Ninety-seven grafts were lost during follow-up (see Table S1, SDC, http://links.lww.com/TP/A668). Overall, there was no significant difference in graft survival between the groups (Fig. 1B and Table 2). Graft survival was higher in the group with 3 g/d MMF compared with the group with 2 g/d MMF, although this was of borderline statistical significance (hazard ratio [HR], 0.61; 95% confidence interval [CI], 0.37–1.00; P=0.05). In the first year after transplantation, there were six graft losses in the AZA group (two with acute rejection, two with rejection after reduction of immunosuppression, two with other causes) compared with four losses in the group with 2 g/d MMF (including one death) and three losses in the group with 3 g/d MMF (including two deaths).
Death-censored graft survival was highest in the group with 3 g/d MMF and lowest in the group with 2 g/d MMF (Fig. 1C and Table 2). The group with 3 g/d MMF was consistently superior to that with 2 g/d MMF (HR, 0.38; 95% CI, 0.19–0.79, P=0.01). The six early graft losses in the AZA group led to a nonsignificantly inferior death-censored graft survival compared with the group with 3 g/d MMF in the first year (Table 2; HR for MMF vs. AZA, 0.16; 95% CI, 0.02–1.35). However, after the first year, death-censored graft survival in the AZA group was comparable to the group with 3 g/d MMF with approximately parallel survival curves (Fig. 1C; HR for MMF vs. AZA, 0.77; 95% CI, 0.33–1.81).
There were 27 non–skin cancers diagnosed during follow-up; 7 in the AZA group, 9 in the group with 2 g/d MMF, and 11 in the group with 3 g/d MMF. Seventy-six patients were diagnosed with at least one skin cancer during follow-up; 28 in the AZA group, 28 in the group with 2 g/d MMF, and 20 in the group with 3 g/d MMF. There were no differences in either non–skin or skin cancer incidence between the groups (Fig. 2 and Table 2).
Serum creatinine concentrations were very well reported (<1% missing). There were no differences in estimated kidney function between the groups (see Figure S1 and Table S2, SDC, http://links.lww.com/TP/A668). The 15-year mean estimated glomerular filtration rate (eGFR) of functioning grafts was 46.1 mL/min per 1.73 m2 in the AZA group, 45.9 mL/min per 1.73 m2 in the group with 2 g/d MMF, and 50.9 mL/min per 1.73 m2 in the group with 3 g/d MMF (P=0.17). Compared with the AZA group, the mean eGFR in the group with 3 g/d MMF was 4.7 mL/min per 1.73 m2 higher (95% CI, −1.0 to +10.5; P=0.11). After imputing an eGFR of 0 at graft failure, there remained no difference between groups using either linear (P=0.08) or tobit (P=0.07) regression; in all models, mean eGFR was nonsignificantly highest in the group with 3 g/d MMF.
Finally, there were no differences in the slope of eGFR against time. The slope of eGFR was −0.9 mL/min per 1.73 m2 per year in the AZA group, −1.6 mL/min per 1.73 m2 per year in the group with 2 g/d MMF, and −0.7 mL/min per 1.73 m2 per year in the group with 3 g/d MMF (P=0.24). After imputing an eGFR of 0 at graft failure, there remained no difference between the groups using either linear (P=0.08) or tobit (P=0.15) regression; in all models, the slope of eGFR was nonsignificantly worst in the group with 2 g/d MMF.
Not all patients remained on their assigned treatment throughout the entire follow-up (see Figure S2, SDC, http://links.lww.com/TP/A668). Most randomized to AZA remained on AZA for at least 7 years. Over the first 3 years after transplantation, approximately one third of patients assigned to MMF switched to AZA; the remainder generally stayed on MMF in the long term. The actual MMF dose in the group with 3 g/d MMF fell over time (Fig. 3; see Table S3, SDC,http://links.lww.com/TP/A668) but remained significantly higher than the 2-g/d group until 5 years after transplantation. Steroid withdrawal occurred at the same rate in the different groups (P=0.58), with 41% of patients being steroid-free at 3 years and very few withdrawals from steroids thereafter. The doses of steroids and cyclosporine prescribed, per kilogram body weight, were equivalent between the groups at all time points. Mean daily steroid dose was 0.5 mg/kg at baseline and 0.1 mg/kg at 1, 5, and 10 years. Mean daily cyclosporine dose was 5.5 mg/kg at baseline, 2.7 mg/kg at 1 year, 2.5 mg/kg at 5 years, and 2.3 mg/kg at 10 years. No patient commenced tacrolimus during follow-up.
