Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Graft-Versus-Host Disease Prophylaxis after Transplantation: A Network Meta-Analysis

  • Panayiotis D. Ziakas,

    Affiliation Division of Infectious Diseases, Warren Alpert Medical School of Brown University, Providence, Rhode Island, United States of America

  • Fainareti N. Zervou,

    Affiliation Division of Infectious Diseases, Warren Alpert Medical School of Brown University, Providence, Rhode Island, United States of America

  • Ioannis M. Zacharioudakis,

    Affiliation Division of Infectious Diseases, Warren Alpert Medical School of Brown University, Providence, Rhode Island, United States of America

  • Eleftherios Mylonakis

    emylonakis@lifespan.org

    Affiliation Division of Infectious Diseases, Warren Alpert Medical School of Brown University, Providence, Rhode Island, United States of America

Graft-Versus-Host Disease Prophylaxis after Transplantation: A Network Meta-Analysis

  • Panayiotis D. Ziakas, 
  • Fainareti N. Zervou, 
  • Ioannis M. Zacharioudakis, 
  • Eleftherios Mylonakis
PLOS
x

Abstract

Background

Graft-versus-host Disease (GvHD) prophylaxis after allogeneic hematopoietic stem-cell transplantation (HSCT) is an ongoing effort but relative effects of different policies are not systematically explored.

Methods

We systematically reviewed 30-year evidence on GvHD prophylaxis and quantified the relative effect of different policies using a network meta-analysis. We searched PubMed and the Cochrane Library for randomized studies on the topic. The primary outcome of interest was grade II-IV acute GvHD over 0 or I (with odds ratio OR <1 denoting benefit).

Findings

Thirty-three eligible studies that enrolled 3,440 patients (published up to June 2014), provided data on seven immunosuppressive drugs namely cyclosporin A (CsA), methotrexate (MTX), anti-thymocyte globulin (ATG), mycophenolate mofetil (MMF), tacrolimus, sirolimus or corticosteroids and their combinations to calculate 14 direct and 21 indirect effects. The majority of trials (32/33) referred to myeloablative conditioning and sibling transplants (25/33). Tacrolimus/MTX (OR 0.44; 95% 0.27–0.70, number needed to treat to benefit, i.e. to avert a case of II-IV GvHD, NNTB = 5) and ATG/CsA/MTX (OR 0.45; 95%CI 0.26–0.78; NNTB = 5) were superior over CsA/MTX. ATG/CsA/MTX did not differ from tacrolimus/MTX (indirect evidence). Sirolimus-based prophylaxis outperformed CsA/MTX (OR 0.10; 95%CI 0.02–0.49, NNTB = 4) and marginally outperformed tacrolimus/MTX (OR 0.22; 95%CI 0.05–1.11). Add-on corticosteroids had no benefit over CsA/MTX.

Conclusions

Tacrolimus/MTX and ATG/CsA/MTX were the outperformers over CsA/MTX, but sirolimus-based regimens showed also potential. More randomized data are needed for reduced-intensity conditioning, as well as for MMF and sirolimus-containing regimens.

Introduction

The progress in the field of hematopoietic stem transplantation (HSCT) has resulted in a substantial rise in eligible patients and expanded therapeutic indications of HSCT. In 2010 only, over 12,000 patients received allogeneic transplant across Europe and approximately 7,000 in the US, figures reported by the European Group of Blood and Marrow Transplantation (EBMT) [1] and the Center for International Blood and Marrow Transplant Research (CIBMTR) [2], respectively. Despite the documented progress, graft versus host disease (GvHD) still remains an important constraint in allogeneic HSCT that partially hampers ongoing efforts to expand the pool of eligible candidates. Acute GvHD correlates inversely with both overall survival and treatment related mortality, and II-IV grade represents a clear cut-off in prognosis [3], [4].

Morbidity remains high, treatment is difficult and prevention strategies are far away from being considered optimal [5]. It was not until recently that the European Group for Blood and Marrow Transplantation and the European LeukemiaNet working group (EBMT-ELN) have published pertinent recommendations for GvHD, aiming to standardize prevention and treatment policies [5]. Optimization of prevention for GvHD remains an ongoing effort, as retrospective data analysis – even for data derived from randomized studies – suffers from substantial clinical heterogeneity between studies and inconsistencies of assigned pharmacologic interventions. In that context, we systematically reviewed pertinent randomized data, in order to summarize the relative effects of assigned protocols on GvHD prophylaxis using a network meta-analysis of direct and indirect comparisons.

Methods

We searched PubMed and The Cochrane Library databases for pertinent randomized trials. Last access was on June 13, 2014. The search terms were: “(GvHD OR graft versus host) AND (randomized OR randomised)”. We further scrutinized bibliography of eligible articles for additional studies on the topic. We complemented our search to include the American Society of Hematology (2004-2013) and the European Hematology Association (2006-2014) proceedings for additional randomized trials on the topic. Language restriction was not imposed. We followed the PRISMA guidelines (S1 Checklist in S1 Appendix).

A randomized trial on HSCT was deemed eligible provided that it met all the following conditions: (1) it randomized prophylactic schemes for GvHD, (2) reported acute GvHD as an outcome of interest, and, (3) randomized immunosuppressive drugs or drug combinations that are included in the recent EBMT-ELN working group consensus for a standardized practice in HSCT [5]. A trial was excluded from analysis if it had no extractable data on acute GvHD after prophylaxis, compared different dosing or formulations of the same pharmacologic agent, or used post hoc or historical arms for comparison. In case of follow-up, extension or overlapping studies, only the first published article was included. Studies outside the prophylactic setting, such as upfront or salvage therapies for acute GvHD were not considered.

