This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [18,19,20]. The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) under the identifier [1110538].

Studies were eligible if they included adult recipients undergoing living-donor kidney transplantation and reported comparative outcomes between grafts with MRA and SRA grafts. When studies included both living and deceased donor populations, they were eligible only if outcomes for living-donor kidney transplantation could be clearly extracted or if living donors constituted the predominant cohort. Non-comparative studies, case series without an SRA comparator, and studies lacking extractable outcome data were excluded from quantitative synthesis but could be considered for qualitative discussion. Studies published before January 2020 were excluded because this review was to evaluate contemporary surgical practice and outcomes after living-donor kidney transplantation using modern reconstruction techniques.

A comprehensive and systematic literature search was conducted to identify studies evaluating the clinical outcomes of living-donor kidney transplantation using MRA gtaftscompared with SRA grafts. The following electronic databases were searched: PubMed/MEDLINE, Embase, Scopus, Web of Science Core Collection, and the Cochrane Central Register of Controlled Trials (CENTRAL).

All databases were searched from inception until 31 October 2025. Although the electronic searches covered all available records from database inception to ensure comprehensive retrieval, eligibility was restricted to studies published between January 2020 and 31 October 2025. This timeframe was selected to reflect contemporary living-donor kidney transplantation practice, including advances in laparoscopic donor nephrectomy, vascular reconstruction techniques, perioperative management, and modern immunosuppressive strategies. No language restrictions were applied. In addition, ClinicalTrials.gov and the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) were searched to identify ongoing or unpublished studies. Reference lists of all included articles and relevant review papers were manually screened to identify additional eligible studies.

The search strategy combined controlled vocabulary terms (e.g., MeSH and Emtree) and free-text keywords related to kidney transplantation, living donation, and MRA. Search terms were adapted to each database’s indexing and syntax. The full electronic search strategies for all databases, including exact search strings and the dates of the last search, are provided in Appendix S1.

All identified records were imported into Rayyan QCRI for deduplication and screening [21]. Duplicate records were identified using Rayyan QCRI’s automated duplicate detection function and subsequently verified manually by two independent reviewers before title and abstract screening. The reviewers independently assessed potential duplicate records based on study title, authorship, publication year, study center, and participant characteristics. Any disagreement regarding duplicate classification was resolved through discussion and consensus. Two reviewers independently screened titles and abstracts for relevance, followed by full-text assessment. Discrepancies were resolved by consensus or a third reviewer. When multiple publications appeared to originate from the same cohort or transplant center and had overlapping study periods, the most comprehensive or most recent report was included in the quantitative synthesis to avoid double-counting participants. Additional reports from the same cohort were retained only for qualitative synthesis if they provided supplementary or non-overlapping outcome data. Studies were included irrespective of donor nephrectomy procurement technique (open or laparoscopic). As reporting of procurement approach was inconsistent across the included studies, this variable was considered a potential source of clinical heterogeneity rather than an exclusion criterion. Studies including mixed donor populations (living and deceased donors) were eligible if outcomes for living-donor kidney transplantation were reported separately. When living-donor-specific data could not be clearly extracted, such studies were excluded from quantitative synthesis but retained for narrative analysis. Similarly, non-comparative studies reporting outcomes exclusively in MRA grafts without a single-artery comparator were not included in meta-analysis but were considered in the qualitative synthesis to contextualize surgical feasibility and safety. Therefore, studies included in this systematic review were categorized into two groups: (1) comparative studies with extractable MRA versus SRA outcomes included in quantitative synthesis, and (2) non-comparative or mixed-population studies included only for narrative synthesis. This distinction was maintained throughout the manuscript and evidence tables. Accordingly, only studies with extractable living-donor-specific comparative data were included in the meta-analysis for each outcome. The study selection process will be presented in a PRISMA 2020 flow diagram.

Data extraction was independently performed by two reviewers using a standardized and piloted extraction form. For each eligible study, data were extracted at the study level and outcome level. Disagreements were resolved through discussion, with arbitration by a third reviewer when consensus could not be achieved [22].

