We evaluated outcomes with the sirolimus (SRL) and mycophenolate mofetil (MMF) combination regimen (SRL/MMF) in solitary kidney transplant recipients transplanted between 2000 and 2005 reported to the Scientific Registry of Renal Transplant Recipients. Three-and-a-half percent received SRL/MMF (n = 2040). Six-month acute rejection rates were higher with SRL/MMF (SRL/MMF: 16.0% vs. other regimens: 11.2%, p < 0.001). Overall graft survival was significantly lower on SRL/MMF. SRL/MMF was associated with twice the hazard for graft loss (AHR = 2.0, 95% C.I., 1.8, 2.2) relative to TAC/MMF, also consistent in both living donor transplants (AHR = 2.4, 95% C.I., 1.9, 2.9) and expanded criteria donor transplants (AHR = 2.1, 95% C.I., 1.7–2.5). Among deceased donor transplants, DGF rates were higher in the SRL/MMF cohort (47% vs. 27%, p < 0.001). However, adjusted graft survival was also significantly inferior with SRL/MMF in DGF-free patients (AHR = 1.9, 95% C.I., 1.6–2.3). In analyses restricted to patients who remained on the discharge regimen at 6 months posttransplant, conditional graft survival in deceased donor transplants was significantly lower with SRL/MMF compared to patients on TAC/MMF or CsA/MMF regimens at 5 years posttransplant (64%, 78%, 78%, respectively, p = 0.001) and across all patient subgroups. In conclusion, SRL/MMF is associated with inferior renal transplant outcomes compared with other commonly used regimens.
Antiproliferative agents are commonly used in combination with calcineurin inhibitors and corticosteroids as immunosuppressants in kidney transplantation (1). Inhibitors of the mammalian target of rapamycin (m-TOR inhibitors) such as sirolimus (SRL) and everolimus and agents such as mycophenolate mofetil (MMF) and mycophenolate sodium are the principal antiproliferatives used in renal transplantation. Azathioprine use has progressively dwindled since 1995 (2). MMF inhibits de novo purine synthesis and SRL inhibits m-TOR, preventing transition from G1 to S phase in affected cells (3,4). Both SRL and MMF inhibit proliferation of T and B lymphocytes; a key mechanism thought to underlie their immunosuppressive effect (3,5). Both these agents are also potent inhibitors of smooth muscle and fibroblast proliferation (3,5).
The m-TOR inhibitors were shown in experimental models and phase III trials to exhibit immunosuppressive synergy with calcineurin inhibitors and corticosteroids; and they provided superior protection from acute rejection in comparison to the then standard regimen of cyclosporine (CsA), azathioprine and corticosteroids (6,7). There were many anticipated benefits of the m-TOR inhibitors, the prototype of which is SRL. These included the ability of SRL to inhibit smooth muscle proliferation, inhibit antibody synthesis and promote a tolerogenic immunologic milieu (8,9). Quite expectedly, these anticipated beneficial effects of SRL were thought to be of potential utility in the prevention or amelioration of the vasculopathy that underlies chronic rejection of solid organ transplants (10,11). These forecasted benefits of the m-TOR inhibitors were counterbalanced by concerns of synergistic nephrotoxicity with the calcineurin inhibitors, hyperlipidemia, hematological toxicities and infectious risk (4,7,12–14). For a brief period, the SRL and tacrolimus (TAC) combination was used in kidney transplantation driven largely by experiences from 1 center (15). We and others have subsequently shown in both prospective clinical trials and retrospective analyses that SRL in combination with calcineurin inhibitors provides both inferior graft survival and renal function compared to CsA or TAC with MMF and corticosteroids in kidney transplantation (16–18).
In recent years, the combination of MMF with SRL has increased (2). This shift in the usage of SRL was likely prompted by the quest for a nonnephrotoxic regimen in kidney transplantation which would still provide adequate fulfillment of 1 of the primary aims of immunosuppression in kidney transplantation, the prevention of acute rejection (19). Notably, Flechner et al. have demonstrated excellent allograft function with preservation of kidney structure, freedom from acute rejection and excellent short-term allograft and patient survival in a prospective randomized controlled trial comparing SRL or CsA with MMF and corticosteroids in de novo renal transplant recipients (20,21). However, more recently, Larson et al. have shown no significant differences in graft function at 1 year in a prospective randomized trial comparing TAC or SRL with MMF and corticosteroids in de novo renal transplant recipients (22). Given the growing clinical experience with the SRL/MMF combination over the last 6 years and the accrued follow-up data, we decided to compare outcomes in renal transplants on this regimen to other commonly used immunosuppressant regimens in the United States.
