INTRODUCTION — The selection of a donor is a critical element contributing to the success of hematopoietic cell transplantation (HCT). There are several possible donor options:
●An identical twin (syngeneic, HLA identical)
●A sibling, relative, or unrelated donor (allogeneic, which can be HLA identical, haploidentical, or mismatched)
●Umbilical cord blood (allogeneic, which can be HLA identical, haploidentical, or mismatched)
●The patient (autologous, HLA identical)
This topic will provide an overview of the general issues involved in donor selection for allogeneic HCT. The decision of which donor source to utilize depends, to a large degree, on the clinical situation and the approaches employed at the individual transplant center. Most patients can now find a suitable donor, given the availability of multiple donor sources.
The term "hematopoietic cell transplantation" (HCT) will be used throughout this review as a general term to cover transplantation of progenitor cells from any source (eg, bone marrow, peripheral blood, umbilical cord blood). Otherwise, the source of such cells will be specified (eg, allogeneic peripheral blood progenitor cell transplantation). (See "Sources of hematopoietic stem cells".)
DEFINITIONS — Matching donor and recipient for human leukocyte antigen (HLA) class I (-A, -B, and -C) and class II (-DRB1 and -DQB1) haplotypes is a key part of successful allogeneic HCT. Techniques for HLA typing and the HLA nomenclature have evolved with the increasing use of unrelated donors, the discovery of additional HLA alleles, and improved methods of HLA typing [1].
●Serologic typing – Serologic typing, used for antigen matching, detects HLA proteins using serologic antibody-based assays. Serologic typing may be used for typing within families and depends on the identification of parental HLA assignments such that all four HLA types are unequivocally defined in the family and distinguishable from one another [2]. When the donor and recipient are not related, serologic typing alone does not ensure that those individuals share the same HLA genes. Serological typing has largely been replaced by the more specific molecular typing.
●Molecular typing – Molecular typing, used for allele matching, defines HLA genes by their DNA sequences. Molecular typing is necessary for HLA matching in unrelated donor transplants and is preferred by most centers for HLA matching related donor transplants. Molecular typing can differ by resolution [2]. High throughput DNA sequencing provides complete typing information and is widely used for donor typing. High resolution typing defines sets of alleles that encode the same protein sequence for the HLA molecule's antigen binding site (eg, HLA-A*02:101). In contrast, low resolution typing provides information regarding the allele group, but not the specific HLA protein (eg, HLA-A*02).
The following definitions also apply:
●A 12 of 12 HLA match refers to donor-recipient pairs matched for HLA-A, HLA-B, HLA-C, HLA-DRB1, HLA-DQB1, and HLA-DP1 at the allele level.
●A 10 of 10 HLA match refers to donor-recipient pairs matched for HLA-A, HLA-B, HLA-C, HLA-DRB1, and HLA-DQB1 at the allele level.
●A 9 of 10 HLA match refers to pairs with a single allele or antigen mismatch at either HLA-A, HLA-B, HLA-C, HLA-DRB1, or HLA-DQB1.
●An 8 of 8 HLA match refers to donor-recipient pairs matched for HLA-A, HLA-B, HLA-C, and HLA-DRB1 at the allele level.
●A 7 of 8 HLA match refers to pairs with a single allele or antigen mismatch at either HLA-A, HLA-B, HLA-C, or HLA-DRB1.
●A 6 of 6 HLA match refers to donor-recipient pairs matched for HLA-A, HLA-B, and HLA-DRB1 at the allele level.
Antigen mismatches can be further characterized as being in the "graft-versus-host" direction or the "host-versus-graft" direction:
●A graft-versus-host direction mismatch means that the patient possesses one or more alleles not present in the donor. In such a mismatch, the donor T cells from the graft would be expected to mount an alloreactive response toward the patient (host) tissue.
●A host-versus-graft direction mismatch means that the donor possesses one or more alleles not present in the patient. In such a mismatch, the patient's T cells (host T cells) would be expected to mount an alloreactive response toward the donor (graft) cells.
An HLA-DP mismatch refers to either a single or a double HLA-DPB1 allele mismatch in the graft-versus-host or host-versus graft direction, or both. HLA-DP mismatches can be further divided into permissive or non-permissive mismatches depending on the haplotypes involved and the direction of the mismatch [3]. (See 'HLA-DP haplotype' below.)
GENERAL APPROACH — Outcomes following allogeneic HCT depend on the underlying disease (reason for transplant), the timing of transplant (early versus late), patient comorbidities (including cytomegalovirus [CMV] seropositivity), and the choice of donor. The search for an appropriate donor must consider the urgency of the procedure and potential risks of postponing transplant. All patients likely to require an allogeneic HCT should undergo high resolution human leukocyte antigen (HLA) typing for HLA-A, -B, -C, and -DR soon after diagnosis to allow for the timely identification of an appropriate donor. Performing these tests early in the course of treatment is often critical so that the donor search can be initiated and time is not lost later.
The key factors in predicting survival after transplant are the stage of disease and the degree of HLA matching. There is a progressive decrease in post-HCT survival with each HLA allele mismatch reducing the probability of overall survival (OS) at five years by approximately 10 percent (figure 1) [4-7]. However, the absolute decrease in survival with each HLA allele mismatch lessens as the urgency of transplant increases [4]. As such, a decision regarding the best search plan must take into account the risk of relapse while performing a more extensive search. If the risk of relapse is high and the chance of finding a perfect match is low, then proceeding with a known mismatch or an alternative donor (haploidentical or umbilical cord blood transplant) may be preferable to a lengthy search for a full match.
When available, a matched sibling donor is preferred over other donor sources due to improved clinical outcomes following transplant (eg, less graft-versus-host disease) and the speed and cost-effectiveness of the search. As such, the initial search should center on molecular HLA typing of the patient's full biological siblings (algorithm 1). If a 6 of 6 matched sibling donor is identified and eligible for donation, transplantation should proceed. A choice among several equally matched sibling donors can then take into account other donor characteristics such as age, CMV status (matching for CMV status preferred), gender (males and nulliparous females preferred), and blood type [8]. (See 'Matched sibling donors' below.)
When a 6 of 6 matched sibling donor is not available, the transplant center usually proceeds with an unrelated donor search. If an urgent transplant is required, some clinicians will choose to search for an umbilical cord blood (UCB) graft or a haploidentical donor since these grafts are usually quicker to acquire. An 8 of 8 matched unrelated donor results in similar survival rates to those seen with a matched sibling donor. Identification of an unrelated adult donor and graft procurement may require two or three months, but in some circumstances the search can be expedited for patients with more urgent needs. Delays in time to transplant are associated with worse patient outcome [9]. Alternative donor sources (UCB, haploidentical donors) may permit a shorter time to transplant (eg, <1 month), but with a potentially increased risk of transplant-related complications. Therefore, it is important to consider patient and donor typing as early as possible in the patient's treatment course. (See 'Unrelated donors' below.)
