Efficacy and safety of early tacrolimus conversion to sirolimus after kidney transplantation: Long-term results of a prospective randomized study

Introduction

Although the use of cyclosporine (CsA) and tacrolimus (Tac) has markedly improved 1-year kidney transplant survival and decreased the acute rejection rate in most centers, both of these calcineurin inhibitors (CNIs) are nephrotoxic, and it is well recognized that their long-term use does not solve the problem of chronic rejection. The use of CNIs, CsA, and Tac has been associated with both acute and chronic nephrotoxicity and contribute to the development of chronic allograft nephropathy (CAN).[12]

Biopsy-proven CAN was found in 62% of the Tac- and 72% of the CsA-treated patients within 2 years posttransplantation. Notably, 68% of the biopsy specimens with CAN were obtained from patients who developed CNI nephrotoxicity during the 1^st year posttransplantation.[1] In the short-term, CNI produces renal arteriolar vasoconstriction and a decrease in glomerular filtration rate (GFR) that is dose-related and reversible.[23456] Long-term exposure to CNI causes chronic nonreversible changes that are characterized by interstitial fibrosis and obliterative arteriolar changes due to fibrous intimal thickening.[7] Nankivell et al.[8] showed a histological evidence of CNI toxicity in all renal allografts within 10 years by adopting annual protocol biopsies. Such a long-term attrition effect of CNI is thought to contribute to the pathogenesis of chronic allograft damage despite the serum level being maintained within the therapeutic range. In order to avoid or even to ameliorate this effect, a variety of strategies has been explored. These include complete withdrawal of CNI at some point in the posttransplant period, substitution of CNI with mycophenolate mofetil (MMF), or simply minimizing the CsA maintenance dose.[910111213] Although some success has been achieved with these strategies, withdrawal of CNI as late as 1 year posttransplant is often associated with acute rejection and the risk of late graft damage.[14]

Sirolimus (SRL) has been shown to be effective as a de novo therapy after renal transplantation[15] and as long-term maintenance therapy with steroids.[161718] It may also have a role as an effective substitute for CNI therapy late after transplantation to avoid further CNI nephrotoxicity.[192021222324] However, the potential risk and benefit of this conversion strategy are not yet fully known, especially in the long-term.

The aim of this study was to evaluate the safety and efficacy of conversion to SRL-based immunosuppression in stable kidney transplant recipients 6 months posttransplant.

Patients and Methods

Patients

Patients selected for the study were low-risk kidney transplants between January 2005 and October 2009 receiving Tac-based maintenance treatment and followed-up at our center. They were invited to participate in the study after signing an informed consent. The characteristics of the patients are outlined in Table 1. Inclusion criteria were: first transplant patients aged >21 years old; serum creatinine levels < 1.5 mg/dl; no past history of acute antibody mediated rejection or recent acute cellular rejection 3 months before randomization; unsensitized patients; and had baseline total serum cholesterol < 200 mg/dl, triglycerides < 160 mg/dl, total white blood cell (WBCs) count of more than 3000; platelet counts of more than 100,000; and/or willingness to participate in the study.

Table 1 Baseline demographics and clinical characteristics

Study design

This was a randomized, parallel-group, prospective study comparing continued triple therapy with Tac (Prograf, Fujisawa Healthcare, Al Hekma Inc. Amman, Jordan), corticosteroids and MMF (Tac group; control), with withdrawal of Tac, and addition of SRL (Rapamune, Wyeth-Ayerst Philadelphia, USA) (SRL group). The 6-month time point was chosen to minimize the risk of early acute rejection. Patients were randomly assigned to one of the two treatment groups (1:1) using a computer generated sequence after obtaining informed, written consent for participation in the study [Figure 1]. The study was undertaken in accordance with the Declaration of Helsinki and all subsequent amendments and was approved by the Local Ethics Committees.