Three reviewers (PDZ, IMZ and FNZ) screened titles and abstracts for relevance to the topic. All potentially relevant publications were independently evaluated in full text by the same authors. The following information was sought: first author, publication year, country of origin, sample size, median age, underlying condition, donor type, setting (myeloablative or reduced-intensity conditioning, RIC), conditioning regimens, total body irradiation, GvHD prophylaxis stratification and risk of acute GvHD. The primary outcome of interest was acute GvHD to day +100, dichotomized as II-IV grade over 0-I grading. We chose II-IV over 0-I because this stratification represents a clear-cut off in prognosis [3], [4] and the clinical cut-off to initiate GvHD treatment [5]. For completeness of the analysis we added III-IV as a secondary outcome.

The quality of individual studies was graded using the Cochrane Collaboration's tool for assessing the risk of bias, with the use of 5 pertinent items: random sequence generation and allocation concealment (for selection bias), blinding of participants and personnel (for performance bias), incomplete outcome (attrition bias), and selective reporting (for selection bias) [6].

The evidence synthesis is a network of comparisons, consisting of (a) pairwise direct effects of the prophylactic regimens, and, (b) indirect effects between two treatments against a common comparator [7]. For direct comparisons, the random-effects pooled Odds Ratio (OR), according to DerSimonian and Laird, was calculated and 95% precision estimates were reported [8]. We quantified statistical heterogeneity using the Cochran's Q test and I2 metric [9]. Evidence for small study effects was sought using the Harbord-Egger test [10].

We applied the Grading of Recommendations Assessment, Development and Evaluation (GRADE) classification criteria (accessible at: http://www.gradeworkinggroup.org/publications/JCE_series.htm) as previously described [11], to validate the quality of evidence regarding the risk of acute GvHD. The GRADE rating aims to explore the magnitude of confidence we have regarding our effect estimations, ranking them in descending order as high, moderate, low and very low [12]. In this context, when evidence is deemed as “high” quality, the authors are very confident that the estimates represent the true effect, as opposed to “very low” which implies that the true effect may actually differ from our estimates. The effects derived from randomized trials are initially ranked as “high” but may be downgraded in the presence of up to five limitations, namely the risk of bias, inconsistency, indirectness, imprecision and publication bias [13]. When treatments were not compared directly, effects were calculated using an indirect approach. Two competing strategies, A and B, are indirectly compared using the equation ln(ORAB) = ln(ORAC)−ln(ORBC), where C is an intermediate strategy with which both A and B are directly compared (the common comparator) [14][16]. The corresponding 95% CIs are computed assuming asymptotic normality and lack of covariance as described previously [14][17] Also, we reported the number needed to treat to benefit (NNTB) i.e. to avert a case of II-IV GvHD, and number needed to treat to harm (NNTH) i.e. to develop II-IV GvHD [18], [19]. Stata v13 (College Station, TX, USA) was used for data analysis and graphs of direct comparisons.

Results

Initial search yielded 769 citations from PubMed and 694 citations from the Cochrane Library. After title and abstract screening of 1,028 non-duplicate publications, 975 were excluded on the basis of relevance and a total of 53 articles were retrieved for full-text evaluation. Twenty one were not deemed eligible because they presented extension/overlapping data (n = 12), randomized drugs not advocated by the EBMT-ELN working group (n = 4) [5], or evaluated different dose/schedule of the same drug (n = 4). Also, one study was a matched-pair analysis (not truly randomized). No additional studies were found from conference proceedings.

Thirty two articles [20][51] remained eligible for final analysis on acute GvHD prevention, providing data 3,440 patients from 33 randomized studies (S1 Flow-Chart in S1 Appendix). The majority of studies (25/33, 76%) referred to sibling donors and myeloablative setting (32/33, 97%). Four studies (12%) presented data on both unrelated and sibling donors, and 4 studies (12%) on unrelated donors solely. Of note is that one study had 40% of patients receiving RIC and no study had randomized solely RIC transplants [25]. Bone marrow transplants were used in 26 studies, peripheral blood stem-cells in 2, and 5 enrolled patients receiving bone marrow or peripheral blood stem-cells.

The randomized interventions included the calcineurin inhibitors cyclosporine A (CsA) and tacrolimus, the antimetabolite methotrexate (MTX), the mTOR inhibitor sirolimus, the inosine monophosphate dehydrogenase inhibitor mycophenolate mofetil (MMF) as well as anti-thymocyte globulin (ATG) and corticosteroids. Systemic corticosteroids (Pse) referred to methylprednisolone, prednisolone and/or prednisone. Beclomethasone dipropionate (BDP) was examined separately due to the lack of systemic effects [22]. Study demographics, setting, randomized interventions and outcomes are summarized in Table 1 and the detailed study characteristics are shown on S1 Table S1 Appendix.

thumbnail
Table 1. Summary of included studies and stratified prophylaxis data on outcome.

https://doi.org/10.1371/journal.pone.0114735.t001

Quality Assessment

The quality assessment of individual studies is shown on S2 Table in S1 Appendix, stratified for the pairwise comparisons. All randomized studies were deemed of adequate quality (low or unclear risk of bias) to be included in the analysis. Across each stratum, the majority of studies had also low or unclear risk of bias, and all but one direct comparisons were considered free of serious limitation to warrant downgrading the level of evidence (S3 and S4 Tables in S1 Appendix).

Direct effects

Direct effects were derived from 14 pairwise comparisons. The network of comparisons is shown in Fig. 1 (solid lines). The CsA/MTX combination was compared over CsA monotherapy, MTX monotherapy, tacrolimus/MTX, ATG/CsA/MTX or Pse/CsA/MTX; CsA monotherapy was compared over MMF/CsA, MP/CsA or MTX monotherapy; ATG/MTX and Pse/ATG/MTX were compared over MTX monotherapy; Pse/CsA/MTX over Pse/CsA; tacrolimus/MTX was compared over tacrolimus/sirolimus (±MTX), tacrolimus/MMF or beclomethasone (BPD)/tacrolimus/MTX. The relative effects are shown in Fig. 2, and their grading of evidence on S3 and S4 Tables in S1 Appendix.

thumbnail
Figure 1. Network of direct (solid lines) and indirect comparisons (dashed lines) of different treatments evaluated for acute GvHD prophylaxis.

