This retrospective study entails a cohort of all consecutive KTs performed at ULS São João from January 2016 to December 2023 (n = 523). During this period, 142 KTs were carried out with uDCD donors, 194 from SCD and 187 from ECD donors, defined according to the United Network for Organ Sharing (UNOS) [7].

Clinical data were collected using a standardized case report form.

The KTs were followed up for a minimum of 6 months and a maximum of 96 months. The last follow-up date was 31 December 2023, unless graft loss or death occurred first. The follow-up period analysed was 7 years for at least 1 year of follow-up for each patient.

The main outcomes studied were 7-year graft survival (both uncensored and censored by death), recipient survival and longitudinal analysis of graft function over 8 years.

In addition, donor and recipient pre-transplant factors and post-transplant variables were assessed and analysed concerning their relevance in the above-described outcomes. Donor characteristics were (Table 1): age, gender, cause of death (cardiac arrest, stroke, trauma, and other), preoperative serum creatinine (SCreat, mg/dL), warm ischemia time (WIT, min), time on A-NRP (min) and cold ischemia time (CIT, hours). Recipient characteristics were (Table 2): age, gender, etiology of CKD (polycystic kidney disease, glomerulonephritis, diabetes, and others including undetermined), dialysis modality (haemodialysis or peritoneal dialysis), dialysis vintage, previous kidney transplant, immunosuppressive induction therapy and immunological risk based on panel reactive antibody (PRA, %) and human leukocyte antigen (HLA) mismatch levels.

Post-transplant characteristics included in the analysis were (Table 3): primary non-function (PNF), delayed graft function (DGF), biopsy-proven acute rejection (BPAR), surgical complications, hospitalization days, and months of follow-up.

The study was performed according to the ethical standards in the Helsinki and Istanbul declarations and was approved by the Local Institutional Review Board and Ethics Committee of ULS São João. Due to the retrospective and non-interventional nature of the investigation, the need for specific informed consent was waived.

Notwithstanding the Portuguese law concerning organ retrieval being an opting-out rule system, the local Organ Procurement Office obtained the agreement from each deceased donor’s next of kin. All recipients were informed at the pre-transplant outpatient clinic about the different types of donors and signed an Informed Consent, according to the donor definition, upon admission for the transplant.

Expanded-criteria brain-death donor was defined as a donor >60 years old or >50 years and with 2 of the 3 comorbidities: hypertension (HTA), death due cerebrovascular event, and SCreat >1.5 mg/dL.

Cold ischemia time was defined as the time from the donor´s aortic clamping during the recovery of the organs until the unclamping of the renal artery during the transplant surgery.

Warm ischemia time was defined for uDCD donors as the time elapsed after cardiac arrest until A-NRP was established.

The A-NRP time encompassed the period from the start of A-NRP until organ retrieval.

The immunological risk was assessed based on three PRA groups (0%; 1%–20%; >20%) and the number of mismatches between donor and recipient of HLA-A, HLA-B, and HLA-DR combined; this variable was dichotomized into two groups (0–3; 4–6).

Primary non-function was defined as permanent graft non-functioning leading to the immediate continuation of dialysis therapy, re-transplant or death. Many PNF cases were associated with early technical complications. Due to frequent overlap between events—such as simultaneous arterial and venous thrombosis, or thrombosis with bleeding or hematoma—a precise attribution of a single causal factor was often not possible.

To address this, we grouped all such intraoperative or early postoperative events under a single category: “surgical complications.” This classification reflects the multifactorial nature of early graft failure and minimizes misclassification bias inherent in retrospective analyses.

Delayed graft function was defined as the need for at least one haemodialysis session during the first week post-transplant, with subsequent recovery of kidney function.

All the acute rejection episodes were biopsy-proven and classified as BPARs. In cases of non-satisfactory function, a percutaneous renal graft biopsy was routinely performed post-transplant during the first 10–14 days and repeated 7–10 days after in case of prolonged DGF.

