Immunosuppression with calcineurin inhibitors with respect to the outcome of HCV recurrence after liver transplantation: Results of a meta-analysis



A controversy exists over whether the outcome of a hepatitis C virus (HCV)-infection-related liver transplant differs based on the calcineurin inhibitor (CNI) used. We have performed a systematic review and a subsequent meta-analysis evaluating tacrolimus (Tac)-based vs. cyclosporine A-based immunosuppression in HCV-infected liver transplant recipients. Searches were conducted to locate randomized controlled trials comparing Tac vs. cyclosporine A. Data on HCV liver transplant recipients were obtained, independently of whether the study was specifically designed for patients with this disease or not. A fixed effects model was used for statistical pooling of the relative risks (RR) for the different outcomes. A total of 5 articles (366 patients) fulfilled the inclusion criteria. Statistically significant differences between Tac-based vs. cyclosporine A-based therapies were not found for mortality (P = 0.11; RR = 0.72; 95% confidence interval [CI], 0.49-1.08), graft survival (P = 0.37; RR = 0.86; 95% CI, 0.61-1.21), biopsy-proven acute rejection (P = 0.65; RR = 0.91; 95% CI, 0.61-1.36), corticoresistant acute rejection (P = 0.26; RR = 2.25; 95% CI, 0.55-9.29), and fibrosing cholestatic hepatitis (P = 0.92; RR = 0.96; 95% CI, 0.41-2.26). In 1 study, no differences were detected regarding severe fibrosis at 1 yr. In conclusion, patient and graft survivals in HCV-positive liver transplant patients are similar independently of the CNI selected as basic immunosuppressant. Unfortunately, data on the severity of recurrence and effect on viremia are scarce. Well-designed randomized prospective studies are needed to determine whether there are differences between the 2 CNIs regarding these specific variables. Liver Transpl 13:21–29, 2007. © 2006 AASLD.

Cirrhosis caused by chronic hepatitis C virus (HCV) infection now accounts for approximately half of all liver transplant procedures in most transplant centers.1 Unfortunately, reinfection of the graft is universal, resulting in chronic hepatitis in most patients, graft cirrhosis in 20 to 30% of these, and graft loss in 10% of patients within the first 5 to 7 yr of transplantation.1–3 In fact, recurrence of the original disease highly contributes to the inferior patient and graft survival rates among the HCV-positive liver transplant population.4–6 A few simple variables, including “old donor age” and “potent immunosuppression” are the most significant determinants of outcome.1, 2, 7 The exact definition of “potent immunosuppression,” however, is difficult to translate in terms of type, doses, and duration of immunosuppressive agents.7, 8 While there is evidence that convincingly demonstrates an association between oversuppression in the setting of rejection therapy and HCV-related progressive disease, data regarding the specific effect of the different immunosuppressive agents on HCV replication and/or progression of recurrent hepatitis C are either preliminary or controversial.1, 2, 7, 8 In fact, published studies have serious drawbacks regarding the impact of immunosuppression on HCV recurrence. In the vast majority of studies, HCV disease is not the main goal of the study. In addition, the study design is generally inadequate (retrospective in nature and hence unable to control for confounding factors), the information on HCV ribonucleic acid levels is scarce, there is a multiplicity of immunosuppressive combinations, which limits the ability to assess the effect of each individual drug, the follow-up is too short (less than 6 to 12 months in general) to evaluate an impact on HCV disease, and routine protocol liver biopsies are typically not performed.

Cyclosporine has recently been found to have an antiviral effect with respect to HCV.9, 10 The availability of an HCV subgenomic replicon system has permitted relatively easy screening of drugs for their antiviral effect. Using this methodology, Watashi et al.9 recently showed an inhibitory effect of cyclosporin on HCV protein expression and replicon HCV ribonucleic acid levels, an effect that was not detected with tacrolimus (Tac). However, whether this in vitro data translate into in vivo differences is however still unclear.

The aim of this meta-analysis was therefore to determine whether, in patients undergoing transplantation for HCV-related cirrhosis, there are differences in clinical, virological, and histologic post-transplantation outcome with regard to the calcineurin inhibitor (CNI) used.