Extracted study-level variables included study design, country, study period, sample size, and inclusion of living-donor recipients. When studies included mixed donor populations, only living-donor-specific data were extracted if reported separately; otherwise, the study was excluded from quantitative synthesis.

Extracted donor-and graft-level variables included donor age and sex (when available), number of renal arteries, presence of accessory vessels, and vascular reconstruction strategy (bench reconstruction versus separate arterial anastomoses). Donor nephrectomy technique (laparoscopic versus open) was recorded when reported, but was not used as an inclusion criterion due to inconsistent reporting.

Extracted intraoperative variables included warm ischemia time, cold ischemia time, arterial anastomosis time, and total operative time, as reported by each study.

Extracted outcome variables included DGF, primary non-function, vascular and urological complications, graft survival, patient survival, and eGFR at prespecified follow-up intervals. DGF was defined a priori as the requirement for dialysis within the first post-transplant week. When outcome definitions varied across studies-particularly for vascular and urological complications-only studies with clearly extractable and comparable definitions were included in quantitative pooling for that outcome. Studies reporting outcomes using composite, non-standard, or purely qualitative definitions were retained for narrative synthesis only.

When data were reported as medians with interquartile ranges, means and standard deviations were estimated for meta-analytic pooling using validated statistical methods [23]. When available, data regarding donor nephrectomy technique, number of renal arteries, and vascular reconstruction strategy (bench reconstruction versus separate arterial anastomoses) were extracted to allow qualitative assessment of their potential influence on early vascular and functional outcomes.

Only studies reporting extractable numerical data (event counts for dichotomous outcomes or mean and standard deviation for continuous outcomes) were included in quantitative pooling for each outcome. Studies reporting outcomes qualitatively (e.g., “no significant difference”) or without sufficient numerical detail were excluded from meta-analysis for that outcome but retained for narrative synthesis. Detailed input data used for each meta-analysis are provided in the relevant tables in the Results section.

Risk of bias was assessed independently by two reviewers using the ROBINS-I tool for non-randomized studies of interventions [24]. For RCTs (if any), we applied the Cochrane RoB 2 tool [25]. Each study was evaluated across seven domains: bias due to confounding, participant selection, classification of exposure, deviations from intended interventions, missing data, outcome measurement, and selective reporting [26].

Predefined decision rules were applied to ensure consistency. Studies including mixed donor populations, non-comparative designs, or lacking adjustment for key confounders (e.g., donor age, ischemia time, reconstruction complexity) were rated at higher risk of bias in the confounding and selection domains. Overall risk of bias for each study corresponded to the highest level of risk identified across domains. Disagreements were resolved by consensus or adjudication by a third reviewer.

Effect measures were calculated as risk ratios (RRs) for dichotomous outcomes, mean differences (MDs) for continuous outcomes, and hazard ratios (HRs) for time-to-event outcomes. Pooled estimates were generated using a random-effects model (restricted maximum likelihood, REML) with Hartung-Knapp adjustment [27]. Heterogeneity was assessed with the I² statistic, τ², and Cochran’s Q [28].

Pre-specified subgroup analyses included:

Sensitivity analyses excluded high-risk studies, unadjusted analyses, and mixed donor cohorts.

Assessment of publication bias was not performed because each meta-analysis contained fewer than ten studies, which limits the reliability and interpretability of funnel plot asymmetry and Egger’s regression testing. The potential influence of publication bias was instead considered qualitatively as part of the GRADE assessment [29]. Only studies with clearly extractable and consistently defined outcome data were included in quantitative pooling for each outcome; all others were summarized narratively.

The certainty of evidence for each outcome was graded using the GRADE approach, considering risk of bias, inconsistency, indirectness, imprecision, and publication bias [30]. Results will be summarized in a Summary of Findings (SoF) table.

The systematic search identified 312 records across electronic databases and trial registries, including PubMed (n = 82), Embase (n = 74), Scopus (n = 61), Web of Science (n = 55), Cochrane CENTRAL (n = 28), ClinicalTrials.gov (n = 7), and the WHO International Clinical Trials Registry Platform (n = 5). After automatic identification of duplicate records using Rayyan QCRI followed by manual verification by the reviewers, 67 duplicate records were removed, leaving 245 unique records for title and abstract screening. No additional eligible studies were identified through trial registry searches or manual reference screening.

