Quality of Included Trials
Overall, the quality of the included trials was moderate. APOMYGRE, FDCC, and OPERA had a Jadad score of three, whereas OPTICEPT had a score of two. OPTICEPT did not report a method of allocation concealment. All four RCTs reported completeness of follow-up and an intention-to-treat (ITT) analysis.
Treatment failure was defined as any one of the following events (whichever was reached first): acute rejection, graft loss, death, or MMF discontinuation. Based on the four RCTs, the pooled proportion of KTRs reaching the composite endpoint of treatment failure was 28.1% (246 of 877) and 29.3% (257 of 878) over the follow-up period in the CD and FD groups, respectively. The meta-analysis showed that the incidence of treatment failure in the CD group was not different from that in the FD group (relative risk [RR], 0.95; 95% confidence interval [CI], 0.82–1.10; P=0.52; I2=44%; Fig. 2A).
Additionally, there was no significant between-group difference for any component of treatment failure (Table 3). There was significant heterogeneity among the trials for acute rejection (I2=65%). We first performed a sensitivity analysis by removing OPTICEPT, which had a lower quality score and used MPA trough concentrations to adjust MMF doses (the remaining three RCTs used limited sampling methods), but subsequent analysis showed that heterogeneity still existed (I2=77%). Then, we hypothesized that the heterogeneity may have been caused by APOMYGRE. First, its sample size was the smallest among the four RCTs (130 of 1755). Second, MMF dose adjustments were calculated by a computer program to reach the target MPA AUC, whereas, in the other RCTs, dosing adjustments were left to the discretion of the investigators. After removing APOMYGRE, there was no more heterogeneity (RR, 1.10; 95% CI, 0.90–1.35; P=0.35; I2=0%). However, there was little change in acute rejection after sensitivity analyses performed with and without APOMYGRE.
The common measure of graft function was estimated creatinine clearance, as reported in APOMYGRE and OPERA (377 KTRs). No significant difference was observed between the CD and the FD groups (weighted mean difference [WMD], 2.46; 95% CI, −1.15 to 6.07; P=0.18; I2=0%).
The serum creatinine level was also reported in APOMYGRE (130 KTRs), and again, no significant difference was observed between groups (WMD, −13.00; 95% CI, −30.46 to 4.46; P=0.14).
Gastrointestinal Adverse Events
The incidence of gastrointestinal adverse events (AEs) was reported in APOMYGRE (130 KTRs), with 24.6% (16 of 65) and 20.0% (13 of 65) in the CD and FD groups, respectively (RR, 1.23; 95% CI, 0.65–2.35; P=0.53).
FDCC, OPTICEPT, and OPERA (1624 KTRs) only recorded the incidence of diarrhea: CD versus FD, 28.9% (234 of 809) versus 27.0% (220 of 815; RR, 1.08; 95% CI, 0.92–1.25; P=0.35; I2=0%).
Hematologic Adverse Events
One or more hematologic AEs, including anemia, leucopenia, and thrombocytopenia, were recorded in the four RCTs. The pooled estimate of APOMYGRE, FDCC, and OPERA (1283 KTRs) showed that the overall incidence of anemia was 29.8% (191 of 641) in the CD group and 26.2% (168 of 642) in the FD group, which were not significantly different (RR, 1.13; 95% CI, 0.97–1.32; P=0.12; I2=0%).
The pooled estimate of the four RCTs (1754 KTRs) showed that the overall incidence of leukopenia was 20.4% (178 of 874) in the CD group and 18.3% (161 of 880) in the FD group, which were also not significantly different (RR, 1.12; 95% CI, 0.93–1.35; P=0.25; I2=23%).
The incidence of thrombocytopenia was reported in FDCC (901 KTRs), with 5.1% (23 of 449) and 6.4% (29 of 452) in the CD and FD groups, respectively (RR, 0.80; 95% CI, 0.47–1.36; P=0.41).
Data regarding total infections were reported in all four RCTs (1754 KTRs). The application of CD MMF significantly increased the risk of total infections compared with FD MMF (36.0% vs. 30.9%; RR, 1.16; 95% CI, 1.03–1.30; P=0.01), with low heterogeneity (I2=0%; Fig. 2B).