In order to perform our analysis we utilized data pertaining to renal transplants that were performed between the years 2000 and 2005 in the United States and reported to the Scientific Registry of Transplant Recipients (SRTR). The results of our analysis are reported hereunder.
Data pertaining to immunosuppression and outcomes were derived from the SRTR database. We examined all adult (age >17) solitary kidney transplant recipients between 2000 and 2005 that were on 1 of 5 immunosuppressant regimen combinations [tacrolimus and MMF (TAC/MMF), cyclosporine microemeulsion and MMF (CsA/MMF), tacrolimus and sirolimus (TAC/SRL), cyclosporine microemulsion and sirolimus (CsA/SRL) or sirolimus and MMF (SRL/MMF)]. We categorized the use of induction immunosuppression as no induction, IL-2 receptor blockers (basiliximab and daclizumab), antithymocyte globulin (only when indicated as used as an induction agent), and other types. We defined delayed graft function (DGF) as the need for dialysis in the first week posttransplant or failure to produce more than 40 mL of urine within 24 h of transplantation. Acute rejection rates were derived from a database field indicating treatment for rejection. In the computation of 6-month acute rejection rates, we included events prior to discharge from the patient's initial hospitalization and the subsequent 6-month follow-up period for patients with at least 6 months of graft survival. Regimen discontinuation rates were assessed for patients with both the applicable amount of graft survival and available information in follow-up immunosuppressant files in the SRTR database. Graft and patient survival were analyzed with Kaplan–Meier plots and Cox proportional hazard models. Survival models were constructed both from baseline and for patients on treatment at 6 months posttransplantation (i.e. the same immunosuppressant regimen at 6 months posttransplantation as at discharge). Survival models utilizing the whole cohort were adjusted for donor type (living vs. deceased donor transplants), recipient race, donor race, recipient age, donor age, primary versus retransplant, primary diagnosis, time on dialysis (categorized as preemptive, 0–1 year, 1–2 years, 2–3 years and 3+ years), steroid utilization and type/use of induction. Donor age was categorized in the multivariate models as (0–6, 7–11, 12–17, 18–29, 30–39, 40–49, 50–59, 60–69, 70+). Among the deceased donor transplants, models were further stratified by DGF and extended criteria donor (ECD) status. All analyses were conducted in SAS version 9.1.3 (SAS, Cary, NC).
Evaluable Information was available for 58 131 patients, who constituted the study population. The demographic information pertaining to the study population stratified by immunosuppressant regimen is displayed in Table 1. Overall, 62% of patients were on a TAC/MMF discharge regimen and 23% received a CsA/MMF regimen. Four percent of patients in the study population received a SRL/MMF regimen (n = 2040). SRL/MMF recipients were more likely to be deceased donor transplant recipients, recipients of transplants from older donors, and more likely to receive induction therapy. As can be seen in Table 2, utilization of the TAC/MMF regimen has rapidly grown to over 80% of the proportion of recipients in 2005. The rate of SRL/MMF utilization increased from 1.5% in 2000, to a peak of 4.9% in 2002 and a gradual waning thereafter.
Table 1. Demographic characteristics of study population (n = 58 131)
Rates of acute rejection at 6 months by immunosuppressant regimen are displayed in Figure 1. Both among living and deceased donor transplants, 6-month acute rejection rates were highest for patients on a SRL/MMF discharge regimen (16.4% in the living donor transplants and 15.8% in the deceased donor transplants). The multivariate logistic model adjusted for potential confounders confirmed the association of higher rejection rates with the SRL/MMF regimen. SRL/MMF was associated with over 50% greater likelihood of acute rejection at 6 months relative to the TAC/MMF [adjusted odds ratio (AOR) = 1.53, 95% C.I., 1.33–1.75; p < 0.01]. CsA/MMF was also significantly associated with increased relative odds of 6-month acute rejection relative to TAC/MMF (AOR = 1.16; 95% C.I., 1.09–1.24; p<0.01). These results are displayed in Table 3. Twelve-month acute rejection rates were also highest among the SRL/MMF cohort for both living (18.1%) and deceased donor transplants (17.0%).
