CLINICAL RESEARCH www.jasn.org Meta-Analysis of Calcineurin-Inhibitor-Sparing Regimens in Kidney Transplantation Adnan Sharif, Shazia Shabir, Sourabh Chand, Paul Cockwell, Simon Ball, and Richard Borrows Renal Institute of Birmingham, Queen Elizabeth Hospital, Edgbaston, Birmingham, United Kingdom ABSTRACT Calcineurin-inhibitor-sparing strategies in kidney transplantation may spare patients the adverse effects of these drugs, but the efficacy of these strategies is unknown. Here, we conduct a meta-analysis to assess outcomes associated with reducing calcineurin inhibitor exposure from the time of transplanta- tion. We search Medline, Embase, and Cochrane Register of Controlled Trials for randomized controlled trials published between 1966 and 2010 that compared de novo calcineurin-inhibitor-sparing regimens to calcineurin-inhibitor-based regimens. In this analysis, we include 56 studies comprising data from 11337 renal transplant recipients. Use of the contemporary agents belatacept or tofacitinib, in combi- nation with mycophenolate, decreased the odds of overall graft failure (OR 0.61; 95% CI 0.39–0.96; P ϭ 0.03). Similarly, minimization of calcineurin inhibitors in combination with various induction and adjunc- tive agents reduces the odds of graft failure (OR 0.73; 95% CI 0.58–0.92; P ϭ 0.009). Conversely, the use of inhibitors of mammalian target of rapamycin (mTOR), in combination with mycophenolate, increases the odds of graft failure (OR 1.43; 95% CI 1.08–1.90; P ϭ 0.01). Calcineurin-inhibitor-sparing strategies are associated with less delayed graft function (OR 0.89; 95% CI 0.80–0.98; P ϭ 0.02), improved graft function, and less new-onset diabetes. The more contemporary protocols did not seem to increase rates of acute rejection. In conclusion, this meta-analysis suggests that reducing exposure to calcineurin inhibitors immediately after kidney transplantation may improve clinical outcomes. J Am Soc Nephrol 22: ●●●–●●●, 2011. doi: 10.1681/ASN.2010111160 Discovery of the immunosuppressive properties of ever, kidney function in the early period post trans- the calcineurin inhibitor (CNI) ciclosporin by Borel plantation is a potent determinant of subsequent in 1976,1 and its introduction to the clinical arena graft outcome,8 and, therefore, later studies focused by Calne in 1978,2 heralded a new era in kidney on reducing or completely eliminating CNIs from transplantation. Randomized controlled studies the time of transplantation itself, a strategy made from the early 1980s showed ciclosporin was asso- possible with the development of “non-nephro- ciated with either significant reductions in absolute toxic” immunosuppressants. acute rejection rates or more “benign” presenta- An ever increasing array of such agents may fa- CLINICAL RESEARCH tions of rejection compared with azathioprine, the cilitate reduced CNI exposure early post transplan- mainstay immunosuppressant hitherto.3–5 However, the intrinsic nephrotoxicity of ciclosporin Received November 14, 2010. Accepted June 1, 2011. became apparent in these early trials and is now well Published online ahead of print. Publication date available at established, persisting despite introduction of the www.jasn.org. alternative CNI tacrolimus,6 and so subsequent Correspondence: Dr. Richard Borrows, Renal Institute of Birming- studies attempted to reduce overall CNI exposure ham, Queen Elizabeth Hospital, Birmingham, UK, B15 2WB. Phone: while maintaining reduced rejection rates. Trials of 0044 121 3716099; Fax: 0044 121 6275747; E-mail: richard. the mid and late 1980s evaluated weaning CNIs [email protected]. months or years following transplantation.7 How- Copyright © 2011 by the American Society of Nephrology J Am Soc Nephrol 22: ●●●–●●●, 2011 ISSN : 1046-6673/2210-●●● 1 CLINICAL RESEARCH www.jasn.org tation. The 1990s saw the emergence of the antiproliferative two separate studies (one minimization; one avoidance with agents mycophenolate mofetil and the mammalian target of mTORI/mycophenolate) without double-counting any of the rapamycin inhibitor (mTORI), sirolimus. Post 2000, the im- participants. Study arms consisting of low intensity belatacept munosuppressive armamentarium (both in standard practice (as opposed to moderate intensity) and low dose tofacitinib (as and clinical trials) expanded to include the sirolimus analog, opposed to high dose) were selected for evaluation against everolimus; the anti-CD52 leuco-depleting antibody, alemtu- standard CNI exposure protocols, as future experience is likely zumab; the protein kinase C inhibitor, sotrastaurin (AEB071); to focus on these regimens. the lymphocyte sequestering agent, FTY 720; the janus kinase 3 In the intervention arm, examples of non-CNI immunosup- inhibitor, tofacitinib (CP-690,550); the CD28 