Childhood Hematopoietic Cell Transplantation (PDQ®)—Health Professional Version - National Cancer Institute

Childhood Hematopoietic Cell Transplantation (PDQ®)–Health Professional Version

General Information About Hematopoietic Cell Transplantation (HCT)

Rationale for HCT

Blood and marrow transplantation, or HCT, is a procedure that involves infusion of hematopoietic stem cells (hematopoietic progenitor cells) to reconstitute the hematopoietic system of a patient. The infusion of hematopoietic stem cells generally follows a preparative regimen consisting of agents designed to do the following:

Create marrow space.
Suppress the patient's immune system to prevent rejection.
Eradicate malignant cells in patients with cancer.

HCT is currently used in the following three clinical scenarios:

Treatment of malignancies.
Replacement or modulation of an absent or poorly functioning hematopoietic or immune system.
Treatment of genetic diseases in which an insufficient expression of the affected gene product can be partially or completely overcome by circulating hematopoietic cells transplanted from a donor with normal gene expression.

Autologous Versus Allogeneic HCT

The two major transplant approaches currently in use are the following:

Autologous (using the patient's own hematopoietic stem cells).
Allogeneic (using related- or unrelated-donor hematopoietic stem cells).

An autologous transplant treats cancer by exposing patients to high-dose therapy with the intent of overcoming chemotherapy resistance in tumor cells, followed by infusion of the patient’s previously stored hematopoietic stem cells. The transplant can be performed in a single procedure or tandem sequential procedures. For autologous transplants to work for malignancies, the following must apply:

A dose-intensified chemotherapy/radiation therapy regimen with hematopoietic stem cell support is used to achieve a significantly higher cell kill than could be achieved without the use of hematopoietic stem cell support. This may include increased tumor kill in areas where standard-dose chemotherapy has less penetration (central nervous system).
Meaningful percentages of cure or long-term remission from the disease must occur without significant nonhematopoietic toxicities that would otherwise limit the therapeutic benefit achieved.

Autologous transplants have also been used to reset the immune system in patients with severe autoimmune disorders.

Current pediatric indications for autologous transplant include patients with certain lymphomas, neuroblastoma, and brain tumors. Autologous transplant techniques are also being used to enable engraftment of genetically modified autologous hematopoietic stem cell progenitors to correct/ameliorate inherited disorders (e.g., immunodeficiencies, metabolic disorders, and hemoglobinopathies).

Allogeneic transplant approaches to cancer treatment also may involve high-dose therapy, but because of immunologic differences between the donor and recipient, an additional graft-versus-tumor (GVT) or graft-versus-leukemia (GVL) treatment effect can occur. Although autologous approaches are associated with less short-term mortality, many malignancies are resistant to even high doses of chemotherapy and/or involve the bone marrow, thus requiring allogeneic approaches for optimal outcome.

Determining When HCT Is Indicated: Comparison of HCT and Chemotherapy Outcomes

Because the outcomes using chemotherapy and HCT treatments have been changing over time, regular comparisons between these approaches should be performed to continually redefine optimal therapy for a given patient. For some diseases, randomized trials or intent-to-treat using an HLA-matched sibling donor have established the benefit of HCT by direct comparison.[1,2] However, for very high-risk patients such as those with early relapse of acute lymphoblastic leukemia (ALL), randomized trials have not been feasible because of investigator bias.[3,4]

In general, HCT typically offers benefit only to children at high risk of relapse with standard chemotherapy approaches. Accordingly, treatment schemas that accurately identify these high-risk patients and offer HCT if reasonably HLA-matched donors are available have come to be the preferred approach for many diseases.[5] Less well-established, higher-risk approaches to HCT are generally reserved for only the very highest-risk patients. However, higher-risk approaches such as haploidentical transplantation are becoming safer and more efficacious and are increasingly being used interchangeably with fully matched allogeneic approaches.[6-9] (Refer to the Haploidentical HCT section of this summary for more information.)

When comparisons of similar patients treated with HCT or chemotherapy are made in the setting where randomized or intent-to-treat studies are not feasible, the following issues should be considered:

Remission/disease status: Comparisons between HCT and chemotherapy should include only patients who obtain remission, preferably after similar approaches to salvage therapy, because patients failing to obtain remission do very poorly with any therapy.[10]
To account for time-to-transplant bias, the chemotherapy comparator arm should include only patients who maintained remission until the median time to HCT. The HCT comparator arm should also include only patients who achieved the initial remission mentioned above and maintained that remission until the time of HCT.[10]
High-risk and intermediate-risk patient groups should not be combined because a benefit for HCT in the high-risk group can be masked when outcomes are similar to those achieved in the intermediate-risk group.[10]
Therapy approaches used for comparison: Comparisons should be made with the best or most commonly used chemotherapy and HCT approaches utilized during the time frame under study.
HCT approach: HCT approaches that are very high risk or have documented lower rates of survival should not be combined for analysis with standard-risk HCT approaches.
Criteria for relapse: Risk factors for relapse should be carefully defined, and analysis should be based on the most current knowledge of risk.
Selection bias: Attempts should be made to understand and eliminate or correct for selection bias. Examples include the following:
- Higher-risk patients preferentially undergoing HCT (i.e., patients who take several rounds to achieve remission or who relapse after obtaining remission and go back into a subsequent remission before HCT).
- Sicker patients deferred from HCT because of comorbidities.
- Related to the time-to-transplant bias noted above, patients who undergo HCT after relapse or recurrence are a subset of all patients with a disease recurrence and will be selected from those who are able to obtain a remission and remain healthy enough to undergo HCT.
- Patient or parent refusal.
- Lack of or inability to obtain insurance approval for HCT.
- Lack of access to HCT because of distance or inability to travel.

One source of bias difficult to control for or detect is physician bias for or against HCT. The effects of access to HCT and therapeutic bias on outcomes of pediatric malignancies for which HCT may be indicated have been poorly studied to date.

References

Matthay KK, Villablanca JG, Seeger RC, et al.: Treatment of high-risk neuroblastoma with intensive chemotherapy, radiotherapy, autologous bone marrow transplantation, and 13-cis-retinoic acid. Children's Cancer Group. N Engl J Med 341 (16): 1165-73, 1999. [PUBMED Abstract]
Woods WG, Neudorf S, Gold S, et al.: A comparison of allogeneic bone marrow transplantation, autologous bone marrow transplantation, and aggressive chemotherapy in children with acute myeloid leukemia in remission. Blood 97 (1): 56-62, 2001. [PUBMED Abstract]
Lawson SE, Harrison G, Richards S, et al.: The UK experience in treating relapsed childhood acute lymphoblastic leukaemia: a report on the medical research council UKALLR1 study. Br J Haematol 108 (3): 531-43, 2000. [PUBMED Abstract]
Gaynon PS, Harris RE, Altman AJ, et al.: Bone marrow transplantation versus prolonged intensive chemotherapy for children with acute lymphoblastic leukemia and an initial bone marrow relapse within 12 months of the completion of primary therapy: Children's Oncology Group study CCG-1941. J Clin Oncol 24 (19): 3150-6, 2006. [PUBMED Abstract]
Schrauder A, von Stackelberg A, Schrappe M, et al.: Allogeneic hematopoietic SCT in children with ALL: current concepts of ongoing prospective SCT trials. Bone Marrow Transplant 41 (Suppl 2): S71-4, 2008. [PUBMED Abstract]
Bertaina A, Merli P, Rutella S, et al.: HLA-haploidentical stem cell transplantation after removal of αβ+ T and B cells in children with nonmalignant disorders. Blood 124 (5): 822-6, 2014. [PUBMED Abstract]
Handgretinger R, Chen X, Pfeiffer M, et al.: Feasibility and outcome of reduced-intensity conditioning in haploidentical transplantation. Ann N Y Acad Sci 1106: 279-89, 2007. [PUBMED Abstract]
Huang XJ, Liu DH, Liu KY, et al.: Haploidentical hematopoietic stem cell transplantation without in vitro T-cell depletion for the treatment of hematological malignancies. Bone Marrow Transplant 38 (4): 291-7, 2006. [PUBMED Abstract]
Luznik L, Fuchs EJ: High-dose, post-transplantation cyclophosphamide to promote graft-host tolerance after allogeneic hematopoietic stem cell transplantation. Immunol Res 47 (1-3): 65-77, 2010. [PUBMED Abstract]
Pulsipher MA, Peters C, Pui CH: High-risk pediatric acute lymphoblastic leukemia: to transplant or not to transplant? Biol Blood Marrow Transplant 17 (1 Suppl): S137-48, 2011. [PUBMED Abstract]

Autologous HCT

Collection and Storage of Autologous Hematopoietic Stem Cells

Autologous procedures require collection of growth-factor–mobilized peripheral blood stem cells (PBSCs) from patients by the process of leukapheresis. Bone marrow can be used for autologous transplants, but PBSCs have been shown to lead to quicker blood count recovery and cause less toxicity.

Patients being considered for autologous hematopoietic cell transplantation (HCT) are generally given chemotherapy to determine tumor responsiveness and minimize the risk of tumor contamination in their bone marrow. After a number of rounds of chemotherapy, patients undergo the leukapheresis procedure, either as their blood counts recover from chemotherapy or during a break between chemotherapy treatments. Growth factors such as granulocyte colony-stimulating factor (G-CSF) are used to increase the number of circulating stem and progenitor cells (CD34+ cells). Collection centers monitor the CD34-positive number in the patient and product each day to determine the best time to begin collection and when collection is complete. Patients with poorly mobilized CD34-positive cells can often have their cells successfully collected using alternative mobilization approaches (e.g., plerixafor).[1] The collected PBSCs are cryopreserved for later use. After completion of an intensive preparative regimen using high-dose chemotherapy, which varies according to the tumor type, the PBSCs are administered back into the patient at the time of transplant.

