SEARCH

SEARCH BY CITATION

Keywords:

  • Azathioprine;
  • hepatitis C recurrence;
  • immunosuppression;
  • liver transplantation;
  • mycophenolate mofetil;
  • rejection;
  • renal dysfunction

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Calcineurin inhibitors (CNIs) combined with steroids with or without azathioprine (AZA), have been a standard immunosuppression regimen after liver transplantation (LT). Since 2000 many centers have substituted AZA by mycophenolate mofetil (MMF). However, in LT the superiority of MMF over AZA is not clearly demonstrated. Therefore, we questioned the benefit of MMF versus AZA in LT with regard to rejection, renal dysfunction and hepatitis C virus (HCV) recurrence and survival. Using a literature search, relevant randomized controlled trials (RCT) and cohort studies were identified: two RCTs compared MMF to AZA only for acute rejection. Treated rejection was less with MMF in only one RCT (38.5% vs. 47.7%; p = 0.025), with no difference in patient and graft survival. No RCTs compared MMF and AZA in patients with CNI-related chronic renal dysfunction. Among two studies evaluating MMF, with substitution of AZA, one was stopped due to severe rejection. Recurrent HCV was less severe in 5/9 studies with AZA compared with 2/17 using MMF, six of which documented worse recurrence. Published data in LT show little, if any, clinical benefit of MMF versus AZA. RCTs should reevaluate AZA in LT. Evaluation of HCV replication and recurrence will be particularly important as AZA may have advantages over MMF.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Until recently, dual or triple therapy involving calcineurin inhibitors (CNIs) combined with steroids with or without azathioprine (AZA) has constituted standard maintenance immunosuppression after liver transplantation (1). Since 2000, azathioprine has been substituted by mycophenolate mofetil (MMF) in many centers.

Azathioprine is a prodrug of 6-mercaptopurine, which inhibiting inosine monophosphate dehydrogenase (IMPDH) reduces purine synthesis, affecting DNA, RNA, other nucleotides and protein.

Mycophenolic acid (MPA), the active metabolite of MMF, is a selective, noncompetitive, reversible inhibitor of IMPDH. T and B lymphocytes are dependent for their proliferation on de novo synthesis of purines; MPA is more specific and potent than AZA (2). However, no data have been published comparing, in vitro, the immunopotency of MMF versus AZA.

Azathioprine is more myelotoxic and hepatotoxic than MMF (3) but MMF causes diarrhea in 30% of patients and tissue-invasive cytomegalovirus (CMV) infection, especially with 3 g/day: these side effects are reduced with 2 g/day (3).

In liver transplantation (LT) the clinical benefit of MMF over AZA is questionable. Therefore, we assessed in detail the impact of MMF and AZA on acute cellular rejection, survival, renal function and hepatitis C virus (HCV) recurrence in LT.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Search strategy

We searched for English language publications search using Medline, Cochrane database and Embase until December 2008. In addition, we manually searched abstracts from national and international liver meetings (2005–2008). Key words were mycophenolate mofetil or MMF and azathioprine or AZA or Imuran combined with LT and rejection or renal dysfunction or HCV recurrence.

Randomized controlled trials (RCT) and prospective or retrospective studies were evaluated if comparing MMF versus AZA, or if MMF or AZA were compared (whether or not together with other drugs) to other immunosuppressants. Endpoints were acute cellular rejection, patient and graft survival, renal function and HCV recurrence.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

Prevention of rejection

Only two RCT have compared MMF with AZA (4,5), with one update (6) (Table 1). The first (4) evaluated MMF/AZA in a cyclosporine-based quadruple regimen using antilymphocyte preparations for induction: 28 patients were randomized to MMF (1 g b.d.) and 29 to AZA (1–2 mg/kg daily); median follow-up was short, only 10 ± 3.2 months. The patient and graft survival was not different: MMF 89.3% and 89.7% versus AZA 85.7% and 82.7%, respectively. Liver biopsy was only performed if a rejection episode was clinically suspected. Histologically proven acute rejection was lower with MMF than with AZA: 21.4% (n = 6, 1 steroid resistant) versus 44.8% (n = 13, 2 steroid resistant) (p = 0.06). Thrombocytopenia was more frequent with AZA: 48.3% versus 21.4% (p < 0.05), as was leukopenia 20% versus 7.1% (p = ns). Gastritis and/or diarrhea occurred in six patients (four MMF and two AZA).