HCV, hepatitis C virus; CNI, calcineurin inhibitor; Tac, tacrolimus; RR, relative risk; CI, confidence interval.


Search Strategy

We conducted a search of publications in electronic databases: Biosis, CINAHL, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, DARE, DERWENT, EMBASE (Drugs and Pharmacology), MEDLINE, and SCISEARCH until the end of 2005. In addition, a manual search of most relevant journals, including Hepatology, Journal of Hepatology, Liver Transplantation, Transplantation Proceedings, Transplantation, American Journal of Transplantation, and Transplant International was also performed to review published studies from 1993 to 2005. Key words in our electronic searches included “Neoral” or “sandimmun” or “CyA” or “cyclosporine” or “csa” or “calcineurin inhibitors” as well as “prograf” or “tacrolimus” or “FK506,” combined with “liver transplantation” and HCV-related key words.

Selection Criteria

Inclusion criteria in our study were: 1) randomized controlled trials in which 1 treatment arm consisted in Tac (either alone or combined with other immunosuppressants and independently of dosage or regimen) and another treatment arm received cyclosporine solution or cyclosporine microemulsion (either alone or combined with other Immunosuppressants and independently of dosage or regimen); 2) trials in which a group of HCV patients could be identified (defined as HCV ribonucleic acid posttransplantation-positive); 3) the trial reported outcome measures (see below) for this particular subgroup of patients, whether the trial considered the inclusion of these HCV patients only or considered these patients as a subgroup; 4) time of follow-up should be at least 12 months. Trials were not excluded on the basis of language. When the results from some or all patients in a clinical trial were reported more than once, data on endpoints from the publication with the longest follow-up were extracted.

Data Collection and Analysis

Two of the investigators (A.R. and M.B.) reviewed all the studies independently. Data were extracted on study characteristics, patient clinical characteristics and demographics, duration of follow-up, and the incidence of the following clinical outcomes: patient and graft survival, HCV viral load (quantitative HCV ribonucleic acid assessment), fibrosis (biopsy-proven, indicating whether it was clinically indicated or per protocol), and acute rejection (biopsy-proven). We also extracted information about reported adverse events as secondary outcomes that were tabulated in the 2 treatment groups at the end of controlled follow-up. A meta-analysis was also attempted for those parameters with homogeneity among the different trials.

Methodological Quality of Trials

Two reviewers (A.R. and J.Z.) assessed trials independently. Quality items examined were sample size computation, method used to generate the random allocation sequence, concealment of allocation, blinding, intention to treat analysis, and completeness of follow-up.

Statistical Techniques

We expressed results for dichotomous outcomes as relative risk (RR) with 95% confidence intervals (CIs). Heterogeneity was analyzed among trials with the Cochran Q test and calculating the I-square index,11, 12 to measure the proportion of total variation due to heterogeneity beyond chance. If no heterogeneity was detected, the overall treatment effect size was calculated as a weighted average of the results of each primary study. A sensitivity analysis was conducted using both random and fixed effects models and practically the same results were found. Results reported herein used the fixed-effects model.12 All statistical analyses were performed using Review Manager Software.13


Literature Search

An initial electronic search identified 314 reports (Fig. 1). After a first review, 12 reports were excluded for duplication as well as 255 for either not being a randomized controlled trial or not including data on immunosuppression. The number of potentially relevant clinical trials identified for more detailed evaluation was 47. A total of 43 articles were excluded with a second review because they were not randomized trials (n = 31) or they did not provide information on immunosuppression or HCV infection (n = 12). Therefore, 5 articles were considered appropriate to be included in the systematic review (Table 1), 4 from searches14–17 and 1 available directly from the authors,18 with a total of 366 randomized participants (183 patients included in each treatment arm, Tac-based or cyclosporine-based immunosuppression). All trials were available in full-text form and all of them were in English.

Figure 1.

