Beyond living donation: robot-assisted kidney transplantation using deceased-donor grafts—a scoping review of selection, ischaemia logistics, and allograft outcomes Authors: Rafał Bogdan Drobot, Tomasz Domel, Marcin Lipa, Daniel Ludwiczak Published: 2026 ABSTRACT Deceased-donor robot-assisted kidney transplantation (DD-RAKT) couples an unscheduled graft pathway with a resource-dependent operation. Living-donor-dominated reviews do not resolve DD-specific selection, ischaemia logistics, outcomes, or patient benefit. We aimed to map DD-RAKT selection, preservation-to-reperfusion logistics, outcomes, and the limits of inference about patient benefit. Following PRISMA-ScR, five sources were searched through 26 July 2026; PubMed/MEDLINE and Europe PMC were updated through 21 August 2026. Reports involving human kidney-transplant recipients and explicitly identifying DD-RAKT were eligible. We charted selection, preservation, and operative pathways, ischaemia intervals, delayed graft function (DGF), complications, and graft outcomes. Overlapping reports were clustered; counts and event rates were not pooled. The combined searches returned 2,789 source records; 1,324 unique records remained after deduplication; 64 reports were assessed; and 33 were included. The reports mapped to 19 programme clusters; 18 were deceased-donor-only reports and 15 were mixed-donor reports. The largest single-centre series included 125 deceased-donor recipients: DGF occurred in 38 (30.4%), death-censored graft loss in two (1.6%), and death within 1 year in three (2.4%); median follow-up was 9.9 months. Cold ischaemia showed a graded observational association with DGF in a separate 115-recipient cohort. Selection denominators and linked preservation-to-outcome timelines remained uncommon. Published evidence supports technical feasibility in selected recipients at experienced centres. It does not establish superiority, net patient benefit, or universal thresholds for cold ischaemia, anastomotic time, graft temperature, or pneumoperitoneum pressure. The proposed dataset provides a descriptive audit framework; causal questions about cooling and pressure require prospective controlled studies.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s11701-026-03896-3. Introduction Kidney transplantation is the preferred treatment for suitable people with kidney failure, and most transplant systems depend heavily on deceased donation [1]. A deceased-donor procedure begins with an unpredictable offer and an already running preservation clock. Donation after circulatory death (DCD) may add procurement warm ischaemia and recovery-pathway variation. Transport, back-table reconstruction, recipient preparation, graft insertion, vascular suturing, and reperfusion then add sequential intervals. The relevant question is therefore not simply whether a robot can complete the implantation. It is whether the robotic pathway can be selected and mobilised without avoidable delay or loss of rescue capacity. RAKT has evolved from an assisted cadaveric implantation in 2002 and a fully robotic deceased-donor procedure in 2010 to programmes reporting hundreds of recipient operations [2–3]. The most plausible patient-centred rationale is reduced wound morbidity, particularly for recipients with obesity who may otherwise face restricted transplant access. Robotic magnification and wristed suturing may also assist pelvic reconstruction. These potential benefits are not uniform across patients or centres and may change with learning curves and platform maturation. Adoption may additionally reflect surgeon ergonomics, preference, existing platform capacity, novelty, market differentiation, or commercial visibility. Those factors may be relevant to implementation, but they are not evidence of patient benefit. The available synthesis does not answer the deceased-donor question. A 2024 systematic review identified 44 RAKT reports and approximately 2,402 recipients, but suspected substantial cohort overlap and estimated that roughly 89% of grafts came from living donors [4]. A contemporary narrative review did not map the donor-specific logistics that determine whether an offered deceased-donor kidney can enter a robotic pathway [5]. Subsequent European cohorts, a cold-ischaemia analysis, orthotopic experience, DBD-versus-DCD comparisons, and new US and Chinese reports have expanded the evidence base [6–20]. Broad comparative reviews nevertheless continue to find uncertain patient benefit [21]. No dedicated scoping review of DD-RAKT was identified. The purpose of this review is to provide a structured evidence base for deciding whether and how DD-RAKT should be evaluated, rather than to advocate wider adoption. We asked the following question: In human recipients of deceased-donor kidneys, what evidence defines selection, preservation and implantation logistics, and early allograft outcomes for RAKT? We first sought to identify independent programme clusters and map recipient, graft, and system selection. We then reconstructed the donor-to-reperfusion ischaemia chain, charted outcomes without double counting, and proposed a minimum descriptive dataset for audit and future comparative research. Scope and positionality: This review does not describe a DD-RAKT programme at the authors’ institutions, and no local clinical data are included. Neither affiliated clinical unit currently performs kidney transplantation. The manuscript therefore makes no claim of direct DD-RAKT programme experience; its conclusions derive from the mapped literature. Materials and methods Design and reporting This