Table 3. Multivariate logistic regression model for 6-month acute rejection1
1Reference group includes patients on TAC/MMF at discharge, patients with a minimum of 6-month graft survival; model additionally adjusted for donor type, recipient and donor race and age, induction therapy, primary diagnosis and transplant number.
Kaplan–Meier survival plots for deceased donor transplant recipients stratified by discharge regimen are displayed in Figure 2. The SRL/MMF cohort was associated with significantly reduced overall graft survival relative to other regimens (log-rank p<0.001). In a similar fashion, overall graft survival was reduced among living transplants for patients on a SRL/MMF regimen (1-year graft survival = 89.4%; 3-year graft survival = 81.0%) as compared to TAC/MMF (1-year graft survival = 96.8% and 3-year graft survival = 90.7%). Even though DGF rates were higher in the SRL/MMF cohort among deceased donor transplants (47% vs. 27%, p<0.001), adjusted graft survival was also significantly inferior for patients not experiencing DGF [adjusted hazard ratio (AHR) = 1.9, 95% C.I., 1.6–2.3].
In the multivariate Cox model, we observed significant variation in overall graft and survival by immunosuppressant regimen (displayed in Table 4). In this model, the results of the univariate analysis were confirmed and SRL/MMF was associated with a twofold relative risk for overall graft loss (AHR = 2.01, 95% C.I., 1.83–2.23; p<0.01) and a 75% increased risk for patient death (AHR = 1.75, 95% C.I., 1.53–2.00; p<0.01) relative to the TAC/MMF discharge regimen. We also stratified our analyses in the SRL/MMF patients by the use of antibody induction therapy. In the absence of antibody induction therapy, SRL/MMF was associated with a 74% increased relative hazard for graft loss (AHR = 1.74; 95% C.I., 1.29–2.19). With the use of antibody induction (any kind), the use of SRL/MMF was associated with a two fold increased relative hazard for graft loss (AHR = 2.08; 95% C.I, 1.87–2.33) (Table 5).
Table 4. Multivariate Cox model for posttransplant outcomes1
Adjusted hazard ratio for overall graft loss
Adjusted hazard ratio for patient death
1Model additionally adjusted for transplant type, primary or re-transplantation, donor and recipient age and race, recipient primary diagnosis, induction use and type, steroid utilization and waiting time on dialysis.
Table 5. Adjusted hazard ratio for overall graft loss for patients on SRL /MMF regimen within patient subgroups
Adjusted hazard ratio1
Percentage of patients on treatment at 6 months
Adjusted hazard ratio for patients on treatment at 6 months2
1Reference group is patients on TAC/MMF at baseline; models additionally adjusted for transplant type, primary or re-transplantation, donor and recipient age and race, recipient primary diagnosis, induction use and type, steroid utilization and waiting time on dialysis.
2Reference group is patients on treatment at 6 months on TAC/MMF; proportion based on patients with minimum 6 months overall graft survival and indication of an immunosuppressive regimen at the 6-month follow-up period.
ECD = extended criteria donor.
Deceased donor transplant recipients
Living donor transplant recipients
African American recipients
Patients with a primary diagnosis of diabetes
Deceased donor transplants without DGF
Deceased donor transplants with DGF
Patients receiving any antibody induction
Patients not receiving Antibody Induction
Regimen discontinuation rates
The rates of discontinuation varied significantly by the baseline discharge immunosuppressant regimen. For patients on a TAC/MMF baseline regimen 6- and 12-month discontinuation rates were lowest at 15% and 27%, respectively. For other regimens, rates were significantly higher for the 6- and 12-month periods respectively as displayed in Figure 3: CsA/MMF (28% and 42%), TAC/SRL (30% and 48%), CsA/SRL (27% and 41%), and SRL/MMF (46% and 54%).