Bacterial infection, cytomegalovirus, herpes, and BK virus were the most commonly reported infections after kidney transplantation. Comparing CD with FD, bacterial infection (47.4% vs. 44.2%; RR, 1.07; 95% CI, 0.86–1.33; P=0.53; I2=0%), cytomegalovirus (15.5% vs. 13.8%; RR, 1.12; 95% CI, 0.82–1.53; P=0.49; I2=24%), and BK virus (3.0% vs. 3.4%; RR, 0.89; 95% CI, 0.33–2.43; P=0.83) were not significantly different, but herpes (9.4% vs. 3.7%; RR, 2.54; 95% CI, 1.09–5.95; P=0.03; I2=49%) was significantly different.
FDCC and OPTICEPT reported the incidence of malignancy after kidney transplantation. The pooled incidence of malignancy was 1.3% (9 of 682) in the CD group and 2.2% (15 of 690) in the FD group. A meta-analysis showed that CD MMF did not decrease the probability of posttransplantation malignancy (RR, 0.61; 95% CI, 0.27–1.38; P=0.23; I2=0%).
The cost-effectiveness of CD versus FD MMF was reported in the post hoc analysis of APOMYGRE (18). The mean total yearly cost per KTR was €47,477 in the CD group and €46,783 in the FD group (P=0.7). The three largest expenditure categories (in decreasing order) were hospitalization, outpatient medications (including CNIs and MMF), and transportation. The observed incremental cost-effectiveness ratio was €3757 per treatment failure. The cost for MMF TDM was €452 per KTR, which is less than 1% of the total cost.
Summary of Key Findings
Combined data from the four eligible RCTs involving 1755 KTRs showed that, compared with FD MMF, CD MMF increased the risk of opportunistic infection by approximately 16% without overall benefit to prevent treatment failure (acute rejection, graft loss, death, and MMF discontinuation), reduced the incidence of gastrointestinal AEs, hematologic AEs, or malignancy, or protected graft function. Our findings suggest that TDM for MMF cannot be recommended as routine practice for KTRs. This finding is contrary to transplant clinicians’ anticipation that CD MMF would result in a lower proportion of KTRs achieving treatment failure and MPA-associated AEs.
Only one previous systematic review (19) has addressed this topic, and it reported that the role of MMF TDM was uncertain. The majority of eligible studies were retrospective in nature, and only one RCT (APOMYGRE) was included. Such retrospective studies are unable to justify the adoption of a costly TDM regimen in everyday clinical practice. Identified studies involved adult or pediatric solid organ transplant recipients; thus, the patient population was much more heterogeneous, and an accurate comparison is difficult because the data on KTRs were analyzed together with those of liver and heart recipients. Clearly, the conclusions that can be drawn from these retrospective studies are limited.
Current Methodology of the Pharmacokinetic Measurement of Mycophenolic Acid Exposure
Due to its higher sensitivity and specificity, high-performance liquid chromatography, rather than enzyme multiplied immunoassay technique, is regarded as the gold standard technique for MPA concentration measurement (6). A full AUC0–12 hr measurement is the most accurate predictor of MPA exposure; however, it is laborious, costly, and clinically impractical (10). Conversely, single-point measurement is at the other extreme (4). A practical alternative is the use of limited samplings coupled with multiple regression or Bayesian estimation (11, 12). Barraclough et al. (20) compared the ability of different MPA AUC estimations to predict MPA exposure for KTRs and demonstrated that Bayesian estimation is slightly superior to trough and multiple regression predictions in correlation and accuracy and may be the preferable methodology.
Regarding limited sampling methods, no consensus sampling time points are regarded as optimal (21). In the four eligible RCTs, three sampling time points were used: 20 min and 1 and 3 hr in APOMYGRE and OPERA and 0, 30, and 120 min in FDCC and OPTICEPT. Optimizing TDM does not necessarily require a high frequency of AUC measurements. Monitoring is recommended every week during the first month after transplantation and every 1 to 3 months thereafter for up to 1 year; after 1 year, if graft function has stabilized, MPA exposure can be assessed each time the immunosuppressive regimen is changed or a potentially interacting medication is introduced or withdrawn (6, 21, 22).
Strengths and Limitations
Our systematic review identified all available RCTs evaluating CD versus FD MMF for KTRs. Our methodology was robust, searching all possible studies, even in the abstract form and non-English-language sources, and used a strict assessment of trial quality. To avoid heterogeneity, our study did not include trials in which enteric-coated mycophenolate sodium (EC-MPS) was investigated, because EC-MPS has different pharmacokinetic properties from MMF, and strategies to estimate MPA exposure from EC-MPS have not yet been developed (21). Consequently, the current study provides critical information to guide clinical decisions on TDM for MMF.