Close

Figure 1

Study flow diagram. The diagram illustrates the study enrollment and disposition of the trial participants

Export to PPT

Results

Patient demographics

Entry and participant flow through the study are shown in Figure 1. A total of 124 patients were recruited, of whom 38 proceeded beyond randomization. Forty-one patients refused to participate, and 16 were excluded for reasons other than those specified in the inclusion criteria, including relocation,[4] undergoing investigation for co-morbidity,[8] and delayed decision until after study closure.[4] Baseline demographic, transplant-related, and clinical data of the randomized patients are presented in Tables 1 and 2. No differences in either demographic or transplant data were observed between the two groups. No between-group differences were found as regards clinical and laboratory parameters as shown in Table 2. All patients in both groups completed the 3-year follow-up and were included in the final intention-to-treat analysis.

Table 2 Baseline biochemical and clinical values in both groups

Immunosuppressive regimen

At 1-year postrandomization, 26 (89.7%) patients in the SRL group and 28 (96.6%) patients in the Tac group were still receiving the initially allocated study drug. Afterward, SRL was discontinued in further 3 (10.3%) patients and Tac in further 2 (6.9%) patients. In the two groups, the primary reason for discontinuing Tac or SRL was the occurrence of adverse events. The majority of adverse events occurred in the SRL group during the 1^st year.

All patients in the two groups were treated with MMF and corticosteroids at 3 years posttransplant with no between-group differences as regards their doses. At 1 year after conversion, the mean Tac trough level was 4.7 ± 1.5 ng/ml, while the mean SRL trough level was 4.5 ± 1.3 ng/ml. At 36 months, there was no significant difference in mean MMF (1.6 ± 0.4 vs. 1.6 ± 0.3 g/day) and steroid doses (5.0 ± 0.5 vs. 5.0 ± 0.3 mg/day) in the SRL and Tac groups, respectively. Moreover, the mean therapeutic level of SRL (7.0 ± 2.3 ng/ml) in SRL group and mean Tac level (7.2 ± 2.1 ng/ml) in Tac group was in the therapeutic window.

Acute rejection and immunological complication

During the follow-up, 4 (13.8%) patients in the SRL group and 3 (10.3%) in the Tac group (P = 0.5) developed BPAR. The median time to the first acute rejection was 20.3 months in the SRL group and 23.1 months in the Tac group (P = 0.19), and one patient in the Tac group developed two acute rejection episodes. In SRL group, one developed chronic antibody-mediated rejection with no response to therapeutic treatment including methyl prednisolone pulses in addition to IV immunoglobulin plus plasmapheresis session followed by four doses of rituximab therapy over 4 weeks and the patient returned to dialysis.

Patient and graft survival

Patient survival at 3-years was 93.1% in the SRL group and 100% in the Tac group (P = 0.32) [Figures 2 and 3]. Two patients died between 1 and 3 years in the SRL group, death was due to a brain tumor (n = 1) and cerebral stroke (n = 1). Censored death graft survival at 3-years was 89.7% in the SRL group and 100% in the Tac group (P = 0.11). Three graft losses in the SRL were reported between 1 and 3 years due to death with functioning graft (n = 2) and chronic antibody-mediated graft rejection (n = 1).

Close

Figure 2

Kaplan–Meier graft survival curve in both groups

Export to PPT

Close

Figure 3

Kaplan–Meier patient survival curve in both groups

Export to PPT

Renal outcome

During the 3-year observation period, GFR decreased in the Tac group (from 73.2 ± 8.3 to 71.3 ± 13.6 ml/min 1.73 m², P = 0.07) and increased in the SRL group (from 72.8 ± 7.1 to 74.2 ± 7.5 ml/min 1.73 m², P = 0.06); comparison analysis showed a significantly different 3-year changes in GFR in the two groups [Table 3]. Urinary protein excretion rate increased in the SRL group (from 0.1 ± 0.1 to 0.7 ± 0.5 g/24 h, P = 0.001), whereas nonsignificant changes were observed in the Tac group (from 0.1 ± 0.1 to 0.2 ± 0.3 g/24 h, P = 0.1). In the SRL group, 6 (23%) patients developed de novo proteinuria after conversion with a median value 1.2 g/day, 2 (7.6%) of whom were at a nephrotic level, therefore, SRL was stopped and Tac was reintroduced in these two patients. In the Tac group, 2 (7.4%) patients developed de novo proteinuria after conversion with a median value 0.4 g/day, none of them were at a nephrotic level.

