Outcomes of kidney transplants from donors after cardiac arrest; a UK population-based cohort study Authors: Ffion Dewi, Kamlesh Patel, James Hodson, Dimitra Limnatitou, Christine Biela, Angeliki Kosti, Adnan Sharif, Vasileios K Mavroeidis, Jay Nath Published: 2026 ABSTRACT AbstractBackgroundKidneys from donation after cardiac arrest (DACA) donors are frequently offered for transplantation in the UK but concerns about organ quality may lead to potential reluctance to accept these kidneys. This study aimed to compare outcomes between DACA and other deceased donor (i.e. non-DACA) kidney recipients.MethodsA retrospective analysis, including multivariable analysis, was conducted of UK national registry data for all deceased donor kidney transplants performed between 2016 and 2020.ResultsData were available for 6,239 donors (of which 43.6% were DACA), and 11,375 recipients. DACA donors were found to be significantly more likely to donate after circulatory death (48.1% vs. 33.5%, p < 0.001) and to have higher terminal creatinine (median: 0.80 vs. 0.71 mg/dL, p < 0.001), but were less likely to be extended criteria donors (29.1% vs. 46.8%, p < 0.001), compared to non-DACA donors. On multivariable analysis, DACA was independently associated with a significantly lower rate of delayed graft function (odds ratio: 0.84, 95% confidence interval [CI]: 0.75–0.94, p = 0.002) and lower 12-month serum creatinine levels (adjusted difference: 2.6%, 95% CI: 1.0-4.2%, p = 0.001), despite higher terminal creatinine. However no significant associations between DACA and patient (p = 0.644) or graft (p = 0.075) survival were observed.ConclusionsThis study found no evidence that DACA is associated with inferior recipient outcomes, with some outcomes even demonstrating a small superiority effect in comparison to non-DACA kidneys. However, these findings can only be interpreted in the context of current organ acceptance practices.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12882-026-04969-z.Keywords: Donation after cardiac arrest, Deceased donor kidney transplantation, Donation after circulatory death, Kidney transplant outcomes, Organ acceptance, National registry study Introduction Background Kidneys from donation after cardiac arrest (DACA) donors, namely those suffering cardiac arrest and receiving cardiopulmonary resuscitation (CPR), are frequently utilized for transplantation in the United Kingdom (UK), and this is consistently increasing [1]. As the use of uncontrolled donation after circulatory death (DCD) kidneys in the UK is currently limited to pilot programs [2], this infers that all DACA donors have experienced a return of cardiac output prior to consideration for organ donation. Following admission to hospital, they may then go on to become DCD or DBD (donation after brainstem death) donors. The potentially detrimental effects of cardiac arrest and CPR on organ quality are not well established, with factors such as the duration and quality of CPR often being difficult to determine at the time of the organ offer. During periods of cardiac arrest there is no blood flow to the organs, leading to ischemic injury and formation of glomerular microthrombi [3]. There are concerns relating to the harmful effects of ischemia-reperfusion injury to organs following CPR [4, 5], which may lead to reluctance amongst clinicians to accept kidneys from such donors. The time period for which kidneys are refractory to asystole is not clear, although it is reassuring that good outcomes are achieved in countries with DCD programs with no-touch time (declaration of death, and initiation of insertion of cannulae for organ preservation [6]) of up to 30 min [7]. The evidence base for DACA kidney donation is limited and, to our knowledge, there are currently no analyses of national registry data comparing outcomes of kidney transplants between unselected DACA and non-DACA donors. One explanation for this could be that many national registries (including that of the UK) have only recently included cardiac arrest and CPR status as a recorded datapoint. A recent systematic review [5] showed reassuring outcomes overall for transplants from DACA donors, but the included studies were mostly small retrospective series. The largest cohort study in the current literature analyzed data from the USA for DBD donors (of which 4,971 were DACA), and reported comparable graft survival for DACA and non-DACA recipients [8]. However, no other outcomes were measured, and only patients who had in-hospital cardiac arrest following a neurological event were included; there were also no data provided regarding pre-donation creatinine levels or the duration of CPR. A recent Australian registry study of DCD kidneys reported no significant differences in the incidence of delayed graft function (DGF), post-transplant estimated glomerular filtration rate (eGFR) or allograft loss between the DACA and non-DACA cohorts [9]; however, there were only 335 DACA donors in this study, which limited the statistical power of the analyses. A UK registry analysis of 11,219 transplanted kidneys, of which 17% had acute kidney injury (AKI), has demonstrated that kidneys from donors with AKI can be utilized with good outcomes [10]. However, the impact of raised donor creatinine resulting from renal hypoperfusion secondary to cardiac arrest is not as well studied as AKI from other causes. As such, there are concerns that AKI in the form of raised donor creatinine from DACA donors may not be as reversible as AKI in kidneys from donors with other pathologies. This may lead to disparate practice amongst clinicians, with a higher likelihood of declining organs offered from DACA donors. Aims & objectives The aim of this study was to improve understanding of the effect of DACA on kidney transplant outcomes, to aid clinical decision-making during the offer process. The objectives were as follows: To quantify the rate of DACA in deceased kidney donors, and compare the characteristics of DACA and non-DACA donors. To quantify the rate of DACA in deceased kidney donors, and compare the characteristics of DACA and non-DACA donors. To assess the association between the duration of CPR and donor terminal creatinine levels. To assess the association between the duration of CPR and donor terminal creatinine levels. To compare transplant outcomes between recipients of DACA vs. non-DACA organs. To compare transplant outcomes between recipients of DACA vs. non-DACA organs. To quantify the relationship between donor terminal creatinine and recipient 12-month serum