Flowchart describing the procedure for selection of studies. RCT, randomized controlled trials; HCV, hepatitis C virus

Table 1. Characteristics of Trials Included in Review
StudyUnicenter/ multicenterDouble blind?Randomization criteriaInclusion criteriaProtocol biopsyInterferon treatmentFollow-up (months)No. of PatientsTreatment
  • Abbreviations: ATG, thymoglobulin induction (antithymocyte globulin); Aza, azathioprine; CS, corticosteroids; CsA, cyclosporine; CsA-me, cyclosporine microemulsion; ESLD, end-stage liver disease; Tac, tacrolimus; IFN, interferon; MU tiw, million units thrice weekly.

  • *

    Immunosuppressive treatment was specified in a previous report.60

  • Authors did not specify whether the cyclosporine is oil-based or microemulsion formulation.

  • The randomization criteria was found in the first report of this study28.

Berenguer et al. (2006)18UnicenterNo (open)Not availableHCV+: ESLD for HCVYes: yearly ±4 monthsNo during study period1290TAC + Cs vs. CsA-me + CS
Martin et al. (2004)14MulticenterNo (open)Telephone randomization system, previous to transplantHCV+: ESLD for HCVYes: months 3 and 12Not available1279Tac + CS + Aza vs. sA-me + CS + Aza
Mueller et al. (1995)17UnicenterNo (open)Not availablePrimary liver transplantation (HCV+: ESLD for HCV)NoNot available1235Tac + Cs vs. CsA + ATG + CS + Aza*,
Wiesner et al. (1998)16MulticenterNo (open)Randomly assigned to the treatment according to a computer-generated code that investigators did not knowLiver transplantation for ESLD (HCV+: ESLD for HCV)Yes: days 7, 28, and 360)Not available60113TAC +CS vs. CsA +CS +Aza
Zervos et al. (1998)15UnicenterNo (open)Not availableHCV+: ESLD for HCVYes: months 12 and 18)αIFN 1.5-3 MU tiw postoperatively1849TAC +CS vs. CsA-me +CS

Trials Included

A description of baseline features of trials included is reported in Table 1. Regarding the immunosuppressant combinations evaluated in the trials, an equivalent concomitant immunosuppressive therapy for Tac and cyclosporine arms was used in 3 studies.14, 15, 18 Double therapy was employed in 2 studies,15, 18 while 1 study14 used triple therapy with corticosteroids plus azathioprine. On the other hand, 1 trial17 compared Tac in double therapy vs. cyclosporine in quadruple therapy, although the authors did not specify what the fourth immunosuppressant was. Finally, 1 study16 compared Tac in double therapy with cyclosporine in triple therapy. One of the studies included17 did not specify the formulation of cyclosporine being tested (oil-based or microemulsion). None of the selected studies had used C2 monitoring.

Outcomes reported in the 5 different trials varied significantly (Table 2). Data on histologic HCV disease progression based on protocol liver biopsies was only available in 1 study.18 The effect of viral replication was also described only in 1 study.14 Information on patient and graft survival was available in all studies, whereas data on biopsy-proven rejection was provided in 4 studies.14–16, 18 However, while the 4 studies indicated that rejection episodes were always biopsy-proven, the exact definitions of corticosteroid-resistant rejection or that of refractory rejection were only specified in 1 study16 In addition, while treatment of the first episode of rejection consisted in methylprednisolone in all studies, the number of boluses varied among studies from 1 gm/day/1 day14 to 1 gm/day/2 days.16 Interferon is a drug that is increasingly being used to treat recurrent hepatitis C and that, unfortunately, may trigger rejection episodes. Data on its use though was limited to 2 studies,17, 18 and in 1 of the 2 studies, the authors specified that interferon was never used during the study period.18

Table 2. Outcomes Reported in Trials Included in the Review
Berenguer et al. (2006)18Martin et al. (2004)14Mueller et al. (1995)17Wiesner et al. (1998)16*Zervos et al. (1998)15
  • Abbreviations: Tac, tacrolimus; CsA, cyclosporine; AR, acute rejection; RNA, ribonucleic acid; n/N, number of cases/total number of patients included; CMV, cytomegalovirus.

