Histocompatibility and Donor Selection Issues

Impact of HLA allele mismatch on the clinical outcome in serologically matched related hematopoietic SCT

Article metrics

Abstract

In unrelated hematopoietic SCT (HSCT), HLA allele mismatch has been shown to have a significant role. To clarify the importance of HLA allele mismatch in the GVH direction in related HSCT, we retrospectively evaluated 2377 patients who received stem cells from an HLA serologically matched related donor in the GVH direction using the database of the Japan Society for Hematopoietic Cell Transplantation. The cumulative incidences of grade II–IV and grade III–IV acute GVHD in patients with an HLA allele-mismatched donor (n=133, 5.6%) were significantly higher than those in patients with an HLA allele-matched donor. Multivariate analyses showed that the presence of HLA allele mismatch was associated with increased risks of grade II–IV and grade III–IV acute GVHD. In particular, HLA-B mismatch and multiple allele mismatches were associated with an increased risk of acute GVHD. The presence of HLA allele mismatch was associated with an inferior OS owing to an increased risk of non-relapse mortality (NRM). In conclusion, the presence of HLA allele mismatch in the GVH direction in related HSCT was associated with increased risks of GVHD and NRM, which led to an inferior OS. HLA allele typing is recommended in related HSCT.

Introduction

Previous studies have shown that HLA allele mismatch significantly affects the clinical outcome after unrelated hematopoietic SCT (HSCT).1,2 Several retrospective studies have demonstrated that the presence of HLA allele mismatch is associated with an increased risk of GVHD in unrelated HSCT.3, 4, 5 Although the disparity of HLA molecules in HLA antigen mismatch is greater than that in HLA allele mismatch without HLA antigen mismatch, the impact of HLA mismatch on the clinical outcome was considered to be, for practical purposes, similar between antigen mismatch and allele mismatch, as reported previously.4,6,7 Although the impact of an HLA mismatch at each locus varied among the studies, there is a consensus that an HLA mismatch at any locus, including A, B, C and DRB1, is in general associated with a poor clinical outcome.2

In related HSCT, the importance of HLA allele mismatch has not yet been well established, because an HLA antigen-matched sibling is in most cases an HLA allele fully matched donor. In Japan, HLA compatibility in related HSCT is usually assessed serologically or by low-resolution DNA typing at three loci, including HLA-A, -B and -DR. However, when the donor is not a sibling, such as a parent or child, the probability of HLA allele mismatch between the recipient and the donor is expected to be higher than that between siblings. Furthermore, there may also be an HLA allele mismatch with a sibling if we consider recombination and mutation. The presence of one HLA antigen mismatch has been reported to be associated with a poor overall clinical outcome in related HSCT.8, 9, 10 Therefore, if the impact of allele mismatch is similar to that of antigen mismatch in related HSCT, as it is in unrelated HSCT, we could assume that the presence of HLA allele mismatch adversely affects the clinical outcome in related HSCT.

In this study, we assessed the impact of HLA allele mismatch on the clinical outcome in related HSCT using the database of the Japan Society for Hematopoietic Cell Transplantation (JSHCT), including patients without serological HLA mismatch in the GVH direction.

Patients and methods

Data collection

Data for all patients who received a first allogeneic HSCT from a serologically HLA-A, -B and -DR matched related donor in the GVH direction, irrespective of the number of mismatches in the HVG direction, between 1 January 2000 and 31 December 2011 were obtained from the Transplant Registry Unified Management Program, which includes data from the JSHCT.11 We excluded patients who lacked data on survival status. Overall, 7089 patients satisfied the above criteria. In further analyses, we considered only 2377 patients (33.5%) for whom information on allele typing at the HLA-A, -B, and -DRB1 loci was available. The study was planned by the HLA working group of the JSHCT and was approved by the data management committees of TRUMP and by the institutional review board of Saitama Medical Centre, Jichi Medical University, Saitama, Japan.