creatinine levels. To quantify the relationship between donor terminal creatinine and recipient 12-month serum creatinine levels. Methods Setting, participants and data sources A retrospective, observational study was performed using the National Health Service Blood and Transplant (NHSBT) registry, which includes contemporaneous data of all transplants performed in the UK. Data were extracted for all deceased donor kidney transplants (i.e. DBD or DCD) in the UK between January 2016 and December 2020 (inclusive). For analysis of recipient factors and transplant outcomes, pediatric recipients (age < 16 years) were subsequently excluded. Outcome measures Outcome measures were 12-month serum creatinine, DGF (including primary non-function [PNF]), death-censored graft survival and patient survival. Variables and definitions Donor variables included demographics, medical history, cause of death and terminal creatinine. Details of cardiac arrests occurring immediately prior to donation were extracted, to identify DACA donors. Cardiac arrest details included whether this occurred in or out of hospital; whether it was witnessed; and the durations of both the arrest and CPR. Donor terminal creatinine levels were extracted; since NHSBT used units of umol/L, analyses were performed in the original units, with the results converted to mg/dL using a conversion factor of 88.4. The donor type was classified as DBD or DCD, with the subset of extended criteria donors (ECDs) also being identified. ECD was defined as those that were either aged ≥ 60 years; or aged 50–59 years with at least two of the following features: death due to cerebrovascular accident (CVA), terminal creatinine > 1.5 mg/dL, or a history of hypertension. Whilst donor age was recorded in all cases, there were some missing data for the other features comprising ECD. As such, the ECD status was classified as being unknown in donors aged 50–59 years who had missing data for two or more features. Those aged 50–59 years with missing data for one feature were classified as ECD if the other two features were positive; non-ECD if the other two features were negative; or as having unknown ECD status if only one of the remaining features was positive. Of the transplant outcomes considered, DGF was defined as the need for dialysis within the first week post-operatively. Since recipients with PNF also met this definition, these were included in the outcome of DGF for analysis; however, PNF was additionally considered as a separate outcome. Survival outcomes were assessed using a time-to-event approach, with follow-up commencing on the date of transplant, and recipients being censored at the most recent clinic appointment where the event of interest did not occur. Graft loss was defined as recommencement of long-term dialysis or re-transplant, with patients who died with a functioning graft being censored at this point (i.e. death-censored graft survival). For patient survival, the NHSBT registry had missing data for some cases, despite graft survival data being recorded. As such, the patient survival for these cases was imputed based on the reported graft survival. Specifically, the follow-up time for patient survival was assumed to be the same as that of graft survival. Patients recorded as having died with a functional graft were treated as having died at this point, whilst those recorded to be alive with or without a functioning graft were censored at this point for the analysis of patient survival. Statistical methods Analyses of donor factors were performed on the donor-level data, such that donors of two kidneys were only included once. Analyses of recipient factors and transplant outcomes were performed on the transplant-level data. Comparisons between the DACA and non-DACA subgroups were performed using Mann-Whitney U tests for ordinal or continuous variables, and Chi-square tests for categorical variables. Associations between CPR duration and both donor terminal creatinine and recipient 12-month serum creatinine levels were then assessed using linear regression models. Since creatinine levels followed a skewed distribution, values were log10-transformed prior to this analysis, in order to improve model fit. Non-DACA organs were initially assigned a CPR duration of 0 min for this analysis, but were subsequently excluded, if this resulted in poor model fit. Univariable analyses of predictors of transplant outcomes were then performed using Cox regression for survival outcomes; general linear models (GLMs) for (log10-transformed) 12-month serum creatinine levels; and binary logistic regression for DGF and PNF. For continuous variables, the goodness-of-fit of the resulting models was assessed graphically; where poor fit was identified, variables were divided into categories, either based on established thresholds or quintiles, and treated as nominal for analysis. To minimize exclusions due to missing data, an “unknown” category was included for variables with > 5% missing data. Multivariable models were then produced, with DACA status entered into the model at the first step, and a backwards stepwise approach (removal at p > 0.1) used to identify other independent predictors of outcome. The univariable models of outcomes by DACA status were also expanded to include either the donor type or donor terminal creatinine and an interaction term, to compare the effect size of DACA status between subgroups. For the outcome of 12-month serum creatinine, the interaction effect between DACA status and terminal creatinine was further interrogated by treating terminal creatinine as a continuous variable. Since non-linear trends were indicated, this effect was visualized using generalized additive models (GAMs) for the DACA and non-DACA subgroups separately, with the smoothing function of log10[donor terminal creatinine] as a covariate, and log10[recipient 12-month serum creatinine] as the dependent variable. Continuous variables are reported using mean ± standard deviation (SD) where approximately normally distributed, or as median (interquartile range [IQR]) otherwise. When visualizing the GLMs, creatinine levels were represented using geometric means and 95% confidence intervals (95% CIs). For survival outcomes, event rates were estimated using Kaplan-Meier curves. For univariable analyses, cases with missing data were excluded from the analysis of the affected factor, with the included sample size reported in the tables; multivariable analyses used a complete-cases approach. Analyses were performed using IBM SPSS 24 (IBM Corp. Armonk, NY), and