Withdrawal of calcineurin inhibitors (CNI) followed by mycophenolate mofetil (MMF) monotherapy after liver transplantation (LT) remains controversial due to the increased risk of acute rejection and graft loss. The aim of the present study, performed in a large cohort of liver-transplanted patients with severe CNI-induced side effects, was to assess renal function recovery, and safety in terms of liver function, of complete CNI withdrawal and replacement by MMF monotherapy. Fifty-two patients treated with MMF monotherapy for CNI-induced toxicity were analyzed. Mean estimated glomerular filtration rate (eGFR) increased significantly during the period of MMF monotherapy, from 37 ± 10 to 44.7 ± 15 mL/min/1.73 m2 at 6 months (p = 0.001) corresponding to a benefit of +17.4% in renal function. eGFR stabilized or improved in 86.5%, 81% and 79% of cases, and chronic renal dysfunction worsened in 13.5%, 19% and 21% of cases, at 6, 12 and 24 months after CNI withdrawal, respectively. Only two patients experienced acute rejection. MMF monotherapy may be efficient at reversing/stabilizing CRD, and appears relatively safe in terms of liver graft function in long-term liver-transplanted patients. However, clinicians must bear in mind the potential risk of rejection and graft loss, and should be very cautious in the management of such ‘difficult-to-treat patients’.
The classical immunosuppressive regimen widely used after liver transplantation (LT) includes calcineurin inhibitors (CNI), which have significantly improved both organ and recipient survival since the 1980s. However, they induce subsequent severe side effects such as chronic renal dysfunction (CRD) in long-term-transplanted patients (1,2). In this setting, a combination involving mycophenolate mofetil (MMF) has been proposed in many protocols to improve long-term outcome in patients with or without HCV (3), and to recover renal function through a progressive decrease in the CNI daily dose (also referred to as ‘sparing strategy’) (4,5). Most studies showed improvement in renal function, but recovery was more pronounced when CNI were totally withdrawn (6,7). To improve management of patients with CNI-induced severe side effects, complete CNI withdrawal which was then followed by MMF monotherapy had been previously evaluated in small-cohort-size studies, which reported an average 15% risk of acute rejection (6,8–15). Therefore, complete replacement of CNI followed by MMF monotherapy remains highly controversial (9,15).
The aim of the present study, performed in a large cohort of liver-transplanted patients with severe CNI-induced side effects, was to assess renal function recovery, and safety in terms of liver function, of complete CNI withdrawal and replacement by MMF monotherapy.
Patients and Methods
Patients and study design
We performed a retrospective chart review of all adult liver-transplanted patients treated with either MMF alone or MMF with the addition of low doses of corticosteroids for severe toxicity related to CNI (CRD, severe arterial hypertension, dyslipidemia) in two French transplant centers from 1991 to 2004. In the overall population, liver graft and renal function were the two main endpoints. The following variables were collected: parameters of liver graft function (prothrombin time (PT), bilirubin, gamma-glutamyltransferase (γGT), alanine aminotransferase (ALT), aspartate aminotransferase (AST)) and renal function (estimated glomerular filtration rate: eGFR). To avoid bias and to homogenize the analysis, data collection was performed at the following intervals: (1) at the time of LT; (2) the third month after LT; (3) at the start of MMF; (4) at CNI withdrawal and (5) at 6, 12 and 24 months after CNI withdrawal. Informed consent was obtained in all cases.
Immunosuppressive regimen and withdrawal protocol
The immunosuppressive regimen was based on an initial combination of CNI, corticosteroids and azathioprine until September 2000. During the first month, target windows for the CNI concentration in sera were between 150 and 250 ng/mL for cyclosporine and between 6 and 12 ng/mL for tacrolimus. From October 2000 on, MMF replaced azathioprine in the primary immunosuppressive regimen. Therefore, the time of start-up of MMF varied in the population studied. MMF was introduced during the first year post-LT in 10 cases, from 1 to 10 years in 30 cases and after 10 years post-LT in 12 cases. CNI was progressively reduced by 20–25% stepwise every month and withdrawn within 6 months.
