Hepatitis C virus (HCV)–induced cirrhosis is a leading indication for liver transplantation (LT).1 Graft reinfection is universal,2, 3 leading to cirrhosis in 20% or more by 5 years after LT.2, 4, 5 Several factors are associated with the severity of recurrent hepatitis C: older donors,6–8 changes in immunosuppression,9–11 and diabetes after transplantation.12
Although immunosuppression does influence HCV recurrence,13–17 a regimen that results in less severe recurrence is not known.18 Cyclosporine has activity against HCV in vitro.19 However, in patients, a meta-analysis demonstrated no difference between cyclosporine and tacrolimus.20 Nevertheless, tacrolimus-based immunosuppression reduces graft loss in comparison with cyclosporine.21 Another meta-analysis suggested that steroids may be detrimental,18 in contrast to a recent study.22 Complete steroid withdrawal (except for autoimmune diseases) is now common practice, reducing complications and not worsening patient or graft survival.23, 24 A randomized study25 of calcineurin inhibitor monotherapy (MT) showed safety and effectiveness with respect to acute/chronic rejection rates, immune graft loss, and patient/graft survival, and this has been confirmed by others.26, 27
Thus, we randomized recipients who underwent transplantation for HCV, assessing tacrolimus MT versus tacrolimus, azathioprine, and prednisolone triple therapy (TT) and hypothesizing that MT was less immunopotent and, without maintenance steroids, would be less deleterious for recurrent HCV.
Inclusion/Exclusion Criteria, Randomization, and Endpoints
From January 2000 to June 2007, in 3 LT centers [Royal Free Hospital (RFH), Edinburgh Royal Infirmary, and St. Vincent's Hospital], consecutive transplant recipients older than 18 years receiving cadaveric liver grafts were randomized if they had HCV-related cirrhosis with serum HCV RNA positivity. Patients with a possible or confirmed/concomitant alcoholic etiology or hepatocellular carcinoma were not excluded. Patients gave informed written consent at listing for transplantation. Exclusion criteria were retransplantation, multiorgan or split/auxiliary transplants, contraindications to tacrolimus or azathioprine, and a refusal to participate. The study protocol was approved by the hospital ethics committee at each site.
For randomization, sealed opaque envelopes were used; they were opened in a numbered sequence containing the allocated treatment in a 1:1 proportion derived from a random number table with a blocked code for each center. Follow-up stopped at death, retransplantation, or January 31, 2008.
The primary endpoint was whichever of the following occurred first: (1) progression to Ishak stage 4 (S4), (2) graft failure resulting in retransplantation or patient death, or (3) immunological failure (more than 2 histologically confirmed episodes of cellular rejection failing to respond to therapy) as MT could theoretically lead to deleterious rejection and decreased survival, obviating any benefit with respect to fibrosis.
Secondary endpoints included patient survival, acute cellular rejection, chronic rejection, steroid-resistant rejection regardless of changes in immunosuppression, recurrent HCV (defined by an Ishak inflammation score ≥ 4), and withdrawal from randomized allocation.
The sample size was calculated to be 110 patients: S4 was estimated to occur in 35% by 3 years with a projected decrease to 10%, and this required a sample size of 99. Considering a dropout rate of 10% (including early deaths), we concluded that 110 patients would be needed with a 2-tailed test with significance at 5% and a power of 80% (1-B).28
Tacrolimus (Prograf, Fujisawa, Ltd., Killorglin, Ireland) at 0.1 mg/kg/day was given nasogastrically (within 6 hours from LT) in 2 divided doses in both the MT and TT groups and was started within 6 hours after transplantation. Azathioprine was given intravenously and then orally (1 mg/kg/day), and methylprednisolone (16 mg/day intravenously) was given until oral intake was established; then, 20 mg/day prednisolone was used. Tacrolimus dosing (evaluated every other day) was adjusted with the goal of maintaining a whole blood level of 5 to 10 ng/mL by microparticle enzyme immunoassay (ImxTacrolimus II, Abbott Laboratories, Abbott Park, IL), particularly with poor renal and/or graft function. The azathioprine dose was not changed unless neutropenia developed. Prednisolone was gradually tapered from 3 weeks and then stopped between 3 and 6 months.