Table 1.  Randomized controlled trials comparing azathioprine (AZA) and mycophenolate mofetil (MMF) after liver transplantation with respect to survival, rejection and adverse effects
Author (ref) yearImmunosuppression Regimen (dose)No. of patientsMedian follow-up (Months)SurvivalAcute rejection1 (%)Adverse effects
Patient (%)Graft (%)Infections (%)GI symptoms (%)Thrombocytopenia (%)Leucopenia (%)
  1. 1Biopsy-proven acute rejection with no protocol liver biopsies.

  2. 2Biopsy-proven and treated rejection or graft loss.

  3. 3 Reference 6 is the updated trial of Reference 4.

  4. *p = 0.02; **p < 0.05.

Sterneck et al. (4) 2000CsA + steroids + MMF (1 g b.d.) 2810 ± 3.289.389.721.432 14.221.4** 7.1
CsA + steroids + AZA (1–2 mg/kg/day) 29 85.782.744.834  6.848.3  20 
Fischer et al. (6)3 2000CsA + steroids + MMF (1 g b.d.) 3110 ± 3.290.387.519.432 12.919.4** 6.5
CsA + steroids + AZA (1–2 mg/kg/day) 32 83.981.340.631  6.246.9  18.8
Wiesner et al. (5) 2001CsA + steroids + MMF (1.5 g b.d.)278At least 1288 85.3  38.52*45.551.36.9  3.6
CsA + steroids + AZA (1–2 mg/kg/day)287 87.185.447.743.249.812.5   0.7

The update (6) had 31 MMF and 32 AZA patients. Median follow-up remained short: 10 ± 3.2 months; patient and graft survival were unchanged. Paradoxically, the rejection episodes remained numerically the same: MMF 19.4% (n = 6) versus AZA 40.6% (n = 13) (p = ns). The lack of statistical difference in rejection rates could be due to the small number of patients, but in any case the absence of any increase in rejection episodes in the updated report (6) makes interpretation of differences in rejection between AZA and MMF difficult.

The registration study for MMF in liver transplantation was an international multicenter randomized double-blind comparison of MMF versus AZA, each combined with cyclosporine and steroids (5): 278 patients assigned to MMF (1.5 g b.d.) and 287 to AZA (1–2 mg/kg/day). Liver biopsies were obtained only if rejection was suspected clinically. The principal endpoint was whichever occurred first posttransplantation: (1) biopsy-proven and treated acute rejection occurring within the first 6 months and (2) graft loss (death/retransplantation) occurring within the first year. Patients losing grafts within 6 months, without rejection, were counted for the rejection endpoint. The MMF group had a significant reduction in combined endpoint compared to AZA (38.5% vs. 47.7%; p = 0.025), but its two components were not separately described. However, the frequency of more than one acute rejection episode was similar: 7.7% AZA and 5.8% MMF. Moreover, 1-year patient and graft survival rates were not significantly different: MMF 88.8% and 85.3% versus AZA 87.1% and 85.4%, respectively. An absolute neutrophil count <500 occurred in 3.6% MMF but only 0.7% AZA patients (p = not reported) contrary to the increased leucopenia with AZA in the first study (4). Platelet count <25.000 occurred in 12.5% AZA and in 6.9% MMF patients (p = not reported). Opportunistic infections were similar: 45.5% (MMF) versus 43.2% (AZA).

