The Impact of Mycophenolate Mofetil on Long-Term Outcomes in Kidney Transplantation

Abstract

Mycophenolate mofetil (MMF) has been used in kidney and pancreas transplantation for almost 10 years. In the pivotal phase III trials, MMF use was accompanied by a dramatic reduction of rejection rates in kidney transplantation; however, the impact on graft and patient was undetermined. Analyses of the United States Renal Data System and the Scientific Registry of Transplant Recipients databases later provided a valuable measure of the impact of MMF in improving outcomes. In this review, we provide a synopsis of the prospective studies, including but not limited to the pivotal MMF approval trials, and analyses of the national transplant registries relevant to the long-term impact of MMF in kidney transplantation. Indeed, a substantial body of evidence has shown MMF treatment improves patient survival, graft survival, and death-censored graft survival in kidney transplantation. The beneficial effects of MMF have been particularly notable in high-risk recipients such as African Americans. In coming years, these benefits will require reevaluation in the context of the growing use of novel protocols combining MMF with tacrolimus or sirolimus.

Mycophenolate mofetil (MMF) has been used routinely for maintenance immunosuppression in renal transplantation in the United States for almost 10 years (^1,2). Despite the introduction of cyclosporine (CsA) in the 1980s and the resulting improvements in acute rejection rates and 1-year graft survival, transplant centers commonly reported renal allograft rejection rates of up to 40% even with the concomitant use of azathioprine (AZA) and corticosteroids (hereafter referred to as steroids) (^3,4). An important unmet clinical need thus existed at the time that MMF was introduced (^3,5). MMF increasingly has become the agent of choice for most centers, usually in combination with steroids and a calcineurin inhibitor (CNI). An analysis of the Scientific Renal Transplant Registry (SRTR) showed that the proportion of patients receiving MMF as one of their maintenance immunosuppressants increased from 11% in 1995 to 77% in 1999 (¹). In this review, we shall focus on the long-term outcomes in renal and pancreas transplantation utilizing MMF.

EARLY CLINICAL STUDIES AND THE PIVOTAL TRIALS

In a phase I study of MMF, Sollinger et al. (⁶) showed that the incidence of acute rejection inversely correlated with the dose in recipients of cadaveric renal transplants on standard doses of CsA and steroids. MMF also stabilized or improved function in almost 70% of patients with acute rejection refractory to pulse-dose corticosteroids and OKT3 (^7,8). Lack of significant adverse events related to MMF in these trials and the promising clinical outcomes set the stage for randomized, clinical trials of MMF in renal transplantation.

At the time the pivotal trials of MMF in kidney transplantation were conceived, 1-year graft and patient survival approximated 85% and 95%, respectively (²). Therefore, the primary efficacy endpoint utilized in the MMF phase III trials was the incidence of biopsy-proven rejection or treatment failure (graft loss, death, or premature termination for any reason) during the initial 6 months posttransplant (²). Incidence of acute rejection was chosen as the primary endpoint based on considerations that significant benefits in graft and patient survival directly attributable to MMF within a 6- to 12-month follow-up were unlikely given the practical limits of sample size and a low overall expected failure rate. Also, acute rejection episodes and their consequent treatment were considered important and clinically significant events, with even a single episode of acute rejection adversely impacting long-term graft survival (^2,9).

Three pivotal multicenter, double-blind trials of MMF were conducted in the United States (U.S. Renal Transplant Mycophenolate Mofetil Study Group) (¹⁰), Europe (European Mycophenolate Mofetil Cooperative Study Group) (¹¹), and Canada, Australia, and Europe (Tricontinental Mycophenolate Mofetil Renal Transplantation Study Group [TRI]) (¹²). These phase III trials together involved 1493 patients.

Each of the three phase III trials of MMF in renal transplantation clearly established the superiority of MMF over the control group for the primary endpoint of acute rejection. However, the number of patients in each of these trials was insufficient to allow analysis of small differences in other measures of clinical interest. Subsequently, Halloran et al. (⁴) combined data from the three trials to permit an increase in the sample size that might detect small differences in graft and patient survival. For purposes of this analysis, the AZA group was considered equivalent to the placebo group from the European study based on the overall similarity of data with respect to efficacy measures and the lack of convincing evidence to support the superiority of AZA over placebo when it was used in combination with CsA with steroids. Analysis of data pooled from 1,493 patients then was directed at endpoints measuring graft loss, patient death, incidence and treatment of rejection episodes, and allograft function (as reflected in the serum creatinine [SCr]) at 1 year posttransplantation. Even in the pooled analysis, no significant difference was observed in graft survival at 1 year between the MMF 2 g (90.4%), the MMF 3 g (89.2%), and the AZA-placebo group (87.6%). MMF treatment did, though significantly reduce the incidence of acute rejection, which was 40.8% with AZA-placebo versus 19.8% and 16.5% in patients receiving MMF 2 g and 3 g, respectively. Furthermore, renal function was better in the MMF treatment groups compared with the AZA-placebo treatment group.