Liver Biopsy and Histological Assessment
Protocol liver biopsy was performed between days 5 and 7 after transplantation and thereafter if clinically indicated and at yearly intervals. All were evaluated for acute cellular rejection, hepatitis and chronic rejection, recurrent HCV, and cytomegalovirus (CMV)29 by liver histopathologists in each center, who were blinded to immunosuppression details.
Liver biopsy samples were formalin-fixed, paraffin-embedded, and stained with hematoxylin and eosin. Gordon and Sweet staining for reticulin and chromotrope aniline blue staining was used to assess fibrosis (in Edinburgh Sirius red and Van Gieson stains). The Ishak stage and grade30 were used for histological evaluation.
We calculated a biopsy index per patient (months of follow-up divided by the biopsy number from 1 year onward) to assess bias in the histological ascertainment and fibrosis progression rate.31
Acute cellular rejection was graded with the RFH score32 or Banff criteria in Edinburgh and Dublin. An RFH score ≥ 4 established the diagnosis of acute cellular rejection; when the score was >7, rejection was treated.
Corticosteroid-resistant cellular rejection was defined if there was no histological improvement in a biopsy sample 5 days after the first despite 1 g of intravenous methylprednisolone daily for 3 days. Immunological failure was defined as an absence of histological improvement after a second identical cycle of methylprednisolone, after which mycophenolate mofetil (MMF) was used (it was also given if renal impairment developed).
Acute hepatitis C was diagnosed when alanine aminotransferase levels increased to ≥2 times the upper normal limit and there were histological changes consistent with hepatitis (predominantly lobular inflammation and/or scattered parenchymal apoptosis) without diagnostic features of cellular rejection, bile duct loss, or other causes of liver injury.
The hepatic venous pressure gradient (HVPG) was measured only at RFH with standard techniques33 associated with transjugular biopsy in 64 of 77 patients who survived the first year (33 MT patients and 31 TT patients). Clinically significant portal hypertension was defined as HVPG ≥ 10 mm Hg.34–37
Virological Assays and Antiviral Therapy
Serum samples before transplant, at 1 and 3 months post-transplant, and at the time of liver biopsy were collected, stored at −70°C, and analyzed for HCV RNA by polymerase chain reaction; the HCV genotype was evaluated as previously described.8 At RFH, CMV viremia was screened for by polymerase chain reaction assay twice weekly.38 Two consecutive positive CMV DNA samples (and positive samples later on if there were increasing titers)29 were an indication to treat a patient with ganciclovir/valgancyclovir for 14 days.39 When the study was designed, standard interferon for HCV was not planned because of reports of rejection. Currently, pegylated interferon and ribavirin are not licensed in the United Kingdom for use after transplantation, but patients were considered for off-label therapy if S4 was reached.
Results are expressed as medians and ranges. Categorical variables were compared with the χ2 test or Fisher's exact test. Continuous variables were compared with the Student t test or, if not normally distributed (Kolmogorov-Smirnov test), with the Mann-Whitney test.
Using intention-to-treat analysis, we evaluated, in a time-dependent fashion, survival, the interval to reach histologically proven S4 (excluding and including those who had defined immunological failure; for patients who did not reach S4, we used the last biopsy date), and the time to reach HVPG ≥ 10 mm Hg.
Comparisons were made by life tables (Kaplan-Meier) with the log-rank test for statistical differences. Variables associated with HCV recurrence in the literature were evaluated by Cox regression analysis: the age and gender of the recipient and donor, hepatocellular carcinoma and/or concomitant alcoholic liver disease, histological acute hepatitis, rejection episodes and their treatment, liver and renal function, cold and warm ischemic times, use of antithymocyte globulin, CMV viremia (needing therapy), CMV hepatitis, diabetes mellitus pre-transplant and post-transplant, and viral load pre-transplant and after LT.
Bio-Medical Data Processing (SPSS 11) was used. A P value < 0.05 was considered statistically significant.