The rejection rate with MMF (3 g/day) (5) was twice as high as in the open-label RCT (4) in which 2 g/day was used, despite similar use of cyclosporine and steroids and indications for biopsy for suspected rejection. Currently, most units use only 2 g/day MMF, so the comparison with AZA in the registration study does not mirror today's clinical practice. Only Sterneck et al. (4) compared 2 g/day MMF with AZA, with a nonsignificant reduction in rejection rates with MMF and no differences in graft or patient survival. Thus, in LT the evidence for a significant benefit in terms of cellular rejection of MMF over AZA is very poor.

Improving CNI-related nephrotoxicity

In patients with CNI-related nephrotoxicity after LT, the complete replacement of CNIs with MMF, although associated with improvement in serum creatinine in 77–100% of patients and improvement of creatinine clearance of about 9–11 mL/min, resulted in more acute cellular rejection. Rejection was frequent when CNI withdrawal occurred within 6 months at 50% (7), or 60% within 12 months (8). If CNI discontinuation (9) or reduction (9–11) occurred after 12 months, including other citated papers (11), acute rejection rates ranged from 0% to 30%.

Two reports document AZA as the primary immunosuppressive agent associated with CNI withdrawal (12,13). In one report (12), 12 patients with cyclosporine A (CsA)-related renal dysfunction at least 12 months after transplant received AZA (2 mg/kg/day) combined with CsA withdrawal. Improvement in renal function was minimal (mean follow-up: 18 months), and six patients (50%) had biopsy-proven acute rejection (two steroid resistant) and two died.

In the second study (13), the results were substantially different. Here, 26 patients received AZA associated with discontinuation or reduction of CsA (median time after LT: 10 months, range 18–57). When CsA was discontinued (n = 14), the mean serum creatinine decreased from 2.42 ± 0.48 to 1.72 ± 0.39 mg/dL (p = 0.00004) with no rejection; when CsA was reduced (n = 12), the mean serum creatinine decreased less markedly (2.08 ± 0.34 to 1.85 ± 0.41 mg/dL; p = 0.69) with one rejection episode.

In LT, no RCT has compared MMF and AZA in patients with CNI-related nephrotoxicity. There is only one retrospective comparative analysis (14) reporting 33 patients with CNI-related nephrotoxicity converted to MMF 1 g b.d. (n = 23) or to AZA 2 mg/kg/day (n = 10) after a mean of 63 ± 51 months. An improvement in creatinine clearance of more than 10% occurred in 16 patients (49%), all receiving MMF (significant univariately, but multivariate analysis was not performed). Six episodes of acute rejection occurred (three MMF and three AZA; p = ns). Side effects were more frequent with MMF: 10 versus 0 (p < 0.05).

Two studies report outcomes after replacing AZA by MMF in patients with renal dysfunction with reduction (15) or discontinuation of CNIs (8). The first (15) evaluated 32 patients, who had MMF (1 g b.d.) added and AZA discontinued at a mean interval of 25.6 ± 34.7 (range 1.2–127.8) months after LT. Mean follow-up after starting MMF was 4.8 ± 0.6 (range 3.1–6) years, with only two biopsy-proven acute rejection episodes. Discontinuation of MMF was necessary in three patients (9.4%), only one due to rejection. At the end of the study, 13 patients (41%) had normal creatinine values. The second study (8) was an RCT: 18 patients with creatinine >150 μmol/L were randomized to MMF (500 mg b.d., increased to 1 g b.d. at 2 weeks) and CNI dose reduction, and then discontinuation at 3 months (study group, n = 9) or to continue to receive AZA and CNIs without dose reduction (control group, n = 9); unfortunately, the failure to reduce CNI negates a useful comparison with MMF. Serum creatinine concentrations fell in six of eight patients treated for more than 1 month following CNI reduction. However, two of five patients receiving MMF who completed 3 months follow-up, developed severe ductopenic rejection requiring retransplantation, and another patient had steroid-responsive severe rejection. Therefore, the study was stopped. It is unclear whether the dosage of MMF was insufficient (2 g/day) and/or whether CNI discontinuation was the major factor.