R Statistical Software v4.3.2 (R Core Team 2023), with p < 0.05 deemed to be indicative of statistical significance throughout. Guidelines This study was conducted and reported in accordance with the STROBE guidelines [11]. Results Donor characteristics Data were available for a total of 6,469 donors, from whom kidneys were transplanted into 11,992 recipients (Fig. 1). Cardiac arrest status was recorded in 6,239 donors, of whom 43.6% (N = 2,718) were DACA, with a median cardiac arrest duration of 30 min and the majority (89.3%) occurring out of hospital (Table 1). DACA donors were found to be significantly younger (median: 50 vs. 55 years, p < 0.001) and more likely to be male (61.0% vs. 51.8%, p < 0.001) than non-DACA donors (Table 2). As such, DACA donors had less disease generally associated with older age, including hypertension (22.4% vs. 33.6%, p < 0.001) and death due to CVA (21.5% vs. 82.8%, p < 0.001) and, as a consequence, were less likely to be ECD donors (29.1% vs. 46.8%, p < 0.001). DACA donors had significantly higher terminal creatinine levels (median: 0.80 vs. 0.71 mg/dL, p < 0.001) and were more likely to be DCD (rather than DBD) donors (48.1% vs. 33.5%, p < 0.001). Study flowchart Table 1Details of donor cardiac arrests (in 2,718 donors)FactorTotal NStatisticCardiac arrest out of hospital26982408 (89.3%)Cardiac arrest witnessed25601717 (67.1%)Cardiac arrest duration (mins)251430 (18–40)CPR administered at time of cardiac arrest25332247 (88.7%)Duration of CPR (mins) a202025 (15–36) < 10 min 282 (14.0%) 10–30 min 1101 (54.5%) > 30 min 637 (31.5%)Analyses are on the donor-level data, and include the 2,718 donors where cardiac arrests were recorded, unless stated otherwise. Data are reported as N (column %), or as median (interquartile range). a In donors receiving CPRCPR, cardiopulmonary resuscitation Details of donor cardiac arrests (in 2,718 donors) Analyses are on the donor-level data, and include the 2,718 donors where cardiac arrests were recorded, unless stated otherwise. Data are reported as N (column %), or as median (interquartile range). a In donors receiving CPR CPR, cardiopulmonary resuscitation Table 2Donor factors by cardiac arrest status (in 6,239 donors)Donor Factor N Non-DACADACAp-ValueAge (Years)623955 (45–65)50 (34–60) < 0.001 Sex (% Male)62391823 (51.8%)1657 (61.0%) < 0.001 BMI (kg/m2)623926.9 ± 5.727.0 ± 6.20.733Ethnicity61700.376 White 3234 (92.9%)2517 (93.6%) Asian 117 (3.4%)70 (2.6%) Black 45 (1.3%)33 (1.2%) Other 85 (2.4%)69 (2.6%)Blood group62380.169 O 1673 (47.5%)1238 (45.6%) A 1404 (39.9%)1145 (42.1%) B 333 (9.5%)264 (9.7%) AB 111 (3.2%)70 (2.6%)CMV (% Positive)62011773 (50.7%)1235 (45.7%) < 0.001 Toxoplasmosis (% Positive)6150588 (16.9%)365 (13.6%) < 0.001 Alcohol use60730.033 a Non-drinker/Stopped drinking 606 (17.6%)528 (20.0%) Rarely/Occasionally 1134 (33.0%)854 (32.4%) Drinks lightly 577 (16.8%)418 (15.9%) Drinks moderately 533 (15.5%)442 (16.8%) Drinks heavily 250 (7.3%)170 (6.4%) Drinks very heavily 337 (9.8%)224 (8.5%)History of smoking62002189 (62.5%)1725 (64.0%)0.221History of CVD6108414 (12.0%)346 (13.1%)0.207History of pulmonary disease6139378 (10.9%)376 (14.1%) < 0.001 History of diabetes6196231 (6.6%)213 (7.9%)0.054History of illegal drug use6108572 (16.5%)696 (26.3%) < 0.001 History of liver disease6122188 (5.5%)73 (2.7%) < 0.001 History of tumors6168215 (6.2%)77 (2.9%) < 0.001 History of UTI6105229 (6.6%)143 (5.4%) 0.041 History of hypertension61571166 (33.6%)603 (22.4%) < 0.001 Donor type (% DCD)62391180 (33.5%)1306 (48.1%) < 0.001 Terminal creatinine (mg/dL)59500.71 (0.58–0.90)0.80 (0.62–1.11) < 0.001 Cause of death6027 < 0.001 Cerebrovascular accident 2818 (82.8%)563 (21.5%) Hypoxic brain Injury 169 (5.0%)1786 (68.1%) Trauma 130 (3.8%)70 (2.7%) Others 286 (8.4%)205 (7.8%)Extended criteria donor61501619 (46.8%)784 (29.1%) < 0.001 Both kidneys transplanted62393096 (87.9%)2455 (90.3%) 0.003 Pancreas transplanted6239487 (13.8%)491 (18.1%) < 0.001 Liver transplanted62392227 (63.2%)1611 (59.3%) 0.001 Small bowel transplanted623842 (1.2%)44 (1.6%)0.152Heart transplanted6239501 (14.2%)360 (13.2%)0.264Lungs transplanted6239512 (14.5%)248 (9.1%) < 0.001 Analyses are on the donor-level data, and are based on 6,239 donors, unless stated otherwise. Continuous variables are reported as mean ± standard deviation, or median (interquartile range), with p-values from Mann-Whitney U tests. Categorical variables are reported as N (column %), with p-values from Chi-square tests. Bold p-values are significant at p < 0.05. a p-Value from Mann-Whitney U test, as the factor is ordinalBMI, body mass index; CMV, Cytomegalovirus; CVD, cardiovascular disease; DACA, donation after cardiac arrest; DCD, donation after circulatory death; UTI, urinary tract infection Donor factors by cardiac arrest status (in 6,239 donors) Analyses are on the donor-level data, and are based on 6,239 donors, unless stated otherwise. Continuous variables are reported as mean ± standard deviation, or median (interquartile range), with p-values from Mann-Whitney U tests. Categorical variables are reported as N (column %), with p-values from Chi-square tests. Bold p-values are significant at p < 0.05. a p-Value from Mann-Whitney U test, as the factor is ordinal BMI, body mass index; CMV, Cytomegalovirus; CVD, cardiovascular disease; DACA, donation after cardiac arrest; DCD, donation after circulatory death; UTI, urinary tract infection Recipient and transplant characteristics After excluding pediatric recipients (N = 196), analyses of transplant characteristics were based on 11,375 transplants (Fig. 1), of which 43.9% (N = 4,988) used DACA organs. Comparisons of recipient characteristics between DACA and non-DACA kidneys found the subgroups to be more similar than for donor characteristics (Table 3a). However, some small but significant differences included a slightly younger age (median: 52 vs. 54 years, p < 0.001) and shorter times on the waiting list (median: 23 vs. 26 months, p < 0.001) for recipients of DACA kidneys. Of the transplant characteristics considered (Table 3b), the proportions of organs from DACA donors were found to increase significantly over time (p = 0.004). There was also a small, but significant, difference in the operative warm ischemic time (WIT) for the subgroup of DCD organs, which was shorter in transplants of DACA organs (median: 17 vs. 18 min, p = 0.036). Table 3aRecipient factors by donor cardiac arrest status (in 11,375 transplants)Recipient Factor N Non-DACADACAp-ValueAge (Years)11,37554 (44–63)52 (40–61) < 0.001 BMI (kg/m2)924727.1 ± 4.926.8 ± 4.90.053Sex (% Male)11,3723913 (61.3%)3164 (63.5%) 0.017 Blood Group11,375 0.049 O 2776 (43.5%)2056 (41.2%) A 2464 (38.6%)2048 (41.1%) B 835 (13.1%)649 (13.0%) AB 312 (4.9%)235 (4.7%)Ethnicity11,2370.896 White 4400 (69.9%)3462 (70.1%) Asian 1121 (17.8%)854 (17.3%) Black 614 (9.7%)488 (9.9%) Other 164 (2.6%)134 (2.7%)Primary renal disease83290.188 Diabetes 1058 (22.7%)915 (24.9%) Polycystic disease 643 (13.8%)487 (13.3%) Glomerulonephritis 476 (10.2%)339 (9.2%) Hypertension 363 (7.8%)275 (7.5%) Pyelonephritis 334 (7.2%)248 (6.8%) Others 1782 (38.3%)1409 (38.4%)Waiting time (Months)11,29226 (13–44)23 (11–41) < 0.001 CRF at transplant11,3680.165 a 0% 3630 (56.9%)2879 (57.8%) 1–49% 1168 (18.3%)947 (19.0%) 50–94% 1078 (16.9%)790 (15.8%) 95–100% 507 (7.9%)369 (7.4%)HLA mismatch level b11,3550.008 a Level 1 510 (8.0%)317 (6.4%) Level 2 1762 (27.6%)1416 (28.4%) Level 3 3174 (49.8%)2407 (48.3%) Level 4 930 (14.6%)839 (16.9%)Dialysis at transplant11,3645381 (84.3%)4167 (83.6%)0.340Analyses are on the transplant-level data, and are based on 11,375 transplants, unless stated otherwise. Continuous variables are reported as mean ± standard deviation, or median (interquartile range), with p-values from Mann-Whitney U tests. Categorical variables are reported as N (column %), with p-values from Chi-square tests. Bold p-values are significant at p < 0.05. a p-Value from Mann-Whitney U test, as the factor is ordinal. b HLA mismatch status is defined according to the NHSBT allocation scheme, and is based on mismatching at the HLA-A, -B and -DR loci. The four categorical levels in the UK system reflect increasing immunological disparity in terms of both the number of mismatches and locus-specific patternsBMI, body mass index; CRF, calculated reaction frequency; DACA, donation after cardiac arrest; HLA, human leukocyte antigen Recipient factors by donor cardiac arrest status (in 11,375 transplants) Analyses are on the transplant-level data, and are based on 11,375 transplants, unless stated otherwise. Continuous variables are reported as mean ± standard deviation, or median (interquartile range), with p-values from Mann-Whitney U tests. Categorical variables are reported as N (column %), with p-values from Chi-square tests. Bold p-values are significant at p < 0.05. a p-Value from Mann-Whitney U test, as the factor is ordinal. b HLA mismatch status is defined according to the NHSBT allocation scheme, and is based on mismatching at the HLA-A, -B and -DR loci. The four categorical levels in the UK system reflect increasing immunological disparity in terms of both the number of mismatches and locus-specific patterns BMI, body mass index; CRF, calculated reaction frequency; DACA, donation after cardiac arrest; HLA, human leukocyte antigen Table 3bTransplant factors and outcomes by donor cardiac arrest status (in 11,375 transplants)Factor N Non-DACADACAp-ValueYear of transplant11,3750.004 a 2016 1233 (19.3%)893 (17.9%) 2017 1330 (20.8%)965 (19.3%) 2018 1351 (21.2%)1082 (21.7%) 2019 1387 (21.7%)1166 (23.4%) 2020 1086 (17.0%)882 (17.7%)CIT (Hours)11,35413 (10–16)13 (10–16)0.975fWIT (Mins) b263518 (14–23)17 (14–23) 0.036 sWIT (Mins) b379413 (11–15)13 (11–15)0.112Machine perfusion11,320173 (2.7%)161 (3.2%)0.104Delayed graft function10,2531448 (25.2%)1129 (25.0%)0.836Primary non-function10,253178 (3.1%)110 (2.4%) 0.047 Serum creatinine at 12 months (mg/dL)91191.48 (1.18–1.90)1.39 (1.11–1.79) < 0.001 Death-censored graft survival (at five years) c11,12985.3%87.8%< 0.001 cPatient survival (at five years) c11,13586.1%85.2%0.416 cAnalyses are on the transplant-level data, and are based on 11,375 transplants, unless stated otherwise. Continuous variables are reported as mean ± standard deviation, or median (interquartile range), with p-values from Mann-Whitney U tests. Categorical variables are reported as N (column %), with p-values from Chi-square tests, unless stated otherwise. Bold p-values are significant at p < 0.05. a p-Value from Mann-Whitney U test, as the factor is ordinal. b In recipients of DCD organs. c Rates represent Kaplan-Meier estimates at five years, and p-values are from univariable Cox regression modelsCIT, cold ischemic time; DACA, donation after cardiac arrest; DCD, donation after circulatory death; f/sWIT, functional/second warm ischemic time Transplant factors and outcomes by donor cardiac arrest status (in 11,375 transplants) Analyses are on the transplant-level data, and are based on 11,375 transplants, unless stated otherwise. Continuous variables are reported as mean ± standard deviation, or median (interquartile range), with p-values from Mann-Whitney U tests. Categorical variables are reported as N (column %), with p-values from Chi-square tests, unless stated otherwise. Bold p-values are significant at p < 0.05. a p-Value from Mann-Whitney U test, as the factor is ordinal. b In recipients of DCD organs. c Rates represent Kaplan-Meier estimates at five years, and p-values are from univariable Cox regression models CIT, cold ischemic time; DACA, donation after cardiac arrest; DCD, donation after circulatory death; f/sWIT, functional/second warm ischemic time Associations between duration of CPR and creatinine levels Donor terminal creatinine was found to increase progressively with the duration of CPR (Fig. 2a, p < 0.001), by an average of 6.0% (95% CI: 5.3%-6.8%) per 10 min of CPR (equivalent to approximately 0.05 mg/dL per 10 min). Regression analysis in the DACA cohort demonstrated no significant association between duration of CPR and 12-month serum creatinine levels (gradient: -0.6% per 10 min of CPR, 95% CI: -1.3%, + 0.2%, p = 0.131, Fig. 2b). Association between duration of CPR and creatinine levels. Analysis of donor terminal creatinine was performed on the donor-level data, whilst analysis of recipient 12-month serum creatinine was performed on the transplant-level data. Each analysis initially included all cases for which both the respective creatinine levels and durations of CPR (where applicable) were recorded (5,276 and 8,089 for donor/recipient creatinine, respectively). The red lines represent linear regression models, with log10[creatinine] as the dependent variable, and the duration of CPR as a continuous covariate. Points represent the geometric mean creatinine levels for non-DACA donors and for subgroups of DACA donors with CPR durations of 1–14, 15–29, 30–59 and 60 + minutes, and are plotted at the mean CPR duration for the subgroup; whiskers represent 95% confidence intervals. The y-axes are truncated, to more clearly visualize trends. a