Post-mortem organ donation can be performed after brain death (Donation after Brain Death, DBD) or following circulatory death (Donation after Circulatory Death, DCD). DCD is categorized based on the Maastricht classification into four types, of which DCD-I, DCD-II, and DCD-IV are classified as uncontrolled donations; DCD-III is classified as donation following withdrawal of life-sustaining therapy and is a controlled donation [1]. Since July 2017, DCD-I and DCD-II procedures are no longer performed in the Netherlands.

In ODE, a patient dies in a controlled manner, following administration by a physician of euthanasia drugs. However, the dying process of these patients differs from that of patients who donate their organs after circulatory death (i.e., DCD) following withdrawal of life-sustaining treatments in the ICU (i.e., DCD-III) or following brain death (i.e., DBD), as these patients are critically ill [2]. Hence, the DCD classification recently proposed to include ODE patients in a separate category of highly controllable DCD: DCD-V [3, 4].

As the present study describes data from Netherlands, we use organ donation after euthanasia (ODE) as terminology rather than MAID (Medical Assistance in Dying), which is used in other countries [5, 6]. The ethical, legal, and logistical implications of ODE, in general, have been extensively discussed in both the scientific literature and public media [7, 8].

The possibility of ODE is expanding to more countries, and the number of procedures is increasing annually in most countries where ODE is already available [9, 10]. Organ donation after euthanasia may increase the number of donor organs and thus aid in narrowing the gap between the demand and availability of organs for transplantation. The next question to consider is whether the outcomes of the transplanted organs after ODE are sufficient to continue the procedure.

Data on the outcome of kidneys transplanted following ODE is scarce [11]. A conference abstract reported graft function of transplanted kidneys following ODE that was comparable to DCD-III and DBD over a 5-year follow-up period [12]. We hypothesized that transplant outcomes after ODE have favorable initial graft function, favorable estimated glomerular filtration rate, and less graft failure compared to transplant outcomes from DCD-III and DBD over a 10-year follow-up period. We investigated whether this was independent of a comprehensive set of donor, recipient, and transplant variables. This investigation provides the results of kidney transplants following ODE compared to kidney transplants from other forms of donation during a 10-year study period.

Data from the Dutch Transplant Foundation (Nederlandse Transplantatie Stichting, NTS) are recorded in the Netherlands Organ Transplantation Registry (NOTR), which includes all kidney transplantations performed in the Netherlands. The authors requested and obtained data on transplantations between 1st January 2012 (the year of the first ODE retrieval in the Netherlands), and 31st December 2021, from the NTS registry NOTR in accordance with their data registry governance.

These data were used to construct a retrospective cohort of patients who underwent a kidney transplantation, to compare the graft function between ODE, DCD-III, and DBD derived grafts.

Next, we excluded organ transplantations from donors younger than 18 years, and DCD-I, DCD-II, and DCD-IV donations, as defined in the Maastricht Category [1], and living donation retrievals. No donors in the dataset were represented in multiple transplantation categories (e.g., a living transplantation followed by a post-mortem donation). This resulted in the following categories to be studied: organ donation after euthanasia (ODE, DCD-V); organ donation after circulatory death, Maastricht Category III (DCD-III); and organ donation after brain death (DBD) [1, 3].

Recipients may undergo multiple kidney transplantations during their disease course. For the primary investigation, we included and characterized the most recent transplant (i.e., the latest transplant) in a recipient. In this way any recipient with multiple kidney transplantations was included only once in the analyses.

We described the donors, the recipients, and transplantation and graft characteristics for ODE (DCD-V), DCD-III and DBD categories. Initial graft function, as well as estimated glomerular filtration (eGFR) rate over 10 years and graft failure, were described for ODE, DCD-III and DBD categories.

Of the 9,208 kidney transplantations performed in the Netherlands between 1st January 2012 and 31st December 2021, 9,070 were from donors aged 18 years and older. After excluding 4,790 transplantations from living donors and 13 transplantations from DCD-I, DCD-II, and DCD-IV donors, and after excluding previous transplantations within the recipients, 4,111 kidney transplant recipients remained, with kidney transplants originating from 2,730 unique donors (Figure 1). In total, 148 recipients received a kidney from a donor after ODE, 2,118 from a donor after DCD-III, and 1,845 from a donor after DBD (Figure 1).

Donor recipient and transplantation baseline characteristics are presented in Table 1. ODE donors had lower serum creatinine concentrations (p = 0.046) compared to DBD donors. ODE recipients were younger (p = 0.034) than DCD-III recipients. A minor, not clinically relevant difference, was found in WIT between ODE transplantations and DCD-III transplantations (p = 0.022). As expected, WIT was longer in ODE as compared to DBD transplantations (p < 0.001). CIT was shorter in ODE transplantations as compared to DBD (p < 0.001), no difference was found for CIT in ODE and DCD-III (Supplementary Table S3).

