However, models A to E did not account for whether each kidney was pumped, a practice that more than tripled from 1999 to 2009. Panel F (Figure 5) shows that after adjusting for the increased use of pumping throughout the 2000s (Figure 4), a statistically significant (P = 0.007), residual KDR trend re-emerged. In other words, if more kidneys had not been pumped during this decade, it is estimated that the observed KDR trend would have risen even more sharply than the observed 5.7% rise.
The full model (Figure 5, panel F), which included all available donor factors, the use of biopsies, % glomerulosclerosis, and the percent of kidneys pumped, explained 82.5% of the 10-year increase in the KDR. Had the distribution of all modeled factors remained stable throughout this period, this method estimates that the KDR would still have increased, but only by about 1% (from 15.8% to 16.8%). On the odds scale, this translates into a 12% increased odds had these factors remained stable, compared with the observed 54% rise in the odds of discard from 1999 to 2009 (OR10, Table 1).
The results of this progressive model-building (steps A-F) are summarized in Table 1. Factor-specific odds ratios from the Model F appear in Table 2.
Propensity Analysis (October 1999-2000 vs 2009)
To complement and validate the multivariable regression findings above, we performed a propensity analysis comparing matched samples of recovered kidneys from the beginning (subperiod 1: October 1999 to December 2000) and end (subperiod 2: January 2009 to December 2009) of the 10-year period. Of the 12 911 and 14 394 kidneys recovered in subperiods 1 and 2, respectively, 7276 from each subperiod were matched based on propensity score (Table 3).
Characteristics of the matched samples were nearly identical, with all absolute standardized differences falling well below the customary 10% threshold26 and none reaching statistical significance (Table 3). Before matching, the 2009 KDR was 19.2% compared with 14.7% for the 1999 to 2000 subperiod, an odds ratio of 1.38. After matching, the 2009 KDR was 16.5% compared with 16.2% for the 1999 to 2000 subperiod, a residual odds ratio of 1.07 after accounting for same-donor clustering. This finding of a small, residual KDR increase after adjusting for changes in the available characteristics of recovered kidneys is consistent with the results from regression model F.
The residual odds ratio of 1.07 represents an 82.1% reduction from 1.38. In terms of both relative risk and risk difference, the proportion explained is also about 80%, which is very similar to the 82.5% explained and estimated by the regression approach. The lack of statistical significance (P = 0.24) in this residual difference could be a byproduct of the reduced sample sizes attributable to matching; it may also suggest that, all else being equal, the likelihood of discard for a kidney recovered in 2009 was no different than for an identical kidney recovered in 2000.
Together, the regression and propensity analyses suggest that at least 80% of the long-run increase in KDRs can be attributed to changes in the available characteristics of kidneys offered for transplantation. However, a residual increase was evident in both analyses.
OPO-Recorded Discard Reasons Over Time (1996-2015)
Biopsy findings has consistently been the most common reason recorded by OPOs for the discard of kidneys. Despite the proportion of kidneys biopsied more than doubling during 1999 to 2009 (Figure 4), the percent of discards with biopsy findings cited as the primary reason remained relatively stable at approximately 40% during this period (Figure 6).
Starting in 2008, the percentage of discards attributed to no recipient located—list exhausted (indicating the OPO attempted but was unable to find a transplant center willing to accept the kidney) began to rise steadily, increasing from less than 10% to nearly 30% in 2015. And although biopsy findings was still the most often-cited discard reason in 2015, the percentage has declined to 32.9%.
The 3 other most common discard reasons have been poor organ function (10.9%), anatomical abnormalities (8.0%), and diseased organ (4.9%).
OPO-Specific Trends in the Kidney Discard Rate (1987-2009)
Most OPOs experienced a similar rising trend in KDRs during this 22-year period (Figure 7A). OPO-specific odds ratios (OR) (OR10), quantifying the increased odds of discard per decade, ranged from 0.90 to 3.94, with the vast majority (78%) between 1.25 and 2.25 (Figure 7B). Three OPOs saw notably larger increases than others, with OR10 values near 4.0. No OPOs experienced a statistically significant decline in KDRs during this period.
