INTRODUCTION — Allogeneic hematopoietic cell transplantation (HCT) is a potentially curative therapy for a wide variety of malignant and non-malignant hematologic disorders. The pluripotent hematopoietic stem cells and lymphocytes required for this procedure are usually obtained from the bone marrow or peripheral blood of a related or unrelated donor. Historically, the best results of allogeneic HCT have been obtained when the stem cell donor is a human leukocyte antigen (HLA)-matched sibling.
Given the small family sizes in developed nations and the 25 percent likelihood that any sibling is fully HLA-matched to the patient, an HLA-matched sibling can be found for only approximately 30 percent of patients. For patients who lack an HLA-matched sibling, alternative sources of donor grafts include suitably HLA-matched or partially mismatched adult unrelated donors, umbilical cord blood stem cells, and partially HLA-mismatched, or HLA-haploidentical, related donors. Substantial improvements in haploidentical transplantations have occurred in the last decade and there is growing use of haploidentical family donors. 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.
The major challenge of HLA-haploidentical HCT is intense bi-directional alloreactivity leading to high incidences of graft rejection and graft-versus-host disease (GVHD). Advances in graft engineering and in pharmacologic prophylaxis of GVHD have reduced the risks of graft failure and GVHD after HLA-haploidentical HCT, and have made this stem cell source a viable alternative for patients lacking an HLA-matched sibling.
This topic will discuss the advantages and disadvantages of HLA-haploidentical HCT and the selection of an HLA-haploidentical donor. A general approach to donor selection for allogeneic HCT is discussed separately. (See "Donor selection for hematopoietic cell transplantation".)
DEFINITIONS — An HLA-haploidentical donor is one who shares, by common inheritance, exactly one HLA haplotype with the recipient and is mismatched for a variable number of HLA genes, ranging from zero to six (HLA-A, -B, -C, -DRB1, DQB1, and -DPB1), on the unshared haplotype. Potential HLA-haploidentical donors include biological parents; biological children; full or half siblings (figure 1); and even extended family donors such as aunts, uncles, nieces, nephews, cousins, or grandchildren.
Only a related donor can be deemed to be HLA-haploidentical to a recipient because the Sanger-based sequencing methods widely used for high resolution HLA typing cannot determine whether nucleotides at polymorphic sites are located on the same (cis-position) or on a different (trans-position) chromosome , and so alleles are assigned to haplotypes only with an assumption of relatedness. Next-generation sequencing permits direct assignment of alleles to haplotypes and may reveal previously unknown relatedness between individuals in the future.
DECIDING TO USE A HAPLOIDENTICAL DONOR
Choice between alternative donor sources — The search for an appropriate stem cell source must consider the urgency of the procedure and potential risks of postponing transplant. Our general approach to donor selection is described in more detail separately (algorithm 1). (See "Donor selection for hematopoietic cell transplantation", section on 'General approach'.)
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 [GVHD]) and the speed and cost-effectiveness of the search. HLA typing of siblings and a preliminary search for an HLA-matched unrelated donor are often conducted concurrently, in case the family typing reveals no HLA-matched sibling. Alternative donor sources (HLA-haploidentical HCT, umbilical cord blood transplant) may be considered if there is an urgent need to proceed to transplantation or if the preliminary search indicates a low likelihood of finding an eight of eight allele-matched unrelated donor.
Despite an ever-increasing number of volunteers in the unrelated donor registries, unrelated adult donor HCT is performed in less than half of patients for whom an unrelated donor search has been activated. The following studies illustrate this issue:
●In a study of 326 consecutive Italian patients activating a donor search, only 121 patients (37 percent) underwent unrelated donor transplantation at a median of 169 days (range 68 to 772 days) after search activation . Searches were stopped in 192 patients (59 percent) because of patient death (n = 100), disease progression (n = 34), unlikelihood of finding a donor (n = 11), withdrawal of consent (n = 8), or choice of an alternative program (n = 39).
●In a similar study from Holland, unrelated donor transplants were performed in 86 of 240 patients (36 percent) activating a donor search .
The United States Blood and Marrow Transplant Clinical Trials Network conducted a phase 3 trial of reduced intensity conditioning and transplantation of either double unrelated donor umbilical cord blood (UCB) or HLA-haploidentical bone marrow for chemosensitive lymphoma or acute leukemia in remission (BMT CTN 1101; NCT01597778). Two-year progression-free survival, the primary endpoint, was 35 percent (95% CI 28-42 percent) compared with 41 percent (95% CI 34-48 percent) after UCB and haploidentical transplants, respectively. Two-year nonrelapse mortality was higher after UCB (18 percent; 95% CI 13-24 percent) than after haploidentical transplantation (11 percent; 95% CI 6-16 percent). This led to lower two-year overall survival after UCB compared with haploidentical transplantation, 46 percent (95% CI 38-53) and 57 percent (95% CI 49-64 percent), respectively. Though these results favor haploidentical over UCB transplantation, further follow-up is required to make definitive conclusions.
Advantages and limitations of haploidentical donors — Substantial improvements in haploidentical transplantations have permitted the growing use of haploidentical family donors [4,5]. Sources of stem cells for allogeneic HCT include HLA-matched siblings, suitably HLA-matched unrelated donors, HLA-haploidentical donors, and unrelated umbilical cord blood. When compared with the other stem cell sources, the major advantages of the HLA-haploidentical donor option include:
●Near universal availability of highly motivated donors – Patients have an average of 2.7 potential HLA-haploidentical donors among first degree relatives. In comparison, only approximately 30 percent of patients will have a HLA-matched sibling, and availability of an unrelated donor genotypically matched at eight of eight alleles (HLA-A, -B, -C, and -DRB1) ranges from 16 to 75 percent depending on the recipient's ethnic background .