Nevertheless, there are potential caveats affecting the validity of our findings. First, the proportion of KTRs achieving the target therapeutic window was similar between the two groups throughout the follow-up period (CD vs. FD, 68.6% vs. 63.6%) in FDCC (with the largest sample size among the four RCTs, 901 of 1755) because transplant clinicians were sometimes reluctant to make dosage changes according to the results of the MPA AUC. The failure to achieve the target therapeutic window may have contributed to the inability to detect significant differences between the CD and FD MMF arms in efficacy outcome measures. Second, the four RCTs differed in how they evaluated MPA exposure. APOMYGRE and OPERA used limited sampling methods combined with Bayesian estimation to make dosing adjustments, FDCC used limited sampling methods combined with multiple regression, and OPTICEPT used trough concentration measurements. Given the different predictive values for MPA exposure, uniformity in efficacy and safety across trials could not be expected and was not observed. Third, all KTRs included in our study were receiving CNI-based immunosuppression regimens. If there are differences in the efficacy and safety of MMF TDM between CNI-treated and non-CNI-treated KTRs, then our findings may not apply to non-CNI-treated KTRs.
The relationship between MPA exposure and the risk of acute rejection in KTRs treated with MMF has been confirmed by numerous pharmacokinetic studies (8). CD MMF may overcome the problems of interpatient variability and time-dependent variation of MPA pharmacokinetics. Based on the four RCTs, the mean MPA AUC was higher in the CD group in the early posttransplantation period (<1 month) but generally was not maintained thereafter; furthermore, a higher proportion of KTRs had an MPA AUC more than 60 mg hr/L in the CD group. Thus, CD MMF administration does not translate into overall benefit to prevent treatment failure but is prone to develop posttransplantation infections. In the current clinical practice, routine MMF TDM cannot be recommended for KTRs. Our meta-analysis highlights selectively performing TDM for MMF only for the high-risk KTRs (22). It may be a cost-effective strategy. Unfortunately, limited data are available on the assessment of high-risk factors for deciding MMF TDM. The current consensus indications for MMF TDM in KTRs are delayed graft function, immunosuppressive protocols excluding induction therapy, steroids or CNIs, and CNI minimization (9). To better define subpopulations that might benefit, a post hoc exploratory analysis of FDCC and OPTICEPT was performed, revealing that other potential indications were body weight less than 50 or more than 100 kg, pediatric KTRs, second or third transplantation, panel-reactive antibodies more than 15%, four or more human leukocyte antigen mismatches, and black race (7, 23–26). The identification of risk factors eligible for MMF TDM poses a potentially useful field of research in clinical transplantation.
In summary, CD MMF based on TDM is, to some extent, prone to result in some AEs without overall benefit to prevent treatment failure. Therefore, TDM for MMF cannot be recommended as routine practice for KTRs, and it may be targeted toward high-risk KTRs, who must be defined in the future.
MATERIALS AND METHODS
The electronic databases Medline, Embase, and the Cochrane Library were searched from the start date of each resource up to June 2012 (date of literature search) with logical combinations of the following terms: kidney, renal transplantation, allograft, MMF, CellCept (Roche, Basel, Switzerland), MPA, FD, CD, AUC, TDM, limited to humans, and RCTs. No restriction of report language, age of KTRs, or immunosuppressive protocols was applied to the search strategy. In addition, the results were supplemented by searching clinical trial registries (including but not limited to WHO International Trial Registry Network, ClinicalTrials.gov, and Australian & New Zealand Clinical Trials Registry), bibliographies of reviews and identified RCTs, transplant textbooks, abstracts of transplant conference proceedings, and directly contacting the manufacturers of MMF (CellCept).
