The impact of HLA-matching on reduced intensity conditioning regimen unrelated donor allogeneic stem cell transplantation for acute myeloid leukemia in patients above 50 years—a report from the EBMT acute leukemia working party

Background Data comparing fully matched and mismatched-unrelated-donor (M- and mM-URD) allogeneic hematopoietic stem cell transplant (allo-SCT) following reduced intensity conditioning regimens for acute myeloid leukemia are limited. Methods We retrospectively compared the outcome of 3398 patients above the age of 50 years who underwent 10/10 M-URD (n = 2567), 9/10 (n = 723), or 8/10 (n = 108) mM-URD allo-SCT for acute myeloid leukemia after reduced intensity conditioning regimen between 2000 and 2013. The Kaplan-Meier estimator, the cumulative incidence function, and Cox proportional hazards regression models were used where appropriate. Results HLA matching had no impact on engraftment (p = 0.31). In univariate analysis, in comparison to 10/10 M-URD, mM-URD was associated with higher incidence of grade II–IV acute graft-versus-host disease (GVHD) (p = 0.0002), similar rates of chronic GVHD (p = 0.138) but increased incidence of its extensive form (p = 0.047). Compared to 10/10 M-URD, patients transplanted in the first complete remission (CR1) with a 9 or an 8/10 mM-URD had decreased 2-year leukemia free (LFS) (p = 0.005) and overall survivals (OS) (56.7, 46.1, and 50.2 %, respectively, p = 0.005), while outcomes were comparable between all groups for patients transplanted beyond CR1. In multivariate analysis, 9/10 versus 10/10 URD was associated with higher non-relapse mortality (HR 1.34, p = 0.001), similar risk of relapse and chronic GVHD and inferior LFS (HR 1.25, p = 0.0001), and OS (HR 1.27, p = 0.0001). There was no difference in adjusted transplant outcomes between 9/10 and 8/10 mM-URD. Conclusions Reduced intensity conditioned allo-SCT with a 10/10 M-URD remains the preferable option for AML patients above the age of 50 years. The use of a 9/10 or an 8/10 mM-URD in patients not having a fully matched donor represents an alternative therapeutic option that should be compared to other alternative donor transplant strategies. Electronic supplementary material The online version of this article (doi:10.1186/s13045-016-0295-9) contains supplementary material, which is available to authorized users.


(Continued from previous page)
Conclusions: Reduced intensity conditioned allo-SCT with a 10/10 M-URD remains the preferable option for AML patients above the age of 50 years. The use of a 9/10 or an 8/10 mM-URD in patients not having a fully matched donor represents an alternative therapeutic option that should be compared to other alternative donor transplant strategies.
As only about one third of the patients that are in need of allo-SCT have a matched-related donor and donor registries have increased the probability to find a suitable unrelated donor (URD), increasing numbers of patients are receiving matched (M-URD) and mismatched unrelated donor (mM-URD) allo-SCT [14,15]. The outcomes of M-URD and mM-URD transplant have significantly improved in the last decade [14,16,17]. Although the MRC AML15 Trial has recently reported improved overall survival with matched related donors (MRD) compared to M-URD due to reduced NRM [11], several studies have shown similar outcomes of MRD in comparison to M-URD or mM-URD after RIC allo-SCT for AML [6,[18][19][20]. Among M-URD, comparative data between M-URD and mM-URD are limited, in particular in the context of RIC allo-SCT for older adults with AML.
Acute leukemias in the elderly population are characterized as more aggressive diseases due to selection of more primitive clones with higher prevalence of complex karyotype and presence of multidrug resistance [21,22]. Disease control after RIC allo-SCT mainly relies on the anti-leukemic effect of allogeneic NK and T cells [23,24]. Since HLA mismatching might improve graft-versus-leukemia effect in high-risk acute leukemias [25], one might hypothesize that HLA-mismatched RIC-allo SCT for AML in patients above the age of 50 years could offset the higher risk of NRM by reducing the need of myeloablative doses of chemotherapy aiming in reducing the risk of relapse. In the present study, we analyzed the outcome of 3398 adult patients above the age of 50 years who underwent matched or mismatched URD allo-SCT with RIC regimen for AML.