The thickness of connecting lines is proportional to the number of available direct comparisons.

https://doi.org/10.1371/journal.pone.0114735.g001

thumbnail
Figure 2. Forest plot of direct comparisons for the primary outcome (II-IV GvHD).

https://doi.org/10.1371/journal.pone.0114735.g002

The use of MTX over CsA monotherapy [40], [42], [44][46] had no significant impact on II-IV GvHD risk (OR 0.85; 95%CI 0.40–1.82; I2 = 55.2%), an effect derived exclusively from studies in the myeloablative setting with sibling donors and bone marrow transplants. The estimates did not alter (OR 0.76; 95%CI 0.29–2.03) after excluding the sole study exclusively on chronic myelogenous leukemia. MTX was inferior to CsA/MTX combination (OR 7.8; 95% CI 1.79–34.07, NNH = 3) in a single trial exclusively in aplastic anemia [41].

The use of tacrolimus/MTX over CsA/MTX [30], [32], [35] was associated with significant reduction of acute II-IV GvHD (OR 0.44; 95%CI 0.27–0.70, NNTB = 5; I2 = 45.1%), with high quality of evidence. All contributing studies used myeloablative conditioning, but differed in donor type. The association persisted (OR 0.35; 95%CI 0.19–0.63) after exclusion of [35] that showed imbalance in baseline disease characteristics (advanced or non-advanced malignancy) that potentially affected survival. The use of CSA monotherapy over CsA/MTX combination [28], [33], [36], [43], [47] significantly increased the risk of II-IV GvHD (OR 2.03; 95%CI 1.31–3.15, NNTH = 6), an effect consistent across studies (I2 = 0). The effect was derived from studies with comparable setting (myeloablative conditioning, bone marrow allografts and sibling donors), with the exception of one study that randomized solely aplastic anemia transplants [33].

A total of six studies examined the use of add-on corticosteroids to prophylaxis regimens, four as add-on to CsA/MTX combination [31], [34], [48], [49], one [22] as add-on to tacrolimus/MTX combination, and one as add-on to CsA [37]. Add-on corticosteroids to CsA/MTX (OR 0.73, 95%CI 0.34–1.57; I2 = 63%) showed no benefit regarding II-IV GvHD. The pooled effect was derived from four studies in myeloablative setting, using bone marrow transplants from sibling donors. One study enrolled exclusively adult patients [31] and the remaining three enrolled both pediatric and adult patients. After excluding [31], the effect remained insignificant (OR 1.11; 95%CI 0.69–1.78) and was consistent across studies (I2 = 0). Add-on corticosteroids to tacrolimus/MTX (OR 0.95;0.45–2.02) or to CsA monotherapy (OR 0.53; 95% 0.24–1.15) had no benefit regarding II-IV GvHD. It should be noted that one study [22] used oral beclomethasone propionate (BDP), which is regarded as having minimal systemic effects.

Six articles examined the use of add-on ATG to prophylaxis, four as add-on to CSA/MTX combination [24][26], [29] and the remaining two [50], [51] as add-on to MTX monotherapy. ATG add-on to CSA/MTX prevented acute GvHD (OR 0.45; 95%CI 0.26–0.78; NNB = 5; I2 = 30.1%), an association of high quality evidence (effect derived from four studies). Three out of 4 studies used rabbit ATG. The association did not alter after excluding the single study using equine ATG [26] (OR 0.41; 95%CI 0.19–0.89; I2 = 0%). ATG add-on to MTX monotherapy had no effect (OR 1.61; 95% CI 0.64–4.07), with the two consisting trials using equine ATG. Also, two trials on sirolimus/tacrolimus combinations for acute GvHD prophylaxis over tacrolimus/MTX [20], [21] suggested protection (OR 0.22; 95%CI 0.05–1.11, NNTB = 5, effect of marginal significance).

MMF-based combinations were not pooled to calculate direct effects [23], [27], as they were deemed highly heterogeneous with regard to donor type, donor source and assigned prophylaxis. MMF/tacrolimus was comparable to tacrolimus/MTX (OR 0.99; 95%CI 0.36–2.74) [23]; MMF/CsA was also comparable to CsA/MTX (OR 1.56;95%CI 0.44–5.53) [27]. The triple combination Pse/CsA/MTX was superior over Pse/CsA (OR 0.35; 0.13–0.89, NNTB = 7) [38] and Pse/ATG/MTX was superior to MTX monotherapy (OR 0.23; 95% CI 0.06–0.90, NNTB = 5) [39], both effects derived from single trials.

Direct effects for the secondary outcome (III-IV GvHD) are presented in Fig. 3. The tacrolimus/MTX regimen was superior to CsA/MTX (OR 0.62; 95%CI 0.41–0.95; NNTB = 12), an effect consistent across studies (I2 = 0). ATG add-on to CsA/MTX prevented III-IV GVHD (OR 0.39; 95%CI 0.17–0.88;I2 = 58%), an effect derived solely from studies using rabbit ATG. In subgroup analysis, this effect was more pronounced across the two studies that used ≥15 mg/kg total dose of ATG (OR 0.26; 95%CI 0.09–0.74). MTX was inferior to CsA/MTX while MMF/tacrolimus may have resulted in increased III-IV GvHD over tacrolimus/MTX. The remaining comparisons were not statistically significant (Fig. 3).

thumbnail
Figure 3. Forest plot of direct comparisons for the secondary outcome (III-IV GvHD).

