Trends in Usage and Outcomes for Expanded Criteria Donor Kidney Transplantation in the United States Characterized by Kidney Donor Profile Index

Abstract

There has been increasing concern in the kidney transplant community about the declining use of expanded criteria donors (ECD) despite improvement in survival and quality of life. The recent introduction of the Kidney Donor Profile Index (KDPI), which provides a more granular characterization of donor quality, was expected to increase utilization of marginal kidneys and decrease the discard rates. However, trends and practice patterns of ECD kidney utilization on a national level based on donor organ quality as per KDPI are not well known. We, therefore, performed a trend analysis of all ECD recipients in the United Network for Organ Sharing (UNOS) registry between 2002 and 2012, after calculating the corresponding KDPI, to enable understanding the trends of usage and outcomes based on the KDPI characterization. High-risk recipient characteristics (diabetes, body mass index ≥30 kg/m², hypertension, and age ≥60 years) increased over the period of the study (trend test p<0.001 for all). The proportion of ECD transplants increased from 18% in 2003 to a peak of 20.4% in 2008 and then declined thereafter to 17.3% in 2012. Using the KDPI >85% definition, the proportion increased from 9.4% in 2003 to a peak of 12.1% in 2008 and declined to 9.7% in 2012. Overall, although this represents a significant utilization of kidneys with KDPI >85% over time (p<0.001), recent years have seen a decline in usage, probably related to regulations imposed by Centers for Medicare & Medicaid Services (CMS). When comparing the hazards of graft failure by KDPI, ECD kidneys with KDPI >85% have a slightly lower risk of graft failure compared to standard criteria donor (SCD) kidneys with KDPI >85%, with a hazard ratio (HR) of 0.95, a confidence interval (CI) of 0.94-0.96, and statistical significance of p<0.001. This indicates that some SCD kidneys may actually have a lower estimated quality, with a higher Kidney Donor Risk Index (KDRI), than some ECDs. The incidence of delayed graft function (DGF) in ECD recipients has significantly decreased over time from 35.2% in 2003 to 29.6% in 2011 (p=0.007), probably related to better understanding of the donor risk profile along with increased use of hypothermic machine perfusion and pretransplant biopsy to aid in optimal allograft selection. The recent decline in transplantation of KDPI >85% kidneys probably reflects risk-averse transplant center behavior. Whether discard of discordant SCD kidneys with KDPI >85% has contributed to this decline remains to be studied.

Keywords: clinical outcomes; deceased donor kidneys; expanded criteria donor; kdpi: kidney donor profile index; kdri: kidney donor risk index; kidney graft survival; kidney transplant; post-transplant mortality; standard criteria donor.

Figure 1. Modified Bailey-Makeham model for risk of graft failure

Formula: Risk(t) = αe^-gt (short-term) + βt + δ (long term). The y-intercept signifies inital excess risk "α," the vertical slope is indicative of risk of decline "γ," and the horizontal slope implies long-term risk of graft failure βt + δ.

Figure 2. Trend in donor characteristics by the ECD definition

Figure 3. Trend in utilization of ECD and KDPI >85% kidneys across the eras

Rates of use of ECD and KDPI >85% kidneys increased during the time period of interest.

Figure 4. Projected linear regression trend line for ECD kidney utilization

Trend line 2003-2008; y=0.0075x + 0.17. A linear regression trend line between 2003 and 2008 estimates the rate of ECD utilization to be around 23.8% in 2012. On the contrary, the proportion of ECD transplants has declined to 17.3% in 2012.

Figure 5. Trend of the use of kidneys by ECD and KDPI classification

The use of kidneys with KDPI >85% has declined in the most recent era (2007-2012), despite the upward trend in use in the previous two eras (2000-2002 and 2003-2006).

Figure 6. Percentage of ECD kidney transplants with DGF by year of transplant

Incidence of DGF has significantly decreased from 35.2% in 2003 to 27.7% in 2010, with a slight increase to 29.6% in 2011 (p=0.007). On multivariable logistic regression, the odds of DGF have significantly decreased per incremental year of transplantation (OR: 0.98; CI: 0.96–0.997; p=0.018).

Figure 7. Kaplan-Meier graft survival curves for the entire cohort: ECD, KDPI 85%

Kaplan-Meier analysis demonstrates the improvements in graft survival for all ECD kidneys (KDPI < and > 85%) across the eras.

Figure 8. Kaplan-Meier graft survival curves for KDPI >85%

Kaplan-Meier curve demonstrating improvement in graft survival across the eras among allografts with KDPI >85%.

Figure 9. Risk of graft failure for the entire cohort across the eras

The modified Bailey-Makeham model for risk of graft failure shows: 1) the initial risk of graft loss (α, the y-intercept) drops across eras; 2) non-significant change in the rates of decline in graft loss (γ, the slope) across eras in approximately the first year; and 3) improvement in the long-term risk over time (βt + δ) in the recent era (2007-2012).

Figure 10. Risk of graft failure characterized by KDPI 85%

Modified Bailey-Makeham model for risk of graft failure for kidneys with KDPI >85% demonstrates no difference in initial risk (α, y-intercept) and a slower decline in the risk of graft loss in the first year (γ, the slope) with the risk of graft loss over time accelerating beyond that of lower KDPI organs (βt + δ).

Figure 11. Risk of graft failure for KDPI >85% across the eras

Modified Bailey-Makeham model for risk of graft failure from the time of transplant for KDPI >85% across eras: 1) Era Two, 2003-2006, has a lower initial risk of graft loss (α, the y-intercept), while the initial risk is the same for the first and third eras; 2) Era Three, 2007-2012, has a prominent rapid decline in risk of graft failure within the first two months (γ, the slope); and 3) minimal change is seen in the long-term risk of graft loss over the years (βt + δ).