End-stage renal disease (ESRD) is an increasing worldwide health concern, and kidney transplantation is the best treatment for survival and quality of life of people with ESRD.1,2 The ageing of donors (currently achieving a mean of 59.2 years in Spain3) and the high discarding rate of organs by biopsy4 are among the main obstacles for transplantation. Performing a right assessment of kidney quality and feasibility is a current challenge to reduce discarding rate of organs potentially valid.

There are several methods to assess the quality of kidneys5 and, for organ allocation, the Kidney Donor Risk Index (KDRI) has acquired special relevance. The KDRI combines 14 donor and transplant factors and provides an estimation of the relative risk of graft failure after kidney transplant from a deceased donor compared to the reference donor.6 Although its discriminatory power is moderate, the KDRI represents a step forward in pretransplantation kidney assessment and allograft prediction.

A new allocation policy based on the KDRI was implemented by the end of 2014 in the United States. Since some transplant factors are generally unknown at the time offer is made, this policy uses a new index, the Kidney Donor Profile Index (KDPI)7 based on the donor-only KDRI version including 10 donor factors. The KDPI is the numerical ranking of the organ to be evaluated compared to all kidneys recovered the previous year. During the allocation process, the KDPI is an aid tool to decide whether to accept an offer of a deceased donor kidney, often discarding organs with KDPI>85% (considered high risk donors).

Last 10 years in Spain, and concretely in Andalucía, the acceptance of a kidney from an expanded criteria donor (ECD)8 has primary based on the result of preimplantation biopsy.9 But biopsy continues having a controversial role on the assessment of renal graft feasibility. Meanwhile some studies justify its use specially for double transplant,10 others doubt about its prediction capacity for renal functioning.11–13

Using a risk index (such as the KDPI) could help the decision-making in Spain for ECD cases when decision depends on the results of the biopsy. Several studies support this hypothesis. Querard et al.8 showed that the relative differences between ECD and standard criteria donors (SCD) were lower in Europe than in North America, particularly for death-censored graft failure. Another study concluded that high-KDPI transplantation offers a better survival versus remaining on the waitlist for recipients older than 50 years when wait time is higher than 33 months.14

This study aims: (1) to check the adequacy of the KDPI to discriminate differences on graft and patient survival for Spain recipients, and (2) to identify predictor factors of graft survival that could contribute to create a Spanish or European index.

A retrospective analysis was performed on data from a database held by the Regional Transplant Coordination of Andalusia (SICATA). Data were retrieved for single kidney transplants performed between January 2006 and December 2015. Graft failure was defined by return to chronic maintenance dialysis. The final follow-up date was 31 December 2015. Statistical analyses were performed using SPSS 24.0 (IBM SPSS Statistics for Windows, Version 24.0. IBM Corp., Armonk, NY, USA).

Since in Spain it is no possible to calculate KDPI as in US, our approach is to use the study population itself to map KDPI from donor factors (see Section Discussion for a comment about this limitation). Cases were ranked, categorised in quartile, and compared for graft survival (death-censored and uncensored) and patient survival using Kaplan–Meier log-rank survival analysis. The associations between the KDPI as a continuous variable and survival of graft and patient were determined using Cox Regression univariate analysis. Cases resulting on patient's death with functioning graft were excluded since this is a competitive event for graft failure and it may alter the results. Furthermore, the discriminative ability of the KDPI was tested with Harrell's C stat for Cox models.

Cox multivariate regression was used for identifying predictor factors of graft survival. Recipient risk factors significantly associated with graft failure using univariable screening at the level of P<0.05 were included in the final multivariable model. These were age, gender, height, diabetes as primary renal disease, time on renal replacement therapy (RRT), HIV and HCV status. Donor variables included HLA mismatch and elements conforming KDPI. Since ethnicity of Spain population is highly uniform, donor and recipient race were not included.

The KDPI may mean a significant improvement for decision-making, but its discriminatory power on graft survival is limited (Harrell's C, 0.56) and the feasibility of high-KDPI grafts (sometimes discarded) on recipients older than 60 years is uncertain. Results obtained are in accordance with those of a recent study validating donor-only KDRI (i.e., KDPI) on a Dutch population and obtaining a discriminatory ability (Harrell's C) of 0.62.19

Our work states that KDPI discriminates graft and patient survival for recipients of 18–59 years, with differences between recipients of optimal donors (KDPI<25%) and the rest. For recipients older than 60 years, the KDPI prediction power is equally limited, but even then, it rightly discriminates between the first, the third and the fourth quartiles (i.e., between optimal and marginal donors). The KDPI is also useful for discriminating patient survival among quartiles. It is noteworthy in both subpopulations the poor discrimination above the median (third and fourth quartiles); this might suggest a lack of accuracy of the index from a certain level of quality, or that others characteristics not collected by the index (e.g., biopsy) are involved.