https://doi.org/10.1371/journal.pone.0114735.g003

Indirect Effects

A total of 21 indirect comparisons were available for the primary outcome. The network of indirect comparisons is shown in Fig. 1 (dashed lines) and the relative effects are presented in Table 2. Nine treatment comparisons yielded a significant association. The bulk of indirect evidence highlights the inferiority of monotherapies (CsA or MTX) over double or triple combination regimens (Table 2). Moreover, substitution of CsA for ATG was inferior to CsA/MTX (OR 12.4;95%CI 2.19–70, NNTH = 2) and addition of Pse to ATG/MTX was superior to ATG/MTX (OR 0.14; 95% CI 0.03–0.74, NNTB = 3). Finally combining sirolimus/tacrolimus with or without MTX was superior over CsA/MTX combination (OR 0.10; 95% CI 0.02–0.49, NNTB = 4). ATG add-on to CsA/MTX was comparable to tacrolimus/MTX (OR 1.02; 95% CI 0.49–2.12).

thumbnail
Table 2. Summary of indirect effects for the primary outcome (II-IV GvHD).

https://doi.org/10.1371/journal.pone.0114735.t002

Discussion

We performed a meta-analysis of randomized trials to quantify the effect of different preventive policies in acute GvHD prophylaxis after HSCT. Direct and indirect evidence further underscored the inferiority of CsA or MTX monotherapies over combined prophylaxis. Monotherapies represent outdated practices and are of no significance in clinical practice. However, their inclusion in this analysis is mandated for reason of completeness and integrity of the analysis. The CsA/MTX combination was the most widely adopted prophylaxis used and was considered the standard of care. There was substantial direct evidence derived from the myeloablative setting, that CsA/MTX outperforms CsA monotherapy. The evidence was deemed to be of high quality and supported the recommendation of CsA/MTX as the standard prophylaxis in myeloablative setting by the EBMT-ELN working group [5].

Interestingly, we found direct evidence on the superiority of tacrolimus/MTX over the standard CsA/MTX regimen. In the context of these findings, tacrolimus/MTX should also be another preferred strategy. However, tacrolimus is not widely used over CsA across Europe and clinical experience is limited [52]. For example, across the 72 centers that responded to a recent survey tacrolimus/MTX was ranked fourth (used merely in 5% of centers that responded to the survey), well below CsA/MTX (87%), CsA-MMF (11%) and CsA monotherapy (7%) [52]. The EBMT-ELN working group noted that the lack of experience in Europe precludes a firm recommendation; nevertheless tacrolimus/MTX is considered as equivalent alternative to CsA/MTX. A recent registry data analysis further supports the preferential use of tacrolimus/MTX over CsA/MTX-based regimens for siblings (adjusted OR 0.65;95%CI 0.53–0.80) or unrelated donors (adjusted OR 0.79;95% 0.67–0.94) [53]. It appears that European centers, most of which have extensive experience in using CsA-based prophylaxis, are reluctant to substitute CsA with tacrolimus. However, within the context of available evidence that change of policy may be warranted.

We also found high quality evidence that ATG add-on to CsA/MTX prophylaxis significantly reduced the risk of GvHD grade II-IV and may further justify the inclusion of ATG for unrelated donor transplantation [5]. A previous Cochrane review [54] has also demonstrated a significant decline in II-IV GvHD (risk ratio 0.68; 95% CI 0.55 to 0.85) after pooling all the pertinent studies, but no subgrouping was performed on the referent prophylaxis (MTX or CsA/MTX). As our data show, ATG has significant benefit as add-on to CsA/MTX, but no effect as add-on to MTX monotherapy. ATG add-on to CsA/MTX and tacrolimus/MTX (over the standard CsA/MTX) also demonstrated significant benefit regarding III-IV GvHD prevention. Moreover, indirect evidence suggested that ATG add-on to CsA/MTX is comparable to tacrolimus/MTX. Consequently, ATG/CsA/MTX should be regarded as equivalent to tacrolimus/MTX in terms of GvHD prevention. ATG use, as opposed to tacrolimus use, is popular across European centers [52] but was not widely adopted across American centers [55]. If ATG add-on is selected, caution on the policy of a center regarding CMV prevention is warranted, as ATG is an independent risk factor for CMV reactivation [56].

Of note is that sirolimus plus tacrolimus combination may outperform tacrolimus/MTX (direct evidence) and CsA/MTX (indirect evidence). However, the risk of post-transplant thrombotic microangiopathy (TMA) and sinusoidal obstruction syndrome [57], [58] are a concern and sirolimus dosage adjustment is strongly warranted [57]. The risk of sinusoidal obstruction syndrome further increases with busulfan-based conditioning regimens [58]. Nevertheless, sirolimus/tacrolimus combinations arise as potential alternatives to standard regimens, provided that the potential harms are taken into account.

Importantly, direct effects did not support the inclusion of add-on corticosteroids to the prophylaxis regimen, with overall effect pointing to insignificant decline. Corticosteroids are the mainstay of care for established acute GvHD, and early initiation is warranted at a dose of 2 mg/kg methylprednisolone upon clinical symptoms and signs (grade II or higher) by both randomized data [59] and the expert opinion [5]. However, it should be emphasized that the individual studies are clinically heterogeneous with respect to timing of corticosteroid administration, CsA dose and folinic acid rescue [31], [34], [48]. Moreover, corticosteroid add-on has resulted in late occurrence of acute GvHD [31], [48] and intensification of immunosuppression may lead to increased risk of relapse as shown in previous studies [60], [61]. The risks of corticosteroid treatment that include hypertension, infections or avascular bones necrosis may outweigh any potential benefits, and EBMT registry data have shown that such an aggressive approach may result in more deaths from infection and graft failures and reduced overall survival [62]. The lack of significant effects in preventing GvHD coupled with the risks of untoward effects and toxicities underscore that systemic corticosteroids should be not considered as part of the initial regimen for GvHD prophylaxis.