Two important concerns emerged from the pivotal trial data (^12–14). Given the higher risk of acute rejection found in African-American renal allograft recipients, might MMF treatment afford specific benefits in this subpopulation? This issue was of particular interest to the North American transplant community. Secondly, would the reduced incidence in acute rejection rates seen at 6 months and 1 year translate into durable benefits in terms of graft and patient survival at 3 years posttransplantation and beyond? Further analysis of the pivotal trial data provided some evidence in support of both of these concerns.

Subset Analysis of the U.S. Study: Outcomes in African-American Recipients

Neylan (¹³) conducted a post hoc racial subgroup analysis of African-American patients enrolled in the U.S. trial. African Americans receiving AZA exhibited the highest incidence of biopsy-proven rejection/treatment failure (57.5% African Americans versus 43.5% non-African Americans). African Americans receiving MMF 3 g showed a significant reduction in biopsy-proven rejection/treatment failure rate (57.5% versus 24.2%; P=0.0008; AZA versus MMF). Biopsy-proven rejections occurred more frequently in both in the AZA (47.5%) and MMF 2 g (31.8%) groups compared with the MMF 3 g (12.1%) group. This finding stood in sharp contrast to the non-African Americans in both MMF dosage groups (AZA, 35.5%; MMF 2 g, 15.7%; MMF 3 g, 18.8%). African Americans in the AZA-treatment group experienced biopsy-proven rejection/treatment failure earlier than African Americans receiving MMF 3 g (median onset 64 days [AZA] versus 183 days [MMF 3 g]; P=0.0012). Overall, African Americans had more severe rejection episodes and higher SCr levels at 6 months posttransplant regardless of treatment group. These results demonstrated that African-American patients require a higher dose of MMF to adequately suppress acute rejection when compared with non-African Americans and led to the clinical practice of using MMF 3 g instead of MMF 2 g in this patient population.

Three-Year Efficacy Results From the European Trial of MMF in Renal Transplantation

The question of long-term efficacy of MMF in renal transplantation was addressed in an analysis of 3-year results from the European study (¹⁴). In this report, all 491 participants who received the study drug were included in the intent-to-treat analysis. Overall graft loss, death-censored graft loss, and death were described by univariate estimators. The 3-year patient survival was 88.9% in the placebo group, 92.7% in the MMF 2 g group, and 91.8% in the MMF 3 g group (P=NS). The overall 3-year graft survival was 78.0% in the placebo group, 84.8% in the MMF 2 g group, and 81.2% in the MMF 3 g group (P=NS). When compared with placebo, MMF 2 g had a significant favorable effect on 3-year death-censored graft loss (8.7% and 16.0%, respectively), with a significant difference in survival curves over the 3-year period (P=0.03).

Acute allograft rejection was the principal cause of graft loss across all groups (placebo, 10.8%; MMF 2 g, 4.6%; MMF 3 g, 6.3%). Furthermore, episodes of acute rejection were strongly associated with 3-year graft loss, with 31.5% of patients who experienced biopsy-proven acute rejection within 6 months of transplantation losing their grafts by the end of the 3-year study period compared with only 6.6% graft loss in patients who had no early acute rejection episodes (¹⁴). These findings supported the selection of acute rejection as a primary endpoint in the pivotal trials. More recently, it has been recognized that reduction in acute rejection rates early in the posttransplant period does not always impact favorably on long-term graft survival.

Diarrhea, anemia, cytomegalovirus infections, and leukopenia were the most commonly observed clinically relevant adverse events associated with MMF, occurring more frequently (as may be reasonably expected) in the MMF 3 g group. Importantly, the overall incidence of malignancies was similar across all treatment groups (¹⁴).

