Everolimus, a mammalian target of rapamycin (mTOR) inhibitor, has been shown in kidney1-6 and heart transplant patients7-11 to offer effective immunosuppression and to preserve renal function in de novo1, 2, 4-6 and maintenance settings.3, 9, 11, 12 The initiation of everolimus with the discontinuation of calcineurin inhibitors (CNIs) or reduced exposure to CNIs can preserve renal function after kidney or heart transplantation if this initiation occurs early enough after transplantation3, 9, 12, 13 and before CNI-related nephrotoxic effects have become irreversible.14
Experience with everolimus in liver transplant recipients, however, is more limited. There are no published data on the de novo use of everolimus after liver transplantation, although this is under investigation.15 One randomized trial assessed the impact of converting 78 patients from CNI therapy to everolimus very early (day 30) after liver transplantation.16 Renal function was significantly better at 1 year in the everolimus cohort, and this improvement was accompanied by a lower incidence of chronic kidney disease (stage 3 or higher); however, the rate of acute rejection was similar to the rate for the CNI arm. Single-arm studies have demonstrated that maintenance liver transplant patients can be converted from CNIs to everolimus safely at later time points with a low rate of acute rejection.17-22 Like kidney transplant recipients,23 liver transplant recipients with kidney dysfunction apparently need to be switched from CNI immunosuppression to everolimus either early after transplantation16 or before the establishment of severe dysfunction.17
In clinical practice, some centers now initiate everolimus in maintenance liver transplant patients who develop kidney dysfunction or, less frequently, posttransplant neoplasms and attempt to withdraw CNI therapy. The aim of this multicenter, retrospective study was to analyze the current indications for the conversion of liver transplant recipients to everolimus, the employed regimens and exposure levels, and the impact on efficacy and safety.
CNI, calcineurin inhibitor; C0/C2, cyclosporin trough blood level/2-hour cyclosporine blood level; eGFR, estimated glomerular filtration rate; HCC, hepatocellular carcinoma; MELD, Model for End-Stage Liver Disease; MPA, mycophenolic acid; mTOR, mammalian target of rapamycin; SD, standard deviation.
PATIENTS AND METHODS
This was a retrospective, multicenter analysis undertaken at 9 liver transplant centers in France. The analysis included all maintenance liver transplant patients in whom everolimus treatment was initiated from October 2005 to November 2008, regardless of the preceding immunosuppressive regimen or the indication for the introduction of everolimus. No Institutional Review Board approval was necessary.
Data were obtained from patients' medical records at the baseline (ie, the day of their conversion to everolimus), 1 month after conversion, and every 3 months after conversion until the end of the study period. The following information about the original transplant procedure and each patient's status at the time of transplantation was obtained: the time since transplantation, the indication for primary transplantation or retransplantation, the Model for End-Stage Liver Disease (MELD) score, the serum creatinine and total bilirubin levels, the international normalization ratio, the body weight, the presence or absence of diabetes and the type of antidiabetic therapy, the presence of hypertension, the donor type and age, the graft type, and the recipient/donor cytomegalovirus status. The reason for the introduction of everolimus (including details about the location of any cancer leading to everolimus initiation and previous cancer treatment) was recorded. On day 0 and at all post-baseline visits, the following data were collected: the current types, doses, and blood concentrations of immunosuppressive agents; hematological data; the serum creatinine and liver enzyme levels (total bilirubin, aspartate aminotransferase, alanine aminotransferase, and gamma-glutamyltransferase); the prothrombin time activity; the presence or absence of diabetes and the type of antidiabetic therapy; the glucose, total and low-density lipoprotein cholesterol, and triglyceride levels; the use of antihyperlipidemic therapy; the 24-hour urinary protein level; the body weight; the arterial blood pressure and the presence of hypertension; and the alpha-fetoprotein and hepatitis C RNA levels in patients with hepatocellular carcinoma (HCC) and in patients with hepatitis C infections, respectively. The glomerular filtration rate was estimated with the Cockcroft-Gault formula.24 The estimated glomerular filtration rate (eGFR) was analyzed for the total population and specifically for 2 subpopulations of patients: (1) those who had been converted to everolimus because of chronic renal failure (established by the investigator) and (2) those who had a baseline eGFR value < 60 mL/minute.