In general, searches for adult donors take into account HLA-A, -B, -C, and -DRB1 at high resolution. When an 8 of 8 matched unrelated adult donor is not available, options include a 7 of 8 matched unrelated adult donor, a 4 of 6 matched UCB unit with adequate cell dose, and a haploidentical donor (eg, parent, child, sibling). A choice among these must take into consideration the likelihood of confirming the match; obtaining the specimen; time to transplant; and individual center preferences, experience, and outcomes. The latter two may be faster, but also may be associated with increased complications. Increasing HLA disparity is associated with worse outcomes post HCT; however, recipients of UCB transplants are able to tolerate greater HLA disparity than those of other graft sources. As such, minimum matching criteria for unrelated adult donor transplant require an at least 7 of 8 HLA match, while UCB transplant allows for a 4 of 6 matched UCB unit. Haploidentical donors are matched at 3 of 6 loci. UCB has a higher risk of graft failure and does not allow for additional grafts from the same donor, if needed. Haploidentical grafts allow for additional grafts if needed, but have a high rate of graft-versus-host disease or relapse depending on the approaches utilized. These two graft sources are being compared in randomized trials performed by The Blood and Marrow Transplant Clinical Trials Network (BMT CTN). These and other advantages and disadvantages of UCB in relation to an unrelated adult donor are presented separately. (See "HLA-haploidentical hematopoietic cell transplantation" and "Selection of an umbilical cord blood graft for hematopoietic cell transplantation", section on 'Features of UCB grafts'.)
A choice among several equally matched donors is further narrowed based on the type of graft planned (bone marrow versus peripheral blood), the mismatched HLA gene, further typing of other HLA genes, and an evaluation of the recipient for donor-specific HLA antibodies. As examples:
●Type of graft planned and mismatched HLA gene – For patients undergoing bone marrow transplant, allele or antigen mismatch at HLA-B or HLA-C may be better tolerated than HLA-A or HLA-DR [4]. In contrast, for patients undergoing peripheral blood progenitor cell transplant, HLA-C antigen mismatch appears to convey a worse prognosis [10].
●Other HLA genes – Typing for HLA-DP, -DQB1, and -DRB 3/4/5 may improve outcomes. If the grafts are otherwise considered equal, donors with the greatest number of matches for these genes are preferred.
●Donor-specific HLA antibodies – The presence of donor-specific HLA antibodies is associated with an increased risk of graft failure and should be avoided [11].
Our approach is generally consistent with guidelines published by the National Marrow Donor Program (NMDP, "Be The Match") for adult donor and cord blood matching [12]. Despite the availability of guidelines, donor selection is a complicated process that requires specific expertise. Help search strategy teams are available through the NMDP to provide commentary and donor recommendations for NMDP searches.
MATCHED RELATED DONORS — When available, a matched related donor is generally the preferred source for allogeneic HCT (algorithm 1) due to a quicker and less expensive evaluation, and improved clinical outcomes. However, a related donor may be less desirable in certain circumstances; as examples, a related donor may be older than other potential graft sources, or may have the same familial condition that predisposed the patient to the disease being treated by transplantation.
Matched sibling donors — While survival rates following fully matched unrelated donor transplantations are similar to those of matched sibling donor transplantations, sibling donor transplants are associated with less morbidity, including a lower rate of acute and chronic graft-versus-host disease (GVHD) [13]. As such, the patient's full biological siblings should undergo typing of human leukocyte antigen (HLA)-A, -B, and -DR to identify potential sibling donors; specific typing for HLA-C is often not necessary because HLA-C and HLA-B are usually inherited together due to their tight linkage on chromosome 6. In some settings, a related donor may be less desirable than other potential graft sources because of the age or sex of the potential donor.
Related donors are generally siblings, because they have the opportunity to inherit the same HLA genes (due to their proximity on chromosome 6). There is a 25 percent chance that a given sibling will be HLA matched at the A, B, and DR loci, which are the principal loci evaluated in this setting (figure 2). Unless the parents happen to have very common haplotypes or are consanguineous, it is unlikely that matched related donors will be anything other than siblings (eg, fully HLA-matched first cousins). In matched siblings, serologic typing is adequate, since a match at these loci makes it highly likely that the same genes were inherited from the parents. However, many centers perform low resolution molecular typing to match siblings. (See "Human leukocyte antigens (HLA): A roadmap".)
In some circumstances, a related donor may be less desirable than other potential graft sources, based on age or sex. Related sibling donor/recipient pairs are usually of similar age and the incidence of clonal hematopoiesis of indeterminate potential (CHIP) increases with age. Some institutions screen older potential donors for CHIP. Increased donor age has been associated with decreased overall and disease-free survival, it is not known whether this is due to the increased incidence of CHIP, nor is it clear if it is preferable to choose an older matched sibling donor versus a younger matched unrelated donor [14]. For sex-mismatched transplantation, especially for female donors to male recipients (FtoM), the choice of bone marrow versus peripheral blood grafts may affect outcomes. (See 'Age, sex, and parity' below.)
For some diseases, donors should be screened for a relevant familial condition. As examples, donors for patients with chronic lymphocytic leukemia (CLL) should be screened for monoclonal B cell lymphocytosis, and donors for patients with a familial myeloid malignancy should be evaluated for the underlying mutation. (See "Monoclonal B cell lymphocytosis", section on 'Stem cell/blood donation' and "Familial disorders of acute leukemia and myelodysplastic syndromes", section on 'Management'.)
Identical twin donors — Rarely, patients who are candidates for HCT have an identical twin as a potential donor. These patients do not require post-transplant immunosuppression and do not develop GVHD. However, they are at a higher risk of relapse of any underlying malignant disease than similar patients transplanted with HLA-matched but nonidentical sibling donors [15]. This is especially true for patients transplanted for acute myeloid leukemia (AML) and chronic myeloid leukemia (CML), less so for patients with acute lymphoblastic leukemia (ALL) [16], and has not been demonstrated in patients transplanted for non-Hodgkin lymphoma [17].
The reason for this disparity is related to the ability of the donor lymphocytes to recognize the recipient tumor cells, known as a graft-versus-tumor reaction [18]. These relationships were illustrated in a review from 163 transplant centers that compared the results of 103 identical twin (syngeneic) and 1030 HLA-identical sibling (allogeneic) transplants for leukemia [16]. The three-year probability of relapse of leukemia was substantially higher in the syngeneic than allogeneic transplants in AML (52 versus 16 percent) and CML (40 versus 7 percent); the graft-versus-tumor effect was less apparent and not statistically significant in ALL (36 versus 26 percent). Interestingly, survival was similar with the two types of transplants because the increased incidence of leukemia relapse with identical twin transplants was counterbalanced by less treatment-related mortality, in particular GVHD. (See "Biology of the graft-versus-tumor effect following hematopoietic cell transplantation".)
Molecular scrutiny of the blood of the healthy twin (and potential transplant donor) is particularly relevant if the malignant clone in the unhealthy twin may have been present in utero, as has been shown for acute leukemia [19,20]. Under such circumstances, identical twins with a shared placental circulation may be harboring the identical malignant clone, with one twin developing clinically evident leukemia in advance of the other.