The secondary as-treated analysis demonstrated a benefit of MMF over AZA in the first year for both graft survival (HR for MMF, 0.28; 95% CI, 0.08–0.92; P=0.04) and death-censored graft survival (HR, 0.15; 95% CI, 0.03–0.73; P=0.02). However, no benefit was demonstrated beyond the first year for either graft survival (HR, 1.31; 95% CI, 0.81–2.10; P=0.27) or death-censored graft survival (HR, 1.36; 95% CI, 0.71–2.62; P=0.35).
In the Australian arm of the Tricontinental Mycophenolate Mofetil Renal Transplantation Study, there were no detectable differences in patient or graft survival, cancer incidence, or estimated kidney function at 15 years after transplantation. This is the first study reporting such long-term follow-up of a randomized trial comparing MMF with AZA for kidney transplantation. Such studies are important because it is not clear whether early surrogate outcomes, such as acute rejection and eGFR, are adequate predictors of ultimate graft outcome. Indeed, most of the gains in kidney transplantation over the last two decades have led to reduced early but not late graft loss (12), suggesting that analyzing surrogate outcomes is insufficient to appreciate the long-term impact of a treatment strategy.
The main strength of this study is that we report “hard” outcomes out to 15 years with only one patient lost to follow-up. With 97 grafts lost and 75 deaths, there were many more hard outcomes than in the original reports of the landmark MMF trials (1–3). This is also the first study to report long-term cancer incidence in a trial of MMF versus AZA. Another strength of this study is the completeness of the data, including nearly complete reporting of medications and serum creatinine values.
The biggest weaknesses of this study are the relatively small numbers and the crossover between treatment groups. Although most patients reached a hard outcome, the small sample size necessarily leads to some uncertainty about the difference between the groups and a type II error cannot be excluded.
Crossover is also likely to bias the results toward there being no difference between the treatment groups. It is somewhat reassuring that the as-treated analysis of graft survival also failed to demonstrate a difference between the groups beyond 1 year, but such an analysis is subject to confounding by indication, especially as the indication for crossover is not recorded by the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry. Although crossover is a major issue for our study, the only feasible ways to obtain such long-term data are to link a randomized trial to registry data, as we have performed, or to conduct a purely observational study. Such an observational study would incorporate the biases of our study but would not account for unmeasured confounders; these can only be accounted for by randomization. Thus, although clearly limited by crossover, our study design is the strongest practical design for analyzing such long-term outcomes.
Another weakness of our study is its dependence on registry data, in this case leading to limited capture of rejection data and no information on proteinuria. Rejection episodes have only been recorded by ANZDATA since 1997. Nevertheless, most rejection episodes occur in the first 6 months after transplantation, and these data from the Tricontinental Study have already been reported (2, 13).
Our results are comparable with those reported by others. The Tricontinental MMF Study found no difference in graft survival at 3 years (13); the original US MMF study reported no difference in graft or patient survival at 3 years (14); and a follow-up of the Mycophenolate Steroids Sparing study reported no overall benefit at 5 years (9). A pooled analysis of the three landmark MMF trials did not demonstrate any difference in graft survival but reported fewer rejection and slightly better (extrapolated) renal function at 12 months (15). Conversely, a more recent systematic review found that graft survival may be increased with MMF (16).
Several observational studies have reported no overall benefit of MMF compared with AZA (8, 10, 11), although Ojo et al. (17) reported improved long-term graft survival with MMF and Meier-Kriesche et al. reported that MMF is associated with fewer late rejection episodes (18) and improved kidney function (19). The era of the studies may be important because the relative benefit of MMF versus AZA seems to be related to the intensity of concomitant immunosuppression; this was demonstrated by the study of Knight et al. (16), wherein the improved outcomes with MMF seemed to be predominantly driven by the findings of studies using the Sandimmune formulation of cyclosporine rather than the microemulsion form of cyclosporine or tacrolimus. Most MMF versus AZA trials were conducted before the widespread use of tacrolimus and induction therapy with either anti-CD25 or T-cell–depleting antibodies.
