Chapter 23: Hematopoietic Cell Transplantation

The successful clinical application of hematopoietic cell transplantation (HCT) required a century of key developmental discoveries (Table 23–1). Between 1868 and 1906, European and American investigators established that marrow cells were the source of blood cell production. In 1939, the first documented human marrow transplant was performed in a patient with gold-induced marrow aplasia.¹ The patient was infused intravenously with marrow from a brother with an identical ABO blood type. The transplant was not successful, and the patient died 5 days after the marrow infusion.

In 1922, a Danish investigator modified radiation injury in guinea pigs by shielding their femora against radiation, preventing the typical radiation-induced thrombocytopenia and hemorrhage.² This work went essentially unnoticed for more than 2 decades. The period of 1949 to 1954 was marked by a political climate concerned with the threat of continued atomic warfare, which stimulated support for experiments studying the effects of irradiation and led to the development of the field of organ and marrow transplantation. Jacobson and colleagues found that mice could survive an otherwise lethal irradiation exposure if the spleen (a hematopoietic organ in the mouse) was protected by lead foil.³ Soon afterward, Lorenz and colleagues showed that lethally irradiated mice and guinea pigs were protected by the administration of syngeneic marrow after irradiation, thereby demonstrating the therapeutic efficacy of allogeneic and xenogeneic marrow suspensions.⁴ These investigators and others considered that chemicals and/or components from the shielded spleen or infused marrow stimulated endogenous hematopoietic cell recovery after total-body irradiation (TBI).^5,6,7 In 1954, Barnes and Loutit showed that if mice were immunized against marrow cells from mice of another strain and then lethally irradiated, no protection was observed by the injection of marrow cells from the strain to which they were immunized. However, if nonimmunized mice were lethally irradiated and injected with the same marrow cells, normal protection was seen and all mice survived more than 60 days. This experiment supported the cellular hypothesis of hematopoiesis and was the first to consider that hematopoietic recovery resulted from cellular repopulation and not from humoral factors.⁸

In 1956, Barnes and associates described the treatment of murine leukemia by supralethal irradiation and marrow grafting.⁹ Researchers pointed out that irradiation alone would not kill all leukemia cells, but that residual leukemia cells might be eliminated by transplanted cells through immunologic mechanisms, and the term adoptive immune therapy was coined. Their publication stimulated tremendous interest, and the period from 1956 to 1959 was characterized by an increasing appreciation of the potential application of marrow grafting to treat individuals exposed to lethal irradiation and to treat human leukemia. Thomas and colleagues began clinical studies in patients with terminal leukemias, and in 1957 described six patients treated with irradiation and intravenous infusion of marrow from healthy donors.¹⁰ Only two patients developed transient detectable donor hematopoietic engraftment, and none of the six survived beyond 100 days from the cell infusion. In 1959, Thomas and associates described a patient with terminal leukemia who received total body irradiation (TBI) and an intravenous infusion of marrow from an identical twin.¹¹ The patient showed prompt hematopoietic recovery and disappearance of the leukemia for 4 months, confirming for the first time that lethal irradiation followed by compatible marrow could have an antileukemic effect and restore normal marrow function. In the same year, Mathé and associates reported the infusion of marrow into six patients exposed to potentially lethal irradiation in a reactor accident in Belgrade, Serbia.¹² Five patients survived, yet there was no clear evidence of donor hematopoietic engraftment and thus no firm agreement over the contribution of marrow transplantation to patient recovery.

The first attempts at autologous marrow transplantation appeared during this time as well. In 1958, Kurnick and colleagues described two patients with metastatic cancer whose marrow was collected and stored by freezing.¹³ Following intensive radiation therapy, the marrow was thawed and infused intravenously. One patient died from transplantation complications, while the other showed hematopoietic recovery after a prolonged period of pancytopenia. In Philadelphia, an autologous marrow transplant was carried out after high-dose nitrogen mustard conditioning in a patient with malignant lymphoma who lived for more than 30 years after transplantation, the majority of that time in complete remission.¹⁴

Despite the many successful preclinical models of marrow transplantation and the predictive value of in vitro histocompatibility testing, the period of 1960 to 1967 was marked by increasing pessimism about allogeneic marrow grafting in humans. In a published compendium of 203 human allogeneic marrow grafts carried out throughout the 1950s and 1960s, none were considered successful.¹⁵ The first positive results came from studies of children with severe combined immunodeficiency (SCID). In 1968, Gatti and colleagues performed the first successful allogeneic marrow HCT in a child with SCID.¹⁶ The lymphoid elements of the donor graft corrected the immunodeficiency, and two similar cases were reported shortly thereafter.^17,18 These patients remained alive and well 25 years later.¹⁹

These successes stimulated a resurgence of enthusiasm for marrow transplantation, and by 1975 strikingly improved results were published by the Seattle team.²⁰ These investigators reported the outcomes of 37 patients with aplastic anemia and 73 patients with leukemia who had reached an advanced stage of their disease before transplantation. This study stressed the importance of histocompatibility and proper preparation of the patient before transplantation, detailed the technique of marrow transplantation, emphasized the role of posttransplant immunosuppression and supportive care, and raised the possibility of using unrelated donors. This report ushered in the modern era of allogeneic HCT. In 1977 and 1980, the first successful HCT procedures from unrelated marrow donors were reported.^21,22 At the end of 1978, the first series of successful autologous HCT for lymphoma were reported.^23,24 In 1990, the Nobel Prize in Physiology or Medicine was awarded to E. Donnall Thomas in recognition of his pioneering work in the field of marrow transplantation. By 2013, more than 700,000 patients worldwide had undergone transplantation during the previous 3 decades and more than 19,000 transplants were being performed annually.²⁵

At the single-cell level, stem cells self-renew and give rise to progeny that differentiate into functional cells carrying out specific functions (Chap. 18).²⁶ Progenitor cells can be multipotent, oligopotent, or unipotent, but lack self-renewal capabilities. Hematopoietic stem cells (HSCs) are cells that give rise to more HSCs and form all elements of the blood. HSCs are entirely responsible for the development, maintenance, and regeneration of blood-forming tissues for life, and are the most important, if not the only, cells required for successful engraftment in hematopoietic transplantations.²⁶ In the adult mouse marrow, all HSC activity is contained in a population marked by the composite phenotype of c-kit⁺, Thy-1.1^lo, lineage marker^–/lo, and Sca-1⁺ (designated KTLS).^27,28 When transplanted at the single-cell level into irradiated mice, KTLS HSCs gave rise to lifelong hematopoiesis, including a steady state of thousands of HSCs with more than 10⁹ blood cells produced daily in the mouse.^27,28,29 In humans, the combination of positive selection for CD34, Thy-1, and negative selection for lineage markers identified a homologous HSC population.³⁰

Following the success in rodent models, purified populations of human HSCs were tested in three separate clinical trials of patients with myeloma, non-Hodgkin lymphoma (NHL), and metastatic breast cancer.^31,32,33 The goal of these trials was to purify HSC and thereby reduce the risk of occult malignant cells contaminating the autografts. These trials presented technical challenges primarily because of the rarity of HSC in marrow and granulocyte colony-stimulating factor (G-CSF)–mobilized blood. However, adequate numbers of HSC could be isolated that were tumor-free in the majority of patients. The times to neutrophil and platelet recovery following purified HSC infusion were comparable to those seen with unmanipulated marrow, but T-cell recovery (especially that of CD4+ T cells) was delayed by up to 6 months in almost all patients. A number of patients developed unusual infections (e.g., severe cases of influenza, respiratory syncytial virus, cytomegalovirus [CMV] and Pneumocystis pneumonia), thus raising concern over “pure” HSCs as the sole source of hematopoietic reconstitution in clinical transplantation. Although these studies were not powered to detect an impact of purified HSCs on relapse or overall survival, these outcomes appeared favorable compared to historical controls.³⁴ Cotransplantation of antigen-specific mature T cells is being investigated as an approach to address the T-cell–specific immunodeficiency seen in patients receiving purified HSC autografts.³⁵

The ability of HSCs to migrate from marrow to blood and back has been conserved throughout evolution. Although the biologic role and physiologic significance of constitutive HSC circulation remains unclear, this capacity to traffic leads to hematopoietic cell reconstitution and is essential for the success of HCT in the treatment of hematologic and nonhematologic diseases.

The restoration of adequate hematopoiesis after transplantation requires a series of balanced interactions between the infused HSCs and the complex supporting marrow microenvironment (Chap. 5). Initially, infused HSCs must adhere to the marrow endothelium with sufficient strength to overcome the shear forces of blood flow.³⁶ Adhesion and arrest of HSCs are mediated primarily by the selectin ligand P-selectin glycoprotein ligand-1 (PSGL-1) and by the hematopoietic cell L- and E-selectin ligands, which interact principally with endothelial E-selectin.^37,38 Other HSC surface adhesion molecules that mediate adherence to the marrow endothelium include members of the integrin superfamily, principally very late antigen-4, integrin α4β7 and lymphocyte function antigen-1, that interact with endothelial immunoglobulin (Ig) superfamily receptors (e.g., vascular cell adhesion molecule [VCAM]-1), and the hyaluronate receptor CD44.^39,40 HSCs that are null for the β integrins cannot migrate to their marrow niche even though they proliferate and differentiate in the fetal liver.⁴¹ Following firm adherence, the transendothelial movement and intraparenchymal homing to hematopoietic niches within the inner endosteal surface of the bone are predominantly regulated by a gradient of extracellular-matrix-bound stromal cell–derived factor-1 (also known as CXCL12).¹² Mice deficient in CXCR4 develop fetal liver hematopoiesis but die prenatally as a consequence of the lack of marrow hematopoiesis.42 The requirement for CXCR4 expression on HSCs for homing and engraftment is well-documented,⁴³ and has led to the development of CXCR4 antagonists such as plerixafor, which help mobilize marrow stem cells for clinical use.⁴⁴

Following successful homing, the initial adhesion of HSC within the hematopoietic niche appears to be regulated at least in part by annexin II.⁴⁵ The marrow niche is a complex biologic unit that includes potentially self-renewing mesenchymal stromal cells (MSCs), regulatory T cells (T_reg), and cells with the defined phenotype of parathyroid-hormone-receptor-bearing osteoblasts.^46,47 MSCs promote engraftment when cotransplanted with HSCs.⁴⁸ Osteoblasts, possibly in conjunction with sinusoidal endothelial cells, also appear to play a pivotal role in the regulation of HSC engraftment by producing a number of molecules, such as annexin II, VCAM-1, intercellular adhesion molecule-1, CD44, CD164, and osteopontin, which promote engraftment.^49,50 Stimulation of osteoblasts with parathyroid hormone results in expansion and mobilization of the HSC pool in animals,⁵¹ although a clinical trial in human cord-blood recipients did not demonstrate a benefit.⁵² In addition to the regulators of HSC adhesion and homing, the function of HSCs is further regulated by intrinsic genetic programs for quiescence, self-renewal, proliferation, differentiation, and apoptosis that are dependent on communication with a network of interacting cells in the marrow microenvironment, including various T-cell subpopulations, adipocytes, and fibroblasts. Given the complexity of HSC trafficking and control, it is surprising that clinical HCT has a relatively low rate of graft failure.

In humans, HSCs for transplantation can be collected from several sources, including directly from the marrow; from the blood after mobilization; and from umbilical cord blood (UCB) obtained at the time of delivery.

MARROW

Marrow has historically been the traditional source of HSCs for allogeneic and autologous transplantation. Marrow for transplantation is typically aspirated by repeated placement of large-bore needles into both sides of the posterior iliac crest, generally involving 50 to 100 aspirations per side, with the patient under regional or general anesthesia. The lowest cell dose which ensures stable long-term engraftment has not been defined with certainty; typical collections contain at least 2 × 10⁸ total nucleated marrow cells per kg of recipient body weight. Current guidelines indicate that collection of up to 20 mL/kg of donor body weight is considered safe.

Marrow harvesting is considered a very safe procedure, and serious side effects are rare. A review of almost 10,000 healthy adult volunteer unrelated donors by the National Marrow Donor Program (NMDP) found that the risk of serious adverse events was 2.38 percent, most of which were mechanical or anesthesia-related and self-limited. Unexpected, life-threatening, or chronic complications occurred in 0.99 percent of donors.⁵³ Evaluation of pediatric marrow donor safety is more limited, but a safety review of 453 pediatric donors by the European Group for Blood & Marrow Transplantation (EBMT) found no serious adverse events; pain was the most common complaint but lasted a median of only 1 day after donation.⁵⁴ A survey of pediatric transplantation hematologists confirmed that 90 percent of centers were willing to perform a marrow harvest on children, even on those younger than 6 months old.⁵⁵

BLOOD

HSCs are normally present in the blood at very low levels. However, a number of different stimuli, including chemotherapy, hematopoietic growth factors, and inhibitors of certain chemokine receptors, result in the mobilization of HSC from marrow to blood. Once mobilized, HSC can be collected by apheresis; this product has been termed “peripheral blood progenitor cells (PBPCs)” to differentiate it from “blood stem cells,” a term that should be reserved for instances where the HSC population itself has been isolated. Agents used to mobilize HSCs include G-CSF, granulocyte-monocyte colony-stimulating factor (GM-CSF), interleukin (IL)-3, thrombopoietin, and the CXCR4 antagonist plerixafor.^44,56,57

Autologous PBPCs are most commonly mobilized with G-CSF, with or without additional chemotherapy. In contrast, PBPCs for allogeneic HCT are typically mobilized with G-CSF alone, so as to avoid exposing healthy donors to chemotherapy. PBPC mobilization and collection is very safe; a review of nearly 7000 healthy unrelated PBPC donors performed by the NMDP found that the rate of serious adverse events was 0.56 percent, making PBPC donation significantly safer than marrow donation.⁵³ The most common side effect of PBPC mobilization and collection is bone pain as a result of G-CSF administration. More serious side effects, such as splenic rupture or intracranial hemorrhage, have been described in case reports but are extremely rare.^58,59

Theoretical concern exists about the potential of short-term growth factor therapy to increase the risk of leukemia in normal donors. However, long-term followup of healthy adult PBPC donors has shown no late effects of G-CSF administration or donation—in particular, no increased risk of cancer, autoimmune disease, or stroke.⁵³ Detailed white cell subset analysis by fluorescent-activated cell sorting of healthy donors at 1 year after donation shows no changes in B, T, and natural killer (NK) cells or monocytes and neutrophils compared with analysis before G-CSF administration.⁶⁰

The adequacy of PBPC products is generally measured through the absolute number of CD34+ cells per kg of recipient body weight. Most laboratories measure CD34+ cell content by fluorescent-activated cell sorting. A threshold of greater than 2 × 10⁶ CD34+ cells/kg is often considered the minimum acceptable dose for PBPC products, although successful engraftment can occur with lower doses.⁶¹ Platelet recovery appears most impacted by low PBPC CD34+ cell dose. Higher CD34+ cell doses are associated with more rapid engraftment, and thus a dose of equal to or greater than 4 × 10⁶ CD34+ cells/kg is considered optimal.⁶¹ The impact of very high CD34+ cells doses in allogeneic HCT is somewhat unclear; some studies have associated doses greater than 8 × 10⁶ CD34+ cells/kg with a higher risk of extensive chronic graft-versus-host disease (GVHD) in matched-related-donor HCT, although this association was not confirmed in unrelated-donor allogeneic HCT. Because there is no evidence of benefit for CD34+ cell doses greater than 8 × 10⁶/kg in allogeneic HCT, this threshold is sometimes, although not universally, used as a maximum.⁶¹