UNRELATED DONORS — When a human leukocyte antigen (HLA)-matched sibling donor is not available, the transplant center usually proceeds with an unrelated donor search (algorithm 1). The adoption of high resolution, HLA allele-level typing for screening unrelated donors has been associated with improved outcomes in transplants using matched unrelated donors. Overall survival (OS), relapse-free survival (RFS), and treatment-related mortality (TRM) following transplant from carefully matched unrelated donors now approach those following matched related donors [13,14,21-24].
Overview — A higher degree of HLA matching is associated with superior survival rates. However, an unrelated donor search can take several months to identify an unrelated adult donor and procure the graft. Delays in time to transplant are associated with worse patient outcome, which can be due to progression of the underlying disease (eg, relapse, progression) or a change in comorbidities that impact transplant eligibility [9]. As such, if an urgent transplant is required, some clinicians will choose to search for an umbilical cord blood (UCB) graft or a haploidentical donor since these grafts are usually quicker to acquire.
In 1986, the National Marrow Donor Program (NMDP) was established as a repository for HLA typing information so that unrelated donors and recipients could be matched [25]. Many international registries have also formed that allow for the worldwide search for appropriate donors for those in need of a bone marrow transplant procedure [26,27]. The depth of resolution and number of alleles typed varied among and within the various registries.
At present, there are over 25 million donors registered worldwide [28]. These potential donors have undergone HLA-A and -B serologic typing. A significant percentage has also undergone molecular typing. The number of potential donors is increasing with greater public awareness of the need. Such a large number of donors is necessary because there is great diversity in the HLA system with over 5500 class I alleles (HLA-A, -B, and -C) and over 1600 class II alleles (HLA-DRB1 and -DQB1) resulting in several million potential HLA combinations [28]. Many potential matches may be subsequently mismatched when higher resolution typing is performed.
It is more difficult to find a donor for patients from certain ethnic and racial backgrounds. As an example, it is much more likely that a matched unrelated graft will be found for a White recipient than for minorities in need in the United States [29,30]. This is in part due to underrepresentation in the data bank; in addition, African Americans and Asian Americans are more polymorphic than White Americans with respect to HLA and have a relatively large number of haplotypes that are specific to their racial groups, making it more difficult to find donors at any registry size [31,32]. Donor registries are actively seeking to expand their donor pool to increase diversity. In addition, HLA-based predictive algorithms (eg, HapLogic) can be used to predict the likelihood of finding a donor for an individual, thereby aiding in the search strategy. If predictive algorithms suggest that it is highly unlikely that a match will be found, it may be better to proceed with finding a mismatched or alternative donor.
Initiating the search — To initiate a preliminary search of the National Marrow Donor Program (NMDP), contact the NMDP Office of Patient Advocacy at the following toll-free telephone number: 1-888-999-6743. Alternatively, a preliminary search request form can be submitted through the NMDP website [33,34].
Time is an issue, as it may take several months to complete a search for a matched unrelated donor, however, a preliminary search can be performed quickly and will give a rough idea of how likely it is that a donor can be found. This is largely due to the lack of complete typing data for the entire donor population, and the large expense required to perform all of the typing tests on all possible donors. In one study of 240 patients of Dutch origin, for example, an appropriate donor could be identified in approximately 70 percent of cases [35]. The search required six months. Approximately one-third of patients were transplanted, one-third had a potential donor but did not receive the transplant, and one-third did not have a suitable donor.
The long search times make this approach feasible only for those diseases with suitable clinical courses. This includes myelodysplastic syndromes, aplastic anemia, chronic myeloid leukemia, and when searches are initiated early in the course of treatment of patients with high-risk leukemias. Therefore, early discussion and initiation of an unrelated donor search and/or referral to the transplant center are essential for those patients in whom transplantation is a likely treatment option.
One aid in searching for an appropriate donor is the HapLogic program developed by the NMDP. This is a matching algorithm in which the probability of matching at HLA-A, -B, -C, -DRB1 and -DQB1 is calculated [36]. The sorting order is for HLA-A, -B, and -DRB1. These probabilities allow the physician to decide whether further testing of possible donors is warranted. If the probability of matching is high, then further testing of donors would be useful, but if the probability is zero or low, alternative donors should be pursued.
Once an appropriate donor is identified, the donor must be screened for general health and suitability. The bone marrow or mobilized peripheral blood progenitor cells are harvested locally and then transported to the transplant center. It is feasible to collect donor hematopoietic cells virtually anywhere in the world. (See 'Other considerations' below.)
HLA gene haplotypes
Initial matching — Matching donor and recipient for HLA class I (-A, -B, and -C) and class II (-DRB1 and -DQB1) haplotypes is a key part of successful allogeneic HCT. Among patients undergoing unrelated donor bone marrow or mobilized peripheral blood progenitor cell HCT, there is a progressive decrease in post-HCT survival with each HLA allele mismatch. Each HLA mismatch may reduce the probability of OS at five years by approximately 10 percent (figure 1) [4-6]. In addition, HLA mismatching is associated with higher rates of graft-versus-host disease (GVHD) in patients transplanted for malignant disorders and higher rates of graft failure in patients transplanted for nonmalignant disease [37-40].
As an example, a retrospective study from the Center for International Blood and Marrow Transplant Research (CIBMTR) evaluated the outcomes of patients with acute myeloid leukemia (AML) who underwent a matched related donor (624 patients), 8 of 8 HLA locus matched unrelated donor (1193 patients), or 7 of 8 HLA locus matched unrelated donor (406 patients) HCT between 2002 and 2006 [41]. Recipients of a matched related graft had a lower 100-day cumulative incidence of acute GVHD (33 versus 51 and 53 percent). When compared with recipients of a matched related donor graft, recipients of an 8 of 8 matched unrelated donor graft had similar survival rates at one (42 versus 55 percent) and three (37 versus 39 percent) years post-HCT. When compared with the other two groups, recipients of a 7 of 8 matched unrelated donor graft had a lower rate of OS at one year (45 percent), but similar rate of survival at three years (34 percent).
Reducing the degree of mismatch with sophisticated genotyping has, as noted above, resulted in decreased mortality following unrelated bone marrow transplants [37,38,42]. However, requiring more specific testing will undoubtedly reduce the number of donors available for an individual patient and may increase the amount of time required to locate a match, since the vast majority of patients in the registries have been typed with serologic assays.
Molecular typing of 10 HLA alleles (HLA-A, HLA-B, HLA-C, HLA-DRB1, and HLA-DQB1) has been considered the standard at most centers. However, the clinical significance of HLA-DQB1 is unclear. In contrast, there has been increased interest in the impact of the HLA-DP haplotype (a class II HLA), which appears to depend on the specific haplotypes and the direction of the mismatch. The degree of HLA disparity that is tolerable depends on many factors, including the clinical situation of the recipient [43].
A large retrospective study evaluating the impact of HLA matching on clinical outcomes in 3857 donor-recipient pairs undergoing myeloablative HCT in the United States between 1988 and 2003 [4]. The vast majority received T cell-replete bone marrow grafts and calcineurin inhibitor-based GVHD prophylaxis. At a median follow-up of six years, the following outcomes were reported:
●When compared with fully matched (10 of 10) transplants, mismatch at a single locus (9 of 10) was associated with significantly worse OS (relative risk [RR] 1.17, 95% CI 1.06-1.33), disease-free survival (DFS; RR 1.16, 95% CI 1.05-1.28), TRM (RR 1.31, 95% CI 1.16-1.47), and acute GVHD (RR 1.35, 95% CI 1.19-1.56). Single locus mismatch did not appear to impact relapse rate, engraftment, or chronic GVHD.