●Rapid availability – The time to identify and mobilize an adult unrelated donor can be longer than three months for up to 25 percent of patients. An HLA-haploidentical donor can be identified and mobilized in two weeks to one month.
●Adequate doses of hematopoietic stem cells (HSCs) – HLA-haploidentical grafts have sufficient doses of HSCs for transplantation and of memory T cells for immune reconstitution. In contrast, the total dose of nucleated cells in a single umbilical cord blood unit may be suboptimal for engraftment in larger adults and immune reconstitution is delayed. (See "Strategies for immune reconstitution following allogeneic hematopoietic cell transplantation", section on 'Source of hematopoietic stem cells' and "Selection of an umbilical cord blood graft for hematopoietic cell transplantation", section on 'Cell dose'.)
●Low cost of graft acquisition – The costs of acquiring grafts from adult unrelated donors and especially from umbilical cord blood banks can be substantially higher than that of related donors.
●Availability of the donor for repeated donations of HSCs or lymphocytes to treat relapse – In contrast, umbilical cord blood is a non-recurring source of cells. If a patient suffers relapse of the underlying hematologic malignancy after umbilical cord blood transplantation, donor lymphocytes to treat the relapse are not available. (See "Immunotherapy for the prevention and treatment of relapse following allogeneic hematopoietic cell transplantation".)
●Graft-versus-leukemia effect – For patients with high-risk acute leukemia, HLA-haploidentical HCT may be associated with a stronger graft-versus-leukemia effect compared with HLA-matched sibling HCT, resulting in a lower cumulative incidence of relapse  and an improved overall survival .
The major challenge of HLA-haploidentical HCT is the high frequency of host and donor T cells reactive to HLA alloantigens resulting in intense bi-directional alloreactivity and, in the absence of effective prophylactic measures, high incidences of fatal graft rejection or severe or fatal GVHD.
The consequences of intense bi-directional alloreactivity after mismatched transplants were illustrated in an analysis of over 2000 allogeneic HCTs performed between 1985 and 1991 and reported to the International Bone Marrow Transplant Registry . When compared with HLA-matched sibling HCT, two HLA antigen-mismatched related donor transplants resulted in higher rates of the following adverse transplant outcomes:
●Transplant-related mortality (55 versus 21 percent at three years among patients with leukemia)
●Graft failure (16 versus 1 percent)
●Grade II to IV acute GVHD (56 versus 29 percent)
●Severe (grade III/IV) acute GVHD (36 versus 13 percent)
●Chronic GVHD (60 versus 42 percent)
Attempts at T cell depletion of the donor graft reduced the incidence of acute GVHD, but at the cost of increased incidence of graft rejection, and did not improve leukemia-free survival . Subsequently, several advances in graft engineering and pharmacologic modulation of alloreactivity have reduced the incidences of GVHD and non-relapse mortality, improved overall and progression-free survival, and made this graft source an acceptable alternative for patients lacking an HLA-matched sibling or unrelated donor. (See "Prevention of graft-versus-host disease".)
DONOR SELECTION FOR HAPLOIDENTICAL HCT
Our approach — When a decision is made to proceed with HLA-haploidentical HCT, the immediate donor options include biological parents, biological children, and full or half siblings. Initial matching efforts will have already identified the HLA type of the patient and potential full sibling donors. If the initial typing of the full siblings has identified an HLA-haploidentical donor and that sibling is eligible and willing to donate, it may not be necessary to perform HLA typing of parents, children, or half siblings. However, even if there is at least one HLA-haploidentical sibling, our general practice is to type up to four potential HLA-haploidentical first degree relatives to optimize donor selection and to increase the likelihood that at least one donor will pass eligibility screening.
Most patients will have more than one HLA-haploidentical first degree relative willing and able to donate. Several of the criteria for selecting donors, such as ABO blood type and cytomegalovirus (CMV) serostatus of the donor and recipient, sex mismatch, and donor age and parity, apply to HLA-haploidentical transplant but are not unique to partially HLA-mismatched HCT and are discussed separately. (See "Donor selection for hematopoietic cell transplantation".)
Unique considerations for partially HLA-mismatched HCT include donor-specific HLA antibodies, donor relationship, donor-recipient HLA mismatch, non-inherited maternal antigens, and natural killer cell alloreactivity. These are discussed in the following sections. Donor selection guidelines that are used at our institution are listed in the table (table 1).
The only absolute contraindications to the use of a specific HLA-haploidentical donor are if the donor is medically or psychologically unfit or if the recipient has anti-donor HLA antibodies of sufficient strength to result in a positive crossmatch result by flow cytometry or by complement-dependent cytotoxicity assay. Using modern graft-versus-host disease (GVHD) prophylaxis protocols, there is no need to minimize the extent of mismatch between donor and recipient. As with other graft sources, clinicians prioritize ABO compatible donors, those with CMV immunoglobulin G (IgG) serologic status that matches the patient, younger adults rather than older adults , and male or nulliparous female donors rather than multiparous female donors for male recipients. Blood from donors who ≥50 years should be tested by next generation sequencing to identify clonal hematopoiesis of indeterminate potential (CHIP), which may be associated with a higher incidence of chronic graft-versus-host disease  or donor-derived leukemia .
Donor-specific HLA antibodies — A potential donor should not be used if the patient has donor-specific HLA antibodies (DSA) against that donor. If the patient has no HLA-matched donors and is sensitized to all potential HLA-haploidentical donors, strong consideration should be given to finding an unrelated donor or umbilical cord blood graft to which the patient has not been sensitized. If no other graft source can be identified, desensitization to lower the concentration of DSA and facilitate engraftment can be considered in centers with the appropriate expertise in HLA antibody testing and interpretation of results [14-16].