Several outliers were apparent. Compared with other OPOs, a far higher percentage of kidney donors recovered in OPO 1 and OPO 2 had Kidney Donor Profile Index exceeding 85%, which appears to explain their sharp rise in KDR during the late 2000s. The increasing slopes for these 2 OPOs were statistically significant at the 0.01 level even after adjustment for multiple comparisons (Bonferroni method).27 Despite a sharp increase in less-than-ideal kidney donors recovered in the 2000s (mirroring the national trend), the KDR for OPO 3 remained stable (OR10 = 0.94, P = 0.45). The abnormally high KDR for OPO 3 in the 1980s and 1990s does not appear to be explained by recovery of kidneys from older donors (Figure 7).
The vast majority (at least 80%) of the KDR rise in the 2000s can be explained by a changing donor pool and practices related to biopsy and pumping. A small, residual increase in the KDR cannot be explained by changes in available factors and hence derives from other cause(s). The regression modeling results suggest that the odds of discard for a kidney recovered in 2009 were 12% higher than an identical kidney (including whether biopsied or pumped) recovered in 1999.
It is conceivable that changes in other, unreported donor characteristics could be responsible for the residual increase in the KDR. Increased risk status (according to the Public Health Services28,29) and pumping resistance, for example, were not available for the entire time period and thus excluded from the analysis. However, the proportion of pumped kidneys with resistance exceeding 0.4 mm Hg/mL per minute and donors labeled Public Health Services increased risk changed a little from 2004 to 2000; still, it is possible that the underlying reasons donors received the increased risk label changed over time. Furthermore, though kidney biopsies and pumping may be done routinely, they may also be performed based on indication; as such, observed effects associated with these factors may be partially attributable to unmeasured factors triggering the decision to biopsy or pump. And though we were able to adjust for history of hypertension, we could not account for potentially varying degrees of severity among donors.
Despite these limitations, the veracity of our central findings is strengthened by the use of 2 complementary statistical methods that produced similar results. Furthermore, to alter our key findings, an unaccounted-for factor would have to be highly associated with the likelihood of discard, predominantly uncorrelated with all factors we did include, and have had a substantial distributional change over time among recovered kidney donors.
Other study limitations include the fact that OPOs can only record a single discard reason even though the underlying cause may be multifactorial. Reporting of these reasons is subjective and varies across OPOs, so although the sharp changes in reporting patterns may reflect a changing dynamic in the underlying causes of kidney discards, these findings should not be overinterpreted.
A residual, unexplained increase of 1% in the KDR translates into an excess of 938 discards during 2010 to 2015. This estimate is derived by hypothesizing a KDR of 17.5% instead of the observed 18.5% during this period.
With a growing gap between supply and demand during the 2000s, why would a kidney donated in 2009 be more likely to be discarded than an identical kidney donated in 1999? Increased risk aversion, manifested in transplant programs lowering their acceptance rates for less-than-ideal kidneys, is possible. In an effort to promote quality improvement, the OPTN, Centers for Medicare and Medicaid Services, and private insurers have increasingly relied on program-specific reports that grade performance based on transplant graft and recipient survival rates.30,31 Though these outcome measures have been shown to be well risk-adjusted,32 some centers may still have concerns about unadjusted confounders.33,34 Since 2007, when CMS issued its conditions of participation,35 which deemed transplant hospitals with lower than expected posttransplant survival rates to be out of compliance, some hospitals have become more selective in accepting kidneys36 or reduced their transplant volume.11
Because of these concerns, the OPTN's Membership and Professional Standards Committee is exploring revisions to routine performance monitoring to avoid discouraging the acceptance and utilization of imperfect but transplantable kidneys. A Health Resources and Services Administration–sponsored initiative led by UNOS—the Collaborative Innovation and Improvement Network—is investigating an entirely new paradigm for measuring, monitoring, and improving quality in transplant programs.37 Reese et al4 effectively highlight other potential program oversight, financial reimbursement, and allocation policy levers for maximizing kidney utilization. Ex vivo organ repair or other donor management interventions38 that improve organ quality may also hold promise for optimizing the supply of kidneys.