With comparable efficacy, an important consideration when choosing between immunosuppressive agents is their adverse effect profile. An important long-term adverse effect of immunosuppression is malignancy, the risk of which is increased several-fold after transplantation (20). This study was unable to demonstrate a difference between MMF and AZA in cancer incidence. There did not seem to be a dose-response for MMF, although power for this comparison was limited.
Although estimated kidney function is not a hard endpoint, it is, nevertheless, a clinically important outcome. Mycophenolate mofetil was found to be protective in several experimental models of chronic allograft injury. Mycophenolate mofetil potently inhibited the development of proteinuria, cellular infiltration, interstitial fibrosis, and tubular atrophy in the F344-to-Lewis rat model of chronic kidney allograft damage (6). In an aortic allograft model, MMF attenuated adventitial inflammation, intimal thickening, cellularity, and smooth muscle cell proliferation (7). Such data, combined with pilot data from human studies (21), underscored the potential for MMF to prevent chronic allograft injury and thereby improve long-term outcomes after kidney transplantation. The lack of an impact of treatment group on graft survival, mean eGFR, or the slope of eGFR against time is therefore an important finding. We used the abbreviated modification of diet in renal disease (MDRD) formula to estimate kidney function (22). Although this formula has not been rigorously validated in kidney transplant patients, alternative formulas are likely to be at least as inaccurate in this patient group (23), and it is likely that any inaccuracies would occur equally in each group. ANZDATA does not collect blood urea nitrogen or serum albumin, preventing use of the six-variable MDRD formula (24) or the formula by Nankivell et al. (25).
In summary, despite MMF’s superiority over AZA in preventing acute rejection early after kidney transplantation (15), our study is consistent with, and extends, most published studies in failing to demonstrate any definite long-term differences between MMF and AZA in the hard outcomes of patient death or graft loss. As with other immunosuppressive agents in transplantation, the optimal choice between MMF and AZA is likely to require tailoring therapy to each patient’s immune risk and reproductive status, and the adverse effect profile, risk of drug interactions, cost, and overall safety of each agent. Because registries provide the only feasible way of monitoring the very long-term outcomes of clinical trials, future trial protocols should ideally include a mechanism whereby trial participants can prospectively be identified in the relevant registry to enable future analyses.
MATERIALS AND METHODS
The Tricontinental Mycophenolate Mofetil Renal Transplantation Study was a double-blind randomized placebo-controlled trial of MMF versus AZA, used in combination with cyclosporine and steroids, for deceased donor kidney transplantation and has been described in detail elsewhere (2). In brief, 503 adults receiving a deceased donor kidney transplant were randomized in equal groups to receive AZA (100 mg daily if body weight <75 kg; otherwise, 150 mg daily), 2 g/d MMF, or 3 g/d MMF. Randomization was stratified by graft number and center. All patients were also treated with cyclosporine and corticosteroids; steroid withdrawal was permitted after 6 months in patients with stable condition. Patients did not receive antibody induction therapy. Recruitment was during 1992 to 1993, and blinding was maintained for 3 years. The primary endpoint was treatment failure defined as biopsy-proven acute rejection, graft loss, patient death, or discontinuation of the study drug within 6 months of transplantation. The study was conducted in Australia, Canada, and Europe, and in this analysis, we report the long-term outcomes of the 133 Australian participants (26% of the total cohort).
The ANZDATA Registry collects deidentified information on consenting (>99%) patients treated with dialysis or kidney transplantation in Australia and New Zealand. The Australian participants of the Tricontinental Study were prospectively flagged at trial entry, and the relevant information regarding donor and recipient baseline characteristics, transplant characteristics, and long-term recipient outcomes were obtained from the registry.