To be eligible, the following inclusion criteria had to be met: (a) the publication type was RCT, (b) participants were recipients of a single kidney allograft (first or repeat) from a living or deceased donor, (c) the immunosuppression regimen consisted of MMF, and (d) participants were randomly allocated to either the CD group or the FD group. To obtain a uniform study population and to minimize heterogeneity, participants who received another solid organ in addition to a kidney transplant were excluded. Participants who were randomized but subsequently did not receive a kidney transplant were considered justifiable exclusions from the ITT population (so-called “modified ITT”, which was put forward by the Centre for Evidence in Transplantation; http://www.transplantevidence.com/library_rct.php). When results from the same participants in a clinical trial were published more than once, the first complete publication was identified (the index publication) and used as the primary data source. However, any other additional publications that included additional outcomes of interest also contributed to the meta-analysis.
All citations identified by the search strategy were independently evaluated by X.W. and X.Q. using titles, abstracts, and, where required, the full text to identify eligible trials. Data extraction was independently conducted by the same two reviewers using a predesigned data extraction form, including study design, participant demographics, interventions, and outcome measures. Trial authors or sponsors were contacted for missing data. X.W. and X.Q. then met to combine their findings, and the information was subsequently entered into Review Manager (RevMan version 5.1.6; The Nordic Cochrane Centre, Cochrane Collaboration, Copenhagen). The primary treatment endpoint was the proportion of KTRs experiencing treatment failure, which was defined as any one of the following events (whichever was reached first): acute rejection, graft loss, death, or MMF discontinuation. Acute rejection was measured in three ways: biopsy-proven acute rejection, subclinical acute rejection at the protocol biopsy, and acute rejection diagnosed clinically without biopsy. The secondary treatment outcomes were graft function, gastrointestinal AEs, hematologic AEs, infections, malignancies, and cost-effectiveness.
Methodological Quality of Trials
The quality of the trials was independently assessed by X.W. and Y.W. based on the Jadad score (0–5) across four items: randomization, blinding, withdrawals, and dropouts (27, 28). A score of three or more is considered good quality. Additionally, allocation concealment and ITT analysis were assessed. Disagreements on data collection and quality assessment were resolved by consensus or, when necessary, by consulting T.L.
All statistical analyses were performed using RevMan version 5.1.6. P<0.05 was considered to be statistically significant. Dichotomous measures were expressed as RR, and continuous variables were expressed as WMD. All summary effects are presented with 95% CI. Meta-analysis was undertaken using the fixed- or random-effects models (depending on the absence or presence of heterogeneity) to estimate effect sizes for each outcome measure. Heterogeneity across RCTs was investigated via forest plots, Cochrane’s Q (P<0.1), and I2 statistics (I2>50%). If significant heterogeneity existed, a sensitivity analysis was additionally used to explore possible sources of heterogeneity. Publication bias was assessed by analyzing funnel plot asymmetry.
The authors thank the editors and reviewers for the constructive comments and suggestions that were of great help in improving the quality of article.
Guerra G, Ciancio G, Gaynor JJ, et al. Randomized trial of immunosuppressive regimens in renal transplantation. J Am Soc Nephrol 2011; 22: 1758.
Sunder-Plassmann G, Reinke P, Rath T, et al. Comparative pharmacokinetic study of two mycophenolate mofetil
formulations in stable kidney transplant recipients. Transpl Int 2012; 25: 680.
Gourishankar S, Houde I, Keown PA, et al. The CLEAR study: a 5-day, 3-g loading dose of mycophenolate mofetil
versus standard 2-g dosing in renal transplantation. Clin J Am Soc Nephrol 2010; 5: 1282.
Miura M, Niioka T, Kato S, et al. Monitoring of mycophenolic acid predose concentrations in the maintenance phase more than one year after renal transplantation. Ther Drug Monit 2011; 33: 295.
van Gelder T. Mycophenolate blood level monitoring: recent progress. Am J Transplant 2009; 9: 1495.
Saint-Marcoux F, Vandierdonck S, Premaud A, et al. Large scale analysis of routine dose adjustments of mycophenolate mofetil
based on global exposure in renal transplant patients. Ther Drug Monit 2011; 33: 285.
van Gelder T, Silva HT, de Fijter H, et al. How delayed graft function impacts exposure to mycophenolic acid in patients after renal transplantation. Ther Drug Monit 2011; 33: 155.
Glander P, Sommerer C, Arns W, et al. Pharmacokinetics and pharmacodynamics of intensified versus standard dosing of mycophenolate sodium in renal transplant patients. Clin J Am Soc Nephrol 2010; 5: 503.
Le Meur Y, Borrows R, Pescovitz MD, et al. Therapeutic drug monitoring
of mycophenolates in kidney transplantation
: report of The Transplantation Society consensus meeting. Transplant Rev (Orlando) 2011; 25: 58.
