Kidney transplantation is the treatment of choice for patients with end-stage kidney disease (ESKD), offering superior survival and quality of life compared with dialysis.1-3 Current standard-of-care immunosuppression, containing a calcineurin inhibitor (CNI), mycophenolate and corticosteroids, has enabled excellent short- and medium-term patient and graft survivals and a low incidence of acute rejection.4 Despite these improvements, rates of mortality and graft loss beyond the first year have not significantly changed.2,4
Challenges to long-term graft survival include chronic allograft injury and death with a functioning graft, principally due to cancer and cardiovascular disease.5 Calcineurin inhibitors are thought to contribute to the development of these outcomes. Calcineurin inhibitors are well established as being nephrotoxic,6,7 and are also associated with unfavorable side effects, including hypertension, new-onset diabetes mellitus after transplantation, and dyslipidemia,8 all of which are all potent risk factors for cardiovascular disease. Attempts to reduce overall CNI exposure in kidney transplant recipients has become the focus of many immunosuppression trials. A key component of such attempts has been the use of mammalian target of rapamycin inhibitors (mTORI) to enable either CNI elimination or CNI minimization. Randomized controlled trials (RCTs) examining conversion from CNI to mTORIs early after transplantation have shown comparable graft function to CNI-based controls with up to 5 years of follow-up9-14; however, the long-term efficacy and safety of mTORI-based regimens is unclear. RCTs in kidney transplantation have been limited by short durations of follow-up and consequent reliance on intermediate endpoints, such as acute rejection and kidney function, rather than patient-relevant and important outcomes including patient and graft survival.15 Linkage of trial patient data to established national registries may provide a more efficient and powerful tool to capture long-term outcomes after trial follow-up has ended.16 The Australian and New Zealand Dialysis and Transplantation (ANZDATA) Registry prospectively captures clinical outcomes for all patients with ESKD requiring dialysis or transplantation. By linking individual participant data to the registry, we aimed to compare the long-term outcomes of graft and recipient survival, graft function and death-censored graft loss in all Australian and New Zealand trial participants randomized to an everolimus-containing regimen, compared with controls on standard CNI-based triple therapy.
MATERIALS AND METHODS
We included all randomized, multicenter, prospective studies with Australian and New Zealand participants that compared standard triple therapy (cyclosporine or tacrolimus, mycophenolic acid, and corticosteroids) with an everolimus-containing regimen (Figure 1). Ethical approval for the study was provided separately for each trial and collectively for the follow-up data under the ANZDATA Registry patient consent, which permits use of de-identified data for purposes including research.
Data Collection and Items
The study sponsor provided the following identifiers to allow deterministic linkage with ANZDATA: date of birth, sex, date of transplantation, date of randomization and transplant center. Australian and New Zealand Dialysis and Transplantation is a government-funded registry, which prospectively collects and reports the outcomes of dialysis and kidney transplantation for patients with ESKD in Australia and New Zealand. From the registry, we extracted data on recipient factors including age at transplantation, ethnicity, cause of ESKD, prior allografts, smoking history, time-on-dialysis and comorbidities present at time of transplantation (coronary artery disease, peripheral vascular disease, diabetes mellitus, cerebrovascular disease, and chronic lung disease); and transplant factors of donor type, donor demographics (age, sex, cause of death), ischemia time, HLA mismatch, maximum cytotoxic antibodies and induction therapy.
Risk of Bias Assessment in Individual Studies
All studies were assessed for risk of bias using the Cochrane Risk of Bias assessment tool17 by 2 authors (T.Y. and S.J.C. or W.L.) independently. Disagreements were resolved by discussion. The tool assesses; random sequence generation, allocation concealment, blinding of participants and study personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other threats to validity, such as the role of industry sponsorship and registration of studies in clinical trials registry (for trials published after 2005).
Primary outcomes of this study were all-cause graft loss, mortality and death-censored graft loss. Australian and New Zealand Dialysis and Transplantation records the date and causes of death (broadly categorized into cardiac, vascular, cancer, infection, social, and miscellaneous) and cause-specific graft loss (chronic allograft nephropathy, recurrent glomerulonephritis, noncompliance, etc.) in all recipients. The date of death-censored graft loss is defined as the date of return to dialysis.