Continuous variables are reported as mean (standard deviation) and were compared using analysis of variance or the Kruskal-Wallis test as appropriate. Categorical variables are reported as proportions and compared using chi-square or Fisher exact test as appropriate. Crossover (switching from MMF to AZA or vice versa) rates were calculated using the Kaplan-Meier method; stopping one antimetabolite without commencing the other was not considered a crossover.
The primary outcomes for this analysis were patient and graft survival; graft failure was defined as death or permanent loss of graft function. Secondary outcomes were death-censored graft survival, cancer incidence, and estimated kidney function. All comparisons were made on an intention-to-treat basis. Patient and graft survival were compared using the Kaplan-Meier method and log-rank test. Univariate Cox regression was used to calculate CIs for HRs. A secondary as-treated analysis compared graft and death-censored graft survival in a Cox regression model, with MMF versus AZA modeled as a time-varying binary covariate. For graft survival, separate models were constructed for the first year and subsequent years owing to nonproportional hazards.
ANZDATA records all cancers in patients undergoing dialysis and transplantation, and cancer reporting to the registry was recently validated (26). Comparison of cancer incidence was performed separately for non–skin and skin cancers and used a competing risks regression (27) to account for the competing risk of death before cancer diagnosis.
Serum creatinine is reported to ANZDATA at 1, 2, 3, and 6 months; at 1, 2, 3, 5, 7, and 10 years; and then every 5 years until the graft fails. The eGFR at each time point was calculated with the abbreviated MDRD formula (22), then compared using linear mixed models, taking into account the repeated measures per patient and clustering by donor because some donors gave kidneys to separate trial recipients. In these analyses, we constructed 3-level models—eGFR (level 1), clustered by patient (level 2), clustered by donor (level 3). Random intercepts were used for donors (reflecting initial nephron mass) and patients (reflecting patient-level characteristics affecting the starting eGFR). For the analysis of slope of eGFR against time, an additional patient-level random slope was included, and the difference between slopes was tested as a treatment group-by-time interaction. An unstructured covariance matrix was used for the latter analysis. As a sensitivity analysis, we repeated the analysis of eGFR slope after imputing an eGFR of 0 at the time of graft failure and additionally conducted a random-effects tobit analysis (incorporating a random intercept for patients) with left censoring of eGFR at 10 mL or less per minute per 1.73 m2. All analyses were conducted in Stata version 11.2 (StataCorp LP, College Station, TX).
The authors thank the Australian and New Zealand renal units, patients, and staff for their cooperation and contributions to ANZDATA.
1. Placebo-controlled study of mycophenolate
mofetil combined with cyclosporin and corticosteroids for prevention of acute rejection. European Mycophenolate
Mofetil Cooperative Study Group. Lancet 1995; 345: 1321.
2. A blinded, randomized clinical trial of mycophenolate
mofetil for the prevention of acute rejection in cadaveric renal transplantation
. The Tricontinental Mycophenolate
Mofetil Renal Transplantation
Study Group. Transplantation
1996; 61: 1029.
3. Sollinger HW. Mycophenolate
mofetil for the prevention of acute rejection in primary cadaveric renal allograft recipients. U.S. Renal Transplant Mycophenolate
Mofetil Study Group. Transplantation
1995; 60: 225.
4. Chadban S. Chapter 8: Transplantation
. ANZDATA Registry
Report 2002. Adelaide, Australia: Australia and New Zealand Dialysis and Transplant Registry
5. Meier-Kriesche H-U, Li S, Gruessner RWG, et al.. Immunosuppression
: evolution in practice and trends, 1994–2004. Am J Transplant 2006; 6: 1111.
6. Azuma H, Binder J, Heemann U, et al.. Effects of RS61443 on functional and morphological changes in chronically rejecting rat kidney allografts. Transplantation
1995; 59: 460.
7. Räisänen-Sokolowski A, Vuoristo P, Myllärniemi M, et al.. Mycophenolate
mofetil (MMF, RS-61443) inhibits inflammation and smooth muscle cell proliferation in rat aortic allografts. Transpl Immunol 1995; 3: 342.
8. Opelz G, Döhler B, Study CT. Influence of immunosuppressive regimens on graft survival and secondary outcomes after kidney transplantation
2009; 87: 795.