The effectiveness of MMF-based regimens remains an unresolved issue given that few randomized studies have compared MMF-based regimens in myeloablative allo-HSCT. The individual study data suggested the non-inferiority these MMF-based regimens but treatment comparators and donors differed across studies and data could not be pooled [23],[27]. We had no randomized trials on MMF-based prophylaxis in the RIC setting, but it should be noted that recent non-randomized data found no difference in II-IV aGVHD rates after RIC, between patients receiving MMF/CsA (38%) compared to CsA/MTX prophylaxis (33%, p = 0.5) [63]. The use of MMF/CsA is the most accepted prophylaxis in RIC across Europe, given the sustained complete chimerism and prolonged remission achieved after RIC [64].

Our findings are consistent with previous meta-analyses on the topic. More specifically, Ram et al. [65] documented the superiority of tacrolimus/MTX over CsA/MTX and the inferiority of CsA monotherapy over CsA/MTX. The authors found no effect from adding corticosteroids to prophylaxis. Also, Theurich et al. [54] noted a significant reduction of II-IV when ATG was added for GvHD prophylaxis, as we did. The consistency of our associations with previous studies strengthens the validity of our data. Moreover, our analysis provides some additional insights by finely stratifying the referent arms (this analysis permits relative effects to be derived for single agents and combinations) and provides relative effects for treatments that are not examined directly (indirect comparisons). In this context, our analysis is a comprehensive network of comparisons that provides –within the limits of the methodology- the best available evidence.

Our analysis did not address longitudinal outcomes, including relapse rates, non-relapse mortality and overall survival. Variability in the duration of follow-up, censored observations, lack of uniform reporting and between-study clinical heterogeneity (with regard to conditioning regimens, the use of TBI, BM/PBSCT transplantation and underlying conditions- all of which affect these outcomes [53], [66][69]), preclude a robust comparison of long-term effects.

A number of limitations should be noted. Clinical heterogeneity remains an important constraint. Meta-analysis cannot account for confounding variables and the reported relative effects are unadjusted for potentially influential covariates. Consequently, adjusting the effects for age, different conditioning regimens, the use of TBI, different dosing and way of administration of the same drug is not feasible. Additionally, the impact of differences in supportive therapy could not be addressed. Second, performance bias is a concern since, in most studies, arm assignment was not blinded to investigators, physicians or patients. Third, outcomes may alter depending on if GvHD was a primary or a secondary outcome in a contributing study. Fourth, the analysis refers to a single aspect of immunosuppressive regimen selection, that is the risk of GvHD. Other aspects that may influence the choice of immunosuppressive drugs such as age, donor type, degree of matching, and preparative regimens could not be accounted for. For example, T-cell depletion limits the problem of GvHD at the expense of relapse, whereas more effective regimens may have high mortality from infection negating the benefit of reduced GvHD. Fifth, III-IV GvHD was an under-reported outcome across studies, as 27% of trials did not report the pertinent data. The field of transplant medicine is moving towards individualizing treatment and this study adds to our knowledge, providing a framework of relative effects for regimens directly compared (direct effects) as well for those that were never compared in a randomized trial (indirect effects). This matrix of effects can serve as a basis for further research on the topic.

In summary, we systematically assessed the effects of GvHD prevention policies after transplantation. The direct and indirect evidence accumulated from randomized control trials ranked tacrolimus/MTX and ATG/CsA/MTX as outperformers of prevention policies, noted the potency of tacrolimus/sirolimus combination, verified the inferiority of monotherapies and highlighted the paucity of randomized data in specific settings, including non-myeloablative transplant recipients.

Supporting Information

S1 Appendix.

The appendix includes the following items: S1 Checklist. PRISMA checklist. S1 Flow-Chart. Flow-chart of study selection process. S1 Table. Characteristics of included studies. S2 Table. Quality assessment of the individual studies graded using the Cochrane Collaboration's tool, stratified by pairwise comparison: (a). MTX vs. CsA (b). MTX vs. CsA/MTX (c). Tacrolimus/MTX vs. CsA/MTX (d). CsA vs. CsA/MTX (e). Pse/Csa/MTX vs. CsA/MTX (f). Beclomethasone/tacrolimus/MTX vs. tacrolimus/MTX (g) Pse/CsA vs.CsA (h) ATG/CsA/MTX vs. CsA/MTX (i)ATG/MTX vs.MTX (j).Sirolimus/Tacrolimus(MTX) vs. Tacrolimus/MTX (k). MMF/Tacrolimus vs. Tacrolimus/MTX (l). MMF/CsA vs. CsA/MTX (m). Pse/CsA/MTX vs. Pse/CsA (n).Pse/ATG/MTX vs. MTX. S3 Table. GRADE Evidence Profile (EP) on the relative effects of pharmacologic prophylaxis on the risk of II-IV GvHD. S4 Table. GRADE Summary of findings (SoF) table on the relative effects of pharmacologic prophylaxis on the risk of II-IV GvHD.

https://doi.org/10.1371/journal.pone.0114735.s001

(DOCX)

Author Contributions

Conceived and designed the experiments: PDZ EM. Analyzed the data: PDZ. Contributed reagents/materials/analysis tools: PDZ IMZ FNZ. Wrote the paper: PDZ IMZ FNZ EM. Performed literature search and collected data: PDZ FNZ IMZ.