  • *

    Wiesner study presents long term results (5 yr).

1-1.5 yrn/Nn/Nn/Nn/Nn/Nn/Nn/Nn/Nn/Nn/N
Patient mortality6/466/446/388/411/170/1812/5722/567/258/24
Graft lost10/467/4410/3811/41  15/5723/568/259/24
AR (biopsy-proven)4/462/4411/3816/4110/175/18  6/2512/24
Corticosteroid resistant AR    5/171/18  0/251/24
Severe fibrosis at 1 yr posttransplantation9/468/44        
Cholestatic hepatitis5/464/44      4/255/24
Median change in RNA-HCV levels from baseline to month 12 million meq/m  4.518.8      
Conversions        0/2511/24
Retransplant        4/251/24
CMV infection        5/256/24
Epstein-Barr virus viremia        0/251/24
Bacterial or fungal infections        17/2517/24
Treatments for headaches        2/250/24
Neurological complications/neurotoxicity    10/176/18  0/254/24

Quality of Included Trials

Reporting of methods was incomplete for most trials. Two trials14, 16 gave information about the randomization method and reported adequate allocation concealment and the remaining trials were randomized but gave no indication of the method used. No trials were blinded. Intention to treat analysis was confirmed for 2 trials.14, 16 Only 1 trial16 reported rationale for the sample size calculation. All trials showed adequate completeness of follow-up.


Hepatitis C recurrence

Fibrosing cholestatic hepatitis was reported in 2 trials15, 18 (Fig. 2) and no statistical significant differences were observed between both groups of study (P = 0.92; RR = 0.96; 95% CI, 0.41-2.26). A meta-analysis to assess other endpoints associated with hepatitis C recurrence was not feasible since the data on each of the 2 major HCV-related endpoints (viral load and histologic disease severity based on protocol biopsies) were only available in 1 study for each endpoint.14, 18 Results of the individual studies are provided (Table 2). In 1 study,14 median change in viremia from baseline was significantly higher at 12 months posttransplantation for patients with cyclosporine than with Tac (18.8 million mEq/mL vs. 4.5 million mEq/mL; P = 0.032). Severe fibrosis at 1 yr was equally present in both groups in another study.18

Figure 2.

Metanalysis of randomized clinical trials comparing tacrolimus and cyclosporine in liver transplant patients: fibrosing cholestatic hepatitis.

Graft and patient survival

Patient survival was reported in 5 trials.14–18 There were no differences in outcome with regard to mortality between the cyclosporine group vs. the Tac group (P = 0.11; RR = 0.72; 95% CI, 0.49-1.08) (Fig. 3). In addition, graft survival (reported in 4 trials14, 15, 18) was also similar between groups (P = 0.37; RR = 0.86; 95% CI, 0.61-1.21) (Fig. 4).

Figure 3.

Metanalysis of randomized clinical trials comparing tacrolimus and cyclosporine in liver transplant patients: mortality.

Figure 4.

Metanalysis of randomized clinical trials comparing tacrolimus and cyclosporine in liver transplant patients: graft loss.


The incidence of biopsy-proven acute rejection (reported in 4 trials14, 15, 17, 18) was similar in both groups (P = 0.65; RR = 0.91; 95% CI, 0.61-1.36) (Fig. 5). In addition, corticoresistant rejection also occurred at the same rate in both arms (reported in 2 trials15, 17) (P = 0.26; RR = 2.25; 95% CI, 0.55-9.29) (Fig. 6).

Figure 5.

Metanalysis of randomized clinical trials comparing tacrolimus and cyclosporine in liver transplant patients: acute rejection.

Figure 6.

Metanalysis of randomized clinical trials comparing tacrolimus and cyclosporine in liver transplant patients. corticorresistant rejection.

Other endpoints

The rate of conversion of CNIs (reported in 1 trial15) was significantly higher with cyclosporine than with Tac (P = 0.03; RR = 0.04; 95% CI, 0.00–0.67). Neurological complications or neurotoxicity events were reported in 2 studies.15, 17 Although no differences between groups were found (P = 0.92; RR = 1.03; 95% CI, 0.52-2.06), the results in both trials were highly heterogeneous (P = 0.04; I-square = 76.6%) (Fig. 7). Only 1 study15 reported more information regarding other adverse effects and no differences between treatment groups were found.