The secondary outcome of interest was graft function. Australian and New Zealand Dialysis and Transplantation records the SCr at prespecified time points after transplantation (mo 1, 2, 3, and 6 and y 1, 2, 3, 5, 7, and 10). One trial (ASCERTAIN n = 70) was excluded from the graft function analysis as trial participants were randomized to everolimus >5 years after transplantation, thus providing insufficient estimated glomerular filtrate rate (eGFR) measures on follow-up. We converted SCr to the eGFR, using the CKD-EPI creatinine equation.18
All primary analyses were based on intention-to-treat. We considered all P values <0.05 as statistically significant. Baseline participant characteristics were reported in numbers and proportions (%) and mean (standard deviation [SD]) for normally distributed data or median (interquartile range) for skewed data. Comparison between 2 categorical outcomes were tested using the Pearson χ2 test.
The median follow-up time was calculated using the reverse Kaplan-Meier method.19 For the main outcomes of all-cause graft loss, mortality and death-censored graft loss, we analyzed all patients using a 1-step meta-analysis approach and calculated a hazard ratio (HR) for each outcome using a Cox proportional hazards model with random effects (shared frailty). For the outcomes of all-cause graft loss and mortality, survival time was calculated from the date of randomization to the date-of-graft loss, date-of-death or date-of-last follow-up. For death-censored graft loss, we censored patients on the date-of-death or date-of-last follow-up. We included treatment group, age, sex, and smoking in all models a priori and screened out any explanatory variables with P <0.1 in the univariable stage. We used backward elimination to derive the final model. We explored effect modification between the treatment arms and other covariates in the final model, which were considered significant at a level of P <0.01.
For the outcome of graft function, we compared eGFR between treatment groups using a multilevel mixed effects (random coefficient) model with an unstructured covariance matrix. Each repeated measure of eGFR (level 1) was clustered within a patient (level 2), and each patient was clustered within a trial (level 3). The model was adjusted for the person-effect (random-intercept), and each person had a random slope for time.
To explore treatment heterogeneity, we performed subgroup analyses by treatment regimen; everolimus with reduced-dose CNI or CNI withdrawal versus control; and by timing of mTORI initiation; early everolimus initiation (de novo or conversion to everolimus within 6 months) or late-conversion (>6 months) to everolimus versus control. We also performed an “on-treatment analysis” for the patients who remained on their assigned treatment groups at 2 years after randomization. The ANZDATA registry recorded medication details at prespecified timepoints after transplantation, thus we considered participants to have remained “on-treatment” if the immunosuppression drugs documented by ANZDATA at 2 years posttransplant were the same as the original group assignment. We performed sensitivity analyses to examine the effects of different model assumptions including: (1) repeating the primary analysis by performing a fixed-effects (stratified) Cox regression model; (2) performing a “delayed-entry” stratified Cox model, to account for participants in the late conversion trial, who were randomized on average 5 years posttransplantation (vs at the time of transplantation in de novo and early conversion trials); and (3) repeating the graft function analysis, by imputing an eGFR of 0 at the time of graft failure. All analyses were performed on Stata, version 14.1 (StataCorp LP, College Station TX).
Study and Participant Characteristics
Trial and participant characteristics are presented in Tables 1 and 2. Five studies with 351 patients were included. Four RCTs (n = 303) compared an everolimus-based regimen with CNI-based triple therapy and 1 study (n = 49) randomized participants to either 1.5 mg/d or 3 mg/d of everolimus. Two patients were excluded as they could not be linked to the ANZDATA registry, yielding a total of 349 patients for inclusion in the study (Figure 1). Of these, 242 were randomized to everolimus and 107 were randomized to control (Figure 1, Table 1). All controls received a CNI (cyclosporine or tacrolimus), mycophenolate or azathioprine and corticosteroids. The baseline characteristics of all patients stratified by randomization are presented in Table 2. The median age was 47 years (SD, 12.4 y) in the everolimus arm compared with 50 years (SD, 12.1 y) in controls. Trial participants were mostly male (66% everolimus vs 69% control) and white (87% vs 84%), with glomerulonephritis as the most frequent cause of ESKD in both arms (50% vs 49%). Half of the patients had received a deceased donor kidney transplant (55% vs 52%), with most of the transplants occurring between 2000 and 2008.