Histocompatibility

Histocompatibility data for serological and genomic typing for the HLA-A, -B and -DR loci were obtained from reports obtained from the institution at which the transplantation was performed. To reflect current practice in Japan, HLA matching in related donors was assessed by serological data for HLA-A, -B, and -DR loci. When the recipient’s antigens or alleles were not shared by the donor, this was considered an HLA mismatch in the GVH direction; when the donor’s antigens or alleles were not shared by the recipient, this was considered a mismatch in the host-versus-graft (HVG) direction.

End points and statistical analyses

The primary end point was the cumulative incidence of acute GVHD. Secondary end points included the cumulative incidences of neutrophil engraftment and non-relapse mortality (NRM) and the probability of OS. The physicians who performed transplantation at each center diagnosed and graded acute GVHD according to the standard criteria.12

A descriptive statistical analysis was performed to assess the patients’ characteristics. Medians and ranges are provided for continuous variables, and the percentages are shown for categorical variables. Patient’s characteristics were compared by using the Chi-squared test or the Fisher’s exact test for categorical variables. The probability of OS was calculated by the Kaplan–Meier method. A Cox proportional-hazards regression model was used to analyze OS. The cumulative incidences of NRM, GVHD and relapse were evaluated using the model of Fine and Grey13 for univariate and multivariate analyses. In the competing risk models for GVHD, relapse and death before these events were defined as competing risks. In the competing risk models for NRM, relapse was defined as a competing risk. Factors that were associated with a two-sided P-value of <0.10 in the univariate analysis were included in a multivariate analysis. We used a backward stepwise selection algorithm and retained only statistically significant variables in the final model. A two-sided P-value of <0.05 was considered statistically significant. The variables evaluated in these analyses were as follows: sex mismatch (female to male vs others), patient’s age at the time of HSCT (age 50 years vs age <50 years), disease risk (standard risk vs high risk), stem cell source (BM vs PBSC), relation to donor (sibling or others), ABO mismatch, use of in vivo T-cell depletion, performance status (0–1 vs 2–4), intensity of the conditioning regimen (myeloablative vs reduced intensity), GVHD prophylaxis (CYA based vs tacrolimus based), year of transplant (2007 vs <2007) and HLA disparity as assessed by allele typing of HLA A, B and DRB1. Standard risk was defined as the first or second CR of acute leukemia, the first or second chronic phase of chronic myeloid leukemia, myelodysplastic syndrome refractory anemia or refractory cytopenia with multilineage dysplasia, malignant lymphoma in CR or PR or non-malignant disease. High risk was defined as some other status of malignancy. All statistical analyses were performed with EZR (Saitama Medical Centre, Jichi Medical University, Saitama, Japan; http://www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmedEN.html), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria, version 2.13.0).14

Results

Patient characteristics

The patient characteristics are summarized in Table 1. The median age was 40 years (range, 0–74). Compared with recipients with an HLA allele-matched donor (Match group, n=2244), recipients with an HLA allele-mismatched donor (Mismatch group, n=133) were more likely to have a poor performance status, to receive a transplantation from a non-sibling donor, to receive a transplantation at an earlier time period, to receive tacrolimus for GVHD prophylaxis and to receive an in vivo T-cell depletion (Table 1). More patients in the Mismatch group received a transplant from a donor with an HLA mismatch in the HVG direction. In the Match group, the number of antigen mismatches in the HVG direction was 0 in 96.4%, 1 in 2.0%, 2 in 1.0% and 3 in 0.5%. In the Mismatch group, the number of antigen mismatches in the HVG direction was 0 in 56.3%, 1 in 33.1%, 2 in 7.5% and 3 in 3.0%. Information on HLA-C allele mismatch was available in only 1152 of 2377 (48.5%).