LONG-TERM OUTCOMES OF MMF IN RENAL TRANSPLANTATION BEYOND THE PIVOTAL TRIALS

The pivotal phase III trials clearly documented that MMF provides a significant beneficial effect in reducing the rate of acute rejection when compared with AZA in a CsA-based, triple-therapy regimen. Although it was hoped that this benefit might translate into improved long-term outcomes, a clear-cut effect of MMF on overall graft and patient survival could not be discerned from these trials (^4,14). Conversely, both experimental and clinical evidence indicated that MMF might have a beneficial effect on chronic allograft nephropathy (CAN) (^15,16). The failure to detect a beneficial effect of MMF on graft survival in the phase III trials was felt to be in part due to relatively small numbers of subjects and a consequent lack of power to detect small but clinically important effects. Indeed, the fact that even pooled analysis of data did not show any differences between MMF and AZA groups in graft and patient survival may have been in part accounted for by the heterogeneity of patient populations and the nonuniformity of the control group regimen (⁴).

These unanswered questions prompted the subsequent subanalyses of a large database maintained by the United Network for Organ Sharing (UNOS), enriched with data from sources such as the Social Security Death Master File and the Centers for Medicare and Medicaid Services and distributed by the United States Renal Data System (USRDS). The UNOS data are distributed also to the SRTR, a unique database that contains all Organ Procurement and Transplant Network data compiled since 1987 (^17,18). Both databases include information dating back to 1992 on patients receiving MMF as part of their maintenance immunosuppressive regimen (^19–24). In contrast to data from codified clinical trials, these databases reflect the use of immunosuppressives in the relatively uncontrolled setting of clinical practice. Nevertheless, endpoints such as graft loss and patient death are reported with surprising accuracy (¹⁷).

The limitations of these data need to be recognized, however. First, they are biased to the choice of immunosuppressive agents used in practice; and second, particularly with the USRDS and SRTR databases, information about dosage or drug concentrations is lacking. For these reasons, any associations derived from these data only reflect the pattern of clinical use of particular drug combinations during the historic time frames analyzed. Although retrospective analyses of these types of data can provide evidence of associations between risk factors (e.g., treatment regimen or race) and outcomes (e.g., time to graft loss or death) and do yield valuable measures of the strength and significance of such associations (²⁵), proof of causality relationships between a risk factor and outcome cannot be automatically assumed (²⁶).

Impact on Overall Long-term Graft and Patient Survival

The first issue to be addressed in using these data was the impact of MMF treatment on long-term graft loss in renal transplantation. To measure the impact of MMF on late allograft failure, USRDS records from 66,744 renal transplant recipients were studied (¹⁹). Only patients receiving a solitary kidney transplant between 1988 and 1997 were included. For lack of more precise data, an operational definition for chronic allograft failure (CAF), which differs from the traditional pathologic criteria that are used to diagnose CAN, was used in this study. CAF, the primary study endpoint, was defined as graft loss beyond 6 months posttransplant, censoring for death, rejection, and other reasons for graft loss with a discrete attributable cause. Secondary study endpoints were graft and patient survival.

As expected, given that MMF had been commercially available for a shorter time period than AZA, a greater number of records of patients receiving AZA (48,436 versus 8,435 for MMF) were included in the analysis. Univariate analysis revealed that death-censored graft survival at 4 years was significantly better in the MMF-treatment group than in the AZA group (85.6% versus 81.9%, respectively; P<0.001) and that patient survival in the MMF group was superior to the AZA group (91.4% versus 89.8%, respectively; P=0.002). The cumulative risk for CAF increased in the AZA-treated patients as opposed to those on MMF (Fig. 1)(¹⁹).

FIGURE 1.: Cumulative risk for chronic allograft failure (CAF). Azathioprine (AZA) is associated with an increasing cumulative risk for CAF compared to mycophenolate mofetil (MMF). Cox model estimates. From Meier-Kriesche et al. (¹⁹) with permission.

Risk Factors for Graft Loss

Multivariate models then were used to measure the relative impact of other covariates on graft loss and to adjust for confounding variables. Cox proportional hazards analysis demonstrated that acute rejection was the strongest risk factor for late graft loss due to CAF and conferred a risk ratio of 2.41. After controlling for acute rejection, the risk for graft loss due to CAF was reduced by 27% for MMF-treated patients (risk ratio=0.73) compared with the AZA-treated group (P<0.001). A subgroup of patients who had never experienced acute rejection was then examined using the same Cox model to see if MMF therapy was associated with a protective effect on long-term graft survival independent of its effects on acute rejection. Even in rejection-free patients, the risk of CAF was noted to be 20% lower in the MMF treatment group (²⁴). MMF was found to provide a long-term renal allograft survival benefit independent of its effects on acute rejection (²⁴).