General Indications for Autologous Procedures/Preparative Regimens/Tumor Purging

In pediatrics, the most common autologous transplant indications are the following:

High-risk neuroblastoma. (Refer to the PDQ summary on Neuroblastoma Treatment for more information.)
Relapsed Hodgkin lymphoma and non-Hodgkin lymphoma. (Refer to the PDQ summaries on Childhood Hodgkin Lymphoma Treatment and Childhood Non-Hodgkin Lymphoma Treatment for more information.)
High-risk and relapsed brain tumors.
Relapsed or resistant germ cell tumors. (Refer to the PDQ summaries on Childhood Central Nervous System Germ Cell Tumors Treatment and Childhood Extracranial Germ Cell Tumors Treatment for more information.)

Tumor-specific regimens are described in disease-specific PDQ treatment summaries.

The tumor-specific activity and intensity of agents used for autologous regimens have been shown to be important in improving survival. This is not the case for allogeneic procedures.

One concern with autologous approaches for these and other tumor types has been the contamination of the collected stem cell product by persistent tumor cells. Although many techniques have been developed to remove or purge tumor cells from products, most studies have shown no benefit to tumor purging.[2]

References

Patel B, Pearson H, Zacharoulis S: Mobilisation of haematopoietic stem cells in paediatric patients, prior to autologous transplantation following administration of plerixafor and G-CSF. Pediatr Blood Cancer 62 (8): 1477-80, 2015. [PUBMED Abstract]
Kreissman SG, Seeger RC, Matthay KK, et al.: Purged versus non-purged peripheral blood stem-cell transplantation for high-risk neuroblastoma (COG A3973): a randomised phase 3 trial. Lancet Oncol 14 (10): 999-1008, 2013. [PUBMED Abstract]

Allogeneic HCT

Improved Outcomes After Allogeneic Transplantation

Over the past one to two decades, significant advances have led to improved outcomes after allogeneic hematopoietic cell transplantation (HCT).[1-3] The most significant improvements in survival occurred in unrelated and alternative donor procedures.[4-6] Possible explanations for these improvements in survival include improved patient selection, better supportive care, refined treatment regimens, improved approaches specific to stem cell sources, and better HLA typing. All of these factors may have contributed to better outcomes; however, the section below focuses on modifiable aspects of HCT (i.e., optimization of HLA typing and selection of stem cell sources).

HLA Matching and Hematopoietic Stem Cell Sources

HLA overview

Appropriate matching between donor and recipient HLA in the major histocompatibility complex located on chromosome 6 is essential to successful allogeneic HCT (refer to Figure 1 and Tables 1 and 2).

Human lymphocyte antigen (HLA) complex; drawing shows the long and short arms of human chromosome 6 with amplification of the HLA region, including the class I A, B, and C alleles, and the class II DP, DQ, and DR alleles. — Figure 1. HLA Complex. Human chromosome 6 with amplification of the HLA region. The locations of specific HLA loci for the class I B, C, and A alleles and the class II DP, DQ, and DR alleles are shown.

HLA class I (A, B, C, etc.) and class II (DRB1, DQB1, DPB1, etc.) alleles are highly polymorphic; therefore, finding appropriately matched unrelated donors is a challenge for some patients, especially those of certain racial groups (e.g., African Americans and Hispanics).[7,8] Full siblings of cancer patients have a 25% chance of being HLA matched.

Early serologic techniques of HLA assessment defined a number of HLA antigens, but more precise DNA methodologies have shown HLA allele-level mismatches in up to 40% of serologic HLA antigen matches. These differences are clinically relevant because the use of donors with allele-level mismatches affects survival and rates of graft-versus-host disease (GVHD) to a degree similar to that in patients with antigen-level mismatches.[9] Because of this, DNA-based allele-level HLA typing is standard when unrelated donors are being chosen.