Table 1 Patient characteristics

GVHD

The cumulative incidences of grade II–IV acute GVHD were 29.5% (95% confidence interval (CI) 27.6–31.4%) in the Match group and 40.6% (95% CI 32.2–48.8%) in the Mismatch group (P=0.0018, Figure 1a). A multivariate analysis showed that the presence of at least one allele mismatch was associated with an increased risk of grade II–IV acute GVHD (hazard ratio (HR) 1.77, 95% CI 1.31–2.38, P=0.0002, Table 2). An increase in the number of HLA mismatches was associated with a statistically significant increase in the risk of grade II–IV acute GVHD. The cumulative incidences of grade II–IV acute GVHD were 38.8% (95% CI 29.9–47.6%) and 52.9% (95% CI 26.5–73.8%) in patients with one allele mismatch and multiple allele mismatches, respectively (P=0.0020, Figure 1b). Compared with the Match group, both the one allele-mismatched and multiple allele-mismatched cohorts were associated with an increased risk of grade II–IV acute GVHD in multivariate analyses (one allele mismatch: HR 1.61, 95% CI 1.17–2.22, P=0.0035; multiple allele mismatches: HR 3.52, 95% CI 1.64–7.59, P=0.0013). We also assessed the impact of each locus excluding patients with multiple allele mismatches. The cumulative incidences of grade II–IV acute GVHD were 25.0% (95% CI 11.6–41.0%) in HLA-A mismatch, 50.0% (24.8–70.9%) in HLA-B mismatch and 42.4% (30.3–54.0%) in HLA-DRB1 mismatch (Figure 1c). In a multivariate analysis, the presence of HLA-B or -DRB1 mismatch was associated with an increased risk of grade II–IV acute GVHD (HLA-A: HR 0.86, 95% CI 0.40–1.84, P=0.69; HLA-B: HR 2.33, 95% CI 1.18–4.63, P=0.015; HLA-DRB1: HR 1.83, 95% CI 1.22–2.72, P=0.0033).

Figure 1
figure1

Cumulative incidence of acute GVHD. Cumulative incidences of grade II–IV (ac) and grade III–IV (df) acute GVHD grouped according to (a, d) allele mismatch, (b, e) the number of allele mismatches and (c, f) locus of allele mismatches.

Table 2 Multivariate analysis

The cumulative incidences of grade III–IV acute GVHD were 9.5% (95% CI 8.3–10.8%) in the Match group and 21.8% (95% CI 15.2–29.2%) in the Mismatch group (P<0.0001, Figure 1d). A multivariate analysis showed that the presence of at least one allele mismatch was associated with an increased risk of grade III–IV acute GVHD (HR 2.39, 95% CI 1.60–3.58, P<0.0001, Table 2). Other factors that were associated with an increased risk of grade III–IV acute GVHD were use of PBSC and high disease risk. An increase in the number of HLA mismatches was associated with a significantly increased risk of grade III–IV acute GVHD. The cumulative incidences of grade III–IV acute GVHD were 19.8% (95% CI 13.1–27.6%) and 35.3% (95% CI 13.8–57.8%) in patients with one allele mismatch and multiple allele mismatches, respectively (P<0.0001, Figure 1e). Compared with the Match group, both the one allele mismatch and multiple allele mismatched cohorts were associated with an increased risk of grade III–IV acute GVHD in multivariate analyses (one allele mismatch: HR 2.12, 95% CI 1.36–3.30, P<0.0001; multiple allele mismatches: HR 4.73, 95% CI 1.88–11.87, P<0.0001). We also assessed the impact of each locus, excluding patients with multiple allele mismatches. The cumulative incidences of grade III–IV acute GVHD were 9.4% (95% CI 2.3–22.6%) in HLA-A mismatch, 38.9% (16.7–60.8%) in HLA-B mismatch and 19.7% (11.1–30.2%) in HLA-DRB1 mismatch (Figure 1f). In a multivariate analysis, the presence of HLA-B mismatch or HLA-DRB1 mismatch was associated with an increased risk of grade III–IV acute GVHD (HLA-A: HR 0.89, 95% CI 0.29–2.68, P=0.830; HLA-B: HR 4.74, 95% CI 2.00–11.28, P<0.0001; HLA-DRB1: HR 2.16, 95% CI 1.22–3.85, P=0.0009).