scoping review followed the PRISMA extension for Scoping Reviews (PRISMA-ScR), PRISMA 2020 flow terminology, and updated JBI guidance [22–24]. A scoping design was chosen because the literature comprised case reports, programme series, mixed-donor cohorts, comparative studies, and repeated analyses of overlapping recipients. The aim was to map concepts and reporting, not to estimate a pooled treatment effect. The protocol was documented during review conduct and was not prospectively registered. Following peer review, it was registered retrospectively on the Open Science Framework on 21 August 2026 (https://doi.org/10.17605/OSF.IO/KRCNH). Retrospective registration does not confer the safeguards of prospective registration. Eligibility criteria Eligibility was structured using population, concept, and context (Table 1). The population comprised human kidney-transplant recipients in a report that explicitly identified at least one deceased-donor graft implanted with robotic assistance. The concepts of interest were recipient and graft selection; procurement, preservation, and operating-room logistics; operative approach and ischaemia intervals; and perioperative or allograft outcomes. All countries, healthcare settings, ages, heterotopic or orthotopic implantation sites, and robotic platforms were eligible. Peer-reviewed full-text journal articles and case reports were included. Mixed living- and deceased-donor cohorts were eligible only when a deceased-donor denominator or a deceased-donor-specific event was explicit. Where outcomes remained aggregate, they were labelled as such and were not attributed to deceased donation. Reports limited to living donation, donor nephrectomy, renal autotransplantation, nonhuman or simulation models, conference abstracts, editorials, protocols, and secondary reviews were excluded. No date or language restriction was applied at search. Data were charted from English-language full text, an English-language abstract, or an official translation; non-English information unavailable in English was not inferred. Table 1Population–concept–context eligibility frameworkElementIncludedExcluded or constrainedPopulationHuman kidney-transplant recipients with ≥ 1 explicitly identified deceased-donor RAKTLiving-donor-only recipients; donor nephrectomy; autotransplantation; no explicit DD datumConceptSelection, preservation/logistics, ischaemia intervals, operative approach, complications, DGF, graft/patient outcomesReports without recipient implantation or without relevant clinical/operational dataContextAny country, age, robotic platform, heterotopic or orthotopic approach; peer-reviewed journal reportNonhuman/simulation studies; conference abstracts; protocols; editorials; secondary reviews Table 1Population–concept–context eligibility frameworkElementIncludedExcluded or constrainedPopulationHuman kidney-transplant recipients with ≥ 1 explicitly identified deceased-donor RAKTLiving-donor-only recipients; donor nephrectomy; autotransplantation; no explicit DD datumConceptSelection, preservation/logistics, ischaemia intervals, operative approach, complications, DGF, graft/patient outcomesReports without recipient implantation or without relevant clinical/operational dataContextAny country, age, robotic platform, heterotopic or orthotopic approach; peer-reviewed journal reportNonhuman/simulation studies; conference abstracts; protocols; editorials; secondary reviews Population–concept–context eligibility framework Information sources and search PubMed/MEDLINE, Europe PMC, OpenAlex, Web of Science Core Collection, and Scopus were searched from inception through 26 July 2026. Full strategies, export metadata, and source counts are in Online Resource 1. To capture publications appearing during revision, the prespecified PubMed/MEDLINE and Europe PMC strategies were rerun for 27 July–21 August 2026. The update returned nine PubMed and ten Europe PMC records; six overlapped across those sources, leaving 13 unique records. Four met the eligibility criteria [17–20]. OpenAlex, Web of Science, and Scopus were not rerun for this targeted update. Crossref was used only to verify DOI and citation metadata and was not counted as an identification source. Reference lists of included reports and recent broad RAKT reviews were checked. No authors were contacted. Selection and deduplication Records were merged using PMID, normalised DOI, and normalised title, with manual adjudication of potential near-duplicates. Titles and abstracts were screened independently in duplicate: R.B.D. assessed the complete record set, while T.D., M.L., or D.L. provided a second assessment of every record. Potentially eligible reports were evaluated independently by two reviewers. Disagreements were resolved by consensus and, when necessary, adjudicated by a third reviewer. R.B.D. performed the primary extraction, source verification, and overlap assessment; T.D., M.L., or D.L. independently verified these processes and all inclusion and full-report exclusion decisions. The decision log for all 1,324 unique records is Online Resource 3. A report was considered eligible on the basis of an indexed abstract alone when deceased-donor inclusion was unambiguous; inaccessible variables were marked not reported (NR). Data charting A standardised form was piloted on five reports and applied to all included sources. Programme, recruitment period, donor type, preservation, recipient and graft selection, robot and team access, and operative approach were charted. Timing fields included cold ischaemia, out-of-ice/rewarming, vascular anastomosis, repeated