Risk of Bias
The overall risk of bias was low for 3 studies,9,12,20 and unclear for 2 studies.21,22 All 5 trials were industry-sponsored; however, most studies did not provide information on the level of involvement by the sponsor in the design, conduct, analysis, or article preparation. For trials published after 2005, 2 studies provided information on trial registration12,22 and 1 study did not.9
A summary risk of bias across all included studies is provided in Figure 2. The risk of bias for random sequence generation and allocation concealment was low or unclear. Four studies9,12,20,22 described a robust method of random sequence generation, and 3 studies9,12,20 provided adequate information on allocation concealment. The remaining studies were rated unclear as they did not describe these methods in the article. None of the studies were blinded to either the participant or trial personnel, thus all were judged to have a high risk of performance bias. Biopsy-proven acute rejection was the key primary outcome for all trials, but only 1 trial described blinding of the outcome assessor,12 thus the risk of detection bias was judged to be unclear for the remaining 4 trials. There was a high risk of attrition bias for 1 study,22 due to a high rate of trial discontinuation for the study drug. The risk of selective reporting bias was considered low for 3 trials,9,12,22 where the study protocols were available, and there were no departures from the prespecific outcomes, and unclear for 2 trials,20,21 where no study protocol was available.
All-cause Graft Loss
Patients were followed for a median of 8.8 years (range, 0.02–17.3 years) after randomization, representing 2822 patient-y of follow-up. Five patients (3 everolimus, 2 control) were lost to follow-up. Ninety grafts were lost during follow-up, 68 (28%) in everolimus and 22 (22%) in controls. There was no significant difference in graft survival between groups (Figure 3). The unadjusted HR for everolimus versus control was 1.24 (95% confidence interval [CI], 0.75-2.04; P = 0.4) and after adjusting for age, male sex, smoking, and years on dialysis, the HR was 1.16 (95% CI, 0.69-1.94; P = 0.6) (Table 3).
A total of 54 patients died, 41 (17%) in the everolimus arm versus 13 (12%) in controls. A higher proportion of patients in the everolimus arm died due to cardiovascular disease compared with controls (5.4% vs 0.9%, P = 0.04), whereas death due to cancer, infection and other causes were similar between the groups (Table S1, SDC, http://links.lww.com/TP/B648). There were no associations between treatment groups and the risk of mortality in the unadjusted and adjusted Cox regression models (Figure 4). Compared with control, the unadjusted HR for everolimus was 1.19 (95% CI, 0.63-2.24); P = 0.6. After adjusting for age, male sex, smoking and years on dialysis, the HR for everolimus was 1.51 (95% CI, 0.78-2.93; P = 0.2) compared with control (Table 3).
Death-censored Graft Loss
Fifty-four patients returned to dialysis during the follow-up period; 39 (16%) in everolimus and 12 (11%) in controls. Chronic allograft nephropathy was the most common cause of graft loss in everolimus patients and controls (72% vs 75%, respectively). There was no association between treatment groups and the risk of death-censored graft loss (Figure 5). Compared with control, the unadjusted HR for everolimus was 1.34 (95% CI, 0.68-2.63, P = 0.4). After adjusting for age, male sex, smoking, and years on dialysis, the HR for everolimus was 1.00 (0.50-2.01, P = 1.0) compared with control (Table 3).
For 4 trials (n = 279) where everolimus was initiated either de novo or within 6 months of transplantation, repeated SCr measures were well reported in the registry (0.7% missing). There was no difference in graft function between the groups out to 10 years. Compared with control, the mean eGFR in the everolimus group was 2.4 mL/min per 1.73m2 lower (95% CI, −6.5 to +1.8, P = 0.3). The decline in the eGFR did not significantly differ between the groups (mean difference in decline rate, 0.01 mL/min per 1.73 m2; 95% CI, −0.06 to +0.09; P = 0.7) (Figure 6).