Although inadequate mobilization of healthy donors is rare, patients with malignancies undergoing mobilization for autologous HCT often have difficulty collecting adequate numbers of CD34+ cells because of marrow damage from previous chemotherapy or radiation therapy. Approximately 10 to 20 percent of patients preparing for autologous HCT do not mobilize sufficient numbers of CD34+ cells using G-CSF alone or in combination with chemotherapy. Unfortunately, it has proven difficult to identify these individuals prospectively. For patients with NHL, Hodgkin lymphoma (HL), and myeloma who fail mobilization with G-CSF alone, the majority proceed to collect a transplantable dose of CD34+ cells (>2 × 10⁶ cells/kg) when remobilized with plerixafor plus G-CSF.⁶² Studies of allogeneic HCT using plerixafor-mobilized grafts have also confirmed prompt and stable donor cell engraftment.⁶³ Animal models suggest that plerixafor-mobilized PBPCs have a different phenotype and cytokine profile than G-CSF–mobilized PBPCs and may be associated with a higher risk of acute GVHD; however, the relevance of these findings to human allogeneic HCT is unclear.⁶⁴ For patients who are thought to be at high risk of poor mobilization, strategies include the upfront use of plerixafor and/or chemotherapy to supplement G-CSF mobilization, as well as large-volume leukapheresis.⁶¹ In 2014, the American Society for Blood & Marrow Transplantation (ASBMT) published guidelines on the mobilization and collection of PBPCs for autologous and allogeneic HCT.⁶¹

There is some evidence that circadian activity in the hypothalamus regulates the cyclic release of HSCs by altering the expression of CXCL12 in the marrow microenvironment, with the peak time for HSC release in humans in the evening.⁶⁵ Preliminary clinical data suggest that CD34+ yield is higher in donors collected in the later afternoon compared to the morning, and more abundant PBPC collections were reported from healthy donors when apheresis was performed at 8:00 PM as opposed to 8:00 AM.⁶⁶ Efforts to exploit this circadian rhythm dependence to increase HSC yield in PBPC products have thus far been inconclusive.⁶⁷

Mobilized Peripheral Blood Progenitor Cells versus Marrow

In the setting of autologous HCT, the superiority of PBPCs over marrow as a stem-cell source is clear. Randomized trials have shown that PBPC autografts in this setting are associated with more rapid engraftment, better quality of life, and lower costs compared to marrow autografts.^68,69,70 On the basis of these and other results, most transplantation centers use mobilized PBPCs for autologous HCT and have adopted a minimum CD34+ cell of 2 × 10⁶ CD34+ cells/kg.⁶¹

In the allogeneic setting, the situation is considerably more complex. PBPC grafts contain approximately 10-fold more T cells compared to marrow grafts, leading to concern over a potentially increased incidence and severity of GVHD. At the same time, G-CSF can induce functional immune tolerance in healthy individuals, and T cells from G-CSF–mobilized PBPC grafts show a predominantly immune-tolerant profile with upregulation of genes related to T-cell helper type 2 (Th2) and T_reg cells, and downregulation of genes associated with Th1 cells, cytotoxicity, antigen presentation, and GVHD.⁷¹

A number of randomized clinical trials have compared PBPC and marrow grafts in the setting of allogeneic HCT.^72,73,74,75 These studies have consistently reported similar or better overall and disease-free survival with PBPCs compared to marrow allografts. Most, although not all, of these randomized trials found a higher risk of chronic GVHD with PBPCs compared to marrow allografts, and one reported a longer duration of immunosuppression in patients receiving a PBPC graft.⁷³ PBPC allografts were also associated with faster engraftment and a lower risk of graft failure. Systematic reviews and meta-analyses have similarly reached varying conclusions.^76,77,78 A 2014 systematic review performed by the Cochrane Library found that overall survival was similar with PBPC and marrow allografts, and that PBPC allografts were associated with faster engraftment but also a higher incidence of chronic GVHD. The effects of stem cell source on relapse and on acute GVHD were unclear.⁷⁶

Currently, the choice between PBPCs and marrow allografts is generally individualized and depends on patient, donor, and institutional considerations. Patients with advanced or high-risk hematologic malignancies may preferentially be given PBPC grafts to reduce their risk of relapse, a strategy with some support in the literature.⁷⁹ Conversely, patients with standard-risk malignancies are often given marrow allografts to reduce their risk of chronic GVHD. Likewise, patients transplanted for nonmalignant diseases such as aplastic anemia are typically given marrow allografts to reduce their risk of chronic GVHD, as they derive no benefit from T-cell–mediated graft-versus-malignancy effects. In settings where the risk of GVHD is particularly high—for example, with human leukocyte antigen (HLA)-mismatched unrelated donors—many institutions prefer to use marrow allografts to mitigate this risk. For patients receiving reduced-intensity conditioning (RIC), most centers use PBPC allografts exclusively, as engraftment in this setting is largely dependent on the presence of donor T cells in the allograft. Donor factors also play a role in the choice of allograft product, as some donors may specifically decline either marrow donation or PBPC mobilization and collection. Finally, with the recent predominance of PBPC as a graft source, institutional resources for and expertise in marrow collection have decreased, sometimes limiting the availability of marrow allograft products.

ALTERNATIVE SOURCES OF HEMATOPOIETIC STEM CELLS

One of the most significant advances in allogeneic HCT in the past 10 years is the increasing experience with alternative donors for patients who lack HLA-identical siblings or suitably HLA-matched unrelated donors. Each full sibling has a 25 percent chance of being HLA-identical with another, so the likelihood of finding an HLA-identical sibling donor is proportionate to the number of siblings available. HLA-matched unrelated donors can be identified for approximately 75 percent of patients of northern European ancestry, but the odds of finding a suitable unrelated donor are much lower for patients who belong to ethnic groups that are underrepresented in donor registries, those of mixed ethnicity, and those with uncommon HLA haplotypes.⁸⁰ As a result, a substantial fraction of patients who would benefit from allogeneic HCT lack “conventional” related or unrelated donors. For such patients, there are two widely used alternative sources of HSC for allogeneic HCT: UCB and HLA-haploidentical family members.

Umbilical Cord Blood

UCB, collected from the umbilical vessels in the placenta at the time of delivery, is a rich source of HSCs. Because these cells are immunologically naïve, it is feasible to cross major histocompatibility barriers, thus extending the donor pool to individuals for whom finding suitably HLA-matched adult donors can be difficult or impossible.⁸¹ The first successful allogeneic HCT using UCB, in a child with Fanconi anemia, was reported by Gluckman and colleagues in 1989.⁸² Since then, hundreds of thousands of UCB units have been collected and stored; searchable registries have been established to facilitate identification of suitable UCB units for transplantation; and more than 20,000 allogeneic HCTs have been performed using UCB.⁸³ Cord blood units are fully HLA-typed before cryopreservation, and thus suitable units can be rapidly identified (as compared to unrelated-donor searches, which can take 2 to 6 months to complete).

Most UCB units are HLA-typed as HLA-A and HLA-B using low-resolution (serologic) methods, and as HLA-DRB1 using high-resolution (molecular) methods. UCB units matched at equal to or greater than 4/6 HLA loci are generally considered suitable for use, although UCB units matched at 5/6 or 6/6 HLA loci should be used if available because they are associated with lower transplant-related mortality.^84,85 The role of HLA matching between UCB units in double UCB transplantation remains somewhat unclear, and unit-to-unit HLA matching does not appear to impact long-term engraftment rates or GVHD incidence.^85,86

A major limitation has been the relatively small number of HSC available in cord blood units relative to the size of the average adult recipient.⁸⁷ As a result, most UCB transplants in adults use two UCB units rather than one. The minimum acceptable cell doses for single-unit UCB transplantation are generally set at equal to or greater than 2.5 × 10⁷ total nucleated cells or equal to or greater than 2 × 10⁵ CD34+ cells/kg of recipient weight.^84,88 It is difficult to locate units meeting these criteria for average-sized adult recipients, and thus most adult UCB transplants are performed using two UCB units. With double UCB transplantation, transient mixed donor chimerism from the two units is often observed, but ultimately one UCB unit dominates, eradicates the other unit, and is responsible for establishment of long-term hematopoiesis.⁸⁶ Despite extensive investigation, the factors determining which unit becomes dominant remain somewhat unclear,⁸⁹ although CD8+ T-cell responses against the nonengrafting unit have been implicated.⁹⁰ For recipients with a single suitably sized UCB unit, single-unit UCB transplantation is preferable to double UCB transplantation, because of better platelet recovery and a lower risk of GVHD.⁸⁸ For the majority of adult recipients, who lack a suitable single UCB unit, double UCB transplantation is typically used and produces equivalent overall survival.⁸⁹ For children, a single UCB unit appears superior to double UCB transplantation.⁸⁸ A major area of active research in UCB transplantation is the expansion of HSC or other hematopoietic progenitor cells in UCB products, with the goal of improving engraftment rates, shortening the period of preengraftment neutropenia, and reducing the need for double-unit UCB transplantation. Several approaches have been described in the literature, including Notch-mediated or prostaglandin-mediated expansion of progenitor cells and ex vivo mesenchymal cell coculture.^91,92,93

Recipients of UCB allografts have a higher risk of opportunistic infections—particularly viral infection or reactivation—compared to recipients of PBPC or marrow allografts. Presumably the immunologic naïveté of UCB, which allows its use across HLA barriers, also contributes to impaired antiviral immunity and immune reconstitution after allogeneic HCT.⁹⁴ CD8+ T-cell recovery is significantly delayed after UCB compared to marrow allotransplantation (median time to reach >0.25 × 10⁹ CD8+ T cells/L, 7.7 months vs. 2.8 months, respectively), whereas CD4+ T cell and NK cell recovery is similar with these two graft sources.⁹⁵ A novel syndrome of cord colitis has been described in UCB recipients and tentatively linked to Bradyrhizobium enterica, a newly identified bacterium,^96,97 although other groups have questioned the existence of a distinct cord colitis syndrome and instead attributed the findings in question to conventional GVHD.^98,99

Human Leukocyte Antigen–Haploidentical Donors

Virtually all patients have HLA-haploidentical family members—including any parent, any child, and some siblings—available as donors. Haploidentical related donor HCT has been evaluated for more than 2 decades as an alternative source of HSCs. However, as a result of the substantial HLA disparity involved, early attempts at haploidentical HCT were associated with severe GVHD in T-cell-replete transplants and with graft rejection in T-cell–depleted transplants.^100,101 Extensive ex vivo depletion of CD3+ and CD19+ lymphocytes, coupled with megadose CD34+ cells and antithymocyte globulin, can successfully overcome the barriers to engraftment.¹⁰² The extensive T-cell depletion used in these protocols to prevent GVHD would also be expected to result in weak or no graft-versus-malignancy effects. Yet despite the lack of T-cell–mediated alloreactivity and the unfavorable prognostic features at the time of transplantation, relapse rates with this approach remained at 18 to 30 percent in patients with acute leukemia transplanted in first complete remission (CR1).¹⁰³ The low rate of relapse was attributed to a strong antitumor effect mediated by donor NK cell alloreactivity. Transplantation from NK-cell-alloreactive donors was associated with a significantly lower leukemia relapse rate and improved overall survival, leading some authorities to recommend selection of haploidentical donors based on NK cell alloreactivity.¹⁰⁴ However, this approach to haploidentical HCT remains hampered by prolonged immune reconstitution and a high risk of serious infection.¹⁰⁵

More recently, the group at Johns Hopkins has pioneered the use of posttransplantation cyclophosphamide (CY) as GVHD prophylaxis in T-cell-replete haploidentical HCT.¹⁰⁶ In this approach, unmanipulated marrow from an HLA-haploidentical donor is infused after nonmyeloablative conditioning. A period of 48 to 72 hours elapses after infusion, during which alloreactive donor T-cell clones become activated and proliferate. CY is then administered on days +3 and +4 after allogeneic HCT, preferentially eradicating the activated alloreactive donor T-cell clones while leaving other, nonalloreactive clones relatively untouched.¹⁰⁷ This approach has been remarkably well-tolerated and results in very low rates of GVHD and transplant-related mortality (TRM).¹⁰⁸ Additionally, because this approach avoids indiscriminate T-cell depletion, immune reconstitution is relatively robust, and the typical complications of T-cell-depleted allotransplantation (such as posttransplantation lymphoproliferative disorder [PTLD]) are not seen.¹⁰⁹ The major limitation of this approach is a relatively high rate of relapse, perhaps driven by the eradication of the alloreactive donor T-cell clones which would mediate graft-versus-tumor (GVT) effects along with those mediating GVHD.¹⁰⁸

Retrospective studies have examined the question of selecting the optimal HLA-haploidentical donor when several such donors are available. Early evidence suggested that lower TRM was seen with HLA-haploidentical sibling donors compared to parental donors, while maternal donors were associated with less chronic GVHD and better overall survival than paternal donors.^110,111 In contrast, a 2014 paper reporting outcomes of HLA-haploidentical HCT in China found less GVHD and better overall survival with paternal compared to maternal donors.¹¹² As with previous studies, haploidentical siblings (particularly those disparate for noninherited maternal antigens) were associated with the best outcomes. There is some preclinical evidence that exposure to noninherited maternal antigens through breastfeeding may reduce the risk of GVHD in maternal-donor haploidentical HCT,¹¹³ and thus the disparate clinical results with maternal donors may be explained in part by variations in demographic patterns of breastfeeding. In view of the conflicting state of the existing literature, there is no universal standard approach to selecting an HLA-haploidentical donor; the decision is often guided by donor availability and health status, although the literature arguably contains weak support to prefer HLA-haploidentical siblings over parents, and mothers over fathers, as donors.

Comparison of Alternative Donor Options

Although both UCB and HLA-haploidentical HCT have been established as feasible and effective options for patients lacking conventional donors, the optimal alternative-donor source remains unclear. Given the lack of robust comparative data, the choice between UCB and haploidentical HCT is often guided by institutional experience, comfort level, and research priorities. To address the relative merits of these approaches, the United States Blood & Marrow Transplant Clinical Trials Network (BMT-CTN) conducted parallel multicenter phase II clinical trials of RIC followed by either UCB or haploidentical HCT (the latter using T-cell-replete marrow as a graft source and posttransplantation CY as GVHD prophylaxis). These trials reported similar overall and disease-free survival at 1 year with the two approaches.¹⁰⁸ UCB transplantation was associated with a higher risk of GVHD and nonrelapse mortality, while haploidentical HCT was associated with a higher risk of relapse. Based on these results, the BMT-CTN is currently conducting a national, multicenter phase III clinical trial randomizing participants between UCB and haploidentical HCT. In addition to clinical outcomes, this trial is designed to measure quality-of-life and cost differences between the two alternative-donor approaches in a comprehensive effort to clarify their relative advantages and disadvantages.

AUTOLOGOUS HEMATOPOIETIC CELL TRANSPLANTATION

The relationship between the dose of chemoradiotherapy administered and the number of tumor cells killed has been extensively studied in vitro and in preclinical animal models, and forms the rationale for myeloablative autologous HCT. For chemosensitive tumors (including most hematologic malignancies), a steeply rising dose–response curve is observed. The curative potential of autologous HCT is, therefore, derived from high-dose chemotherapy or chemoradiotherapy administered as transplant conditioning to enhance tumor cell kill and overcome drug resistance. This level of dose escalation is possible in autologous HCT because dose-limiting toxicities to the hematopoietic system are circumvented by the infusion of autologous HSCs, which rebuild hematopoiesis after high-dose conditioning.