●Subset analysis revealed that HLA-DQ did not impact outcome as a single mismatch or when paired with another mismatch. In contrast, mismatch at HLA-A, -B, -C, and -DR was associated with worse clinical outcomes. A single mismatch detected either by low or high resolution DNA testing at one of these four loci (ie, a 7 of 8 match) was associated with higher mortality (RR 1.25, 95% CI 1.13-1.38). Single mismatches at HLA-B or HLA-C appeared to be better tolerated than mismatches at HLA-A or HLA-DRB1, while mismatches at HLA-DP or HLA-DQ loci and donor factors other than HLA type were not associated with reduced survival.
●Each additional HLA mismatch reduced OS by approximately 10 percent, being 52, 43, and 33 percent for those with 8 of 8, 7 of 8, and 6 of 8 matches, respectively (figure 1). The absolute decrease in survival was more pronounced among patients with early stage disease. In patients with high-risk disease, the increased risk due to mismatch appeared to be abrogated by the high risk of relapse from the underlying disease.
These data support the use of high resolution matching of HLA-A, -B, -C, and -DRB1 when searching for an unrelated bone marrow donor. When an 8 of 8 matched donor is not available, a haploidentical or UCB source should be considered. In choosing among mismatched related and unrelated donors, a single mismatch at HLA-B and HLA-C is preferred over a mismatch at HLA-A or -DRB1.
The impact of HLA matching appears to be similar for patients receiving an unrelated peripheral blood progenitor cell (PBPC) transplant except that HLA-C antigen mismatches are generally avoided since they appear to convey a worse prognosis. The impact of HLA matching in PBPC transplant was evaluated in a retrospective analysis of 1933 patients undergoing unrelated PBPC transplant [10]. This study had a significant number of patients undergoing reduced intensity conditioning and most transplants were performed in patients with AML. As in bone marrow transplant, HLA-DQB1 mismatch did not appear to impact outcome. Antigen mismatching at HLA-C was the only antigen that demonstrated a significant increase in the risk for mortality, DFS, TRM, and GVHD. However, the number of patients with other single mismatched alleles was low, thereby limiting the power to detect a difference.
The direction of antigen mismatch may impact patient outcome. In one study, patients with a bidirectional 7 of 8 HLA mismatch or a graft-versus-host 7 of 8 HLA mismatch had significantly increased transplant-related mortality and decreased OS and DFS when compared with patients with an 8 of 8 HLA match [44]. In contrast, patients with a host-versus-graft 7 of 8 HLA mismatch had similar outcomes to those with an 8 of 8 HLA match.
The location of antigen mismatch within an HLA allele may also impact patient outcome. In one study, the specific peptide position involved in the HLA mismatch determined whether a class I mismatch was permissive or nonpermissive [45]. Nonpermissive mismatches were associated with higher rates of GVHD and/or mortality and occurred at positions 116 or 99 of HLA-C and at position 9 of HLA-B. A 7 of 8 HLA mismatch that did not include amino acid substitutions at one of these positions had outcomes similar to those seen with 8 of 8 HLA matches. The importance of permissive and nonpermissive mismatches of the HLA-DP haplotype is described below.
Patient outcome may also be impacted by the expression level of the mismatched antigen. As an example, in a study of 1975 patients receiving an unrelated donor graft with a single mismatch at HLA-C, increased expression of the mismatched HLA-C allele was associated with increased risks of acute GVHD, nonrelapse mortality, and overall mortality [46]. Mismatch of HLA-C alleles with low expression were associated with better outcomes.
HLA-DP haplotype — Initial studies of unrelated donor HCT did not consider the class II HLA-DPB1 haplotype in the selection of donors. Retrospective analyses looking at the impact of HLA-DP haplotype on HCT outcomes suggested that mismatching at the HLA-DP loci was not associated with reduced survival, but was associated with an increased risk of acute GVHD and a reduced risk of disease relapse [4,7,47]. However, subsequent studies suggested that the impact of an HLA-DP mismatch varies depending on the specific haplotypes, the direction of the mismatch, and the degree of haplotype expression [48,49]. A choice among several donors equally matched for HLA-A, -B, -C, and -DRB1 may be further narrowed by evaluation of the HLA-DPB1 haplotype. Selection of a matched or permissive HLA-DP haplotype appears to decrease early nonrelapse mortality, but its impact long-term survival is unclear.
In an international retrospective analysis of 8539 unrelated donor and recipient pairs matched for 10 of 10 HLA alleles (64 percent) or 9 of 10 HLA alleles (36 percent), HLA-DPB1 status was identified as matched (20 percent), non-permissive mismatched (31 percent), or permissive mismatched (49 percent) [50]. The effect of HLA-DPB1 status on clinical outcomes was dependent on the degree of HLA mismatch and type of HLA-DPB1 mismatch:
●Among pairs with a 10 of 10 HLA match, when compared with permissive mismatches, non-permissive mismatches were associated with a significantly increased risk of overall mortality (hazard ratio [HR] 1.15, 95% CI 1.05-1.25), nonrelapse mortality (HR 1.28, 1.14-1.42), and severe acute GVHD (odds ratio [OR] 1.31, 95% CI 1.11-1.54), but not relapse (HR 0.89, 95% CI 0.77-1.02).
●Also, among pairs with a 10 of 10 HLA match, when compared with permissive mismatches, HLA-DPB1 matches were associated with lower rates of nonrelapse mortality (HR 0.86, 95% CI 0.75-0.98) and higher rates of relapse (HR 1.34, 95% CI 1.17-1.54), but similar overall mortality (HR 0.96, 95% CI 0.87-1.06) and acute GVHD (OR 0.84, 95% CI 0.69-1.03).
●Among pairs with a 9 of 10 HLA match, when compared with permissive matches, non-permissive mismatches increased the risk of overall mortality (HR 1.10, 95% CI 1.00-1.22), nonrelapse mortality (HR 1.19, 1.05-1.36), and severe acute GVHD (OR 1.37, 95% CI 1.13-1.66), but a similar risk of relapse (HR 0.93, 95% CI 0.78-1.11).
●The clinical outcomes among pairs with a 10 of 10 HLA match with non-permissive HLA-DPB1 mismatches was similar to those with a 9 of 10 HLA match with permissive HLA-DPB1 mismatches or HLA-DPB1 matches.
The increased GVHD and transplant-related mortality associated with a nonpermissive HLA-DPB1 allele mismatch was confirmed in another large retrospective analysis of 8003 patients undergoing myeloablative unrelated allogeneic HCT [51]. In this latter study, this increase in transplant-related mortality appeared to result in a small increase in overall mortality among those with an 8 of 8 HLA match or 10 of 10 HLA match.