Figure 7.

Metanalysis of randomized clinical trials comparing tacrolimus and cyclosporine in liver transplant patients: neurotoxicity.


The outcome of recurrent hepatitis C is highly heterogeneous.1–3 Several factors related to the host, virus, donor, and environment have been implicated in disease progression.7 Of these, both donor age and immunosuppression are considered the major factors determining outcome differences.1, 2, 4, 7, 8, 19–22 Organs from donors older than 50 are associated with progressive disease and reduced survival.4, 8, 19, 23 In addition, indirect data point to the importance of immunosuppression in shaping outcome. Indeed, HCV-related disease progression is faster in immunosuppressed liver transplant recipients or human immunodeficiency virus coinfected individuals than in immune competent patients both before and after the development of cirrhosis,1–3, 24, 25 and worse outcomes reported in recent years parallel changes in immune suppression with the introduction of newer and more potent immunosuppressive agents.1–4, 26, 27 While the negative impact of additional immunosuppression for the treatment of rejection on hepatitis C is established,1–3, 7, 8 the effect of agents used for induction immunosuppression or maintenance immunosuppression is highly inconclusive.1, 7, 8, 20–22, 27, 60 In fact, results of the potential association between the type of administered immunosuppression and disease severity are difficult to prove due to the multiplicity of immunosuppressive regimes together with the changes in immunosuppressive drugs in individual patients over time. In addition, endpoints in prospective randomized studies assessing the effect of immunosuppression have basically focused on rejection rate and/or short-term survival. Finally, histologic data on recurrent disease based on protocol liver biopsies are generally missing, a fact that represents a serious limitation given the lack of correlation between liver enzymes and the extent of histologic liver damage.1–3

CNIs have been the cornerstone of immunosuppression in liver transplantation. However, as with prior agents, most prospective studies comparing Tac to cyclosporine have not adequately addressed the problem of HCV recurrence but rather have evaluated rejection rate and/or short-term survival.28–30 In retrospective studies analyzing the potential factors associated with outcome, no consistent differences between cyclosporine-based or Tac-based immunosuppressive regimens and recurrent disease have been noted in most studies (Table 3). However, some studies have suggested that the worse outcome observed in recent years may be related with the use of more potent first line immunosuppressive drugs such as Tac.27

Table 3. Outcomes Reported in Retrospective Studies in Which Induction Immunosuppression Was Included as a Potential Predictive Factor
ReferencesEndpointResultsProtocol biopsy
  1. Abbreviation: CsA, cyclosporine.