References

  1. 1. Passweg JR, Baldomero H, Gratwohl A, Bregni M, Cesaro S, et al. (2012) The EBMT activity survey: 1990–2010. Bone Marrow Transplant 47:906–923.
  2. 2. Pasquini MC, Wang Z, Horowitz MM, Gale RP (2010) 2010 report from the Center for International Blood and Marrow Transplant Research (CIBMTR): current uses and outcomes of hematopoietic cell transplants for blood and bone marrow disorders. Clin Transpl: 87–105.
  3. 3. Gratwohl A, Hermans J, Apperley J, Arcese W, Bacigalupo A, et al. (1995) Acute graft-versus-host disease: grade and outcome in patients with chronic myelogenous leukemia. Working Party Chronic Leukemia of the European Group for Blood and Marrow Transplantation. Blood 86:813–818.
  4. 4. Dignan FL, Clark A, Amrolia P, Cornish J, Jackson G, et al. (2012) Diagnosis and management of acute graft-versus-host disease. Br J Haematol 158:30–45.
  5. 5. Ruutu T, Gratwohl A, de Witte T, Afanasyev B, Apperley J, et al. (2014) Prophylaxis and treatment of GVHD: EBMT-ELN working group recommendations for a standardized practice. Bone Marrow Transplant 49:168–173.
  6. 6. Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, et al. (2011) The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 343:d5928.
  7. 7. Ziakas PD, Kourbeti IS, Mylonakis E (2014) Systemic antifungal prophylaxis after hematopoietic stem cell transplantation: a meta-analysis. Clin Ther 36: 292–306 e291.
  8. 8. DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177–188.
  9. 9. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 327:557–560.
  10. 10. Harbord RM, Egger M, Sterne JA (2006) A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints. Stat Med 25:3443–3457.
  11. 11. Kourbeti IS, Ziakas PD, Mylonakis E (2014) Biologic therapies in rheumatoid arthritis and the risk of opportunistic infections: a meta-analysis. Clin Infect Dis 58:1649–1657.
  12. 12. Guyatt GH, Oxman AD, Kunz R, Vist GE, Falck-Ytter Y, et al. (2008) What is "quality of evidence" and why is it important to clinicians? BMJ 336:995–998.
  13. 13. Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, et al. (2011) GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol 64:383–394.
  14. 14. Hadjigeorgiou GM, Doxani C, Miligkos M, Ziakas P, Bakalos G, et al. (2013) A network meta-analysis of randomized controlled trials for comparing the effectiveness and safety profile of treatments with marketing authorization for relapsing multiple sclerosis. J Clin Pharm Ther 38:433–439.
  15. 15. Lumley T (2002) Network meta-analysis for indirect treatment comparisons. Stat Med 21:2313–2324.
  16. 16. Bucher HC, Guyatt GH, Griffith LE, Walter SD (1997) The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials. J Clin Epidemiol 50:683–691.
  17. 17. Caldwell DM, Ades AE, Higgins JP (2005) Simultaneous comparison of multiple treatments: combining direct and indirect evidence. BMJ 331:897–900.
  18. 18. Laupacis A, Sackett DL, Roberts RS (1988) An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 318:1728–1733.
  19. 19. Chatellier G, Zapletal E, Lemaitre D, Menard J, Degoulet P (1996) The number needed to treat: a clinically useful nomogram in its proper context. BMJ 312:426–429.
  20. 20. Pulsipher MA, Langholz B, Wall DA, Schultz KR, Bunin N, et al. (2014) The addition of sirolimus to tacrolimus/methotrexate GVHD prophylaxis in children with ALL: a phase 3 Children's Oncology Group/Pediatric Blood and Marrow Transplant Consortium trial. Blood 123:2017–2025.
  21. 21. Pidala J, Kim J, Jim H, Kharfan-Dabaja MA, Nishihori T, et al. (2012) A randomized phase II study to evaluate tacrolimus in combination with sirolimus or methotrexate after allogeneic hematopoietic cell transplantation. Haematologica 97:1882–1889.
  22. 22. Martin PJ, Furlong T, Rowley SD, Pergam SA, Lloid M, et al. (2012) Evaluation of oral beclomethasone dipropionate for prevention of acute graft-versus-host disease. Biol Blood Marrow Transplant 18:922–929.
  23. 23. Perkins J, Field T, Kim J, Kharfan-Dabaja MA, Fernandez H, et al. (2010) A randomized phase II trial comparing tacrolimus and mycophenolate mofetil to tacrolimus and methotrexate for acute graft-versus-host disease prophylaxis. Biol Blood Marrow Transplant 16:937–947.
  24. 24. Finke J, Bethge WA, Schmoor C, Ottinger HD, Stelljes M, et al. (2009) Standard graft-versus-host disease prophylaxis with or without anti-T-cell globulin in haematopoietic cell transplantation from matched unrelated donors: a randomised, open-label, multicentre phase 3 trial. Lancet Oncol 10:855–864.
  25. 25. Bacigalupo A, Lamparelli T, Milone G, Sormani MP, Ciceri F, et al. (2010) Pre-emptive treatment of acute GVHD: a randomized multicenter trial of rabbit anti-thymocyte globulin, given on day+7 after alternative donor transplants. Bone Marrow Transplant 45:385–391.
  26. 26. Champlin RE, Perez WS, Passweg JR, Klein JP, Camitta BM, et al. (2007) Bone marrow transplantation for severe aplastic anemia: a randomized controlled study of conditioning regimens. Blood 109:4582–4585.
  27. 27. Bolwell B, Sobecks R, Pohlman B, Andresen S, Rybicki L, et al. (2004) A prospective randomized trial comparing cyclosporine and short course methotrexate with cyclosporine and mycophenolate mofetil for GVHD prophylaxis in myeloablative allogeneic bone marrow transplantation. Bone Marrow Transplant 34:621–625.