Patients' Demographic and Transplant Data
One hundred three patients (89 at RFH, 8 at the Royal Infirmary, and 6 at St. Vincent's Hospital) were randomized: 54 received MT, and 49 received TT. Recipient/donor characteristics pre-transplant are shown in Table 1, and postoperative variables are shown in Table 2. Two were patients coinfected with hepatitis B/D (they remained hepatitis B virus DNA–negative throughout the study).
Table 1. Characteristics of Patients in Randomized Groups: Monotherapy and Triple Therapy Just Before LT
NOTE: Monotherapy consisted of tacrolimus alone, whereas triple therapy consisted of tacrolimus, azathioprine, and prednisolone. The cold ischemia time was the interval from the donor cross-clamp to removal from cold storage. The warm ischemia time was the interval from removal from cold storage to venous reperfusion.
The groups did not have significant differences in preoperative, perioperative, or postoperative variables. The mean follow-up was 53.5 months (range: 1–86): 54 months for MT and 52.7 months for TT.
Only 2 patients were withdrawn (both MT patients) because of side effects of the treatment (days 12 and 27 post-LT).
During the first 3 months, 22 MT patients had 42 rejection episodes, and 16 (30%) received 21 methylprednisolone courses; 31 TT patients had 64 rejection episodes, and 24 (49%) received 30 methylprednisolone courses. These proportions are similar to those in our preliminary report on rejection episodes40; the relatively high rates are due to protocol biopsy, diagnosis, and repeat biopsy to check histological resolution.40
Tacrolimus Concentrations, Changes in Immunosuppression, Side Effects, and Antiviral Therapy
The tacrolimus concentrations at the first protocol biopsy were higher in the MT group; the median was 7.7 ng/mL (range: 3–26 ng/mL) for MT versus 5 ng/mL (range: 2.5–14 ng/mL) for TT (P = 0.02), and on day 21, the median was 10 ng/mL (3.3–18 ng/mL) for MT versus 6 ng/mL (3–20 ng/mL) for TT (P = 0.01). The median levels at 1 month (8 ng/mL for MT versus 8.4 ng/mL for TT), 2 months (7 ng/mL for MT versus 7.5 ng/mL for TT), and 3 months (8 ng/mL for MT versus 7 ng/mL for TT) were not significantly different.
During follow-up, tacrolimus was discontinued in 6 MT patients (5 had severe renal impairment with multiorgan failure shortly before death, and 1 had neurotoxicity; cyclosporine/sirolimus was started); among 4 TT patients, 3 had neurotoxicity (2 were converted to cyclosporine, and 1 was converted to MMF substituting for azathioprine and continued on steroids), and 1 had renal impairment (converted to sirolimus).
Serum creatinine rose to ≥130 μmol/L (normal range: 66–112 μmol/L) within 3 months after transplant in 66% of MT patients versus 60% of TT patients. Serum creatinine levels at 6 weeks and 3 months were not significantly different, and neither was new-onset diabetes (29% for MT and 27% for TT; Table 2), bacterial or fungal infections, or courses of antibiotics (data not shown).
Antiviral treatment for HCV recurrence (after S4 was reached) was used in 6 MT and 8 TT patients, with 3 in each group achieving sustained virological response. The last biopsy in the sustained virological response patients showed the fibrosis stage: 5, 4, and 3 in TT and 4, 4, and 4 in MT.
Primary Endpoints and Factors Associated with Early Survival and Overall Survival
Overall mortality was not significantly different: there were 9 MT deaths (7 within 3 months, including 2 after re-LT), including 5 from sepsis/multiple organ failure (0.3, 0.3, 0.6, 1, and 4 months), 1 from graft failure (0.5 months), 1 from pulmonary hypertension (0.2 months), 1 from ductopenic rejection (18 months), and 1 from recurrent HCV disease (30 months), and 6 TT deaths [3 within 3 months: 1 after retransplantation (sepsis at 2 days) and 2 at 1 month from sepsis/multiple organ failure], including 1 due to non–liver-related causes at 6 months and 2 from recurrent HCV disease (30 and 42 months). There were 6 immunological failures: 2 MT patients and 4 TT patients.
The difference in reaching the primary endpoint was significant (P = 0.045). Multivariately, only randomization to MT (odds ratio = 0.6, 95% confidence interval = 0.064–0.85) was associated with the primary endpoint. The hazard function is shown in Fig. 1.