Table 3 1-year and 3-year changes in laboratory parameters after randomization in both groups

Blood pressure

At the start of the study, the mean systolic blood pressure (128 ± 16 mmHg and 129 ± 17 mmHg; P = 0.2) and diastolic blood pressure (77 ± 13 mmHg and 81 ± 13 mmHg; P = 0.3) in the SRL and Tac groups, respectively, was not significantly different. Most of the patients in both groups were controlled with one to two drugs (20 patients, and 22 patients in Tac and SRL groups, respectively) with a mean number of drugs, 1.5 ± 0.7 and 1.4 ± 0.7 in Tac and SRL groups, respectively. At 12 month, the mean systolic blood pressure (130 ± 15 mmHg and 138 ± 19 mmHg; P = 0.03) and diastolic blood pressure (73 ± 11 mmHg and 80 ± 12 mmHg; P = 0.04) in the SRL and Tac groups, respectively. At 36 month, the mean systolic blood pressure (132 ± 19 and 141 ± 21 mmHg; P = 0.001) and diastolic blood pressure (74 ± 14 mmHg and 82 ± 10 mmHg; P = 0.001) in the SRL and Tac groups, respectively. At 36 month, 44% and 30% of the patients were not taking anti-hypertensive medications in the SRL group and the Tac group, respectively, (P = 0.02 [Table 4] and 17 patients in Tac group versus 5 patients in SRL group were controlled more than two drugs with a mean number of drugs, 2.3 ± 0.5 and 1.1 ± 0.6 in Tac and SRL groups, respectively).

Table 4 Principal side effects following randomization

Hyperlipidemia

Before randomization, the blood cholesterol (224 ± 38 and 220 ± 23 mg/dl; P = 0.1) and triglycerides (141.6 ± 70.8 mg/dl and 150.4 ± 70.8 mg/dl; P = 0.5) were not significantly different in the SRL and Tac groups, respectively. At 12 months, the mean serum cholesterol level was 224 ± 38 mg/dl versus 181 ± 19 mg/dl (P = 0.02), and at 36 months, 226 mg/dl versus 192 mg/dl (P = 0.01) in the SRL and the Tac group, respectively. At 12 months, the mean serum triglyceride level was 204 ± 53 mg/dl versus to 159 ± 53 mg/dl (P = 0.01) and at 36 months, 221 ± 53 mg/dl versus 142 ± 44 mg/dl (P = 0.01), in the SRL and the Tac group, respectively. At 36 months, the mean serum cholesterol (236 ± 19 mg/dl) and triglyceride (177 ± 27 mg/dl) levels increased significantly in the SRL group, compared to TAC group (212 ± 27 mg/dl) and (142 ± 27 mg/dl) (P < 0.05) and 24% of the patients in the Tac group, and 58% of the patients in the SRL group were taking lipid-lowering medications (P < 0.0001) [Tables 3 and 4].

Hematology

Mean hemoglobin concentrations were similar between groups at baseline [Table 2] significantly different from 12 months (11.9 and 12.7 g/dl, P = 0.02) to 36 months (11.4 and 12.3 g%, P = 0.01), in the SRL and Tac groups, respectively. Erythropoietic drug use was similar at baseline 4.7% and 4.8%, increased to 12.7% versus 5.6% in the SRL and Tac groups, respectively, at 36 months. The total leucocyte count fell significantly at 12 months in patients on SRL (P = 0.04) and at 36 months (P = 0.01). There was no significant difference in the change of platelet count between groups at either time point.