The Effect of Treatment Regimen (Everolimus With Reduced-dose CNI or CNI Withdrawal Versus Control) on Graft Loss, Mortality, and Death-censored Graft Loss
Patients on the everolimus arm were further stratified into different treatment regimens; everolimus with reduced-dose (RD)-CNI (n = 156) or everolimus with CNI withdrawal (n = 86) (Table 1). Neither treatment regimen was significantly associated with graft loss, mortality or death-censored graft loss versus controls (Table S2, SDC, http://links.lww.com/TP/B648). For graft loss, the adjusted HRs for RD-CNI were 1.26 (95% CI, 0.73-2.16; P = 0.4) and 0.96 (95% CI, 0.48-1.92, P = 0.9) for CNI withdrawal. For death, the adjusted HRs for RD-CNI were 1.37 (95% CI, 0.69-2.72, P = 0.4) and 1.20 (95% CI, 0.52 - 2.79, P = 0.7) for CNI withdrawal. With respect to death-censored graft loss, the adjusted HRs for RD-CNI were 1.16 (95% CI, 0.56-2.39; P = 0.7) and 0.63 (95% CI, 0.21-1.81, P = 0.4) for CNI withdrawal.
The Effect of Early Versus Late Everolimus Initiation on Graft Loss, Mortality, and Death-censored Graft Loss
There were no significant differences in the risk of graft loss, mortality or death-censored graft loss in patients on early (de novo or <6 months conversion) (n = 270) or late everolimus initiation trials versus controls (n = 70) (Table S3, SDC, http://links.lww.com/TP/B648). Patients in the early everolimus initiation trials9,12,20,21 did not have an increased risk of graft loss (adjust HR, 1.42; 95% CI, 0.74-2.73; P = 0.3), mortality (adjusted HR, 1.41; 95% CI, 0.63-3.12; P = 0.4) or death-censored graft loss (adjusted HR, 1.43; 95% CI, 0.58-3.52; P = 0.4) versus controls. In the late conversion trial,22 patients who converted to everolimus >6 months after transplantation did not incur an increased risk of graft loss (adjusted HR, 0.73; 95% CI, 0.29-1.83), death (adjusted HR, 1.21; 95% CI, 0.37-3.94; P = 0.8), or death-censored graft loss (adjusted HR, 0.95; 95% CI, 0.32-2.8; P = 0.9) compared with controls.
Of the 279 patients in the early everolimus initiation trials, there were missing medication details for 32 patients (31 everolimus, 1 control) from ANZDATA. At 1-year posttransplantation, 117 (70.5%) remained on everolimus compared with 82 (95%) in control. At 2 years posttransplantation, 95 (59%) remained on everolimus versus 77 (90%) in control. There were no significant differences in the risk of graft loss (adjusted HR, 0.86; 95% CI, 0.34-2.16; P = 0.7), mortality (adjusted HR, 1.02; 95% CI, 0.35-3.02; P = 1.0), or death-censored graft loss (HR, 1.06; 95% CI, 0.31-3.67; P = 0.9) (Table S4, SDC, http://links.lww.com/TP/B648). For the analysis of graft function, patients who remained on-treatment at 2 years (n = 172) showed no significant differences in the mean eGFR (mean difference, 0.61 mL/min per 1.73 m2; 95% CI, −4.08 to +5.30; P = 0.8) or slope of the eGFR (0.02 mL/min per 1.72 m2; 95% CI, −0.07 to +0.11; P = 0.6).