Factors Associated with Reaching S4
The median biopsy index was 14.7 (10.4–20) for MT and 16.25 (10.4–20.3) for TT (P = 0.21); this eliminated a histological ascertainment bias. The median number of biopsies in each group was 3.2 for MT and 3 for TT. No histological features of alcohol-induced liver damage were seen.
In the MT group, 15 patients reached S4 as a primary endpoint, and 8 (53%) progressed to stage 5/6 after a median of 32 months post-LT. In the TT group, 8 patients reached S4, and 3 (37.5%) progressed to stage 5/6 after a median of 38 months after LT. The median time to reach S4 was 26 months for MT and 39 months for TT. Another 2 patients in each group reached S4 from the 6 with immunological failure, and the 2 TT patients among those 6 died. Thus, a total of 17 MT patients and 10 TT patients reached S4. At 1 year, the median Ishak stage was 2.2 (0–4) in the MT group and 1.79 (0–4) in the TT group. Deaths occurring before the diagnosis of S4 were censored. The hazard function for reaching S4 either as the primary endpoint or in combination with the 4 patients who had previous immunological failure was significantly different: P = 0.044 and P = 0.04, respectively (Fig. 2). Only 3 variables listed in the methods were associated univariately with a fibrosis stage ≥ 4: randomization to MT (P = 0.007), episodes of acute hepatitis (P = 0.009), and patients with 1 or more histologically proven rejection episodes versus none. Only 2 were significant multivariately: randomization to MT (odds ratio = 0.7, 95% confidence interval = 0.066–0.847) and diagnosis of acute hepatitis (odds ratio = 3.59, 95% confidence interval = 1.108–9.823). The latter was more common in MT patients (17 of 54) than TT patients (8 of 49; P = 0.027) and occurred at a median of 4.4 months (1–6 months) in the MT group and at a median of 4.8 months (1.5–7 months) in the TT group.
The rate of S4 at 3 years post-LT was 23% for the TT group and 44% for the MT group (Fig. 2).
Azathioprine was discontinued between 12 and 36 months after LT in 8 patients, and S4 was not significantly different: 1 of 8 (12.5%) versus 7 of 38 who continued azathioprine (18%).
Factors Associated with Portal Hypertension
HVPG measurement was performed in 33 MT patients and 31 TT patients, who represented 78% of MT patients and 72% of TT patients, surviving at 1 year at RFH.
HVPG was not measured in 9 MT patients (2 retransplanted) and 6 TT patients (1 retransplanted) not surviving 12 months or in another 1 MT patient and 2 TT patients because of retransplantation.
Among the 33 MT patients with HVPG, 13 (39%) reached S4, and among the 5 patients without HVPG, 2 (20%) reached S4.
Among the 31 TT patients with HVPG, 7 (23%) reached S4, and among the 10 patients without HVPG, 1 (10%) reached S4.
Ten MT patients versus 2 TT patients reached HVPG ≥ 10 mm Hg. Kaplan-Meier curves (Fig. 3) were significantly different (P = 0.045).
Cox regression analysis identified only a stage ≥ 4 as an independently associated variable from those listed in the methods with HVPG ≥ 10 mm Hg.
Recurrent HCV infection after LT remains a major challenge as a significant proportion of patients develop progressive liver disease.41–43
We compared in a randomized trial a steroid-free and less immunopotent tacrolimus regimen (MT) versus a standard tacrolimus, prednisolone, and azathioprine regimen (TT) used from the outset after LT. The MT group had higher tacrolimus concentrations immediately after transplant, with values similar to those of the TT group at 1 month. Rejection episodes assessed by protocol biopsies that were histologically proven44 and/or required methylprednisolone (30% versus 49%) were less frequent in the MT group. A salient feature is that we measured HVPG, which is associated with progressive fibrosis.44, 45
There were no survival differences between the 2 treatment arms, with 11 deaths (10.7%) within 1 year. Retransplantation rates (7.8% for TT and 9.6% for MT) and chronic rejection rates (2% for TT and 3.8% for MT) were not different and were consistent with the eras in which the transplants occurred, suggesting a standard course post-LT.