The primary objective of the analysis was to describe the current indications for everolimus use in maintenance liver transplant patients in routine clinical practice. The secondary objectives were (1) to determine the effect of everolimus conversion on renal function in patients with or without chronic renal failure, (2) to examine the effect of everolimus conversion on the risk of biopsy-proven acute rejection, (3) to assess the doses and trough levels of everolimus, (4) to identify the modifications made to the CNI-based regimen, and (5) to evaluate adverse events and tolerability after everolimus conversion in this setting.
For comparisons between patients, the χ2 test was used for categorical data, and the independent sample t test was used for continuous data. Biological values and immunosuppressive doses on day 0 and at subsequent time points were compared with the paired t test. Overall survival probabilities were determined with the life-table method and were compared with the log-rank test. A P value <0.05 was considered significant. Statistical analyses were performed with SAS 9.1.3 (SAS Institute, Inc., Cary, NC).
Two hundred forty of 1680 liver transplant patients were converted to everolimus therapy during the study period and were included in the analysis. The majority of the patients (70.4%) were male. Over half (52.5%) had undergone transplantation for alcoholic cirrhosis, whereas HCC was present in 42.5% (Table 1). Almost all the patients (n = 234) had undergone liver transplantation for the first time, and 90.8% received a deceased donor graft. The mean time from transplantation to the introduction of everolimus was 4.9 ± 5.2 years (median = 3.0 years, range = 0.1-23.3 years). The most frequent reasons for the introduction of everolimus were chronic renal failure (48.6%), the prevention of HCC recurrence (18.3%), the management of recurrent HCC (15.8%), and acquired de novo malignancies (14.2%; Table 1). The mean duration of follow-up after the introduction of everolimus was 15.3 ± 13.1 months (median = 13.6 months, range = 0.6-113 months).
Table 1. Baseline Characteristics and Reasons for the Introduction of Everolimus (n = 240)
Patients could have more than 1 indication for liver transplantation.
Patients could have more than 1 reason for the introduction of everolimus.
At the time of conversion, almost all patients (226/240 or 94.2%) were receiving CNI therapy, over half were receiving mycophenolic acid (MPA; 126/240 or 52.5%), and almost a third (77/240 or 32.1%) were receiving steroids. A dual-therapy regimen was being used in 49.6% of the patients, and monotherapy was being used in 32.1% (Table 2).
Table 2. Immunosuppressive Drugs and Regimens at the Time of Everolimus Introduction (n = 240)
NOTE: All continuous variables are shown as medians and ranges.
Steroids, n (%)
Cyclosporine, n (%)
Tacrolimus, n (%)
MPA, n (%)
Mycophenolate mofetil, n (%)
Enteric-coated mycophenolate sodium, n (%)
Triple therapy, n (%)
Cyclosporine, steroids, and MPA
Tacrolimus, steroids, and MPA
Dual therapy, n (%)
Tacrolimus and steroids
Cyclosporine and steroids
Steroids and MPA
Cyclosporine and MPA
Tacrolimus and MPA
Monotherapy, n (%)
Everolimus was introduced at a mean dose of 2.4 ± 0.8 mg/day (median = 2 mg/day). The numbers of patients receiving everolimus monotherapy 1, 3, 6, and 12 months after conversion were 11, 26, 49, and 49, respectively; these represented 5.1%, 13.2%, 27.5%, and 33.6% of all patients for whom data on immunosuppression were available at each time point. The mean everolimus trough level was 7.3 ± 4.1 ng/mL at month 1 and 8.1 ± 4.7 ng/mL at month 12 across the total population, with higher values in the everolimus monotherapy cohort (8.8 ± 4.9 ng/mL at month 12; Table 3).