Autologous HCT is associated with relatively low TRM. The use of mobilized PBPCs rather than marrow as a graft source has decreased the duration of neutropenia, and when combined with improved supportive care and patient selection has reduced the TRM associated with autologous HCT from approximately 8 to 10 percent in historical studies to 1 to 3 percent at most centers. In addition, the reduced risks and faster engraftment times seen with PBPC have enabled many centers to pursue outpatient autologous HCT, further easing the logistical burden on patients as well as treatment costs.

Tumor Contamination in the Autograft

A consistent concern in autologous HCT is the possibility that residual tumor cells may contaminate the HSC product and contribute to relapse. The relative contributions of autograft contamination and residual disease in the patient are difficult, if not impossible, to distinguish. Several investigators have approached this question by marking the HSC product at the time of harvest and then assaying for the marker gene in malignant cells at the time of subsequent relapse. Studies using this approach have been performed in patients with leukemia, lymphoma, and myeloma, and have reached conflicting conclusions about the contribution of autograft contamination to relapse.^114,115

A number of ex vivo and in vivo purging strategies have been investigated in autologous HCT. These include administration of rituximab before autologous HSC collection in patients with CD20+ malignancies; ex vivo positive selection for CD34+ cells; chemotherapy-based purging using CY derivatives; and purging via oncolytic viruses.^{116,117,118,119,120} Single-arm or uncontrolled studies have suggested a possible benefit in some of these instances, but there is little or no evidence from randomized clinical trials at present to support the efficacy of autograft purging, and its use remains investigational. One of the few randomized clinical trials in the field concluded that in vivo purging with rituximab administered before autologous HSC collection was as effective, and likely safer, than ex vivo CD34+ selection.¹²¹ Another randomized clinical trial found that CD34+ selection significantly reduced autograft contamination with myeloma cells, but did not improve clinical outcomes.¹²² Purging strategies carry some potential for harm, as they deplete the autograft of mature T cells and increase the risk of infectious complications, particularly CMV or other viral reactivation, after autologous HCT.^123,124 At present, given the lack of convincing evidence of clinical benefit, purging strategies are not widely used and most centers collect and infuse unmanipulated autologous HSC (although these cells are often mobilized with a regimen containing chemoimmunotherapy, which arguably represents a form of in vivo purging). Because relapse remains a major concern after autologous HCT, ongoing research is focused on identifying more efficient and clinically effective means of autograft purging.

ALLOGENEIC HEMATOPOIETIC CELL TRANSPLANTATION

Allogeneic HCT is a considerably more complicated procedure than autologous HCT. It involves more pretransplantation preparation, poses a greater risk of complications to the patient, is associated with a significantly higher rate of TRM, and requires at least temporary postgrafting immunosuppression to enable engraftment and prevent GVHD. The decision to pursue allogeneic HCT is based on diagnosis, prognosis, and remission status, as well as the availability of an appropriate donor and the psychosocial resources of the patient to cope with the demands of the process. Decisions about eligibility for allogeneic HCT are typically made on a center-by-center and case-by-case basis, with inherent elements of subjectivity and patient and physician judgment.

A major obstacle to successful allogeneic HCT is the immune competence of the recipient. The potential to reject infused donor cells is mediated predominantly through regimen-resistant host T and NK cells. Strategies to reduce host immunity and promote donor hematopoietic cell engraftment include pretransplantation conditioning (most often chemotherapy and/or radiation) and postgrafting immunosuppressive medications. Donor T cells in the allograft play a key role in hematopoietic engraftment, and depletion of T lymphocytes from the allograft before transplantation is associated with a substantial increase in the occurrence of graft rejection.^125,126 With modern conditioning regimens, rates of allograft rejection are low (typically <5 percent) for patients who have received previous chemotherapy, which weakens their immune response to the allograft. In contrast, graft rejection in seen more often in patients who have not received cytotoxic chemotherapy before allogeneic HCT (for instance, some patients with myeloproliferative neoplasms) or those with nonmalignant diseases such as aplastic anemia, thalassemia, or sickle cell disease, who are often highly sensitized against donor antigens by virtue of being heavily transfused before allogeneic HCT.

Graft-Versus-Tumor Effects

The dominant mechanism of cancer eradication following allogeneic HCT is the immunologic recognition and destruction of residual host tumor cells by donor-derived immune cells. This phenomenon, termed the GVT effect, has been conclusively demonstrated and represents one of the most significant biologic findings of the past half-century, with implications well beyond the transplantation setting. The existence of GVT effects is supported by the following lines of evidence:

• Tumor relapse is lower after allogeneic than after syngeneic HCT: The appreciation of alloreactive GVT effects stemmed from the observation that recipients of genotypically identical (syngeneic) grafts had significantly higher rates of disease relapse than did patients who received grafts from HLA-identical siblings.^127,128

• Tumor relapse is higher in recipients of T-cell–depleted grafts: Further support for the allogeneic GVT effect came from studies of T-cell depletion of the graft.¹²⁹ T-cell depletion was performed in the expectation that it would reduce the risk of GVHD. However, the later observation that these patients had a much higher risk of disease recurrence as well as graft rejection was unanticipated. These results linked GVHD with GVT effects, and supported the concept that patients who developed some degree of alloreactivity, as manifested by clinically apparent GVHD, had a reduced risk of disease relapse.

• Donor lymphocyte infusions can induce remission: Perhaps the most definitive evidence for the existence of GVT effects came from the application of donor lymphocyte infusions (DLIs). In the early 1990s, Kolb and others demonstrated that patients with relapsed malignancies after allogeneic HCT could, in some cases, be returned to complete remission by the simple infusion of donor-derived lymphocytes.^130,131,132 With increasing experience, it became clear that some diseases (such as chronic myelogenous leukemia [CML]) respond very well to DLI whereas others (for instance, acute lymphoblastic leukemia [ALL]) are much less responsive. Long-term followup of patients successfully treated with DLIs revealed that responders had remarkably durable remissions and excellent outcomes.¹³² These observations definitively established the GVT effect as a biologic entity capable of controlling an otherwise lethal condition such as leukemia.

Targets and Effector Cells in Graft-Versus-Tumor Reactions

The biology of the GVT effect remains incompletely understood. A number of immunologic targets recognized by donor immune effector cells have been proposed, including alloantigens (such as major or minor histocompatibility antigens depending upon donor–recipient genetic differences), lineage-specific antigens, and malignancy-specific antigens such as products of chromosomal translocations. Donor T cells clearly play a key role in GVT, and there is emerging evidence that NK cells are also responsible for tumor cell control, especially in the setting of T-cell-depleted HLA-haploidentical HCT.¹³³ Humoral immunity has also been implicated as playing a role in the GVT effect.¹³⁴ Based on preclinical models, the common effector cells that could potentially mediate a clinical GVT effect include (1) CD8+ cytotoxic T lymphocytes that recognize tumor-associated antigens in context of class I major histocompatibility complex (MHC) antigens; (2) CD4+ T cells that recognize tumor-associated antigens in context of class II MHC antigens and mediate their effects via Th1 cytokines such as interferon (IFN)-γ and IL-2, upregulating expression of class I MHC antigens and promoting expansion and activation of CD8+ cytotoxic T lymphocytes; and (3) NK cells that recognize stress ligands and cells lacking MHC expression.^{135,136,137,138,139,140,141} The impact of NK cells seems especially pronounced in HLA-haploidentical or mismatched allotransplantation.^104,142

A major question continues to be whether the subset of T cells that induce GVHD is the same population of T cells responsible for the GVT effect. One of the central aims of research in allogeneic HCT is to separate the beneficial GVT effects from deleterious GVHD. Clinical evidence suggests that, in principle, the two should be separable, as some patients experience an apparent GVT effect in the absence of apparent GVHD.¹⁴³ Although numerous approaches have succeeded in separating GVT effects from GVHD in preclinical and animal models, none of these approaches have yet been translated successfully to widespread clinical use. One approach supported by preclinical models involves the use of T_reg,¹⁴⁴ and recent studies in the HLA-haploidentical setting are supportive of this concept.^145,146

The transplant preparative regimens used in HCT must accomplish two goals. Because the majority of autologous and allogeneic HCT are performed in individuals with cancer, these regimens were designed, at least initially, to maximize tumor cytoreduction and disease eradication. In the case of allogeneic HCT, the regimen must be sufficiently immunosuppressive to overcome host rejection of the graft. In autologous HCT, where efficacy depends on exploiting the dose–response curve, high-dose conditioning regimens are universally used. In contrast, in allogeneic HCT much or all of the clinical benefit derives from donor alloimmunity, enabling the use of RIC designed to facilitate donor engraftment with minimal toxicity.

TOTAL-BODY IRRADIATION

TBI has been a primary component of many autologous and allogeneic HCT preparative regimens since the inception of the field. TBI has excellent activity against a variety of hematolymphoid malignancies, has pronounced immunosuppressive properties, and is able to treat sanctuary sites like the testicles and the central nervous system. Aside from one very early study of high-dose TBI alone, most preparative regimens combine TBI with cytotoxic agents such as CY. Dose-finding studies suggest that higher TBI doses are associated with a lower risk of relapse with dose escalation as high as 15.75 Gy, but doses above 12 Gy are associated with higher risks of GVHD and TRM, which offset the reduced risk of relapse.¹⁴⁷ Currently, most high-dose TBI-based conditioning regimens use a dose between 12 and 13.2 Gy. Long-term concerns with TBI-based regimens include the development of cataracts and hypothyroidism, impairment of growth and development in children, and secondary malignancies.^148,149,150

Hyperfractionated TBI, in which relatively small dose fractions are given two to three times a day over a few days, minimizes leukemia regrowth and reduces lung and gastrointestinal toxicity, allowing higher TBI doses to be administered safely. Several clinical studies confirm decreased overall lung toxicity with fractionation.^151,152 Excellent results are also reported with the combination of fractionated TBI and VP-16 (etoposide), particularly promising results in patients with ALL.^153,154

An alternate form of irradiation is radioimmunotherapy, which involves the use of antibodies to deliver locally acting radionucleotides to targeted sites. In theory, this strategy could provide excellent targeted antitumor effects without increased systemic toxicity. Clinical trials incorporating anti-CD45 monoclonal antibodies conjugated to radioactive iodine (¹³¹I) or yttrium (⁹⁰Y) show promising early results in both the autologous and the allogeneic setting.^{155,156,157,158} One approach to further improving the targeting of radiation is the use of α-particle emitters. Alpha particles have a very short effective range but carry immense kinetic energy, making them an attractive option to maximize malignant cell killing while minimizing bystander damage. Preclinical studies using bismuth-213 (²¹³Bi) and astatine-211 (²¹¹At) have demonstrated efficacy,^{159,160,161,162} and this approach is currently being translated into clinical trials.

CHEMOTHERAPY-ONLY REGIMENS

Autologous

Patients with NHL or HL have often received prior intensive local radiotherapy, often to the mediastinum, which results in a high incidence of fatal interstitial pneumonitis following TBI.¹⁶³ Therefore, non-TBI-containing conditioning regimens were developed. The choices of drugs include agents that can be significantly dose-escalated and have improved tumor cell killing at higher doses, yet have nonoverlapping toxicities; for example, a common preparative regimen for autologous HCT in lymphoma includes 1,3-bis(2-choloroethyl)-1-nitrosurea (carmustine or bis-chloroethylnitrosourea [BCNU]), VP-16, and CY.¹⁶⁴ The dose-limiting non-hematologic toxicity of BCNU affects the lungs; VP-16 affects the liver; and CY, the heart. Consequently, using these drugs below the maximally tolerated doses results in relatively low regimen-related toxicity but maximizes tumor cell kill to overcome drug resistance. Many other chemotherapy-only regimens have been developed for autologous HCT according to similar principles, including BCNU, etoposide, cytarabine, and melphalan (BEAM) (for lymphomas) and high-dose melphalan (for myeloma).

Allogeneic

A variety of chemotherapy-only conditioning regimens have been developed for allogeneic HCT. The most widely used combines oral busulfan (BU), at a total dose of 16 mg/kg given over 4 days, with CY at a total dose of 120 mg/kg given intravenously over 2 days (referred to as BU/CY).^165,166 A randomized comparison between fractionated TBI plus CY (TBI/CY) and BU/CY in patients transplanted for CML demonstrated that the BU/CY regimen was better tolerated, but there was no significant difference in overall or event-free survival, TRM, or GVHD incidence.¹⁶⁷ The development of intravenous BU further improved the availability and tolerability of this regimen, although steady-state plasma concentrations remain variable after intravenous BU and therapeutic drug monitoring arguably remains necessary.^168,169 Interestingly, pretreatment with BU may deplete hepatic glutathione and thus potentiate the toxicity of CY.^170,171 Reversing the sequence of conditioning agents (from BU/CY to CY/BU) has been studied as a means of reducing regimen-related toxicity, and this alteration to the regimen has been associated with reduced exposure to toxic CY metabolites and a lower risk of hepatotoxicity.¹⁷²

Other high-dose chemotherapy-only conditioning regimens remain in use on an institution-, disease-, and patient-specific basis. The most common modification to BU/CY is the substitution of fludarabine (FLU) for CY (BU/FLU). This regimen has been proposed to avoid the hepatotoxicity of CY and to reduce regimen-related toxicity compared to BU/CY. However, two recent randomized clinical trials comparing BU/CY and BU/FLU have found increased risks of graft rejection and pneumonitis, as well as decreased overall and disease-free survival, with BU/FLU, raising the concern that this regimen is inferior to BU/CY.^173,174

REDUCED-INTENSITY TRANSPLANTATION

The demonstration that immune-mediated mechanisms are critical in eradicating malignancy after allotransplant challenged the rationale for relatively toxic full-dose conditioning. A number of reduced-intensity regimens have been developed which have lower regimen-related toxicity but are sufficiently immunosuppressive to allow full donor engraftment, shifting the responsibility for tumor eradication largely or entirely to immunologic GVT effects. The development of RIC is one of the most transformative advances in the field of allogeneic HCT in the past several decades, as it has allowed the expansion of eligibility for allogeneic HCT to older and less medically fit patients who would otherwise be ineligible for high-dose conditioning. Additionally, patients with relatively indolent disease may not require the immediate cytoreductive capacity of an aggressive preparative regimen and may be particularly suitable candidates for transplantation using RIC. These regimens are also useful in patients with nonmalignant diseases where the goal is strictly the establishment of donor hematopoiesis; examples include genetic disorders, autoimmune diseases, and the induction of tolerance in combined solid-organ and same-donor marrow transplantation.

A variety of RIC regimens with differing dose intensities have been developed. One significant regimen came from detailed studies in a canine model using a backbone of low-dose 2 Gy TBI followed by postgrafting immunosuppression with mycophenolate mofetil (MMF) and cyclosporine (CSP).¹⁷⁵ This work was translated to patients with a variety of malignancies and resulted in reliable donor engraftment with the addition of intravenous FLU 90 mg/m².¹⁷⁶ This reduced-intensity regimen has been used successfully in more than 1000 patients with a wide variety of malignancies, especially in older patients and those with more indolent diseases such as follicular NHL or chronic lymphocytic leukemia (CLL).^177,178,179 Another commonly used RIC regimen consists of intravenous FLU (between 90 and 150 mg/m²) and CY (between 900 and 2,000 mg/m²).¹⁸⁰ This regimen, combined with rituximab or ⁹⁰Y ibritumomab tiuxetan, has produced excellent long-term disease-free survival in patients with indolent lymphoma.¹⁸¹ A third common reduced-intensity regimen was developed in a rodent model using fractionated low-dose total lymphoid irradiation (TLI) combined with depletive T-cell antibodies (antithymocyte serum), which showed that recipients were protected from GVHD induction by donor-derived T cells.¹⁸² Rodents conditioned with this approach did not develop lethal acute GVHD despite the infusion of megadoses of donor T lymphocytes. TLI and antithymocyte serum altered residual host T-cell subsets to favor regulatory NK/T cells which suppress GVHD by polarizing the infused donor T cells toward secretion of noninflammatory cytokines such as IL-4, and by promoting expansion of donor CD4+CD25+FoxP3+ T_regs.¹⁸³ This approach was successfully translated to clinical transplantation; patients conditioned with TLI and antithymocyte globulin (ATG) developed sustained donor-derived hematopoiesis with very low incidences of acute GVHD and TRM.^184,185 The relative merits of RIC versus high-dose conditioning continue to be studied; a national multicenter prospective randomized trial comparing high-dose to RIC in patients with acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS) is ongoing under the auspices of the BMT-CTN.

