Prophylactic antithrombotic management in adult and pediatric kidney transplantation: A systematic review and meta‐analysis

RGT is a major cause for early graft loss after KTx. Although evidence‐based recommendations are lacking, aP is often used to prevent RGT. This systematic review aimed to determine the effectiveness and safety of aP in adult and pediatric KTx recipients.


| INTRODUC TI ON
Kidney transplant recipients are at risk for developing RGT. 1,2 RGT is characterized by thrombus formation in either the renal artery or vein of the renal allograft and predominantly occurs within the first days after KTx. 3 With a reported incidence of 0% to 6.1% in adults, and 0% to 13.0% in pediatric patients, RGT accounts for up to 10.0% of early graft failure. [4][5][6][7][8] In order to minimize the risk of renal graft loss due to thrombosis, aP is widely used in KTx. Current management strategies favor antiplatelet and anticoagulant drugs. 8,9 The clinical impact of aP for KTx recipients is still a matter of debate, primarily due to the paucity of available studies with partially conflicting results. [9][10][11] In the absence of evidence-based recommendations and consensus guidelines, current antithrombotic strategies are often characterized by incongruent, non-standardized, and center-specific protocols for aP. 12,13 Therefore, this systematic review and meta-analysis were performed to collate, summarize, and quantify the available evidence on beneficial and adverse effects of aP in adult and pediatric KTx recipients.

| Protocol
This systematic review was developed using guidance from the PRISMA statement and recommendations of the Cochrane Collaboration. 14, 15 The protocol was registered in the International Prospective Register of Systematic Reviews (http://www.crd.york. ac.uk/PROSPERO). 16,17

| Data sources
The electronic databases CENTRAL (Cochrane Central Register of Controlled Trials, Issue 4, 2020), MEDLINE, and EMBASE (from their inception to April 21, 2020) were searched without language restriction (search strategy in Table S1). Additionally, clinical trial registries (for ongoing or recently completed trials), available conference proceedings (Table S2), and reference lists of included studies and potentially relevant other articles were searched using predefined keywords. 17

| Eligibility criteria
All types of studies were included if they compared thromboprophylaxis with either placebo, no aP, or with different antithrombotic agents. Exclusion criteria were as follows: combined organ transplantation, ABO-incompatible KTx, aP for other reasons than KTx, and case reports with ≤3 participants.

| Outcomes
The primary outcome of interest was the event of RGT confirmed by ultrasound, angiography or any other equivalent imaging method, or by renal histology and graft loss, respectively. Secondary outcomes were (1) thrombosis other than RGT, (2) thromboembolism, (3) adverse outcomes related to aP (bleeding, hematuria, surgical re-intervention, blood transfusions, heparininduced thrombocytopenia, allergic reaction, peptic ulcer, and others), (4) serious adverse events (prolonged length of hospital stay, persistent or significant disability, life-threatening events, death), (5) graft function (based on estimated glomerular renal function/blood creatinine concentration), and (6) patient survival.
Subgroup and sensitivity analyses were planned to be performed as prespecified. 17

| Study identification and selection
The literature search was conducted by pairs of reviewers (SB, MW).
Conflicts were resolved by a third reviewer (MZ). In case of studies reporting results in more than one publication, the most recent and comprehensive article was selected. If studies included data on patients in whom eligibility criteria for the review were unclear or additional information was needed, study authors were contacted to obtain required data. The full texts of identified studies were analyzed by three authors (SB, MZ, MW). Studies ineligible for inclusion in the review were excluded with reasons.

| Data extraction
Two authors (SB, MZ) carried out data extraction by using a standard data extraction form including baseline characteristics and outcomes. Any disagreements were resolved by consensus or, if necessary, by a third party (MW). Further information required from the original authors was requested by written correspondence.

| Quality assessment
The Cochrane risk of bias (RoB) tool (RoB 2 18 for randomized controlled trials [RCT] and ROBINS-I 19 for non-randomized studies) was used to appraise included studies, which was done by three independent reviewers (SB, KB, MW). 20

| Data synthesis and analysis
The RR was calculated from the number of events in each group using a Random effects model and the Mantel-Haenszel method. 21 Heterogeneity was investigated using forest plots and the I 2 statistic. 22 Sensitivity analyses were carried out to assess the statistical effect of a trial design. The investigation of publication bias was estimated by funnel plots. 23,24 All data were transferred into R 3.6.3., and statistical analysis was performed using the additional package meta 4.12-0. 25

| Search results
The selection process for the studies included in the systematic review and meta-analysis is presented in Figure 1. After removing 1829 duplicates, the search strategy yielded 6361 records. From these identified studies, 113 full-text articles were retrieved for full-text review after title and abstract screening. 101 studies were excluded with reason (Table S3) leaving 12 studies (six pediatric and six adult studies) fulfilling the criteria for inclusion in the systematic review and meta-analysis. [8][9][10][11][26][27][28][29][30][31][32][33] No additional studies were found in available clinical trial registries, conference proceedings, and reference lists.
Additional data were retrieved by personal communication.