Patients may become sensitized to HLA alloantigens by prior blood transfusions, allogeneic transplantation, or pregnancy. Among patients referred for allogeneic HCT, approximately 20 to 23 percent will have detectable HLA antibodies and 15 percent will have DSA [14,15,17]. In a study of 296 candidates for haploidentical HCT, of whom 38 percent were female, the overall incidence of DSA was 15 percent . DSA against at least one HLA-haploidentical donor was found in 5 percent (9 out of 185) of males, 13 percent (6 out of 48) of nulliparous females, and 43 percent (27 out of 63) of parous women.
DSA have consistently been associated with an increased risk of graft failure after HLA-haploidentical HCT [14,15,17]. In one study, graft failure occurred in three of four haploidentical HCT recipients with DSA as compared with 1 of 20 patients without DSA  and, in another study, in three of five patients with high levels of DSA .
Donor-recipient HLA mismatch — Early studies of HLA-haploidentical HCT using T cell-replete bone marrow grafts showed significantly worse outcomes with increasing donor-recipient HLA disparity [9,18]. However, with modern supportive care including graft engineering and GVHD prophylactic regimens developed specifically for HLA-haploidentical HCT, the deleterious effect of HLA mismatching on transplantation outcome has been substantially reduced or even eliminated. Using modern GVHD prophylaxis protocols, there is no need to minimize the extent of mismatch between donor and recipient.
Examples of studies that have examined the association of HLA mismatching with transplant outcomes include:
●A study of 509 recipients of unmanipulated haploidentical HCT performed for acute leukemia found no association between the cumulative number of HLA mismatches and GVHD, transplant-related mortality, disease-free survival, or relapse . The risk of grade ≥2 acute GVHD was associated with antigenic (but not allelic) mismatch at the HLA-DRB1 locus in patients who received post-transplantation cyclophosphamide (PTCy; 313 patients), but not in those who received anti-thymocyte globulin (196 patients).
●Among 481 recipients of T cell-replete bone marrow plus peripheral blood from filgrastim-stimulated donors, HLA-B mismatch was associated with an increased incidence of acute, grades II to IV GVHD and with an increased risk of non-relapse mortality; however, increasing overall HLA disparity was not associated with an increased risk of non-relapse mortality or worse leukemia-free survival .
●Increasing HLA disparity between donor and recipient had no detrimental impact on the outcome of 185 hematologic malignancy patients treated with nonmyeloablative conditioning, T cell-replete bone marrow transplantation, and GVHD prophylaxis including PTCy . In this study, the presence of an HLA-DRB1 antigen mismatch in the graft-versus-host direction was associated with a decreased risk of relapse and improved survival. The favorable association of HLA Class II (HLA-DRB1, -DQB1, and non-permissive -DPB1) mismatching on relapse and survival was found in an independent, single center study of 208 patients undergoing haploidentical HCT with PTCy .
●Three retrospective analyses of HLA-haploidentical HCT, two using PTCy as GVHD prophylaxis [23,24] and one employing the GIAC protocol , have shown no significant difference in overall or disease-free survival between recipients of grafts from HLA-matched siblings versus HLA-haploidentical donors, supporting the hypothesis that these transplantation platforms have nullified the detrimental impact of HLA mismatching on outcome.
Non-inherited maternal HLA antigens — HLA-haploidentical siblings share inheritance of either the paternal or the maternal HLA haplotype (figure 1). Those who share paternal HLA antigens are mismatched for both inherited and non-inherited maternal HLA haplotypes. In utero exposure of the fetus to maternal cells may induce immunologic hyporesponsiveness to non-inherited maternal HLA antigens (NIMA), which may result in reduced alloreactivity against an NIMA-mismatched, HLA-haploidentical sibling after HCT. While further studies are needed to clarify the role of NIMA in donor selection, if other factors are equal, mismatching of maternal, rather than paternal, antigens is better tolerated in haploidentical transplant. Selection based on NIMA requires HLA typing of at least one parent to determine the origin of the inherited and non-inherited HLA haplotypes.
The impact of NIMA on outcome after transplant is illustrated in the following studies:
●The following outcomes were noted in a study of 269 subjects receiving one or two antigen-mismatched sibling or parental non-T cell-depleted transplants, using data from the International Bone Marrow Transplant Registry :
•Sibling transplants mismatched for paternal antigens had similar rates of transplant-related mortality and graft failure, but higher rates of acute GVHD than maternal-mismatched transplants (relative risk [RR]: 1.86, 95% CI 1.15-3.05).
•In the first four months post-transplant, father-to-child transplants involved significantly more chronic GVHD than mother-to-child transplants (RR: 2.44, 95% CI 1.12-5.34).
•When compared with maternal-mismatched sibling transplants, a transplant from either parent was equally likely to be associated with an increase in treatment-related mortality (RR: 2 for each).
●A retrospective analysis of 118 consecutive patients with acute leukemia evaluated the effect of sex of the parent on subsequent survival following T cell-depleted haploidentical transplantation from a parent . Results included:
•For a control group of leukemic patients receiving transplants from a haploidentical sibling, donor sex had no effect on post-transplant outcome.
•For patients receiving transplants from a haploidentical parent, five-year event-free survival was significantly better in those who received transplants from the mother, rather than the father (51 versus 11 percent, respectively).
The protective effect of a maternal haploidentical donor was seen in both male and female recipients, suggesting that maternal exposure to the child's alloantigens inherited from the father may have influenced these outcomes.