We separately examined long-term outcomes of patient and graft survival for patients that remained on baseline discharge immunosuppressive regimens at 6 months posttransplant and for those individuals having a minimum of 6 months of graft survival. The Kaplan–Meier survival plots for deceased donor transplant recipients on treatment at 6 months for the TAC/MMF, CsA/MMF and SRL/MMF regimens are displayed in Figure 4. SRL/MMF was associated with inferior graft survival compared to both MMF–calcineurin inhibitor combinations. Furthermore, the separation in survival curves began early and continued to increase with increasing follow-up time. A similar pattern was observed with living donor transplant recipients and was statistically significant (p<0.001; plots not shown). The multivariate model for the 6-month on-treatment analysis showed significant differences in graft survival between treatment regimens and confirmed the findings of the univariate analysis. SRL/MMF was associated with a 63% increased relative hazard for overall graft loss (AHR = 1.63; 95% C.I., 1.34–1.99; p<0.01). Table 5 displays the adjusted relative hazard for graft loss from baseline and from 6 months posttransplant on treatment in patient subgroups for patients on SRL/MMF relative to patients on TAC/MMF. The proportion of patients on the SRL/MMF regimen by subgroups is also displayed in Table 5. Rates of treatment continuation on SRL/MMF were highest among living transplant recipients and relatively lower among recipients of expanded criteria donor (ECD) transplants and African American recipients.
Transplant center regimen utilization and outcomes
Centers with up to 10% of patients on a SRL/MMF discharge regimen (low-use) constituted 73% of the study population. An additional 7% of patients were transplanted at centers with 11–30% of patients on the regimen (intermediate-use) and 1% of patients were transplanted at centers with more than 30% of patients on the SRL/MMF discharge regimen (high-use). Nineteen percent of centers had no patients on this regimen. Overall graft survival rates by immunosuppressant regimens for deceased donor transplants by centers' proportion of utilization of SRL/MMF are displayed in Table 6. As shown, overall graft survival rates for patients with the SRL/MMF combination were lower across all subgroups, though not attaining statistical significance in the high-use centers. Multivariate models for overall graft survival by the utilization groups also indicated higher hazards associated with the SRL/MMF. For centers with low-use of the SRL/MMF regimen, the adjusted hazard ratio for overall graft loss relative to TAC/MMF was 2.0 (95% C.I., 1.8–2.4), for intermediate-use centers, 1.7 (95% C.I., 1.3–2.3) and in high-use centers, 2.2 (95% C.I., 0.8–6.0).
Table 6. Overall graft survival in deceased donor transplants by center utilization of the SRL /MMF regimen
Proportion of patients on SRL /MMF discharge regimen
In this retrospective analysis, SRL/MMF de novo maintenance immunosuppression in solitary kidney transplant recipients was associated with the highest risk for acute rejection at 6 months and substantially inferior graft and patient survival in both living and deceased donor transplants compared to the reference regimen of TAC/MMF. Multivariate models and subgroup analyses confirmed these results.
The central finding in our study is that SRL/MMF is associated with inferior graft survival compared to the calcineurin inhibitor/MMF and even calcineurin inhibitor/SRL combinations. This difference in overall graft survival is evident very early in the posttransplant period and increases gradually over time. This difference might be in part driven by the lack of efficacy exemplified by the inferior rejection prophylaxis in addition to possible tolerability issues.
Any intent to treat analysis is potentially plagued by treatment switches that make the assessment of the real impact of a drug regimen challenging. Intent to treat designs certainly make detection of a potentially beneficial effect more challenging; but they also render the finding of a potential detrimental effect more robust. On the other hand, as this is a retrospective analysis and there is a real potential of upfront selection biases in the choice of which regimen was chosen for each patient, on-treatment analysis becomes important (23). In our present study, the on-treatment analysis for graft and patient survival only included those subjects who accrued a minimum of 6 months of graft survival and who remained on the same immunosuppressant regimen as at discharge. Even utilizing this conservative technique minimizing any early selection biases, the patterns and magnitude of graft and patient survival observed were very similar to those noted with the baseline graft survival analysis. In addition to mitigating a potential early selection bias, this model excludes all early failures and patients switched because of efficacy or tolerability issues, and concentrates only on those patients who still were on the same regimen at 6 months, which for the SRL/MMF cohort was only about half of the original intent to treat population. Notably this analysis documents a sustained higher attrition rate of grafts in the long term associated with the SRL/MMF regimen. In part, the inferior graft survival in the SRL/MMF group is likely driven by the increased acute rejection rates and potentially, ongoing subclinical rejection and/or repeat episodes of acute rejection late in the post-transplant period. The high rates of discontinuation of the SRL/MMF regimen likely reflect varying contributions of lack of efficacy and poor tolerability of the regimen.
The association of increased risk of death with SRL/MMF is possibly in part related to increased incidence of rejection episodes and graft loss and/or suboptimal renal function. The SRTR data pertaining to cause of death are incomplete and this makes the analysis of the exact causes of death by regimen difficult.
