Reduced-intensity conditioning allogeneic hematopoietic-cell transplantation for older patients with acute myeloid leukemia

Abstract

Elderly patients (>60 years) with acute myeloid leukemia have a poor prognosis with a chemotherapy-alone approach. Allogeneic hematopoietic-cell transplantation (HCT) can improve overall survival (OS). However, myeloablative regimens can have unacceptably high transplant-related mortality (TRM) in an unselected group of older patients. Reduced-intensity conditioning (RIC) or nonmyeloablative (NMA) conditioning regimens preserve the graft-versus-leukemia effects but reduce TRM. NMA regimens result in minimal cytopenia and may not require stem cell support for restoring hematopoiesis. RIC regimens, intermediate in intensity between NMA and myeloablative regimens, can cause prolonged myelosuppresion and usually require stem cell support. A few retrospective and prospective studies suggest a possibility of lower risk of relapse with myeloablative HCT in fit older patients with lower HCT comorbidity index; however, RIC and NMA HCTs have an important role in less-fit patients and those with significant comorbidities because of lower TRM. Whether early tapering of immunosuppression, monitoring of minimal residual disease, and post-transplant maintenance therapy can improve the outcomes of RIC and NMA HCT in elderly patients will require prospective trials.

Introduction

Acute myeloid leukemia (AML) is the most common type of acute leukemia in adults, with an estimated 20,000 new cases of AML in 2016 [Siegel et al. 2016]. It is particularly common in older patients and has a median age of 65–70 years at diagnosis [Siegel et al. 2016]. Most of the institutional registries, including the Cancer and Leukemia Group B (CALGB) have defined old age in the context of AML and hematopoietic-cell transplantation (HCT) as more than 60 years [Devine et al. 2015; Giles et al. 2007; Wallen et al. 2005]. The management of AML in older individuals can be particularly challenging due to the presence of multiple comorbidities, inability to tolerate intensive chemotherapy [Schiffer, 2010], and a higher incidence of unfavorable cytogenetics and drug resistance [Appelbaum et al. 2006]. Most of the reported trials involving older patients with AML have resulted in complete remission (CR) rates ranging between 35% and 70% [Ferrara et al. 1998; Gardin et al. 2007; Leoni et al. 1997; Lowenberg et al. 2009; Vey et al. 2004]. After a CR is achieved, AML relapse may occur in most cases within 4–6 months in the absence of HCT. Postremission therapy aims at reducing such relapses by destroying any leukemia cells that may have survived induction therapy. This involves the use of either consolidation chemotherapy or allogeneic HCT. The use of standard-dose consolidation chemotherapy in older patients is limited by a high incidence of treatment-related toxicities, with dose modification or reduction required in most cases [Gardin et al. 2007; Mayer et al. 1994; Lowenberg et al. 1998]. The use of HCT, by its graft versus leukemia (GVL) effect, has the advantage of offering long-term disease-free survival (DFS), especially in patients with high-risk disease [Blaise et al. 1992; Chang et al. 2007; Vicente et al. 2007].

Myeloablative regimens incorporate alkylating agents with or without total body irradiation (TBI) at doses that will not allow autologous stem cell recovery [Bacigalupo et al. 2009]. In older patients, the use of myeloablative chemotherapy regimens can lead to a high transplant-related mortality (TRM) of approximately 20% at 100 days and 40% at 3 years [Wallen et al. 2005]. The risk of TRM depends on HCT comorbidity index, which is higher in older patients [Sorror et al. 2005]. This has led to an increasing interest in utilizing nonmyeloablative (NMA) and reduced-intensity conditioning (RIC) regimens that preserves GVL effect but reduces TRM. The use of NMA regimen causes minimal cytopenias but is immunosuppressive to the extent that it results in full engraftment of donor stem cells [Bacigalupo et al. 2009]. By strict definition, NMA regimens do not require stem cell support. Conversely, RIC regimen, intermediate in intensity between myeloablative and NMA regimens, causes prolonged cytopenia and requires stem cell support. In RIC regimens, the dose of alkylating agents or TBI is reduced by at least 30% [Bacigalupo et al. 2009]. The definition of Center for International Blood and Marrow Transplant Research (CIBMTR) for RIC regimens includes the use of less than 500 cGy of TBI as a single fraction or 800 cGy in fractionated doses, a busulfan dose of less than 9 mg/kg, a melphalan dose of less than 140 mg/m², or a thiotepa dose of less than 10 mg/kg [Giralt et al. 2009]. Both NMA/RIC regimens allow the use of allogeneic HCT in older patients [Bacigalupo et al. 2009]. Despite the availability of NMA/RIC regimens, older AML patients are infrequently referred to transplant centers, and HCT is underutilized in this population. This article aims to review the existing evidence on the role of NMA and RIC HCT in older patients with AML.