Interestingly, one of the independent risk factors for CAF in this analysis was race; recipients of the African-American race were at higher risk for CAF. Because this study did not specifically address the impact of MMF therapy in African Americans, a subsequent separate analysis of the USRDS database was conducted to examine this issue (²⁰). The primary endpoints evaluated were patient death with a functioning graft and death-censored renal allograft survival. Secondary endpoints were acute rejection within the first 6 months posttransplant and chronic renal allograft failure. The study sample included 57,926 Caucasians and African Americans who underwent primary solitary renal transplantation between 1988 and 1997. Cox proportional hazard models, corrected for potential confounding factors, were employed to evaluate the interaction between transplant recipient race (African American versus Caucasian) and MMF or AZA therapy. Among Caucasian patients receiving AZA, 25.3% experienced an episode of rejection in the first 6 months posttransplantation compared with 15.3% of those receiving MMF (P<0.001). Among African-American patients, a more significant reduction in acute rejection rates was observed in patients receiving MMF (20.5%) versus AZA (32.8%) within the first 6 months posttransplant. Although there was a similar patient survival benefit associated with the use of MMF in Caucasian and African-American renal transplant recipients (Figs. 2A and B) (²⁰), for death-censored graft survival the relative benefit associated with MMF was stronger in African-American recipients (Figs. 3A and B) (²⁰). The interactions between MMF use, race, and the study endpoints were confirmed in the multivariate analysis (Table 1) (²⁰). In summary, the use of MMF was associated with clinically important and significant reductions in the risk of death with a functioning graft, death-censored graft loss, and CAF for Caucasians and even more often for African-American renal transplant recipients.

FIGURE 2.: (A) Kaplan-Meier estimated patient survival in Caucasian patients who were receiving azathioprine (AZA) vs. mycophenolate mofetil (MMF). (B) Kaplan-Meier estimated patient survival in African-American patients who were receiving AZA vs. MMF. From Meier-Kriesche et al. (²⁰) with permission.

FIGURE 3.: (A) Kaplan-Meier estimated death-censored graft survival in Caucasian patients who were receiving azathioprine (AZA) vs. mycophenolate mofetil (MMF). (B) Kaplan-Meier estimated death-censored graft survival in African American patients who were receiving AZA vs. MMF. From Meier-Kriesche et al. (²⁰) with permission.

TABLE 1: The interactions between mycophenolate mofetil use, race, and the study endpoints

Another important finding from this retrospective analysis was that the beneficial effects of MMF on acute rejection and graft loss were not accompanied by an increase in the risk of death. Thus, based on the evidence from this study (²⁰), it appeared that MMF not only decreases the risk of graft loss but also may be associated with a wider therapeutic index compared to AZA (paralleling theoretical expectations based on the relatively specific and potent immunosuppressive properties of MMF) (^27,28).

The impact of MMF on variables well known to enhance graft survival, such as human leukocyte antigen (HLA) matching, also was studied (²⁹). An analysis of USRDS data found that MMF did not obviate the beneficial effects of HLA matching; HLA matching and MMF therapy were additive factors in reducing the risk for renal allograft loss (²¹). This analysis confirmed the findings of a previous study (¹⁹) on the graft survival benefits associated with MMF versus AZA in a relatively contemporary cohort of patients who were better matched with regard to relevant practice patterns that may influence outcome.

Impact of the Occurrence of Acute Rejection 1 Year Posttransplantation

Although the effect of MMF in preventing early episodes of acute rejection was well documented, its effect on late episodes was less clear. As described previously, CAF rates were found to be lower even in MMF-treated subjects who had never experienced an episode of acute rejection (²⁴). Could the effect of MMF on CAF be attributed, at least in part, to its ability to prevent episodes of acute rejection beyond 1 year posttransplantation? This question was addressed in an analysis of USRDS data that included 47,693 primary renal allograft recipients transplanted between 1988 and 1998 (²²). MMF also was found to be associated with a reduced risk of developing late acute rejection when compared with AZA (relative risk [RR]=0.35; confidence interval [CI], 0.27–0.45; P<0.001) (Fig. 4) (²²). Furthermore, the incidence of late acute rejection episodes at both 2 and 3 years posttransplantation was significantly lower in MMF-treated patients than in the AZA-treatment group (0.9% versus 6.1% at 2 years; 1.1% versus 9.3% at 3 years, respectively; P<0.001).

FIGURE 4.: Continued treatment ≥12 months with mycophenolate mofetil (MMF) vs. azathioprine (AZA) is associated with decreased late rejection. Kaplan-Meier estimates. From Meier-Kriesche et al. (²²) with permission.