Thus, despite more rejection episodes, more boluses of steroids, and lower trough tacrolimus concentrations within 1 month, the group that had long-term azathioprine and maintenance low-dose steroids for 3 to 6 months had slower progression to S4.
The fibrosis rates in the 2 trial groups were similar to or lower than the fibrosis rates described by others.4, 31, 46 Importantly, the fibrosis progression rate in the TT group was lower than others published.4, 31
In our study, the only independent factors associated with the time to reach S4 among those reported in the literature to be associated with fibrosis were episodes of acute hepatitis (biochemically and histologically proven), which were significantly less frequent by 50% in TT patients, and allocation to MT. The acute hepatitis episodes were diagnosed within the first year after transplantation; these have been shown to be an important and independent risk factor for fibrosis progression.6, 8 The diagnosis was histological, and thus the nature of the infiltrate was interpreted to be hepatocyte-attacking, but there is the possibility that the infiltrate includes tolerogenic cells and thus a less vigorous response to HCV, which may then be related to the persistence of the virus and continuing damage to the graft. Alternatively, if all the infiltrate is involved in attacking infected cells, it is not sufficient, and continuing damage ensues because of the persistence of infected hepatocytes.
Randomization in this study eliminated potential biases of donor and recipient age and gender differences as well as subtle changes in surgical and medical care with time that could have influenced the recurrence of HCV disease.
Our results are consistent with observational studies published during the conduct of our trial,14, 22, 47 showing benefit from prolonged steroid therapy for graft survival and disease recurrence. Although a recent review18 suggests that steroid avoidance results in less severe recurrence of HCV, there is important heterogeneity in the selected studies, and the steroid-containing arms18 had considerably more immunosuppressive potency than our TT arm; this complicates the interpretation of the use of steroids. A recent randomized study of cyclosporine and basiliximab48 with or without steroids found no difference in the progression of fibrosis, but both groups had considerable immunosuppression. Our trial is of particular interest because the nonsteroid group received the least potent immunosuppression in comparison with the reported nonsteroid groups,18 with a 30% treated rejection rate based on protocol biopsies.
The study design was made in 1998/1999 when MMF was not used in our unit, so azathioprine was our standard adjunctive immunosuppression. MMF has been associated with a worsening effect or no effect on fibrosis.49
The use of azathioprine has been associated with less fibrosis;49; this finding has been documented in our whole population8 and 6 others,6, 14, 50–53 with only 3 studies54–56 showing no difference in HCV recurrence and none showing worsening. Experimentally, an antiviral effect of azathioprine on HCV exists,57 but this has not been shown in patients, although lower viral loads with azathioprine versus mycophenolate have been documented.58 In the first year, virtually no patient stopped azathioprine, but steroids were tailed off in most. This might suggest that the effect of azathioprine is more important than low-dose maintenance steroids. Indeed, in a recent randomized trial48 in which azathioprine was not used, the steroid group had more fibrosis.
The evaluation of histological fibrosis was not biased as the histopathologists did not have knowledge of the immunosuppression regimens and the biopsy index showed that similar numbers of biopsy procedures were performed in both groups; this removed an ascertainment bias.
The histological findings favoring the TT group are supported by the HVPG measurements, which provide internal validation despite the relatively small size of the total cohort randomized. More MT patients developed HVPG ≥ 10 mm Hg (ie, clinically significant portal hypertension),34, 59 with a significantly shorter time to reach HVPG ≥ 10 mm Hg in comparison with the TT group (P = 0.045). Our study confirms the utility of HVPG measurement in assessing the progression of chronic hepatitis C5, 44, 45, 60 after LT.
In summary, the use of long-term low-dose azathioprine and shorter term (3–6 months) low-dose prednisolone, in addition to tacrolimus, in HCV cirrhosis recipients resulted in a slower onset of histologically proven severe fibrosis and portal hypertension in comparison with tacrolimus alone, and this was independent of other factors known to affect fibrosis. This randomized trial supports and extends the few observational reports of the benefit of low-dose and slowly tapered steroids22 as well as azathioprine6, 8, 14, 49–53 after LT for HCV-positive recipients. New regimens should be compared to ones containing azathioprine and steroids.