History of nonmyeloablative and reduced-intensity conditioning regimens in acute myeloid leukemia

In allogeneic HCT, antitumor effect is mediated by donor lymphocytes as well as chemotherapy [Childs and Srinivasan, 2002]. Syngeneic HCT has a lower incidence of graft-versus-host disease (GVHD), thereby resulting in a higher incidence of relapse. NMA and RIC regimens, compared with myeloablative regimens, rely more on the GVL effect and less on chemotherapy-mediated destruction of leukemia cells. NMA and RIC regimens use conditioning regimens intense enough to prevent rejection, but allow preservation of the GVL effect [Storb, 2007]. Some of the retrospective registry analyses have shown encouraging results with such approaches. In 2006, a multicenter study of 122 AML patients from various centers in the United States and Europe utilized low-dose TBI with or without fludarabine as an RIC prior to allogeneic HCT. Of 122 patients, 18 had secondary AML, 51 were in first CR (CR1), 39 were in second CR (CR2), and 32 were beyond CR2. The use of RIC HCT resulted in 2-year overall survival (OS) of 48%, with TRM of 22% and 10% at 2 years for unrelated and related recipients, respectively [Hegenbart et al. 2006]. Another retrospective registry-based analysis from the European group for Blood and Marrow Transplantation (EBMT) comparing 361 RIC allogeneic HCT to autologous HCT, demonstrated less risk of relapse with the former [relative risk (RR), 0.77; 95% CI, 0.63–0.95; p = 0.013). In a multivariate analysis, RIC allogeneic HCT was shown to have a better DFS and OS as compared with autologous HCT in patients who achieved CR2 [Herr et al. 2007]. However, the reduced risk of relapse associated with allogeneic HCT in patients in CR1 did not translate into a survival advantage, likely secondary to increased TRM.

One of the largest retrospective studies from the CIBMTR registry analyzed 545 AML and 535 myelodysplastic syndrome (MDS) patients who underwent RIC HCT. All patients with AML were between the ages of 40–79 years and underwent allogeneic HCT in CR1. Multivariate analysis revealed no impact of age on TRM, DFS, and OS. The 2-year OS in age groups 40–54, 55–59, 60–64, and 65 years or over were as follows: 44% (95% CI, 37–52%), 50% (95% CI, 41–59%), 34% (95% CI, 25–43%), and 36% (95% CI, 24–49%), respectively (p = 0.06) [McClune et al. 2010]. Unfavorable cytogenetics, poor pre-HCT performance status, and greater human leukocyte antigen (HLA) disparity were associated with worse survival in this study. Sorror and colleagues conducted a large multicenter study of 372 patients aged 60–75 years with various hematologic malignancies (n = 109 for acute leukemia) who underwent HCT using NMA regimens. The 5-year cumulative incidences of TRM and relapse were 27% (95% CI, 22–32%) and 41% (95% CI, 36–46%), respectively, leading to 5-year OS and progression-free survival (PFS) of 35% (95% CI, 30–40%) and 32% (95% CI, 27–37%), respectively [Sorror et al. 2011]. Moreover, these outcomes were not statistically significant when stratified by age. However, like many other retrospective studies, this study might have the potential for selection bias and does not examine AML patients specifically. Nevertheless, it points out that old age should not be a sole factor for excluding the use of HCT.