We have shown that most of the long-term rise in kidney KDRs can be explained by expansion of the donor pool. The push to identify and recover more donors in the early 2000s may have increased discards,13-16 but it also spurred a sharp rise in transplants. Compared with the 8540 deceased donor kidney transplants in 2002, 10 660 were performed in 2006, a 25% rise in just 4 years. More transplants at the cost of more discards is a better reality for patients than a stagnant transplant practice.
Still, these and others' results suggest that a significant number of opportunities for kidney transplant are missed.4 Not only has the donor pool gradually expanded, but the pool of candidates added to the kidney waiting list has also become older and increasingly comorbid.39 Such patients may realize a survival advantage from a less-than-ideal kidney compared with remaining on dialysis.40,41 Despite more waiting list candidates who stand to benefit from such kidneys, center practices on whether to even receive such offers vary substantially.42
Our findings suggest that the increase in the percent of kidneys pumped from 10% to 30% during the 2000s prevented the KDR from rising even higher than the observed 19.2%. Had pumping rates not increased, our model predicts the KDR would have risen by an additional 1%, translating into 878 kidneys spared from discard and instead transplanted due to pumping during 2010 to 2015. Though pumping may on occasion cause discard by revealing concerning performance measures (eg, high vascular resistance), the therapeutic43 and logistical (allowing more flexibility in scheduling surgery) benefits appear to more than compensate for any negative impact on utilization.
Consequently, future changes in pumping practice may have potential to reverse the kidney discard trend. In 2015, the percentage of recovered kidneys pumped ranged from 0% to 74% across the 58 OPOs, revealing marked disparities in practice. KDRs may have increased in the wake of the new kidney allocation system,44 and a decline in pumping may have contributed.45 The cost savings associated with minimizing delayed graft function can exceed the cost of pumping,46-48 suggesting expanded use of pumping nationwide for less-than-ideal kidneys with a high likelihood of discard is not untenable.
By contrast, the decision to perform a biopsy is associated with markedly increased risk of discard,23,24 despite questionable reliability and diagnostic value.49 Our results suggest that increased biopsy rates contributed to the long-term rise in the KDR. Kidney transplantation without routine use of biopsies has proven successful in Europe,50 suggesting reduced reliance on biopsy findings in clinical decision making might increase kidney utilization in the United States. Alternatively, biopsy interpretation could be centralized and performed by expert pathologists.
Why did the long-term rise in the KDR abruptly stop after 2009? The long-term rising trends in donor age and KDRI came to a halt at that same time, and in fact, median donor age actually fell from 43 to 39 from 2009 to 2015 (Figure 2). Any potential drop in the KDRs due to this trend reversal may have been offset by a continuation of the residual KDR increase that we have found, possibly due to increased risk aversion or allocation inefficiency.
We have found that though the long-term rise in the KDR is largely explained by expansion of the donor pool, a residual increasing trend is evident. This increase may be due to behavioral and/or other systemic factors, such as heightened aversion among transplant programs to accepting imperfect kidneys and allocation system inefficiencies. Our and others' findings also suggest that more routine pumping of these kidneys, which has been shown to improve outcomes and increase utilization, may be a potent and cost-effective way to increase the organ supply by reducing discards. Overcoming logistical barriers to pumping in an era of increased kidney shipping44 may be particularly important.
The authors would like to acknowledge Frank Delmonico, whose ideas for related work spawned this study, Heather Neil for article preparation and submission, and the rest of the UNOS Research department for their behind-the-scene contributions.
1. Organ Procurement and Transplantation Network online data reports. https://optn.transplant.hrsa.gov/data/view-data-reports/national-data/ Accessed 5/2/2016.