We repeated the primary analyses using a stratified Cox (fixed-effects) model. The treatment estimates did not substantially alter in direction or magnitude of the effects (Table S5, SDC, http://links.lww.com/TP/B648). There was no association between everolimus with graft loss (adjusted HR, 1.15; 95% CI, 0.68-1.97; P = 0.6)), mortality (adjusted HR, 1.60; 95% CI, 0.80-3.2; P = 0.2), or death-censored graft loss (adjusted HR, 0.93; 95% CI, 0.45-1.93; P = 0.9) compared with controls. As participants in 1 study (ASCERTAIN, n = 70) were randomized on average >5 years posttransplantation, we repeated the stratified Cox model using a delayed-entry (or left-truncation) analysis to take into account the effect of “late-entry” of these patients. The treatment estimates were similar to the stratified Cox model (Table S6, SDC, http://links.lww.com/TP/B648). There was no association between everolimus with all-cause graft loss (adjusted HR, 1.15; 95% CI, 0.67-1.96; P = 0.6), mortality (adjusted HR, 1.34; 95% CI, 0.67-2.65; P = 0.4) or death-censored graft loss (adjusted HR, 1.01; 95% CI, 0.49-2.11; P = 0.9) compared with control.
For the outcome of graft function, imputing eGFR of 0 at time of graft loss did not significantly alter the slopes between the groups (mean difference in decline rate, 0.01 mL/min per 1.73 m2; 95% CI, −0.05 to +0.08; P = 0.7).
Our study found that after a decade of follow-up, there were no observed differences in the risk of all-cause graft loss, mortality or death-censored graft loss in Australian and New Zealand patients randomized to everolimus-based immunosuppression versus standard of care. However, discontinuation of everolimus was high within the first 1 to 2 years after transplantation. Stratification by treatment regimen (everolimus with reduced-dose CNI or CNI withdrawal) or timing (early vs late initiation) did not alter the results. For patients on de novo or early conversion trials, there was also no difference in long-term graft function between the groups.
Long-term registry studies and 1 individual patient data meta-analysis reported inferior recipient survival in patients on mTORI-based regimens; however, the mechanisms behind such increases in risk remains unclear. Isakova et al23 and Badve et al24 both reported an increased risk of all-cause mortality with mTORI use in large, observational, registry cohort studies containing United States and Australian and New Zealand kidney transplant recipients, respectively. The main limitation for registry-based studies is the issue of confounding by indication. More compelling evidence of an increased risk of mortality with mTORI use comes from an individual patient data meta-analysis of sirolimus RCTs. Knoll et al25 found an increased risk of death in patients in both de novo and conversion trials of sirolimus compared with control. The study included trials of sirolimus, which targeted higher drug concentrations (>10 ng/mL) than present-day practice. This is further supported by the subgroup analysis, which found a significant mortality risk for high-dose sirolimus trials versus control, but no increased risk in low-dose sirolimus trials. Our study differs from Knoll et al to include more contemporary trials of everolimus-base immunosuppression and found no increase in 9-year risk of mortality for everolimus-based immunosuppression compared with control. Of note, our study did find a higher proportion of deaths due to cardiovascular disease in the everolimus group, whereas death rates from cancer and infection were similar between the groups. The number of cardiovascular deaths was small in our study; therefore, it is possible that these results may have occurred by chance. Similarly, Knoll et al25 also found a higher proportion of deaths attributed to cardiovascular disease in the sirolimus arm; however, as the study did not have patient-level data on total drug exposure, the results may have been due to indication and survival bias. Although mTORIs are known to have side effects associated with cardiovascular disease, such as proteinuria and dyslipidemia, it could be argued that CNIs also increase cardiovascular risk.8 Thus, the finding of an increase in cardiovascular death in patients randomized to mTORIs remains exploratory and should be interpreted with caution.