Acute rejection episodes were further stratified as primary late rejection episodes where rejection was first noted 12 or more months posttransplantation and as secondary late rejection episodes where late rejection was preceded by an acute rejection episode within 12 months posttransplantation. In these subgroup analyses, MMF-treatment was associated with a significantly lower risk of late primary and secondary acute rejection (60%; RR=0.40; P<0.001 and 72%, respectively; RR=0.28; P<0.001) compared to the AZA-treatment group. Importantly, in African Americans, the late acute rejection risk was 70% lower in MMF patients than in AZA-treated patients (RR=0.30; P<0.001) (²²). In comparison with AZA, MMF therapy was associated with significantly lower incidence of late acute rejection episodes in Caucasian and African-American recipients (Fig. 5) (²²). In fact, it appears likely that the reduction in late acute rejection episodes might be one mechanism by which MMF benefits long-term graft survival.

FIGURE 5.: Continued treatment ≥ 12 months with mycophenolate mofetil (MMF) vs. azathioprine (AZA) is associated with decreased late rejection. Kaplan-Meier estimates stratified by MMF vs. AZA and African American (AA) vs. white recipients. From Meier-Kriesche et al. (²²) with permission.

Preservation of Renal Function

An intrinsic expectation from the findings of these database analyses cited above is that MMF treatment should be associated with preservation of renal function over time (²³). In this analysis, the investigators defined a primary endpoint as at least a 20% decline below a 6-month baseline in the 1/SCr (slope of reciprocal creatinine) at or beyond 1 year posttransplantation. A secondary endpoint was attainment of a SCr value of 1.6 mg/dL or greater. Using follow-up data to 5 years posttransplantation, the continuing use of MMF versus AZA beyond the first year was associated with a protective effect against declining renal function as measured by the slope of the reciprocal creatinine (RR=0.84; CI, 0.78-0.91; P<0.001) (Fig. 6) (²³). When used for 24 months and beyond and compared to AZA, MMF was associated with a further decreased risk for a decline in renal function (RR=0.66; CI, 0.57-0.77; P<0.001). Furthermore, MMF was associated with a protective effect against reaching the SCr threshold of 1.6 mg/dL (RR=0.80; P<0.001) when use was continued beyond 12 months posttransplantation (²³). These retrospective, cohort studies suggest that there is a sustained beneficial effect on renal function when MMF treatment is continued for at least a year. Taken together, findings of the two aforementioned studies on stability of renal allograft function (²³) and decreased risk of late acute rejection (²²) indicate that some of the benefit of MMF in preserving long-term renal allograft function may in fact arise from preventing late episodes of acute rejection especially in high-risk populations such as African Americans.

FIGURE 6.: Continued mycophenolate mofetil (MMF) for 12 months or greater is associated with better preservation of renal function. Kaplan-Meier estimates of progressive decline in renal function defined as 20% or greater decline in slope of 1/Cr. From Meier-Kriesche et al. (²³) with permission. AZA, azathioprine.

MMF IN COMBINATION WITH TACROLIMUS

In the studies reviewed thus far, MMF had been used in combination with CsA according to approved product labeling and based on the results of phase III trials. In recent years, though, CsA has been largely replaced in many transplant centers by tacrolimus (²). Importantly, this shift reflected a gradual change in practice patterns based on empiric observations and expectation of less nephrotoxicity with tacrolimus that were only subsequently supported by actual published results of formal studies evaluating this combination (³⁰). Retrospective analyses and clinical investigation support the notion that renal function is superior in patients receiving MMF with tacrolimus as opposed to CsA (^{31, 32}). Johnson et al. (³³) have reported a 3-year follow-up of their experience with 223 recipients of first cadaveric kidney allografts who were randomized to receive tacrolimus with MMF, tacrolimus with AZA, or CsA (microemulsion) with MMF. Each of these regimens included steroids. Monoclonal antibody induction was used only in instances of delayed graft function (DGF) within the first day of transplant. Of note, the efficacy endpoints in this study were incidences of biopsy-proven acute rejection episodes and acute rejection requiring treatment with antilymphocyte antibody. After 1 year, the rate of steroid-resistant acute rejection was lowest in the tacrolimus/MMF treatment group, without significant differences in overall incidence of acute rejection (³³). Interestingly, in this study, patients with DGF treated with tacrolimus and MMF experienced an increase in 3-year allograft survival compared with patients receiving CsA and MMF (84.1% versus 49.9%; P=0.02). However, the selective use of induction therapy in the patients with DGF obfuscates this finding. Furthermore, patients randomized to either treatment arm containing tacrolimus exhibited lower SCr concentrations than those receiving CsA. Given these results, it is quite conceivable that long-term graft survival can be expected to be superior when MMF is used in combination with tacrolimus as opposed to CsA.