A small prospective study assessed the outcomes of 50 patients (30 AML and 20 MDS patients) with a median age of 57 years, who underwent NMA allogeneic HCT using fludarabine and cyclophosphamide [Nelson et al. 2010]. A total of 43% of the patients were in CR1, 30% in CR2, 23% were not in remission, and 3% in relapse at the time of HCT. The study estimated 4-year DFS and OS to be 37% and 41%, respectively [Nelson et al. 2010]. In another EBMT study, clofaribine was used as a component of the RIC regimen in 69 AML patients (median age 43; age range 18–69 years). This study demonstrated a 2-year leukemia-free survival (LFS) of 30%, OS of 35% and TRM of 31% [Chevallier et al. 2012]. In a British study, the investigators attempted to reduce the TRM by reducing the dose of alkylating agent that was used in prior trials [Davies et al. 2013]. A total of 56 patients with AML (n = 41) and MDS (n = 15) were treated with a reduced dose of cyclophosphamide (~45 mg/kg versus 120 mg/kg in earlier trials). After a median follow up of 5 years, the DFS was reported to be 45% (overall) and 55% (patients transplanted in remission), whereas TRM was 9%. These data are encouraging for the use of NMA or RIC regimens for older patients who may have several comorbidities.

A recent prospective multicenter phase II trial from CALGB assessed the role of RIC HCT in AML patients between the ages of 60 and 74 years (median age 65 years), who were transplanted in CR1. Of 114 total patients, 52% received unrelated donor HCT. At 2 years, DFS and OS were 42% and 48% respectively for the entire cohort, and 40% and 50%, respectively, for the unrelated donor group. The 2-year cumulative relapse rate and TRM for the entire group were 44% and 15%, respectively [Devine et al. 2015]. This recent study demonstrates promising results of RIC HCT in older AML patients in CR1.

Chemotherapy versus reduced-intensity conditioning or nonmyeloablative conditioning

When assessing the role of RIC or NMA conditioning in elderly AML patients, it is imperative to compare it with postremission chemotherapy. In the past, chemotherapy was the preferred post-remission therapy for many older patients, partly secondary to increased TRM with HCT. A retrospective study from the CIBMTR registry compared the outcomes of 190 older patients (60–70 years) with AML receiving RIC HCT in CR1 with patients who received consolidation chemotherapy. [Farag et al. 2011] The patients who received HCT were younger (63 versus 65 years) and had a longer duration from diagnosis to CR1 (44 versus 38 days). At 3 years, the study demonstrated that RIC allogeneic HCT was associated with a significantly lower risk of relapse (32% versus 81%; p < 0.001), higher TRM (36% versus 4%; p < 0.001), and longer LFS (32% versus 15%; p = 0.001). However, the study was able to show only a trend towards increased OS in patients who received HCT (37% versus 25%; p = 0.08) [Farag et al. 2011]. The retrospective design with potential selection bias and differences in the characteristics of the two groups may limit the conclusion of the study.

Kurosawa and colleagues retrospectively compared outcomes in 152 older Japanese AML patients (50–70 years) who received allogeneic HCT (myeloablative or RIC) after achieving CR1 with patients who received chemotherapy alone (n = 884) [Kurosawa et al. 2011]. The HCT group had a lower cumulative incidence of relapse as compared with the chemotherapy group at 3 years from CR1 (22% versus 62%; p < 0.001), but a higher incidence of TRM (21% versus 3%; p < 0.001). The 3-year relapse-free survival (RFS) (56% versus 29%; p < 0.001) and OS (62% versus 51%; p = 0.012) were also higher in the HCT group [Kurosawa et al. 2011]. In a subgroup analysis among HCT recipients, 37% underwent myeloablative conditioning and 61% underwent RIC, with no significant OS difference, however, the number of patients was relatively modest. A prospective trial from MD Anderson Cancer Center for patients with AML and unfavorable cytogenetics utilizing RIC in first CR was conducted in patients over 50 years of age. The transplant team evaluated 53 eligible patients who achieved CR, but only 14 patients underwent HCT. These patients were compared with 43 matched patients who received postremission chemotherapy alone. The study suggested that RIC HCT was associated with increased RFS but not OS. [Estey et al. 2007]

Randomized trials comparing postremission chemotherapy versus HCT regardless of the donor availability are difficult to conduct. However, a French study utilized a ‘genetic randomization’ technique, wherein donor-versus-no-donor comparison was made in an intention-to-treat trial based on the availability of an HLA-identical sibling donor [Mohty et al. 2005]. The study included 95 AML patients in CR. The median age of the patients was 52 years (26–65 years). RIC regimens included fludarabine, busulfan and antithymocyte globulin. LFS was significantly higher in the ‘donor’ group as compared with the ‘no donor’ group that received conventional chemotherapy (54% versus 30% at 4 years; p = 0.01). The difference in LFS attained greater significance when patients who actually received HCT were compared with those who did not (62% versus 31% at 4 years; p = 0.001) [Mohty et al. 2005]. Additionally, OS was significantly higher (approximately 15%) in the ‘donor’ group as compared with the ‘no donor’ group (p = 0.04). However, in 25 patients who actually received allogeneic HCT, 13 patients had received high-dose cytarabine and autologous transplantation prior to the allogeneic HCT (details unavailable). Additionally, this study included many younger patients and patients with less comorbidities.