Other significant events

There were 11 cases of posttransplant diabetes mellitus in the Tac group (37.9%) and 10 in the SRL group (34.5%) (P = 0.3). There was one case of polyomavirus infection in the Tac group. One case of posttransplant malignancy (brain tumor) was diagnosed in the SRL group. One case of systemic cytomegalovirus infection in the SRL group and one case of H1N1 virus infection in the Tac group were diagnosed during the study period. Other cases which required admissions were two cases of pneumonia in each group. Other adverse events were shown in Table 4.

Discontinuation from the study

SRL was discontinued in six patients in SRL group (20.7%; pregnancy, acute humoral rejection, and heavy proteinuria) versus in three patients in Tac group (10.3%; chronic interstitial fibrosis in two and polyoma virus nephropathy in one; P = 0.2).

Discussion

Prolonging renal allograft survival remains one of the most important challenges in kidney transplantation. Indeed, long-term kidney transplant survival rates have not kept pace with the striking improvements achieved in short-term outcomes.[25] A major cause of long-term allograft injury, fibrosis and functional decline, is CNI toxicity.[20] This study examines the safety and efficacy of converting patients with a stable renal function from Tac-based regimen to an SRL-based regimen after kidney transplantation.

SRL-based immunosuppressive therapy is thought to be less nephrotoxic and continues to be evaluated as a CNI sparing therapy. In the de novo setting, SRL therapy combined with mycophenolic acid has not provided superior outcomes or adequate protection from acute rejection.[18] However, conversion from CNI to SRL therapy after transplantation improved short-term renal function but did not decrease allograft fibrosis.[2627] Moreover, in conversion trials, SRL-treated patients typically experienced higher rejection rates and adverse events, further confounding the results.[2829]

In our study, we found the incidence of BPAR was 13.8% in SRL group and 10.3% in Tac group patients (P = 0.5). Most of the patients in both groups experienced one rejection episode. However, in SRL group, one developed chronic antibody-mediated rejection with no response to therapeutic modalities and patient returned back to dialysis. A previous meta-analysis of mammalian target of rapamycin inhibitor (mTORi) use in kidney transplant recipients had been shown no difference in acute rejection and superior graft function when SRL are used as CNI replacement.[30] In a study,[24] on conversion of CsA to everolimus at 4.5 months posttransplant found that the incidence of BPAR from randomization to month 36 was significantly higher in the everolimus group 13% versus 4.8% in the CsA arm, P = 0.015. On the other hand, Lebranchu et al.[31] although they noticed a similar pattern of BPAR after SRL conversion, however, two graft losses due to acute rejection were observed in the SRL group during the 1^st year of the follow-up period. However, although, they found no difference in the incidence of de novo or persistent anti-human leukocyte antigen antibodies between the groups; 12.3% of patients in the SRL group and 21.1% in the CsA group, they concluded that SRL may expose to the risk of graft loss due to resistant acute rejection.

Our results show that the rates of graft loss (including death with a functioning graft) were significantly higher in the SRL group, despite improved graft function in those surviving with functioning grafts. Two patients died between 1 and 3 years in the SRL group; death was due to brain tumor and cerebral stroke. Results from large registry database analyses[18] suggest that Tac/SRL- or SRL/MMF-based immunosuppression may be inferior to Tac/MMF-based immunosuppression in long-term graft survival. In symphony study,[14] it was found that allograft survival differed significantly among the four groups (P = 0.02) and was highest in the low-dose Tac group (94.2%), followed by the low-dose CsA group (93.1%), the standard-dose CsA group (89.3%), and the low-dose SRL group (89.3%). In a recent large UNOS-based observational study of 139,370 kidney transplant patients, de novo use of mTORi was associated with increased allograft loss and mortality throughout 8 years of longitudinal follow-up.[32] The higher incidence of allograft loss may be due to the increased rate of acute rejection seen with mTORi and the known association of acute rejection with allograft loss. Interestingly, in this study, that mortality did not correlate with acute rejection suggesting a mechanism independent of effect on allograft function. This phenomenon was also observed in a prior observational investigation of Hungarian allograft recipients, in which mTORi were associated with increased mortality, but not worse allograft outcomes.[33] Further studies are needed to firmly establish this association and the responsible mechanism for high mortality among these patients. However, in spare the nephron study,[12] they found better graft and patient survival after 2 years of follow-up of their patients and explained this for lower incidence of acute rejection (9.5% and 11.3% in MMF/SRL group and MMF/CNI group, respectively), and fewer deaths in the SRL group. Compared to our results, it may be explained by the shorter period of follow-up, the higher number of patients allocated and the multicenter nature of the study design.