The permanent return to dialysis or re-transplant defined the death‐censored graft loss. Graft survival encompassed graft loss to dialysis or re-transplant as well as recipient death. Patient survival was determined until death, censored to return to dialysis (considering the deaths during the first 3 months post-transplantation), or until the last date of maximum follow-up, 8 years post-transplantation.

The graft function was assessed by the estimated glomerular filtration rate (eGFR, in mL/min/1.73m2) using the CKD-EPI equation [8]. The eGFR value following transplant was measured at 1 month, 3 months, every 6 months until the end of the third year, and annually thereafter.

The potential uDCD donors in our center must be ≥18 and ≤60 years old, although this is not an exclusion criterion. In addition, standard criteria for selecting uDCD donors were followed regarding neoplastic, infectious, or potentially transmissible diseases [9].

The timeline and order of events for unsuccessful resuscitation before identifying potential uDCD donors were carried out according to the ULS São João protocol [6]. Patients with out-of-hospital cardiac arrest are transported to the emergency department of our hospital on ventilation support and continuous mechanical chest compressions (LUCAS, Physio-Control Inc., Sweden). In case of unsuccessfully advanced life support without inclusion criteria for E-CPR, the intensivist of the emergency department declares death after 10 min (min) of no-touch when no signs of circulation or electrical activity are found. After this, mechanical chest compressions are restarted, and percutaneous cannulation of a femoral artery and vein is performed. An occlusion balloon is placed on the thoracic aorta through the contralateral femoral artery to exclude brain circulation from the extracorporeal circuit. After the establishment of the extracorporeal circuit, chest compressions is terminated.

The nRP is performed using a Maquet CardioHelp system (Maquet, Germany). This technique restores blood flow and oxygenation to the abdominal organs and allows the recovery of kidneys after a prolonged period of WIT. The target pump flow for nRP is 1.75–2.5 L/min, and a constant temperature of 37°C is maintained. Blood samples are obtained every 30 min for blood gases, and biochemistry analysis including serum lactate levels and hematocrit. The maximum preservation time of nRP is 240 min, but it can be extended to 360 min if blood gases, biochemistry, and hematocrit parameters are adequate for organ recovery.

The extraction of organs is performed through a median laparotomy, maintaining nRP support during surgery. Thereafter the kidneys are included in a preservation fluid (Custodiol HTK, Franz Kohler, Germany) and placed in “static” cold storage at 4°C. Implantation surgery is performed within the next 18 h.

In our center, the characteristics of the recipients considered for receiving a kidney from uDCD or DBD donors are similar. However, we avoid performing transplants with uDCD donors in candidates i) waiting for a retransplant, ii) with high immunological risk, iii) with heart failure with depressed ejection fraction and iv) in those undergoing anticoagulation therapy.

All uDCD transplanted patients and DBD recipients with high immunologic risk received rabbit anti-thymocyte globulin (ATG-Fresenius, Fresenius Biotech GmbH, Germany) as induction immunosuppression, at a maximum dose of 2.5 mg/kg/day, with a cumulative target dose of 12.5 mg/kg. High immunologic risk was defined in cases of cPRA >20%, as well as in recipients with prior transplants, the presence of donor-specific antibodies (even if weakly positive), and those with high mismatch levels (HLA 4–6 mismatches). This broader immunologic risk stratification justified the use of ATG in 44.3% of SCD and 33.7% of ECD recipients. In the remaining DBD recipients, we used anti‐CD25 monoclonal antibodies (basiliximab) as induction therapy.