Mixed Chimerism Following Reduced-Intensity Conditioning

A feature common to all RIC protocols is the incomplete eradication, at least initially, of host hematopoietic elements. As a consequence, a significant percentage of patients have mixed donor–recipient chimerism for months after transplantation before fully converting, if ever, to complete donor type. Most reports suggest that persistent mixed chimerism is a significant risk factor for disease relapse.^185,186 Interventions such as withdrawal of immunosuppressive medications, CD34-selected donor cell boost, and DLI have been used to convert mixed chimerism to complete donor type, although these interventions are not without risk and can precipitate GVHD. A significant percentage of patients who received alemtuzumab-containing conditioning regimens required DLI to promote donor engraftment and protect against immune-mediated graft rejection, but the risk of developing post-DLI GVHD was significant.¹⁸⁷ Mixed chimerism is not unique to RIC; prior to the addition of ATG, almost 60 percent of patients who received high-dose conditioning and allogeneic HCT for severe aplastic anemia had mixed chimerism, of whom two-thirds eventually converted to complete donor hematopoiesis while the remainder experienced late graft failure.¹⁸⁸

Persistent mixed chimerism may have a role in allogeneic transplantation for noncancer patients. In organ transplantation, tolerance, defined as immunosuppressive drug withdrawal without graft rejection, was achieved in kidney transplant recipients who received the same donor marrow when sustained mixed chimerism was established, and not in recipients who experienced donor hematopoietic graft loss.¹⁸⁹

Most transplantation candidates are referred by hematologists or oncologists to the tertiary center where HCT will be performed. Patients considered for transplantation require in-depth evaluation and counseling by experienced transplantation physicians, nurses, and social workers. A detailed review of initial diagnostic studies, previous drug and radiation treatments, and responses to these interventions, as well as a psychosocial assessment of the patient and their caregivers, are of the utmost importance. Table 23–2 highlights the issues and topics that should be addressed during counseling meetings with transplantation candidates and their caregivers.¹⁹⁰ Important factors which consistently impact outcomes following HCT include, but are not limited to, disease status at transplantation, type and compatibility of donor, recipient age, and comorbid medical conditions. Early referral for transplantation consultation is critical, particularly if allogeneic HCT is under consideration, because of the time required to identify a suitable donor.

DISEASE STATUS AT THE TIME OF TRANSPLANTATION

Disease status at the time of transplantation is perhaps the most powerful predictor of long-term disease-free survival following allogeneic and autologous HCT. Early studies of allogeneic HCT were performed using predominantly patients with refractory and progressive disease.¹⁹¹ Although a small percentage of these patients were salvaged, transplantation was unsuccessful in the majority of patients due to progressive malignancy. Patients with acute leukemias transplanted in complete remission (CR) have substantially better outcomes compared to those transplanted with active disease.¹⁹² Even very minimal amounts of residual disease are associated with significantly higher relapse risk in patients undergoing allogeneic HCT for AML, regardless of conditioning intensity,^193,194,195 underscoring the importance of disease status at transplant as a prognostic factor. Likewise, disease status as determined by positron emission tomography (PET) prior to transplant is an important predictor of progression-free survival for patients with diffuse large B-cell lymphoma and HL undergoing autologous HCT.^{196,197,198,199} These scenarios highlight a truism: patients who have advanced poorly controlled disease at the start of transplant conditioning have significantly inferior outcomes compared to patients transplanted earlier in the course of their disease and to those who have achieved good, albeit temporary, control of their disease. On the other hand, attempts to salvage patients with advanced disease who have failed multiple therapies are rarely successful, and transplantation is often best considered early in the course of therapy. These discussions are complex, as earlier transplantation, especially allogeneic, carries significant risks to the patient. A number of other disease-specific considerations are important in determining the appropriate timing for transplantation, including the presence and/or persistence of cytogenetic and molecular abnormalities, the immune phenotype, and evidence of extramedullary or extranodal disease. Advanced genetic characterization of leukemia and lymphoma may provide improved insight into cohorts of patients for which HCT should be performed earlier in the course of disease.

AGE

Historically, older age was a significant barrier to allogeneic HCT, since older patients suffered severe and often prohibitive toxicity after high-dose conditioning regimens.²⁰⁰ However, the impact of older age is mitigated by the increasing use of RIC for allogeneic HCT.²⁰¹ High-dose conditioning and allogeneic HCT is generally reserved for patients 60 years of age or younger, whereas allotransplantation with RIC has been performed successfully in patients into their eighth decade of life. Most centers in the United States do not have a stringent age cutoff for allogeneic HCT, although careful screening for comorbid medical conditions, such as heart, lung, kidney, and liver disease, is particularly important in older patients, and allotransplantation in patients 70 years of age and older remains somewhat controversial. Existing data suggest that allogeneic HCT can be safely performed in selected patients age 60 to 75 years, with a 5-year overall survival of 35 percent.^202,203 Some investigators have proposed that age is a poor and imprecise prognostic marker, and instead advocate the use of comorbidity assessment and scoring to determine eligibility for allogeneic HCT.²⁰⁴ However, caution is warranted since this approach has not been prospectively validated; retrospective cohort studies necessarily suffer severely from patient selection bias, as they include only those older patients who were deemed appropriate candidates to proceed to allogeneic HCT. Outcomes for this selected group of older patients cannot be generalized to the population of older adults as a whole.

In contrast to allogeneic HCT, autologous HCT relies on high-dose conditioning for its antitumor efficacy. Thus, there is no way to reduce conditioning intensity without sacrificing some degree of efficacy. As a result, age limitations are often stricter for autologous HCT than for allogeneic HCT, as candidates for the former must be able to tolerate intensive, high-dose chemotherapy. It is unusual for autologous HCT to be offered to patients older than 75 years of age.^205,206

COMORBID MEDICAL CONDITIONS

Comorbid medical conditions have a significant impact on transplantation outcomes. Routine screening of heart and lung function to detect occult abnormalities is of critical importance, especially in older patients. Evaluation of liver and kidney function, as well as exposure to potential pathogens such as CMV, hepatitides B and C, herpes viruses, and HIV are routine and should be performed in all patients. Another major factor is the nutritional status of the patient, as extremes such as cachexia or obesity require special considerations and adversely impact TRM.^207,208

Several scoring instruments have been devised to allow quantification and comparison of pretransplantation comorbidities. The most widely used of these are the EBMT Risk Score,²⁰⁹ the Pretransplant Assessment of Mortality (PAM) score,²¹⁰ and the HCT-specific Comorbidity Index (HCT-CI; later modified to incorporate age).^204,211 A number of efforts have been made to validate, revise, or combine these scores in diverse populations, with variable success.^{212,213,214,215,216} Several caveats are important when considering quantitative scoring of pretransplantation comorbidities. First, because these scoring instruments were derived from retrospective cohorts of patients, they suffer from an inescapable selection bias, as only patients who were deemed fit for transplantation were included in their derivation. This selection bias limits the ability to generalize their use to unselected patient populations. Second, transplantation outcomes have not remained stable over time; instead, TRM has steadily decreased over time with the availability of better supportive care and other refinements.^217,218 Thus, it is conceivable that the relative impacts of specific comorbidities on transplantation outcomes are not fixed, but may vary over time. Efforts are currently underway to prospectively validate these comorbidity scoring instruments.

Every effort should be made to encourage potential patients to maintain good health practices, including discontinuation of alcohol use, tobacco smoking, and illicit drug use (if applicable). Centers vary in their approach to abstinence, but it is common to require that patients cease all tobacco use permanently as a condition of proceeding to autologous or, especially, allogeneic HCT. The risk of pulmonary complications from chemotherapy (e.g., BCNU) or from chronic GVHD involving the lung is potentiated by smoking, as is the risk of secondary cancers of the lung and other organs. A special situation deserving consideration is the use of marijuana, which is increasingly legal, available, and widely used by cancer patients to combat chemotherapy-related nausea and anorexia. Anecdotally, an increasing number of patients referred for transplantation consultation are actively using marijuana to control these symptoms during their pretransplantation chemotherapy. Cases of severe or fatal pulmonary aspergillosis from inhaled spores have been reported in immunosuppressed patients using marijuana.^219,220 Transplant center policies regarding medical marijuana use and abstinence have lagged behind the rapidly changing legal status of marijuana in the United States, creating a challenge in assessing and counseling patients.

Numerous malignant and nonmalignant hematologic disorders, as well as selected solid tumors, may be treated with HCT. The results obtained with transplantation are reviewed in detail in the disease-specific chapters of this book, and are discussed only briefly here.

In general terms, autologous HCT is recommended for patients whose malignancy exhibits chemosensitivity to conventional dose therapy and does not extensively involve the marrow; included are most lymphomas, germ cell tumors, and other selected pediatric tumors. In these instances, tumor eradication is a result of dose escalation of cytotoxic therapy in the conditioning regimen, and the autograft serves as hematopoietic cell rescue. In contrast, allogeneic transplantation is generally pursued for hematologic malignancies and disorders that primarily originate in the marrow, such as acute and chronic leukemias, aplastic anemia, MDSs, and myeloproliferative neoplasms. For some diseases with extensive marrow involvement, such as the low-grade lymphomas and myeloma, the decision to pursue autologous versus allogeneic HCT is more complex. In these settings, allogeneic transplantation has generally been more successful in controlling disease recurrence and reducing relapse risk. However, the associated risks, including GVHD and prolonged immunosuppression, result in a higher TRM compared to autologous HCT. Thus, the decision to pursue an allogeneic or autologous HCT for patients with these diseases depends on the combination of patient characteristics such as comorbidities and age, availability of a suitable donor, disease-specific characteristics, and often patient preference. For some hematologic conditions, such as MDSs, myeloproliferative neoplasms, and aplastic anemia, only allogeneic transplants are generally appropriate.

In addition, patients with selected solid tumors, such as testicular cancer, neuroblastoma, and other pediatric tumors, have had successful outcomes with autologous HCT.^{221,222,223,224} Extensive studies in women with breast and ovarian carcinoma, and more limited studies in patients with renal cell carcinoma and small cell lung cancer, have failed to demonstrate a role for HCT.^225,226 Outside of the investigational setting, there are no currently accepted indications for allogeneic HCT to treat nonhematologic solid tumors.

A variety of congenital and acquired nonmalignant disorders can be successfully treated with HCT. The most well-established nonmalignant indication is for allogeneic HCT in patients with severe aplastic anemia, where outstanding results have been achieved, particularly for younger patients with HLA-matched sibling donors, where long-term disease-free survival rates of 88 to 100 percent have been reported.^227,228 Hematopoietic cell transplantation for patients with clinically significant hemoglobin disorders, such as thalassemia major, has been very successful, especially in patients without significant liver disease.^229,230 Likewise, allogeneic HCT is considered a treatment option for young patients with severe forms of sickle cell disease.^231,232 Guidelines for patient selection and management of patients with thalassemia or sickle cell disease were published in 2014 by EBMT.²³³ In patients with hemoglobin disorders, transplantation serves as a form of gene therapy, using allogeneic hematopoietic cells as vectors for genes essential for normal hematopoiesis. Eventually the vector may well be autologous stem cells transformed by the insertion of normal genes.²³⁴

For patients with SCID syndrome and other congenital immunodeficiencies, allogeneic HCT remains the treatment of choice.^235,236 The role of allogeneic HCT for patients with storage diseases, a diverse group of disorders that typically involve a single gene defect in a lysosomal hydrolytic enzyme or peroxisomal function, is evolving and it appears that subsets of mucopolysaccharidoses derive the most benefit.²³⁷

SELECTED RESULTS OF HEMATOPOIETIC CELL TRANSPLANTATION

A comprehensive discussion of transplantation outcomes is beyond the scope of this chapter; please refer to other disease-focused chapters of this text for more complete information. A brief overview of transplantation-related outcomes in a number of diseases is presented here.

Acute Myeloid Leukemia

Allogeneic HCT has a major role in the management of patients with AML. For patients in CR1, the decision to proceed to allogeneic HCT as opposed to chemotherapy-based consolidation is predicated on prognostic markers and donor availability, as well as patient preference. For patients beyond CR1, allogeneic HCT often offers the only potentially curative treatment option.

First Complete Remission

A question of significant practical importance is how best to treat a younger AML patient who achieves a CR1 following induction chemotherapy. Several large prospective trials have been conducted using so-called genetic randomization: that is, patients in CR1 with an HLA-identical sibling received allogeneic HCT, while those without an HLA-identical sibling received chemotherapy-based consolidation or autologous HCT.^238,239 Two meta-analyses have focused on the comparative outcomes of allogeneic HCT versus chemotherapy, both of which found that allogeneic HCT in CR1 yielded better overall and disease-free survival for patients with intermediate- and poor-risk cytogenetics, but not for those with favorable-risk cytogenetics.^240,241 For patients with poor-risk cytogenetics, there is little dispute that allogeneic HCT is the preferred postremission therapy for medically fit patients with available donors. Conversely, patients younger than 60 years of age with favorable-risk cytogenetics who promptly achieve CR1 with induction chemotherapy are generally not considered for allogeneic HCT, and instead should receive chemotherapy-based consolidation. For younger patients with intermediate-risk cytogenetics, allogeneic HCT in CR1 from an HLA-identical sibling donor is the best available postremission therapy, but its advantage over other approaches is not as substantial as in the setting of poor-risk cytogenetics. One caveat is that the studies demonstrating the superiority of allogeneic HCT were performed using genetic randomization, and thus the results are, strictly speaking, only applicable to patients with HLA-identical sibling donors. That said, single-center and registry data suggest that outcomes with HLA-matched unrelated donor allotransplantation for AML in CR1 are similar to those with HLA-identical sibling donors,^242,243 so it is reasonable to extrapolate the results of the genetic-randomization studies to patients with HLA-matched unrelated donors.

Patients older than 60 years of age typically have substantially higher relapse rates with chemotherapy alone (60 to 80+ percent)^244,245 and have poorer outcomes than younger patients with equivalent cytogenetics, suggesting that older adults may benefit from allogeneic HCT in CR1. There are no prospective, genetically randomized trials in older adults with AML comparing allogeneic HCT to chemotherapy in CR1, but retrospective comparisons suggest that allogeneic HCT can reduce relapse risk and improve outcomes in this demographic.^244,245 The decision to proceed to allogeneic HCT in older adults is often based less on disease risk factors and more on medical comorbidities and the estimated risk of TRM, which may be prohibitive.

In the genomic era of AML risk stratification, patients with intermediate-risk disease are often categorized based on the results of molecular testing for mutations of FLT3, NPM1, and CEBPA. Patients with intermediate-risk cytogenetics and biallelic CEBPA mutations or NPM1 mutations in the absence of FLT3 mutations have good-risk disease and are often not considered for allogeneic HCT in CR1.²⁴⁶ In contrast, patients with internal tandem duplications in FLT3 have poorer prognoses and are often considered for allogeneic HCT in CR1.