Study design and data collection
This is a retrospective multicenter analysis using the data set of the acute leukemia working party (ALWP) of the European Society of Blood and Marrow Transplantation (EBMT) group registry. The EBMT is a voluntary working group of more than 500 transplant centers that are required to report all consecutive stem cell transplantations and follow-ups once a year. Audits are routinely performed to determine the accuracy of the data. The study was planned and approved by the ALWP of the EBMT. In addition, the study protocol was approved by the institutional review board at each site and complied with country-specific regulatory requirements. The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Since 1990, patients provide informed consent authorizing the use of their personal information for research purposes. Eligibility criteria for this analysis included patients with AML above 50 years old who underwent a first allo-SCT from an HLA-matched (10/10) or mismatched (9/10 or 8/10) unrelated donor (mM-URD) following RIC regimen between 2000 and 2013. All donors were HLA matched (10/10) or mismatched at one or two loci (9/10 or 8/10) (−A, −B, −C, DRB1, −DQB1). HLA typing was determined at all loci by high-resolution techniques. Exclusion criteria were previous allogeneic or cord blood transplantation, ex vivo T cell-depleted stem cell graft. Regimens were classified as RIC based on EBMT criteria [26]. Variables collected included recipient and donor characteristics (age, gender, CMV serostatus, recipient's Karnofsky status at transplant), disease characteristics and status at transplant, year of transplantation and interval from diagnosis to transplantation, transplantrelated factors including conditioning regimen, pretransplant in vivo T cell depletion, stem cell source (bone marrow (BM) or peripheral blood (PB)), posttransplant graft-versus-host disease (GVHD) prophylaxis, and outcome variables (acute and chronic GVHD, relapse, NRM, LFS, OS, and causes of death). Grading of acute GVHD was performed using established criteria [27]. Chronic GVHD was classified as limited or extensive according to published criteria [28]. For the purpose of this study, all necessary data were collected according to the EBMT guidelines, using the EBMT Minimum Essential Data forms. The list of institutions reporting data included in this study is provided in Additional file 1: Table S1.

Statistical analysis
Primary endpoints of the study were incidences and severity of acute and chronic GVHD, NRM, and disease relapse incidence (RI). Secondary endpoints included engraftment, OS, and LFS. Start time was the date of transplant for all endpoints. LFS was defined as survival without relapse or progression and NRM as death without relapse/progression. Cumulative incidence functions (CIFs) were used to estimate RI and NRM in a competing risk setting, because death and relapse compete with each other. For estimating the cumulative incidence of chronic GVHD, we considered relapse and death to be competing events. The three groups according to level of HLA matching were compared by the chi-square method for qualitative variables, whereas the Mann-Whitney test was applied for continuous parameters. Univariate comparisons were done using the log-rank test for OS, LFS, and the Gray's test for RI, NRM, and GVHD cumulative incidences. Multivariate analyses were performed using logistic regression for grade II-IV acute GVHD rate and Cox proportional hazards model for all other endpoints. Factors differing in terms of distribution between the three groups and all factors known as potentially risk factors were included in the final model. Factors included in the Cox models included HLA matching, patient age (analyzed by 10-year scale), and Karnofsky status (≥80 % versus below 80 %), year of transplantation, time from diagnosis to transplantation (per 6 months), disease status at transplantation, secondary AML versus de novo AML, low-dose TBI-based versus chemotherapy-based RIC regimens, use of in vivo T cell depletion, female donor to male recipient versus other gender combinations, and CMV risk (high-risk seropositive recipient with seronegative donor versus others combinations). All tests were two sided. The type I error rate was fixed at 0.05 for determination of factors associated with time to event outcomes. Statistical analyses were performed with SPSS 22.0 (IBM Corp., Armonk, NY, USA) and R 3.1.1 software packages (R Development Core Team, Vienna, Austria).