Of the 245 records screened, 203 were excluded for being irrelevant to the study question. Forty-two full-text articles were assessed for eligibility. Thirty-two studies were excluded at the full-text stage for the following reasons: non-comparative study design (n = 9), mixed donor populations without extractable living-donor-specific data (n = 8), absence of relevant clinical outcomes (n = 7), duplicate or overlapping cohorts (n = 5), and conference abstracts without a full report (n = 3).

Ultimately, 10 studies published between 2020 and 2025 met the inclusion criteria and were included in the qualitative synthesis, encompassing approximately 2,100 living-donor kidney transplant recipients. All included studies were retrospective cohort or case-control analyses conducted in high-volume transplant centers across Europe, Asia, and North America [31]. Studies with extractable numerical data contributed to the respective meta-analyses, depending on the availability of outcomes. The study selection process is summarized in the PRISMA 2020 flow diagram (Figure 1).

The 10 included studies comprised a total of 2,105 kidney transplant recipients (MRA: approximately 600; SRA: approximately 1,500). Among these studies, seven comparative studies provided extractable data on living-donor kidney transplantation. They contributed to quantitative synthesis for at least one outcome. In contrast, three studies were included only in the narrative synthesis because they lacked a direct SRA comparator or contained mixed donor populations without separable living-donor-specific data. The contribution of each study to individual meta-analyses varied depending on the avialibility of outcomes and reporting completeness. Sample sizes ranged from 30 to 330 patients. Most studies were retrospective single-center cohorts, and one study used a matched case-control design. Follow-up duration ranged from 12 months to 3 years. Detailed characteristics of all included studies and their contribution to quantitative or narrative synthesis are summarized in Table 1.

Surgical strategies varied across studies and included end-to-side arterial reconstruction, ex vivo vascular reconstruction, and selective ligation of accessory arteries [36]. Operative and vascular reconstruction characteristics are presented in Table 2.

Studies included only in the narrative synthesis were retained because they provided clinically relevant information regarding surgical feasibility, reconstruction strategies, and perioperative management of complex renal vascular anatomy in living-donor kidney transplantation. Importantly, the number of studies contributing to each pooled analysis was limited. Only three studies provided extractable data for delayed graft function analysis, while four contributed to analyses of graft survival and renal function. Therefore, pooled estimates should be interpreted with caution, given the limited evidence base and wide confidence intervals.

Risk of bias assessment was performed using the ROBINS-I tool for non-randomized studies. Most included studies were judged to have a moderate risk of bias, mainly due to their retrospective observational design and potential confounding factors. Two studies were considered to have a serious risk of bias because of non-comparative designs or mixed donor populations without extractable living-donor-specific data. The detailed risk of bias assessment is presented in Table 3.

Six studies reported rates of DGF, defined as the requirement for dialysis within the first post-transplant week. Pooled analysis of three studies with extractable event data [6,13] demonstrated no statistically significant difference in DGF rates between MRA and SRA recipients (RR 1.23, 95% CI 0.24-6.22, I² = 0%; Figure 2) [37]. Three studies provided extractable numerical event data for DGF and were included in the meta-analysis. Other studies were excluded from quantitative pooling due to non-standard definitions or absence of numerical event counts. Detailed study-level input data for the DGF meta-analysis are provided in Table S1.

Four studies provided extractable data on 1-year graft survival [6,13,33,35]. The pooled risk ratio for graft loss showed no statistically significant difference between MRA and SRA grafts (RR 1.03, 95% CI 0.51-2.07, I² = 0%; Figure 3) [38]. Four studies reported extractable graft loss data at 12 months and contributed to the pooled analysis. Studies lacking explicit graft-loss counts or reporting mixed donor populations without separable living-donor data were excluded from the quantitative synthesis. Detailed study-level input data for the 1-year graft survival meta-analysis are provided in Table S2.

Patient survival was reported in 8 studies, consistently exceeding 95% and no significant difference between MRA and SRA recipients [39]. None of the included studies demonstrated an increased mortality risk associated with MRA.