  28. 28. Lee KH, Choi SJ, Lee JH, Kim S, Seol M, et al. (2004) Cyclosporine alone vs cyclosporine plus methotrexate for post-transplant immunosuppression after HLA-identical sibling bone marrow transplantation: a randomized prospective study. Bone Marrow Transplant 34:627–636.
  29. 29. Bacigalupo A, Lamparelli T, Bruzzi P, Guidi S, Alessandrino PE, et al. (2001) Antithymocyte globulin for graft-versus-host disease prophylaxis in transplants from unrelated donors: 2 randomized studies from Gruppo Italiano Trapianti Midollo Osseo (GITMO). Blood 98:2942–2947.
  30. 30. Hiraoka A, Ohashi Y, Okamoto S, Moriyama Y, Nagao T, et al. (2001) Phase III study comparing tacrolimus (FK506) with cyclosporine for graft-versus-host disease prophylaxis after allogeneic bone marrow transplantation. Bone Marrow Transplant 28:181–185.
  31. 31. Ruutu T, Volin L, Parkkali T, Juvonen E, Elonen E (2000) Cyclosporine, methotrexate, and methylprednisolone compared with cyclosporine and methotrexate for the prevention of graft-versus-host disease in bone marrow transplantation from HLA-identical sibling donor: a prospective randomized study. Blood 96:2391–2398.
  32. 32. Nash RA, Antin JH, Karanes C, Fay JW, Avalos BR, et al. (2000) Phase 3 study comparing methotrexate and tacrolimus with methotrexate and cyclosporine for prophylaxis of acute graft-versus-host disease after marrow transplantation from unrelated donors. Blood 96:2062–2068.
  33. 33. Locatelli F, Bruno B, Zecca M, Van-Lint MT, McCann S, et al. (2000) Cyclosporin A and short-term methotrexate versus cyclosporin A as graft versus host disease prophylaxis in patients with severe aplastic anemia given allogeneic bone marrow transplantation from an HLA-identical sibling: results of a GITMO/EBMT randomized trial. Blood 96:1690–1697.
  34. 34. Chao NJ, Snyder DS, Jain M, Wong RM, Niland JC, et al. (2000) Equivalence of 2 effective graft-versus-host disease prophylaxis regimens: results of a prospective double-blind randomized trial. Biol Blood Marrow Transplant 6:254–261.
  35. 35. Ratanatharathorn V, Nash RA, Przepiorka D, Devine SM, Klein JL, et al. (1998) Phase III study comparing methotrexate and tacrolimus (prograf, FK506) with methotrexate and cyclosporine for graft-versus-host disease prophylaxis after HLA-identical sibling bone marrow transplantation. Blood 92:2303–2314.
  36. 36. Zikos P, Van Lint MT, Frassoni F, Lamparelli T, Gualandi F, et al. (1998) Low transplant mortality in allogeneic bone marrow transplantation for acute myeloid leukemia: a randomized study of low-dose cyclosporin versus low-dose cyclosporin and low-dose methotrexate. Blood 91:3503–3508.
  37. 37. Deeg HJ, Lin D, Leisenring W, Boeckh M, Anasetti C, et al. (1997) Cyclosporine or cyclosporine plus methylprednisolone for prophylaxis of graft-versus-host disease: a prospective, randomized trial. Blood 89:3880–3887.
  38. 38. Chao NJ, Schmidt GM, Niland JC, Amylon MD, Dagis AC, et al. (1993) Cyclosporine, methotrexate, and prednisone compared with cyclosporine and prednisone for prophylaxis of acute graft-versus-host disease. N Engl J Med 329:1225–1230.
  39. 39. Ramsay NK, Kersey JH, Robison LL, McGlave PB, Woods WG, et al. (1982) A randomized study of the prevention of acute graft-versus-host disease. N Engl J Med 306:392–397.
  40. 40. Deeg HJ, Storb R, Thomas ED, Flournoy N, Kennedy MS, et al. (1985) Cyclosporine as prophylaxis for graft-versus-host disease: a randomized study in patients undergoing marrow transplantation for acute nonlymphoblastic leukemia. Blood 65:1325–1334.
  41. 41. Storb R, Deeg HJ, Farewell V, Doney K, Appelbaum F, et al. (1986) Marrow transplantation for severe aplastic anemia: methotrexate alone compared with a combination of methotrexate and cyclosporine for prevention of acute graft-versus-host disease. Blood 68:119–125.
  42. 42. Biggs JC, Atkinson K, Gillett E, Downs K, Concannon A, et al. (1986) A randomized prospective trial comparing cyclosporine and methotrexate given for prophylaxis of graft-versus-host disease after bone marrow transplantation. Transplant Proc 18:253–255.
  43. 43. Storb R, Deeg HJ, Whitehead J, Appelbaum F, Beatty P, et al. (1986) Methotrexate and cyclosporine compared with cyclosporine alone for prophylaxis of acute graft versus host disease after marrow transplantation for leukemia. N Engl J Med 314:729–735.
  44. 44. Ringden O, Backman L, Lonnqvist B, Heimdahl A, Lindholm A, et al. (1986) A randomized trial comparing use of cyclosporin and methotrexate for graft-versus-host disease prophylaxis in bone marrow transplant recipients with haematological malignancies. Bone Marrow Transplant 1:41–51.
  45. 45. Storb R, Deeg HJ, Thomas ED, Appelbaum FR, Buckner CD, et al. (1985) Marrow transplantation for chronic myelocytic leukemia: a controlled trial of cyclosporine versus methotrexate for prophylaxis of graft-versus-host disease. Blood 66:698–702.
  46. 46. Torres A, Martinez F, Gomez P, Herrera C, Rojas R, et al. (1989) Cyclosporin A versus methotrexate, followed by rescue with folinic acid as prophylaxis of acute graft-versus-host disease after bone marrow transplantation. Blut 58:63–68.
  47. 47. Mrsic M, Labar B, Bogdanic V, Nemet D, Pavletic Z, et al. (1990) Combination of cyclosporin and methotrexate for prophylaxis of acute graft-versus-host disease after allogeneic bone marrow transplantation for leukemia. Bone Marrow Transplant 6:137–141.
  48. 48. Storb R, Pepe M, Anasetti C, Appelbaum FR, Beatty P, et al. (1990) What role for prednisone in prevention of acute graft-versus-host disease in patients undergoing marrow transplants? Blood 76:1037–1045.