Table 1. Level of HLA Typing Currently Used for Different Hematopoietic Stem Cell Sourcesa,b,c

ENLARGE

	Class I Antigens			Class II Antigens
BM = bone marrow; PBSCs = peripheral blood stem cells.
aHLA antigen: A serologically defined, low-resolution method of defining an HLA protein. Differs from allele-level typing half of the time. Designated by the first two numbers (i.e., HLA B 35:01—antigen is HLA B 35).
bHLA allele: A higher-resolution method of defining unique HLA proteins by typing their gene through sequencing or other DNA-based methods that detect unique differences. Designated by at least four numbers (i.e., HLA B 35:01—35 is the antigen and 01 is the allele).
cConsensus recommendations for HLA typing, including extended class II typing of mismatched donors, have been published by the National Cancer Institute/National Heart, Lung, and Blood Institute–sponsored Blood and Marrow Transplant Clinical Trials Network.[10]
dSiblings need confirmation that they have fully matched haplotypes with no crossovers in the A to DRB1 region. If parental typing is performed and haplotypes are established, antigen-level typing of class I is adequate. With no parental haplotypes, allele-level typing of eight alleles is recommended.
eParent, cousin, etc., with a phenotypic match or near-complete HLA match.
Stem Cell Source	HLA A	HLA B	HLA C	HLA DRB1	HLA DQB1;HLA DPB1; HLA DR3,4,5
Matched siblingdBM/PBSCs	Antigen or allele	Antigen or allele	Optional	Allele
Mismatched sibling/other related-donoreBM/PBSCs	Allele	Allele	Allele	Allele	Recommended, if mismatches are present
Unrelated-donor BM/PBSCs	Allele	Allele	Allele	Allele	Recommended, if mismatches are present
Unrelated-donor cord blood	Antigen (allele recommended)	Antigen (allele recommended)	Allele recommended	Allele

Table 2. Definitions of the Numbers Describing HLA Antigens and Alleles Matching

If These HLA Antigens and Alleles Match:	Then the Donor Is Considered to be This Type of Match:
A, B, and DRB1	6/6
A, B, C, and DRB1	8/8
A, B, C, DRB1, and DQB1	10/10
A, B, C, DRB1, DQB1, and DPB1	12/12

HLA matching considerations for sibling and related donors

The most commonly used related donor is a sibling from the same parents who, at a minimum, is HLA matched for HLA A, HLA B, and HLA DRB1 at the antigen level. Given the distance between HLA A and HLA DRB1 on chromosome 6, there is approximately a 1% possibility of a crossover event occurring in a possible sibling match. Because a crossover event could involve the HLA C antigen and because parents may share HLA antigens that actually differ at the allele level, many centers perform allele-level typing of possible sibling donors at all of the key HLA antigens (HLA A, B, C, and DRB1). Any related donor that is not a full sibling should have full HLA typing because similar haplotypes from different parents could differ at the allele level.

Although single-antigen mismatched related donors (5/6 antigen matched) were used interchangeably with matched sibling donors in some studies, a large Center for International Blood and Marrow Transplant Research (CIBMTR) study in pediatric HCT recipients showed that the use of 5/6 antigen-matched related donors resulted in rates of GVHD and overall survival (OS) equivalent to rates in 8/8-allele-level-matched unrelated donors and slightly inferior survival than in fully matched siblings.[11] Any siblings with single mismatches should have extended typing to ensure that if the mismatch is caused by a crossover, it only occurs with one antigen. If clinicians choose siblings with multiple antigen mismatches as donors, haploidentical approaches may be warranted.

HLA matching considerations for unrelated donors

Optimal outcomes are achieved in unrelated allogeneic marrow transplantation when the pairs of antigens at HLA A, B, C, and DRB1 are matched between the donor and the recipient at the allele level (termed an 8/8 match) (refer to Table 2).[12] A single antigen/allele mismatch at any of these antigens (7/8 match) lowers the probability of survival between 5% and 10%, with a similar increase in the amount of significant (grades III–IV) acute GVHD.[12] Of these four antigen pairs, different reports have shown HLA A, C, and DRB1 mismatches to potentially be more highly associated with mortality than the other antigens,[9,12,13] but the differences in outcome are small and inconsistent, making it very difficult to conclude that one can pick a more favorable mismatch by choosing one type of antigen mismatch over another. Many groups are attempting to define specific antigens or pairs of antigens that are associated with either good or poor outcomes. For example, a specific HLA C mismatch (HLA-C*03:03/03:04) has outcomes similar to a match; therefore, selection of this mismatch is desirable in an otherwise matched donor/pair combination.[14]

It is well understood that class II antigen DRB1 mismatches increase GVHD incidence and worsen survival.[13] Subsequent data have also shown that multiple mismatches of DQB1, DPB1, and DR3,4,5 lead to worse outcomes in the setting of less than 8/8 matches.[15] DPB1 mismatches have been extensively studied and classified as permissive or nonpermissive based on T-cell epitope matching. Patients with 10/10 matches and nonpermissive DPB1 mismatches have more transplant-related mortality but have survival rates similar to those with DPB1 matches or permissive matches. Those with 9/10 matches who have nonpermissive DPB1 mismatches had worse survival than did those with permissive mismatches or DPB1 matches.[16-18]

With these findings in mind, although a 7/8- or 8/8-matched unrelated donor can be used routinely, centers may be able to further improve outcomes by the following:

Extended typing of DQB1, DPB1, and DR3,4,5, especially in the context of a less-than-8/8-matched donor.[16-18]
Extended HLA testing to select appropriate donors in the context of HLA-sensitized patients to avoid the potential risk of graft failure.[19,20] HLA sensitization is detected by testing for the presence of specific anti-HLA antibodies and avoidance of donors who have any HLA antigens associated with the antibodies present in the recipient.
Use of younger donors.[10]
Matching cytomegalovirus (CMV)-positive or CMV-negative recipients with positive or negative donors.[21]
Use of blood type–compatible unrelated donors.[10]

Chart showing HLA allele duplication and type of match between donor and recipient: an allele match (0201 and 0401 for both donor and recipient); a mismatch (0201 for both donor and recipient and 0201 for donor, 0401 for recipient) shown by an arrow pointing in a direction that promotes GVHD (GVH-O); a mismatch (0201 for both donor and recipient and 0401 for donor, 0201 for recipient) shown by an arrow pointing in a direction that promotes rejection (R-O); and a bidirectional mismatch (0201 for donor, 0301 for recipient, and 0401 for both donor and recipient) shown by arrows pointing in two directions, a direction that promotes rejection (R-O) and a direction that promotes GVHD (GVH-O). — Figure 2. HLA allele duplication in a donor or recipient results in a *half* match and a mismatch that will either occur in a direction that promotes GVHD (GVH-O) or a direction that promotes rejection (R-O).

If a donor or recipient has a duplication of one of their HLA alleles, they will have a halfmatch and a mismatch only in one direction. Figure 2 illustrates that these mismatches will occur in either a direction that promotes GVHD (GVH-O) or a direction that promotes rejection (R-O). When 8/8-matched unrelated donors are compared with 7/8 donors mismatched in the GVH-O direction, 7/8 mismatched in the R-O direction, or 7/8 mismatched in both directions, the mismatch in the R-O direction leads to rates of grades III and IV acute GVHD similar to rates in the 8/8 matched and better than in the other two combinations. The 7/8 mismatched in only the R-O direction is preferred over GVH-O and bidirectional mismatches.[22] It is important to note that this observation in unrelated donors differs from observations in cord blood recipients, outlined below.

HLA matching and cell dose considerations for unrelated cord blood HCT

Another commonly used hematopoietic stem cell source is that of unrelated umbilical cord blood, which is harvested from donor placentas moments after birth. The cord blood is processed, HLA typed, cryopreserved, and banked.

Unrelated cord blood transplantation has been successful with less stringent HLA matching requirements compared with standard related or unrelated donors, probably because of limited antigen exposure experienced in utero and different immunological composition. Cord blood matching has traditionally been performed at an intermediate level for HLA A and B and at an allele level (high resolution) for DRB1. This means that until just recently, attempted matching of only six antigens has been necessary to choose units for transplantation.

Although better outcomes occur when 6/6 or 5/6 HLA-matched units are used,[23] successful HCT has occurred even with 4/6 or less HLA-matched units in many patients. In a large CIBMTR/Eurocord study, better matching at the allele level using eight antigens (matching for HLA A, B, C, and DRB1) resulted in less transplant-related mortality and improved survival. Best outcome was noted with 8/8 allele matching versus 4/8 to 7/8 matches, with poor survival in patients with five or more allele mismatches. Patients receiving 8/8-matched cord blood did not require higher cell doses for better outcomes; however, those with one to three allele mismatches had less transplant-related mortality with total nucleated cell counts higher than 3 × 107/kg, and those with four allele mismatches required a total nucleated cell count higher than 5 × 107/kg to decrease transplant-related mortality.[24] This observation was noted to be especially important in cord blood transplantation for nonmalignant disorders, where any mismatching below 7/8 alleles led to inferior survival.[25] Many centers will type additional alleles and use the best match possible, but the impact of DQB1, DPB1, and DR3,4,5 mismatches has not been studied in detail.

As in unrelated peripheral blood stem cells (PBSCs) or bone marrow donors, extended HLA testing can support the selection of appropriate cord blood units in HLA-sensitized patients to avoid the potential risk of graft failure.[26,27] Evidence also suggests that selecting a mismatched cord blood unit, where the mismatch involves a noninherited maternal antigen, may improve survival.[28,29]

As with unrelated donors, individuals can occasionally have duplicate HLA antigens (e.g., the HLA A antigen is 01 on both chromosomes). When this occurs in a donor product and the antigen is matched to one of the recipient antigens, the recipient immune response will see the donor antigens as matched (matched, in the rejection direction), but the donor immune response will see a mismatch in the recipient (mismatched in the GVHD direction). This variation of partial mismatching has been shown to be important in cord blood transplant outcomes. Mismatches that are only in the GVHD direction (GVH-O) lead to lower transplant-related mortality and overall mortality than in those with recipient direction only (R-O) mismatches.[30] R-O mismatches have outcomes similar to those of bidirectional mismatches.[31] Although these studies suggest that using unidirectional mismatching as a criteria for cord blood selection may be beneficial, a Eurocord–European Society for Blood and Marrow Transplantation analysis disputes the value of this type of mismatching.[32]