| Main outcome
The overall RR for developing RGT was 0.24 (95% confidence interval [CI] 0.12-0.49) in the group with aP compared to the control group, with a low to moderate heterogeneity (I 2 = 36.0%, p = .12; Figure 2A). Figure 3 shows

| Subgroup analyses
The following subgroup analyses were carried out: adult studies after

| Sensitivity analyses
Sensitivity analyses were carried out for the following prespecified

| Heterogeneity and publication bias
A random effects model was used for meta-analysis. There was low to moderate statistical heterogeneity between studies for the main outcome RGT. The major sources of clinical heterogeneity between studies were selection of study participants, insufficient reporting of confounders, and measurement of outcomes. The included studies showed an asymmetric appearance with a prominent gap in the right lower bottom corner, suggesting a publication bias likely ( Figure S1).

| Risk of bias
The RoB was assessed for each study and across all studies. The overall RoB of the included RCTs is presented in Figure 4  ASS was the second most frequent used drug for aP. Albeit the use of ASS in the prevention of RGT seems to be slightly superior to heparin, the fact that only two studies were available for the single drug use of ASS, the deduction from this finding cannot be made. 30,35 Thereby, it remains unclear whether antiplatelet drugs offer net advantages over anticoagulant regimens, which resembles the body of evidence in other areas of application. 36,37 Two studies used heparin and ASS simultaneously for a short time period without demonstrating beneficial effects in contrast to single drug use. 28,30 However, this may be more due to the study design than to inferior effectiveness of these drugs when used in combination. The same applies to the non-superior effects of heparin if it was used subcutaneously.
Not surprisingly, all included studies commenced aP in the peri-or early postoperative phase due to the fact that the majority of RGT occur within the first days following KTx. 3  Within this context, it is also a remarkable finding that only few studies utilized drug monitoring with respect to potential subtherapeutic dosing, safety aspects, and risk of prescribing errors. 39

F I G U R E 4
Overall risk of bias of the included randomized controlled trials. The risk of bias in the included randomized controlled trials was assessed by the Risk of Bias 2 Tool. 18 Figure S1 shows the risk of bias assessment per study Different dosing regimens and variable routes of drug administration hinder the comparability of the outcome measures, particularly for the most often used heparin with hypothetically available drug monitoring. 40 Studies reporting on side effects and complications of aP only were not considered to be relevant for the study objective.
Although the majority of the trials provided some information about bleeding complications or other adverse effects associated with aP, it was not possible to make a reliable overall statement due to the imprecise reporting of these outcome parameters.
However, it is of utmost importance to carry out a critical assessment of the benefit-risk balance for aP as for any other medical intervention, and therefore, further studies need to address this issue for a better-informed decision-making in the daily clinical routine. 41,42 From our point of view, preventing RGT represents the main reason for aP in renal transplant patients. For that reason, trials focusing on other types of thrombosis or embolisms were not considered relevant for the study objective. A recently published systematic review assessing the efficacy and safety of aP in renal transplant patients included patients with any type to thromboembolism. 43 As a consequence of the different eligibility criteria, included studies for the outcome RGT were partially different to our systematic review; however, the authors also assumed that aP may reduce the rate of RGT but were unable to draw final conclusions on the benefits and harms of aP in KTx due to the heterogeneity of currently existing data. 43 The risk for developing RGT is assumed to be higher in children than in the adult population. The RGT incidence as well as the more frequent occurrence of RGT in children in the included study populations were comparable to previously published data.
Several modifiable and non-modifiable factors such as surgical challenges due to small donor and recipients may play a pivotal role. 4 4-46 This consideration might support the observation that the suggested benefit of aP in KTx recipients could not be demonstrated anymore when children were excluded. Contrariwise, if the single study using dipyridamole was not considered for inclusion, aP still seems to be superior compared to the control group which counteracts the hypothesis of superior effectiveness in pediatric patients only.
Of note, after excluding studies with a non-randomized controlled design, the intervention and control group showed no differences. However, this observation might be rather due to the poor quality of the RCTs with moderate to high risk of bias than a true estimated effect of the intervention. All other sensitivity analyses did not affect the overall results.
Interestingly, although the study population showed a highly clinical heterogeneity without addressing adequately potential confounders, the statistical heterogeneity of the included studies was only low to moderate. The overall risk of bias, however, was mostly classified as "high" in both, cohort studies and randomized controlled trials, predominantly due to selection of participants, measurement of outcomes, and confounding. Furthermore, the funnel plot shows an asymmetric appearance with a gap in the right bottom corner, which suggests the presence of unpublished studies without intervention effects of aP. 47 This finding is even more important as realistic estimates of the effect of aP in KTx recipients derived from the included studies seem to be very difficult. Therefore, the true effect of intervention may be substantially different. The critical judgment of the quality of the studies may not be only traced back to the fact that many studies were initiated before the Consolidated comprehensive reporting about associated harms and risks. In addition, the primary outcome RGT needs to be assessed objectively ensured by an adequate imaging method or histological diagnosis.

| CON CLUS ION
The findings from this systematic review and meta-analysis might suggest that aP could reduce the risk of RGT in KTx recipients.
However, based on the poor quality of the data and the incongruent management protocols for aP, recommendation cannot be made in favor or against the use of aP in adult and pediatric KTx.
Notwithstanding, the presented data clearly demonstrate the urgent need of future high-quality clinical trials and a common built consensus between KTx centers for an aP intervention protocol.

ACK N OWLED G M ENTS
We thank Sabine Klein, the search coordinator at the main library of the University of Zurich (Switzerland), for developing and running the search strategy used in this review.