●In a Japanese study of 35 patients with advanced hematologic malignancies who underwent HLA-2-antigen- or HLA-3-antigen-incompatible stem cell transplant from a microchimeric NIMA-mismatched donor, NIMA mismatch in the graft-versus-host direction was associated with a lower risk of severe grade III to IV acute GVHD when compared with non-inherited paternal antigen mismatch .
●Among 1210 recipients of myeloablative conditioning, HLA-haploidentical HCT, and intensive pharmacologic GVHD prophylaxis including anti-thymocyte globulin, cyclosporine, methotrexate, and mycophenolate mofetil, NIMA-mismatched sibling donors were associated with the lowest incidence of acute GVHD compared with parental donors and noninherited paternal antigen-mismatched sibling donors .
Enhanced natural killer cell alloreactivity — Natural killer (NK) cells may play a role in preventing or treating relapse after HLA-haploidentical HCT. The killer-cell immunoglobulin-like receptor (KIR) gene complex encodes up to 15 genes for NK cell immunoglobulin-like receptors that recognize epitopes of HLA-A, HLA-B, and HLA-C, which are also called KIR ligands. The potential benefit of KIR ligand incompatibility or of KIR haplotype B donors needs to be evaluated prospectively in the haploidentical HCT setting before the widespread incorporation of these factors on selecting donors.
The "missing self" hypothesis states that "self" major histocompatibility complex (MHC) molecules deliver inhibitory signals through NK cell receptors and that the absence of "self" MHC on a virally infected or tumor target permits NK cell cytotoxicity against that target. After HLA-haploidentical HCT, HLA mismatch between donor and recipient creates a situation in which recipient leukemia cells lack inhibitory "self" HLA ligands for KIRs, on donor NK cells, thereby permitting an NK cell-mediated graft-versus-leukemia effect. Absence of HLA ligands on recipient cells for donor inhibitory KIRs (KIR ligand incompatibility) predicted NK cell alloreactivity in vitro and was associated with a potent anti-leukemia effect after T cell-depleted, "megadose" HLA-haploidentical HCT for acute myeloid, but not acute lymphoblastic leukemia [30,31]. In contrast, beneficial effect of KIR ligand incompatibility was not found after T cell-replete, HLA-haploidentical HCT, perhaps because GVHD-producing T cells mask the beneficial effects of NK cell alloreactivity .
KIR genes are inherited as haplotypes. Among patients with hematologic malignancy treated with nonmyeloablative, HLA-haploidentical HCT with high dose, post-transplantation cyclophosphamide, those who were homozygous for the KIR "A" haplotype, which encodes only one activating KIR, had an improved overall survival (hazard ratio [HR] = 0.30; CI 0.13-10.69), event-free survival (HR = 0.47; CI 0.22-1.00), and non-relapse mortality (cause-specific HR = 0.13; CI 0.017-0.968) if their donor expressed at least one KIR B haplotype, which encodes several activating KIRs . KIR haplotype B donors also conferred a reduced risk of relapse after HLA-haploidentical HCT using reduced intensity conditioning and grafts depleted of CD3+ and CD19+ cells , and after HLA-haploidentical HCT for children with acute lymphoblastic leukemia . HLA-haploidentical transplantation from KIR ligand-mismatched donors with at least one KIR B haplotype was also associated with reduced non-relapse mortality, largely infection related, and improved survival .
The impact of KIR gene haplotype on the selection of matched donors is discussed in more detail separately. (See "Donor selection for hematopoietic cell transplantation", section on 'KIR gene haplotype'.)
SELECTION OF GRAFT TYPE — Either bone marrow (BM) or peripheral blood (PB) is suitable as the graft source for haploidentical HCT. There is no evidence that the graft source influences overall survival (OS), but it may influence rates of non-relapse mortality (NRM), leukemia relapse, graft-versus-host disease (GVHD), and cytokine release syndrome (CRS).
No randomized studies have directly compared BM versus PB as the graft source for haploidentical HCT and data regarding outcomes are derived from retrospective analyses. In general, selection of the type of graft source is influenced by institutional expertise and preference.
Informative studies regarding the type of graft source for haploidentical HCT include:
●A large multi-institutional study compared outcomes in 681 patients with hematologic malignancies who underwent haploidentical HCT followed by post-transplant cyclophosphamide (PTCy) . Rates of OS, NRM, and hematopoietic recovery were comparable with either type of graft source, but BM grafts were associated with a lower risk of grade II to IV acute GVHD (hazard ratio [HR] 0.45; 95% CI 0.34-0.61) and chronic GVHD (HR 0.35; 95% CI 0.25-0.49). BM grafts were also associated with a higher risk of relapse, but this effect was limited to patients with leukemia (HR 1.73; 95% CI 1.2-2.4), but not lymphoma. Another study of 451 patients who underwent haploidentical HCT with PTCy for acute leukemia also reported that BM grafts were associated with a lower risk of acute GVHD, but the graft source did not affect rates of chronic GVHD, relapse, or NRM .
●A retrospective, single institution study of 66 patients who underwent haploidentical HCT reported that use of PB grafts was an independent risk factor for higher grade CRS . Compared with patients who received a BM graft, those who received PB grafts had higher risk of grade ≥2 CRS (odds ratio 9.8; 95% CI 3.0-32.5); no other risk factor (eg, performance status, conditioning regimen, dose of CD3+ or CD34+ cells) was independently associated with CRS. Other studies have also reported high rates of CRS with PB grafts [40-45].
Diagnosis and management of CRS are discussed separately. (See "Cytokine release syndrome (CRS)".)