9. Remuzzi G, Cravedi P, Costantini M, et al.. Mycophenolate
mofetil versus azathioprine
for prevention of chronic allograft dysfunction in renal transplantation
: the MYSS follow-up randomized, controlled clinical trial. J Am Soc Nephrol 2007; 18: 1973.
10. Schold JD, Kaplan B. AZA/tacrolimus is associated with similar outcomes as MMF/tacrolimus among renal transplant recipients. Am J Transplant 2009; 9: 2067.
11. Shah S, Collett D, Johnson RJ, et al.. Long-term graft outcome with mycophenolate
mofetil and azathioprine
: a paired kidney analysis. Transplantation
2006; 82: 1634.
12. Lamb KE, Lodhi S, Meier-Kriesche H-U. Long-term renal allograft survival in the United States: a critical reappraisal. Am J Transplant 2011; 11: 450.
13. Mathew TH. A blinded, long-term, randomized multicenter study of mycophenolate
mofetil in cadaveric renal transplantation
: results at three years. Tricontinental Mycophenolate
Mofetil Renal Transplantation
Study Group. Transplantation
1998; 65: 1450.
mofetil in cadaveric renal transplantation
. US Renal Transplant Mycophenolate
Mofetil Study Group. Am J Kidney Dis 1999; 34: 296.
15. Halloran P, Mathew T, Tomlanovich S, et al.. Mycophenolate
mofetil in renal allograft recipients: a pooled efficacy analysis of three randomized, double-blind, clinical studies in prevention of rejection. The International Mycophenolate
Mofetil Renal Transplant Study Groups. Transplantation
1997; 63: 39.
16. Knight SR, Russell NK, Barcena L, et al.. Mycophenolate
mofetil decreases acute rejection and may improve graft survival in renal transplant recipients when compared with azathioprine
: a systematic review. Transplantation
2009; 87: 785.
17. Ojo AO, Meier-Kriesche H-U, Hanson JA, et al.. Mycophenolate
mofetil reduces late renal allograft loss independent of acute rejection. Transplantation
2000; 69: 2405.
18. Meier-Kriesche H-U, Steffen BJ, Hochberg AM, et al.. Long-term use of mycophenolate
mofetil is associated with a reduction in the incidence and risk of late rejection. Am J Transplant 2003; 3: 68.
19. Meier-Kriesche H-U, Steffen BJ, Hochberg AM, et al.. Mycophenolate
mofetil versus azathioprine
therapy is associated with a significant protection against long-term renal allograft function deterioration. Transplantation
2003; 75: 1341.
20. Webster A, Wong G, McDonald S. Chapter 10: Cancer Report. ANZDATA Registry
Report 2009. Adelaide, Australia: Australia and New Zealand Dialysis and Transplant Registry
21. Weir MR, Anderson L, Fink JC, et al.. A novel approach to the treatment of chronic allograft nephropathy. Transplantation
1997; 64: 1706.
22. Levey A, Greene T, Kusek J, et al.. A simplified equation to predict glomerular filtration rate from serum creatinine [abstract]. J Am Soc Nephrol 2000; 11: 155A.
23. Mariat C, Alamartine E, Barthelemy J-C, et al.. Assessing renal graft function in clinical trials: can tests predicting glomerular filtration rate substitute for a reference method? Kidney Int 2004; 65: 289.
24. Levey AS, Bosch JP, Lewis JB, et al.. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999; 130: 461.
25. Nankivell BJ, Gruenewald SM, Allen RD, et al.. Predicting glomerular filtration rate after kidney transplantation
1995; 59: 1683.
26. Webster AC, Supramaniam R, O’Connell DL, et al.. Validity of registry
data: agreement between cancer records in an end-stage kidney disease registry
(voluntary reporting) and a cancer register (statutory reporting). Nephrology (Carlton) 2010; 15: 491.
27. Fine J, Gray R. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999; 94: 496.
Mycophenolate; Azathioprine; Immunosuppression; Transplantation; Randomized controlled trial; Registry
Supplemental Digital Content
Copyright © 2012 Wolters Kluwer Health, Inc. All rights reserved.