Inferior graft survival for mTOR-based regimens compared with CNI-based regimens has been reported in a cohort study23 and a meta-analysis of RCTs.26 Isakova et al23 reported an increased risk of death-censored graft failure in patients receiving mTORI without CNIs, especially within the first 2 years posttransplantation. In that study, 99% of the mTORI-treated patients received sirolimus. Similarly, the combination of mTORI and mycophenolate was associated with an increased risk of all-cause and death-censored graft failure when 16 RCTs were pooled.26 Although our study did not detect statistically significant differences in the risk of all-cause or death-censored graft loss, the Kaplan-Meier survival curves for both outcomes suggest a separation out to ~6 years, which raises the possibility that our study was underpowered to detect a true effect. With >40% of patients in the everolimus arm discontinuing the study drug at year 2, as opposed to only 10% in the control arm, it is also possible that the high discontinuation rate in the everolimus arm may have biased the results toward the null. Therefore, although we did not detect a statistically significant difference in the risk of graft failure in everolimus compared with the control, the long-term effects of everolimus on graft survival remains uncertain.
Graft function is a clinically relevant outcome for both clinicians and patients. In our study, we found no significant differences in either the slope of the decline or the mean differences in eGFR between the groups. The long-term deleterious effects of CNI exposure on graft function are of significant concern to the transplant community.6,27,28 Conversion from CNI to mTORI, with the aim of improving graft function, has been investigated as a strategy to reduce CNI toxicity; however, such studies have yielded conflicting results. Earlier studies of sirolimus in combination with a CNI showed inferior graft function in animals29 and humans.30 Later, trials of an everolimus-based, CNI-free regimen in kidney transplant recipients reported better graft function at 12 months,31 which was sustained to 5 years32 compared with cyclosporine-based therapy. Evidence generated from a systematic review and meta-analyses of conversion to mTORI studies showed significantly better eGFR at 1 year and 2 to 5 years after transplantation in both early and late conversion trials.13 Gatault et al33 found better mean eGFR for patients randomized to de novo sirolimus + MMF versus cyclosporine + MMF after 8 years of follow-up. In contrast, our study has found that within the limits of fewer eGFR measures between 8 and 10 years after transplantation (Figure 6), there was no significant difference in 10-year graft function for trials which initiated everolimus either de novo or within 6 months of transplantation compared with CNI-based triple therapy. The high discontinuation of rate of 30% at 1 year and 40% at 2 years in our study may have contributed to an absence of benefit in eGFR compared with previous trials.10,33 The recently completed TRANSFORM study reported a similar discontinuation rate of 27% in the everolimus plus reduced-dose CNI arm at 1-year follow-up.34 The results of our study suggests that the improved eGFR in patients on everolimus found in some studies may be mitigated by a high discontinuation rate, thus limiting the use of everolimus-based immunosuppression as routine de novo immunosuppression a large proportion of patients.
This is the first study to report long-term follow-up beyond 5 years for RCTs comparing everolimus with the standard-of-care. Long-term follow-up studies are important because most trials in kidney transplantation use surrogate outcomes, such as eGFR or biopsy-proven acute rejection, which have limited capacity to predict the hard outcomes of patient and graft survival.35 Graft loss, graft function, and death are critically important outcomes for patients and clinicians,36 thus these outcomes should always be reported in trials in kidney transplantation. However, patient-relevant outcomes are often difficult to capture beyond 5 years and remain a significant challenge due to logistics and cost. Until validated surrogate outcomes are established in clinical trials, linking long-term outcomes from trials to registry data remains an innovative and practical solution to such problem.
Strength and Limitations
Our study has several strengths. We included only prospective multicenter randomized trials. Randomization to treatment arm at baseline reduces the risk of selection bias and confounding by indication, both of which are major potential sources of bias in observational registry studies. Inclusion of individual patient data for all Australian and New Zealand participants (except 2 individuals that were not linked), along with the completeness of the baseline and follow-up data also strengthened our study. We reported clinically relevant hard outcomes and minimal patients were lost to follow-up. The main limitation of this study was the high rate of discontinuation in the everolimus arm which may have resulted in the absence of treatment effect in our intention-to-treat analysis. The results of the study must therefore be interpreted with this limitation in mind. Nevertheless, our study also highlights the issue of mTORI tolerability, which has important clinical implications when clinicians are faced with choosing the most appropriate immunosuppression regimen for their patients. Our study is also limited by a relatively small number of participants, thereby limiting the power to detect a true difference. However, as immunosuppression trials which are powered on long-term patient and graft survival are unlikely to be conducted, investigative approaches are required to explore long-term outcomes. Our study demonstrates that linkage to an established national registry is a feasible method to explore long-term outcomes in kidney transplantation trials with minimal loss to follow-up.