All uDCD and DBD recipients received maintenance immunosuppression with steroids, tacrolimus and mycophenolate mofetil. We start with a 500 mg dose of iv methylprednisolone before and immediately after surgery and transition to 1 mg/kg/day of oral prednisolone on the first-day post-transplantation, with subsequent tapering. Mycophenolate mofetil is started orally on the first day post-transplant at a 500 mg dose twice a day. After the end of the induction therapy, this dose is increased to 750 mg twice a day. Tacrolimus is started at a 0.2 mg/kg/day dose, 2 days before the predicted last dose of anti-thymocyte globulin, to ensure adequate levels by the end of the induction therapy. All patients receive chemoprophylaxis with cotrimoxazole and nistatin. Valganciclovir is administered during the first 6 months after transplant in recipients that receive induction therapy with ATG or in cases of CMV donor (+)/receptor (−) serostatus.

Categorical variables were expressed as absolute (n) and relative (%) frequencies. Continuous variables were expressed as mean ± standard deviation (SD) or median with interquartile range (IQR: first and third quartiles) if the hypothesis of normality of the distribution was not verified. The hypothesis of normal distribution was tested using the Shapiro-Wilk test.

Categorical variables were compared between groups using Pearson’s chi-square test or Fisher’s exact test when the first cannot be used. Continuous variables were compared using Student´s t-test when data were normally distributed or using the Wilcoxon-Mann-Whitney´s non-parametric test otherwise. The Bonferroni correction was the method used to counteract the multiple comparison problems.

Survival curves for graft and patient survival were obtained using the Kaplan-Meier method, and differences between the groups were compared using the Log-rank test.

PNF and DGF risk factors were analysed using a binary logistic regression model. Relative risks are reported as odds ratios with 95% confidence intervals. Multivariable models included all significant factors in the univariable models and were determined with a forward stepwise procedure.

A p-value <0.05 was considered significant in all tests. The statistical analysis was undertaken using IBM SPSS Statics software (version 26.0) and R software (version 4.2.2).

Tables 1 and 2 show an overview of the donor and recipient characteristics according to donor group. The mean age of both donors and recipients in the ECD group was significantly higher (65.3 ± 6.4 and 60.7 ± 7.9 years, respectively) than in the other two groups. However, no significant difference was observed between uDCD (46.9 ± 11.2 and 52.6 ± 11.0 years) and SCD (44.3 ± 11.6 and 50.4 ± 10.4 years).groups regarding the median age of both donors and recipients. Male donors predominated across all groups (uDCD: 71.8%, SCD: 61.8%, ECD: 64.2%), as did male recipients (uDCD: 69.0%, SCD: 59.3%, ECD: 65.8%).

As can be observed in Table 1, the donor cause of death in uDCD group was mainly due to cardiac arrest (97.2%), whereas the donor cause of death in ECD group was mainly due to stroke (59.4%). Preoperative serum creatinine (SCreat) was significantly higher in uDCD donors (1.2 ± 0.33 mg/dL) compared to both SCD (0.88 ± 0.33 mg/dL) and ECD (0.75 ± 0.22 mg/dL) groups. Cold ischemia time (CIT) was significantly lower in the SCD group (13.1 ± 5.8 h) compared to both ECD (14.8 ± 5.8 h) and uDCD (14.5 ± 3.5 h) groups. Kidneys from uDCD donors had a mean warm ischemia time (WIT) of 93.8 min and a mean abdominal normothermic regional perfusion (A-NRP) duration of 177.7 min.

Recipient characteristics of the three groups are presented in Table 2. Recipients from ECD group had an increased percentage of diabetic aetiology (20.3%) in comparison with the other two groups. On the other hand, the uDCD group’s dialysis vintage (50.5 ± 20.5 months) was lower than the other two groups. As expected, PRA in uDCD group (85.9% had 0%) was lower than in the other two groups. So, according to the protocol, the recipients from uDCD group received more commonly ATG (97.8%) as induction therapy in comparison with the other two groups. No differences were observed among recipients from the three groups concerning previous kidney transplants, dialysis modality, and HLA mismatches.

