate of rejection is higher.

Table 6.  Pooled outcomes of IFN–RIB for prophylactic treatment of HCV infection post-LT
OutcomeIFN–RIB
No. of studiesPooled proportion (95% CI)
  1. ETBR = end of treatment biochemical response; ETVR = end of treatment virological response; ETHR = end of treatment histological response; EFBR = end of follow-up biochemical response; SVR = sustained virological response; EFHR = end of follow-up histological response; n/a = not available or applicable.

ETBR223% (13–32%)
ETVR430% (22–38%)
ETHR352% (42–62%)
EFBR133% (18–49%)
SVR331% (21–41%)
EFHR133% (18–49%)
Compliance—full duration, full dose550% (39–61%)
Compliance—full duration, reduced dose429% (21–37%)
Compliance—early cessation421% (1–40%)
Acute rejection36% (2–11%)

Subgroup analyses of SVR based on stratifying variables are shown in Table 7. Of the subgroups studied, only a few could be calculated due to the small total number of studies. SVR was statistically different if the result was published in full paper versus abstract form, and also by the dose of ribavirin. No variables were statistically significant in the multivariate logistic regression model, likely owing to the small number of studies.

Table 7.  Univariate analysis of predictor variables for SVR for prophylactic treatment of HCV infection post-LT
VariableIFN–RIB
No. of studiesPooled proportion (95% CI)p-Value
  1. IFN = interferon; Rib = ribavirin; GF = growth factors (erythropoietin and/or colony stimulating factors); G-1 = genotype 1; LT = liver transplantation; n/a = not available or applicable.

Published papers vs.237% (24–49%)0.02
Abstracts114% (0–29%) 
IFN > 3 MU sc/week vs.143% (22–64%)0.18
IFN = 3 MU sc/week226% (15–37%) 
Rib > 800 mg/day vs.114% (0–29%)0.02
Rib ≤ 800 mg/day237% (24–49%) 
GF used vs.143% (22–64%)0.18
GF not used226% (15–37%) 
≥75% G-1 patients vs.237% (24–49%)n/a
<75% G-1 patients0n/a 
Previously treated vs.0n/an/a
Not previously treated331% (21–41%) 
6 months therapy vs.0n/an/a
≥12 months therapy331% (21–41%) 
Therapy started <6 months from LT vs.331% (21–41%)n/a
6–12 months from LT vs.0n/a 
>12 months from LT0n/a 

In the one randomized controlled trial of prophylaxis, 12 months of IFN–RIB therapy was compared to placebo (67). No patients in the control group achieved SVR (vs. 14% in the treatment group).

Discussion

Treatment of chronic HCV infection in the immunocompetent, non-transplant population has been well studied with large, randomized controlled trials. Meta-analyses of these trials have demonstrated that combination IFN–RIB therapy produces an overall SVR rate of approximately 41% (69), while combination PEG–RIB therapy produces an overall SVR rate of approximately 55% (70). Treatment of HCV infection post-transplantation has been inferred from the non-transplant setting, but without similar large, controlled clinical trials proving efficacy. Instead, the published literature consists of numerous small cohort studies. Therefore, it has been difficult to provide patients with an accurate estimate of the efficacy and tolerability of treatment. The present review of data from 53 trials of combination anti-viral therapy post-LT is unique, in that it is the first to our knowledge to quantitatively consolidate and pool the individual results of multiple studies using meta-analytical techniques in this patient population. Because most studies did not include a control group, we were unable to calculate pooled odds ratios or mean differences in comparison to placebo or other therapies.

The primary outcome chosen was virological response because the vast majority of patients achieving SVR post-LT will continue to have undetectable HCV RNA with long-term histological improvement up to 3 years after cessation of therapy (71). Our results suggest that approximately 24% of patients treated with IFN–RIB, and 27% of patients treated with PEG–RIB, for recurrence of HCV infection post-LT will obtain an SVR. Approximately 31% of patients will achieve SVR when treated prophylactically early post-LT. Thus, when compared to the non-transplant setting, both therapies have reduced efficacy when used post-LT. In particular, the therapeutic advantage of PEG–RIB over IFN–RIB appears to be lost. Potential factors responsible for the reduced virological response are prior non-response to therapy, inclusion of sicker patients requiring treatment, predominance of genotype 1 infection in the studies, effect of immunosuppression on drug efficacy and tolerance, intolerance to side effects of medications, as well as other unknown factors. Compliance to therapy was also low. In the non-transplant setting, combination therapy is associated with higher frequencies of side effects and lower compliance than monotherapy (69), and is likely similar post-LT as well. Immunosuppression or side effects from immunosuppressive medications may further contribute to decreased tolerability. It is important to emphasize that with the high rates of discontinuation, and the fact that many patients with HCV infection post-LT are not suitable for treatment (72), it appears that the full benefits of combination anti-viral therapy may apply to only the minority (and not the majority) of the overall hepatitis C liver transplant population. Interferon use has also been associated with increased rates of acute and chronic allograft rejection when used in the renal transplant setting (73,74). Importantly, our analysis demonstrates low rates of acute cellular rejection. However, given the observational nature of many of these studies, it is impossible to determine a cause-and-effect relationship between anti-viral therapy and rejection.