Evaluation of renal function
Renal function was evaluated by the eGFR using Modification of Diet in Renal Disease (MDRD) formula. This formula is more accurate than use of a serum creatinine measurement alone and/or the Cockroft–Gault formula, which may underestimate the true incidence of CRD. CRD is multifactorial in origin (hepatitis virus C infection, diabetes, ischemic nephropathy), but the main cause was related to CNI chronic nondose-dependent nephrotoxicity. Patients with CRD prior to LT were excluded. CNI-related CRD was defined by a persistent decrease in eGFR < 60 mL/min/1.73 m2 measured on at least two consecutive occasions; absence of significant hematuria/proteinuria on dipstick analysis and absence of renal artery stenosis or urinary tract disease. Renal biopsy and measurement of urinary protein excretion (proteinura/24 h) were not routinely performed in the management of patients.
We specifically sought to assess renal function variation (ΔeGFR) during period of MMF monotherapy. Renal recovery was defined by a Δ eGFR of more than +10% during MMF monotherapy. Renal stabilization was defined by a ΔeGFR between a target window of −10 to +10% and CRD progression by a ΔeGFR beyond −10%.
Continuous variables are expressed as mean ± SEM with their range. We used the paired t-test to compare quantitative variables between each period of analysis. Statistical analysis was performed using NCSS 2001 software (NCSS, Kaysville, UT).
Characteristics of the population
We identified 52 patients treated with MMF monotherapy and followed from the time of LT for a mean period of 121 ± 53 months. Main characteristics of the population are provided in Table 1. Mean time with the usual dose of CNI during which CNI toxicity occurred was 70 ± 53 months and mean length of MMF monotherapy was 41.4 ± 20 months.
Table 1. Characteristics of 52 patients treated by MMF monotherapy
Age at LT (years), mean ± SEM (range)
47.2 ± 9 (26–65)
Gender M/F, number (%)
43 (83%)/9 (17%)
Cause of liver disease
Alcoholic cirrhosis, number (%)
Viral cirrhosis, number (%)
Other causes, number (%)
Presence of HCC, number (%)
Type of CNI toxicity
Renal insufficiency, number (%)
Metabolic disorders, number (%)
Patients were treated with MMF alone in 32 cases (61.5% of cases) and MMF with the addition of low doses of corticosteroids in 20 cases. The mean daily corticosteroid dose was 3.3 ± 0.7 mg (from 10 to 20 mg in 11 cases, 7.5 mg in two cases and less than or equal to 5 mg in seven cases). There were no statistical differences in terms of clinical characteristics between patients treated with MMF alone and those with MMF and low doses of corticosteroids (data not shown). The main reason for terminating MMF monotherapy was the presence of CRD (n = 46). Only one of the 46 patients with CRD underwent liver biopsy. The only available biopsy showed typical features related to CNI-induced chronic toxicity, with arteriolopathy and tubulointerstitial fibrosis.
Thirteen patients were diabetic, among whom five required insulin therapy and the remaining eight were treated with metformine. In the overall population, 26 (50% of cases) were treated for hypertension. Eleven were on angiotensin-converting enzyme inhibitors (fosinopril, six cases; lisinopril, three cases; captopril, one case; perindopril, one case) and five were on angiotensin receptor blockers (losartan, two cases;valsartan, two cases; irbesartan, one case).
Kinetics of liver and renal function tests
The first endpoint was graft function during the phase of MMF monotherapy, based on liver function tests. The course of liver function tests is shown in Table 2. Mean values of AST, ALT, γGT, PT and bilirubin remained unchanged in the 52 patients during MMF monotherapy, suggesting the absence of wide-scale graft dysfunction. In the 46 patients with CRD, the course of renal function according to the mean eGFR is provided in Figure 1. Briefly, mean eGFR significantly dropped during the first two periods of study after LT (79 ± 35 mL/min/1.73 m2) to the third month after LT (57 ± 23 mL/min/1.73 m2; p = 0.001), and from the third month (57 ± 23 mL/min/1.73 m2) to start of MMF (41 ± 15 mL/min/1.73 m2; p = 0.001), respectively. Mean eGFR also decreased during the sparing strategy period, to 37 ± 10 mL/min/1.73 m2 (p = 0.045). Once CNI was withdrawn, mean eGFR increased significantly during the period of MMF monotherapy, from 37 ± 10 to 44.7 ± 15 mL/min/1.73 m2 at 6 months (p = 0.001), corresponding to a benefit of +17.4% in renal function. The recovery of CRD was maintained at 12 and 24 months after CNI withdrawal with a mean eGFR at 45.5 ± 17 (p = 0.001) and 44 ± 18 mL/min/1.73 m2 (p = 0.002), respectively. Finally, eGFR stabilized or improved in 86.5%, 81% and 79% of cases, and CRD worsened in 13.5%, 19% and 21% of cases, at 6, 12 and 24 months after CNI withdrawal, respectively. Among patients whose condition worsened, dialysis was required in five cases (11%).