MODERN HAPLOIDENTICAL STRATEGIES
Selecting a strategy — Over the past several decades, numerous approaches to HLA-haploidentical HCT have been developed. The significant heterogeneity in study design and patients treated complicates any comparisons between these approaches. Due to the lack of prospective comparative studies, this decision should be based on the expertise of a particular center.
The three most developed modern approaches used are:
●T cell depletion (TCD) with "mega-dose" CD34+ cells – TCD haploidentical HCT is, in general, associated with high rates of non-relapse mortality (NRM) secondary to slow immune reconstitution and infectious complications. If choosing a TCD platform, the center has to have expertise in the manipulation of stem cells and other cellular products and have in place a strategy for adoptive immunotherapy to accelerate immune reconstitution and reduce the risk of infections. (See 'T cell depletion with mega-dose CD34+ cells' below.)
●The "GIAC" strategy uses GCSF-stimulation of the donor; Intensified immunosuppression post-transplantation; Anti-thymocyte globulin added to conditioning to help prevent GVHD and aid engraftment; and Combination of peripheral blood stem cell and bone marrow allografts – The GIAC strategy is relatively inexpensive and requires no expertise in graft manipulation. When compared with high dose, post-transplantation cyclophosphamide, GIAC appears to be associated with higher rates of acute and chronic graft-versus-host disease (GVHD). There is limited experience with this approach outside of China or Italy. (See 'The GIAC strategy' below.)
●High dose, post-transplantation cyclophosphamide (PTCy) – PTCy is relatively inexpensive and requires no graft manipulation. PTCy has been extended safely to the myeloablative conditioning setting and to the use of peripheral blood progenitor cells as the donor source. (See 'High dose post-transplantation cyclophosphamide' below.)
As with other allogeneic HCT strategies, all haploidentical HCT studies have significantly worse clinical outcomes for patients transplanted in relapse compared with those transplanted in remission (eg, estimated event-free survival rates of ≤14 percent after TCD in relapse).
T cell depletion with mega-dose CD34+ cells — TCD with "mega-dose" CD34+ cells is a haploidentical strategy that requires graft manipulation and is associated with slow immune reconstitution and infectious complications. Initial studies of TCD required negative selection of CD3+ cells by soybean agglutination and erythrocyte rosetting, while subsequent studies have used grafts manipulated with CD34+ positive selection [46-48]. These more modern studies utilizing "mega-dose" CD34+ grafts following intensive conditioning and no additional post-grafting GVHD prophylaxis are associated with engraftment rates of 90 to 95 percent and rates of acute and chronic GVHD of <10 percent; however, NRM is in range of 37 to 53 percent and largely attributed to infection [46-48]. The conditioning regimen used with "mega-dose" CD34+ grafts have evolved through different studies. The initial conditioning included total body irradiation (TBI) (8 Gy in single fraction), thiotepa, cyclophosphamide, and rabbit antithymocyte globulin (ATG). Subsequent modifications included replacement of cyclophosphamide by fludarabine and some patients were given alemtuzumab instead of thymoglobulin [46,48,49].
A European study reported the outcomes of TCD haploidentical HCT in 266 adult patients with de novo acute leukemia . For patients transplanted in remission, engraftment rate was 91 percent, grade II to IV acute GVHD 10 percent, and grade III to IV acute GVHD 6 percent. Chronic GVHD was seen in 14 percent of patients surviving beyond 100 days. Two-year NRM ranged between 36 to 66 percent depending on the disease type and stage at allogeneic HCT. Patients with acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) in first complete remission had two-year disease-free survival (DFS) of 48 and 13 percent, respectively, while outcomes were much worse for patients with advanced disease with two-year DFS of 1 and 7 percent for AML and ALL patients, respectively. Similarly, high NRM and poor outcomes for ALL patients not in complete remission were also reported in pediatric-specific reports [51,52]. Two small studies from North American groups reported poor outcomes with high NRM using the TCD haploidentical HCT approach for patients with advanced hematologic malignancies [53,54].
Efforts to improve immune reconstitution after TCD haploidentical HCT include CD3/CD19 negative selection ; depletion of alpha/beta but not gamma/delta T cells [56,57]; the infusion of cytotoxic T cell lines with viral-specificity for the prevention or treatment of viral infections ; and reintroduction of lower levels of both conventional and regulatory T cells . Another approach combining TCD haploidentical HCT with gene therapy has involved infusing donor lymphocytes expressing suicide genes that could be activated if GVHD developed [60-62].
The GIAC strategy — The GIAC approach developed by the Peking University group in Beijing, China, involves four main components, prompting the acronym "GIAC": GCSF-stimulation of the donor; Intensified immunosuppression through post-transplantation cyclosporine (CsA), mycophenolate mofetil (MMF), and short-course methotrexate; ATG added to conditioning to help prevent GVHD and aid engraftment; and Combination of peripheral blood stem cell (PBSC) and bone marrow (BM) allografts. There is limited experience with this approach outside of China. Conditioning used with this strategy is usually a modified busulfan plus cyclophosphamide regimen with ATG, cytarabine, and semustine (Me-CCNU).
In the initial study, engraftment was achieved in all 171 patients, with the cumulative incidences of grade II to IV and grade III to IV acute GVHD at 100 days of 55 and 23 percent, respectively . The cumulative incidences of chronic GVHD and extensive chronic GVHD at two years were 74 and 47 percent. The two-year probabilities of NRM, relapse, and DFS were 20, 12, and 68 percent for standard-risk disease patients and 31, 39, and 42 percent for high-risk disease patients, respectively.