Four studies [13,32,33,35] reported eGFR at 12 months. Meta-analysis revealed a pooled mean difference of -1.75 mL/min/1.73 m² (95% CI -4.50 to 1.01, I² = 0%; Figure 4) [40]. This indicates no statistically significant difference in long-term graft function between MRA and SRA. Four studies reporting mean eGFR values with measures of dispersion at 12 months were included in the meta-analysis. Studies reporting renal function qualitatively or without variance estimates were excluded from pooling. Detailed study-level input data for the eGFR at 12 months meta-analysis are provided in Table S3.

The incidence of reported vascular complications was low across included studies, with no consistent difference between MRA and SRA grafts. Choudhary et al. [11] reported an increased risk of clinically significant transplant renal artery stenosis (cTRAS) among recipients requiring complex vascular reconstruction (HR 2.5, p = 0.01) [41]. Reports of cTRAS were infrequent across studies and varied in definition and reporting methods [42]. Additional narrative outcome data and postoperative complications not suitable for quantitative synthesis are summarized in Table 4. A detailed summary of outcome reporting and data extractability across included studies is presented in Table 5.

Heterogeneity estimates were low across pooled analyses (I² = 0% for DGF, graft survival, and eGFR. However, these estimates should be interpreted cautiously because heterogeneity statistics are unreliable when based on a small number of studies. The observed low I² may reflect limited statistical power rather than true clinical homogeneity. Considerable clinical variability persisted across the included studies, including differences in vascular reconstruction techniques, definitions of outcomes, donor nephrectomy approaches, and duration of follow-up. Therefore, the absence of statistical heterogeneity should not be interpreted as evidence of identical treatment effects across settings. Formal assessment of small-study effects was not performed because the number of studies contributing to each meta-analysis was below the recommended threshold for reliable interpretation. Publication bias therefore remains difficult to exclude [43].

The certainty of evidence for each outcome was assessed using the GRADE approach. The overall judgments and supporting rationale are summarized in Table 6.

Our findings provide reassurance to transplant surgeons and programs that the presence of MRA should not, in itself, preclude the use of otherwise optimal living kidney donors. Although grafts with MRA are associated with increased operative time and anastomotic complexity, this technical burden does not consistently translate into inferior short-term graft function, patient survival, or 1-year graft survival when contemporary reconstruction techniques are employed. Accordingly, expanding the living donor pool by accepting MRA kidneys appears justified, particularly in settings facing persistent donor shortages, provided meticulous surgical planning and expertise are available.

Donor nephrectomy procurement technique represents an important potential confounder that was inconsistently reported across studies. Differences between laparoscopic and open procurement may influence arterial length, vascular handling, and warm ischemia time, particularly in grafts with multiple renal arteries. This limitation should be considered when interpreting early vascular and anastomotic outcomes. Planned subgroup analyses according to procurement technique, number of renal arteries, or reconstruction strategy could not be robustly performed due to limited and inconsistent reporting across studies.

Several limitations should be noted. First, all included studies were observational, and residual confounding cannot be excluded. Second, several outcomes were extracted from reported percentages rather than raw counts, potentially introducing minor imprecision. Third, several pooled outcomes had wide confidence intervals with lower bounds that crossed unity, reflecting imprecision and limited statistical power. Consequently, the absence of statistically significant differences should not be interpreted as definitive equivalence between multiple and single renal artery grafts. Clinical heterogeneity related to vascular reconstruction technique, number of renal arteries, and center-specific surgical practices likely contributes to variability across studies and limits the feasibility of robust subgroup analyses. All included studies were observational in design and therefore at least at moderate risk of bias according to ROBINS-I. However, no study was judged to be at critical risk of bias. Meta-analysis was performed in accordance with established guidance, with cautious interpretation of pooled estimates and downgrading of certainty of evidence for risk of bias. Finally, long-term outcomes beyond 1-3 years remain insufficiently reported in the contemporary literature.

Future research should focus on prospective multicenter registries and standardized reporting of vascular and urological complications. Studies incorporating long-term graft function and survival, as well as cost-effectiveness analyses of operative complexity, would provide further insight into the implications of MRA in LDKT.