Table 2. Evolution of liver function test results during MMF monotherapy following CNI withdrawal (n = 52)
In terms of potential selection bias, it should be pointed out that 701 patients at the two centers received LT from 1991 to 2004. Among them, 31 patients (4.4% of cases) underwent retransplantation for primary nonfunction or hepatic artery thrombosis in 15 cases, and recurrent viral cirrhosis in 16 cases. Over this 13-year period, 284 deaths occurred in the overall population of liver recipients; 21 of these (7.4% of cases) involved rejection leading to graft loss. None of the deceased patients were treated with MMF monotherapy.
For our 52 patients, seven deaths occurred during the phase of MMF monotherapy: two patients died from myocardial infarction at 71 and 119 months after LT; two patients died from recurrent alcoholic cirrhosis at 55 and 67 months; one patient succumbed to recurrent cholangiocarcinoma at 47 months and two patients died from septic shock at 45 and 152 months. Two patients experienced biopsy-proven acute rejection (incidence of 4%) during this phase, leading to reintroduction of CNI in one case. No secondary graft loss was observed. The patient who returned to CNI died from myocardial infarction at 119 months after LT. His renal function had worsened and eGFR dropped from 40 to 21 mL/min/1.73 m2 at the time of death. The second patient with acute rejection was treated with corticosteroid boluses followed by low oral dose of steroids; MMF was maintained. CRD continued and the patient required dialysis and then kidney transplantation 5 years after CNI withdrawal.
Numerous factors are involved in the increasing rate of CRD after LT, including length of exposure to CNI, increased recipient age, male gender and the presence of pretransplantation hypertension, diabetes mellitus and hepatitis C (1). Clinicians face difficulties in delaying or reversing these complications in long-term-transplanted patients. The development of a new immunosuppressive schedule is challenging in difficult-to-treat patients such as those with CRD (6,7). In the present study, we focused on a large cohort of patients treated with MMF monotherapy with long follow-up patients and we observed that this strategy was efficient for recovery of renal function and relatively safe for liver graft.
Previous experience with MMF monotherapy had been inconclusive and highlighted an unacceptable risk of acute rejection (9,14–17). We did not observe such major adverse effects for several reasons: first, MMF monotherapy was delayed, and was achieved more than 5 years after LT; second, the main cause of liver disease was alcoholic cirrhosis, which is a recognized favorable immunological context, with a lower risk of rejection (18); third, replacement of CNI was progressive and required several months to once again minimize this risk. The course of liver function tests was satisfactory, though liver biopsies were not routinely performed so as to definitively avoid infra-clinical rejection.
Moreover, a minority of patients (18 patients) who were stabilized in their MMF regimen underwent plasma mycophenolic acid (MPA) assessment (data not shown). They displayed a mean MPA AUC of 48 μg h/mL, which was within the high part of the target window known to be protective against rejection following renal transplantation (19). Our data on therapeutic drug monitoring of MPA involved too few patients to enable firm conclusions, but should encourage future studies on the potential advantage of monitoring MPA in liver transplant recipients with CRD. We also recognize that the main draw back of our approach is that we followed an uncontrolled pragmatic schedule but our results gave a clinical rationale to promote future controlled studies in this setting.
Finally, MMF monotherapy may be efficient at reversing/stabilizing CRD, and appears relatively safe in terms of liver graft function in long-term liver-transplanted patients. However, clinicians must bear in mind the potential risk of rejection and graft loss, and should be very cautious in the management of such ‘difficult-to-treat patients’. In the future, therapeutic drug monitoring of MPA may help to monitor MMF monotherapy under acceptable safety conditions.
The database from which results were extracted was supported by Roche France.