2. United States Renal Data System. http://http://www.usrds.org
/qtr/default.aspx Accessed 4/19/2016.
3. Rodrigue JR, Schold JD, Mandelbrot DA. The decline in living kidney donation in the United States: random variation or cause for concern? Transplantation
4. Reese PP, Harhay MN, Abt PL, et al. New solutions to reduce discard of kidneys donated for transplantation. J Am Soc Nephrol
5. Singh SK, Kim SJ. Epidemiology of kidney discard from expanded criteria donors undergoing donation after circulatory death. Clin J Am Soc Nephrol
6. Hall IE, Schroppel B, Doshi MD, et al. Associations of deceased donor kidney injury with kidney discard and function after transplantation. Am J Transplant
7. Klassen DK, Edwards LB, Stewart DE, et al. The OPTN deceased donor potential study: implications for policy and practice. Am J Transplant
8. Matas AJ. Sometimes zero is the correct answer. Am J Transplant
9. Shapiro R, Halloran PF, Delmonico FL, et al. The ‘two, one, zero’ decision: what to do with suboptimal deceased donor kidneys. Am J Transplant
10. Schold JD, Kayler LK, Howard RJ, et al. Increased regulatory oversight of center performance led to rare declines in kidney transplantation rates? American Transplant Congress Abstracts
. 2009 Abstract 272.
11. Schold JD, Buccini LD, Srinivas TR, et al. The association of center performance evaluations and kidney transplant volume in the United States. Am J Transplant
12. Massie AB, Zeger SL, Montgomery RA, et al. The effects of DonorNet 2007 on kidney distribution equity and efficiency. Am J Transplant
13. Metzger RA, Delmonico FL, Feng S, et al. Expanded criteria donors for kidney transplantation. Am J Transplant
. 2003;3(Suppl 4):114–125.
14. Shafer TJ, Wagner D, Chessare J, et al. US organ donation breakthrough collaborative increases organ donation. Crit Care Nurs Q
15. Khan AS, Shenoy S. What did we really learn from the collaborative? Is it in our best interest to use “every organ every time” in kidney transplantation? Curr Transplant Rep
16. Sung RS, Guidinger MK, Lake CD, et al. Impact of the expanded criteria donor allocation system on the use of expanded criteria donor kidneys. Transplantation
18. Rao PS, Schaubel DE, Guidinger MK, et al. A comprehensive risk quantification score for deceased donor kidneys: the Kidney Donor Risk Index. Transplantation
19. Pasta David J, Cisternas Miriam G. “Estimating Standard Errors for CLASS Variables in Generalized Linear Models Using PROC IML,” Proceedings of SUGI 28. Paper
20. Haukoos JS, Lewis RJ. The Propensity Score. Jama
21. Rosenbaum PR, Rubin DB. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am Stat
22. Yang D, Dalton JE (2012). A unified approach to measuring the effect size between two groups using SAS®. In SAS Global Forum (Vol. 1).
23. Massie AB, Desai NM, Montgomery RA, et al. Improving distribution efficiency of hard-to-place deceased donor kidneys: predicting probability of discard or delay. Am J Transplant
24. Sung RS, Christensen LL, Leichtman AB, et al. Determinants of discard of expanded criteria donor kidneys: impact of biopsy and machine perfusion. Am J Transplant
25. Schold JD, Kaplan B, Howard RJ, et al. Are we frozen in time? Analysis of the utilization and efficacy of pulsatile perfusion in renal transplantation. Am J Transplant
26. Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Communications in Statistics-Simulation and Computation
27. Rice J. Mathematical statistics and data analysis: Cengage Learning. 2006.
28. Guidelines for preventing transmission of human immunodeficiency virus through transplantation of human tissue and organs. Centers for Disease Control and Prevention. MMWR Recomm Rep
29. Seem DL, Lee I, Umscheid CA, et al. United States Public Health S. PHS guideline for reducing human immunodeficiency virus, hepatitis B virus, and hepatitis C virus transmission through organ transplantation. Public Health Rep
30. Edwards E. 1991 Center Specific Graft and Patient Survival Rates. Washington, DC: US Dept of Health and Human Services
. United Network for Organ Sharing: Richmond, VA; 1992.