Table 3 summarizes the post-transplant outcomes across the three donor groups. PNF occurred significantly less frequently in the SCD group (4.6%) compared to the uDCD (16.9%) and ECD (13.4%) groups. No significant difference in PNF rates was observed between uDCD and ECD. Surgical complications were the leading cause of PNF in all groups, particularly in uDCD (9.2%) and ECD (10.2%).

Regarding patient outcomes following PNF, in the SCD group, 5 out of 9 patients (55.6%) were relisted and all underwent re-transplantation. In the ECD group, 15 of 25 patients (60.0%) were relisted, and 7 (28.0%) received a second transplant. Among the 24 uDCD recipients with PNF, 19 (79.1%) were relisted, and 9 (37.5%) underwent re-transplantation during the study period. Among the 19 patients who were not relisted—either due to medical contraindications or personal decision—all were older than 60 years, with 11 (57.9%) being over 65. Finally, the overall rate of PNF following re-transplantation was 10.6% in the uDCD group (15 cases), 2.1% in the SCD group (4 cases), and 9.6% in the ECD group (18 cases).

As expected, DGF in recipients from the uDCD group (69.7%) was significantly higher than in the other two groups. In addition, DGF in the ECD group (43.9%) was non-significantly higher than that of the SCD group (37.2%). Accordingly, uDCD recipients had a higher median number of HD sessions and days of hospital stay in comparison with the other two groups. On the other hand, no significant difference was observed in the number of BPARs among the three groups.

As can be observed in Table 4, in univariable analysis, both preoperative serum creatinine (OR 5.042; p = 0.048) and WIT >60 min (OR 1.021; p = 0.016) were significant risk factors for PNF in the uDCD group. In contrast, in the BDD group (SCD + ECD), donor age (OR 1.055; p = 0.001) and CIT (OR 1.082; p = 0.021) were significantly associated with PNF. In multivariable analysis, WIT remained a risk factor for PNF in the uDCD group (OR 1.018; p = 0.047), whereas donor age remained significant in the BDD group (OR 1.056; p = 0.002). Additionally, CIT was a significant risk factor for DGF in both uDCD (OR 1.194; p < 0.001) and BDD (OR 1.091; p < 0.001) groups (Table 4).

The graft survival rate, both death-censored and uncensored, was significantly higher in the SCD group than in the ECD group (Table 5; Figures 2A,B). Death-censored graft survival at 7 years was 92.2% in the SCD group, 79.9% in the uDCD group, and 73.3% in the ECD group (p = 0.003 for SCD vs. ECD; p = 0.082 for SCD vs. uDCD). Similarly, overall graft survival not censored for death was also superior in the SCD group (78.9%) compared to uDCD (64.6%) and ECD (53.5%), with statistically significant differences (p < 0.001 for SCD vs. ECD; p = 0.010 for SCD vs. uDCD).

However, when excluding cases of primary non-function (PNF), the death-censored graft survival improved notably in all groups, reaching 91.4% in uDCD, 96.2% in SCD, and 82.7% in ECD. Under these conditions, no significant difference in death-censored graft survival was observed among the three groups (p = 0.121), reinforcing the impact of early graft loss on long-term outcomes (Table 5; Figure 2C).

The longitudinal analysis of allograft kidney function in the three groups is shown in Figure 3 and Table 5. As observed, mean eGFR was consistently higher in the SCD group (69.37 mL/min/1.73 m² at 12 months) than in the ECD group (50.28 mL/min/1.73 m²). In the uDCD group, eGFR increased from 1 to 3 months and stabilized around 60.66 mL/min/1.73 m² at 1 year, remaining stable thereafter. This resulted in a significantly higher eGFR in the uDCD group compared to the ECD group at 1 year and beyond. No significant difference was observed in eGFR between SCD (67.80 mL/min/1.73 m²) and uDCD (59.90 mL/min/1.73 m²; p = 0.396)groups over the 7-year follow-up (Table 5; Figure 3)univariablemultivariablemultivariable