LT for HCV-related cirrhosis

AZA and MMF both inhibit IMPDH and have some common features with ribavirin, which enhances the activity of interferon in the treatment of HCV. However, whether MMF or AZA are beneficial in HCV-infected liver transplant recipients and whether MMF is more effective compared to AZA is not clear.

Firpi et al. (16) evaluated MMF and HCV viral load randomizing 30 nontransplanted patients with chronic HCV hepatitis to four regimens (1000 mg b.d., 500 mg b.d., 250 mg b.d. or a matched oral placebo b.d.) for 8 weeks. No subject who was HCV–RNA positive on entry had a one-log decrease in virus concentration or became virus negative and no subject normalized serum alanine aminotransferase (ALT) level. There were no significant differences within or between groups. A prospective 9-month cross-over study evaluated MMF (1 g/b.d.), used as substitute for AZA, in 13 transplanted patients with recurrent chronic HCV (17), with a background of CNI and steroids. In all patients at the end of the MMF treatment period, the mean viral load had significantly increased compared to baseline (1.64 × 106± 1.34 × 106 vs. 0.74 × 106± 0.47 × 106 mRNA copies/mL; p = 0.02). When AZA was reintroduced the mean viral load declined and at the end of the study, the mean values were not statistically different to baseline AZA values. However, the increase in viral load compared to baseline (58.2 ± 37.5 vs. 80 ± 69 U/L) was not associated with ALT flares.

Nine studies between 1996 and 2008 evaluated AZA and HCV recurrence (18–26). AZA was associated with reduced severity of HCV recurrence in five studies (56%) (18,19,22,23,25), evaluated in a univariate analysis in one, multivariate in three, and with chi-square test in one, whereas in four studies (44%) (20,21,24,26) there was similar severity. However, no study with AZA showed increased severity of recurrent HCV. In contrast in 17 studies evaluating MMF and HCV recurrence between 2001 and 2007 (5,19,23,24,26–38), only two studies (12%) (31,33) found decreased severity of HCV recurrence, and in one (31) there was no multivariate analysis. Nine studies (53%) (5,19,24,26–30,32) documented similar severity of HCV recurrence, but six (35%) (23,34–38) showed increased severity of HCV recurrence (Tables 2 and 3).

Table 2.  Studies with AZA (5/9) and MMF (2/17) documenting reduced severity of HCV recurrence
AuthorNo. of patients total/HCVOutcomeVariablesResultsp-Value
  1. Analysis performed with: 1chi-square test; 2univariate analysis; 3multivariate analysis.

Azathioprine
 Hunt et al. (22) 200165/65HCV recurrenceUse vs. nonuse of AZA16/17 vs. 37/48<0.005 
 Progression of HCV recurrence 1/6 vs. 18/370.014
 Berenguer et al. (19) 2002522/283Cirrhosis (fibrosis stage 4)Induction without AZA2Associated in the univariate analysis
 Berenguer et al. (23) 2003554/554Severe HCV recurrence (Fibrosis 3 and 4 within the first 2 years)AZA use <12 months3OR 3.24; 95%CI 1.51–6.960.003
 Samonakis et al. (18) 2005193/193Overall survivalAZA at 3 months3OR 0.3; 95%CI 0.18–0.64
 3 months survivalNo maintenance AZA3OR 0.3; 95%CI 16–0.64
 Severe fibrosisNo maintenance AZA20.029
 Eid et al. (25) 200792/92Cirrhosis (Fibrosis stage 4)Use of AZA3OR 0.37; 95%CI 0.14–0.920.033
Mycophenolate mofetil
 Bahra et al. (31) 200580/80Fibrosis score (after 24 months of treatment)MMF vs. no MMF1 ± 0.4 vs. 2.05 ± 1.02<0.013 
 Sanchez-Bueno et al. (33) 2006476/142HCV recurrenceNo maintenance MMFOR 530.015
Table 3.  Studies with MMF (6 of 17) documenting increased severity of HCV recurrence
AuthorNo. of patients Total/HCVOutcomeVariableResultsp-Value
  1. Analysis performed with: 1chi-square test; 2univariate analysis; 3multivariate analysis.