Patient, disease, and transplant characteristics
Details of patients, disease, and transplant characteristics are summarized in Table 1. Three thousand three hundred ninety-eight patients with AML were included in the study. Two thousand five hundred sixty-seven patients (75.5 %) received a HLA 10/10 fully matched, while 723 (21.3 %) received a 9/10 and 108 (3.2 %) received an 8/10 mismatched unrelated donor (mM-URD) allo-SCT between 2000 and 2013. All patients were older than 50 years and median age was comparable between the three groups (60 to 61 years with ranges between 50 and 78 years) ( Table 1). Patients receiving an HLA 8/10 mM-URD had been transplanted more lately than the two other groups (median year of transplantation 2009 versus 2011, p = 0.001) and had a longer follow up of surviving patients (median 34 versus 24 months, p = 0.042). Interval from diagnosis to transplantation was shorter in the HLA 10/10 in comparison to the 9/10 and 8/10 mM-URD groups (212 versus 250 and 295 days, respectively, p = 0.0001). Patients in the mM-URD groups had been more frequently transplanted with a female donor (p < 10 −4 ) and more transplants were performed from female donor to male recipient (p = 0.01). Secondary AML was more frequent in the 8/10 mM-URD group (44 versus <30 %, p = 0.001). The proportions of poor cytogenetics were equally distributed between the three groups although cytogenetic analysis was missing in 52 to 55 %, of the patients. Significantly higher numbers of patients were transplanted in CR1 in the 10/10 URD in comparison to the 9/10 and 8/10 mM-URD groups (55 versus 46.1 and 44.4 %, respectively, p = 0.0002). Peripheral blood represented the major source of stem cells in all groups. TBI-based RIC was less frequently used in the 8/10 HLA mM-URD group (p = 0.02). Patients in the mM-URD groups had received more frequently in vivo T cell depletion in comparison to 10/10 matched URD SCT (86 % in 8/10, 83.1 % in 9/10 versus 73.5 % in 10/10 groups, p < 10 −4 ). There was higher proportions of patients with high CMV reactivation risk (negative donor with positive recipient CMV serologies) in the 9/10 group in comparison to the others (p = 0.029). The choice of conditioning and GVHD prophylaxis was dependent on centers' protocols and strategies of transplantation.

Discussion
Unrelated donors represent the most frequent stem cell source for allo-SCT in Europe and are used in more than 50 % of RIC allo-SCT for AML patients [15]. In the absence of HLA-matched URD, a significant proportion of patients is transplanted with a 9/10 and to a less extends, with an 8/10 mM-URD. Although outcomes of RIC allo-SCT with HLA 10/10 or 8/8 MUD have been reported as comparable to transplants performed with a matched related donor [6,[18][19][20], the outcomes of M-URD versus mM-URD for older adults with AML have been poorly explored. This large, multicenter, registry study showed superior outcome using HLA-matched (10/10) donor compared to mM-URD (9/10 or 8/10) allo-SCT for AML in patients above the age of 50 years. Patients receiving mM-URD had significantly higher incidence of acute GVHD (both grades II-IV and II-IV) and NRM. A larger series of patients transplanted with RIC or MAC regimens for AML reported by the CIBMTR similarly showed increased risk of NRM with 7/8 mM-URD (n = 406) compared to 8/8 M-URD (n = 1193) or MRD (n = 624) due to increased incidence of acute GVHD in M-URD versus MRD [29]. We did not observe any impact of HLA matching on the overall incidence of chronic GVHD. However, in univariate analysis, the use of an 8/10 mM-URD was associated with increased risk of overall and extensive chronic GVHD particularly for patients transplanted in advanced phase disease. In our study, the other factors associated with a higher risk of both acute and chronic GVHD were active disease at transplantation and the absence of in vivo T cell depletion. In addition, patient age above 60 years and Karnofsky performance status below 80 % had a negative impact on NRM. Thus, the choice of a mM-URD for elderly patients transplanted with RIC-allo SCT requires the use of ATG and should take into consideration the higher risk of GVHD and NRM, in particular for patients above 60 years, not in CR and/ or with an decreased performance status at allo-SCT.