Non-DACA donors were assigned a CPR duration of 0 min, and included in the regression analysis. b Including recipients of non-DACA organs resulted in poor model fit; hence, the regression analysis was performed for the DACA subgroup only. CPR, cardiopulmonary resuscitation; DACA, donation after cardiac arrest Transplant outcomes by DACA status Transplant outcomes were assessed using both univariable and multivariable analyses. The resulting multivariable models are reported in full in Supplementary Tables 1–5; summarized in Table 4; and described below: Table 4Summary of multivariable analyses of donor cardiac arrest (in 11,375 transplants)OutcomeUnivariable AnalysisMultivariable AnalysisModelTotal N(Events) aCoefficient(95% CI)p-ValueTotal N(Events) aCoefficient(95% CI)p-ValueDelayed graft functionBLR10,253 (2577)0.99 (0.91–1.08)0.8357910 (2012)0.84 (0.75–0.94) 0.002 Primary non-functionBLR10,253 (288)0.78 (0.61–0.99) 0.045 7910 (222)0.87 (0.59–1.28)0.474Serum creatinine at 12 months bGLM9119-6.3% (-7.7%, -4.8%) < 0.001 7003-2.6% (-4.2%, -1.0%) 0.001 Death-censored graft survivalCox11,129 (948)0.80 (0.70–0.91) < 0.001 8557 (725)0.83 (0.67–1.02)0.075Patient survivalCox11,135 (754)0.94 (0.81–1.09)0.4168562 (579)1.04 (0.88–1.23)0.644Analyses are on the transplant-level data, and results are from BLR, GLM, or Cox regression models, as appropriate. Initially, univariable models are reported, which included the DACA status as a single covariate. Coefficients are reported for DACA vs. non-DACA, and represent hazard ratios for Cox models, odds ratios for BLR models, and percentage differences for GLMs. Multivariable models were then produced, with the DACA status entered at the first step, and a backwards stepwise approach used to select independent predictors of the outcome for inclusion. Further details of the multivariable models are reported in Supplementary Tables 1–5. Bold p-values are significant at p < 0.05. a The total number of cases included in the analysis, after excluding those with missing data, with the number in brackets representing the number of events (for Cox and BLR models). b Creatinine levels were log10-transformed prior to analysis, to reduce the level of skew in the distribution; the coefficients from the resulting models were then anti-logged, and converted to percentage differences between groupsBLR, binary logistic regression; CI, confidence interval; DACA, donation after cardiac arrest; GLM, general linear model Summary of multivariable analyses of donor cardiac arrest (in 11,375 transplants) Analyses are on the transplant-level data, and results are from BLR, GLM, or Cox regression models, as appropriate. Initially, univariable models are reported, which included the DACA status as a single covariate. Coefficients are reported for DACA vs. non-DACA, and represent hazard ratios for Cox models, odds ratios for BLR models, and percentage differences for GLMs. Multivariable models were then produced, with the DACA status entered at the first step, and a backwards stepwise approach used to select independent predictors of the outcome for inclusion. Further details of the multivariable models are reported in Supplementary Tables 1–5. Bold p-values are significant at p < 0.05. a The total number of cases included in the analysis, after excluding those with missing data, with the number in brackets representing the number of events (for Cox and BLR models). b Creatinine levels were log10-transformed prior to analysis, to reduce the level of skew in the distribution; the coefficients from the resulting models were then anti-logged, and converted to percentage differences between groups BLR, binary logistic regression; CI, confidence interval; DACA, donation after cardiac arrest; GLM, general linear model DGF and PNF Rates of DGF were not found to differ significantly between recipients of DACA and non-DACA organs on univariable analysis (25.0% vs. 25.2%, p = 0.836, Table 3b). However, after adjustment for confounding factors, DACA was found to be independently associated with significantly lower rates of DGF on multivariable analysis (odds ratio [OR]: 0.84, 95% CI: 0.75–0.94, p = 0.002). Further investigation found that this discrepancy was almost entirely a result of the adjustment for the donor type (i.e. the greater proportion of DCD donors in the DACA group). Rates of PNF were found to be marginally lower in the DACA cohort (2.4% vs. 3.1%, p = 0.047); however, this effect did not remain significant after adjustment for confounders on multivariable analysis (OR: 0.87, 95% CI: 0.59–1.28, p = 0.474). Patient and graft survival Analysis of long-term survival outcomes included a median of 24 months (IQR: 12–44) of follow-up, during which time there were 754 deaths and 948 (death-censored) graft losses. Patient survival was not found to differ significantly between recipients of DACA vs. non-DACA kidneys on univariable analysis (hazard ratio [HR]: 0.94, 95% CI: 0.81–1.09, p = 0.416, Fig. 3), and remained non-significant on multivariable analysis (HR: 1.04, 95% CI: 0.88–1.23, p = 0.644). Recipients of DACA kidneys were found to have significantly longer death-censored graft survival on univariable analysis (HR: 0.80, 95% CI: 0.70–0.91, p < 0.001, Fig. 3), with this effect narrowly missing significance on multivariable analysis (HR: 0.83, 95% CI: 0.67–1.02, p = 0.075). Kaplan-Meier curves of post-transplant survival outcomes by donation after cardiac arrest status. Analyses were performed on the transplant-level data. The x-axis is truncated at five years, due to declining numbers of patients at risk, and the y-axis is truncated at 50%, to more clearly visualize the difference between groups. p-Values are from univariable Cox regression models, and bold p-values are significant at p < 0.05. DACA, donation after cardiac arrest 12-month serum creatinine Recipients of DACA organs had significantly lower 12-month serum creatinine levels than the non-DACA cohort on univariable analysis (median: 1.39 vs. 1.48 mg/dL, p < 0.001, Table 3). Interestingly, this difference persisted after adjustment for confounders on multivariable analysis, with DACA being associated with 2.6% (95% CI: 1.0%-4.2%, p = 0.001) lower serum creatinine levels (Table 4). This was despite higher terminal creatinine levels in the DACA cohort. Subgroup analysis by donor type (DBD/DCD) and donor terminal creatinine Subgroup analyses were then performed with donors stratified by donor type (DBD or DCD) and terminal creatinine, to assess whether these factors mediated the effect size associated with DACA (Table 5). These analyses were performed for all outcomes, except for PNF, due to there being too few events for reliable analysis. No significant interaction effects between DACA status and donor type were detected for any of the outcomes considered, implying that the relative differences in recipient outcomes between the DACA and non-DACA groups was similar for DBD and DCD donors. However, subgroup analysis by donor terminal creatinine identified significant interaction effects for the outcomes of DGF (p = 0.008), 12-month serum creatinine (p < 0.001) and death-censored graft survival (p < 0.001). In each case, the difference between the DACA and non-DACA subgroups increased progressively across the subgroups with donor terminal creatinine of ≤ 1.0, 1.1-2.0 and 2.1-3.0 mg/dL. No significant interaction effect was observed for the outcome of patient survival (p = 0.124). Table 5Transplant outcomes by donor cardiac arrest status within donor subgroups (in 11,375 transplants)FactorNon-DACADACADACA vs. Non-DACAInteractionp-Value N Statistic N StatisticStatistic (95% CI) Delayed Graft Function Donor type0.251 DBD 3805772 (20.3%)2328406 (17.4%)0.83 (0.73–0.95) DCD 4120676 (34.9%)2182723 (33.1%)0.93 (0.81–1.05)Donor terminal creatinine (mg/dL) 0.008 ≤ 1.0 44401059 (23.9%)2943712 (24.2%)1.02 (0.91–1.14) 1.1-2.0 895270 (30.2%)1058262 (24.8%)0.76 (0.62–0.93) 2.1-3.0 7030 (42.9%)21059 (28.1%)0.52 (0.30–0.91) > 3.0 267 (26.9%)11844 (37.3%)1.61 (0.63–4.15) Recipient Serum Creatinine at 12 Months (mg/dL) Donor type0.797 DBD 33701.45 (1.14–1.84)21131.33 (1.06–1.69)-7.5% (-9.4%, -5.7%) DCD 17101.57 (1.24–2.01)19261.45 (1.15–1.86)-7.2% (-9.4%, -4.9%)Donor terminal creatinine (mg/dL) < 0.001 ≤ 1.0 39641.47 (1.17–1.87)26531.39 (1.11–1.80)-4.2% (-5.9%, -2.4%) 1.1-2.0 7741.56 (1.22–2.02)9331.39 (1.11–1.79)-11.5% (-14.7%, -8.3%) 2.1-3.0 561.62 (1.24–2.22)1841.33 (1.05–1.61)-22.7% (-31.2%, -13.2%) > 3.0 261.36 (1.09–1.67)1031.28 (1.06–1.47)-9.0% (-21.3%, 5.1%) Death-Censored Graft Survival (at Five Years) Donor type0.212 DBD 417686.7%254688.9%0.70 (0.58–0.85) DCD 208182.5%232686.7%0.84 (0.69–1.01)Donor terminal creatinine (mg/dL) < 0.001 ≤ 1.0 485286.7%320787.4%0.99 (0.85–1.16) 1.1-2.0 96981.7%113388.4%0.51 (0.39–0.68) 2.1-3.0 7174.1%21886.4%0.27 (0.12–0.58) > 3.0 2990.9%12096.4%0.93 (0.10–8.34) Patient Survival (at Five Years) Donor type0.185 DBD 417788.1%254687.3%0.82 (0.67–1.01) DCD 208482.3%232883.0%1.01 (0.82–1.25)Donor terminal creatinine (mg/dL)0.124 ≤ 1.0 485687.0%320983.2%1.04 (0.87–1.23) 1.1-2.0 96982.3%113385.8%0.74 (0.54–1.01) 2.1-3.0 7187.9%21891.7%0.42 (0.16–1.11) > 3.0 2996.4%12089.5%1.31 (0.16–10.88)Analyses are performed on the transplant-level data for all outcomes, except for primary non-function, which was not considered in the analysis due to the small number of events. The numbers included in each subgroup analysis are reported in the “N” columns. Outcomes are initially summarized for the DACA and non-DACA cohorts separately, using N (%) for delayed graft function; median (interquartile range) for creatinine levels; and Kaplan-Meier estimated five-year survival rates for the survival outcomes. Separate univariable models were then produced for each of the subgroups defined by the donor type/terminal creatinine, using the approach described in Table 4. These are summarized as odds ratios for delayed graft function; percentage differences for creatinine levels; and hazard ratios for survival outcomes, and are for DACA relative to non-DACA. Models were then produced for the whole cohort, with either the donor type or terminal creatinine, DACA status, and an interaction term as covariates. The p-values for the interaction terms are reported, and interpreted as comparing the magnitude of the difference between the DACA and non-DACA cohorts across the subgroups. Bold p-values are significant at p < 0.05DACA, donation after cardiac arrest; DBD, donation after brainstem death; DCD, donation after circulatory death Transplant outcomes by donor cardiac arrest status within donor subgroups (in 11,375 transplants) Analyses are performed on the transplant-level data for all outcomes, except for primary non-function, which was not considered in the analysis due to the small number of events. The numbers included in each subgroup analysis are reported in the “N” columns. Outcomes are initially summarized for the DACA and non-DACA cohorts separately, using N (%) for delayed graft function; median (interquartile range) for creatinine levels; and Kaplan-Meier estimated five-year survival rates for the survival outcomes. Separate univariable models were then produced for each of the subgroups defined by the donor type/terminal creatinine, using the approach described in Table 4. These are summarized as odds ratios for delayed graft function; percentage differences for creatinine levels; and hazard ratios for survival outcomes, and are for DACA relative to non-DACA. Models were then produced for the whole cohort, with either the donor type or terminal creatinine, DACA status, and an interaction term as covariates. The p-values for the interaction terms are reported, and interpreted as comparing the magnitude of the difference between the DACA and non-DACA cohorts across the subgroups. Bold p-values are significant at p < 0.05 DACA, donation after cardiac arrest; DBD, donation after brainstem death; DCD, donation after circulatory death This relationship was explored further by repeating the interaction analysis treating creatinine as a continuous variable (Fig. 4). For the non-DACA cohort, the model found increasing donor terminal creatinine to be associated with an increase in recipient 12-month serum creatinine levels. However, this reached a plateau at a donor terminal creatinine of approximately 1.6 mg/dL, beyond which further increases in donor terminal creatinine did not appear to impact recipient 12-month serum creatinine. Whilst the DACA cohort showed a similar trend initially, the plateau occurred considerably earlier, at approximately 0.8 mg/dL. Above this threshold a negative trend was observed, with further increases in donor terminal creatinine corresponding to reductions in recipient 12-month serum creatinine. Association between donor terminal creatinine and recipient 12-month serum creatinine. Analyses were performed on the transplant-level data. Trend lines are from generalized additive models, with log10[recipient 12-month serum creatinine] as the dependent variable, and the smoothing function of log10[donor terminal creatinine] as a covariate. Separate models were produced