Several confirmatory studies, including pediatric-specific reports, showed that the GIAC protocol could produce complete engraftment, acceptable NRM, and favorable DFS after T cell-replete haploidentical HCT [63-65]. However, relatively high rates of severe acute and chronic GVHD are associated with this approach. In an initial report, the consortium of Italian investigators modified this approach through using only BM allografts and adding basiliximab . This modified GIAC strategy resulted in a lower rate of chronic GVHD (17 percent) and slightly higher rates of graft failure (7 percent) and NRM (36 percent).
High dose post-transplantation cyclophosphamide — High dose post-transplantation cyclophosphamide (PTCy) is a strategy for haploidentical HCT that is relatively inexpensive and requires no graft manipulation. PTCy has been extended safely to the myeloablative conditioning setting and to the use of peripheral blood progenitor cells as the donor source.
Non-myeloablative haploidentical HCT with PTCy has historically incorporated the following [67,68]:
●Low dose pre-transplant cyclophosphamide
●Nonablative conditioning with fludarabine and low dose TBI
●GVHD prophylaxis consisting of post-transplant cyclophosphamide (50 mg/kg given on each of days +3 and +4), mycophenolate mofetil (MMF, given from day 5 to 35, inclusive) and tacrolimus (given from day 5 to 180, inclusive)
With this regimen, retrospective studies suggest that the significant HLA disparity of haploidentical transplant is not associated with increased acute GVHD or worsened progression-free survival (PFS) in acute leukemias or lymphomas [23,69-72]. Furthermore, there have been no cases of post-transplantation lymphoproliferative disease within the first post-transplant year among 785 patients treated with PTCy .
Support for the use of this regimen comes largely from small prospective trials and retrospective analyses. As examples:
●A retrospective analysis of 372 patients treated with haploidentical HCT and PTCy reported cumulative incidences of NRM and severe acute GVHD at six months of 8 and 4 percent, respectively . The cumulative incidence of chronic GVHD was 13 percent at two years. PFS and overall survival (OS) rates at three years were 40 and 50 percent. When a disease risk index was applied to stratify across histologies, three-year OS rates ranged from 35 to 71 percent. Relapse and OS estimates were comparable to those seen with HLA-matched HCT.
●The Blood and Marrow Transplant Clinical Trials Network (BMT CTN) sponsored a multicenter phase II trial of haploidentical HCT (CTN 0603) for high-risk hematologic malignancies after reduced-intensity conditioning (RIC), which was run in parallel with a phase II trial (CTN 0604) of RIC and transplantation of two units of unrelated umbilical cord blood (dUCB) as the donor source confirmed the results of initial study . Overall, haploidentical HCT with PTCy approach after nonmyeloablative conditioning is associated with low rates of NRM, acute GVHD, and particularly chronic GVHD.
●A retrospective analysis of the Center for International Blood and Marrow Transplant Research (CIBMTR) database reported the outcomes of patients with lymphoma who underwent reduced-intensity transplantation followed by haploidentical HCT with PTCy (185 patients) or HLA-matched unrelated donor transplantation (732 patients) . Haploidentical HCT was associated with lower rates of grade III/IV acute and chronic GVHD and similar rates of relapse, NRM, PFS, and OS.
●Another analysis of the CIBMTR database reported the outcomes of patients with lymphoma who underwent RIC transplantation followed by haploidentical HCT with PTCy versus transplants from HLA-matched siblings . Haploidentical HCT was associated with lower rates of chronic GVHD and similar rates of relapse, NRM, PFS, and OS.
In an effort to reduce relapse rates, other investigators have pursued intensifying the conditioning of this transplantation platform. In two studies from other institutions, myeloablative conditioning prior to haploidentical HCT with PTCy was associated with similar rates of acute GVHD, slightly higher but still favorable rates of chronic GVHD (26 and 35 percent, respectively), similar NRM (18 and 10 percent, respectively), and lower relapse (22 and 40 percent, respectively) [23,24]. The first of these two studies incorporated PBSCs instead of BM as the allograft source . The second study also had spaced the PTCy to days +3 and +5 and started the MMF and cyclosporine before the PTCy . Unlike the preclinical data, tolerance was still induced when CsA was started before PTCy as evidenced by low rates of grade II-IV acute GVHD (12 percent) and chronic GVHD. The second study above also had incorporated PBSCs instead of BM as the allograft source . The majority of patients transplanted by several groups received Seattle-based non-myeloablative conditioning regimen [23,24,67,68,77]. The myeloablative or reduced conditioning regimens used with PTCy after haploidentical allogeneic HCT ranged from standard busulfan/cyclophosphamide, or fludarabine/TBI, fludarabine/melphalan and thiotepa, busulfan, to fludarabine-reduced conditioning .
IMMUNE RECONSTITUTION AFTER HLA-HAPLOIDENTICAL HCT — The breadth and kinetics of immune reconstitution are highly related to the transplantation platform employed, in particular to the extent of donor T cell depletion.
Megadose T cell-depleted (TCD) HCT employs the greatest depletion of donor T cells and has been associated with slow immune reconstitution and infectious mortality approaching 40 percent [47,48]. A variety of approaches have been taken to improve immune reconstitution after TCD HLA-haploidentical HCT  including infusion of pathogen-specific T cells [58,76,79], polyclonal T cell populations depleted of alloreactive T cells , or polyclonal T cells engineered with suicide genes to reduce the consequences of graft-versus-host disease (GVHD) [60,61]. Selective depletion of megadose grafts of B cells and of T cells bearing the alpha/beta heterodimeric receptor retains T cells expressing gamma/delta receptors in the graft. Gamma/delta T cells have been shown to participate in the elimination of human cytomegalovirus (CMV) infection [81,82] and to mediate anti-leukemia effects [83,84], so this novel form of graft engineering may improve immune reconstitution and relapse after HLA-haploidentical HCT.