Staatz CE, Tett SE. Maximum a posteriori Bayesian estimation of mycophenolic acid area under the concentration-time curve: is this clinically useful for dosage prediction yet? Clin Pharmacokinet 2011; 50: 759.
Musuamba FT, Rousseau A, Bosmans JL, et al. Limited sampling models and Bayesian estimation for mycophenolic acid area under the curve
prediction in stable renal transplant patients co-medicated with ciclosporin or sirolimus. Clin Pharmacokinet 2009; 48: 745.
Zhao W, Fakhoury M, Baudouin V, et al. Limited sampling strategy for estimating individual exposure of tacrolimus in pediatric kidney transplant patients. Ther Drug Monit 2011; 33: 681.
Kuypers DR, Le Meur Y, Cantarovich M, et al. Consensus report on therapeutic drug monitoring
of mycophenolic acid in solid organ transplantation. Clin J Am Soc Nephrol 2010; 5: 341.
van Gelder T, Silva HT, de Fijter JW, et al. Comparing mycophenolate mofetil
regimens for de novo renal transplant recipients: the fixed-dose concentration-controlled trial. Transplantation 2008; 86: 1043.
Le Meur Y, Buchler M, Thierry A, et al. Individualized mycophenolate mofetil
dosing based on drug exposure significantly improves patient outcomes after renal transplantation. Am J Transplant 2007; 7: 2496.
Gaston RS, Kaplan B, Shah T, et al. Fixed- or controlled-dose mycophenolate mofetil
with standard- or reduced-dose calcineurin inhibitors: the Opticept trial. Am J Transplant 2009; 9: 1607.
Le Meur Y, Thierry A, Glowacki F, et al. Early steroid withdrawal and optimization of mycophenolic acid exposure in kidney transplant recipients receiving mycophenolate mofetil
. Transplantation 2011; 92: 1244.
Rousseau A, Laroche ML, Venisse N, et al. Cost-effectiveness analysis of individualized mycophenolate mofetil
dosing in kidney transplant patients in the APOMYGRE trial. Transplantation 2010; 89: 1255.
Knight SR, Morris PJ. Does the evidence support the use of mycophenolate mofetil therapeutic drug monitoring
in clinical practice? A systematic review. Transplantation 2008; 85: 1675.
Barraclough KA, Isbel NM, Staatz CE. Evaluation of the mycophenolic acid exposure estimation methods used in the APOMYGERE, FDCC, and Opticept trials. Transplantation 2010; 90: 44.
Tett SE, Saint-Marcoux F, Staatz CE, et al. Mycophenolate, clinical pharmacokinetics, formulations, and methods for assessing drug exposure. Transplant Rev (Orlando) 2011; 25: 47.
van Gelder T, Le Meur Y, Shaw LM, et al. Therapeutic drug monitoring
of mycophenolate mofetil
in transplantation. Ther Drug Monit 2006; 28: 145.
Kaplan B, Gaston RS, Meier-Kriesche HU, et al. Mycophenolic acid exposure in high- and low-weight renal transplant patients after dosing with mycophenolate mofetil
in the Opticept trial. Ther Drug Monit 2010; 32: 224.
Hocker B, van Gelder T, Martin-Govantes J, et al. Comparison of MMF efficacy and safety in paediatric vs. adult renal transplantation: subgroup analysis of the randomised, multicentre FDCC trial. Nephrol Dial Transplant 2011; 26: 1073.
van Gelder T, Tedesco Silva H, de Fijter JW, et al. Renal transplant patients at high risk of acute rejection benefit from adequate exposure to mycophenolic acid. Transplantation 2010; 89: 595.
van Agteren M, Armstrong VW, van Schaik RH, et al. AcylMPAG plasma concentrations and mycophenolic acid-related side effects in patients undergoing renal transplantation are not related to the UGT2B7-840G>A gene polymorphism. Ther Drug Monit 2008; 30: 439.
Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17: 1.
Pengel L, Morris P. The transplant library of randomized controlled trials and systematic reviews. Transplantation 2011; 92: 613.
Keywords:© 2013 by Lippincott Williams & Wilkins
Area under the curve; Kidney transplantation; Meta-analysis; Mycophenolate mofetil; Therapeutic drug monitoring