A recent time-dependent analysis of four prospective phase III trials in AML from Dutch-Belgian Hemato-Oncology Cooperative Group and the Swiss Group for Clinical Cancer Research (HOVON-SAKK) collaborative study group compared outcomes of various postremission strategies in AML patients aged 60 years and over, who had achieved CR following induction chemotherapy. In a multivariate analysis, the use of RIC allogeneic HCT resulted in better 5-year OS as compared with other nonallogeneic HCT therapies (such as chemotherapy, autologous HCT, or no further therapy) [hazard ratio (HR), 0.71; 95% CI, 0.53–0.95] [Versluis et al. 2015]. Taken together, the above studies suggest a superiority of RIC allogeneic HCT over consolidation chemotherapy alone in terms of RFS and disease control.

Alternative donor transplant in older patients

Availability of a matched donor is an important barrier to allogeneic HCT. This has led to an increasing interest in the evaluation of alternate donor sources such as partially HLA-mismatched (haploidentical) donors or unrelated umbilical cord blood. In a prospective study from Minnesota, 98 patients with AML or MDS aged 55 years or over underwent RIC HCT using matched-sibling donor (n = 38; median age 63 years) or cord blood (n = 60; median age 61 years). There were 26 AML patients in the sibling donor group (14 in CR1, 12 in CR2–3) and 44 AML patients in the cord blood group (27 in CR1, 17 in CR2–3) [Majhail et al. 2012]. Of the patients receiving cord blood HCT, 22 were 5/6 antigen matched and 31 were 4/6 antigen matched. On a multivariate analysis, there was no difference with the use of cord blood HCT in OS (RR, 1.3; 95% CI, 0.8–2.3], LFS (RR, 1.2; 95% CI, 0.7–2.1), relapse rate (RR, 0.7; 95% CI, 0.3–1.3), and TRM (RR, 1.2; 95% CI, 0.6–2.7), compared with sibling donor HCT. In addition, cord blood HCT recipients had a lower incidence of chronic GVHD at 2-years post-transplant (61 versus 33%, p = 0.04) [Majhail et al. 2012]. Another small retrospective study from Minnesota assessed the use of cord blood RIC HCT in very old patients (age ⩾ 70 years) with AML (n = 7) or MDS (n = 3). This was compared with nine patients (age ⩾ 70 years) with AML (n = 6) or MDS (n = 3) who received RIC HCT using matched-sibling donor [Sandhu et al. 2016]. The study demonstrated comparable 2-year TRM (20% versus 24%) and OS (30% versus 49%) outcomes with the use of cord blood versus sibling donor HCT. A recent Japanese registry analysis compared outcomes in 2091 patients aged over 50 years with AML (n = 1346), acute lymphoblastic leukemia (ALL, n = 308), and MDS (n = 437) who underwent matched-sibling donor marrow HCT (n = 319) [Konuma et al. 2016], matched-sibling donor peripheral blood HCT (n = 462), or unrelated cord blood HCT (n = 1310). The study demonstrated similar relapse rates among all three donor groups, but lower overall mortality in the marrow HCT (HR, 0.67; p < 0.001) and peripheral blood HCT (HR, 0.75; p = 0.002) as compared with cord blood HCT.

In a recently reported large retrospective study from Johns Hopkins (n = 271), 65 AML patients aged between 50 and 75 years received NMA haploidentical donor HCT with post-transplant high-dose cyclophosphamide [Kasamon et al. 2015]. In AML patients aged 60 years and over, the 3-year relapse rate, PFS and OS were 60%, 31%, and 38%, respectively. The overall cumulative incidence of 1-year TRM for the entire study population of 271 patients with various hematological malignancies were as follows: 10% for patients aged 50–59 years, 14% for those aged 60–69 years and 11% for those aged 70–75 years (p = 0.20). This indicated that the NMA haploidentical HCT might be well tolerated in older patients regardless of age [Kasamon et al. 2015].

