The imbalance between supply and demand of kidneys for transplantation and consequent drive to increase kidneys available for transplantation has led to a progressive expansion of deceased donor acceptance criteria. Utilization of older donors, donors with comorbidities, and donors after cardiac death (DCD) have all been part of this process—the first two of these often have been further categorized as expanded criteria donors (ECD), based on the criteria for additional listing in the US system. However, understanding the outcomes from transplantation using kidneys from more marginal donors is key to evaluating whether and how these kidneys should be used.
The poorer outcomes from ECD kidneys (compared with standard criteria donor kidneys) have been well documented, together with the risk/benefit profile for individuals whose alternative is long-term dialysis . Similarly, there have been analyses of outcomes of kidney transplantation from DCD donors [2, 3]. In this issue, Singh and Kim  extend these analyses to a group with both risk factors for poor outcomes—kidneys from ECDs who donated after cardiac death (DCD ECD kidneys). Underlying this question is a lingering concern that donor kidneys which are already affected by age-related and vascular disease-related changes might be more sensitive to the warm ischemia and other traumas of donation after circulatory standstill than “ideal” donor kidneys. Framed as a statistical question, the issue is whether there is an interaction between ECD- and DCD-associated risks.
Utilization of DCD has been increasing in recent years, with about 10% DCD kidneys in the Singh and Kim series; in the United Kingdom over 2000–2007, 1/3 of kidneys were from DCD donors . However, as demonstrated, uptake has been from donors with a lower risk profile. Use of DCD donors fitting ECD criteria in the United States and Australia has not been common until recently. In the current series in 2009, of kidney DCD, 10% were ECD, in contrast to 20% of nonDCDs. In the United States in the period examined by the authors (2000–2009), these donors comprised less than 1% of all deceased kidney donors. These are subject to selection bias—the usage described in the United States is among those with lower PRA. In the United Kingdom, use of these donors is more common (Figure 1); Summers et al. reported 88/627 (14%) DCD donors ≥60 years . Grafts from those donors were associated (in a multivariate analysis) with a twofold increased risk of graft failure compared with those <40 years; a similar age effect is seen among DCD in the United States .
The stated conclusion from the Singh and Kim study is that there is no interaction between the risks associated with DCD and ECD donor status—that is, the risks associated with ECD and DCD are additive rather than multiplicative. This result is consistent with expectations, but not definitive. Interactions are notoriously difficult to demonstrate; the challenges of statistical power are demonstrated in the current study—despite a total of almost 68 000 recipients, there were less than 600 DCD ECD recipients, most of whom had less than 3 years follow-up. The p value (0.14) for the principal interaction is a test of the difference between the DCD-associated relative risks between non-ECD (1.07) and ECD (1.21) and could easily be a false-negative result (type II error). Although interactions were similarly statistically insignificant for other endpoints, the direction of effect was consistent.
How then should these results affect the decisions of organ donation agencies, recipients and their clinicians? Singh and Kim advocate the use of these kidneys “judiciously” to expand the donor pool. This is already happening, with rapid expansion in the use of DCD ECD kidneys (Figure 1). Which patients will benefit from these transplants? This is a crucial area yet unexplored; given the literature on ECD kidneys we can predict a more restricted group than for ECD-only kidneys. The overall risk of graft failure in the current study is 19% higher for ECD non-DCD kidneys (compared with non-ECD non-DCD kidneys); the risk for ECD DCD kidneys is a similar quantum higher again, with a relative risk of graft failure 43% higher compared with non-ECD non-DCD kidneys. With this increased risk, there are likely to be recipient groups for whom transplantation from this source is not beneficial or possibly even detrimental to overall survival.
The rapidly increasing utilization of these donors also raises questions about how to mitigate this increased risk—issues such as shortening cold ischemia time, pulsatile perfusion and transplanting both kidneys into a single recipient are some possibilities, which will all need further exploration. Performing adequately powered analyses to respond to clinical questions about the best way to use these donors will challenge available numbers. While ideal, RCTs to answer these questions are unlikely to be performed. They may be best addressed in collaborative multinational Registry studies. This needs consistency of definitions and metrics. Although simple, there remains variation in definitions of organ donation, transplantation and graft failure throughout the literature. The WHO and TTS have started a process to harmonize some core definitions and metrics, but progress has been slow. In the meantime, two items will inform whether a silk purse can be formed from the sow's ear of DCD ECD kidneys. First, regular updates of analyses such as those of Singh and Kim are needed. Outcomes are likely to change, as acceptance practices change and the results from the initial experience with highly selected donors and recipients may not be replicated. Insights might be gleaned from temporal changes to date in a larger multinational dataset. Second, evaluation of how to best utilize the resource of these kidneys in the allocation process. This demands high quality analyses of a variety of scenarios, to ensure individual recipients and society as a whole benefit from this expansion of the donor pool. Thirdly, Registries and other observational studies need to critically evaluate whether they are collecting the right information—in the case of DCD warm ischemic time is of particular importance. A final challenge will be ensuring equity in utilization of this resource; Singh and Kim demonstrate substantial variability in the use of these organs, unsurprising given the inconsistencies seen with listing for ECD donors .