The effect of HLA match on renal graft survival has become controversial as has the policy of mandatory sharing of kidneys.
The effect of HLA match on renal graft survival has become controversial as has the policy of mandatory sharing of kidneys.
We performed a retrospective analysis of HLA matched (M) and mismatched (MM) kidney transplants in our center. Tacrolimus, mycophenolic acid, and steroids were used as maintenance therapy and basiliximab induction was added for high-risk patients.
A total of 229 kidney transplants were included with median follow-up of 5.1 years. The 5-year death-censored graft survival by Kaplan-Meier method was significantly higher in the M group than in the MM group for deceased-donor kidney transplants (log-rank, p = .018). This graft survival advantage was detected in patients with a peak panel reactive antibody (PRA) greater than 20% (p = .023), but not in those with a PRA level of less than 20% (p = .32). The graft survival was not statistically different for live donor kidney transplants (p = .077). A mismatched kidney was an independent risk for graft loss (hazard ratio: 2.27, 95% confidence interval: 1.009–5.09, p = .047) and acute rejection was a significant cause of graft loss in mismatched deceased-donor transplants (p = .035).
Acute rejection remains a significant cause of graft loss in HLA-6-antigen mismatched deceased-donor kidney transplants. Our data support mandatory sharing of HLA-matched kidneys in sensitized patients with a PRA level greater than 20%.
Kidney transplant is the preferred treatment for selected patients with end-stage renal disease, as it prolongs life expectancy, improves quality of life, and decreases cost of long-term medical care compared with maintenance dialysis treatment.1–5 The best method of allocating the limited number of available deceased-donor kidneys remains controversial.6 In 1987, the United Network for Organ Sharing (UNOS) established a national program of sharing HLA-matched deceased-donor kidneys (3 pairs of HLA-A, B, and DR antigens identical in both donor and recipient). Superior graft survival was well-documented with these shared kidneys.7 In August 1990, the HLA-matching criteria was expanded to “phenotypically matched kidneys” (e.g., donor A2, A2, B8, B13, DR3, DR4 to recipient A2, A3, B8, B13, DR3, DR4: donor has homozygous antigen A2 and is phenotypically matched to recipient).7 It was again expanded in March 1995 to “zero mismatched kidneys” (e.g., donor A2, – B8, B13, DR3, DR4 to recipient A2, A3, B8, B13, DR3, DR4: failure to identify the 2nd A is usually due to the presence of homozygous antigen).8
The lower incidence of rejection and the better graft survival of these HLA-matched kidneys was demonstrated in earlier studies.7–11 However, a recent analysis indicated that the impact of HLA matching has steadily declined as more potent immunosuppressive agents have been used.12 It has also been suggested that the mandatory sharing beyond the local allocation area for low sensitized patients with a panel reactive antibody (PRA) level less than 20% created logistical inefficiencies and tended to benefit white patients more than African American patients.13 In January 2009, UNOS updated its policy that HLA-matched kidneys continue to be nationally shared for sensitized adult patients with PRA levels greater than 20%, and the locally matched patients have the highest priority regardless of PRA level.
In this study, we analyzed the outcome of all HLA-matched kidney transplants that received potent immunotherapy with tacrolimus (TAC), mycophenolic acid (MFA), and steroids. We sought to answer the question of whether the aforementioned changes in policy would be supported using a single center's data in this era of modern immunosuppressive regimens.
We retrospectively compared the outcomes of HLA-matched and HLA-mismatched kidney transplants from January 1997 to December 2007. Any kidney transplant recipient who also received a pancreas, liver, or heart transplant was excluded from this study. Based on the 6 antigens at HLA-A, B, and DR loci, the HLA-matched group (M group) included all of the HLA identical, phenotypically matched, and zero mismatched kidneys. The HLA-mismatched group (MM group) was used as comparison and it represented no match at any of the 6 HLA antigens (6-antigen mismatch). Both kidney transplants from deceased donors and living donors were studied for graft survival analysis, and deceased-donor kidney transplants were further subgrouped into those with peak PRA levels greater than 20% and those less than 20%.