In the univariate and multivariate analyses, the inclusion of studies with patients who had received prior anti-viral therapy after transplantation influenced the pooled proportion of patients achieving SVR for both IFN–RIB and PEG–RIB, but the results were conflicting (Table 4). This finding requires verification; but if true, possible explanations for the inconsistent results can be postulated. Most studies did not report the specific therapy that patients had previously received nor the actual number of patients who had previously been treated. It is likely that patients who were previously treated received IFN–RIB (as it is usually the ‘first-line’ therapy in most centers). The decision to re-treat patients with IFN–RIB (vs. PEG–RIB) likely meant that they had responded to IFN–RIB and have relapsed. The fact that they had a prior response may predict a future-favorable response to IFN–RIB therapy again. In the PEG–RIB group, re-treated patients likely failed prior IFN–RIB therapy (non-responders), which may predispose to failing subsequent anti-viral therapy with PEG–RIB. The data to confirm these hypotheses are not available and need further investigation.

The results of our quantitative analysis should be interpreted in the context of study limitations. One limitation is the inclusion of non-controlled and non-randomized trials. Although our eligibility criteria were designed to limit bias and heterogeneity, there was considerable variability in the patients and treatment regimes between the studies, and thus our results may be subject to confounding from individual studies and the potentially high degree of heterogeneity between them. These include differences in the definition of recurrent HCV infection, inclusion criteria for patients, dosages of drugs, timing of therapy, and immunosuppressive regimes. For this reason, we calculated outcomes using random-effects models. We also performed subgroup analyses for the primary outcome, looking at some of the between-study differences to determine their influence on the pooled SVR. Ultimately, we believe that this can be viewed both as a limitation and strength of our review. Limiting our analysis to only studies of a specific group of patients or intervention regimens may have provided a more focused answer; however, we chose to assess the effect of combination therapy under a variety of circumstances, to increase the degree of generalizability of the results and better reflect true practice. Immunosuppressive regimes also varied greatly among patients within and between studies. It would have been useful to assess whether the efficacy and tolerability of therapy were influenced by the type and amount of immunosuppression, but that information was unavailable. Further, the outcomes of our current review are based on surrogate endpoints (biochemical, virological and histological responses). The question of whether combination therapy post-LT affects liver-related mortality remains unanswered. Unfortunately, these limitations are a reality of the current post-LT literature.

The ideal method to control for confounding is a large, controlled randomized trial. Given the aggressive course of recurrent HCV infection post-LT, it may be unethical to include a placebo-control group in such a trial, since no patients in the placebo groups of the three controlled studies (28,49,67) we reviewed achieved SVR. It is also unlikely to include a monotherapy-control group, given the proven benefit of combination therapy over monotherapy in non-transplant HCV infection. Thus by default, IFN–RIB or PEG–RIB may become the standard treatment post-LT to which new therapies will be compared despite a lack of knowledge of its true efficacy. Our review begins to address this problem; however, we believe there needs to be direct, comparative randomized trials between IFN–RIB and PEG–RIB.

In conclusion, the present review demonstrates that combination therapy has a beneficial effect on virological, biochemical and histological responses in patients with HCV infection post-LT. IFN–RIB and PEG–RIB therapies have similar outcomes, tolerability and safety. We believe this review further advances the field of transplant hepatology by providing a better understanding of the benefits and risks of HCV anti-viral therapy post-LT. Future research focusing on direct comparative trials, strategies to improve therapeutic efficacy and tolerability, and the effect on liver-related morbidity and mortality are needed in this important patient population.

Acknowledgments

This project was supported by an unrestricted research grant from Hoffman-LaRoche Canada Limited. Dr. Carlo A. Marra is supported by a Michael Smith Foundation for Health Research Scholar Award. Dr. Eric M. Yoshida has received unrestricted research grants from Hoffman-LaRoche Canada, GlaxoSmithKline Canada, Fujisawa Canada, Pfizer Canada, Jansen-Ortho Canada. He has received honoraria for CME lectures sponsored by Hoffman-LaRoche Canada, Schering Canada. He has attended Advisory Board meetings and received an honorarium from Hoffman-LaRoche Canada, Novartis Canada. He has also been an investigator in clinical trials sponsored by Hoffman-LaRoche Canada, Schering Canada, Fujisawa Canada, Novartis Canada, Wyeth-Canada, Migenix Inc, Idenix Inc, Human Genome Sciences Ltd. Dr. Carlo A. Marra has received unrestricted research grants from Schering Plough Inc, Pfizer Canada, Bristol-Meyers Squibb Canada, and Merck Canada. He has received consulting fees from Schering Plough Inc. and Schering Canada.

Ancillary