or taken-down anastomoses, and conversion. Other fields covered DGF, complications, graft and patient outcomes, follow-up, comparator, and cohort overlap. Candidate and graft-offer denominators, reasons for choosing robotic or open implantation, learning-curve position, cooling, pneumoperitoneum pressure, and patient- or system-level value were charted when reported. Data were rechecked against accessible full text, tables, and supplements. No value was imputed. Automation and generative-AI use Deterministic, non-AI scripts were used only to parse database exports, normalise identifiers, deduplicate records, generate audit files, and propagate prespecified counts into figures and tables. No record was excluded by an automated tool. Generative AI and other AI tools were not used in the design or conduct of the review, including search-code development, bibliographic organisation, evidence retrieval, study selection, eligibility assessment, data extraction, analysis, or interpretation, and were not used to draft the submitted manuscript. All decisions were made by the authors, who reviewed and approved the final text and take full responsibility for the work. Cohort overlap and evidence synthesis Reports were clustered when author groups, institutions, recruitment dates, denominators, or explicit cross-citations indicated likely patient overlap. The largest or most donor-specific report served as the cluster anchor, while earlier reports retained value for technical evolution or longer follow-up. Because individual overlap could not always be resolved, we did not calculate a total patient denominator or pooled event rate, nor did we perform a meta-analysis or formal between-study comparison.Results were mapped by evidence architecture, selection, ischaemia logistics, technique, outcomes, and reporting gaps. Consistent with the mapping purpose of this scoping review, no formal risk-of-bias appraisal was undertaken; design, comparator status, donor-specific extractability, and overlap were mapped as evidence-maturity attributes. Results Search and evidence architecture The original five-source searches returned 2,770 records. The targeted PubMed/MEDLINE and Europe PMC update added 19 source records, of which 13 were unique after cross-source deduplication. Across both phases, 2,789 source records yielded 1,324 unique records after removal of 1,465 duplicates. We excluded 1,260 records at title/abstract screening, assessed 64 reports, and included 33 (Fig. 1). Thirty-one full-report exclusions comprised absence of an extractable deceased-donor subgroup (n = 16), living-donor-only evidence (n = 8), secondary reviews (n = 5), and nonhuman or simulation studies (n = 2). Fig. 1PRISMA 2020flow diagram official template for new systematic reviews with database and register searches [23]. DD, deceased donor; RAKT, robot-assisted kidney transplantation. The original five-source search identified 2,770 records; a PubMed/MEDLINE and Europe PMC update identified 19 additional source records (13 unique across the update sources). Crossref was used for DOI and metadata verification and was not counted as an identification source PRISMA 2020flow diagram official template for new systematic reviews with database and register searches [23]. DD, deceased donor; RAKT, robot-assisted kidney transplantation. The original five-source search identified 2,770 records; a PubMed/MEDLINE and Europe PMC update identified 19 additional source records (13 unique across the update sources). Crossref was used for DOI and metadata verification and was not counted as an identification source The 33 reports were published from 2002 to 2026 and mapped to 19 programme clusters (Fig. 2; Table 2; Online Resource 1). Twenty-four reports (73%) appeared from 2020 onward. Eighteen were deceased-donor-only and 15 were mixed-donor reports. Eight described a single recipient. The remainder ranged from small technical series to multicentre or registry-comparator cohorts. Seven reports used matched, open, or registry comparison groups relevant to recipient outcomes [6, 10, 25–29]. Fig. 2Evidence topology by programme cluster and publication year. Each marker is a report, not an independent cohort; connected markers indicate reports from the same programme cluster and may contain overlapping recipients. Reports sharing a publication year within one programme are horizontally offset within that year to prevent marker overlap. Marker area approximates the visible deceased-donor denominator and is capped at 125 recipients. Overlap was inferred from centres, author groups, recruitment windows, and explicit cross-report statements. Evidence topology by programme cluster and publication year. Each marker is a report, not an independent cohort; connected markers indicate reports from the same programme cluster and may contain overlapping recipients. Reports sharing a publication year within one programme are horizontally offset within that year to prevent marker overlap. Marker area approximates the visible deceased-donor denominator and is capped at 125 recipients. Overlap was inferred from centres, author groups, recruitment windows, and explicit cross-report statements. Table 2Programme-level architecture of the included DD-RAKT evidence (33 reports; 19 clusters)Programme clusterReports (years)Visible DD denominatorSelection/approachKey DD-relevant signalParis, France1 (2002)1Assisted cadaveric implantationFirst human DD report; transient acute tubular necrosisChicago, USA5 (2010–2024)≤ 115; overlappingObesity-focused; transperitoneal; regional coolingCold-ischaemia/DGF