Table 3. Evolution of Immunosuppression After the Introduction of Everolimus
Day 0 (n = 216)
Month 1 (n = 215)
Month 3 (n = 197)
Month 6 (n = 178)
Month 12 (n = 146)
Mean ± SD
2.4 ± 0.8
2.7 ± 1.2
2.9 ± 1.4
2.9 ± 1.3
2.9 ± 1.43
Monotherapy, n (%)
C0, ng/mL, mean ± SD
7.3 ± 4.1
7.5 ± 4.9
7.7 ± 4.1
8.1 ± 4.7
5.2 ± 2.4 (n = 9)
8.7 ± 8.1 (n = 20)
8.8 ± 4.9 (n = 40)
8.8 ± 4.9 (n = 40)
7.4 ± 4.1 (n = 137)
7.3 ± 4.1 (n = 112)
7.1 ± 3.6 (n = 79)
7.5 ± 4.4 (n = 42)
Steroids, n (%)
Cyclosporine, n (%)
C0, ng/mL, mean ± SD
138 ± 112
151 ± 181
149 ± 112
111 ± 76
111 ± 45
Tacrolimus, n (%)
C0, ng/mL, mean ± SD
6.6 ± 3.5
5.0 ± 3.1
4.4 ± 2.8
4.7 ± 2.5
5.8 ± 3.5
Mycophenolate mofetil and enteric-coated mycophenolate sodium, n (%)
Twelve months after the introduction of everolimus, 61.6% of the patients (90/146) were no longer receiving CNI therapy. The proportion of patients receiving steroids was unchanged, but the use of MPA had decreased from 50.5% to 26.7% (Table 3).
In the entire population, the mean serum creatinine level was 140 ± 69 μmol/L on day 0, decreased to 132 ± 69 μmol/L at month 1 (P = 0.009 versus day 0), and remained stable thereafter (month 3, 129 ± 69 μmol/L, P = 0.04 versus day 0; month 6, 131 ± 70 μmol/L, P = 0.11; and month 12, 134 ± 74 μmol/L, P = 0.43). The mean eGFR value showed a significant improvement from day 0 (64.2 ± 30.0 mL/minute) to month 12 (68.4 ± 32.5 mL/minute, P = 0.007); this change was evident by month 1 (Fig. 1A). In the subpopulation of patients for whom everolimus was introduced because of chronic renal failure (established by the investigator), eGFR was calculated for 103 of 119 patients. The mean eGFR values improved significantly by month 1 and remained elevated at month 12 (day 0, 47.2 ± 17.5 mL/minute; month 1, 53.8 ± 23.1 mL/minute, P = 0.0004 versus day 0; month 3, 56.3 ± 25.3 mL/minute, P = 0.0004; month 6, 53.8 ± 20.5 mL/minute, P = 0.0002; and month 12, 55.8 ± 26.5 mL/minute, P = 0.02). When eGFR was reanalyzed in the subpopulation of patients known to have poor renal function on day 0 as determined by eGFR < 60 mL/minute (n = 92), there was also a significant improvement from day 0 (39.8 ± 11.1 mL/minute) to month 12 (44.8 ± 15.8 mL/minute, P = 0.0006; Fig. 1B). For patients whose eGFR was ≥ 60 mL/minute on day 0, a nonsignificant trend of improvement was observed in the mean eGFR values from day 0 to month 12 (85.7-89.5 mL/minute, P = 0.052). The change in eGFR was also analyzed according to the time of everolimus conversion. When everolimus was started after the first year post-transplant (n = 172), eGFR improved from 59.1 to 64.6 mL/minute (P = 0.01), but it increased from 77.5 to 90.0 mL/minute (P = 0.04) for the patients in whom everolimus was initiated during the first year after transplantation (n = 68).
Using a repeated measures approach for patients who had eGFR measurements available for all study visits, we found that the mean eGFR in the global population (n = 91) increased significantly from 63.4 mL/minute at the baseline to 68.2 mL/minute at 12 months (P = 0.007). In the 43 patients with eGFR values < 60 mL/minute at baseline, the increase remained significant (41.0-45.9 mL/minute, P = 0.004), but it was not significant in the 48 patients with eGFR values ≥ 60 mL/minute (83.4-88.2 mL/minute, P = 0.30).