Initial graft function was available for ODE 127 out of 148 (86%), for DCD-III 1940 out of 2,118 (92%), and for DBD 1636 out of 1845 kidneys (89%) (Table 2). First-week post-transplantation outcomes showed that immediate graft function was higher in ODE, which was similar to DBD, when compared to DCD-III (overall p-value: p < 0.001) (Table 2). DCD-III showed more delayed graft function as compared to ODE and DBD, whereas primary non-function was similar (overall p-value: p < 0.001) (Table 2; Supplementary Table S1).

Mean graft function over time by eGFR is shown in Table 2. Longitudinal mixed-effects regression analyses adjusted for donor category and time (model 1, Table 3) showed that, compared to DBD, longitudinal eGFR for ODE was (β: 95% CI) 1.64 mL/min/1.73 m2 (−1.84; 5.13) and for DCD-III was −1.47 mL/min/1.73m2 (−2.74;-0.20) over the 10-year period. After additional adjustments for donor sex, donor age, donor smoking, recipient age, recipient sex, cold ischemic period, anastomosis time, and initial graft function (model 2, Table 3), and further adjustment for donor hypertension, donor diabetes mellitus, WIT and transplant PRA (model 3, Table 3) this association disappeared. Mixed-effects analyses with random slopes for time showed similar results (model 1–3, Supplementary Table S4).

After exclusion of primary non-functioning grafts, over a median follow-up period of 4.0 years (IQR 2.0–6.6), 1810 grafts were transplanted, which included 72 ODE, 958 DCD-III, and 780 DBD grafts. Over the median follow-up period, 213 grafts (11.8%) failed. This included 7 grafts (9.7%) for ODE, 93 grafts (9.7%) for DCD-III, and 113 grafts (14.5%) for DBD. Median follow-up periods of each donor category were 3.7 years (IQR 2.0–5.8) for ODE, 4.0 years (IQR 2.0–6.3) for DCD-III and 4.1 years (IQR 2.1–7.0) for DBD grafts.

When studying the association between graft failure and donor category, compared to DBD, the hazard ratio for ODE was (HR: 95% CI) 0.67 (0.33–1.36) and for DCD-III was 0.71 (0.57–0.88) using crude Cox regression analysis (model 1, Table 4). After adjustment for donor sex, donor age, donor smoking, recipient age, recipient sex, cold ischemic period, anastomosis time and initial graft function, the hazard ratio was, compared to DBD, 0.57 (0.25–1.29) for ODE and 0.56 (0.43–0.73) for DCD-III. After additional adjustments for donor hypertension, donor diabetes mellitus, WIT, and transplant PRA, the statistically significant difference between DBD and DCD-III grafts disappeared. The proportional hazards assumption was met (Supplementary Figures S1–S3).

The four sensitivity analyses, re-analyzing the above models by replacing the recipients for organs donated (i.e., including all kidneys transplanted in each recipient; sensitivity analysis 1, Supplementary Table S2); in which recipient death was not censored, but included as an event; (sensitivity analysis 2, Supplementary Table S2); in which primary non-function was included (sensitivity analysis 3, Supplementary Table S2); and in which only the first transplantation of each recipient was included (sensitivity analysis 4, Supplementary Table S2), all showed similar results as the primary analyses.

This study addresses the transplant outcomes of kidneys donated after euthanasia over a 10-year study period, compared to DCD-III and DBD, and has three main findings. First, immediate graft function was higher in ODE, when compared to DCD-III, and similar to DBD. Second, longitudinally, eGFR for ODE did not differ from eGFR for DBD and eGFR for DCD-III over 10 years, after adjustment for donor sex, donor age, donor smoking, recipient age, recipient sex, CIT, AT, initial graft function and donor hypertension, donor diabetes mellitus, WIT and transplant PRA. Third, graft failure for ODE did not differ from graft failure for DBD and graft failure for DCD-III, after adjustment for donor sex, donor age, donor smoking, recipient age, recipient sex, CIT, AT, and initial graft function, donor hypertension, donor diabetes mellitus, WIT and transplant PRA.

Human kidney transplantation remains the treatment of choice for the majority of patients with end-stage renal failure [19–21]. Despite increased numbers of donor organs due to expanded donor criteria, organs from living donors, and donation after circulatory death, the gap between the demand and availability of kidneys for transplantation remains substantial [22–25]. Although the results for ODE kidneys regarding longitudinal eGFR and graft failure were not statistically significantly different compared to those for DBD or DCD-III kidneys, the overall results for ODE support the concept that ODE kidneys are a promising extension of the donor pool. Notably, extension of the donor pool is not the primary goal of the procedure, because ODE is the patient’s final altruistic wish.

Previous research on the outcomes of kidney transplantations following ODE was done in smaller cohorts and case series [11, 12], while no study has assessed longitudinal eGFR and graft function over 10 years. In contrast to others who included data that did not measure an extensive set of potential confounders [18] or only studied ODE compared to DCD-III [14], we focused on DBD, DCD-III, and ODE, using comprehensive data from a nationwide registry.

With regard to other organs donated following ODE, preliminary studies on graft function of transplanted lungs after ODE [26–28] yielded outcome results comparable to DCD-III grafts, and similar results were reported for transplanted livers [29–31]. Recently, the first successful heart transplantation after donation after euthanasia was published [32]. The current study thus found comparable transplant outcomes between ODE and DCD-III [29] and between ODE, DCD-III, and DBD [26] regarding graft failure for kidneys transplanted.