Two aspects of umbilical cord blood HCT have made the practice more widely applicable. First, because a successful procedure can occur with multiple HLA mismatches, more than 95% of patients from a wide variety of ethnicities are able to find at least a 4/6-matched cord blood unit.[7,33] Second, as mentioned above, adequate cell dose (minimum 2–3 × 107total nucleated cells/kg and 1.7 × 105 CD34+ cells/kg) has been shown to be associated with improved survival.[34,35] Total nucleated cells are generally used to judge units because techniques to measure CD34-positive doses have not been standardized. Because even large single umbilical cord blood units are only able to supply these minimum doses to recipients weighing up to 40 kg to 50 kg, early umbilical cord blood HCT focused mainly on smaller children. Later studies showed that barriers of this size could be overcome by using two umbilical cord blood units, as long as each of the units is at least a 4/6 HLA match with the recipient; because two cords provide higher cell doses, umbilical cord blood transplantation is now used widely for larger children and adults.[36]

If a single unit provides an adequate cell dose, there may be disadvantages to adding a second unit.[37][Level of evidence: 1iiA] Two randomized trials showed that in children who had adequately sized single units, the addition of a second unit did not alter relapse, transplant-related mortality, or survival rates, but was associated with higher rates of extensive chronic GVHD.[37,38]

Investigators have shown that by using combinations of cytokines and other compounds to expand cord blood for a period of time before infusion, engraftment of cord blood cells can occur more rapidly than after standard approaches.[39-42] Although some studies using multiple units or split units have shown that expanded units will engraft early and then give way to nonexpanded units for long-term reconstitution,[43] other studies are showing persistence of expanded cells, implying preservation of stem cells through the expansion process.[41,42] A number of these approaches are currently under investigation; their effect on efficacy and survival of children using cord blood as a stem cell source has yet to be established and none are approved by the U.S. Food and Drug Administration (FDA).

Comparison of stem cell products

Currently, the following three stem cell products are used from both related and unrelated donors:

Bone marrow.
PBSCs.
Cord blood.

In addition, bone marrow or PBSCs can be T-cell depleted by several methods, and the resultant stem cell product has very different properties. Finally, partially HLA-matched (half or more antigens [haploidentical]) related bone marrow or PBSCs can be used after in vitro or in vivo T-cell depletion, and this product also behaves differently from other stem cell products. A comparison of stem cell products is presented in Table 3.

Table 3. Comparison of Hematopoietic Stem Cell Products

ENLARGE

	PBSCs	BM	Cord Blood	T-cell–Depleted BM/PBSCs	Haploidentical T-cell–Depleted BM/PBSCs
BM = bone marrow; EBV-LPD = Epstein-Barr virus–associated lymphoproliferative disorder; GVHD = graft-versus-host disease; HCT = hematopoietic cell transplantation; PBSCs = peripheral blood stem cells.
aAssuming no development of GVHD. If patients develop GVHD, immune reconstitution is delayed until resolution of the GVHD and discontinuation of immune suppression.
bIf a haploidentical donor is used, longer times to immune reconstitution may occur.
T-cell content	High	Moderate	Low	Very low	Very low
CD34+ content	Moderate–high	Moderate	Low (but higher potency)	Moderate–high	Moderate–high
Time to neutrophil recovery	Rapid: median, 16 d (11–29 d) [44]	Moderate: median, 21 d (12–35 d) [44]	Slower: median, 23 d (11–133 d) [38]	Rapid: median, 16 d (9–40 d) [45]	Rapid: median, 13 d (10–20 d) [46]
Early post-HCT risk of infections, EBV-LPD	Low–moderate	Moderate	High	Very High	Very High
Risk of graft rejection	Low	Low–moderate	Moderate–high	Moderate–high	Moderate–high
Time to immune reconstitutiona	Rapid (6–12 mo)	Moderate (6–18 mo)	Slow (6–24 mo)	Slow (6–24 mo)	Slow (9–24 mo)b
Risk of acute GVHD	Moderate	Moderate	Moderate	Low	Low
Risk of chronic GVHD	High	Moderate	Low	Low	Low

The main differences between the products are associated with the numbers of T cells and CD34-positive progenitor cells present; very high levels of T cells are present in PBSCs, intermediate numbers in bone marrow, and low numbers in cord blood and T-cell–depleted products. Patients receiving T-cell–depleted products or cord blood generally have slower hematopoietic recovery, increased risk of infection, late immune reconstitution, higher risks of nonengraftment, and increased risk of Epstein-Barr virus (EBV)–associated lymphoproliferative disorder. This is countered by lower rates of GVHD and an ability to offer transplantation to patients where full HLA matching is not available. Higher doses of T cells and other cells in PBSCs result in rapid neutrophil recovery and immune reconstitution, but also increase rates of chronic GVHD.