Berenguer et al. (37) 2000284/284Fibrosis progression2Induction with MMF0.05 
Fasola et al. (38) 2002125/125Severe fibrosis1No/low-dose MMF vs. high-dose MMFNumber of patients: 26 vs. 00.07 
Burak et al. (36) 200293/93Graft failure3Use of MMF vs. AZARR 2.75; 95%CI 1.14–6.610.024
Berenguer et al. (23) 2003554/554Severe HCV recurrence3 (Fibrosis 3 and 4 within the first 2 years)Induction with MMFOR 8.89; 95%CI 2.67–29.60.001
Kornberg et al. (34) 200521/21Stage of fibrosis (6 months vs. pre-MMF)1Use of MMFStage increasing (1.5 ± 0.5 vs. 2.3 ± 0.5)0.04 
 
Kornberg et al. (35) 200719/19Stage of fibrosis (12 months vs. pre-MMF)1Use of MMFStage increasing (1.5 ± 0.5 vs. 2.3 ± 0.7)0.02 

A single direct comparison between MMF and AZA was evaluated in 56 HCV-positive liver transplanted patients—an RCT subgroup (5): 27 received CsA, steroids and MMF (1 g b.d. i.v. for 4 to 10 days, then oral 1.5 g b.d.) and 27 received CsA, steroids and AZA (1–2 mg/kg/day i.v. then by oral administration). The combined incidence at 6 months after transplantation of acute cellular rejection and/or graft loss was significantly reduced with MMF compared to AZA (30.6% vs. 41.4%; p < 0.04). A multivariate analysis showed that MMF was superior in preventing rejection within 6 months. Graft loss rates were 16.1% (AZA) and 9.4% (MMF). Histological HCV recurrence, associated with HCV–RNA in serum, was 18.5% MMF and 29.1% AZA (p = not reported) at 6 months. However, despite 1-year data being available for the main study (5), the 1-year and other long-term data for the HCV subgroup were not available. In particular, HCV recurrence rate and its severity were not reported.

Recently, Kornberg et al. (34) performed a prospective study, with contrasting early and delayed results: 21 patients received quadruple induction CsA-based immunosuppression, augmented by MMF (n = 12) or by AZA (n = 9). Recurrent HCV disease was diagnosed earlier with MMF, than with AZA (50 ± 35 vs. 35 ± 35 weeks), but patients taking MMF had less severe allograft fibrosis at diagnosis of disease recurrence (Ishak–Knodell) (1.5 ± 0.5 vs. 2.2 ± 1.2). However, stage of fibrosis significantly increased with MMF, during 6 months of antiviral treatment compared to AZA (1.5 ± 0.5 vs. 2.3 ± 0.5; p = 0.04). As details of the adequacy of liver biopsy samples were not given, and this is important for staging, these paradoxical results of earlier recurrence with less severe disease are difficult to interpret (39).

The data on MMF and AZA with respect to HCV recurrence need to be better known, given that 41% of 36 US transplant centers responding to a questionnaire about patients transplanted for HCV-related cirrhosis.(40), used MMF, and that azathioprine use was not mentioned at all.