Our standard triple immunosuppression regimen consisted of steroids, TAC, and MFA. High risk patients defined as prior transplant recipients, HLA-6-antigen mismatch, and those with PRA levels greater than 20% received basiliximab induction therapy. Intravenous methylprednisolone was administrated prior to reperfusion and tapered to maintenance oral prednisone. TAC doses were adjusted to keep the 12-hour trough levels between 10 to 12 ng/mL for the first 3 months, 7 to 10 ng/mL for the remainder of the first year, and 4 to 7 ng/mL thereafter. Each patient received either mycophenolate mofetil (MMF) at 1 gram or enteric coated sodium mycophenolate (Myfortic) at 720 mg twice daily.
Acute rejection was confirmed by kidney biopsy. The severity of rejection was defined according to Banff criteria. Rejection of grade 1 or below was initially treated with IV methylprednisolone. Thymoglobulin was used for steroid resistant rejection, or as initial therapy for any rejection of Banff grade 2 or higher. Thymoglobulin, plasmapheresis, and intravenous immune globulin (IVIG) were used for antibody mediated rejection. Standard antifungal, antibacterial, and cytomegalovirus prophylaxis were administered per protocol.
Outcomes measures included: (1) death-censored graft survivals over 5 years, (2) incidence of biopsy confirmed and treated acute rejection, (3) quality of graft function as assessed by estimated glomerular filtration rate (eGFR) using the Modification of Diet in Renal Disease equation, and (4) etiologies for the graft loss. Renal graft loss was defined by primary non-function or loss of renal function requiring chronic dialysis. Patient death included all mortalities from the time of kidney transplantation. Death with a functioning graft (DWFG) was excluded from graft survival estimate (death-censored). Graft failure was excluded from graft function calculation.
Statistical analyses were performed using SAS version 9.1.3 software (Cary, North Carolina). A x2 or Fisher exact test was used for count data and a t-test for continuous measures. Product-limit estimates of survival curves were generated by Kaplan-Meier method and the survival difference was analyzed by log-rank test. A multivariable Cox proportional hazard regression analysis with a stepwise variable selection was performed to examine the risk factors for the graft loss. A p-value < .05 was considered statistically significant.
A total of 229 consecutive kidney transplants that met the study criteria were identified during the 11-year period. Median follow-up was 5.1 years (range, 12.1– 132 months) as of December 2008 and all were transplanted more than 1 year before this study. TableI summarizes the transplant recipient demographic characteristics of the 2 groups. There was no difference in age, sex, body mass index (BMI), warm ischemia time (WIT), or cold ischemia time (CIT) in either deceased-donor kidney transplants or live-kidney transplants between the 2 groups. However, African American ethnicity represented an overwhelmingly higher proportion in the MM group, while white ethnicity was the majority of the M group (p < .0001). The peak PRA was significantly higher in the M group than the MM group for deceased-donor kidney transplants. These different distributions become less significant in live donor kidney transplants (TableI).
|Deceased-donor Kidneys||(n = 73)||(n = 98)||p-value|
|Age (years)||48.33 ± 14.96||45.25 ± 15.17||.23|
|Female||33 (45.2%)||36 (36.7%)||.26|
|Male||40 (54.8%)||62 (63.3%)|
|African American||24 (32.9%)||72 (73.5%)||<.0001|
|White||49 (67.1%)||23 (23.5%)|
|Others||0 (0%)||1 (3%)|
|BMI||26.9 ± 6.04||27.36 ± 6.76||.64|
|Peak PRA (%)||37.90 ± 41.70||21.83 ± 33.8||.01|
|Previous Transplants||14 (19.2%)||11 (11.2%)||.14|
|CIT (hours)||18.91 ± 6.56||17.45 ± 6.96||.17|
|WIT (minutes)||28.92 ± 6.1||29.86 ± 8.52||.41|
|Live Kidneys||(n = 26)||(n = 32)||p-value|
|Age (years)||36.81 ± 12.45||39.88 ± 15.99||.53|
|Female||11 (42.3%)||13 (40.6%)||.92|
|Male||15 (57.7%)||19 (59.4%)|
|African American||4 (15.4%)||13 (40.6%)||.23|
|White||20 (74.1%)||17 (54.8%)|
|Others||2 (7.4%)||2 (6.5%)|
|BMI||27.66 ± 6.81||27.18 ± 7.64||.82|
|Peak PRA (%)||17.23 ± 30.33||9.75 ± 21.17||.29|
|Previous Transplants||5 (19.2%)||0 (0%)||.01|
|CIT (hours)||2.75 ± 6.38||1.61 ± 3.72||.43|
|WIT (minutes)||28.05 ± 7.07||27.23 ± 7.47||.67|
There were no differences in the graft function as measured by eGFR between the M and MM groups for both deceased-donor and live-kidney transplants (Table II). The 5-year cumulative incidence of biopsy confirmed and clinically treated acute rejection were significantly higher in the MM group than in the M group for both deceased-donor (Figure1A, 27.6% versus 9.6%, p = .003) and live-kidney transplants (Figure1B, 34 .4% versus 3.8%, p = .01). The relative risk (RR) of acute rejection was 2.94 (95% confidence interval [CI]: 1.32–6.23, p = .003) in mismatched deceased-donor kidney transplants, and 9.09 (95% CI: 1.23–64.9, p = .01) in mismatched live-kidney transplants.