gradient; repeated nested cohortsPGIMER dual-graft, India3 (2015–2020)≤ 6; overlappingDual expanded-criteria graftsNo conversion in six-case update; DGF 2/6Tours, France1 (2018)8Combined extra-/intraperitoneal; extraperitoneal graftDGF 2/8; ureteric reintervention 1/8Mumbai, India2 (2018–2022)≥ 9; likely overlappingComplete extraperitonealisationDD-specific outcomes incompletely separatedFlorence/ERUS, Italy4 (2018–2022)≤ 21; overlappingHeterotopic; DBD/DCD; regional coolingProspective DBD comparator; DGF and graft-loss signalsKrasnodar, Russia1 (2020)2Heterotopic; intra-abdominal coolingTwo national first cases; satisfactory functionERUS orthotopic1 (2024)9 DD/16 totalLeft orthotopic for calcified iliac vesselsConversion 2/16; DGF 2/16Germany1 (2024)1Heterotopic; postmortem graftNo dialysis or major complication at 60 daysShandong, China2 (2024–2026)≤ 8; overlappingTransperitoneal; regional cooling; paired-donor comparisonNo reported DD DGF or major complicationKunming, China1 (2025)≥ 109 DD/138 RAKT13 versus 7 mmHg; graft ultrasonographyNo causal pressure inference; DD allocation incompletely visibleCopenhagen, Denmark1 (2025)2 DD/16 totalPilot before ORAKTxDD cold ischaemia 19.7 and 23.0 hERUS multicentre2 (2025)≤ 67; overlappingSeven-centre DD programme; heterotopic/orthotopicLargest European DD cohort and matched comparisonIndian tertiary programme1 (2025)NR; ≥2Includes cadaveric en-bloc graftTwo DD DGF events; denominator not reportedColorado, USA3 (2026)≤ 69 DD; overlappingDBD/DCD; NRP; single versus multiple arteriesExploratory donor/recovery analyses; one aggregate conversionSeoul, Korea1 (2026)1Transperitoneal; regional hypothermiaStable function at 24 monthsCleveland Clinic, USA1 (2026)125 DD/228 totalSingle-port extraperitoneal or multiport transperitonealDD DGF 30.4%; graft loss 1.6%; no DD conversionUS left-sided multicentre1 (2026)14 DD/17 totalLeft-sided for retransplant or right-side contraindicationFive DGF, all DCD; no conversion; centre-specific coolingYantai, China1 (2026)1Third DD transplant; common-iliac implantationCold ischaemia 180 min; no DGF/complication at 30 daysDD, deceased donor; DGF, delayed graft function; DBD, donation after brain death; DCD, donation after circulatory death; NRP, normothermic regional perfusion; NR, not reported. Denominators are not additive because cohorts overlap within and across programme clusters Table 2Programme-level architecture of the included DD-RAKT evidence (33 reports; 19 clusters)Programme clusterReports (years)Visible DD denominatorSelection/approachKey DD-relevant signalParis, France1 (2002)1Assisted cadaveric implantationFirst human DD report; transient acute tubular necrosisChicago, USA5 (2010–2024)≤ 115; overlappingObesity-focused; transperitoneal; regional coolingCold-ischaemia/DGF gradient; repeated nested cohortsPGIMER dual-graft, India3 (2015–2020)≤ 6; overlappingDual expanded-criteria graftsNo conversion in six-case update; DGF 2/6Tours, France1 (2018)8Combined extra-/intraperitoneal; extraperitoneal graftDGF 2/8; ureteric reintervention 1/8Mumbai, India2 (2018–2022)≥ 9; likely overlappingComplete extraperitonealisationDD-specific outcomes incompletely separatedFlorence/ERUS, Italy4 (2018–2022)≤ 21; overlappingHeterotopic; DBD/DCD; regional coolingProspective DBD comparator; DGF and graft-loss signalsKrasnodar, Russia1 (2020)2Heterotopic; intra-abdominal coolingTwo national first cases; satisfactory functionERUS orthotopic1 (2024)9 DD/16 totalLeft orthotopic for calcified iliac vesselsConversion 2/16; DGF 2/16Germany1 (2024)1Heterotopic; postmortem graftNo dialysis or major complication at 60 daysShandong, China2 (2024–2026)≤ 8; overlappingTransperitoneal; regional cooling; paired-donor comparisonNo reported DD DGF or major complicationKunming, China1 (2025)≥ 109 DD/138 RAKT13 versus 7 mmHg; graft ultrasonographyNo causal pressure inference; DD allocation incompletely visibleCopenhagen, Denmark1 (2025)2 DD/16 totalPilot before ORAKTxDD cold ischaemia 19.7 and 23.0 hERUS multicentre2 (2025)≤ 67; overlappingSeven-centre DD programme; heterotopic/orthotopicLargest European DD cohort and matched comparisonIndian tertiary programme1 (2025)NR; ≥2Includes cadaveric en-bloc graftTwo DD DGF events; denominator not reportedColorado, USA3 (2026)≤ 69 DD; overlappingDBD/DCD; NRP; single versus multiple arteriesExploratory donor/recovery analyses; one aggregate conversionSeoul, Korea1 (2026)1Transperitoneal; regional hypothermiaStable function at 24 monthsCleveland Clinic, USA1 (2026)125 DD/228 totalSingle-port extraperitoneal or multiport transperitonealDD DGF 30.4%; graft loss 1.6%; no DD conversionUS left-sided multicentre1 (2026)14 DD/17 totalLeft-sided for retransplant or right-side contraindicationFive DGF, all DCD; no conversion; centre-specific coolingYantai, China1 (2026)1Third DD transplant; common-iliac implantationCold ischaemia 180 min; no DGF/complication at 30 daysDD, deceased donor; DGF, delayed graft function; DBD, donation after brain death; DCD, donation after circulatory death; NRP, normothermic regional perfusion; NR, not reported. Denominators are not additive because cohorts overlap within and across programme clusters Programme-level architecture of the included DD-RAKT evidence (33 reports; 19 clusters) DD, deceased donor; DGF, delayed graft function; DBD, donation after brain death; DCD, donation after circulatory death; NRP, normothermic regional perfusion; NR, not reported. Denominators are not additive because cohorts overlap within and across programme clusters The number of publications exceeded the number of independent