MMF COMPARED WITH SIROLIMUS

In Cyclosporine-Based Immunosuppressive Regimens

Recent years have experienced an increasing use of MMF or sirolimus with CNIs with very few centers using AZA in newly transplanted recipients (³). Thus, a more current and relevant comparison is MMF and sirolimus in CNI-based immunosuppressive regimens. In this recently published retrospective study, data from 23,016 primary recipients reported to the SRTR between 1998 and 2003 were analyzed (³⁴). The regimen combining CsA and sirolimus was associated with significantly lower graft survival (74.6% versus 79.3% at 4 years; P=0.002) and death-censored graft survival (83.7% versus 87.2%; respectively; P=0.003) than CsA and MMF. In multivariate analyses, the CsA-sirolimus combination was associated with a significantly increased risk for graft loss (Fig. 7) (³⁴), death-censored graft loss (Fig. 8) (³⁴), and decline in renal function (hazard ratio [HR]=1.22; P=0.002; HR=1.22; P=0.018; and HR=1.25; P<0.001, respectively) (Fig. 9) (³⁴). Furthermore, regimens combining MMF with CsA or tacrolimus were associated with better serial creatinine clearances than regimens combining sirolimus with CsA. The association of lower creatinine clearances with the CsA-sirolimus combination was thought to reflect the effects of full-dose CsA with sirolimus (³⁴), a regimen known to be associated with significantly poorer renal function in the phase III sirolimus studies (^35–38).

FIGURE 7.: Mycophenolate mofetil (MMF) with cyclosporine (CsA) is associated with better graft survival than CsA with sirolimus (SRL). Kaplan-Meier estimates of graft survival. From Meier-Kriesche et al. (³⁴) with permission.

FIGURE 8.: Mycophenolate mofetil (MMF) with cyclosporine (CsA) is associated with better death-censored graft survival than CsA with sirolimus (SRL). From Meier-Kriesche et al. (³⁴) with permission.

FIGURE 9.: Mycophenolate mofetil (MMF) combined with calcineurin inhibitors (tacrolimus [TAC] or cyclosporine [CsA]) affords better serial renal function than sirolimus (SRL) combined with a calcineurin inhibitor. From Meier-Kriesche et al. (³⁴) with permission.

More recently, Ciancio et al. (³⁹) have compared MMF in combination with tacrolimus with regimens combining CsA or tacrolimus with sirolimus in a long-term, prospective, randomized study of 150 first cadaveric and HLA-nonidentical living donor renal transplants. All patients received maintenance steroids and daclizumab induction (³⁹). The protocol involved a progressive dose-reduction regimen for tacrolimus and CsA over 1 year, with MMF doses and target sirolimus trough levels being maintained (⁴⁰). At 1-year, reduced-dose tacrolimus with sirolimus had a similar effect on renal function as did the tacrolimus and MMF combination (³⁹). However, statistically significant trends of a declining creatinine clearance and SCr levels were noted in the 1-year follow-up in the CsA-sirolimus group (³⁹). This finding parallels inferior SCr levels reported in the sirolimus phase III trials, albeit in the context of more modern comparator regimens (^{35, 36}). The incidence of biopsy-proven rejection was comparable between the tacrolimus-MMF and the tacrolimus-sirolimus treatment groups but higher in the CsA-sirolimus group (³⁹). This is a worrisome finding despite higher sirolimus trough levels observed in the CsA-treated group (³⁹). Whether preservation of renal function with the sirolimus-tacrolimus combination will continue to endure in the long term remains to be seen.

In Tacrolimus-Based Immunosuppressive Regimens

The perception that tacrolimus is associated with a lower level of nephrotoxicity and the concerns about the adverse effects of the CsA-sirolimus combination have, in recent years, led to increasing use of tacrolimus in combination with sirolimus (^{36, 41–43}). The only published randomized, prospective study comparing sirolimus and MMF in combination with tacrolimus currently yields 1-year follow-up data (⁴⁴). In this multicenter study, 361 renal transplant patients were included and randomized 1:1 to receive steroids and either tacrolimus and sirolimus (n=185) or tacrolimus and MMF (n=176). Renal function, as measured by SCr levels, was significantly worse in the patients randomized to the tacrolimus and sirolimus-treatment group with 20% of patients having a SCr level exceeding 2.0 mg/dl versus 11.2% in the MMF-treatment group. A trend toward poorer graft survival also was observed in the group receiving tacrolimus and sirolimus (91% graft survival versus 94% with tacrolimus and MMF; P=0.22) (⁴⁴). The failure to attain statistical significance could, in part, be a function of the limited sample size, few events, short follow-up time, and the consequent lack of power to detect small differences.