|Deceased-Donor Kidneys Graft Function|
|eGFR at 1 year||56.12 ± 18.75||60.53 ± 21.71||.18|
|eGFR at 3 year||58.89 ± 21.1||55.88 ± 24.43||.53|
|eGFR at 5 year||60.93 ± 22.16||60.47 ± 28.23||.94|
|Causes of Graft Loss|
|DWFG||10 (55.5%)||9 (28.1%)||.1|
|CAN||5 (27.7%)||7 (21.8%)||.9|
|Rejections||0 (0%)||9 (28.1%)||.035|
|Others||3 (16.6%)||7 (21.8%)||.70|
|eGFR at 1 year||57.31 ± 15.83||61.2 ± 26.75||.53|
|eGFR at 3 year||56.43 ± 16.74||58.21 ± 20.33||.77|
|eGFR at 5 year||54.62 ± 20.51||58.45 ± 25.51||.68|
|Causes of Graft Failure|
|DWFG||2 (33.3%)||2 (22.2%)||.53|
|CAN||2 (33.3%)||3 (33.3%)||.62|
|Acute Rejections||0 (0%)||3 (33.3%)||.22|
|Others||2 (33.3%)||1 (11.1%)||.25|
The 5-year death-censored graft survival estimated by the Kaplan-Meier method was significantly higher in the M group than in the MM group for deceased-donor kidney transplants (Figure2). The estimated deceased-donor graft survivals at 1, 3, and 5-years were 100%, 89%, and 84% in the M group, and 93%, 79%, and 70% in the MM group (log-rank, p = .018). We further sub-analyzed deceased-donor graft survivals by their peak PRA levels (PRA <20% versus >20%). There was no survival difference between the M and MM groups for PRA levels less than 20% (Figure3A). The estimated graft survival at 1, 3, and 5-years were 100%, 82%, and 78% in the M group, and 94%, 81%, and 69% in the MM group (log-rank, p = .32). However, for sensitized patients with PRA levels greater than 20%, HLA-matched deceased-donor kidney transplants had a superior graft survival compared with mismatched transplants (Figure3B). The estimated graft survival at 1, 3, and 5-years were 100%, 96%, and 90% in the M group, and 88%, 75%, and 75% in the MM group (log-rank, p = .023). For live-kidney transplants, there was no statistical significance in the graft survival between the M and MM groups (Figure4). The estimated living-donor graft survival at 1, 3, and 5-years were 100%, 100%, and 89% in the M group, and 93%, 81%, and 72% in the MM group (log-rank, p = .077). The causes of graft loss are summarized in TableII. DWGF and chronic allograft nephropathy (CAN) were the main causes of graft loss. Acute rejection did not cause any graft loss in the M groups of both deceased-donor and live donor kidneys, but it was an important cause of graft loss in the MM groups, especially in the deceased-donor kidney transplants (p = .035).
The risk factors for graft loss were also examined by Cox's proportional hazard regression analysis and significant risk factors were further analyzed by a stepwise variable selection model. The risk factors included recipient age, sex, race, BMI, PRA, previous transplant, CIT, WIT, and mismatched kidneys. Mismatched kidneys and recipient age were found to be independent risk factors for graft loss in deceased-donor kidney transplants with a hazard ratio (HR) of 2.27 (95% CI: 1.009–5.09, p = .047) and 0.96 (95% CI: 0.093–0.97, p = .0001), respectively. After adjusting for acute rejection, the effect of mismatched kidneys loses statistical significance (HR: 1.84, 95% CI: 0.84–4.21, p = .142) indicating a mediating effect.