Our data do stand in contrast to the excellent early results reported by Flechner et al. utilizing the SRL/MMF combination with corticosteroids compared with CsA/MMF and corticosteroids in de novo renal transplant recipients (20). Notably, in their study, induction immunosuppression was used in all recipients, which could be the reason underlying their reported low rate of acute rejection. As described earlier in this paper, the SRL/MMF patients in this study were less likely to receive induction immunosuppression. In addition, one must keep in mind that the study reported by Flechner et al. used a CsA/MMF comparator; a combination that may be associated with slightly inferior allograft function compared to the TAC/MMF combination (24). Indeed, with the more contemporary comparator regimen of TAC/MMF (vs. SRL/MMF) in de novo kidney transplants, Larson et al. do not report inferior renal function in the TAC/MMF patients (22).
While our present study does yield measures of the strength and significance of associations of the various studied regimens to outcomes, interpretation of the results of a retrospective study such as ours should necessarily be tempered by the following caveats (25). First, our data reflect the use of drug combinations in the relatively uncontrolled setting of clinical practice as opposed to the codified restriction inherent to a clinical trial conducted in a carefully predefined population. Secondly, many biases govern the choice of immunosuppressive agents used in clinical practice. The reasons for regimen selection in particular patient subgroups are not discernible from registry data. It is very possible that higher risk transplants were selected to receive SRL/MMF or that this regimen was used preferentially for kidneys perceived to be at higher risk for nephrotoxic insults, such as kidneys from older donors and those from ECDs. Our statistical analysis attempted to correct some of the selection biases to the limits possible for the multivariate methods using such measures as on-treatment analysis of outcomes and analysis of outcomes across all patient subgroups. However, the nature of the database does not allow exclusion of others. An additional important limitation of this study is that the SRTR database does not contain any dosing or drug concentration data that can help interpret the effects of drug doses and exposure on transplant outcomes. Thus, any associations derived in analyses such as ours, should only be extrapolated to the pattern of clinical use of particular drug combinations during the historic timeframes analyzed.
Importantly, retrospective studies themselves do not establish cause and effect relationships (25). Proof of causality and the establishment of mechanistic links between risk factors and outcomes should necessarily include prospective experimental evidence (25). In that regard, corroborative evidence that is consistent with the results of our report come from (1) the Symphony study and (2) 2 clinical trials conducted in renal transplant recipients by Wyeth–Ayerst Pharmaceuticals. In the Symphony study standard-dose CsA based regimens are being compared to low-dose CsA, TAC or SRL in combination with MMF, daclizumab and corticosteroids in renal transplantation (26). In the reported 1-year results of this study, biopsy-proven acute rejection at 6 months in SRL/MMF patients was 33% versus 11% with TAC/MMF (p<0.01) and 22% with CsA/MMF. With regard to allograft function, calculated glomerular filtration rate was 57.3 mL/min with SRL/MMF versus 65.4 mL/min with TAC/MMF (p<0.0001). Lastly, 1-year graft survival was significantly inferior in SRL/MMF patients (TAC/MMF: 94%; SRL/MMF: 89%; p = 0.017) (26). Notably, targeted SRL concentrations in the Symphony study may be lower than those reported previously by Flechner et al. (20,21), but the higher acute rejection rates and the worse graft survival certainly mirror our findings from the U.S. retrospective data.
In addition, concerns about the de novo efficacy of the SRL/MMF combination arose from Wyeth–Ayerst Pharmaceuticals studies, known commonly as the 318 Trial (27) and ORION (28). Each of these trials included a SRL/MMF arm. Recently, the SRL/MMF arm of ORION was terminated and the 318 Trial halted both due to higher than expected acute rejection rates (Communication to Investigators from Wyeth–Ayerst Pharmaceuticals; June 2006), again in line with both the U.S. retrospective data and the mostly European Symphony trial.
In conclusion, the SRL/MMF regimen, in the context of clinical practice, is associated with inferior outcomes compared to other commonly used maintenance regimens in kidney transplantation in our retrospective study. These results need to be interpreted in the context of results of prospective trials and accruing clinical experience.
The data reported here have been supplied by the University Renal Research and Education Association (URREA) as the contractor for the SRTR. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government. IRB approval or exemption determination is the responsibility of the authors as well.