One study compared immune reconstitution among 50 patients with hematologic malignancy receiving HLA-haploidentical HCT according to the GIAC protocol versus 25 recipients of HLA-matched transplants . Non-relapse mortality, relapse, leukemia-free survival, and overall survival were similar between the two groups, although the cumulative incidence of CMV antigenemia was higher among recipients of HLA-haploidentical as compared with HLA-matched sibling HCT (50 versus 13 percent). Compared with recipients of HLA-matched HCT, recipients of HLA haploidentical HCT had lower T cell and dendritic cell counts in the first 90 days after transplantation, particularly among naïve T cells, whereas B cell and monocyte reconstitution were comparable.
In another study, immune reconstitution was compared among 459 consecutive patients with hematologic malignancies treated with HCT from HLA-matched siblings (SIB; n = 176), matched unrelated donors (MUD; n = 43), mismatched unrelated donors (mMUD; n = 43), unrelated umbilical cord blood (UCB; n = 105), or HLA-haploidentical donors (n = 92) . GVHD prophylaxis comprised cyclosporine and methotrexate for matched siblings, anti-thymocyte globulin for all unrelated donor transplants, and post-transplantation cyclophosphamide (PTCy), cyclosporine, and mycophenolate mofetil for haploidentical HCT. The following outcomes were noted:
●The cumulative incidence of developing CMV antigenemia was 58 percent in the SIB group, 60 percent in the MUD group, 60 percent in the mMUD group, 68 percent in the UCB group, and 74 percent in the haploidentical group.
●The median blood CD4+ cell count at 100 days after transplantation was 229/microL in the SIB group, 106/microL in the MUD group, 90/microL in the mMUD group, 63/microL in the UCB group, and 190/microL in the haploidentical group.
●Infectious mortality was 11 percent in the haploidentical group versus 4 percent in the sibling group, though total non-relapse mortality at 1000 days after transplantation was 24 percent in the SIB group, 33 percent for the MUD group, 35 percent for the mMUD group, 35 percent for the UCB group, and 18 percent for the haploidentical group.
Taken together, the results suggest that immune reconstitution is slightly slower after HLA-haploidentical HCT using the GIAC protocol or PTCy when compared with matched sibling HCT, though non-relapse mortality is not significantly impaired. Both haploidentical platforms appear to achieve superior immune reconstitution when compared with the megadose, T cell-depleted stem cell approach.
MANAGEMENT OF RELAPSE AFTER HLA-HAPLOIDENTICAL HCT — Loss of expression of the mismatched HLA haplotype has been described to occur in as many as 25 percent of patients with relapsed acute leukemia after HLA-haploidentical HCT, including patients treated with post-transplantation cyclophosphamide (PTCy) [86-90].
This loss of HLA molecule expression only occurs on the mismatched haplotype and is significantly less common after HLA-matched sibling  or well-matched unrelated donor HCT [92,93], indicating that the loss of HLA expression is a mechanism of leukemia escape from immune surveillance. The mechanism of HLA loss was loss of heterozygosity in an extensive region on chromosome 6p via acquired uniparental disomy, in which the HLA locus from the shared chromosome replaces the mismatched HLA locus on the unshared chromosome six. Loss of mismatched HLA occurs more frequently among patients relapsing nine or more months after HLA-haploidentical HCT, and leukemia cells lacking mismatched HLA expression are not expected to respond to infusions of lymphocytes from the original HLA-haploidentical HCT donor.
Donor lymphocyte infusions (DLI) are capable of inducing sustained remissions of hematologic malignancies in relapse after HLA-haploidentical HCT if the mismatched HLA haplotype is maintained. Cases that have lost expression of the mismatched HLA haplotype are candidates for a second haploidentical HCT from a relative who is HLA-mismatched to the original donor.
Our general practice is to reserve DLI for patients with documented relapse of the underlying hematologic malignancy. Patients with overt leukemia receive re-induction chemotherapy followed by 106 CD3+ T cells/kg, one or two days after the completion of chemotherapy. Patients with less aggressive malignancies may be treated with DLI alone. In the absence of graft-versus-host disease (GVHD) and disease response, the dose may be escalated to a maximum of 107 CD3+ T cells/kg.
This approach is supported by the following studies:
In one study, 40 patients in relapse after HLA-haploidentical HCT with PTCy were treated with a total of 52 DLIs . The most commonly used dose for the first DLI was 106 CD3+ T cells per kilogram of recipient ideal body weight. Doses were escalated if patients failed to achieve remission or GVHD. Acute GVHD developed in 10 patients (25 percent), six with grade III to IV, and three patients developed chronic GVHD. Twelve (30 percent) patients achieved a complete response with a median response duration of 11.8 months.
In another report of DLI after HLA-haploidentical HCT with PTCy, 42 patients received a total of 108 DLI, with a median interval from transplant to DLI of 266 days :
●Twenty patients with leukemia in molecular relapse received DLI alone (n = 17) or with azacitidine (n = 3) with a response rate of 45 percent. Acute grade II to III GVHD occurred in 15 percent. Two-year actuarial survival was 43 percent.
●Twelve patients with leukemia in hematologic relapse received chemotherapy followed by DLI (n = 11) or DLI alone (n = 1) with a response rate of 33 percent. Acute grade II to III GVHD occurred in 17 percent. Two-year actuarial survival was 19 percent.
●Ten patients with Hodgkin lymphoma received DLI following one to three courses of chemotherapy with a response rate of 70 percent. Acute grade II to III GVHD occurred in 10 percent. Two-year actuarial survival was 80 percent.