  49. 49. Atkinson K, Biggs J, Concannon A, Dodds A, Young S, et al. (1991) A prospective randomised trial of cyclosporin and methotrexate versus cyclosporin, methotrexate and prednisolone for prevention of graft-versus-host disease after HLA-identical sibling marrow transplantation for haematological malignancy. Aust N Z J Med 21:850–856.
  50. 50. Doney KC, Weiden PL, Storb R, Thomas ED (1981) Failure of early administration of antithymocyte globulin to lessen graft-versus-host disease in human allogeneic marrow transplant recipients. Transplantation 31:141–143.
  51. 51. Weiden PL, Doney K, Storb R, Thomas ED (1979) Antihuman thymocyte globulin for prophylaxis of graft-versus-host disease. A randomized trial in patients with leukemia treated with HLA-identical sibling marrow grafts. Transplantation 27:227–230.
  52. 52. Ruutu T, van Biezen A, Hertenstein B, Henseler A, Garderet L, et al. (2012) Prophylaxis and treatment of GVHD after allogeneic haematopoietic SCT: a survey of centre strategies by the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant 47:1459–1464.
  53. 53. Jagasia M, Arora M, Flowers ME, Chao NJ, McCarthy PL, et al. (2012) Risk factors for acute GVHD and survival after hematopoietic cell transplantation. Blood 119:296–307.
  54. 54. Theurich S, Fischmann H, Shimabukuro-Vornhagen A, Chemnitz JM, Holtick U, et al. (2012) Polyclonal anti-thymocyte globulins for the prophylaxis of graft-versus-host disease after allogeneic stem cell or bone marrow transplantation in adults. Cochrane Database Syst Rev 9:CD009159.
  55. 55. Bacigalupo A (2005) Antithymocyte globulin for prevention of graft-versus-host disease. Curr Opin Hematol 12:457–462.
  56. 56. George B, Kerridge IH, Gilroy N, Huang G, Hertzberg MS, et al. (2012) A risk score for early cytomegalovirus reactivation after allogeneic stem cell transplantation identifies low-, intermediate-, and high-risk groups: reactivation risk is increased by graft-versus-host disease only in the intermediate-risk group. Transpl Infect Dis 14:141–148.
  57. 57. Shayani S, Palmer J, Stiller T, Liu X, Thomas SH, et al. (2013) Thrombotic microangiopathy associated with sirolimus level after allogeneic hematopoietic cell transplantation with tacrolimus/sirolimus-based graft-versus-host disease prophylaxis. Biol Blood Marrow Transplant 19:298–304.
  58. 58. Cutler C, Stevenson K, Kim HT, Richardson P, Ho VT, et al. (2008) Sirolimus is associated with veno-occlusive disease of the liver after myeloablative allogeneic stem cell transplantation. Blood 112:4425–4431.
  59. 59. Van Lint MT, Uderzo C, Locasciulli A, Majolino I, Scime R, et al. (1998) Early treatment of acute graft-versus-host disease with high- or low-dose 6-methylprednisolone: a multicenter randomized trial from the Italian Group for Bone Marrow Transplantation. Blood 92:2288–2293.
  60. 60. Weaver CH, Clift RA, Deeg HJ, Storb R, Appelbaum FR, et al. (1994) Effect of graft-versus-host disease prophylaxis on relapse in patients transplanted for acute myeloid leukemia. Bone Marrow Transplant 14:885–893.
  61. 61. Frassoni F, Bacigalupo A, Sessarego M, Gualandi F, Lamparelli T, et al. (1991) The dose of total body irradiation (TBI), and GvHD prophilaxis have a crucial role on relapse risk and survival after BMT for leukaemia. Bone Marrow Transplant 7 Suppl 2: 130.
  62. 62. Ruutu T, Hermans J, van Biezen A, Niederwieser D, Gratwohl A, et al. (1998) How should corticosteroids be used in the treatment of acute GVHD? EBMT Chronic Leukemia Working Party. European Group for Blood and Marrow Transplantation. Bone Marrow Transplant 22:614–615.
  63. 63. Pinana JL, Valcarcel D, Fernandez-Aviles F, Martino R, Rovira M, et al. (2010) MTX or mycophenolate mofetil with CsA as GVHD prophylaxis after reduced-intensity conditioning PBSCT from HLA-identical siblings. Bone Marrow Transplant 45:1449–1456.
  64. 64. Niederwieser D, Maris M, Shizuru JA, Petersdorf E, Hegenbart U, et al. (2003) Low-dose total body irradiation (TBI) and fludarabine followed by hematopoietic cell transplantation (HCT) from HLA-matched or mismatched unrelated donors and postgrafting immunosuppression with cyclosporine and mycophenolate mofetil (MMF) can induce durable complete chimerism and sustained remissions in patients with hematological diseases. Blood 101:1620–1629.
  65. 65. Ram R, Gafter-Gvili A, Yeshurun M, Paul M, Raanani P, et al. (2009) Prophylaxis regimens for GVHD: systematic review and meta-analysis. Bone Marrow Transplant 43:643–653.
  66. 66. Gratwohl A, Hermans J, Goldman JM, Arcese W, Carreras E, et al. (1998) Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation. Chronic Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Lancet 352:1087–1092.
  67. 67. Sorror ML, Maris MB, Storb R, Baron F, Sandmaier BM, et al. (2005) Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood 106:2912–2919.
  68. 68. Omer AK, Ziakas PD, Anagnostou T, Coughlin E, Kourkoumpetis T, et al. (2013) Risk factors for invasive fungal disease after allogeneic hematopoietic stem cell transplantation: a single center experience. Biol Blood Marrow Transplant 19:1190–1196.
  69. 69. Versluis J, Labopin M, Niederwieser D, Socie G, Schlenk RF, et al. (2014) Prediction of non-relapse mortality in recipients of reduced intensity conditioning allogeneic stem cell transplantation with AML in first complete remission. Leukemia.