As discussed above, retrospective analyses of data from large transplant databases can help surmount these limitations (²⁵). A recent analysis of SRTR data confirms that allograft recipients treated with tacrolimus and MMF exhibit better graft survival than those receiving CsA and MMF or tacrolimus and sirolimus (⁴⁵). This database analysis almost duplicates the 1-year graft survival rates noted in the prospective study (tacrolimus and sirolimus, 91.8%, versus tacrolimus and MMF, 94.2%) but yields a statistically significant difference (P<0.01). In our retrospective analysis, we also have documented a progressive separation of the survival curves over time, with a clinically significant difference at 3 years after transplantation (tacrolimus and sirolimus, 80.3%, versus tacrolimus and MMF, 85.9%; P<0.001). This difference in graft survival was similar in magnitude to that previously discussed between CsA and sirolimus and CsA and MMF in both the univariate and multivariate analyses (³⁴).

MMF AND SIROLIMUS IN COMBINATION

Flechner et al. (⁴⁶) have demonstrated excellent results in kidney transplants using a CNI-free regimen of MMF and sirolimus with steroids designed to avoid nephrotoxicity. A prospective, randomized trial was carried out in 61 kidney transplant recipients, each of whom received basiliximab induction, steroids, and MMF 2 g/day. Patients then were randomized to receive sirolimus (n=31) or CsA (n=30). One-year patient and graft survival and rejection rates did not differ between the CsA- and sirolimus-treated groups. However, at both the 6, and 12-month time points, SCr levels and, as an extension, calculated creatinine clearance rates were worse in the CsA group (⁴⁶). Thus, it appears that sirolimus in combination with MMF used without a CNI is associated with preserved renal allograft function and excellent graft survival in the short term. Whether these differences in renal function represented merely a hemodynamic effect of CsA on renal function or were accompanied by structural events was not readily apparent. Interestingly, sirolimus-treated patients exhibited higher MMF trough levels (probably reflecting absence of the known association of lower mycophenolic acid [MPA] trough levels) (⁴⁶). The comparator CsA-based regimen in this study is known to be associated with a lower exposure to MPA. Whether these results would be similar if the sirolimus-MMF combination is compared to a tacrolimus-MMF combination with consequently higher exposures to MPA is uncertain (^47,48). As a follow-up to this study, protocol biopsies were obtained at 2 years and analyzed both by Banff criteria and deoxyribonucleic acid microassays (⁴⁹). Both calculated creatinine and iothalamate clearances showed steady improvement in the sirolimus-MMF group and a slow decline in the CsA-MMF group. Of note is that 2-year protocol biopsies in this study with higher Banff chronic scores exhibited upregulation of several genes associated with inflammation and fibrosis. Again, the question remains open as to how these results would compare with a control arm that included tacrolimus with MMF (⁴⁹). Long-term results from studies such as these on larger numbers of patients are critical in gaining a better understanding of outcomes associated with MMF in combination with newer immunosuppressants (⁵⁰).

MMF IN DIABETIC TRANSPLANT RECIPIENTS

Posttransplantation cardiovascular mortality is disproportionately higher in the diabetic transplant recipient (^{51, 52}). As reviewed above, beneficial effects of MMF (over AZA) include improved graft and patient survival in high-risk populations such as African Americans, decreased incidence of CAF, decreased late acute rejection, and better preservation of allograft function (^{19, 20, 22}). It also has been previously shown that better renal function after transplantation lessens cardiovascular mortality (⁵³). Thus, the effect of MMF on preserving graft function could be expected to translate into improvements in cardiovascular mortality after transplantation. The hypothesis that MMF (versus AZA) treatment would reduce cardiovascular mortality in diabetic transplant recipients was tested in a recent study by David et al. (⁵⁴) analyzing SRTR data. The study population was comprised of 17,145 adult primary transplant patients with diabetes receiving MMF (n=14144) or AZA (n=3001). The endpoints in this study were acute rejection, late acute rejection, patient survival, and rates of graft loss. The study specifically investigated cardiovascular mortality, infectious mortality, and mortality due to malignancy. MMF patients exhibited decreased rates of acute rejection and late acute rejection compared to those on AZA (as expected from previous studies). Interestingly, MMF was associated with a 20% decrease in the risk for cardiovascular death compared with AZA treatment (RR 0.80; 95% CI 0.67-0.97; P=0.020). MMF-treated patients in this study also had a lower risk for malignancy. Whether this incidence of malignancy would change with an increase in the follow-up time is an open question.