evidence units. Five Chicago reports spanned the first fully robotic DD procedure through a 115-recipient cold-ischaemia analysis [3, 6, 25, 27, 30]. Florence and the broader ERUS network generated partly nested case, programme, comparative, orthotopic, and multicentre reports [7, 10–11, 26, 31–33]. Three reports described one dual-graft programme [34–36], and three arose from the Colorado programme [13, 16, 20]. A two-centre left-sided cohort included Colorado cases whose patient-level overlap could not be resolved [18]. Cleveland, Yantai, and the left-sided collaboration added three programme clusters [17–19]. This topology precluded a defensible pooled recipient count. Recipient and graft selection Obesity was the most frequently reported indication for choosing RAKT. The Chicago programme aimed to provide access for candidates often excluded from open transplantation because of wound risk; the earliest analysed cohort had a mean body mass index of 42.6 kg/m² [25]. Later Chicago reports focused on severe obesity, while the 2024 cold-ischaemia cohort excluded clinically important iliac atherosclerosis [6]. An Indian programme excluded high anaesthetic risk, significant iliac atherosclerosis, and multiple previous abdominal operations [12]. Cleveland included recipients aged at least 16 years and selected single- or multiport systems according to availability and anatomical complexity [17]. Few reports supplied denominators for candidates assessed, graft offers considered or declined, or cases redirected to open surgery. Vascular anatomy created several selection branches. Iliac calcification was commonly treated as a contraindication to heterotopic RAKT. Conversely, a multicentre orthotopic series used extensive external iliac calcification as the indication for a left renal-fossa approach; nine of the 16 grafts were from deceased donors, two procedures were converted to open surgery , and two recipients developed DGF [7]. A new two-centre cohort used left-sided heterotopic implantation for 15 retransplants and two right-side contraindications; 14 of the 17 grafts were from deceased donors [18]. A separate case placed a third deceased-donor graft on the right common iliac vessels when both fossae were occupied [19]. The evidence supports anatomy-specific pathways, not a single eligibility rule. Donor selection broadened gradually. Expanded-criteria grafts were implanted as dual kidneys in six recipients [34–36]. Among the 67 ERUS recipients, 54 grafts were from DBD donors, six were from controlled-DCD donors, and seven were from uncontrolled-DCD donors; 26 grafts met expanded-criteria definitions [11]. Colorado reported 29 DBD and 40 DCD grafts and an exploratory comparison of normothermic regional perfusion (NRP) with rapid recovery [16]. Cleveland reported 42 DBD and 83 DCD grafts among 125 DD recipients, including dual and simultaneous kidney-pancreas procedures [17]. Donor quality scores, machine perfusion, and discard or nonselection denominators remained inconsistently reported. The donor-to-reperfusion ischaemia chain Terminology was unstable. ‘Warm ischaemia’ variously denoted donor procurement injury, out-of-ice time, vascular anastomosis, or the complete intraperitoneal rewarming interval. Some reports separated arterial and venous suturing; others reported only one rewarming value. Few distinguished an uninterrupted anastomosis from one that was taken down, repeated, or prolonged by rescue. Similarly labelled intervals were therefore not necessarily comparable and were not pooled. Cold-ischaemia exposure ranged widely. The largest European DD-specific series reported a median of 14.8 h (interquartile range 11.0–19.0) [11]. Cleveland reported a median DD cold ischaemia of 21 h 4 min [17], and the left-sided cohort reported an aggregate median of 1,207 min [18]. A highly selected third-transplant case reported 180 min [19]. These values reflect allocation, transport, preservation, and recipient preparation as well as operative efficiency; they cannot be interpreted as a robotic effect. The clearest within-programme signal came from 115 Chicago DD-RAKTs. Median cold ischaemia was 13.46 h, and DGF increased from 15.8% below 11 h to 34.1% at 11–15 h and 48.5% above 15 h [6]. This observational gradient is susceptible to donor and recipient confounding. It nevertheless reinforces the established clinical importance of total preservation exposure. Reporting component intervals is intended to locate potentially modifiable delay and harmonise definitions, not to replace reporting of total ischaemia or assign independent causation to each component. Regional hypothermia using ice slush or a cooling jacket was the dominant intra-abdominal strategy, but graft temperature, cooling duration, and the time at which the graft left effective cold storage were rarely reported together. No validated temperature threshold or adjusted temperature–outcome relation was identified [37]. One nonrandomised cohort compared 13- and 7-mmHg pneumoperitoneum [29], but the literature did not establish an optimal pressure. Cooling method and pneumoperitoneum pressure therefore remain candidate exposures for controlled evaluation, not evidence-based standards. Operative approaches and rescue Most reports used a transperitoneal heterotopic technique with regional cooling and subsequent retroperitonealisation. Alternatives included combined extra-/intraperitoneal placement [38], complete extraperitonealisation [39], and single-port extraperitoneal implantation [17]. Other variants were left-sided heterotopic implantation [18], orthotopic left