In conclusion, our study provides insights into the impact of everolimus on long-term hard outcomes of graft and patient survival. After 10 years of registry follow-up, Australian and New Zealand patients randomized to everolimus-based immunosuppression within 6 months after transplantation showed similar long-term graft function compared with CNI-based triple therapy. Long-term patient and graft survival among patients on everolimus are less certain due to the high discontinuation rate, which may limit its routine use. As linkage to registry studies provides an efficient and feasible way of collecting long-term outcomes of RCTs, future trials should consider combining outcomes from different international registries to increase power and generalizability.
The authors are grateful to all nurses, physicians, surgeons and database staff for maintaining the ANZDATA. The authors would also like to thank Chris Davies, biostatistician at ANZDATA for his assistance with data linkage and Dr Peter Bernhardt, medical affairs director for Transplantation at Novartis for providing deidentified patient data.
1. Tonelli M, Wiebe N, Knoll G, et al. Systematic review: kidney transplantation compared with dialysis in clinically relevant outcomes. Am J Transplant. 2011;11:2093–2109.
2. Lamb KE, Lodhi S, Meier-Kriesche HU. Long-term renal allograft survival in the United States: a critical reappraisal. Am J Transplant. 2011;11:450–462.
3. Laupacis A, Keown P, Pus N, et al. A study of the quality of life and cost-utility of renal transplantation. Kidney Int. 1996;50:235–242.
4. McDonald S, Russ G, Campbell S, et al. Kidney transplant rejection in Australia and New Zealand: relationships between rejection and graft outcome. Am J Transplant. 2007;7:1201–1208.
5. ANZDATA Registry. 39th Report Chapter 8: Transplantation. 2017. Available at http://www.anzdata.org.au/
6. Nankivell BJ, Borrows RJ, Fung CL-S, et al. The natural history of chronic allograft nephropathy. N Engl J Med. 2003;349:2326–2333.
7. Davies DR, Bittmann I, Pardo J. Histopathology of calcineurin inhibitor-induced nephrotoxicity. Transplantation. 2000;69:SS11–SS13.
8. Halloran PF. Immunosuppressive drugs for kidney transplantation. N Engl J Med. 2004;351:2715–2729.
9. de Fijter JW, Holdaas H, Øyen O, et al. Early conversion from calcineurin inhibitor- to everolimus-based therapy following kidney transplantation: results of the randomized ELEVATE trial. Am J Transplant. 2017;17:1853–1867.
10. Budde K, Lehner F, Sommerer C, et al. Five-year outcomes in kidney transplant patients converted from cyclosporine to everolimus: the randomized ZEUS study. Am J Transplant. 2015;15:119–128.
11. Tedesco-Silva H Jr, Vitko S, Pascual J, et al. 12-month safety and efficacy of everolimus with reduced exposure cyclosporine in de novo renal transplant recipients. Transpl Int. 2007;20:27–36.
12. Chadban SJ, Eris JM, Kanellis J, et al. A randomized, controlled trial of everolimus-based dual immunosuppression versus standard of care in de novo kidney transplant recipients. Transpl Int. 2014;27:302–311.
13. Lim WH, Eris J, Kanellis J, et al. A systematic review of conversion from calcineurin inhibitor to mammalian target of rapamycin inhibitors for maintenance immunosuppression in kidney transplant recipients. Am J Transplant. 2014;14:2106–2119.
14. Liu J, Liu D, Li J, et al. Efficacy and safety of everolimus for maintenance Immunosuppression of kidney transplantation: a meta-analysis of randomized controlled trials. PLOS ONE. 2017;12:e0170246.
15. O’Connell PJ, Kuypers DR, Mannon RB, et al. Clinical trials for Immunosuppression in transplantation: the case for reform and change in direction. Transplantation. 2017;101:1527–1534.