Detailed guidelines have been published by the European LeukemiaNet AML Working Party to guide the use of allogeneic HCT in AML patients in CR.²⁴⁷ Based on existing data, many experts would recommend allogeneic HCT for all medically fit patients younger than 60 years old with AML in CR1 except for those with favorable-risk disease (including those with intermediate-risk cytogenetics and favorable molecular markers) who achieve CR1 with their first cycle of induction and have no evidence of minimal residual disease. For all other younger CR1 patients—including those with intermediate-risk cytogenetics and negative molecular markers, and those in any cytogenetic risk group who do not promptly achieve CR with induction or who have minimal residual disease—allogeneic HCT should be strongly considered as the best postremission option.²⁴⁸ For older adults with AML in CR1, allogeneic HCT should be considered for all medically fit patients except those with favorable-risk disease, although there are fewer data to support this recommendation and medical comorbidities play an increasingly important role in decision making as patients age.

The role of consolidation chemotherapy before allotransplantation in patients in CR1 is unclear. Many transplant physicians recommend at least 1 to 2 cycles of consolidation, especially for patients slated to receive RIC, in order to maximize pretransplantation cytoreduction. However, two recent retrospective studies have questioned the benefit of consolidation chemotherapy, and instead favored moving forward to allogeneic HCT as soon as a suitable donor is available.^249,250 Consolidation chemotherapy is to maintain remission during prolonged donor searches, but its benefit is less clear in patients who have a donor identified.

Aml Beyond First Complete Remission

For patients with AML who relapse after attaining a CR1, allogeneic HCT is the treatment of choice. While no prospective randomized trials have compared allogeneic HCT to salvage chemotherapy alone in this setting, retrospective data strongly support the use of allogeneic HCT.²⁵¹ Autologous HCT has been used historically in patients who lacked suitable allogeneic donors,²⁵² but with the advent of improved alternative-donor options the use of autologous HCT for AML has become rare except in developing countries without the capability to perform allogeneic HCT. Patients with AML beyond CR1 who lack suitably HLA-matched donors are candidates for HLA-haploidentical or UCB allotransplantation, as retrospective studies have associated these approaches with equivalent overall and disease-free survival compared to allotransplantation from conventional donor sources.^253,254

The likelihood of long-term survival is very low in AML patients whose disease fails to achieve CR1 following induction therapy (primary induction failure) and in those whose disease is chemorefractory at relapse. Allogeneic HCT with myeloablative conditioning has been reported to cure approximately 19 to 30 percent of such patients in retrospective analyses.^192,255 These retrospective reports undoubtedly suffer from substantial selection bias; presumably patients with relapsed or refractory AML were not transplanted indiscriminately, but only if their physicians felt they had an above-average chance of responding. These success rates are therefore very unlikely to pertain to the overall population of adults with relapsed or refractory AML. Attempts have been made to construct prognostic tools to identify the subset of patients most likely to benefit from allogeneic HCT,¹⁹² and some authorities have argued that allogeneic HCT is underused in this population.²⁴⁸ Currently there is no single standard of care for treatment of patients with AML in refractory relapse or primary induction failure; reasonable options include additional salvage chemotherapy to induce remission; enrollment on a clinical trial; allogeneic HCT (in carefully selected patients); and palliative care, and the optimal choice must be individualized based on patient characteristics and donor availability.

Acute Lymphoblastic Leukemia

Allogeneic HCT is widely used in adult patients with ALL, especially those patients with high-risk features (most often defined as white blood cell count at diagnosis >30,000 [or >100,000 in T-cell ALL], adverse cytogenetics, progenitor B-cell immunophenotype, age >60 years, or failure to achieve CR within 4 weeks of induction chemotherapy).²⁵⁶ Patients with none of these features are considered to have standard-risk disease; there is no clear “favorable-risk” population in adult ALL. While the advisability of allogeneic HCT is uncontroversial as postremission therapy for adults with high-risk ALL, it is also the treatment of choice for eligible adults with standard-risk ALL in CR1. Several large genetically randomized prospective trials have demonstrated a survival benefit for allogeneic HCT in CR1 for adults with standard-risk ALL.^257,258 A 2013 meta-analysis confirmed the benefit of upfront allogeneic HCT for adults 35 years of age or younger; for older adults, the benefit was more difficult to demonstrate owing to increasing TRM.²⁵⁹ As with all genetically randomized clinical trial data, these findings apply only to patients with HLA-identical sibling donors, strictly speaking, although many authorities extrapolate the observed benefit to HLA-matched unrelated donors and possibly even alternative donors on the basis of generally equivalent outcomes.

On the basis of existing data, we believe that myeloablative allogeneic HCT is the optimal therapy for all eligible adults with ALL in CR1 and available HLA-identical sibling donors—a recommendation supported by a recent Cochrane Library systematic review.²⁶⁰ Allogeneic HCT is likely also the optimal therapy for those with HLA-matched unrelated donors, although this recommendation relies more on extrapolation from data in the HLA-identical sibling setting. The decision to proceed with alternative-donor allogeneic HCT in a patient with standard-risk ALL in CR1 is less clear-cut and depends on patient preference and institutional comfort level. In any case, we believe that all adults with ALL should be referred for transplantation consultation early in their course to assist in determining the optimal treatment approach.

Conditioning-regimen intensity appears to play a major role in the success of allogeneic HCT for ALL. Results with nonmyeloablative allogeneic HCT for Philadelphia chromosome–negative (Ph–) ALL beyond CR1 have been poor, with virtually no survivors.²⁶¹ For older adults with Ph– ALL who are not candidates for intensive conditioning, the optimal approach remains unclear, particularly for those beyond CR1.

Philadelphia Chromosome–Positive Acute Lymphoblastic Leukemia

The postremission management of patients with Philadelphia chromosome–positive (Ph+) ALL deserves special attention. Historically, Ph+ ALL carried a very poor prognosis with conventional chemotherapy, and this chromosomal abnormality was considered a high-risk feature. High-dose TBI-based conditioning and allotransplant has historically been the standard of care for postremission therapy of Ph+ ALL in adults, and resulted in 10-year overall survivals of 54 percent and 29 percent for patients transplanted in CR1 and beyond CR1, respectively, in the pre–tyrosine kinase inhibitor (TKI) era.²⁶²

The development of TKIs targeting bcr/abl has markedly improved the prognosis of Ph+ ALL. With the incorporation of TKI therapy into pretransplant induction and posttransplant maintenance, Ph+ ALL now carries one of the best prognoses of all forms of adult ALL, even with nonmyeloablative conditioning.^261,263 Most authorities continue to recommend allogeneic HCT as postremission therapy for adults with ALL in the TKI era. The pediatric literature suggests that it may be safe to forgo allotransplantation in favor of TKI-based chemotherapy alone in Ph+ ALL,²⁶⁴ but these data cannot be extrapolated to the adult setting because of the very different clinical behavior of adult versus pediatric ALL. For adults with Ph+ ALL who are not candidates for allotransplantation, TKI-containing chemotherapy regimens coupled with close monitoring of minimal residual disease may be an alternative, but this approach has not been tested in clinical trials.

Myeloma

Autologous HCT is not curative in myeloma, but a number of randomized clinical trials have demonstrated prolongation of event-free survival by a median of approximately 1 year when single or double autologous HCT is incorporated into the treatment program.^265,266,267 Some of these trials also reported improved overall survival with autologous HCT. In addition, early autologous HCT is associated with an improved quality of life in patients with myeloma, likely because it is associated with a shorter period of chemotherapy.²⁶⁸ However, these studies were performed before the introduction of modern highly active antimyeloma therapies such as thalidomide, lenalidomide, and bortezomib. In 2014, results were published from a randomized trial comparing upfront tandem autologous HCT to lenalidomide-based maintenance therapy after induction with lenalidomide and dexamethasone.²⁶⁹ Patients randomized to upfront autologous HCT had significantly improved 4-year overall survival compared to those randomized to the nontransplant arm (81.6 percent vs. 65.3 percent, p = 0.02; Fig. 23–1).²⁶⁹ Patients randomized to the nontransplantation arm remained eligible for autologous HCT at the time of myeloma progression, but only 62.8 percent actually received autologous HCT, suggesting that delayed autologous HCT is not feasible for many patients as a result of rapid clinical deterioration following relapse. These results support the continued role of upfront autologous HCT as the standard of care for medically fit patients with myeloma in the modern era, although several caveats apply. First, the maintenance regimen in the nontransplantation arm consisted of low-dose melphalan in combination with lenalidomide and dexamethasone, and it is unclear whether these findings can be generalized to other, more commonly used maintenance regimens. Second, patients randomized to upfront autologous HCT received planned tandem autologous HCT with intravenous melphalan 200 mg/m² conditioning for each transplant. This approach is relatively uncommon, as tandem autologous HCT has fallen out of favor based on evidence that it does not improve overall survival.²⁷⁰ In current practice, second autologous HCT is usually reserved for patients who fail to achieve at least a very good partial remission after first autologous HCT,²⁷¹ and is often performed using a lower dose of intravenous melphalan (140 mg/m²) as conditioning. As a result, it is unclear whether the benefit observed in this trial extends to the setting of single autologous HCT. Nonetheless, upfront autologous HCT continues to be widely used in myeloma, and the recent trial adds support for this approach, particularly since nearly 40 percent of patients intended to undergo transplantation never received the intervention.

Allogeneic HCT remains the only therapy thought to be potentially curative in myeloma, and has been widely studied, most often as part of a tandem autologous/allogeneic double-transplant approach. Comparisons of tandem autologous/autologous versus tandem autologous/RIC allogeneic HCT have yielded conflicting results. Improved overall survival with allotransplantation was seen in an Italian randomized study of 162 patients and in long-term followup of the EBMT-NMAM2000 trial involving 257 patients,^272,273,274 although the relapse rate remained high at 60 percent even after allografting.²⁷³ In contrast, three other large randomized trials failed to demonstrate an overall survival advantage with allografting.^275,276,277 The largest trial addressing this question was completed by the BMT-CTN and enrolled 710 myeloma patients randomized between auto-auto and auto-allo approaches.²⁷⁸ This trial found no benefit in overall or progression-free survival with allografting. These results, along with a 2013 meta-analysis of randomized trials which similarly found no benefit to allotransplantation,²⁷⁹ have led to a loss of enthusiasm for allogeneic HCT in myeloma. Some authors have argued that survival improvements with allogeneic HCT require longer followup.²⁷³ Nonetheless, given existing data, allogeneic HCT in myeloma is typically reserved for younger patients with very high-risk disease or those who have failed autologous HCT, and is ideally performed in the context of a clinical trial.

Non-Hodgkin Lymphoma and Hodgkin Lymphoma

Patients with chemosensitive aggressive and highly aggressive lymphomas beyond CR1 have an improved overall survival with high-dose therapy followed by autologous HCT compared to best-of-care salvage chemotherapy.^280,281 The benefit of autologous HCT appears restricted to patients with chemosensitive disease. Patients with negative 18-fluorodeoxyglucose (FDG)-PET imaging prior to transplantation have significantly better outcomes than patients with residual FDG-avid disease at the time of transplant.^197,282,283 Historically, autologous HCT produced long-term disease-free survival in approximately 50 percent of patients with chemosensitive relapsed lymphoma.²⁸⁴ More recent studies have paradoxically demonstrated poorer cure rates (approximately 20 percent) with autologous HCT for relapsed B-cell NHL in the rituximab era.²⁸⁵ One possibility is that rituximab-containing first-line chemotherapy is successful in curing a greater number of patients, but also selects for patients with particularly resistant disease in the setting of relapse.

Autologous HCT is sometimes used as consolidation for high-risk lymphomas in CR1, especially in mantle cell lymphoma, where autologous HCT in CR1 has been shown to extend progression-free survival in a phase III randomized trial.²⁸⁶ Autologous HCT in mantle cell lymphoma is most effective when performed early in the disease course,^287,288 and thus eligible patients with mantle cell lymphoma should be referred for transplant consultation during their induction course to discuss the risks and benefits of consolidative autologous HCT. Outside the setting of mantle cell lymphoma, there are few B-cell lymphomas where data clearly support consolidative autologous HCT in CR1. Nevertheless, patients with so-called “double-hit” lymphomas (those mutated at both c-MYC and either bcl-2 or bcl-6) are viewed as particularly high-risk and sometimes treated with autologous HCT in CR1. However, two recent studies have questioned the need for this approach; in both, patients with double-hit lymphomas who achieved PET-negative CR with induction therapy had excellent overall survival (75 to 83 percent), and autologous HCT in CR1 was not considered beneficial.^289,290 Autologous HCT is also often advocated as consolidation for patients with T-cell lymphomas in CR1, with the exception of anaplastic lymphoma kinase-positive (ALK+) large cell lymphoma. Definitive data supporting this practice are not available, but this approach appears reasonable in view of the very high relapse rate associated with these lymphomas.

Allogeneic HCT is highly active and potentially curative for NHL, particularly indolent NHL. Excellent disease-free survival has been reported with both early and late allogeneic HCT in indolent NHL, even in heavily pretreated patients and those with active disease at allotransplantation.^179,181 The optimal timing of allogeneic HCT in indolent NHL remains to be determined, particularly as indolent NHL can often be treated effectively with conventional chemotherapy for years or even decades. For more aggressive relapsed NHL or for relapsed HL after failed autografting, allogeneic HCT remains the only potentially curative salvage option. However, with these more aggressive malignancies, it is essential that patients be chemoresponsive and, ideally, in a PET-negative CR in order to proceed to allotransplantation, as otherwise relapse is nearly universal before GVT effects can become established.^291,292

Table 23-3 lists the complications associated with HCT; the most common are discussed below. The first 100 days following transplantation are the time of greatest risk for recipients of autologous and allogeneic HCT. Care by physicians skilled in the management of patients undergoing these procedures is of critical importance.

GRAFT FAILURE

Graft failure is defined as the lack of donor hematopoietic cell engraftment following autologous and allogeneic HCT. Criteria are predominantly operational, and graft failure is divided into primary (early) and secondary (late) phases. The consequences of graft failure are significant, and include high risks of death from infection, hemorrhage, or relapsed malignancy.

Primary (Early) and Secondary (Late) Graft Failure

Myeloid engraftment has commonly been defined as the first of three consecutive days on which the absolute neutrophil count exceeds 5 × 10⁸/L. Myeloid engraftment typically occurs within 21 days of the graft infusion, irrespective of graft source. Platelet recovery is more variably defined, often as the first day of a platelet count of at least 20, 50, or 100 × 10⁹/L, sustained without transfusion for 7 days. Platelet recovery may be substantially delayed compared with myeloid recovery, particularly with lower CD34+ cell doses or UCB allografts. A hemoglobin level of at least 8 g/dL without transfusion support is an accepted threshold for red cell engraftment. These criteria were derived from the predictable kinetics seen after myeloablative conditioning. With RIC, many patients never develop severe cytopenias, and thus engraftment in these settings is usually defined, at least in part, by assessment of donor chimerism in the blood or marrow. In these settings, graft loss is typically defined by donor chimerism of less than 5 percent in blood CD3+ cells.

Primary graft failure is defined as failure to achieve these threshold counts or donor chimerism levels at any point beyond day +28. Isolated cytopenias does not necessarily herald graft failure, as they may be transitory phenomena related to infection, medications, lineage-specific immune-mediated cytopenias, or GVHD. Secondary (late) graft failure occurs in patients who initially meet criteria for engraftment but subsequently lose graft function in at least two cell lines. Late graft failure is more often associated with allogeneic HCT than with autologous transplantation; possible causes include graft rejection related to residual host immunity, persistent or progressive malignancy, low donor cell yield, medication side effects, infection, or GVHD.