To exclude the possibility that HLA antigen mismatch in the HVG direction may affect the incidence of acute GVHD, we performed a subgroup analysis that included patients without HLA antigen mismatch in the HVG direction. In this subgroup analysis, the cumulative incidences of grade II–IV and grade III–IV acute GVHD in the Mismatch group were significantly higher than those in the Match group (grade II–IV 41.3% vs 29.5%, P=0.010; grade III–IV 24.0% vs 9.6%, P<0.0001). In multivariate analyses, the presence of an HLA allele mismatch in the GVH direction was still associated with increased risks of grade II–IV and grade III–IV acute GVHD (HR 1.75, 95% CI 1.30–2.35, P=0.0002; HR 2.39, 95% CI 1.60–3.58, P<0.0001, respectively).

Graft failure

The cumulative incidence of neutrophil engraftment at 60 days was 96.3% (95% CI 95.4–97.0%) in the Match group and 90.4% (95% CI 83.6–94.5%) in the Mismatch group (P=0.0044). Although the presence of HLA antigen mismatch in the HVG direction was associated with an increased risk of graft failure in a multivariate analysis (HR of engraftment 0.79, 95% CI 0.65–0.95, P=0.013), the presence of at least one allele mismatch in the GVH direction was not associated with an increased risk of graft failure.

NRM and relapse

The cumulative incidences of NRM at 2 years were 13.7% (95% CI 12.3–15.3%) in the Match group and 19.2% (95% CI 12.8–26.6%) in the Mismatch group (P=0.022, Figure 2a). A multivariate analysis showed that the presence of at least one allele mismatch was associated with an increased risk of NRM (HR 1.64, 95% CI 1.11–2.41, P=0.012, Table 2). The cohort with a one allele mismatch was associated with an increased risk of NRM, compared with the allele-matched cohort, in a multivariate analysis (one allele mismatch HR 1.83, 95% CI 1.18–2.84, P=0.0073; multiple allele mismatch HR 0.93, 95% CI 0.22–3.94, P=0.92). We also assessed the impact of each locus excluding patients with multiple allele mismatches. The cumulative incidences of 2-year NRM were 29.3% (95% CI 14.2–46.2%) in HLA-A mismatch, 23.5% (6.9–45.8%) in HLA-B mismatch and 15.1% (7.3–25.5%) in HLA-DRB1 mismatch (Figure 2b). In a multivariate analysis, the presence of an HLA-A mismatch was associated with an increased risk of NRM (HLA-A: HR 2.73, 95% CI 1.34–5.54, P=0.0056; HLA-B: HR 2.08, 95% CI 0.74–5.88, P=0.17; HLA-DRB1: HR 1.31, 95% CI 0.69–2.50, P=0.41).

Figure 2
figure2

NRM, relapse and OS. Cumulative incidence of NRM grouped according to (a) allele mismatch and (b) locus of allele mismatch. Cumulative incidence of relapse grouped according to (c) allele mismatch and (d) locus of allele mismatch. The probability of OS grouped according to (e) allele mismatch and (f) locus of allele mismatches.

The cumulative incidences of relapse at 2 years were 32.7% (95% CI 30.7–34.7%) in the Match group and 30.1% (95% CI 22.3–38.3%) in the Mismatch group (P=0.54, Figure 2c). The presence of allele mismatch did not affect the incidence of relapse. The cumulative incidences of relapse at 2 years were 22.9% (95% CI 9.7–39.3%) in HLA-A mismatch, 24.2% (6.9–47.0%) in HLA-B mismatch and 35.4% (23.6–47.4%) in HLA-DRB1 mismatch (Figure 2d). There was no statistically significant difference among the four groups.