Renal function deteriorated by 1 stage according to the Risk, Injury, Failure, Loss, and End-Stage classification of chronic kidney disease25 in 10 patients and progressed to chronic kidney disease (stage 4 or higher); the patients required chronic dialysis.
Survival and Causes of Death
Two years after conversion, the patient survival rate was 95.0%. Fourteen patients (5.0%) died between 0.4 and 48 months; the most frequent causes were HCC recurrence (n = 6; 1.1, 4.3, 7.4, 9.8, 13.6, and 29.8 months after conversion), cerebrovascular accident (n = 2; 6.8 and 5 months after conversion), and sepsis (n = 3; 27.2, 31.1, and 52.7 months after conversion).
No HCC recurrence was observed among the 44 patients who underwent transplantation for the prevention of HCC recurrence during a mean follow-up time of 11.2 ± 6.8 months after everolimus conversion.
Four patients (1.6%) developed mild (n = 2) or moderate biopsy-proven acute rejection (n = 2) after the introduction of everolimus (mean time = 2.0 ± 1.8 months). At the time of rejection, 3 of the patients were receiving mycophenolate mofetil, and 1 patient was receiving a low dose of steroids in addition to everolimus. The rejection episodes were resolved after an increase in the everolimus dose (n = 4) and the reintroduction of a CNI (n = 2). All 4 patients were alive with a functioning graft at the last follow-up visit.
Over the study period, 658 adverse events occurred in 203 patients (84.6%) at a mean of 4.3 ± 4.3 months (median = 3 months) after their conversion to everolimus. The most frequently reported adverse events were gastrointestinal disorders (22.9%), cutaneous rash (18.8%), edema (16.3%), hypertriglyceridemia (14.6%), mouth ulcers (14.2%), and hypercholesterolemia (13.3%; Table 4). Seventeen patients developed pneumonia while they were receiving everolimus; for 3 of these patients, the everolimus treatment was discontinued, and for 2, the dose was reduced. The pneumonia was resolved in all but 2 cases. No cases of incisional hernia occurred. When adverse events were compared between centers, only the rates of edema (n = 39) and pneumonia (n = 17) differed significantly; they ranged from 2.3% to 29.2% (P = 0.003) and from 0% to 16.7% (P = 0.045), respectively. As for the 637 adverse events for which data regarding the impact on everolimus treatment were available, 12.9% (82/637) led to the discontinuation of everolimus, and 16.0% (102/637) were followed by an everolimus dose reduction. In all, 47 of 240 patients (19.6%) discontinued their treatment, and the dose of everolimus was reduced in 59 of 240 patients (24.5%).