The degree of HLA-DPB1 haplotype expression varies on the surface of donor and recipient cells and the HLA-DPB1 regulatory region variant rs9277534 can act as a marker of HLA-DPB1 cell-surface expression in an individual. A retrospective analysis of 1441 unrelated donor and recipient pairs matched for 10 of 10 HLA alleles with one HLA-DPB1 mismatch used rs9277534 as a marker of HLA-DPB1 cell-surface expression [52]. When compared with those with low expression (rs9277534A), recipients with high expression (rs9277534G) had a greater likelihood of severe acute GVHD. The graft-versus-host response was highest among recipients with high expression who received a graft from a low expressing donor. Such patients had a high rate of acute GVHD and an increased risk of death due to causes other than disease recurrence.
While these studies suggest that the best clinical outcomes are seen among pairs with a 12 of 12 HLA match (HLA-A, -B, -C, -DRB1, -DQB1, -DPB1), this level of match will be seen in only approximately 20 percent of patients previously identified as a 10 of 10 HLA match (HLA-A, -B, -C, -DRB1, -DQB1). Further delineation of HLA-DPB1 mismatches as permissive and non-permissive or determination of HLA-DPB1 expression allows for more informed donor selection. When all other factors are equal, selection of a 12 of 12 matched donor should be preferred.
MHC-I MICA — The major histocompatability complex (MHC)-encoded class I chain-related gene A (MICA) is a human HLA-encoded transplantation antigen that also influences the risk of GVHD and disease relapse. MICA is a nonconventional MHC class I glycoprotein molecule with at least 105 distinct alleles, which is expressed primarily on epithelial cells; MICA is tightly linked to HLA-B locus [53]. MICA binds NKG2D, an activating receptor on cytotoxic CD8+ T lymphocytes and natural killer cells.
Selection of a MICA-matched pair, when available, may be beneficial when a reduction in GVHD is paramount. A retrospective study of 922 unrelated donors with 10/10 HLA allele matches (HLA-A, B, C, DRB1, DQB1) analyzed compatibility at the MICA locus [54]. The 12 percent of pairs who were mismatched in MICA experienced increased grade III to IV acute GVHD (HR 1.83, 95% CI 1.50-2.23), chronic GVHD (HR 1.50, 95% CI 1.45-1.55), and nonrelapse mortality (HR 1.35, 95% CI 1.24-1.46). The increased risk for GVHD was mirrored by a lower risk for relapse (HR 0.50, 95% CI 0.43-0.59).
Donor-specific HLA antibodies — The presence of donor-specific HLA antibodies is associated with an increased risk of graft failure. When selecting among several otherwise equivalent donors, the recipient should undergo screening for donor-specific HLA antibodies if there is an HLA mismatch. Preference is given for donors with HLA haplotypes to which the recipient has no antibodies. In general, an alternative donor source is preferable to attempting to reduce the titer of donor-specific anti-HLA antibodies, which has had variable results.
A retrospective, case-controlled study compared serum samples from 37 patients who did not attain sustained engraftment following mismatched unrelated HCT with serum samples of 78 case-matched controls who attained sustained engraftment [11]. Controls were matched for disease, disease status, graft type, age, sex, and year of transplant. Approximately 35 percent of patients had anti-HLA antibodies. The presence of anti-HLA antibodies did not predict graft failure, but 9 of 10 recipients with donor-specific anti-HLA antibodies failed to engraft.
KIR gene haplotype — The impact of matching for non-HLA sites is not as well studied. The killer-cell immunoglobulin-like receptor (KIR) gene complex encodes up to 15 genes for natural killer cell immunoglobulin-like receptors that recognize epitopes of HLA-A, HLA-B, and HLA-C, which are also called KIR ligands. KIR haplotypes can be activating or inhibitory. More specifically, individuals vary in the number of KIR genes that they possess, and the complex has been broadly characterized as A and B haplotypes. KIR-haplotype A mainly encode inhibitory receptors and only one activating receptor, while KIR-haplotype B encode more activating receptors (eg, KIR2DS1, KIR3DS1). Initial reports suggest that the donor's KIR gene haplotype may influence transplant outcomes, but this may depend on the specific conditions of transplantation [55-60].
A study of 1409 unrelated transplant donors providing grafts for patients with AML or acute lymphoblastic leukemia (ALL) reported that KIR ligand status affected the outcomes of patients with AML, but not those of patients with ALL [56]. Patients with AML who received grafts from donors with B motifs had a lower rate of relapse and superior survival rates. When compared with those of homozygous A/A donors, grafts from homozygous B/B donors resulted in a lower cumulative incidence of relapse (15 versus 36 percent).
In another study of 1277 patients with AML who underwent unrelated 10 of 10 HLA match or 9 of 10 HLA single mismatch allogeneic transplant, the impact of KIR haplotype on outcomes was at least partially related to the HLA-C haplotype of the donor [58,61]. The rate of relapse was lower among patients whose donor had the KIR2DS1 haplotype (27 versus 33 percent). This benefit from the KIR2DS1 haplotype was primarily seen in transplants from donors who were homozygous or heterozygous for HLA-C1 antigens. In contrast, donor KIR3DS1 positivity did not impact AML relapse rates, but was associated with decreased mortality following transplant (60 versus 67 percent).
The KIR gene haplotype can easily be performed along with HLA genotyping. When there are multiple HLA-matched potential unrelated donors for a patient with AML, analysis of the KIR gene haplotype may provide guidance in selecting the best match.
ALTERNATIVE DONORS — When a 6 of 6 matched sibling donor is not available, the transplant center usually proceeds with an unrelated donor search (algorithm 1). If an urgent transplant is required, some clinicians will choose to search for an umbilical cord blood graft or a haploidentical donor since these grafts are usually quicker to acquire. Alternative donor sources (umbilical cord blood, haploidentical donors) may allow for a shorter time to transplant (eg, <1 month), but with a potentially increased risk of transplant-related complications.
Umbilical cord blood donors — Relatively high numbers of hematopoietic stem cells are present in umbilical cord blood (UCB) collected at the time of delivery. These cells can be processed and cryopreserved in cord blood banks. Units can be searched similar to the way that unrelated donors are identified. Following HLA matching to a recipient, the cord blood unit can be transported to the transplant center with minimal delay and used. (See "Sources of hematopoietic stem cells" and "Collection and storage of umbilical cord blood for hematopoietic cell transplantation".)
Unrelated UCB offers many practical advantages over unrelated donor bone marrow or mobilized peripheral blood progenitor cells as a source of hematopoietic stem cells including an expanded donor pool, ease of procurement and lack of donor attrition, and decreased graft-versus-host disease (GVHD). (See "Selection of an umbilical cord blood graft for hematopoietic cell transplantation", section on 'Features of UCB grafts'.)
Limitations to UCB include an increased risk of graft failure, delayed immune reconstitution, and unavailability of the donor for additional donations (ie, donor lymphocyte infusions). Efforts are underway to develop strategies to expand UCB hematopoietic stem cells. (See "Umbilical cord blood transplantation in adults using myeloablative and nonmyeloablative preparative regimens".)
Mismatched related donors
Single antigen mismatched siblings — In most centers, a complete match at the HLA-A, -B and -DR loci is required for that individual to be used as a related donor. In some cases, siblings are mismatched at a single HLA locus. Transplants from single antigen mismatched siblings in the graft-versus-host direction are associated with a higher risk of GVHD, although overall survival (OS) may not be different from that observed with fully matched siblings [62,63]. Transplants from single antigen mismatched siblings in the host-versus-graft direction are associated with higher graft failure and lower survival [63-65].