renal-fossa transplantation [7], a third transplant onto the common iliac vessels [19], and dual ipsilateral implantation [34–36]. Colorado also reported grafts with multiple arteries, using back-table reconstruction or multiple anastomoses [20]. These variants address different risks and should not be treated as one uniform procedure. Conversion was uncommon but not absent. The 67-recipient ERUS series and the 16-case orthotopic series each reported two conversions [7, 11]. Three conversions were reported in each of the earlier 34-case technical cohort and the later 100-case perioperative cohort from the Mumbai programme [39, 40]. Cleveland reported no conversion among 125 DD recipients, and the left-sided cohort reported none among its 17 recipients, of whom 14 received DD grafts [17–18]. The Colorado multiple-artery analysis reported one conversion overall, but donor source was not specified [20]. The indication, phase, added ischaemia, and clinical consequences of rescue were inconsistently captured. Allograft and patient outcomes Outcome synthesis was descriptive because reports used overlapping recipients, inconsistent denominators, and variable DGF definitions. The 67-recipient ERUS cohort reported DGF in 18 recipients (26.9%) and major postoperative complications in 11 (16.4%). At a median follow-up of 27.1 months, graft survival was 92.6%, patient survival was 97.0%, and median estimated glomerular filtration rate was 55.0 mL/min/1.73 m² [11]. In the 115-recipient Chicago cohort, DGF was 33.2% and increased across cold-ischaemia strata [6]. Cleveland reported DGF in 38 of 125 DD recipients (30.4%) and death-censored graft loss in two (1.6%). Three deaths within 1 year (2.4%) were reported among 125 DD recipients; median follow-up was 9.9 months. [17]. Comparative evidence did not show a consistent survival disadvantage for RAKT, but it did not establish superiority. In the Florence DBD comparison, 21 RAKT recipients had no conversion, DGF in two (9.5%), and major complications in three (14.3%) [26]. In the matched seven-centre comparison, 37 RAKT recipients had shorter vascular and rewarming intervals than 72 open recipients but longer total surgery; postoperative and survival outcomes did not differ detectably [10]. Matching could not remove centre allocation, anatomy, team availability, learning-curve, and offer-selection bias. A recent umbrella review likewise found the comparative benefits uncertain [21]. Small reports frequently described immediate or satisfactory function [2–3, 8, 15, 19, 31, 34–35, 41]. Their value lies in demonstrating technical feasibility and identifying failure modes, not in estimating event rates. In the left-sided cohort, DGF occurred in five of 17 recipients, all after DCD grafts. No conversion or major complication was reported, and graft and patient survival were 100% at a median 12.1 months [18]. This cohort was small, included mixed donors, and partly overlapped with the Colorado experience. The paired-donor Shandong comparison reported no conversion, DGF, or other perioperative complication among eight RAKT recipients [28].In Kunming, 138 RAKT recipients were analysed according to pneumoperitoneum pressure (13 vs 7 mmHg) and were also compared with 105 open-transplant recipients. Thirty-day and one-year serum creatinine levels and complication rates did not differ detectably; differences in graft-flow velocities, which were lower at 13 mmHg, were not statistically significant.[29]. The nonrandomised design, incompletely visible donor allocation, and potential confounding prevent a causal pressure recommendation. Reporting gaps and minimum dataset No report supplied the complete denominator from graft offer and recipient eligibility through final robotic or open implantation. Reasons for not using RAKT, robot and team availability, night or weekend timing, predicted versus realised cold ischaemia, procurement pathway, machine perfusion, graft temperature, and rescue time were rarely linked to outcomes. Fifteen mixed-donor reports did not consistently stratify every relevant endpoint by donor source. These omissions prevent readers from separating technical execution from selective deployment. Table 3 proposes a minimum descriptive dataset across donor and preservation, recipient and access, system logistics, operation and timing, mechanistic exposures, and outcomes and value. It is neither a consensus standard nor a causal design. Its purpose is to make selection and timelines auditable, support deduplication and fair comparisons, and identify questions that require randomised or tightly controlled prospective studies. Table 3Proposed minimum descriptive reporting dataset for DD-RAKT programmes and comparative studiesDomainMinimum variablesPurpose and inferential boundaryDonor and preservationDBD/cDCD/uDCD; procurement warm-ischaemia definition; NRP or rapid recovery; expanded-criteria/KDPI; graft side and vessels; reconstruction; static storage or machine perfusion and durationSeparates preimplantation injury and graft complexity from the implantation pathway; supports stratification, not causal attributionRecipient and accessAge; sex; BMI/body composition; dialysis status; anaesthetic risk; prior abdominal/transplant surgery; vascular calcium grade and location; eligibility decision and reasonDefines who was considered, selected, redirected, or denied and permits evaluation of access and equitySystem logisticsOffer/acceptance times; predicted cold ischaemia; transport; recipient call-in; night/weekend status; robot/team availability; open bailout team; reason RAKT was not