In the future, more patients will request to donate their organs after euthanasia, and it is expected that an increasing number of countries will allow this procedure. Observational data showed that approximately 10% of all patients undergoing euthanasia might be medically eligible to donate at least one organ [33].

However, the vast majority of patients who undergo euthanasia are suffering from end-stage malignancy, which makes them unsuitable as a donor. “Euthanasia donors” often suffer from neurodegenerative or psychiatric disorders, which are not primarily associated with deterioration of organ function of potentially transplantable organs, such as the heart, lungs, liver, kidneys and pancreas [4, 34]. DCD-III donors, who die after withdrawal from life-sustaining therapy, inevitably suffer from hypoxia, hypotension, and inadequate organ perfusion during the progression to circulatory arrest (agonal phase) and the mandatory 5-min period of warm, pulseless ischemia [35, 36]. Donors after brain death suffer from a systemic inflammatory response with a potentially negative impact on graft outcomes [37].

This study has strengths and limitations. A major strength is the large nationwide registry including all donations and transplantations in the Netherlands with a comprehensive set of donor, recipient and transplantation variables that created the opportunity to investigate potential confounding in the associations under investigation. Indeed, different studies have shown associations between either prolonged CIT or prolonged AT or both, and both kidney function and post-transplant graft failure [38, 39]. AT has also been associated with delayed graft function [40]. Hence the adjustments for these variables in the models of the present study. Furthermore, the Cox models were adjusted for transplant PRA as it indicates the risk of transplant failure to the host-response to transplantation [13, 14]. In addition, hypertension and diabetes mellitus, and tobacco exposure, have each been associated to worse eGFR [41]. Therefore, models were adjusted for recipient hypertension and diabetes mellitus and donor smoking behavior. As no information on smoking behavior in recipients was available, residual confounding might have caused us to underestimate the present associations, although most likely recipient smoking behavior is not dependent on ODE, DBD or DCD-III donors [42]. The criteria for HLA-mismatch are different for the first transplantation and for later transplantations and we chose to study the population that comprises the most recent transplantation of recipients in the analyses. Therefore, the associations could not be adjusted for HLA mismatch and this could have led to an underestimation of the present results.

Another strength is a total of four sensitivity analyses that were conducted. In sensitivity analysis 1, all kidney transplantations within recipients were included to investigate whether including multiple transplants for the same recipients changed the results. In the second sensitivity analysis, recipient death with functioning transplant, was included as an event, as graft survival and recipient deaths may be related. In sensitivity analysis 3, primary non-function was included as an event, despite it being considered a short-term transplant outcome. The fourth sensitivity analysis, considering only the first kidney transplantation per recipient, was performed as HLA matching is not independent of the number of transplantations within a recipient. The presence of donor-specific HLA antibodies before transplantation is considered a risk factor for graft rejection. Furthermore, waiting time until transplantation increases the risk of higher sensitization levels. Organ transplantation induces HLA alloimmunization, affecting the matching of a re-transplant and waiting time until transplantation [18].

Another strength lies in the approach of investigating the 10-year post-transplantation outcome in two different ways: using Cox proportional hazards analyses for the association between donor categories and graft failure and longitudinal analyses for the association between donor categories and eGFR, which together increase the validity of the results. The study has several limitations as well. In the Netherlands, post-mortem donation allocation is based on blood and tissue match between the donor and recipient of the organ, the medical urgency of the recipient, and other circumstances related to the condition of the organ or the background of the recipient. Neither the donor nor their relatives are allowed to choose a recipient [43]. One donor could have donated two kidneys to two different recipients. However, we did not take into account the potential dependency between recipients who received a kidney from the same donor, which is a limitation of this study. This dependency between recipients could potentially have affected transplant outcomes, although the direction of its effect is difficult to assess. With regard to sensitivity analysis 1, to investigate whether including multiple transplants for the same recipients changed the results, it needs to be recognized that multiple kidney transplantations within a recipient during the disease course are dependent and this was not accounted for in the Cox models.

Furthermore, the current study’s ODE sample size decreased considerably after 6 years of follow-up, potentially compromising the reliability of the longer-term findings. This limitation requires a cautious interpretation of results during the extended follow-up period of ODE grafts, suggesting that conclusions towards 10 years should be less strongly conveyed. Future research should thus focus on larger cohorts to enhance the robustness of long-term conclusions. Given the contemporary annual increase in the number of ODE procedures, it is, however, estimated that an analysis of the first 300 kidney transplantations will take at least 5 additional years.

Another limitation of the study is that recipient ethnicity could not be used in the MDRD4 equation, due to lack of information. This has potentially led to a small underestimation of follow-up eGFR. However, since the same error has been made within each recipient, this will not make major difference in the trend over time.

In conclusion, kidneys transplanted after euthanasia have a good immediate graft function and a comparable longitudinal eGFR over 10 years and comparable hazard for graft failure when compared to kidneys transplanted after brain death or circulatory arrest. Overall, these results support the concept that ODE kidneys are a promising contribution to the donor pool, and ODE should be continued.