for the DACA and non-DACA subgroups. Points represent the geometric mean recipient 12-month serum creatinine levels for organs within each decile of donor terminal creatinine, and are plotted at the mean donor terminal creatinine of the interval; whiskers represent 95% confidence intervals. The y-axis is truncated, to more clearly visualize trends. DACA, donation after cardiac arrest Discussion What this study adds This is the largest study to date detailing the baseline characteristics and outcomes for kidney transplants from donors following cardiac arrest. Although reassuring outcomes have been previously demonstrated for DACA kidneys, this study is a valuable addition to the existing body of evidence; unlike previous studies, multivariable analyses and subgroup analyses by donor type and donor terminal creatinine were performed to control for the effect of confounding variables. This approach is particularly important, given the large differences in the characteristics of DACA and non-DACA donors. The larger dataset and more robust methodology in this study confirms the reassuring outcomes of DACA kidneys that have been suggested in previous, smaller studies. This study demonstrates two key findings. Firstly, almost half of deceased donors had a cardiac arrest and CPR prior to becoming an organ donor, with the vast majority of these having out-of-hospital CPR. Whilst the primary objective of CPR is not to increase the number of organs available for transplantation, this is no doubt a hugely beneficial consequence, contributing around a thousand kidney transplants per year to recipients in the UK [12]. This is likely to be similar in other countries which have established DCD donor programs. The second key finding is that the outcomes of transplanted DACA kidneys are excellent, and appear to be non-inferior to non-DACA organs. These reassuring outcomes were observed even in the context of significant donor AKI and prolonged CPR duration. Indeed, for several factors (e.g. recipient 12-month serum creatinine, rate of DGF and death censored graft survival), there appears to be a slight superiority in outcomes for recipients of DACA kidneys. This trend towards superior graft outcomes in DACA organs has previously been shown in small, single-center studies [13–16], but this is the first large, methodologically robust, multi-center study to demonstrate this finding. This small superiority effect is interesting and goes further than the existing evidence supporting excellent outcomes for DACA kidneys; however, it warrants further investigation into the mechanism of the effect which may have implications for both organ selection strategies and storage or operative techniques. Interpretation Baseline characteristics Description of the DACA donor cohort demonstrates that DACA donors are more likely to have factors associated with superior outcomes (e.g., younger age and lower likelihood of being ECD); this is consistent with a systematic review of organ transplants from DACA donors [5]. As our study only includes utilized organs, selection bias may play a role, with clinicians potentially accepting lower-risk DACA organs preferentially. There are also other plausible explanations for the baseline differences. The cause of death for DACA donors differed markedly from non-DACA donors: the vast majority of DACA donors in this cohort sustained hypoxic brain injury, including asphyxiation and drug overdose, which generally occur in younger individuals. In contrast, most non-DACA donors died from cerebrovascular accidents, which are more common in older donors. This also explains the higher proportion of DCD donors in the DACA cohort, as hypoxic brain injury often initially spares brainstem function, whereas CVAs, such as intracranial hemorrhage, frequently cause rapid brainstem death, enabling DBD donation. This pattern of proportionally more DCD donors in the DACA context is consistent with other studies [17]. Finally, donation following cardiac arrest may involve an inherent selection bias. Successful return of spontaneous circulation after CPR—particularly in out-of-hospital settings—is more likely in individuals with greater physiological reserve. Therefore, donors who survive arrest and proceed to donation may represent a fitter subgroup, consistent with evidence that CPR success is more likely in younger, healthier individuals [18, 19]. Terminal creatinine and impact of CPR As anticipated, DACA donors had higher terminal creatinine values, which increased progressively with increasing CPR duration. However, this did not translate into poorer post-transplant renal function. On the contrary, recipient 12-month serum creatinine was modestly but significantly lower in the DACA cohort after adjustment for confounding factors. Moreover, interaction analysis showed that this beneficial effect was most pronounced in kidneys from donors with high terminal creatinine values (i.e. 2.1-3.0 mg/dL). This finding supports the hypothesis that transient ischemic injury in the context of cardiac arrest leads to a reversible AKI, which does not impair longer-term graft function in the same way as elevated terminal creatinine in non-DACA donors. It should be noted that additional data which support decision-making surrounding AKI kidney offers, such as historic creatinine values, trajectory of creatinine values and urine output were not available, and so could not be considered in the analysis. It may be that artificial intelligence technologies such as machine-learning applications could utilize these variables to predict outcomes in AKI kidney transplants, thereby aiding decision-making surrounding DACA organ offers [20]. Donor physiology following CPR This study demonstrates that even after prolonged periods of CPR, excellent transplant outcomes can be achieved. The finding that DACA status in itself is not detrimental, and may be protective, in relation to early post-transplant outcomes is notable. DACA donors, who undergo cardiac arrest and receive CPR, present a unique physiological context. One explanation for this is that CPR may confer protective effects through ischemic preconditioning [21], a phenomenon whereby brief ischemic episodes prime the organ to withstand subsequent ischemic stress. This preconditioning could activate cellular mechanisms that mitigate injury, such as upregulating heat shock proteins and antioxidant defenses [22]. However, unlike experimental studies, cardiac arrest is an uncontrolled environment for exposing kidneys to ischemic preconditioning. In addition, pre-clinical