31. Kasiske BL, McBride MA, Cornell DL, et al. Report of a consensus conference on transplant program quality and surveillance. Am J Transplant
32. Snyder JJ, Salkowski N, Wey A, et al. Effects of high-risk kidneys on scientific registry of transplant recipients program quality reports. Am J Transplant
33. Pelletier RP, Phillips GS, Rajab A, et al. Effects of cardiovascular comorbidity adjustment on SRTR risk-adjusted Cox proportional hazard models of graft survival. Transplantation
34. Weinhandl ED, Snyder JJ, Israni AK, et al. Effect of comorbidity adjustment on CMS criteria for kidney transplant center performance. Am J Transplant
35. Centers for Medicare and Medicaid Services. Hospital Conditions of Participation: Requirements for Approval and Re-Approval of Transplant Centers To Perform Organ Transplants https://http://www.cms.gov
/Medicare/Provider-Enrollment-and-Certification/CertificationandComplianc/downloads/transplantfinal.pdf March 30, 2007.
36. Schold JD, Arrington CJ, Levine G. Significant alterations in reported clinical practice associated with increased oversight of organ transplant center performance. Prog Transplant
37. Organ Procurement and Transplantation Network press release on COIIN project. https://optn.transplant.hrsa.gov/news/advisory-council-formed-for-project-to-increase-kidney-utilization/ Published January 28, 2016. Accessed April 26, 2016.
38. Niemann CU, Feiner J, Swain S, et al. Therapeutic hypothermia in deceased organ donors and kidney-graft function. N Engl J Med
39. Matas AJ, Smith JM, Skeans MA, et al. OPTN/SRTR 2013 Annual Data Report: kidney. Am J Transplant
. 2015;15(Suppl 2):1–34.
40. Merion RM, Ashby VB, Wolfe RA, et al. Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA
41. Massie AB, Luo X, Chow EK, et al. Survival benefit of primary deceased donor transplantation with high-KDPI kidneys. Am J Transplant
42. Grams ME, Womer KL, Ugarte RM, et al. Listing for expanded criteria donor kidneys in older adults and those with predicted benefit. Am J Transplant
43. Moers C, Smits JM, Maathuis MH, et al. Machine perfusion or cold storage in deceased-donor kidney transplantation. N Engl J Med
44. Stewart DE, Kucheryavaya AY, Klassen DK, et al. Changes in deceased donor kidney transplantation 1 year after KAS implementation. Am J Transplant
. 2016 doi: 10.1111/ajt.13770.
45. Stewart D, Kucheryavaya A, Brown R, et al. Understanding the Initial Rise in Kidney Discard Rates Observed Post-KAS. In: Accepted abstract, American Transplant Congress
46. Jochmans I, O'Callaghan JM, Pirenne J, et al. Hypothermic machine perfusion of kidneys retrieved from standard and high-risk donors. Transplant Int
47. Balupuri S, Strong A, Hoernich N, et al. Machine perfusion for kidneys: how to do it at minimal cost. Transpl Int
48. Lodhi SA, Lamb KE, Uddin I, et al. Pulsatile pump decreases risk of delayed graft function in kidneys donated after cardiac death. New Solutions to Reduce Discard of Kidneys Donated for Transplantation. Am J Transplant
49. Kasiske BL, Stewart DE, Bista BR, et al. The role of procurement biopsies in acceptance decisions for kidneys retrieved for transplant. Clin J Am Soc Nephrol
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
50. Cecka JM, Cohen B, Rosendale J, et al. Could more effective use of kidneys recovered from older deceased donors result in more kidney transplants for older patients? Transplantation