Table 4. Adverse Events up to the Last Follow-Up Visit (n = 240)
Any adverse event
Biopsy-proven acute rejection
The mean white blood cell count decreased from day 0 (5.8 ± 2.8 × 103/mm3) to month 1 (5.0 ± 2.9 × 103/mm3, P < 0.001 versus day 0), from day 0 to month 3 (5.0 ± 2.2 × 103/mm3, P < 0.001), from day 0 to month 6 (5.1 ± 1.9 × 103/mm3, P < 0.001), and from day 0 to month 12 (5.6 ± 2.4 × 103/mm3, P = 0.033). At the baseline, 19 patients (9.1%) had a leukocyte count < 3000/mm3, whereas 20 patients (11.8%) at month 6 (P = not significant) and 13 of 141 patients (9.2%) at month 12 (P = not significant) did. The mean platelet counts were similar at months 1 (194 ± 96 × 103/mm3) and 12 (201 ± 93 × 103/mm3, P = 0.32). There were no statistically significant changes in liver enzymes from day 0 to month 12. The proportions of patients with diabetes on day 0 (98/216 or 45.4%) and at month 12 (62/146 or 42.5%) were similar; the insulin therapy requirements of patients with diabetes were also similar (46/95 or 48.4% on day 0 and 29/61 or 47.5% at month 12; data were not available for all patients). The mean total cholesterol levels were 4.7 ± 1.3 mmol/L on day 0 and 5.6 ± 1.8 mmol/L at month 12 (P < 0.001). The mean triglyceride levels were 1.9 ± 1.3 mmol/L on day 0 and 2.9 ± 2.4 mmol/L at month 12 (P = 0.007). The use of antihyperlipidemic treatment increased from 17.3% (37/214) to 39.7% (56/141, P < 0.003). The mean blood pressure was 141/79 mm Hg (median = 140/80 mm Hg) on day 0 and 142/81 mm Hg (median = 140/80 mm Hg) at month 12. Hypertension was diagnosed in 69.6% of the patients (149/214) at month 1 and in 76.7% (112/146) at month 12 (P = not significant). After the exclusion of all patients receiving steroids, there were still no significant differences in the rates of diabetes or hypertension. No significant difference was observed when the patients on steroids were removed from the analysis; this was probably related to the low doses of steroids that most of these patients received (5 mg/day) and the low number of patients at the 12-month visit.
This multicenter, retrospective analysis of the current use of everolimus in routine clinical practice indicates that the use of everolimus in maintenance liver transplant recipients allows the discontinuation of CNI and MPA therapy in a significant proportion of patients with a very low risk of rejection and with an acceptable safety profile. In particular, we were able to withdraw CNI therapy from approximately 60% of the patients by month 12. At approximately 7 to 8 ng/mL, the mean trough levels of everolimus were toward the higher end of the recommended target range (3-8 ng/mL), and they reflected the low use of concomitant immunosuppression.
The most frequent reason for the introduction of everolimus was chronic renal failure, a clinically important complication that affects 18% to 58% of liver transplant recipients within 5 years after transplantation.26, 27 When we analyzed the total population, we found that there was a small but significant improvement in the mean eGFR value (∼4 mL/minute or 6.5%) as early as month 1, and this persisted to 12 months after conversion. However, when we analyzed only those patients with chronic renal failure (in whom the greatest benefit might be expected), the increase in the mean eGFR value during the 12 months after the conversion to everolimus was >8 mL/minute (18.2%) in patients diagnosed with renal failure by investigators and >9 mL/minute (21.2%) in patients with serum creatinine levels ≥ 130 μmol/L at the baseline. These improvements can be regarded as clinically relevant.28 De Simone et al.17 reported a mean eGFR increase of 4 mL/minute in a prospective study of 40 maintenance liver transplant patients converted to everolimus with CNI withdrawal. In that study, 90% of the patients were converted because of poor kidney function; they had a mean baseline eGFR value similar to that of the current population and a slightly shorter mean time post-transplant (3.8 versus 4.9 years for the current cohort). A multivariate analysis showed that baseline creatinine clearance was the only clinical variable significantly correlating with the probability of renal function improvement 12 months after the conversion to everolimus.17 Another prospective study by De Simone et al.29 found no change in renal function 6 months after CNI withdrawal (or CNI reduction in 20% of the patients) with the introduction of everolimus, even though the time since transplantation was shorter than the time in the current study (∼3 years), most likely because of the poor baseline renal function (creatinine clearance = 51 mL/minute). Together, these studies suggest that liver transplant patients can receive a renal benefit from everolimus initiation with CNI withdrawal even at a relatively late stage after transplantation unless renal deterioration is profound. The remarkable improvement in eGFR (almost 30 mL/minute) reported by Masetti et al.16 after CNI withdrawal on day 30 in everolimus-treated patients versus CNI-treated patients highlights that early, preemptive conversion may be a more promising strategy.