and clinical studies suggest that the window of benefit is postulated to last up to 24 h after the insult [21, 23]. It is likely that most kidneys in this cohort were transplanted more than 24 h after cardiac arrest, therefore may no longer be subject to this beneficial effect. Emerging evidence suggests that characteristics of donor cardiac arrest and post-arrest management may influence graft outcomes. A recent analysis of the French national transplant registry [24] found that shorter time to procurement after primary cardiac arrest was associated with an increased incidence of delayed graft function. These findings suggest that factors surrounding the arrest, beyond simply its occurrence, may contribute to transplant outcomes and merit further study. Limitations This study is subject to several limitations inherent in registry-based retrospective analyses, including the aforementioned selection bias. Residual confounding cannot be excluded despite adjustment for a wide range of donor and recipient characteristics. Although the national dataset was well populated, missing data reduced the sample size for certain analyses, which could introduce bias if data were not missing at random. The accuracy of registry data – particularly regarding cardiac arrest variables – was dependent on clinician reporting at the time of data entry. In addition, the breadth of available variables was limited by the predefined fields within the national registry. It was not possible to include more granular clinical details such as CPR quality, intra-arrest physiology, post-resuscitation care, or comprehensive donor risk metrics such as the Kidney Donor Profile Index (KDPI). Whilst we acknowledge that contemporary practice is increasingly moving away from a simple standard criteria donor (SCD) versus ECD dichotomy, this classification remains in routine use within the UK for counselling and allocation purposes. We therefore included ECD status as a pragmatic and clinically relevant variable, although it does not capture donor risk with the granularity of composite indices such as KDPI. The generalizability of our findings to other transplant systems, particularly the United States (US), warrants consideration. In our cohort, the median cold ischemia time was approximately 12 h, which is shorter than that commonly reported in US practice [25]. Differences in organ preservation strategies and broader structural characteristics of the US and UK transplant systems may also contribute to variation in organ transport logistics and cold ischemia times. These factors may influence ischemia–reperfusion injury and early graft outcomes; therefore, our findings should be interpreted within the context of the UK transplant framework. Conclusion Under current donor selection criteria, excellent short- and long-term outcomes are achieved for DACA kidneys, which appear to be superior to those from non-DACA donors, even in the context of prolonged CPR. This study also provides reassurance that raised donor creatinine after cardiac arrest is reversible and appears more favorable than raised donor creatinine due to other causes. Clinicians should be reassured by these findings when considering such organ offers. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors would like to thank NHSBT for supplying the data. We would like to thank all transplant centers in the United Kingdom who contributed data on which this article is based. Abbreviations Acute kidney injury Confidence interval Cardiopulmonary resuscitation Cerebrovascular accident Donation after cardiac arrest Donation after brainstem death Donation after circulatory death Delayed graft function Extended criteria donor Estimated glomerular filtration rate Generalized additive model General linear model Interquartile range National Health Service Blood & Transplant Primary non-function Standard criteria donor Standard deviation Warm ischemic time Author contributions F Dewi: Writing of original draft, review and editing, literature search, data analysis. K Patel: Writing of original draft, review and editing, data analysis, conceptualization, supervision. J Hodson: Statistical analysis, writing of original draft, review and editing. D Limnatitou: Literature search, data analysis, review and editing. C Biela: Data analysis, review and editing. A Kosti: Literature search, data analysis, review and editing. A Sharif: Conceptualization, methodology, review and editing. VK Mavroeidis: Conceptualization, methodology, review and editing, supervision. Jay Nath: Writing of original draft, review and editing, data analysis, conceptualization, supervision. All authors read and approved the final manuscript. Funding Not applicable. Data availability Additional supporting information may be found online in the Supporting Information section. The data underlying this study are held by NHS Blood and Transplant (NHSBT) and are subject to governance and data access restrictions. Anonymized data may be made available to qualified researchers upon reasonable request and with appropriate approvals from NHSBT. Requests should be directed to NHSBT via the corresponding author. Declarations Ethics approval and consent to participate This study was conducted in accordance with the ethical principles of the World Medical Association Declaration of Helsinki. The study comprised a retrospective analysis of fully pseudonymized, routinely collected data from the UK national transplant registry held by NHS Blood and Transplant (NHSBT). Approval for data access and use for research purposes was granted by NHSBT (data reference number: HD29379) through its established governance procedures. Local Research and Development approval was obtained from University Hospitals Birmingham (R&D approval number: RRK7778). Due to the retrospective nature of the study and the use of anonymized registry data, individual informed consent from donors and recipients was not required. No organs from prisoners were used in this study and all organs transplanted were obtained via the UK NHSBT national deceased donor scheme. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Additional supporting information may be found online in the Supporting Information section. The data underlying this study are held by NHS Blood and Transplant (NHSBT) and are subject to governance and data access restrictions. Anonymized data may be made available to qualified researchers upon reasonable request and with appropriate approvals from NHSBT. Requests should be directed to NHSBT via the corresponding author.