Economic evaluation

There are no cost-effectiveness studies comparing MMF versus AZA in LT. However, in the United Kingdom the cost of 1-month treatment with MMF (1 g b.d.) is £210 compared to AZA (100 mg/day) at £6.41 (41). Considering that the superiority of MMF over AZA in clinical practice is not clearly demonstrated as outlined above, it is unlikely that a formal cost-effectiveness analysis would be in favor of MMF. Indeed, over £2400/patient/year would be saved by using AZA 100 mg/day in substitution for 2 g/day MMF.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References

In this review, we question the clinical benefit of MMF compared to AZA in LT, not only for rejection, but also in relation to renal dysfunction and HCV recurrence.

The approval to license MMF for use after LT was based on a single registration trial comparing MMF to AZA (5). Although treated rejection rates with histological confirmation were about 25% lower in MMF than in AZA-treated patients (38.5% vs. 47.7%) the dose used was 1.5 g b.d., whereas the commonly used dose is 1 g b.d. as higher doses cause diarrhea. Moreover, with 1.5 g/b.d. there was no improvement in graft or patient survival. Rejection rates with MMF 3 g/day were twice as high as in the only other RCT, which was unblinded (4), which used 2 g/day. The same criteria for obtaining liver biopsies on the basis of suspected rejection were used so that the evaluation of rejection in these studies cannot be considered reliable. Again, graft and patient survival were not improved compared to AZA. Thus, the improvement in rejection of MMF compared with AZA is marginal and, furthermore, the interpretation of the registration RCT is complicated by not being able to separate rejection alone, from graft loss without rejection within 6 months, and graft loss within 12 months, in the combined endpoint. The studies reported in Table 1, comparing MMF versus AZA with respect to incidence of acute cellular rejection, all used cyclosporine-based immunosuppressive regimens. As a vast majority of liver transplant programs, particularly in the United States, use tacrolimus-based immunosuppression any differences between AZA and MMF for rejection might be obviated by the use of tacrolimus.

Even if lower rates of rejection were to be substantiated in future comparisons of MMF (2 g/day) versus AZA (1 mg/kg), it is known that acute cellular rejection is less important with respect to liver graft function and survival, than for other solid organ transplants. Indeed, Wiesner et al. (42) showed that the occurrence and successful treatment of a single rejection episode favored increased survival. Neither AZA nor MMF resulted in different rates of multiple rejection episodes: 7.7% (AZA), 5.8% (MMF). Moreover, mild rejection does not require therapy: the outcome at 3 months was no different between patients who received or did not receive treatment for rejection in a large study (43).

In patients with chronic renal dysfunction, the introduction of MMF followed by a complete withdrawal of CNI leads to an increased incidence of rejection, more likely if CNI are reduced or withdrawn within 12 months of transplantation (7,8). Even if rejection is not increased, although one study (8) was stopped due to 50% of patients experiencing severe rejection, the improvement in renal function is mostly marginal. Indeed, stoppage of CNI may be dangerous as chronic rejection is reported (8). Last, no randomized controlled trials are reported directly comparing MMF and AZA in association with either reduction or discontinuation of CNI. It could be that AZA may be as good or as bad as MMF in this context. There are no adequate studies.

The comparison between MMF and AZA is most intriguing with regard to the severity of HCV recurrence after LT. Both drugs in vitro have antiviral activity against HCV, and in one in vivo study AZA appeared to have more HCV antireplicative potential (17). However, no histological correlate was evaluated. In contrast, in clinical studies, albeit all observational, the rate of HCV recurrence and/or its severity was less in 56% of studies with AZA (5/9) with no study showing worsening, whereas this was the case in only 12% (2/17) with MMF, and 6 (35%) showed worsening histology.

The data described in this review suggest that the perceived clinical benefits of MMF over AZA are not supported by good evidence. Randomized controlled studies are needed to reevaluate the place of azathioprine for maintenance immunosuppression in LT especially with a tacrolimus-based immunosuppressive regimen. Studies will need to be investigator led and supported by funding bodies. Evaluation of HCV replication and HCV recurrence in the allograft will be of particular importance as AZA may have advantages over MMF for patients transplanted with HCV cirrhosis.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. References