It is not known how single antigen mismatched sibling donors compare with fully matched unrelated donors. A retrospective study from the Japanese Hematopoietic Cell Transplant Registry compared outcomes among 327 patients who received a serologically defined single antigen 5 of 6 (HLA-A, -B, or -DRB) mismatched related donor to those of 327 patients who received an 8 of 8 (HLA-A, -B, -C, -DR) allele matched unrelated donor [66]. Transplantation with a single antigen mismatched related donor was associated with lower survival at two years (44 versus 59 percent; HR 1.49, CI 1.19-1.86). Therefore, in those unusual situations where there is a single antigen mismatched related donor, attempting to find a fully matched unrelated donor may be preferred.
The use of one antigen mismatched siblings in the graft-versus-host direction is considered in some centers, depending on the clinical situation (eg, high- versus standard-risk patient, amount of time available for a matched unrelated donor search), with results that are generally similar to those obtained using fully matched siblings or fully matched unrelated donors [63,67].
Haploidentical donors — Haploidentical related hematopoietic cell donors are those that are mismatched at 3 of 6 loci (HLA-A, -B and -DR). Encouraging data have been obtained from a number of groups using a variety of different approaches [68-74]. Ongoing studies are evaluating novel approaches to induce graft-versus-tumor (GVT) activity and decrease GVHD in recipients of haploidentical grafts. The Blood and Marrow Transplant Clinical Trials Network (BMT CTN) is conducting a randomized trial comparing double UCB transplantation to haploidentical transplantation using post-transplant cyclophosphamide as part of the GVHD prevention strategy. The BMT CTN is also conducting a three-arm randomized trial to compare post-transplant cyclophosphamide versus calcineurin inhibitor/methotrexate versus CD34+ cell selection.
A potential advantage of using haploidentical donors is that within a given family there are usually multiple individuals who could serve as potential donors including parents, siblings, and children [75]. As a result, the donor used could be chosen depending on HLA type, cytomegalovirus status, blood type, or other features that could have potential clinical implications [76-79]. Another advantage is equal availability of donors for all ethnic and racial groups, in contrast to matched unrelated donors. (See "HLA-haploidentical hematopoietic cell transplantation".)
LIKELIHOOD OF FINDING A MATCH — With the expanded use of alternative donors, most candidates for HCT will be able to find an HLA-matched or minimally mismatched donor. White individuals of European descent are the most likely to find an optimal donor, whereas Black individuals from Central or South America are the least likely.
The impact of race and ethnicity on the likelihood of finding a match was analyzed by the National Marrow Donor Program (NMDP) using data from over 10 million adult donors available at the end of 2012 [30]. They defined an optimal unrelated adult donor as one matched at HLA-A, HLA-B, HLA-C, and HLA-DRB1 at the allele level (8 of 8 HLA-match). The probability of finding an available 8 of 8 or ≥7 of 8 HLA matched adult donor varied by racial and ethnic group:
●White European – 75 and 97 percent
●Middle Eastern or North African – 46 and 90 percent
●African American – 19 and 76 percent
●African – 18 and 71 percent
●Black Central or South American – 16 and 66 percent
●Black Caribbean – 19 and 74 percent
●Chinese – 41 and 88 percent
●Korean – 40 and 87 percent
●South Asian – 33 and 84 percent
●Japanese – 37 and 87 percent
●Filipino – 40 and 83 percent
●Southeast Asian – 27 and 76 percent
●Vietnamese – 42 and 84 percent
●Hawaiian or Pacific Islander – 27 and 72 percent
●Mexican – 37 and 87 percent
●Hispanic Central or South American – 34 and 80 percent
●Hispanic Caribbean – 40 and 83 percent
●Native North American – 52 and 91 percent
●Native Central or South American – 49 and 87 percent
●Native Caribbean – 32 and 77 percent
●Native Alaskan – 36 and 83 percent
The same study evaluated the impact of race and ethnicity on the likelihood of finding a match among over 180,000 available umbilical cord blood units [30]. Optimal cord blood units were defined as those of adequate cell dose matched at the antigen level at HLA-A and HLA-B and matched at HLA-DRB1 at high resolution (6 of 6 HLA match). Optimally matched cord blood units were uncommon and ranged in likelihood from 1 percent for those with Black Caribbean ancestry to 17 percent among those with White European ancestry. The likelihood of finding an optimal match was higher for children and ranged from 5 percent for those with African ancestry to 38 percent for those with White European ancestry. A ≥4 of 6 HLA matched cord blood unit could be identified in >80 percent of adults and >95 percent of children.
OTHER CONSIDERATIONS
Donor eligibility — Transplant donors must be in generally good health without other comorbid conditions. The donor must have a performance status that will permit the safe collection of the cells, either by bone marrow or peripheral blood progenitor cell (PBPC) collection [80,81]. A study from the National Marrow Donor Program estimated that approximately half of White donors and one-quarter to one-third of other donors are ultimately available for donation [30]. Evaluation of potential donors is discussed in more detail separately. (See "Sources of hematopoietic stem cells" and "Evaluation of the hematopoietic cell transplantation donor".)
Effect of donor characteristics
Age, sex, and parity — Donor age, sex, and parity may impact the incidence of transplant-related complications, but these factors are only relevant when there are multiple potential donors. When available, preference is given to younger donors and to male or nulliparous female donors. For sex-mismatched donors, especially for female donors to male recipients (FtoM), the choice of bone marrow versus PBPC grafts may affect outcomes. Older donors are more likely to exhibit clonal hematopoiesis of indeterminate potential (CHIP), but the impact of transplanting a CHIP-involved graft is not well-defined. CHIP is discussed separately. (See "Clonal hematopoiesis of indeterminate potential (CHIP) and related disorders of clonal hematopoiesis".)
The effect of donor age was best demonstrated in a retrospective analysis from the Center for International Blood and Marrow Transplant Research (CIBMTR) of over 11,000 unrelated transplants performed from 1988 to 2011 that evaluated the effects of various donor characteristics (eg, age, sex, cytomegalovirus [CMV] serologic status, ABO compatibility, race, and parity) on recipient outcome [82]. Following adjustment for patient disease and transplant characteristics, age and donor-recipient HLA match were the only donor traits significantly associated with overall survival (OS). For every 10-year increment in donor age, there was a 5.5 percent increase in the hazard ratio (HR) for mortality. Older donor age was also associated with an increase in acute graft-versus-host disease (GVHD) but not chronic GVHD. Donor sex, parity, and CMV serostatus did not appear to impact survival. However, recipients with female donors who had undergone multiple pregnancies had a higher rate of chronic GVHD than recipients with male donors (HR 1.22) or nulliparous female donors.
Other studies have found that younger donors are associated with a lower incidence of secondary graft failure [83], B cell lymphoproliferative disorders [84], and obstructive lung disease [85] following allogeneic transplantation. A common question is whether it is preferable to utilize a fit older (>60 years) donor rather than a younger matched unrelated donor. Although both are acceptable donor sources, a retrospective analysis of 1174 patients found that recipients of older related donors fared better [8].