Estimation of renal allograft function by calculated GFR revealed a better renal function in group SRL patients as compared to Tac patients at most time points. This finding came in accordance with what had been previously reported by Lebranchu et al.[31] and Weir et al.[12] that CNI-free regimens based on SRL have better renal function than CNI-based regimens. In the Spare the Nephron study,[12] they found at 1 year, the mean percentage change in directly measured GFR was greater in the SRL arm (24.4 vs. 5.2%; P = 0.054), but this benefit was no longer evident at 2 years. The evidence suggests that in patients without markedly compromised kidney function or proteinuria, conversion from a CNI to mTORi may preserve GFR, but offers no definitive benefit on the hard outcomes of mortality or allograft loss.

In our study, as observed at the end of the follow-up observational period, urinary protein was higher in the SRL group. Severe proteinuria was not frequent (7.6%) but may have resulted in the reconversion to CNI in these cases. Thus, the small increase in protein excretion was considered clinically acceptable, especially in the presence of the improved GFR values in the SRL group. However, in view of indications that early low-grade proteinuria may predict subsequent graft loss,[34] urinary protein excretion should be carefully monitoring during the long-term follow-up of these patients.

The decrease in blood pressure in the SRL group versus Tac controls has potentially more advantages on patient and graft survival. The avoidance of CNI with the use of SRL may allow more effective long-term blood pressure control.[35] The profile of adverse events reported here in the SRL group was as expected, including the effect on lipid profile, diabetes mellitus, and hematological values. Cholesterol levels requiring statin therapy were significantly higher in SRL group, as shown in the present trial. SRL has been observed to elevate blood lipids in almost all clinical trials,[15] and the dyslipidemic effect appears to be dose-related. Hyperglycemia has been reported in more than a third of patients on SRL and Tac. CNI therapy is well-known to increase the rate of new-onset diabetes after kidney transplantation, and in few small, single-center studies have suggested that SRL may have a diabetogenic effect. In one retrospective analysis, the authors showed that SRL is associated with a similar risk of diabetes compared to Tac.[36] There was a significant increased incidence of anemia and the percentage of patients receiving erythropoietin in the SRL groups. Augustine et al.[37] reported a prevalence of anemia of 31% in patients on Tac-based compared with 57% on SRL-based therapy in renal transplant patients at 1 year post-transplantation.

Some limitations must be acknowledged in our study. First of all, it was a single-center study with a small sample size (n = 58) and also the relatively short observation period which may lighten the sound of conclusions. The fundamental rationale for SRL conversion is the potential for improved long-term outcome, measured over decades, with respect to proteinuria, renal function, and graft survival. Second, some methodological factors should be taken into account in the interpretation of this study including statistical power which is needed to estimate the total sample size needed based on the expected outcome in each group. Moreover, we did not include protocol biopsy to assess chronic allograft damage index in our study which may have some benefits in diagnosing histopathological changes in grafts over time.

Conclusion

Our experience demonstrates that conversion to SRL from CNI-based therapy may result in better renal function and blood pressure control in renal transplant recipients without an increased risk of acute rejection but was associated with higher discontinuation rate attributable to adverse events. However, these benefits have not resulted in a growing advantage in graft or patient survival.