Only a few studies directly compared outcomes of different stem cell sources/products in pediatric patients.

Evidence (comparison of outcomes of stem cell sources/products in children):

A retrospective registry study of pediatric patients who underwent HCT for acute leukemia compared those who received related-donor bone marrow with those who received related-donor PBSCs.[47]
- Although the bone marrow and PBSC recipient cohorts differed some in their risk profiles, after statistical correction, increased risk of GVHD and transplant-related mortality associated with PBSCs led to poorer survival in the PBSC group.
A retrospective study of Japanese children with acute leukemia compared 90 children who received PBSCs with 571 children who received bone marrow.[48]
- The study confirmed higher transplant-related mortality caused by GVHD and inferior survival among the children who received PBSCs.

These reports, combined with a lack of prospective studies comparing bone marrow and PBSCs, have led most pediatric transplant protocols to prefer bone marrow over PBSCs from related donors.

For those requiring unrelated donors, a large Blood and Marrow Transplant Clinical Trials Network (BMT CTN) trial that included a number of pediatric patients randomly assigned enrollees to receive either bone marrow or PBSCs. This trial demonstrated that OS was identical using either source, but rates of chronic GVHD were significantly higher in the PBSC arm, with a small increase in rejection in the bone marrow arm.[49] Rejections were rare in pediatric patients, and there was an insufficient number of patients to draw specific conclusions about rejection risk in children who received bone marrow.

Published studies comparing unrelated cord blood and bone marrow have been retrospective, with weaknesses inherent in such analyses.

Evidence (comparison of unrelated cord blood versus bone marrow outcomes):

In one study, pediatric patients with acute lymphoblastic leukemia (ALL) who underwent HCT and received 8/8-HLA-allele–matched unrelated-donor bone marrow were compared with those who received unrelated cord blood.[23]
- The analysis showed that the best survival occurred in recipients of 6/6 HLA-matched cord blood; survival after 8/8 HLA-matched unrelated bone marrow was slightly less and was statistically identical to survival for patients receiving 5/6 and 4/6 HLA-matched cord blood units.
In a second study from a single center consisting of mostly adult patients with acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), and ALL, outcomes for cord blood recipients were compared with outcomes for recipients of matched and mismatched unrelated-donor bone marrow/PBSCs.[50]
- Better survival because of less relapse was noted in cord blood recipients, mainly resulting from superior survival in patients with minimal residual disease (MRD) present just before transplant.
- No difference was seen in relapse and survival between patients with pre-HCT MRD and patients without pre-HCT MRD if they received cord blood.
- This result is controversial because it contradicts many other studies that showed that the presence of pre-HCT MRD in cord blood recipients led to increased relapse and inferior survival.[51-54]

On the basis of these studies, most transplant centers consider matched sibling bone marrow to be the preferred stem cell source/product. If a sibling donor is not available, fully matched unrelated-donor bone marrow or PBSCs or HLA-matched (4/6 to 6/6) cord blood leads to similar survival rates. Although adult studies of T-cell–depleted unrelated bone marrow or PBSCs have shown outcomes similar to non–T-cell–depleted approaches, large pediatric trials or retrospective studies comparing T-cell–depleted matched or haploidentical bone marrow or PBSCs have not been conducted.

Haploidentical HCT

Early HCT studies demonstrated progressively higher percentages of patients experiencing severe GVHD and lower survival rates as the number of donor/recipient HLA mismatches increased.[55] Other studies showed that even with very high numbers of donors in unrelated-donor registries, patients with rare HLA haplotypes and patients with certain ethnic backgrounds (e.g., Hispanic, African American, Asian-Pacific Islander, etc.) have a low chance of achieving desired levels of HLA matching (7/8 or 8/8 match at the allele level).[8]

To allow access to HCT for patients without fully HLA-matched donor options, investigators have developed techniques allowing the use of siblings, parents, or other relatives who share only a single haplotype of the HLA complex with the patient and are thus halfmatches. Most approaches developed to date rely on intense T-cell depletion of the product before infusion into the patient. The main challenge associated with this approach is intense immune suppression with delayed immune recovery, which can result in lethal infections,[56] increased risk of EBV-associated lymphoproliferative disorder, and high rates of relapse.[57] This has generally led to inferior survival compared with matched HCT and has resulted in the procedure being practiced mainly at larger academic centers with a specific research focus on studying and developing this approach.