MMF AND CHRONIC ALLOGRAFT NEPHROPATHY (CAN)

MMF mitigates chronic failure of renal allografts. The major causes of late graft failure is thought to be CAN, a term used to describe the histologic changes that accompany long-term deterioration of graft function: fibrointimal thickening of arterioles, fibrosis, and a characteristic glomerulopathy (⁵⁵). It has been postulated that this entity may be mediated by pathways involving humoral immunity (^{56, 57}). The course of CAN may be accelerated by CNI toxicity, which itself can result in progressive fibrosis of the allograft (⁵⁸). It also has been shown in small single-center experiences that antidonor antibody responses can be reduced with tacrolimus and MMF in combination in CAN (⁵⁹). As humoral responses are felt to be an important accompaniment to CAN, this property of MMF also may contribute to its effects on decreasing long-term graft loss. In fact, MMF is devoid of intrinsic renal, cardiovascular or metabolic toxicities (⁵⁴).

In a recent study, Merville et al. (⁶⁰) demonstrate that the prevalence of CAN is lower in 1-year protocol biopsies in CsA-and corticosteroid-treated patients receiving MMF (n=37) as opposed to AZA (n=34). In the intent-to-treat analysis the proportion of patients with CAN was 46% in the MMF group versus 71% in the azathioprine group (P=0.03). The efficacy analysis in this study (evaluating only the 78.8% of patients with 1-year protocol biopsies) showed similar findings, with 31.0% of MMF-treated patients having CAN compared to 63.0% of the AZA-treated patients (P=0.01). These findings corroborate the decreased propensity to CAF reported by Ojo et al. cited earlier (²⁴).

Thus, it appears that MMF therapy is useful in the context of chronic progressive dysfunction of renal allografts by reducing the incidence of CAN and allowing the clinician to mitigate the added nephrotoxic insult conferred by a CNI. However, to date, there is no generally accepted, easy, and fail-safe method to lower exposure to a CNI and yet maintain rejection-free graft survival.

It should be noted that the trials that have been summarized above have used MMF in fixed doses. It is becoming increasingly apparent that changes in concomitant immunosuppression significantly impact MMF pharmacokinetics (^{61, 62}). One mechanism that potentially could account for the superiority of graft function with the tacrolimus-MMF combination could be based on the increased MPA exposure that is obtained when tacrolimus is used instead of CsA (^{48, 63}). This augmentation of MPA exposure by tacrolimus potentially could confer adequate immunosuppression while at the same time allow decreases in exposure to the nephrotoxic effects of tacrolimus. This is an issue under active investigation in ongoing trials.

CONCLUSIONS

The introduction of MMF must be viewed as a significant and clinically important step forward in improving graft and patient survival in kidney and pancreas transplantation. More importantly, better outcomes have occurred with an acceptable risk profile. In fact, MMF is devoid of intrinsic renal, cardiovascular, or metabolic toxicities. Furthermore, the benefits of MMF extend particularly to traditional high- risk groups such as African Americans. Two cautionary notes are in order at this juncture. As we have repeatedly noted, the paradigm that has governed practice, clinical trial design, and outcomes research in transplantation has focused on the prevention of acute rejection. However, reduction in acute rejection rates does not automatically translate into long-term graft survival. In a recent analysis of the SRTR data pertaining to the first adult transplants performed between 1995 and 2000, outcomes for renal transplantation in the most recent era do not match expectations or the progressive improvements noted in previous years (⁶⁴). This finding, despite a reduction in acute rejection rates, is accompanied by a significant trend to increasing death-censored graft loss and lesser numbers of rejection episodes returning to baseline levels of renal function after treatment (⁶⁴). Thus, we may need to rethink endpoints that are utilized in the study design; in addition to acute rejection rates, such parameters as proportion of rejection episodes that return to baseline function after successful treatment may be helpful. A second major factor is the disturbing increase in incidence of polyoma virus nephropathy (⁶⁵). The exact contribution of MMF usage to this epidemic is not known (^{66, 67}) but its occurrence demands a reduction in immunosuppression. Inasmuch as it impacts the overall level of immunosuppression that can be delivered and still maintain efficacy in preventing rejection, an uncertain effect will be exerted on renal transplant outcomes. Future studies must address these important issues even as we pause to reflect on the advances of the last decade.