16. Lauer MS, D’Agostino RB Sr. The randomized registry trial—the next disruptive technology in clinical research? N Engl J Med. 2013;369:1579–1581.
17. Higgins J, Green S; The Cochrane Collaboration, editor. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. 2011: http://handbook.cochrane.org
18. White CA, Akbari A, Doucette S, et al. Estimating glomerular filtration rate in kidney transplantation: is the new chronic kidney disease epidemiology collaboration equation any better? Clin Chem. 2010;56:474–477.
19. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials. 1996;17:343–346.
20. Tedesco Silva H Jr, Cibrik D, Johnston T, et al. Everolimus plus reduced-exposure CsA versus mycophenolic acid plus standard-exposure CsA in renal-transplant recipients. Am J Transplant. 2010;10:1401–1413.
21. Vitko S, Tedesco H, Eris J, et al. Everolimus with optimized cyclosporine dosing in renal transplant recipients: 6-month safety and efficacy results of two randomized studies. Am J Transplant. 2004;4:626–635.
22. Holdaas H, Rostaing L, Serón D, et al. Conversion of long-term kidney transplant recipients from calcineurin inhibitor therapy to everolimus: a randomized, multicenter, 24-month study. Transplantation. 2011;92:410–418.
23. Isakova T, Xie H, Messinger S, et al. Inhibitors of mTOR and risks of allograft failure and mortality in kidney transplantation. Am J Transplant. 2013;13:100–110.
24. Badve SV, Pascoe EM, Burke M, et al. Mammalian target of rapamycin inhibitors and clinical outcomes in adult kidney transplant recipients. Clin J Am Soc Nephrol. 2016;11:1845–1855.
25. Knoll GA, Kokolo MB, Mallick R, et al. Effect of sirolimus on malignancy and survival after kidney transplantation: systematic review and meta-analysis of individual patient data. BMJ. 2014;349:g6679.
26. Sharif A, Shabir S, Chand S, et al. Meta-analysis of calcineurin-inhibitor-sparing regimens in kidney transplantation. J Am Soc Nephrol. 2011;22:2107–2118.
27. Naesens M, Kuypers DRJ, Sarwal M. Calcineurin inhibitor nephrotoxicity. Clin J Am Soc Nephrol. 2009;4:481–508.
28. Chapman JR. Chronic calcineurin inhibitor nephrotoxicity—lest we forget. Am J Transplant. 2011;11:693–697.
29. Andoh TF, Lindsley J, Franceschini N, et al. Synergistic effects of cyclosporine and rapamycin in a chronic nephrotoxicity model. Transplantation. 1996;62:311–316.
30. Kahan BD. Efficacy of sirolimus compared with azathioprine for reduction of acute renal allograft rejection: a randomised multicentre study. The Rapamune US study group. Lancet. 2000;356:194–202.
31. Budde K, Becker T, Arns W, et al. Everolimus-based, calcineurin-inhibitor-free regimen in recipients of de-novo kidney transplants: an open-label, randomised, controlled trial. Lancet. 2011;377:837–847.
32. Budde K, Sommerer C, Rath T, et al. Renal function to 5 years after late conversion of kidney transplant patients to everolimus: a randomized trial. J Nephrol. 2015;28:115–123.
33. Gatault P, Bertrand D, Buchler M, et al. Eight-year results of the Spiesser study, a randomized trial comparing de novo sirolimus and cyclosporine in renal transplantation. Transpl Int. 2016;29:41–50.
34. Pascual J, Berger SP, Witzke O, et al. Everolimus with reduced calcineurin inhibitor exposure in renal transplantation. J Am Soc Nephrol. 2018;29:1979–1991.
35. Clayton PA, Lim WH, Wong G, et al. Relationship between eGFR decline and hard outcomes after kidney transplants. J Am Soc Nephrol. 2016;27:3440–3446.
36. Sautenet B, Tong A, Manera KE, et al. Developing consensus-based priority outcome domains for trials in kidney transplantation: a multinational Delphi survey with patients, caregivers, and health professionals. Transplantation. 2017;101:1875–1886.