OS

The probabilities of OS at 2 years after allogeneic HSCT were 61.7% in the Match group and 54.0% in the Mismatch group (P=0.0090, Figure 2e). A multivariate analysis showed that the presence of at least one allele mismatch was associated with an inferior OS (HR 1.43, 95% CI 1.11–1.85, P=0.0058, Table 2). Other factors that were associated with an increased risk of overall mortality were age (50 years), poor performance status (2–4), use of PBSC and high disease risk. Compared with an allele match, the presence of a one allele mismatch was associated with an inferior OS in a multivariate analysis (one allele mismatch: HR 1.46, 95% CI 1.11–1.90, P=0.0059; multiple allele mismatch: HR 1.25, 95% CI 0.59–2.66, P=0.56). We also assessed the impact of each locus excluding patients with multiple allele mismatches. The probabilities of 2-year OS were 57.6% (95% CI 38.0–72.9%) in HLA-A mismatch, 55.0% (29.8–74.5%) in HLA-B mismatch and 51.0% (37.7–62.9%) in HLA-DRB1 mismatch (Figure 2f). In a multivariate analysis, patients with an HLA-A or HLA-DRB1 mismatch tended to have a worse OS (HLA-A: HR 1.51, 95% CI 0.93–2.45, P=0.094; HLA-B: HR 1.49, 95% CI 0.77–2.87, P=0.24; HLA-DRB1: HR 1.43, 95% CI 1.00–2.03, P=0.050).

Discussion

In this study, we have demonstrated for the first time that HLA allele mismatch in the GVH direction in related HSCT was associated with increased risks of acute GVHD and NRM, which led to a poor OS. No previous study has assessed the impact of HLA allele mismatch in the related HSCT setting, as it is generally believed that HLA is completely matched in serologically HLA-matched related HSCT, especially in sibling donors if the parental HLA types are missing. Our result demonstrated that there is a possibility of HLA allele mismatch even in serologically matched related HSCT (5.6% in an HLA serologically matched donor/recipient combination). Our current result in related HSCT was consistent with the findings in unrelated HSCT, which suggests that serological HLA typing is insufficient to assess HLA compatibility.1,2 Therefore, HLA typing at high resolution (allele-level typing) should be done in all patients, including matched related transplants. The presence of HLA allele mismatch in the GVH direction should be taken into consideration when selecting a stem cell donor and determining the intensity of GVHD prophylaxis. In this study, the presence of HLA-B allele mismatch was associated with a significantly increased risk of severe acute GVHD. The significant impact of HLA-B antigen mismatch seemed to be similar to that in a previous report from Japan that assessed the impact of HLA-one antigen mismatch in related HSCT.10 An important limitation here is the lack of HLA-C information in our current database. The frequency of an HLA-C mismatch in an HLA-B-mismatched group was shown to be substantially higher than those in the HLA-A and -DR antigen-mismatched groups.15,16 In our database, information about the HLA-C allele was available in only 1152 cases (48.5%). Therefore, the impact of HLA-B and -C allele mismatch in related HSCT should be clarified in analyses using larger cohorts with complete HLA-C allele information.

One important issue in this study was the result that the use of PBSC was significantly associated with an increased risk of grade III–IV acute GVHD (HR 1.85, 95% CI 1.41–2.44, P<0.0001, Table 2), which led to an increased risk of NRM and overall mortality. Therefore optimization of GVHD prophylaxis is particularly important in patients who receive PBSC to improve the clinical outcome.

A major limitation of this study is the small sample size in the Mismatch group, which is largely due to the fact that we included patients for whom data on the HLA allele were available. Because of the limited number of cases with HLA allele mismatch, it was difficult to assess the effect of the type of GVHD prophylaxis, such as the use of T-cell depletion, on the incidence of acute GVHD. Although the use of T-cell depletion seems to reduce the risk of GVHD, this association was not statistically significant (data not shown). This may have been due to the limited number of cases with T-cell depletion in this cohort.

In conclusion, our findings suggest that the presence of an HLA allele mismatch in serologically matched related HSCT was associated with increased risks of acute GVHD and NRM, which led to a poor OS. Therefore, HLA typing at high resolution (allele-level typing) should be done in all patients, including matched related transplants. The optimal GVHD prophylaxis in patients who receive stem cells from an HLA allele-mismatched related donor should be explored prospectively.

References

  1. 1

    Park M, Seo JJ . Role of HLA in hematopoietic stem cell transplantation. Bone Marrow Res 2012; 2012: 680841.

  2. 2

    Petersdorf EW . Optimal HLA matching in hematopoietic cell transplantation. Curr Opin Immunol 2008; 20: 588–593.

