Home | Business News | Browse by Publication | J | Journal of the American Statistical Association

Frailty survival model analysis of the national deceased donor kidney transplant dataset using Poisson variance structures.

Article Excerpt
1. INTRODUCTION

More than 350,000 patients are receiving chronic dialysis treatment for end-stage renal disease (ESRD) kidney failure in the United States today. Although kidney transplantation was originally thought to be primarily a quality of life-enhancing treatment, patients receiving kidney transplantation have been found to have substantially longer lifetimes than patients remaining on dialysis (Wolfe et al. 1999). Most transplanted kidneys are obtained from deceased donors, although a substantial fraction are donated by living donors. Transplantation is a critical medical procedure that typically involves quick action to identify a potential donor, assemble a medical team to remove the organ, identify a candidate from among those awaiting a transplant, transport the organ, and implant the organ. The rules used to prioritize the candidates awaiting a transplant have been developed to consider both efficiency and equity. Efficiency goals include ensuring that organs are transplanted before they become unusable and that organs are allocated to candidates who will benefit the most. Equity considerations assure that the allocation system is "fair." Although the details of the rules are complicated and differ by region of the country, candidates generally are registered on a waiting list, and deceased donor kidneys are offered to candidates ordered with respect to their waiting times.

Unfortunately, there are many more patients joining the waiting list for deceased donor kidney transplantation than there are organs available each year, which has led to longer waiting times and more deaths among wait-listed candidates. The number of patients on the national kidney waiting list has increased from 22,063 in 1992 to 51,144 in 2001 (132%), whereas the number of kidney waiting list deaths has increased from 1,077 to 2,918 in those same years (171%). However, the corresponding numbers of deceased donor kidney transplant recipients has increased only moderately, from 7,202 to 8,202 (14%). Although the absolute number of deaths has increased, posttransplant death rates in the first year for patients age 18-34 have declined from 30 to 15 per 1,000 patient years between 1992 and 2001, due in large part to improvements in immunosuppressant therapy (OPTN/SRTR 2003, table 5.7; Merion 2003). One option for meeting the shortage of organs is to use organs that carried too high of a risk of failure in previous years but that have acceptable levels of risk with current transplant outcomes. Kidneys from deceased donors age 60 years or older and from those age 50-59 years with at least two of three defined conditions [cerebrovascular accident as the cause of death (CVA), renal insufficiency (serum creatinine >1.5 mg/dL), and history of hypertension] have been identified as expanded criteria donor (ECD) kidneys (Port et al. 2002), that is, kidneys whose graft loss relative rate was greater than 1.70 when compared with a low-risk group. Until recently, kidneys from such donors were often considered "marginal" or unsuitable. Ojo et al. (2001) showed that although outcomes from ECD kidneys are worse than those from standard kidneys, both life expectancy and quality of life are improved for ECD kidney recipients compared with patients remaining on dialysis treatment.

Increasing the use of higher-risk donor organs is intended to lead to outcomes that are no worse than were deemed acceptable in previous years, while potentially increasing the current number of available donors dramatically. A 70% increase in failure rates compared with today's rates may lead to a post-transplant death rate of 26 per 1,000 patient years, which is lower than the risk reported for 1992. In recognition of the fact that risk varies by more than a factor of 2 among types of donors, the current transplant policy provides informed decision making on the part of potential recipients by identifying which organs are from higher-risk donors. This was recently implemented by creating a separate registration process to be added to the waiting list for these higher-risk organs. Patients on the alternate waiting list are prioritized for ECD kidneys in addition to non-ECD kidneys, which could substantially shorten a candidate's waiting time and thus reduce that candidate's probability of waiting list death. The degree of confidence in the classification of higher-risk organs depends not only on the magnitude of the differences among donors based on their measured characteristics, but also on the remaining unexplained variation among donors. Frailty (or random-effects) survival models offer direct estimates of the variation among donors in recipient death rates, even in this case with only two replications based on two kidneys per donor transplanted into two different recipients. Simulation results reported in this article document the accuracy of such analyses. It is shown here that the ECD characteristics identified by Port et al. (2002) explain nearly 30% of the variability among donors.

In addition to variation among donors, there is substantial variation among transplant centers in recipient survival. The decision to accept an organ for transplantation lies with both the transplant surgeon and the patient. When kidneys become available from a deceased donor, they are offered for specific patients according to a regional and national allocation system that considers donor and recipient matching of blood group and tissue type, level of recipient preformed antibodies, and priority based on each candidate's time on the transplant waiting list. Patients may choose in advance to be eligible for organs with reduced probability of function according to ECD characteristics (a binary variable). The transplant surgeon may turn down an offered organ because of poor quality or other factors. Such a turndown leads to reallocation to the next ranked candidate, usually at another center. A surgeon may choose to accept this offer even if the offer was previously rejected at another center. Thus center factors play an important rule, even though the offers of organs are regulated by well-defined allocation rules. Because some centers tend to use more higher-risk donors than others, it is important to account for both donor and center factors simultaneously when evaluating the magnitude of either factor. Outcomes of organ function and patient survival are remarkably complete; the Scientific Registry of Transplant Recipients (SRTR) ascertains organ failure and death through both reports from transplant centers and from matching with Social Security and Medicare files.

In ascertaining the effectiveness of the ECD criteria in classifying high-risk organs, it is of great interest to find out how much of...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!