The authors ultimately chose 105 T cells/kg as the initial dose for patients with molecular relapse, and 106 T cells/kg for patients with hematologic relapse. Patients failing to achieve a remission with the initial dose escalated to the highest dose of 107/kg.
In another study, 124 patients treated with HLA-haploidentical HCT according to the GIAC protocol received a total of 168 DLIs to prevent (n = 74), pre-empt (n = 47), or treat (n = 47) relapsed hematologic malignancy . Patients received a median of 4 x 107 CD3+ T cells/kg (range 1.3 to 21.1 x 107/kg) from filgrastim-treated donors, and received concurrent GVHD prophylaxis for a median of two to four weeks with cyclosporine (n = 108), methotrexate (n = 35), or both (n = 9). The cumulative incidences of grades II to IV and III to IV acute GVHD were 53 and 28 percent, respectively. Among patients receiving GVHD prophylaxis for <2 weeks, 2 to 4 weeks, 4 to 6 weeks, and >6 weeks, the cumulative incidences of severe GVHD were 50, 32, 14, and 9 percent, respectively. For all patients, the two-year overall survival, transplant-related mortality, and cumulative incidence of relapse were 47, 34, and 35 percent, respectively.
SUMMARY AND RECOMMENDATIONS
●Human leukocyte antigen (HLA)-haploidentical hematopoietic cell transplantation (HCT) refers to the transplantation of hematopoietic cells from a related donor who shares one HLA haplotype with the recipient but is mismatched for a variable number of HLA molecules on the unshared HLA haplotype (figure 1). (See 'Definitions' above.)
●An appropriate candidate for haploidentical HCT is the patient who lacks an HLA-matched sibling donor and for whom a suitably matched unrelated donor cannot be identified and mobilized in a suitable time frame. There are no published randomized comparisons of haploidentical HCT versus either umbilical cord blood HCT or mismatched unrelated donor HCT; the choice between these alternative graft sources depends on urgency of the transplant and upon institutional preference. (See 'Deciding to use a haploidentical donor' above.)
●The major advantages of haploidentical HCT are the near universal availability of highly motivated donors, rapid availability and relatively low cost of the stem cell source, adequate doses of hematopoietic stem cells for transplantation and immune reconstitution, and the availability of the donor for repeated donations of stem cells or lymphocytes to treat relapse. The major challenge of HLA-haploidentical HCT is the high frequency of host and donor T cells reactive to HLA alloantigens resulting in intense bi-directional alloreactivity and, in the absence of effective prophylactic measures, high incidences of fatal graft rejection or severe or fatal graft-versus-host disease (GVHD). (See 'Advantages and limitations of haploidentical donors' above.)
●Potential haploidentical donors include biological parents, biological children, and full or half siblings. Most patients will have more than one HLA-haploidentical first degree relative willing and able to donate. Criteria to consider in selecting the optimal donor include: donor health, age, and sex, relationship to the patient, HLA mismatch in both the host-versus-graft and graft-versus-host directions, ABO blood type, cytomegalovirus (CMV) serostatus, and possibly non-inherited maternal antigen (NIMA) matching and natural killer cell alloreactivity as predicted by donor killer immunoglobulin receptor (KIR) haplotyping and KIR ligand matching with the recipient (table 1). (See 'Our approach' above.)
●The only absolute contraindications to the use of a specific HLA-haploidentical donor are if the donor is medically or psychologically unfit or if the recipient has anti-donor HLA antibodies of sufficient strength to result in a positive crossmatch result by flow cytometry or by complement-dependent cytotoxicity assay.
●Desensitization procedures to reduce the circulating concentration of donor-specific HLA antibodies may be considered for patients with no other potential donors. (See 'Donor-specific HLA antibodies' above.)
●Haploidentical HCT platforms vary by institutional experience and preference; acceptable strategies include transplantation of a (see 'Selecting a strategy' above):
•T cell-depleted (TCD), "megadose" stem cell graft with no pharmacologic prophylaxis of GVHD (see 'T cell depletion with mega-dose CD34+ cells' above)
•Filgrastim-mobilized, T cell-replete graft with GVHD prophylaxis including anti-thymocyte globulin, calcineurin inhibitor, methotrexate, and mycophenolate mofetil, with or without basiliximab (see 'The GIAC strategy' above)
•Unmanipulated or filgrastim-mobilized, T cell-replete graft with GVHD prophylaxis including high dose post-transplantation cyclophosphamide (PTCy), calcineurin inhibitor, and mycophenolate mofetil (see 'High dose post-transplantation cyclophosphamide' above)
●The breadth and kinetics of immune reconstitution are highly related to the transplantation platform employed, in particular to the extent of donor T cell depletion. TCD "megadose" stem cell grafts have been associated with slow immune reconstitution and infectious mortality approaching 40 percent. Although slightly slower than that seen with matched sibling HCT, immune reconstitution using the GIAC protocol or post-transplantation cyclophosphamide is faster than seen with TCD. (See 'Immune reconstitution after HLA-haploidentical HCT' above.)
●The choice of therapy at the time of relapse depends on whether the mismatched HLA haplotype on the cancer cells is maintained or lost. Donor lymphocyte infusions (starting dose of 1 million CD3+ T cells/kg of recipient weight) are capable of inducing sustained remissions of hematologic malignancies in relapse after HLA-haploidentical HCT if the mismatched HLA haplotype is maintained. Cases that have lost expression of the mismatched HLA haplotype are candidates for a second haploidentical HCT from a relative who is HLA-mismatched to the original donor. (See 'Management of relapse after HLA-haploidentical HCT' above.)