  3. 3

    Morishima Y, Sasazuki T, Inoko H, Juji T, Akaza T, Yamamoto K et al. The clinical significance of human leukocyte antigen (HLA) allele compatibility in patients receiving a marrow transplant from serologically HLA-A, HLA-B, and HLA-DR matched unrelated donors. Blood 2002; 99: 4200–4206.

  4. 4

    Woolfrey A, Klein JP, Haagenson M, Spellman S, Petersdorf E, Oudshoorn M et al. HLA-C antigen mismatch is associated with worse outcome in unrelated donor peripheral blood stem cell transplantation. Biol Blood Marrow Transplant 2011; 17: 885–892.

  5. 5

    Kanda Y, Kanda J, Atsuta Y, Maeda Y, Ichinohe T, Ohashi K et al. Impact of a single human leucocyte antigen (HLA) allele mismatch on the outcome of unrelated bone marrow transplantation over two time periods. A retrospective analysis of 3003 patients from the HLA Working Group of the Japan Society for Blood and Marrow Transplantation. Br J Haematol 2013; 161: 566–577.

  6. 6

    Lee SJ, Klein J, Haagenson M, Baxter-Lowe LA, Confer DL, Eapen M et al. High-resolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood 2007; 110: 4576–4583.

  7. 7

    Flomenberg N, Baxter-Lowe LA, Confer D, Fernandez-Vina M, Filipovich A, Horowitz M et al. Impact of HLA class I and class II high-resolution matching on outcomes of unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood 2004; 104: 1923–1930.

  8. 8

    Ciurea SO, Saliba RM, Rondon G, Patah PA, Aung F, Cano P et al. Outcomes of patients with myeloid malignancies treated with allogeneic hematopoietic stem cell transplantation from matched unrelated donors compared with one human leukocyte antigen mismatched related donors using HLA typing at 10 loci. Biol Blood Marrow Transplant 2011; 17: 923–929.

  9. 9

    Valcarcel D, Sierra J, Wang T, Kan F, Gupta V, Hale GA et al. One-antigen mismatched related versus HLA-matched unrelated donor hematopoietic stem cell transplantation in adults with acute leukemia: Center for International Blood and Marrow Transplant Research results in the era of molecular HLA typing. Biol Blood Marrow Transplant 2011; 17: 640–648.

  10. 10

    Kanda J, Saji H, Fukuda T, Kobayashi T, Miyamura K, Eto T et al. Related transplantation with HLA-1 Ag mismatch in the GVH direction and HLA-8/8 allele-matched unrelated transplantation: a nationwide retrospective study. Blood 2012; 119: 2409–2416.

  11. 11

    Atsuta Y, Suzuki R, Yoshimi A, Gondo H, Tanaka J, Hiraoka A et al. Unification of hematopoietic stem cell transplantation registries in Japan and establishment of the TRUMP system. Int J Hematol 2007; 86: 269–274.

  12. 12

    Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant 1995; 15: 825–828.

  13. 13

    Fine JP, Gray RJ . A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999; 94: 496–509.

  14. 14

    Kanda Y . Investigation of the freely available easy-to-use software 'EZR' for medical statistics. Bone Marrow Transplant 2013; 48: 452–458.

  15. 15

    Prasad VK, Heller G, Kernan NA, O'Reilly RJ, Yang SY . The probability of HLA-C matching between patient and unrelated donor at the molecular level: estimations based on the linkage disequilibrium between DNA typed HLA-B and HLA-C alleles. Transplantation 1999; 68: 1044–1050.

  16. 16

    Petersdorf EW . The major histocompatibility complex: a model for understanding graft-versus-host disease. Blood 2013; 122: 1863–1872.

Download references

Acknowledgements

This work was supported in part by a Grant-in-Aid from the Ministry of Health, Labor and Welfare of Japan. We thank all the physicians and data managers at the centers who contributed valuable data on transplantation to the JSHCT. We also thank all the members of the data management committees of the JSHCT for their contributions.

Author information

Correspondence to Y Kanda.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Further reading