In contrast to the reduced risk of relapse observed in the CIBMTR study with the use of 7/8 mM-URD in comparison to 8/8 MRD and M-URD [29], the use of a 9/10 mM-URD in our study was associated with a higher risk of relapse in comparison to 10/10 M-URD in multivariate analysis. There was, however, a trend for reduced relapse incidence with 8/10 in comparison to 9/10 mM-URD (HR 0.67, 95 % CI 0.43-1.02; p = 0.064). These differences might be explained by higher proportions of patients transplanted with active disease in the mM-MUD groups and by more intensive and prolonged immunosuppression administered to patients receiving mM-URD allo-SCT in order to control acute GVHD. This might have affected the development of the graft-versus-leukemia (GVL) effect, otherwise expected at higher level with mismatched donor T cells. Consistent with this hypothesis is the absence of increased chronic GVHD in the 9/10 mM-URD versus the 10/10 M-URD groups, while higher incidence of extensive chronic GVHD in the 8/10 mM-URD was associated with a trend towards reduced relapse incidence, in particular observed in patients transplanted in CR1. Another factor associated with increased risk of relapse, leading to reduced LFS and OS, was the use of a lowdose TBI-based RIC (truly non-ablative) regimen, suggesting that increasing the intensity of the conditioning might help for the control of the disease until the GVL response takes place [30].
As expected, disease status at SCT was another important factor impacting RI, LFS, and OS. In our study in older adults with AML transplanted in CR1, the 2year OS was comparable to prior reports using RIC and MRD [11] and to younger AML patients transplanted with MAC regimens and MRD or M-URD allo-SCT [31]. Although LFS and OS were reduced with mM-URD in patients transplanted in CR, 2-year OS for patients transplanted in CR1 or in >CR1 were between 46 and 50 % with a 9/10 or 8/10 mM-URD transplantation, comparable to results of RIC for AML with MRD and M-URD [3,4,6], suggesting that mM-URD remains a valid option for AML above 50 years old, in particular for those transplanted in >CR1.
The feasibility of haplo-identical SCT performed with T replete stem cell grafts and in vivo T cell depletion based either on ATG [32] or post-transplant cyclophosphamide [33] has been demonstrated in the past 10 years. Comparisons of haplo-identical SCT using the Chinese approach combining T replete G-CSF mobilized bone marrow stem cell graft, a myeloablative conditioning regimen with ATG, to matched related and unrelated allo-SCT for acute leukemias have shown equivalent OS with reduced risk of relapse in high-risk leukemias in the haplo-identical groups [34,35]. Using this approach, the Beijing's group recently reported similar outcomes in fit patients transplanted above 50 years of age in comparison to younger patients [36]. In the setting of posttransplant cyclophosphamide, Blaise et al. reported in patients older than 60 years inferior outcomes of RIC allo-SCT performed with HLA 10/10 and 9/10 URD compared to those transplanted with MRD or haploidentical donors because of higher NRM related to higher incidence of acute and chronic GVHD [37]. Comparisons of haplo-identical SCT at the era of posttransplant cyclophosphamide to URD have shown similar LFS and OS but reduced NRM and chronic GVHD with T replete haplo-SCT [38][39][40][41]. From 2012 onward, there has been increasing numbers of transplants performed from related haplo-identical donor, which is likely mainly due to increased use of haplo-identical donors with the post-transplant cyclophosphamide strategy. In AML patients undergoing allo-SCT without an HLA-matched (related or unrelated) donor, the decision to use one alternative graft source over another is complex. Published data support any one of the three alternative donor allo-SCT options (i.e., mM-URD, CBT, related haplo-identical) currently available for patients without a matched donor. Our current study support this notion as results of mM-URD in AML patients with age above 50 years transplanted in CR1 were inferior to matched URD and therefore other alternative like Haplo-SCT and CBT may be considered.