usedMakes platform availability, opportunity cost, and selection auditableOperation and timingPlatform and learning-curve position; approach; graft insertion; incision/docking; out-of-ice; venous/arterial/total clamp; reperfusion; repeated or taken-down anastomosis; conversion phase/reason; median and IQRLocates delay and rescue events while avoiding an unsupported linear or causal interpretation of anastomotic timeMechanistic exposuresCooling method; temperature site/device and time series; cooling duration; pneumoperitoneum pressure and duration; haemodynamics; graft perfusion measuresDefines exposures for nested randomised or tightly controlled studies; registry observation alone cannot establish effectOutcomes and valueEAUiaiC; Clavien–Dindo; transfusion; wound/lymphocele; DGF by dialysis and functional definitions; primary nonfunction; serial creatinine/eGFR; rejection; graft/patient survival; readmission; patient-reported recovery; costLinks execution to transplant, patient, equity, and system value and tests whether adoption produces benefitThis evidence-derived framework is not a consensus standard or a causal design. It supports transparent selection, audit, and hypothesis generation; mechanistic effects require prospective controlled studies Table 3Proposed minimum descriptive reporting dataset for DD-RAKT programmes and comparative studiesDomainMinimum variablesPurpose and inferential boundaryDonor and preservationDBD/cDCD/uDCD; procurement warm-ischaemia definition; NRP or rapid recovery; expanded-criteria/KDPI; graft side and vessels; reconstruction; static storage or machine perfusion and durationSeparates preimplantation injury and graft complexity from the implantation pathway; supports stratification, not causal attributionRecipient and accessAge; sex; BMI/body composition; dialysis status; anaesthetic risk; prior abdominal/transplant surgery; vascular calcium grade and location; eligibility decision and reasonDefines who was considered, selected, redirected, or denied and permits evaluation of access and equitySystem logisticsOffer/acceptance times; predicted cold ischaemia; transport; recipient call-in; night/weekend status; robot/team availability; open bailout team; reason RAKT was not usedMakes platform availability, opportunity cost, and selection auditableOperation and timingPlatform and learning-curve position; approach; graft insertion; incision/docking; out-of-ice; venous/arterial/total clamp; reperfusion; repeated or taken-down anastomosis; conversion phase/reason; median and IQRLocates delay and rescue events while avoiding an unsupported linear or causal interpretation of anastomotic timeMechanistic exposuresCooling method; temperature site/device and time series; cooling duration; pneumoperitoneum pressure and duration; haemodynamics; graft perfusion measuresDefines exposures for nested randomised or tightly controlled studies; registry observation alone cannot establish effectOutcomes and valueEAUiaiC; Clavien–Dindo; transfusion; wound/lymphocele; DGF by dialysis and functional definitions; primary nonfunction; serial creatinine/eGFR; rejection; graft/patient survival; readmission; patient-reported recovery; costLinks execution to transplant, patient, equity, and system value and tests whether adoption produces benefitThis evidence-derived framework is not a consensus standard or a causal design. It supports transparent selection, audit, and hypothesis generation; mechanistic effects require prospective controlled studies Proposed minimum descriptive reporting dataset for DD-RAKT programmes and comparative studies This evidence-derived framework is not a consensus standard or a causal design. It supports transparent selection, audit, and hypothesis generation; mechanistic effects require prospective controlled studies Discussion Principal findings This review maps 33 reports to 19 programme clusters and reframes DD-RAKT as selection under preservation pressure. The literature has progressed beyond isolated technical feasibility and includes DBD, DCD, expanded-criteria, orthotopic, left-sided, retransplant, dual, and complex-artery grafts [7, 10–11, 16–20, 33, 36]. These reports show that implantation can be completed in selected recipients at experienced centres. They do not establish a universal eligibility threshold, superiority over open transplantation, or net patient benefit. Total ischaemia time is clinically linked to DGF in deceased-donor kidney transplantation [42–45]. The Chicago gradient is therefore more informative than an isolated anastomosis median, although it remains confounded [6]. Component timestamps remain useful for a different reason: they show where delay occurred and expose inconsistent definitions. They should be reported alongside total exposure, not substituted for it. No component should be interpreted as independently causal without an appropriate design. The effective evidence unit is often the programme rather than the paper. Sequential reports document legitimate maturation, but summing their denominators would inflate experience and precision.The concentration of evidence in a small number of programmes, as shown in Fig. 2 , limits transportability. Favourable results may reflect mature transplant governance, trained teams, multiple available platforms, selective allocation, and immediate open rescue as much as the robotic approach itself. Implications for clinical implementation A DD-RAKT pathway requires three explicit decisions. First, the graft decision should record donor pathway, procurement injury, recovery strategy, quality, anatomy, preservation, and