Earlier studies demonstrated superior graft survival of HLA-matched kidney transplants compared with mismatched transplants.7–11 This was mainly due to the lower incidence of rejection and graft loss from rejection in the matched kidneys. Over the last decade, several potent immunosuppressive agents have been introduced into clinical practice that have significantly reduced the incidence of rejection and improved graft survival.14 A recent analysis suggested that the impact of HLA match on graft survival has steadily declined.12 All of our patients received modern triple immunosuppressive agents as maintenance therapy. Our data indicated that HLA-matched deceased-donor kidney transplants still had significantly better 5-year graft survival than the mismatched deceased-donor transplants. However, this survival advantage was limited to the sensitized patients with PRA levels greater than 20%, and there was no survival difference in patients with PRA levels less than 20%. Our result supports UNOSs recently updated policy that mandatory sharing of HLA-matched kidneys should continue for sensitized adults with PRA levels greater than 20%. The intent was to increase kidney transplants for sensitized patients who otherwise would not receive an organ offer or have difficulty in matching an offered organ. Previous data have indicated that 65.7% of matched kidneys are allocated to the candidates with PRA levels less than 20%.13
We also found that there were a significantly higher proportion of African American patients (73.5%) in the MM group with more white patients (67.1%) in the M group, which supports the notion that sharing of matched kidneys benefited the white patients more than the African American patients.13 The peak PRA level was higher in the M group than the MM group, suggesting that kidney sharing may increase kidney transplantation in sensitized patients. The UNOS policy of sharing has been frequently criticized for the inevitably longer CIT and more severe ischemic injury in the shared kidneys.15 However, we did not find significantly longer CIT in the M group compared with the MM group. The graft function, as measured by eGFR, was also similar between the M group and the MM group.
The living-donor kidney transplants provided better short-term (1-year) and long term (5-year) graft survival than the deceased-donor kidney transplants in both the M and the MM groups. We did not see a significant survival difference between the M and MM groups for live donor kidney transplants. This illustrates that the effect of HLA match is less consequential in live donor than in deceased-donor kidney transplants.16–18 Living unrelated donor kidney transplants have better graft survival than deceased-donor kidney transplants despite more HLA mismatches in the former category.17–20 This has been explained by the fact that living-donor kidneys are usually healthier and have less ischemic injury than deceased-donor kidneys.17, 21
The cumulative incidence of acute rejection was significantly higher in the MM group than the M group with a RR of 2.94 for rejection in mismatched deceased-donor kidneys and a RR of 9.09 in mismatched live donor kidneys. Our rejection rates were similar to the rates reported previously.16, 22, 23 Acute rejection did not cause any graft loss in the M groups of both deceased-donor and live donor kidney transplants, but it was a significant cause of graft loss in the MM groups. Interestingly, HLA-mismatch was initially found to be an independent risk for graft loss in deceased-donor kidney transplants with a HR of 2.27. When adjusted for acute rejection, the effect of mismatch lost significance, suggests that acute rejection is the mediating factor of HLA mismatch for graft loss in deceased-donor kidney transplants. An interleukin-2 receptor antibody (basiliximab) was used as an induction for our high risk patients including all patients in the MM group. It was suggested that more potent induction therapy with thymoglobulin can decrease the incidence of acute rejection and may prolong graft survival in high risk patients.24–26 Therefore, induction with a T-cell depleting antibody should be considered for HLA-6-antigen mismatched deceased-donor kidney transplants.
This study provides the first single-center data examining the effect of HLA match on long-term graft survival. Among the strengths of this study is the fact that our patients were treated with 1 modern immunosuppressive regimen, the previous studies were based on either pooled multi-center data or UNOS data, and different immunosuppressive protocols were used in different transplant centers. Our study is limited by its retrospective nature and relatively small sample size in a single center that prevents the study of the impact of various HLA mismatching (1 to 5 HLA- antigen mismatch).
Our data indicate that mandatory sharing of HLA-matched kidneys remains beneficial for sensitized patients with PRA levels greater than 20%. It not only increases the access of kidney transplantation, but also provides superior long-term graft survival in sensitized patients. The impact of HLA match has become less consequential in low sensitized patients as well as living-donor kidney transplants. Acute rejection remains a significant cause of graft loss in HLA-6-antigen mismatched deceased-donor kidney transplants who were treated by basiliximab induction and tacrolimus, mycophenolic acid, and steroids as maintenance. More potent induction therapy may be needed to reduce the incidence of rejection and to improve graft survival in those high risk patients.
We thank the members of the Tulane Abdominal Transplant Institute for maintaining the transplant data base.