Graft Rejection and Poor Graft Function

Graft rejection is a subset of primary or secondary graft failure caused by immune-mediated rejection of donor cells by residual host effector cells. A diagnosis of graft rejection requires analysis of blood or marrow for donor hematopoietic chimerism; graft rejection is defined as the inability to detect a meaningful percentage (usually >5 percent) of donor hematopoietic elements. In contrast, poor graft function describes the failure to achieve adequate blood counts following allogeneic HCT in the presence of substantive donor hematopoietic cell chimerism.

Graft Failure Following Reduced-Intensity Conditioning

Allogeneic HCT following RIC is associated with incomplete eradication of host hematopoiesis. As a consequence, a significant percentage of patients have mixed donor/host hematopoietic chimerism for several months after transplantation before converting to complete donor type.^186,293 Primary engraftment following reduced-intensity allogeneic HCT is defined by neutrophil, platelet, and hemoglobin count recovery as outlined above as well as stable donor T-cell chimerism. As described above, graft failure is said to have occurred when blood donor T-cell chimerism is less than 5 percent at any point after reduced-intensity allogeneic HCT. Donor T-cell chimerism levels greater than 5 percent but less than 95 percent are generally termed “mixed chimerism,” while full donor chimerism is defined by blood donor T-cell chimerism of 95 percent or greater.

Incidence of Graft Failure

The incidence of graft failure varies widely in published reports. To estimate the incidence of graft failure following autologous HCT, consider that in most centers the TRM associated with autologous HCT is less than 5 percent, of which only a small subset can be attributed to graft failure. Another surrogate marker for estimating the incidence of graft failure following autologous HCT is the requirement for hematopoietic cell rescue using a backup autograft product. A study of 300 patients who underwent autologous HCT revealed that 4.7 percent required their backup product.²⁹⁴ Thus, it is reasonable to estimate that the incidence of graft failure following autologous HCT is somewhere between 1 and 5 percent.

Graft failure following allogeneic HCT is more complex, because of confounding factors such as histocompatibility, ABO matching, graft-versus-host and host-versus-graft reactions, and the use of postgrafting immunosuppression. The overall incidence of graft failure after allogeneic HCT is approximately 5 to 6 percent.²⁹⁵ In general, graft failure is uncommon after high-dose conditioning and in patients who are heavily pretreated with cytotoxic chemotherapy before coming to allogeneic HCT. Even in the myeloablative setting, though, the incidence of graft failure varies with conditioning regimen, as illustrated in a randomized trial where graft failure occurred in zero of 64 (0 percent) of patients receiving BU/CY but in five of 62 (8 percent) of patients receiving BU/FLU.¹⁷⁴ The risk of graft rejection is highest in patients who are heavily presensitized or who have autoimmunity directed at hematopoietic cells (as in aplastic anemia), those who receive low CD34+ cell doses,^295,296 and those with diseases such as myelofibrosis where the marrow microenvironment is significantly perturbed.

The consequences of graft failure, and its optimal treatment, depend in large part upon the likelihood of autologous hematopoietic recovery. In patients who have received high-dose conditioning, autologous marrow recovery is likely to be severely delayed if not absent, and graft failure is associated with high mortality rates as a consequence of prolonged cytopenias. Second-salvage allogeneic HCT has been used successfully to treat graft failure in this setting; reported outcomes vary from dismal to encouraging,^297,298 and likely depend substantially on patient selection. There is no consensus on whether to use the same or a different donor for salvage allogeneic HCT for graft rejection, and the decision often depends on donor availability. The time needed to identify and collect a second allograft product are often prohibitive for patients with graft rejection and pancytopenia, and thus readily available HSC sources such as UCB and HLA-haploidentical family members have sometimes been used.

For patients with graft failure after RIC, autologous hematopoietic recovery is more likely. For these patients, the optimal strategy often involves withdrawing postgrafting immunosuppression and awaiting autologous count recovery. However, for patients with malignant disease, the risk of relapse is substantially elevated in the setting of graft failure,²⁹⁵ presumably as a result of a loss of GVT effects.

REGIMEN-RELATED ORGAN TOXICITIES

The severity of organ toxicities associated with HCT is a function of the intensity of conditioning therapy, the amount of prior therapy received, patient comorbidities before transplantation, and posttransplantation factors such as immunosuppressive medication and antimicrobial agents.

Mucositis

Mucositis occurs in more than 90 percent of patients receiving high-dose regimens and is often regarded as the most difficult issue from the patient’s perspective.²⁹⁹ Current management is supportive and includes frequent rinsing with saline solutions and antimicrobials, cryotherapy, pain control (often with continuous intravenous infusions of opioids), and parenteral nutrition when needed. Improvement typically occurs within 10 to 21 days of transplantation, around the time of engraftment. Fully ablative regimens, TBI-based conditioning, and the administration of posttransplantation methotrexate (MTX) for GVHD prevention are associated with more severe mucositis.³⁰⁰ Severe mucositis may result in significant tissue edema, upper airway obstruction, and/or aspiration pneumonitis, although fortunately these complications are rare. Regimen-related gastroenteritis results in nausea, vomiting, and diarrhea, which may persist for weeks after the transplant. Breaches in the mucosal lining predispose to bacterial translocation from the gastrointestinal tract, with increased risk of bacteremia and sepsis. Palifermin (keratinocyte growth factor) can reduce patient-controlled anesthesia (PCA) and total parenteral nutrition (TPN) usage because of mucositis, predominantly in patients receiving TBI-based conditioning,³⁰¹ but at a cost of $5500 to $14,000 per day.³⁰² Initial hopes that this agent would prevent GVHD after allogeneic HCT have not been realized,^303,304 and the benefits of palifermin appear limited to ameliorating mucositis.

Sinusoidal Obstructive Syndrome

Sinusoidal obstructive syndrome (SOS) is a clinical syndrome of regimen-related hepatotoxicity characterized by tender hepatomegaly, fluid retention, weight gain, and elevated serum bilirubin following autologous or allogeneic HCT. This syndrome was formerly called venoocclusive disease (VOD), but this term is no longer used as it inaccurately describes the underlying pathobiology: the liver injury is initiated by damage to hepatic sinusoidal epithelium, and obstruction of hepatic venules is not essential to the development of the syndrome.³⁰⁵ The incidence of SOS varies significantly with the intensity of conditioning and with the stringency of diagnostic criteria, from less than 10 percent to as high as 30 to 40 percent. CY is a key culprit in the development of SOS, and large interpatient variability in CY metabolism may account for the syndrome’s unpredictability.^306,307 Other contributing factors include the preadministration of BU in the BU/CY regimen, which potentiates CY hepatotoxicity¹⁷²; preexisting hepatic fibrosis, as in patients with cirrhosis or with hepatic extramedullary hematopoiesis as in myelofibrosis³⁰⁸; and pretreatment with higher doses of gemtuzumab ozogamicin.^309,310 The incidence of SOS appears to be decreasing over time,³¹¹ likely a result of the prevalence of RIC regimens and the prophylactic use of ursodiol, which prevents SOS and other forms of hepatic injury during allogeneic HCT.^312,313

SOS is generally is classified as mild (clinically apparent yet resolves without treatment), moderate (requiring diuretics and pain medication for abdominal discomfort yet completely resolves before day +100), or severe (not resolving before day +100 or death).³¹⁴ Severe SOS has a tendency to progress to multiorgan failure and is associated with a mortality rate of greater than 80 percent.³¹⁵ Therapy for SOS is supportive and includes management of sodium and water balance with diuretics, preservation of renal blood flow, and paracentesis for ascites associated with significant discomfort or pulmonary compromise. Patients with a poor prognosis can be recognized early after SOS onset by steep rises in serum bilirubin, body weight, and other liver enzymes; hepatic venous pressure measurement greater than 20 torr; development of portal vein thrombosis; and multiorgan failure requiring mechanical ventilation or renal dialysis.³¹⁶

There are few satisfactory therapies for severe SOS; the most commonly used is intravenous defibrotide, a mixture of single-stranded porcine oligodeoxyribonucleotides which induces antithrombotic and fibrinolytic effects in preclinical models.³¹⁷ Its mechanism of action against SOS remains unknown. A randomized phase II dose-finding trial involving 149 patients with severe SOS reported day +100 survival of 42 percent with few adverse events, and a dose of 25 mg/kg/day was selected for ongoing randomized phase III trials.³¹⁸ However, defibrotide is not yet approved by the FDA and remains available only on an investigational or compassionate-use basis in the United States.

Pulmonary Complications

Noncardiogenic and noninfectious diffuse lung injury, also referred to as idiopathic pneumonia syndrome (IPS), remains a significant problem following autologous or allogeneic HCT, occurring in 10 to 15 percent of transplant recipients.³¹⁹ Risk factors for idiopathic IPS include high-dose conditioning, TBI, GVHD, older recipient age, prior history of cigarette smoking, prior thoracic/mediastinal irradiation, and abnormal gas exchange as measured by pretransplant pulmonary function testing.³²⁰ Preclinical models suggest that donor T cells play a key role in the development of IPS, indicating that it may be a form of graft-versus-host reaction.³¹⁹ In a small subset of patients, diffuse alveolar hemorrhage (DAH) develops, characterized by progressive shortness of breath, cough, and hypoxemia. Classically, DAH is defined by the demonstration of progressively bloodier aliquots in bronchoalveolar lavage fluid. Mortality from this complication is high (often >75 percent) despite aggressive treatment. Another subset of patients with IPS develop periengraftment respiratory distress without a bloody bronchoalveolar lavage. In the autologous setting, the IPS often responds promptly to glucocorticoids, whereas in the allogeneic setting response rates are lower, indicating that perhaps some cases may be complicated by GVHD.

The management of suspected IPS begins with bronchoscopy to rule out infectious etiologies and to evaluate for DAH. Care is supportive and aimed at maximizing respiratory function and preventing volume overload or multiorgan failure. Patients are typically treated with high-dose glucocorticoids (methylprednisolone at 2 mg/kg/day or higher), along with broad-spectrum antimicrobials and intensive supportive care. Retrospective studies have suggested that tumor necrosis factor (TNF) blockade with etanercept may be effective as an adjunct to high-dose glucocorticoids,³²¹ but a prospective randomized study failed to find an additive benefit,³²² and thus etanercept cannot be routinely recommended at this time. Patients with IPS who progress to require mechanical ventilation have a very poor prognosis, and a frank discussion of the goals of care is indicated in this setting, particularly in the presence of multiorgan failure.³²³

Lung inflammation following the administration of BCNU is a separate form of noninfectious lung injury seen in HCT recipients who receive this agent as part of their conditioning regimens. BCNU-induced pneumonitis is often characterized by a nonproductive cough with increasing dyspnea and bilateral pulmonary infiltrates on chest radiography, with or without fevers, and often occurs 1 to 2 months after transplantation.³²⁴ Pulmonary function tests reveal a restrictive pattern of lung injury and a decrease in diffusing capacity compared to pretransplant values. Prompt treatment with glucocorticoids reduces mortality and morbidity and is crucial to a successful outcome. If untreated or recognized late, significant pulmonary fibrosis may develop.

INFECTIONS

Susceptibility to infection is a significant challenge in the clinical management of transplant recipients. The essential principles are prevention, judicious monitoring, and expeditious treatment of all bacterial, fungal, and viral infections. These basic principles are widely accepted, yet the day-to-day strategy for implementing them varies widely among transplant physicians and centers. Two universally important measures for reducing infections in immunocompromised transplant recipients are effective handwashing policies and a strategy for preventing transmission of respiratory viruses, including metapneumovirus, respiratory syncytial virus, parainfluenza, and influenza.

The duration of neutropenia and severity of oral and gastrointestinal mucosal damage from the conditioning regimen are risk factors for infection before neutrophil recovery has occurred. Following neutrophil recovery, deficiencies of B- and T-cell–mediated immunity persist and increase susceptibility to opportunistic infections. Immune recovery following autologous HCT is relatively rapid compared to after allogeneic transplantation. Most autologous transplant recipients recover T-cell immunity specific for herpes viruses, including CMV, by 3 months after transplantation.³²⁵ The degree and duration of immunodeficiency following allogeneic HCT are influenced, in part, by the type of immunosuppressive therapy and severity of GVHD. Chronic GVHD is associated with chronic B- and T-cell immune deficiencies that may persist for years, and Ig production and reticuloendothelial function may also be impaired.^326,327,328

Bacterial Infections

Bacterial infections are common during the period of immediate neutropenia that follows the preparative transplantation regimen, and can be caused by both Gram-positive and Gram-negative organisms.³²⁹ The increased risk of bacterial infections is not only caused by neutropenia, but also from loss of epithelial integrity from regimen-related injury, bacterial translocation, and the presence of indwelling intravenous catheters. Some centers institute preventive measures in addition to rigorous hand washing, such as gowning and masking, although there is little evidence that these actions reduce infection risk. Removal of venous catheters is sometimes required for patients who do not respond promptly to treatment. Chapter 24 reviews specific strategies and regimens for treating bacterial infections in neutropenic patients.

Patients who require ongoing immunosuppressive therapy for the control of chronic GVHD are at risk for recurrent bacteremia with encapsulated bacteria and sinopulmonary infections. Preventive strategies differ from institution to institution, although some form of antibiotic prophylaxis is often used in these patients. Infrequent bacterial pathogens which should also be considered, especially in the presence of pulmonary infiltrates or nodules, are Legionella, Nocardia, Mycobacterium tuberculosis, and atypical mycobacteria.

Fungal Infections

Fungal infections are serious and potentially fatal complications following HCT, and are seen most commonly in allogeneic HCT recipients as a result of the requirement for postgrafting immunosuppressive medication. The incidence of fungal infection varies considerably among transplantation centers because of a variety of factors, including geographic location, nearby construction, and prophylactic regimens. Candida and Aspergillus are the most common fungal pathogens; however, other organisms can also cause life-threatening infections. Chapter 24 discusses the treatment and prophylaxis of fungal infections.

Fluconazole prophylaxis decreases the incidence of invasive and superficial Candida albicans infections and may decrease the 100-day mortality in allogeneic HCT recipients.³³⁰ Fluconazole has limited activity against Candida krusei, Torulopsis glabrata, and Aspergillus species, and some centers reported an increased incidence of resistant Candida infections in patients receiving prophylactic fluconazole.³³¹ More aggressive prophylaxis with mold-active agents such as voriconazole or itraconazole can prevent invasive mold infections, including aspergillosis, but these agents are associated with a higher risk of adverse events and no clear benefit to mortality.^332,333

Viral Infections

Infection with herpesviruses can cause significant morbidity and mortality in HCT recipients. Most infections are a result of viral reactivation and follow a relatively predictable temporal pattern in the absence of prophylaxis: Herpes simplex virus (HSV) causes clinically apparent disease at approximately 2 to 3 weeks after HCT, CMV disease usually occurs during the second to third month, and varicella-zoster virus (VZV) recurrences present at a median of 5 months after HCT.³³⁴

CMV is an important viral pathogen in HCT recipients. Infection occurs from reactivation of latent virus or is newly acquired from the donor graft or blood transfusions. Reactivation is a common problem in allogeneic HCT, but is relatively rare after autologous HCT except in the setting of CD34+ selection and T-cell depletion.¹²³ Before effective prevention strategies were introduced, CMV infection developed in 70 percent of CMV-seropositive transplant recipients and 32 percent of CMV-seronegative recipients and was a frequent cause of TRM.³³⁵ Currently, all allogeneic HCT recipients at risk for CMV infection (those who are CMV-seropositive or who have a CMV-seropositive donor) require monitoring with preemptive therapy for CMV reactivation; this approach has markedly reduced the risk of progression to frank CMV disease such as enteritis or pneumonitis.³³⁶ Prophylactic antiviral therapy against CMV is used by some centers, particularly in high-risk situations such as UCB recipients. More commonly, allogeneic HCT recipients are monitored with plasma CMV polymerase chain reaction (PCR) assays at least weekly through at least day +100 after HCT, and are treated preemptively if these studies yield values which exceed institutionally established thresholds.³³⁴ The preemptive approach avoids the toxicity of universal antiviral prophylaxis, reduces the risk of acquired antiviral resistance, and limits overtreatment of CMV while preventing development of tissue disease. However, CMV tissue disease such as enteritis or pneumonitis can, rarely, develop despite negative plasma CMV PCR, and CMV enteritis or pneumonitis cannot be definitively ruled out by blood tests alone.