Regarding recipients older than 60 years, there is an association between graft survival and KDPI. Several studies have analysed different effects caused by expanded or high-KDPI donors according the recipient's age, stating that the influence on survival is lower in older recipients. Ma et al. studied if the use of SCD or ECD kidneys influences differently on patient survival if the recipient's age is higher or lower than 60 years.20 They analysed 3822 cases between 1997 and 2009 stating that in younger recipients there was an excess risk of all-cause mortality (adjusted HR=1.55; 95% CI: 1.23–1.97) and death with functioning graft (adjusted HR=1.72; 95% CI: 1.28–2.29) after transplantation with ECD kidneys compared with SCD kidneys. This conclusion could not be applied to older recipients (adjusted HR=1.11; 95% CI: 0.80–1.54, and adjusted HR=1.30; 95% CI: 0.89–1.89, respectively). The authors set that an excess risk of all-cause mortality on younger recipients of an ECD kidney is caused by the increase of cardiovascular mortality associated to worse renal function of the graft (and reduced GFR) obtained with this kind of donor. Chronic inflammation and the state of immunosuppression associated with uraemia lead to accelerated atherosclerosis that would justify these poor results.21 Our work agrees since the survival of younger recipients is clearly affected by the KDPI (Fig. 4a). It also strengths the importance of maintaining the principle of allocation ‘old for old’.

Hernandez et al. studied the odds of death and graft loss regarding organ quality measured by KDPI, divided in five categories, among cohorts of recipients with different age ranges.22 The survival of recipients from 50 to 69 years receiving a low-quality graft is significantly reduced. However, recipients from 70 to 79 years receiving a very low-quality graft had a lower death risk than recipients from 50 to 69 years. Recipients older than 79 years had the lowest risk. These results are consistent with our study since the KDPI hazard ratio for patient survival is lower in the population over 60 years (HR=1.011) than among those aged 18–59 years (HR=1.013).

A major objective of the US allocation system is to allocate low-KDPI grafts (better quality) to patients with high estimated survival (based on the Estimated Post Transplant Survival, EPTS7). Thus, young recipients benefit from the best-quality donors. When a high-KDPI graft is allocated to a senior recipient, the deleterious effect on patient survival is manifest. However, our results state that recipients older than 60 years can also benefit of extended survival (as Fig. 4b shows). Although allocation systems in Spain and US are different, motivation is alike. The aim is to obtain a longevity matching, so cases of allocation of optimal donors to older recipients are uncommon.

Finally, there is a couple of potential limitations of this work. Firstly, KDPI is calculated by using the study population itself. It would be more appropriate to use a reference European donor population (as it is done in US) but there is no such registry available. The authors think that it is better to use the study cohort itself to calculate the KDPI than using the completely different population published by the US Organ Procurement and Transplantation Network.23 Secondly, for the retrospective registry analysis performed, some cases had to be excluded due to missing data. Missing data is due to random failure in manual registry on data introduction, so there is no evidence to suggest this has a negative effect on interpretation. Therefore, the analysis performed are so-called complete-case analysis, and no data imputation technique has been used to replace missing data.

In summary, this is one of the first studies to validate the use of KDPI in a Spanish cohort. However, due to certain limitations on the KDPI translation to Spain, creating a new index from Spanish or European data would be advisable.24 In this work, some predictive factors of graft survival are identified that can server as a first step in this path.

Although it is correlated significanty with the graft and patients’ survival, more studies should be performed aiming to increase the predictive power of kidney donor scores or complement them with other data sources that mean a real clinical aid to professionals in pretransplantation. Increasingly, recipients demand wider information about the graft quality to receive, and the KDPI is a right step in that direction. However, new technology trends may bring tools that integrate information from different data sources (clinical, biological, histological, …) about donors and recipients to perform survival estimations with higher predictive power.

The authors have no conflicts of interest to declare.