Mueller et al. (1994)32Recurrence of HCVTac worse than CsANo
Sheiner (1995)34Recurrence of HCVEarlier recurrence with TacNo
Gane et al. (1996)35Recurrence of HCVTac = CsAYes
Johnson et al. (1996)36Recurrence of HCV and survivalTac lower survival and greater incidence of cirrhosisNo
Boker et al. (1997)37Recurrence of HCV and survivalTac = CsANo
Charlton et al. (1998)38SurvivalTac = CsANo
Platz et al. (1998)39SurvivalTac = CsANo
Feray et al. (1999)40Survival and recurrence of HCVTac = CsANo
Ghobrial et al. (1999)41SurvivalTac = CsANo
Papatheodoris et al. (1999)42Recurrence of HCV and viral loadTac = CsANo
Testa et al. (2000)43Recurrence of HCV and survivalEarlier recurrence with TacYes
Berenguer et al. (2000)44Recurrence of HCVTac = CsAYes
Pelletier et al. (2000)45Recurrence of HCVTac = CsANo
Sreekumar et al. (2000)46Recurrence of HCVTac = CsAYes
Ghobrial et al. (2001)47SurvivalTac = CsANo
Hunt et al. (2001)48Recurrence of HCVTac = CsANo
Papatheodoris et al. (2001)49Recurrence of HCV and survivalTac = CsA (triple and quadruple associated with worse outcome)No
Bilbao et al. (2001)50Recurrence of HCV and survivalTac = CsANo
Wali et al. (2002)19Recurrence of HCVTac = CsAYes
Berenguer et al. (2002)51Recurrence of HCV and survivalTac associated with more fibrosisYes
Burak et al. (2002)26Recurrence of HCV and survivalTac = CsAYes
Brillanti et al. (2002)21Recurrence of HCV and survivalTac = CsANo
Paik et al. (2002)52Recurrence of HCV and survivalTac = CsANo
Sánchez-Fueyo et al. (2002)53Recurrence of HCV and survivalTac = CsAYes
Berenguer et al. (2003)27Recurrence of HCV and survivalTac = CsAYes
Berenguer et al. (2003)54Recurrence of HCVTac = CsAYes
Wali et al. (2003)55Recurrence of HCV and survivalTac = CsAYes
Neumann et al. (2004)56Recurrence of HCV and survivalTac = CsANo
Neumann et al. (2004)56Recurrence of HCVTac = CsAYes
Rimola et al. (2004)57Recurrence of HCVTac = CsANo
Bahr et al. (2005)58Recurrence of HCV and survivalTac = CsANo
Samonakis et al. (2005)20Recurrence of HCV and survivalTac = CsANo
Lucey et al. (2005)59SurvivalTac = CsANo

Recent in vitro studies point toward an antiviral effect of cyclosporine with regard to HCV, an effect not present with Tac.9 In fact, the antiviral activity appears to be mediated through a different pathway than the immunosuppressive activity.10 Whether these in vitro data translate into differences in both viral load and HCV disease progression in liver transplant recipients is a subject of intense debate. Hence, we decided to perform a meta-analysis using prospective randomized studies to answer the following questions: 1) is viral load different in cyclosporine-treated vs. Tac-treated liver transplant recipients?; 2) is disease outcome more severe in Tac-treated patients compared to those under cyclosporine? The following conclusions can be reached from our search. First, no differences exist in clinical outcomes such as graft rejection or graft or patient survival between the 2 CNIs. Interestingly, while rejection was a major endpoint in published studies, the exact definitions of corticosteroid-resistant rejection and that of refractory rejection were only provided in 1 of the studies. In addition, therapy for the first episode of cellular rejection varied between studies. Also, data on the use of interferon therapy, a drug that may trigger rejection episodes, was provided in only 2 studies. Second, a meta-analysis could not be performed to assess the 2 major questions on viral load and severity of HCV recurrence due to the lack of studies. Based on 1 single study,18 histologic disease severity at 1 yr was similar regardless of the immunosuppressive used. In addition, 1 study reported a greater increase in viremia at 1 yr among cyclosporine-treated patients compared to Tac.14

The mechanisms by which HCV leads to liver injury are incompletely understood.31 Several lines of indirect evidence support a role for the cellular immune response in shaping outcome following transplantation. Increased levels of HCV replication, together with the altered host immune responsiveness, contribute to the pathogenesis of liver damage, particularly to the severe course of disease of the grafted liver.1, 31 In this context, it is likely that more than the choice of a specific CNI, disease severity/progression is the result of the way we modify the immunosuppression over time. Thus, it may be dependent on the patient's immunosuppressive status at the time of transplantation, type and number of concurrent immunosuppressive agents used, and the doses/blood levels of the different immunosuppressive agents used, both initially and as maintenance therapy.33

In summary, patient and graft survivals in liver transplant patients due to HCV infection are similar, independently of the CNI selected as the basic immunosuppressant. Unfortunately, important information is missing on the effect of Tac and cyclosporine on both viral replication and disease progression. Well-designed, randomized, prospective studies with careful clinical, virological, and histological assessments are needed to determine whether there are differences in clinical practice with regard to these variables between both CNIs.


We thank Celia López-Palafox and Beatriz Domínguez-Gil (Medical Department, Astellas Pharma SA) for their bibliographic and technical support.