Prophylaxis with acyclovir or, more potently, valacyclovir can reduce the risk of CMV reactivation after allogeneic HCT, but does not obviate the need for CMV surveillance.^337,338 Intravenous ganciclovir is typically the first-line treatment for CMV reactivation or tissue disease, although oral valganciclovir may be equally effective as preemptive therapy.^339,340 The most common adverse effect of these antiviral drugs is myelosuppression, which often requires growth-factor support. Intravenous foscarnet is as effective as ganciclovir in CMV prevention and treatment,³⁴¹ and does not cause myelosuppression. However, foscarnet is nephrotoxic and requires cumbersome pre- and posthydration, and is thus most often used as second-line therapy in patients who cannot tolerate ganciclovir because of myelosuppression. Patients with CMV reactivation are typically treated with an induction dose of anti-CMV therapy (ganciclovir, valganciclovir, or foscarnet) for at least 2 weeks, followed by maintenance antiviral therapy until CMV assays are persistently negative. It is not uncommon for patients to experience additional late reactivations after antiviral therapy is discontinued, and these patients require ongoing monitoring beyond day +100.

Investigational approaches to CMV prevention and treatment include CMX-001, an oral prodrug that is converted to cidofovir intracellularly and lacks renal toxicity. A placebo-controlled randomized trial of CMX-001 as CMV prophylaxis in allogeneic HCT found that this agent reduced the incidence of CMV activation from 37 percent to 10 percent.³⁴² Diarrhea was the most common adverse effect; myelosuppression and nephrotoxicity were not observed. Letermovir, a novel terminase inhibitor, has also shown efficacy in preventing CMV reactivation in allogeneic HCT in randomized, placebo-controlled clinical trials.³⁴³ Maribavir, another antiviral agent with a novel mechanism of action, was effective in preventing CMV in a randomized dose-finding study,³⁴⁴ but failed to demonstrate a benefit in a randomized phase III clinical trial.³⁴⁵ Donor-derived CMV-specific cytotoxic lymphocytes have also been used investigationally with some success.³⁴⁶

HSV and VZV are two other members of the herpesvirus family that cause significant morbidity in the posttransplantation setting. These viruses share the characteristics of latency, reactivation, and neurotropism. Virtually all HSV disease occurring after HCT is a result of reactivation, and the serologic status of the transplant recipient determines the risk for disease and the requirement for prophylaxis. Oral mucositis, cutaneous infections, esophagitis, genital herpes, and pneumonia are the most common clinical manifestations. Acyclovir is highly effective for the prevention and treatment of HSV and should be administered to all HSV-seropositive transplant recipients, beginning before or during conditioning. Acyclovir prophylaxis is often continued for at least 1 year after HCT (or longer if patients remain on immunosuppressive therapy), as this approach reduces the risk of late HSV recurrence.³⁴⁷ Acyclovir is well tolerated immediately after transplantation with no effect on the recovery of neutrophil counts. Valacyclovir is an acceptable alternative. Patients do not require concomitant acyclovir prophylaxis while receiving maribavir, foscarnet, valganciclovir, ganciclovir, or cidofovir for treatment of another virus, because these agents have adequate anti-HSV activity. Acyclovir resistance is rare in HSV and can be treated using foscarnet.³⁴⁸

Recurrent VZV disease can occur after both allogeneic and autologous HCT. The initial manifestations of recurrence are localized in approximately half of patients. Treatment with acyclovir within 24 to 48 hours of the onset of herpes zoster prevents dissemination and shortens the course of cutaneous disease. The failure of VZV infections to resolve quickly or their recurrence shortly after acyclovir therapy is discontinued is usually a function of the limited host immune response and is generally not a result of acyclovir resistance. For the rare cases of acyclovir-resistant VZV, foscarnet is the most commonly used alternate therapy.³⁴⁸ Acyclovir, given for at least 1 year after HCT, is highly effective in preventing VZV recurrence.³⁴⁹ Pilot studies of vaccination with live attenuated VZV in allogeneic HCT recipients suggest that this approach is safe and effective in selected patients (median of 4 years posttransplantation, off systemic immunosuppression, blood CD4+ cell count >200/μL),³⁵⁰ but larger studies are required before vaccination can be widely recommended. Heat-inactivated VZV vaccination has also been reported as efficacious in recipients of autologous HCT.³⁵¹

ACUTE GRAFT-VERSUS-HOST DISEASE

Acute GVHD remains one of the most serious and challenging complications of allogeneic HCT. The requirements for the development of acute GVHD were described more than 40 years ago: the graft must contain immunologically competent cells, the recipient must express tissue antigens not found in the donor, and the recipient must be immunologically suppressed such that an effective response against transplanted cells cannot occur.³⁵² HLA disparities are potent triggers of GVHD, but in the setting of HLA-matched donor/recipient pairs GVHD is mediated by minor histocompatibility antigen disparities which provoke a donor T-cell response.^353,354 There are two primary classes of MHC antigens in humans: HLA class I antigens have a broad distribution and are expressed on nearly all cells, whereas HLA class II antigen expression is restricted to macrophages, dendritic cells, B cells, and activated T cells. Minor histocompatibility antigens are endogenous cellular proteins which are subject to significant genetic polymorphism and are presented to donor T cells as small peptides bound in the grooves of the major histocompatibility antigens.³⁵⁵ Some minor histocompatibility antigens associated with GVHD include CD31, HA-1, and the male-specific DBY gene.^356,357 There are likely hundreds if not thousands of minor histocompatibility antigens relevant to allogeneic HCT, of which only a handful have been identified to date. Efforts are underway to utilize genome-wide association studies to broaden our knowledge of these determinants of GVT effects and GVHD.³⁵⁸

The most important risk factor for the development of acute GVHD is the degree of HLA disparity between donor and recipient. The increased incidence of acute GVHD with fully HLA-matched unrelated donors compared to HLA-identical sibling donors is likely related to increased disparity in minor histocompatibility antigens or unrecognized disparities in the phenotypically matched major histocompatibility loci.³⁵⁹ Other risk factors for acute GVHD development include conditioning intensity, use of TBI, and possibly graft source (although the effect of graft source on acute GVHD incidence is not consistently observed).^72,359

Classically, acute GVHD was defined temporally by its occurrence before day +100 after allogeneic HCT. With RIC, acute GVHD can occur beyond day +100, and the distinction between acute and chronic GVHD is now based on organ involvement and histology rather than time of onset.³⁶⁰ Acute GVHD affects the skin, gastrointestinal (GI) tract, and liver (although liver involvement is increasingly rare, for reasons which are not entirely clear).²¹⁸ The overall incidence of acute GVHD after allogeneic HCT is 40 to 60 percent,³⁵⁹ although the incidence may vary widely in specific settings depending on conditioning regimen, HLA matching, and graft source.

Skin involvement manifests as a rash, which may be localized and maculopapular or diffusely erythematous with bullae and desquamation in very severe cases. Definitive diagnosis requires skin biopsy and interpretation by an experienced pathologist.³⁶¹ However, skin biopsies are often inconclusive, and the diagnosis is often made on clinical grounds in patients with a skin rash consistent with acute GVHD arising during the appropriate timeframe after allogeneic HCT. Decision analysis supports the concept that the diagnosis of skin GVHD can be made clinically and does not require skin biopsy in patients with a pretest likelihood of acute GVHD of 30 percent or greater (the majority of allogeneic HCT recipients).³⁶²

GI manifestations of acute GVHD can affect the upper GI tract (presenting as nausea, emesis, anorexia, and weight loss), the lower GI tract (presenting as diarrhea with or without abdominal cramping and hematochezia), or both. Upper-GI involvement with acute GVHD is likely underdiagnosed, since the symptoms may be mild and posttransplant anorexia and nausea are often nonspecific and multifactorial. Diagnosis requires upper endoscopy and endoscopic biopsy, with interpretation by an experienced pathologist. The diagnosis of upper-GI acute GVHD is clinically important, because this syndrome often responds dramatically to even low-dose treatment and, if left untreated, can cause significant nutritional compromise.

Lower-GI involvement with acute GVHD is a more serious and feared complication of allogeneic HCT. These patients present with substantial diarrhea, often several liters per day, accompanied by pain and bleeding. The diarrhea of acute GVHD is secretory, related to epithelial injury, and typically persists around the clock. Abdominal computed tomography (CT) findings, particularly bowel-wall thickening, are common in acute GVHD,³⁶³ but CT findings alone are insufficient for diagnosis. Patients with suspected lower-GI GVHD should undergo endoscopic evaluation and biopsy as soon as feasible, although in the setting of myeloablative conditioning it is often difficult to differentiate regimen-related GI injury from acute GVHD endoscopically or histologically before day +20. Flexible sigmoidoscopy is viewed as a sufficient diagnostic test in most cases,^364,365 and is far easier to perform than full colonoscopy as it does not require an aggressive preparatory regimen. Visual inspection of the gut mucosa is often sufficient to advance a diagnosis of acute GVHD and initiate treatment,³⁶⁶ particularly if the endoscopist has experience in evaluating allogeneic HCT recipients, but biopsies remain essential to rule out CMV enteritis and other etiologies.³⁶⁵ Mucosal denudation is a particularly concerning endoscopic finding,³⁶⁶ and is associated with a grim prognosis.

Hepatic GVHD presents with abnormal liver function tests, most commonly elevated bilirubin and alkaline phosphatase. Liver biopsy is required for definitive diagnosis, but is rarely performed because of the low pretest likelihood of hepatic GVHD in the modern era and because of procedural risk. When biopsies are performed, the most common finding associated with hepatic GVHD is bile-duct injury and loss resulting in cholestasis.³⁶⁷ Numerous other etiologies can cause abnormal liver function tests in allogeneic HCT recipients, including SOS, viral hepatitis, regimen-related hepatotoxicity, and medication-related hepatotoxicity (particularly with triazoles and CSP). Hepatic GVHD is relatively uncommon except in the setting of severe multisystem acute GVHD.

Pathophysiology

The current model of acute GVHD requires three steps. In step 1, the transplantation conditioning regimen damages host tissue, leading to increased secretion of inflammatory cytokines such as TNF-α and IL-1. These cytokines enhance alloreactivity of donor T cells by upregulating the expression of major and minor host tissue histocompatibility antigens and also affect other molecules on host antigen-presenting cells (APCs). Regimen-related damage to the GI tract results in leakage of endotoxins such as lipopolysaccharides into the systemic circulation, where they serve as additional inflammatory stimuli.³⁵³ In step 2, resting donor T cells become activated in secondary lymphoid organs by host or donor APCs, which present alloantigens to the T-cell receptor in the context of MHC. Costimulatory signals are required for full T-cell activation. Donor T-cell activation is characterized by cellular proliferation and predominance of Th1 cells and the secretion of IL-2 and IFN-γ. In step 3, cellular effectors mediate tissue injury and destruction in the target organs, resulting in the clinical manifestations of acute GVHD. This step involves the continued release of inflammatory cytokines that direct specific antihost donor-derived T cells to migrate to the target tissues of acute GVHD, namely, skin, liver, and gut. Neutrophils and mononuclear phagocytes contribute to local tissue injury by amplifying the proinflammatory response.

Prevention

Without some form of GVHD prophylaxis, virtually all patients undergoing allogeneic HCT would develop severe or fatal acute GVHD. Immunosuppressive drugs are the mainstay of acute GVHD prevention, and all patients undergoing allogeneic HCT with a T-cell-replete graft require prophylaxis. On the basis of randomized clinical trials published in the 1980s, the most commonly used regimen in myeloablative allogeneic HCT is the combination of a calcineurin inhibitor (CSP or tacrolimus [TAC]) with a short course of MTX, generally given on days +1, +3, +6, and +11 after allotransplantation.^368,369 The addition of prednisone to this backbone paradoxically increased the risk of acute GVHD in a randomized clinical trial, and thus glucocorticoids are rarely used for GVHD prophylaxis.³⁷⁰ Two randomized studies demonstrated that TAC was superior to CSP in preventing acute GVHD after myeloablative allogeneic HCT,^371,372 and thus TAC has largely supplanted CSP in this setting. A recent randomized clinical trial conducted by the BMT-CTN compared TAC/sirolimus (TAC/SRL) to the standard TAC/MTX regimen as GVHD prophylaxis after myeloablative allogeneic HCT. The primary endpoint (day +114 survival free of acute GVHD grades II to IV) was not significantly different in the two arms, nor were overall or progression-free survival.³⁰⁰ The toxicity profiles of the two regimens differed somewhat, but did not impact TRM. Interestingly, there was a strong trend toward a higher incidence of chronic GVHD in the SRL-containing arm (53 percent vs. 45 percent, p = 0.06).

Donor T-cell depletion has been explored as a means of GVHD prevention, using either mechanical ex vivo T-cell depletion or in vivo T-cell depletion in the form of ATG or alemtuzumab. T-cell-depletion strategies have generally been successful in reducing GVHD, but in many instances the reduction in donor T cells contributes to an increased incidence of graft rejection, infection, and relapse which may negate the advantage of GVHD prevention. A recent randomized clinical trial of ATG versus placebo added to standard CSP/MTX GVHD prophylaxis reported lower rates of acute and chronic GVHD in the ATG arm, although TRM and overall survival were not improved.³⁷³ Longer followup from this study confirmed a significantly lower rate of chronic GVHD in the ATG arm (45 percent vs. 12 percent, p <0.0001).³⁷⁴ However, ATG was associated with an increased risk of PTLD, a complication that is virtually nonexistent after T-cell-replete allogeneic HCT. In fact, there were 4 deaths from PTLD in the ATG arm (versus none in the placebo arm). Perhaps as a result, the study did not show improvements in TRM or overall survival with ATG. A subsequent randomized clinical trial aimed at replicating and extending these results with ATG has been completed in the United States.

In the setting of reduced-intensity allotransplantation, a number of immunosuppressive regimens have been used, the most common being CSP plus MMF.^177,185 Posttransplant CY has demonstrated impressive efficacy in the setting of HLA-haploidentical HCT, and has been studied in the setting of HLA-matched allotransplantation as well.^375,376 The combination of TLI and ATG is associated with very low rates of acute and chronic GVHD even with standard CSP/MMF postgrafting immunosuppression,^184,185 suggesting that protective conditioning can play a significant role in GVHD prevention.