Presently, the effect of CHIP in a transplanted graft on recipient outcomes is not well-defined. Using targeted sequencing of a 66 gene panel, a retrospective analysis of 500 healthy related donors ≥55 years reported CHIP in 16 percent; the median variant allele frequency (VAF) was 5.9 percent [86]. With median follow-up >3 years, donor cell leukemia was observed in 2 of the 80 recipients of CHIP-involved grafts, but the donor’s CHIP status did not affect survival. Compared to donors without evidence of CHIP, patients who received CHIP-involved grafts had more chronic GVHD and a lower cumulative incidence of relapse/progression, but there was no effect on nonrelapse mortality rate (NRM), acute GVHD, or CMV reactivation. In a study of 42 long-term survivors of allogeneic HCT (median follow-up 16 years), CHIP was detected in 10 of the donors; donor-engrafted clonal hematopoiesis was observed in five recipients and in one donor-recipient pair, clonal hematopoiesis progressed to myelodysplastic syndrome in both the donor and the recipient [87].
For FtoM HLA-matched related donor transplantation, compared with patients who received bone marrow grafts, patients who received PBPC grafts had inferior survival and higher NRM, based on analysis of a Japanese transplantation registry database [88]. Compared to 95 patients who received bone marrow grafts, 220 patients who received PBPC grafts had inferior two-year OS: 62 versus 76 percent; hazard ratio (HR 1.9; 95% CI 1.1-3.4). PBPC was also associated with higher two-year NRM: 21 versus 10 percent; HR 3.7 (95% CI 1.4-9.6). FtoM HCT with PBPCs was associated with higher rates of fatal GVHD and organ failure. No differences were reported for outcomes with PBPC versus bone marrow for male to female or sex-matched transplantation.
CMV status — CMV status may impact donor selection when there are more than one otherwise equal donor options. CMV-seropositive recipients have worse survival than seronegative recipients as a group. The best outcomes are seen in CMV-negative recipients with a CMV-negative donor. If the recipient is CMV seronegative, a CMV-seronegative donor is preferred since this will greatly reduce the risk of subsequent infection with CMV, assuming that the patient receives CMV-seronegative blood products. If the recipient is CMV seropositive, some centers advocate for the use of a CMV-seropositive donor with the hope that this will result in a faster restoration of CMV immunity.
Retrospective studies have had conflicting reports regarding the impact of donor CMV status on recipient survival. Superior outcomes were seen when a CMV-seronegative donor was paired with a CMV-seronegative recipient in an analysis of over 16,000 patients with de novo acute leukemia who had undergone allogeneic HCT [89]. When compared with donor and recipient pairs that were both CMV seronegative, HCT pairs in which the donor and/or recipient were CMV seropositive had inferior rates of leukemia-free survival (LFS; 44 versus 49 percent) and OS (50 versus 56 percent) at two years and an increased NRM (23 versus 20 percent). On subset analysis, the impact of CMV status appeared to be greater for those with acute lymphoblastic leukemia (ALL) than for those with acute myeloid leukemia (AML). In contrast, there was no identified impact of donor CMV status on mortality in another large retrospective study [82].
EBV status — Epstein-Barr virus (EBV) status is not usually incorporated into donor selection, but it impacts the risk of post-transplantation lymphoproliferative disorder (PTLD) and GVHD. Since the vast majority of the adult population is EBV seropositive, most donors and recipients will be EBV positive.
Patients receiving grafts from EBV-seropositive donors have an increased risk of developing PTLD, acute GVHD, and chronic GVHD, but rates of relapse, NRM, and OS do not appear to be affected [90,91]. (See "Epidemiology, clinical manifestations, and diagnosis of post-transplant lymphoproliferative disorders", section on 'EBV serostatus'.)
ABO and Rh status — ABO and Rh compatibility are not required between the donor and recipient [92,93]. Hemolysis, delayed erythropoietic engraftment, and pure red cell aplasia (PRCA) may complicate ABO-incompatible transplantation after either marrow or PBPC transplantation. However, these complications are not common [94-100]. A meta-analysis of available studies has concluded that there was no adverse association between any ABO mismatching and survival following allogeneic HCT [101].
Anti-D seldom develops in RhD-mismatched recipients of a myeloablative allogeneic HCT, but may occur following nonmyeloablative conditioning [102,103]. Data from three studies suggest that major ABO-incompatibility is not a barrier to successful nonmyeloablative transplantation [104-106], although transplant-related morbidity and mortality were higher in a fourth study [107]. It is not clear whether major or minor ABO mismatches are associated with a higher incidence of GVHD and/or reduced OS [108-110].
If major mismatches are present, additional manipulation of the hematopoietic cell product, such as depleting the product of red cells if bone marrow is being utilized or depleting plasma prior to infusion, may or may not be required [94,111,112].
Some groups have reported a high incidence of severe immune hemolysis and transplant-related mortality in patients receiving ABO-mismatched grafts [107]. However, one study reported that only 1 among 20 patients undergoing nonmyeloablative conditioning developed mild hemolysis following prophylactic exchange using group O red blood cells for a minor and/or bidirectional ABO mismatch [113]. Outcomes were improved compared to their historical experience, but this approach is not widely used.
Cases of delayed hemolysis and/or PRCA have been described in patients with pre-existing alloantibodies against Rh or ABO antigens, despite the presence of 100 percent donor chimerism of the circulating lymphocytes [114,115]. This phenomenon, including persistence of other host-derived antibodies post-transplantation (eg, anti-platelet alloantibodies), may be due to the presence of long-lived antibody-producing recipient plasma cells surviving the transplantation regimen [115-117]. (See "Acquired pure red cell aplasia in adults" and "Hematopoietic support after hematopoietic cell transplantation", section on 'Red blood cell transfusion'.)
Graft content of invariant NKT cells — Graft content of invariant NKT (iNKT) cells may serve as a clinically useful pretransplantation predictor of GVHD after allogeneic HCT, and recovery of iNKT cells after allogeneic HCT may be a useful prognostic marker [118]. iNKT cells express a highly restricted T cell repertoire composed of a single invariant chain (Valpha24Jalpha18 in humans) that responds to glycolipids presented by the nonpolymorphic class I-like molecule, CD1d [119].
Examples of studies that support a role of iNKT cells in GVHD following allogeneic HCT include:
●A total of 117 peripheral blood (72 percent of total) and bone marrow (28 percent) stem cell specimens from HLA-identical siblings (59 percent), matched unrelated donors (36 percent), and mismatched unrelated donors (5 percent) were analyzed prospectively for graft content of CD34+ cells and various T and NK cell subsets [120]. Among the various immunologic subsets, only increased CD4– iNKT cell graft content correlated with a decreased incidence of grade II to IV acute GVHD (HR 0.56; 95% CI 0.38-0.84). Due to variability of absolute numbers of iNKT cells between specimens, in vitro expansion capacity of CD4– iNKT cells above 6.83 versus below that threshold was the best predictor of acute GVHD (9.7 versus 80 percent), while the incidence of relapse at two years was similar.