The second major indication for the introduction of everolimus was the prevention or treatment of malignancies; this reflects current interest in the role of mTOR inhibitors in the management of posttransplant tumors.30-32 The mode of action of mTOR inhibitors33 suggests that they may be associated with a direct impact on malignant cells and may also exert an antiangiogenesis effect,34 as demonstrated in animal models.35-38 A phase 3 trial of everolimus in patients with advanced renal cell carcinoma39 and a series of phase 2 studies in patients with various other types of malignancies40-44 have consistently demonstrated an antitumor effect of everolimus. Although the data remain preliminary, a series of uncontrolled trials have shown markedly lower rates of HCC recurrence after liver transplantation in patients receiving sirolimus-based immunosuppression versus a standard CNI regimen.45-48 Chinnakotla et al.45 documented a 5-year survival rate of 80% in 121 patients receiving sirolimus after they underwent liver transplantation for HCC and a 5-year survival rate of 59% in 106 case controls receiving a CNI (P = 0.001). In a large registry analysis of 2491 patients who underwent transplantation for HCC, sirolimus-based maintenance therapy was also associated with improved survival and conferred a 15% 5-year survival advantage in comparison with CNI-based immunosuppression.49 Here, however, the retrospective nature of the data collection, the heterogeneity of the patients in terms of the type of cancer and the site of recurrence, the differences in the therapeutic strategies (including surgery and chemotherapy regimens) at the various participating centers, and the short duration of follow-up prohibit any conclusions regarding the impact of everolimus therapy on the progression of either HCC recurrence or de novo neoplasms.
Previous reports have described no rejection episodes after the introduction of everolimus,19-21 but in 2 trials in which CNI therapy and adjuvant immunosuppressants were withdrawn relatively aggressively, the rates of rejection were 7%18 and 10%.17 In the current study, in which there were no protocol-specified withdrawals of other agents, there was a very low rate of acute rejection (<2.0%).
Among the few hepatitis C virus–positive patients who underwent hepatitis C virus RNA monitoring after conversion, no significant changes in RNA levels were observed. However, we could not assess the true effect of everolimus on hepatitis C virus recurrence because of the small number of patients, the short duration of follow-up, and the absence of histological data.
The adverse events that were reported after everolimus conversion indicated no novel safety concerns. Total cholesterol and triglyceride levels increased significantly despite the greater use of antihyperlipidemic therapy. Although the reduction in the total leukocyte count was significant, it was not clinically relevant. CNI discontinuation in the majority of the patients did not lead to a reduction in the proportion of patients with diabetes, and there was no evidence for lessened severity of diabetes, as indicated by the use of insulin. Similarly, the blood pressure did not change over the 12 months after conversion. The lack of benefit with respect to diabetes and hypertension was maintained after the exclusion of steroid-treated patients; this was not unexpected because most patients on steroid therapy were receiving a low dose (5 mg/day), and only a small proportion of patients were receiving steroids at the 12-month visit. The proportion of patients who discontinued everolimus therapy because of adverse events was similar to that reported in a recent clinical trial of maintenance liver transplant recipients who were converted from a CNI to everolimus.17 No significant correlation was found between the types of adverse events and the trough level of everolimus.
This retrospective series represents the largest cohort of maintenance liver transplant patients analyzed after everolimus conversion and reflects current routine practice for the use of the drug in this setting. More than 60% of the patients were kept CNI-free 1 year after the introduction of everolimus, and they had a very low risk of acute rejection and an acceptable safety profile. Notwithstanding the limitations of the retrospective study design, these results suggest that everolimus therapy may preserve kidney function in patients on CNI therapy who experience chronic kidney dysfunction after liver transplantation. Randomized trials in which maintenance liver transplant patients are switched to everolimus in response to clinical indications are awaited; further explorations of preemptive conversion early after transplantation are also needed. Specifically, the position of everolimus in the management of patients with HCC recurrence and de novo malignancies requires further investigation and is being evaluated in an ongoing, pivotal randomized trial of de novo liver transplant recipients.
The authors are grateful to Marie Kouassi for data collection, to David Delvart for data management, and to Valérie Delvart for statistical analysis; they are also grateful to the investigators Teresa Antonini, Olivier Boillot, and Christophe Duvoux.