Current approaches, however, are rapidly evolving, resulting in improved outcomes, with some pediatric groups reporting survival similar to that of standard approaches.[58,59] These approaches include the following:

Newer techniques of T-cell depletion and add-back of specific cell populations (e.g., CD3 or alpha-beta CD3/CD19-negative selection) may decrease transplant-related mortality.[60]; [61,62][Level of evidence: 3iiiDii]
Reduced toxicity regimens have led to improved survival.
Better supportive care has decreased the chance of morbidity from infection or EBV-associated lymphoproliferative disorder.[63]
Some patient-donor combinations that have specific killer immunoglobulin-like receptor mismatches have shown decreased likelihood of relapse (refer to the Role of killer immunoglobulin-like receptor mismatching in HCT section of this summary for more information).
Certain techniques, such as using combinations of granulocyte colony-stimulating factor (G-CSF)–primed bone marrow and PBSCs with posttransplant antibody–based T-cell depletion [64] or post-HCT cyclophosphamide (chemotherapeutic T-cell depletion),[65]; [66][Level of evidence: 3iiiA] have made these procedures more accessible to centers because the expensive and complicated processing necessary for traditional T-cell depletion are not used.

Reported rates of survival using many different types of haploidentical approaches varies between 25% and 80%, depending on the technique used and the risk of the patient undergoing the procedure.[57,58,64,65]; [66][Level of evidence: 3iiiA] Whether haploidentical approaches are superior to cord blood or other stem cell sources for a given patient group has not been determined because comparative studies have yet to be performed.[57]

Even more than with other stem cell sources, patients undergoing haploidentical procedures can develop anti-HLA antibodies that, if directed against nonshared haploidentical antigens, can greatly increase the risks of rejection. Clinicians should choose donors with HLA types against whom the recipient does not have an antibody present, if possible. Guidelines on how to best approach this issue have been published.[67]

Other donor characteristics associated with outcome

Although HLA matching has consistently been the most important factor associated with improved survival in nonhaploidentical allogeneic HCT, a number of other characteristics of the donor have been shown in studies to affect key outcomes. Higher cell dose from the donor (refer to the HLA matching and cell dose considerations for unrelated cord blood HCT section of this summary for more information) has also been shown to be important when related, unrelated, or haploidentical bone marrow or PBSC donors are used.[68,69] The effects of donor age, blood type, CMV status, sex, and parity of female donors have also been studied.

Ideally, after HLA matching, transplant centers should select donors based on the following characteristics:

Donor age. The youngest donor available is preferred.[70,71]
CMV status of the recipient. CMV-negative donor matched to CMV-negative recipient and CMV-positive donor matched to CMV-positive recipient are preferred.[72]
Matched donor blood type.[73-75]
Donor sex and parity of female donors. Male or nonparous female donors are preferred over parous female donors.[71,76]

Rarely can a donor/recipient pair fit perfectly into this algorithm, and determining which of these characteristics should be chosen over others has been controversial. A CIBMTR study involving 6,349 patients who underwent transplant for hematological malignancies from 1988 to 2006, with a confirmation cohort of 4,690 patients who underwent transplant between 2007 and 2011, tested the effect of donor characteristics while adjusting for disease risk and other key transplant characteristics.[71,77]

The earlier data set showed that in addition to HLA mismatching, older donor age and major or minor ABO blood-type mismatching increased overall mortality; parous female graft recipients experienced lower rates of relapse; recipients of younger donor grafts had lower rates of acute GVHD; and recipients of parous female grafts had higher rates of chronic GVHD. Recipient CMV status was more important than donor CMV status (recipients who are CMV-positive are at higher risk of mortality independent of the donor CMV status), although a CMV-negative donor to a CMV-negative recipient combination improves survival.[71]
The more recent confirmation cohort was tested by a multivariate analysis for independent predictors of survival in an EBMT study. Older donor age was confirmed to be independently associated with worse OS; every 10 years of donor age increased the risk of mortality by 5.5%. HLA matching continued to have the most important effect on survival; ABO mismatching was not confirmed to have a continuing effect.[77]

Thus, after HLA matching, donor age is likely the most important second factor to optimize, unless the recipient is CMV negative, at which point finding a CMV-negative donor would take priority.

A number of studies have attempted to identify characteristics of the best donors for haploidentical procedures. As with conventional bone marrow transplantation, use of younger donors appears to be beneficial, but data regarding donor sex are inconclusive. Studies involving intense T-cell depletion have noted better outcomes using maternal donors,[78] but studies using posttransplant cyclophosphamide or intense immune suppression seem to favor male donors.[79,80] Further study is needed to clarify this important issue. One large comparison of haploidentical donors showed an effect of ABO incompatibility on engraftment (risk of rejection doubling from 6% to 12%, ABO match vs. ABO major mismatch), and patients receiving bidirectionally mismatched donors had a 2.4-fold increase in grades II to IV acute GVHD.[81]