●Among 80 patients (35 percent HLA-matched related, 50 percent HLA-matched unrelated, and 15 percent HLA-mismatched unrelated donors), those who received higher than median numbers of iNKT cells (>0.11 x 106/kg) experienced improved outcomes as measured by the composite endpoint of GVHD-free and progression-free survival (49 versus 22 percent), but had no difference in the incidence of GVHD or OS [121].
●In a study of 71 adults (60 percent HLA-matched sibling, 30 percent 10 of 10 HLA-matched unrelated, and 10 percent 9 of 10 HLA-matched unrelated donors), post-transplant peripheral blood ratios of iNKT/T cells >10-3 were independently associated with a reduced risk of acute GVHD (HR 0.12, 95% CI 0.03-0.40) [122]. With a median follow up of 3.7 years, patients who reached iNKT/T ratios >10-3 before day 90 experienced reduced nonrelapse mortality (6.1 versus 36.7 percent) and improved OS (90 versus 52 percent) at two years, without an increased risk of relapse. As a predictive tool, an iNKT/T ratio of >0.58 x 10-3 on day 15 was associated with a reduced risk of acute GVHD (HR 0.06, 95% CI 0.01-0.23).
Donor use of statins — Statins may alter immune function in ways that may contribute to their antiatherosclerotic effects and appear to reduce immune responses to auto- and allo-antigens [123]. Of relevance, the use of statins reduced the mortality related to acute GVHD in an animal model [124] and was associated with a trend toward decreased risk of grades II to IV acute GVHD in a retrospective analysis of 67 patients who had received allogeneic HCT [125].
An additional, larger retrospective analysis of allogeneic HCT outcomes was performed according to statin use by 567 patients with hematologic malignancies and their HLA identical sibling donors. Compared with the 464 allografts where neither the donor nor the recipient was treated with a statin at the time of HCT, statin use by the HCT donor and not the recipient was associated with a significantly decreased risk of grades III to IV acute GVHD (HR 0.28, 95% CI 0.1-0.9) [123]. Statin-associated GVHD protection was limited to those patients receiving cyclosporine-based postgrafting immunosuppression, and was not observed among those treated with tacrolimus.
In a phase II trial, atorvastatin (40 mg daily by mouth) was administered to 30 sibling donors starting two to three weeks prior to stem cell collection [126]. Atorvastatin was administered at the same dose to the HCT recipient in addition standard GVHD prophylaxis with tacrolimus and a short course of methotrexate. Atorvastatin was well tolerated by the donors, and the cumulative incidence rates of acute and chronic GVHD were 11 and 52 percent, respectively. There did not appear to be increased rates of infection, relapse, or mortality.
Randomized trials are needed to further evaluate the efficacy and toxicity of statins in this setting. Many questions remain concerning statin use by the HCT donor and its possible protective effect.
Disease status — When a matched sibling donor is not available, the search for an appropriate donor must consider the urgency of the procedure and potential risks of postponing transplant while performing an extensive donor search. If the risk of relapse is high and the chance of finding a perfect match is low, then proceeding with an alternative donor (haploidentical or umbilical cord blood [UCB] transplant) or known mismatch may be preferable to a lengthy search for a full match.
It is not known whether donor selection should be impacted by disease status (eg, refractory disease, detectable measurable residual disease [MRD]; also referred to as minimal residual disease) for reasons other than expediency. However, in one retrospective study, leukemia patients with MRD positivity prior to transplant had less relapse and better survival with an unrelated UCB transplant than with peripheral blood progenitor cell or bone marrow from an HLA-matched unrelated donor or an HLA-mismatched unrelated donor [127]. In contrast, a benefit for UCB transplant was not seen among those with MRD negativity. Further analysis suggested that the use of UCB abrogated the negative impact of MRD positivity on patient outcome. Further trials are needed to confirm these findings and evaluate whether other donor groups (eg, haploidentical transplant) impact outcomes in MRD-positive patients.
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient education" and the keyword(s) of interest.)
●Basics topics (see "Patient education: Allogeneic bone marrow transplant (The Basics)")
●Beyond the Basics topics (see "Patient education: Hematopoietic cell transplantation (bone marrow transplantation) (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●There are multiple options available in selecting a donor and stem cell source for an allogeneic hematopoietic cell transplantation (HCT). The availability of multiple donor options has been a major advance for patients in need of an allogeneic HCT. The clinical problem, overall health of the donor and recipient, infectious history, clinical approach of the transplant center, and other factors are important in deciding what type of donor is selected. (See 'Other considerations' above and "Sources of hematopoietic stem cells".)
●With the expanded use of alternative donors, most patients will be able to find a human leukocyte antigen (HLA)-matched or minimally mismatched donor. White individuals of European descent are the most likely to find an optimal donor, whereas Black individuals from Central or South America are the least likely. (See 'Likelihood of finding a match' above.)
●Matching donor and recipient for HLA class I (-A, -B, and -C) and class II (-DRB1 and -DQB1) haplotypes is a key part of successful allogeneic HCT. Techniques for HLA typing and the HLA nomenclature have evolved with the increasing use of unrelated donors and the discovery of additional HLA alleles. (See 'Definitions' above.)
●The search for an appropriate donor must consider the urgency of the procedure and potential risks of postponing transplant. All patients likely to require an allogeneic HCT should undergo high resolution HLA typing for HLA-A, -B, -C, and -DR soon after diagnosis to allow for the timely identification of an appropriate donor. (See 'Overview' above.)
●When available, a matched sibling donor is preferred over other donor sources due to improved clinical outcomes following transplant (eg, less graft-versus-host disease) and the speed and cost-effectiveness of the search. As such, the initial search should center on molecular HLA typing of the patient's full biological siblings (algorithm 1). If a 6 of 6 matched sibling donor is identified and eligible for donation, transplantation should proceed. (See 'General approach' above.)
●When a 6 of 6 matched sibling donor is not available, the transplant center usually proceeds with an unrelated donor search. An 8 of 8 matched unrelated donor results in similar survival rates to those seen with a matched sibling donor, but with more acute and chronic graft-versus-host disease. (See 'Initial matching' above.)
●When an 8 of 8 matched unrelated adult donor is not available, options include a 7 of 8 matched unrelated adult donor, an at least 4 of 6 matched umbilical cord blood unit with adequate cell dose, and a haploidentical donor (eg, parent, child, sibling). A choice among these must take into consideration the likelihood of confirming the match, obtaining the specimen, and time to transplant. Umbilical cord blood and haploidentical donors may allow for a shorter time to transplant, but with a potentially increased risk of transplant-related complications. (See 'HLA gene haplotypes' above.)
●A choice among several equally matched donors is further narrowed based on the type of graft planned (bone marrow versus peripheral blood), the mismatched HLA gene, further typing of other HLA genes, and an evaluation of the recipient for donor-specific HLA antibodies. (See 'HLA gene haplotypes' above.)
●The choice between available donors is also dependent on the particular expertise and focus of the individual transplant center. The role of haploidentical related donors is under ongoing clinical investigation and is particularly attractive in settings where there is no matched related donor and not time to search for an unrelated donor. (See 'Haploidentical donors' above.)