co-stimulation pressants included sirolimus or everolimus (18 studies, nϭ3155), blocker, belatacept. belatacept (three studies, n ϭ 950), tofacitinib (CP-690550) (two CNI exposure in current clinical practice is lower than that studies, n ϭ 257), FTY720 (two studies, n ϭ 898), sotrastaurin employed historically; however, the safety and efficacy of re- (one study, n ϭ 142) and alemtuzumab induction (four stud- ducing CNI exposure from the time of transplantation has not ies, n ϭ 242). In the control arm 20 studies utilized tacrolimus been subjected to a full and robust data synthesis, with many as the maintenance CNI (n ϭ 3289) and 35 used ciclosporin protocols remaining experimental. The purpose of this sys- (n ϭ 7568), with one study53 incorporating both calcineurin tematic review and meta-analysis was, therefore, to evaluate inhibitors. The individual immunosuppressant regimens and the clinical outcomes associated with strategies designed to study lengths for all of the randomized controlled trials are improve allograft function/survival by reducing, avoiding or summarized in Table 1. delaying introduction of CNI. Graft Failure In the pooled analysis, no difference was identified between RESULTS standard and reduced CNI exposure regarding overall graft failure (OR 1.05 [95% CI 0.85–1.29], P ϭ 0.66, I2 ϭ 54%) or The results of the literature search are illustrated in Figure 1. death-censored graft failure (OR 1.11 [95% CI 0.89–1.38], P ϭ Fifty-six randomized clinical trials, providing data for 11,337 0.36, I2 ϭ 44%). However, significant interstudy heterogeneity renal transplant recipients were identified (Table 1), with the was evident and, therefore, further subgroup analyses were median end-of-study time point of 12 mo. On a JADAD scor- conducted. ing scale for study quality 19 studies scored 1/5, 15 studies No difference in overall graft failure (OR 1.51 [95% CI scored 2/5, and 18 studies scored 3/5 (four trials were not 0.91–2.50], P ϭ 0.11, I2 ϭ 80%) or death-censored graft failure scored due to being in abstract format). (OR 1.59 [95% CI 0.94–2.68], P ϭ 0.08, I2 ϭ 78%) was appar- Total CNI avoidance,3,5,9–38 CNI minimization,17,39–55 and ent when azathioprine or mycophenolate monotherapy was com- delayed introduction of CNI49,50,55–62 were investigated in 32 pared with CNI based regimens (11 studies, n ϭ 1896). However, (n ϭ 5791), 17 (n ϭ 4131), and 10 studies (n ϭ 1519) respec- death-censored graft failure due to acute rejection was more com- tively. Two studies50,55 investigated CNI delay followed by mon in the azathioprine or mycophenolate monotherapy arms minimization: to avoid “double counting” these were analyzed (OR 2.79 [95% CI 1.39 –5.61], P ϭ 0.004, I2 ϭ 65%). as “delay” studies initially, but if subgroup analyses were nec- The combination of mTORI and mycophenolate (16 essary (due to heterogeneity), then the same study was consid- studies, n ϭ 2688) was associated with increased overall ered separately in both the “minimization” and “delay” suba- graft failure (OR 1.43 [95% CI 1.08–1.90], P ϭ 0.01, I2 ϭ nalyses. One four-arm trial17 was suitable for consideration as 19%) (Figure 2) and death-censored graft failure (OR 1.59 [95% CI 1.12–2.25], P ϭ 0.009, I2 ϭ 5%) compared with CNI-based regimens. Similar results were seen when the anal- ysis was repeated comparing mTOR/mycophenolate versus low-dose ciclosporin rather than low-dose tacrolimus for the Symphony study: OR 1.35 [95% CI 1.02 to 1.79], I2 12%, P ϭ 0.03 and OR 1.40 [95% CI 0.98 to 1.99], I2 0%, P ϭ 0.07 for overall graft failure and death-censored graft failure respec- tively. No difference between groups for death-censored graft failure secondary to acute rejection was demonstrated (OR 1.56 [95% CI 0.57–4.25], P ϭ 0.39, I2 ϭ 0%). In contrast, the combination of mycophenolate with newer immunosuppressive agents (belatacept or tofacitinib) (five studies, n ϭ 1207) was associated with reduced overall graft failure (OR 0.61 [95% CI 0.39–0.96], P ϭ 0.03, I2 ϭ 0%) Figure 1. PRISMA flow diagram identifying de novo CNI sparing (Figure 3). No difference in death-censored graft failure rates trials for inclusion in meta-analysis. were observed (OR 0.77 [95% CI 0.46–1.31], P ϭ 0.34, I2 ϭ 2 Journal of the American Society of Nephrology J Am Soc Nephrol 22: ●●●–●●●, 2011 www.jasn.org CLINICAL RESEARCH Table 1. Data for selected randomized controlled trials Study (year) Intervention arm Control arm CNI sparing strategy Study length Andres (2009) IL2 ϩ CϩMMF ϩ P IL2 ϩ lowC ϩ MMF ϩ P DELAY ϩ MINIMISATION 6 months Andres (2009) IL2 ϩ TϩMMF ϩ shortP T ϩ MMF ϩ P DELAY 6 months Asberg (2006) IL2 ϩ MMF ϩ PCϩ MMF ϩ P AVOIDANCE 12 months Buchler (2007) ATG ϩ SϩMMF ϩ P ATG ϩ CϩMMF ϩ P AVOIDANCE 12 months Budde (2010)* AEB ϩ lowT ϩ P AEB ϩ TϩP MINIMISATION