|
|
| |||||||||||||||||||||||||||||||||||||||
Background The methods used to allocate cadaveric kidneys in the United States have been criticized as being unfair to minorities because of an overemphasis on HLA matching. We evaluated a new HLA-matching method that might alleviate this problem.
Methods We used data from the United Network for Organ Sharing (UNOS) Kidney Transplant Registry to evaluate and project the outcome of cadaveric kidney transplantation. An HLA-matching method based on compatibility at 10 key amino acid residues of HLA-A and B molecules and a limited number of HLA-DR types was evaluated with use of the HLA types of the patients currently on the national waiting list and waiting lists in Los Angeles and Birmingham, Alabama.
Results With national kidney sharing, the projected 10-year rate of graft survival for transplants in which there were no HLA-A, B, or DR mismatches was 66 percent, as compared with 39 percent for transplants with more than two HLA-A, B, or DR mismatches. With local sharing, 43 percent of patients could be fully matched, and they had a projected 10-year graft-survival rate of 50 percent. When one HLA-DR mismatch was allowed, the projected 10-year graft survival was 46 percent, and 67 percent of patients waiting locally could receive such grafts. Even in Alabama, where 68 percent of the patients on the waiting list are black, 48 percent of waiting patients could obtain a matched kidney.
Conclusions Inserting two new HLA-matching categories into the UNOS point system for cadaveric kidney allocation would increase the number of patients for whom matches could be found in local pools.
The current kidney-allocation system of the United Network for Organ Sharing (UNOS) incorporates elements of each philosophy, allotting points for waiting time, presence of preformed anti-HLA antibodies, age, and levels of HLA compatibility. The patient with the most points rises to the top of the waiting list. The physician makes the final judgment whether or not that patient will receive an available kidney. When UNOS identifies a perfect HLA match somewhere in the United States, the kidney is sent to that recipient. Otherwise, most kidneys are transplanted locally. Through the UNOS six-antigen matching program, in which the recipient is matched to the donor for two antigens each in the HLA-A, B, and DR loci, projected survival at 20 years is 50 percent for grafts functioning at 1 year4. Such superior long-term graft survival not only benefits the recipient but also decreases the demand for subsequent kidneys because these patients do not need retransplantation.
The problem is that only 5 percent of patients receive a perfectly matched kidney, and the next best match does not yield a sufficiently high success rate to justify national sharing5,6. Gaston et al.7 recently argued that efficiency in terms of optimal transplantation success should not be a primary determinant in kidney allocation, for if such a policy were taken to the extreme, no kidneys would be transplanted to blacks because 50 percent of grafts functioning at one year survive for only five years in blacks, whereas the comparable figure in whites is eight years.
We describe an allocation system that could achieve maximal kidney function while providing equity for patients. It increases the number of matched patients by using the concept of permissible HLA-DR mismatches and matching for 10 widely shared HLA determinants identified by key amino acid residues of HLA-A and B molecules8,9.
Methods
The subjects in the study were 30,672 recipients of first cadaveric kidneys who underwent transplantation between October 1987 and August 1993 and were reported to the UNOS Scientific Renal Transplant Registry.
The humoral response to HLA-A and B molecules is almost always directed toward antigenic determinants shared by several HLA alleles (public determinants) that compose a limited portion of the HLA molecule. In a previous study,9 we determined that donor-recipient compatibility could be achieved by avoiding donor mismatches for the following 10 groups of amino acid residues representing some public determinants: threonine (T) at position 41, phenylalanine (F) at position 67, threonine-asparagine-threonine-glutamine-threonine-aspartate-arginine-glutamate-serine (TNTQTDRES) at positions 69 through 77, serine (S) at position 70, alanine-glutamine-threonine-aspartate (AQTD) at positions 71 through 74, asparagine (N) at position 80, arginine (R) at position 83, glutamine (Q) at position 114, arginine (R) at position 114, and lysine (K) at position 127. We retrospectively estimated the number of public HLA mismatches using a computer program to convert the donor and recipient HLA-A and B locus molecules to their respective sequences, selecting the key residues, and considering any of the key residues present in the donor's type but absent in the recipient as mismatched. Sequence data were available for all HLA molecules reported.
Public determinants are not well characterized for HLA-DR molecules. Two levels of mismatch were used for this antigen to increase the number of compatible donor-recipient combinations. First, HLA-DR mismatches were determined after we had converted all HLA-DR specificities to their broad designations; for example, DR15 or DR16 in the phenotype was converted to DR2. Second, we allowed one HLA-DR mismatch, termed a "permissible HLA-DR mismatch," identified retrospectively as having a low frequency of immunologic loss (with no HLA-A and B mismatches) in transplants from living related donors and shown to predict superior survival in cadaveric transplants in two separate studies8,10.
Graft survival was calculated by the product-limit method, and tests for the significance of differences between groups were performed by comparing differences in proportions with their appropriate standard errors. Asymptotic normality was assumed during the estimation of standard errors. Mismatch projections were calculated as described by Mickey et al.11 The recipient pool included the 32,184 HLA-typed patients currently awaiting transplantation in the United States. The first 2000 organ donors in 1992 were used in projections. Only recipients with their donor's ABO blood type were considered for each match run. The impact of matching was calculated as described by Gjertson et al.12 One-year rates of graft survival and the projected lengths of time at which 50 percent of the grafts functioning after the first year would still be functioning were used to estimate the 5- and 10-year graft survival for each matched group. The overall graft survival was adjusted for the projected matched fractions as previously described12.