Several laboratories have reported that naïve (CD62L+) T cells induce experimental acute GVHD, whereas effector memory (CD62L–) T cells do not.^377,378 Based on this insight, depletion of naïve donor T cells has been investigated as a means of preventing acute GVHD in humans.³⁷⁹ Murine models of marrow transplantation show that simultaneous infusions of T_reg limited the proliferation and clonal expansion of activated donor T cells and protected against acute GVHD development in murine models.¹⁴⁴ Pilot studies of T_reg infusions in humans have demonstrated safety and some efficacy in preventing GVHD,^145,146,380 although larger trials are required to confirm these preliminary findings. Importantly, relapse rates did not appear to be increased by the addition of T_reg, consistent with murine models.^144,146

Another investigational avenue in GVHD prevention involves inhibition of lymphocyte trafficking through CCR5 blockade. A recent trial combining the CCR5 antagonist maraviroc with standard GVHD prophylaxis demonstrated low rates of acute GVHD, particularly GI GVHD.³⁸¹ The BMT-CTN is currently conducting a three-arm randomized clinical trial comparing novel approaches to GVHD prophylaxis: posttransplant CY; maraviroc; and bortezomib-based immunosuppression.

Treatment

The standard first-line therapy for acute GVHD requiring systemic treatment is methylprednisolone or prednisone, at a dose of 1 to 2 mg/kg/day with subsequent tapering once disease activity resolves. Higher doses of methylprednisolone (10 mg/kg per day) do not prevent evolution to grades III to IV acute GVHD or improve survival.³⁸² Patients with acute GVHD grade II or less can be safely treated with a starting dose of 1 mg/kg/day of methylprednisolone, an approach that reduces overall glucocorticoid exposure and toxicity.³⁸³ Complete resolution of acute GVHD is reported in less than 50 percent of patients after first-line treatment with glucocorticoids, and the likelihood for long-term survival is low among individuals who develop glucocorticoid-refractory acute GVHD.³⁸⁴ Various approaches, including third-party mesenchymal stem cells, MMF, and TNF blockade, have been combined with glucocorticoids in upfront treatment of acute GVHD, but none has proven superior to glucocorticoids alone.^385,386

There is no standard second-line therapy for patients with glucocorticoid-refractory acute GVHD. Practices vary widely, although groups such as the ASBMT have published guidelines in an effort to codify and standardize treatment.³⁸⁷ Even though many agents have been used in this setting, none has proven superior or even reliably effective. Options include enrollment on a clinical trial, treatment with a second-line agent of choice, and palliative care. Agents that have been studied include daclizumab, SRL, MMF, ABX-CBL (a CD147-specific monoclonal antibody), anti-TNF agents, visilizumab, ruxolitinib,^388,389 and ATG, among others. In the absence of comparative clinical trial data, the choice of second-line agent is often guided by institutional experience, physician preference, and side-effect profiles. Outcomes with second-line therapy remain poor, and novel treatments for glucocorticoid-refractory acute GVHD are needed.

CHRONIC GRAFT-VERSUS-HOST DISEASE

Chronic GVHD is the major determinant of quality of life in long-term survivors of allogeneic HCT.^390,391 Despite its impact, however, it remains poorly understood and treatment options remain limited and largely empiric. Historically, any form of GVHD occurring beyond day +100 after allogeneic HCT was defined as chronic GVHD. However, with an increasing appreciation of the biologic differences between acute and chronic GVHD and the recognition that late acute GVHD occurs beyond day +100 with RIC, chronic GVHD was redefined on the basis of pathognomonic clinical and histologic criteria by a National Institutes of Health (NIH) consensus conference in 2005.³⁶⁰ The NIH consensus criteria include global and organ-specific scoring to evaluate the severity of chronic GVHD, and these staging tools have been validated to correlate with clinical severity and outcomes.^392,393 The Center for International Blood and Marrow Transplant Research (CIBMTR) has also developed a chronic GVHD risk-stratification algorithm that identifies age, donor–recipient gender mismatch, serum bilirubin, platelet count, donor type, and performance status, among other factors, as predictors of outcome in patients with chronic GVHD.³⁹⁴ The presence of chronic GVHD is consistently associated with more potent GVT effects and a lower risk of posttransplantation relapse,^177,395 although the negative impact of chronic GVHD on TRM may outweigh the benefit against relapse.¹⁷⁷

In contrast to acute GVHD, which is limited to the skin, gut, and liver, chronic GVHD has diverse manifestations which can affect nearly any organ system and which overlap considerably with those of autoimmune disorders such as scleroderma, lichen planus, Sjögren syndrome, and dermatomyositis. The most common clinical features of chronic GVHD include lichenoid skin lesions that may progress to generalized scleroderma, keratoconjunctivitis sicca, lichenoid oral lesions, esophageal and vaginal strictures, intestinal abnormalities, chronic liver disease, and bronchiolitis obliterans. Comprehensive assessment of chronic GVHD requires a detailed history and physical examination, but can be performed in the clinic in less than 20 minutes.³⁹⁶ Instructional videos and training tools demonstrating clinical assessment of chronic GVHD are available.³⁹⁷

The pathophysiology of chronic GVHD remains poorly understood. In contrast to acute GVHD, where there is a track record of translation from relevant animal models to humans, existing murine models of chronic GVHD suffer from serious shortcomings,³⁹⁸ limiting the understanding of its pathobiology and restricting the study of novel therapies. Efforts continue to develop more relevant murine models,³⁹⁹ although translational success remains elusive.

Although chronic GVHD is frequently preceded by acute GVHD, progress in reducing the incidence of acute GVHD has generally not translated into a reduction in the incidence of chronic GVHD. If anything, the incidence of chronic GVHD is likely increasing, as a result of the increasing use of PBPC as a graft source and the increasing at-risk pool of long-term survivors of allogeneic HCT. The only approaches proven to prevent chronic GVHD are the use of marrow as a graft source rather than PBPC,⁷² and the use of ex vivo or in vivo T-cell depletion (which carries increased risks of infection, PTLD, and relapse).^373,400 It has been suggested that posttransplant CY may prevent chronic GVHD, as low rates have been reported in single-arm studies,^375,376 but this comparison is confounded by graft source (predominantly marrow in trials of posttransplant CY) and requires confirmation in randomized clinical trials. Recipient statin use has been associated with a significantly reduced risk of chronic GVHD, but a higher risk of relapse, in patients receiving CSP-based GVHD prophylaxis regimens.⁴⁰¹ Rituximab has been studied as a prophylactic agent because of the demonstrated role of B cells in the genesis of chronic GVHD. Results have been mixed; a clinical trial performed at Stanford University demonstrated that rituximab potently abrogated B-cell alloimmunity, but did not lead to a statistically significant reduction in chronic GVHD incidence.⁴⁰² A nonrandomized study from the Dana-Farber Cancer Institute reported that rituximab decreased glucocorticoid-requiring chronic GVHD (but not overall chronic GVHD incidence) in comparison with a concurrent control cohort,⁴⁰³ a finding which requires confirmation in randomized clinical trials.

Despite decades of investigation of novel therapies, the standard of care for initial systemic treatment of chronic GVHD remains prednisone 1 mg/kg/day, with or without a calcineurin inhibitor.⁴⁰⁴ Numerous investigational agents have demonstrated promise in uncontrolled phase II clinical trials. However, these agents have uniformly proven disappointing when subjected to randomized phase III studies. At least six large randomized trials conducted over the past 20 years testing various alternate regimens, including azathioprine, thalidomide, MMF, and hydroxychloroquine, have failed to demonstrate the efficacy of any novel regimen for treatment of established chronic GVHD.^{405,406,407,408,409,410}

Patients with chronic GVHD typically require prolonged courses of immunosuppressive therapy. The median time to discontinuation of all systemic immunosuppression in patients with chronic GVHD resolution is approximately 2 years.⁴¹¹ Approximately half of patients with chronic GVHD will fail to respond to glucocorticoids and require second-line therapy.⁴¹² As with acute GVHD, there is no single standard of care for glucocorticoid-refractory chronic GVHD, and enrollment on a clinical trial should be strongly considered. Treatment choice depends on patient and physician preference, side-effect profile, and institutional priorities, as there are no data to guide a more systematic approach. Typical second-line agents include extracorporeal photopheresis (ECP), SRL, thalidomide, pentostatin, rituximab, and imatinib, among others.⁴¹² ECP is logistically cumbersome but can be effective and causes relatively few adverse effects⁴¹³; this approach is currently being studied in a randomized clinical trial conducted by the BMT-CTN. Low-dose IL-2 has been reported to facilitate T_reg expansion and homeostasis in a small single-center study of patients with chronic GVHD,^414,415 although the degree of T_reg expansion was not significantly associated with clinical response.⁴¹⁴ Nonetheless, therapies targeting T_regs are under active investigation in the treatment of chronic GVHD, as is the Bruton tyrosine kinase inhibitor ibrutinib.⁴¹⁶ More effective prevention and treatment of chronic GVHD remains a crucial research need in allogeneic HCT.

RELAPSED MALIGNANCY AFTER HCT

Relapse following autologous or allogeneic HCT is an ominous clinical event. Every effort should be made to verify relapse pathologically, as it is common for patients to have residual radiographic abnormalities following transplantation, especially in patients with lymphoma. Patients with myeloma have gradual reductions in biochemical markers of disease which may take several months after autologous HCT to reach maximal response. Following RIC, the allogeneic GVT effect may take weeks to months to result in tumor eradication, and it may be difficult to distinguish persistent yet slowly regressing disease from slowly progressive disease, particularly with indolent NHL or CLL.

Relapse after Autologous Hematopoietic Cell Transplantation

Disease relapse remains the most common cause of treatment failure after autologous HCT. Relapse often occurs at sites of previous disease, suggesting that residual disease within the patient rather than autograft contamination is responsible.⁴¹⁷ Additional cytotoxic chemotherapy alone is highly unlikely to be curative in patients relapsing after autologous HCT, as the disease has already survived supralethal doses of chemotherapy. Treatment options for patients with relapse after autologous HCT include irradiation, immunomodulators, and/or targeted therapies. For selected patients, salvage allogeneic HCT may be feasible using RIC. Strategies to reduce the risk of relapse in high-risk patients undergoing autologous HCT include consolidative involved-field radiotherapy, antitumor vaccination,⁴¹⁸ maintenance therapy with targeted agents,^269,419,420 and planned tandem allogeneic HCT.⁴²¹

Relapse after Allogeneic Hematopoietic Cell Transplantation

Treatment of disease relapse following allogeneic HCT is generally unsuccessful. In particular, patients with high-risk malignancies and early relapse (<100 days after allogeneic HCT) have a dismal prognosis, with 2-year overall survival of less than 5 percent.⁴²² Salvage chemotherapy can result in disease responses, but they are unlikely to be durable. Performing a second myeloablative transplantation procedure has largely been unsuccessful because of excessive toxicity and TRM of greater than 50 percent. More recently, selected patients have been treated with a second allogeneic HCT using RIC. Treatment-related toxicity with this approach is not prohibitive and successful disease eradication has been reported, although relapse remains the major cause of death.⁴²³ There is no consensus on whether to use the same donor or a different donor for second-salvage allogeneic HCT in the setting of relapse. More importantly, the majority of patients with relapse after allogeneic HCT are ineligible for second allotransplant because their diseases cannot be adequately controlled or cytoreduced. Experimental therapies with chimeric antigen receptor-bearing autologous T cells or other investigational approaches warrant consideration in these cases. In our view, palliative care is a reasonable option, particularly in the setting of early or chemorefractory relapse after allogeneic HCT, and should be presented to patients.

In the setting of retained donor T-cell chimerism, posttransplantation relapse has sometimes been treated by rapidly tapering immunosuppressive medications to stimulate a GVT effect. Although this approach is occasionally successful in patients with indolent malignancies such as low-grade NHL or CLL, it is rarely effective against more aggressive diseases such as acute leukemias, and carries a high risk of precipitating severe GVHD.

In patients who are off immunosuppression without evidence of GVHD at the time of relapse, DLI has been used to augment GVT. The mechanism by which DLI works is unclear; it may normalize the T-cell repertoire or reverse so-called T-cell “exhaustion.”^424,425 Historically, the best outcomes for DLI have been reported for CML, although TKIs have supplanted DLI for the most part in this setting.⁴²⁶ DLI alone is typically insufficient to control aggressive hematologic malignancies,⁴²⁷ and thus patients are often treated with reinduction chemotherapy or other cytoreduction before receiving DLI as consolidation.

The major potential adverse effects of DLI are GVHD and marrow aplasia. The risk of GVHD after DLI is at least 20 percent and likely as high as 50 to 70 percent.^428,429 Marrow aplasia typically occurs in the setting of residual host hematopoiesis; when host HSC are eradicated by DLI, there may be too few donor HSC to support hematopoietic recovery and prolonged aplasia (>6 weeks) can ensue.⁴³⁰ Thus, chimerism should be evaluated before DLI, and DLI should be used with caution in patients with significant residual host hematopoiesis. There are limited data on the optimal cell dose of DLI. In a retrospective analysis, doses greater than 1 × 10⁸ CD3+ cells/kg were associated with a high risk of GVHD (55 percent) without a corresponding benefit in disease control, while doses of 1 × 10⁷ CD3+ cells/kg or less were associated with the lowest rates of GVHD (21 percent).⁴²⁸ Modifications of DLI, including selection of CD8+ effector lymphocytes or production of cytokine-induced killer cells,¹⁴¹ are under investigation to improve the safety and efficacy of this approach.

In recent years, research in HCT has helped to establish alternative-donor allotransplantation, improve outcomes through better supportive care,²¹⁸ and expand transplant eligibility to previously ineligible populations using RIC,²⁰² among other advancements. Current research directions in HCT are diverse; two avenues among many are briefly summarized here.

The past several years have seen an explosion of research interest in the microbiome and the role of host/microbiome interaction in regulating immunity. The relevance of this field to allogeneic HCT is immediately obvious, as the gut and skin are key targets of GVHD. Preclinical evidence suggests that gut microbiota play a critical role in the development of GI GVHD.^431,432 In humans, preliminary evidence links changes in the microbiome to respiratory complications,⁴³³ GI GVHD,^432,434 bacteremia,⁴³⁵ and mortality after allogeneic HCT.⁴³⁶ While our understanding of these interactions is still in its infancy and the clinical implications of these findings remain unclear, studies are underway revisiting older strategies, such as total gut decontamination, as well as more modern efforts to tailor the microbiome to optimize transplant outcomes.⁴³⁷

Allogeneic HCT is also under investigation as a platform to promote tolerance of solid-organ allografts. Solid-organ transplants typically require lifelong immunosuppression to maintain graft function, but investigators in the 1990s noted that renal allotransplantation could be accomplished without immunosuppressive therapy in recipients of allogeneic HCT, provided that the kidney allograft was obtained from the original marrow donor.^438,439 The development of reduced-intensity regimens has spurred research in this field, as these regimens are less toxic and often produce mixed rather than complete donor chimerism—a desirable characteristic in this setting since mixed chimerism reduces the risk of GVHD.

Investigators at Northwestern University described a cohort of 15 HLA-mismatched living-donor kidney transplant recipients given RIC and an infusion of donor marrow enriched with facilitator cells. Six of the 15 patients developed sustained chimerism and were completely withdrawn from immunosuppressive medication without renal allograft rejection.^440,441 Investigators in Boston reported that four of 10 HLA-haplotype matched living-donor kidney transplant recipients were able to discontinue immunosuppression for up to 11.4 years without subsequent graft dysfunction after establishment of transient mixed chimerism.⁴⁴² Stanford investigators used TLI-ATG conditioning in 16 living-donor kidney transplant recipients from HLA-matched siblings treated with infusion of CD34-selected hematopoietic progenitor cells and a defined dose of T cells. Fifteen patients developed multilineage mixed chimerism without GVHD, and eight were withdrawn from antirejection medications without subsequent rejection episodes. Four additional chimeric patients were in the midst of drug withdrawal, and only four patients were not withdrawn because of the return of the underlying disease or rejection episodes.^189,443 Case reports have also described the use of allogeneic HCT as a tolerance-induction platform in lung transplantation.⁴⁴⁴