the current cold-ischaemia forecast. Second, the recipient decision should record body habitus, vascular imaging, previous surgery, anaesthetic tolerance, and a credible implantation site. Third, the system decision should confirm the availability of the transplant and robotic teams, platform access, cooling equipment, vascular resources, and immediate open bailout. These decisions should be conditional, not rule-based. Severe iliac calcification may preclude heterotopic RAKT yet support orthotopic implantation in a specialised programme [7]. Prior grafts may justify left-sided or common-iliac alternatives [18–19]. A DCD graft is not automatically unsuitable, but preimplantation ischaemic injury increases the importance of preservation control [11, 16, 17]. Conversely, a favourable recipient should not enter a robotic pathway if mobilisation would create unacceptable delay. The literature does not support a universal cold-ischaemia or rewarming cut-off. Programme expansion should be justified by outcomes that matter to recipients and the health system. Reduced wound morbidity and improved access for people with obesity are plausible benefits, but current comparative evidence is of low certainty [4, 21, 46–47]. Surgeon ergonomics, preference, platform utilisation, training, institutional branding, novelty, and commercial incentives may also influence adoption. These drivers should be disclosed and evaluated separately from patient benefit, equity, cost, and opportunity cost. Anastomotic time should not automatically be assumed to be normally distributed or linearly related to outcomes. Its distribution may be right-skewed by difficult cases involving anastomosis takedown or repetition, vascular repair, or conversion. Studies should therefore report the median and interquartile range, the number and timing of interrupted or repeated anastomoses, and the added ischaemia and consequences of rescue. Intraoperative events should be classified using EAUiaiC, and postoperative events should be classified using Clavien–Dindo[48–49]. Research priorities The proposed dataset is suitable for descriptive reporting, programme audit, selection denominators, and generation of hypotheses. A registry cannot by itself determine whether cooling method, graft temperature, or pneumoperitoneum pressure causes better graft function. Confounding by donor quality, anatomy, case complexity, team experience, and elapsed preservation time would remain. Registry data should therefore support, not replace, causal studies. Cooling and pressure questions require prospective protocols with standardised thermometry, explicit pressure exposure, perfusion measurements, prespecified renal outcomes, and either randomisation or tightly controlled allocation [29, 37]. Comparative programme evaluation should record every DD offer considered, every eligible recipient, the reason robotic or open implantation was chosen, calendar time, and learning-curve position. Target-trial emulation or pragmatic trials could then address selective allocation. Evaluation should follow the IDEAL framework [50]. Strengths and limitations The review asks a donor-specific question and uses a comprehensive five-source search through 26 July 2026 plus a targeted update through 21 August 2026. Mixed cohorts were included only when a deceased-donor datum was identifiable. Further strengths are record-level audit files and clustering of overlapping programmes. The review separates donor, recipient, and system selection and treats ischaemia as a defined sequence rather than an isolated robotic metric. Several limitations remain. Embase and dedicated Chinese-language databases were not searched, so Chinese and other non-English evidence may be underrepresented. The post-26 July update reran only PubMed/MEDLINE and Europe PMC; a very recent report unique to OpenAlex, Web of Science, or Scopus could have been missed. Thirteen included reports were charted partly or wholly from indexed abstracts or public publisher records because full text was inaccessible; unavailable variables were marked NR. Programme overlap was inferred and may be incomplete. The protocol was registered retrospectively after peer review and cannot prevent hindsight bias in the way prospective registration can. Most included evidence was retrospective, selectively deployed, and concentrated in experienced centres. No formal risk-of-bias appraisal or quantitative synthesis was performed, limiting causal inference. The authors’ institutions do not run kidney-transplant programmes, and the review contributes no direct DD-RAKT implementation experience. Finally, the proposed dataset has not undergone formal consensus or prospective validation. Conclusions Acceptable early outcomes after DD-RAKT have been reported in highly selected recipients at experienced programmes across several donor and anatomical scenarios. The evidence does not show that DD-RAKT improves outcomes over open transplantation and does not justify wider adoption on the basis of technical feasibility alone. Future evaluation should link graft offers, recipient selection, total and component ischaemia, team and platform availability, rescue events, and donor-stratified outcomes. The proposed dataset can make those processes auditable. Controlled prospective studies are required for cooling and pneumoperitoneum questions, and programme expansion should depend on demonstrable patient and system benefit. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2 Supplementary Material 3 Supplementary Material 3 Supplementary Material 4 Supplementary Material 4