We tested the racial bias inherent in the various algorithms using the 783 patients awaiting transplantation in Los Angeles during August 1993 and 522 patients awaiting transplantation in Birmingham, Alabama, in October 1993 (data on the latter group were provided by Dr. Arnold Deithelm). Since approximately 200 donor kidneys are allocated in these regions annually, 200 sequential donors from the two regions were used to test the models.
Results
The projected 10-year rate of graft survival for 30,672 recipients of first cadaveric transplants listed according to the number of HLA points accumulated in the current UNOS allocation system is shown in Figure 1. The high rate of success for transplants with no HLA-A, B, or DR mismatches is conspicuous. Unfortunately, patients with optimal matches constituted only a small fraction of the transplant recipients, as shown in the bar graph in Figure 1. More than 75 percent of the patients received transplants with two to four HLA-B and DR mismatches.
|
|
The new HLA-match categories would identify a more racially balanced group of matched candidates in local pools, as shown in Figure 2. There is a comparison to be drawn between the distribution of candidates for transplantation in southern California, where there is a large Hispanic population, and the distribution in Alabama, where the majority of patients are black. The proposed matching categories identified a substantial fraction of blacks and nonwhites who could receive a matched graft. Most important, in both local pools, 30 to 40 percent of all patients in different ethnic or racial groups could obtain a matched graft.
|
|
|
|
The UNOS point system for kidney allocation as practiced in the United States has recently been criticized5,6,7. The most serious charge was that the emphasis on HLA matching discriminates against blacks. One proposed solution was to award extra points to black patients7. We object to this proposal on the grounds that extra points awarded from the first day to anyone on the basis of skin color or claims of ancestry are equally unjust for others. We proposed a two-pronged solution to the problem. First, by adjusting the way in which we look at HLA mismatches, HLA match points awarded for public HLA matching would be distributed more equally between whites and blacks. When matched for these broad HLA groups, patients of both races would benefit in terms of graft survival and the proportion of matched transplants. Second, under the proposed matching scheme, there would be more matched recipients from whom a selection could be made for each donor. This means that the matched patient who had waited the longest would be selected. If the matching system is too restrictive, too few patients are considered for a given donor and those who have been waiting for an especially long time are at a disadvantage.
The UNOS point system allows the adjustment of points for matching and waiting time. Matching, which represents the resource-driven factor, and waiting time, which represents the patient-driven factor, are the major components controlling distribution. On the basis of our findings, we propose altering the UNOS point system as indicated in Table 2. The most important modification is that only two factors would be considered: the HLA match, to improve long-term survival, and waiting time. Gradations in numbers of match points are made according to the projected 10-year rate of graft survival. If subsequent analyses show that a different point allotment for matching is warranted, the points can easily be changed. The basic point structure is the same as that used for the past five years, the only difference being the inclusion of two new categories.
With respect to waiting time, the new calculation allocates 0.1 point per month of waiting. Since many patients may receive the same number of match points in the public HLA matching categories, waiting time would serve as the tiebreaker. Such a system would also ensure that patients who have waited five or more years would be considered with recently listed matched recipients in the top four groups and promoted to the next level with each year of waiting. Candidates for a second or subsequent transplantation and those who have preformed anti-HLA antibodies against many donors may wait longer but will gain points simply by waiting. Patients with rare HLA types could also achieve equity by accumulating waiting points.
It is particularly important to note that the proposed matching method can be very effective in local pools of patients awaiting a transplantation. Shipping kidneys, other than those with zero HLA mismatches, long distances may be unnecessary. If the method proves successful locally, national sharing could be considered for kidneys that could not be matched to a local patient. In the meantime, under the new system, kidneys shipped nationally as a payback for kidneys matched for six antigens could be allocated as a nationally matched group.
The HLA antigens are the main barrier to long-term success in kidney transplantation. The enormous diversity of these antigens makes good matches between donors and recipients rare, and good matches are even rarer when the racial backgrounds of donors and recipients are dissimilar. These studies based on the outcomes of more than 30,000 kidney transplantations suggest that by avoiding mismatches of 10 public HLA-A and B determinants instead of 84 alleles and of 10 HLA-DR antigens instead of 18, we can achieve high rates of long-term success and provide matched kidneys for a more racially balanced population of waiting patients.
Supported in part by a grant (2R01 DK 02375-34A1) from the National Institute of Diabetes and Digestive and Kidney Diseases and a contract from the United Network for Organ Sharing.
We are indebted to Dr. Arnold Deithelm and Ms. Sharon Hudson for providing HLA-typing data for their waiting-list candidates and most recent donors.
Source Information
From the Tissue Typing Laboratory, Department of Surgery, UCLA.
Address reprint requests to Mr. Takemoto at the UCLA Tissue Typing Laboratory, 950 Veteran Ave., Los Angeles, CA 90024.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
HLA Matching in Renal Transplantation
Gaston R. S., Deierhoi M. H., Ayres I., Takemoto S., Terasaki P. I., Gjertson D. W., Cecka J. M., Held P. J., Port F. K., Hunsicker L. G., Kahan B. D., Sanfilippo F.
Extract |
Full Text
N Engl J Med 1995;
332:752-753, Mar 16, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