These studies demonstrate good outcomes of older AML patients who undergo alternative donor HCT. Hence, in older patients without matched donors, haploidentical and cord blood HCT are acceptable alternative donor strategies.

Myeloablative versus reduced-intensity or nonmyeloablative conditioning regimens

There is a relative scarcity of well-designed studies comparing the outcomes of myeloablative conditioning regimens to NMA or RIC regimens. A retrospective study from Seattle analyzed the outcomes of 150 patients with MDS or AML, transformed from MDS, who underwent myeloablative or NMA HCT [Scott et al. 2006]. The NMA regimen used in 38 patients included 2 Gy TBI (n = 2) or fludarabine at a dose of 90 mg/m² (n = 36). Other patients received myeloablative regimens including busulfan and cyclophosphamide. The patients who received the NMA regimen were older (median age of 62 versus 52 years; p < 0.001), had higher-risk disease by International Prognostic Scoring System (53% versus 30%; p = 0.004), and a higher comorbidity index (68% versus 42%; p = 0.01). The study demonstrated similar outcomes in both myeloablative and NMA conditioned groups in terms of 3-year OS (48% versus 27%; p = 0.56), PFS (44% versus 28%; p = 0.60), and TRM (34% versus 41%; p = 0.94) [Scott et al. 2006]. This suggests that the GVL effect can play an important role in reducing the risk of relapse in older patients with MDS and possibly AML.

In a large retrospective EBMT study of AML patients aged over 50 years in CR1, the outcomes of 315 RIC HCT were compared with 407 myeloablative HCT [Aoudjhane et al. 2005]. The most common RIC regimens were fludarabine in combination with either busulfan (53%) or low-dose TBI (24%). After a follow up of 13 months (median), acute GVHD (grade II–IV) and TRM were significantly less (p = 0.01) in the RIC group. However, DFS was similar in both groups, and relapse risk was significantly higher (p = 0.003) in the RIC group. The patients in the RIC group were relatively older, suggesting the probable role of RIC in decreasing the overall mortality at the possible expense of increased relapse risk [Aoudjhane et al. 2005]. Dana Farber Group published a large retrospective study that analyzed outcomes in AML patients aged over 50 years, who received NMA or myeloablative conditioning. A total of 71 patients (21 with high-risk AML) received the NMA regimen that consisted of fludarabine (120 mg/m²) and intravenous busulfan (3.2 mg/kg); 81 patients (13 high-risk AML) received myeloablative conditioning with cyclophosphamide and TBI. [Alyea et al. 2005]. Despite the adverse characteristics, such as active disease and prior myeloablative HCT, there was a trend towards higher OS in the NMA group at 1 (51% versus 38%; p = 0.06) and 2 years (39% versus 29%; p = 0.06). There was no difference in PFS (27% versus 25% at 2 years; p = 0.24) or incidence of grade II–IV acute GVHD (28% versus 27%). At median follow up of about 40 months, the NMA group had a lower cumulative rate of TRM (32% versus 50%; p = 0.01) but at the cost of higher relapse rate (46% versus 30%;p = 0.052) [Alyea et al. 2005].

The role of modified-dose intensity HCT was studied further in a prospective study from Israel involving 112 patients with AML (n = 95) and MDS (n = 17). A total of 45 patients in CR were eligible to receive myeloablative regimen with intravenous busulfan (12.8 mg/kg) and cyclophosphamide (60 mg/kg for 2 days) [Shimoni et al. 2006]. Other patients received RIC with fludarabine (150–160 mg/m²) and intravenous busulfan (6.4 mg/kg; n = 41) or a modified myeloablative regimen with fludarabine (150–160 mg/m²) and intravenous busulfan (12.8 mg/kg;n = 26). Patients who received modified-myeloablative (median age 51 years) or RIC (57 years) regimens tended to be older than the myeloablative group (42 years). The study showed similar 2-year OS and risk of relapse in all three groups. The myeloablative, compared with modified-myeloablative and RIC arms, had a higher incidence of TRM (22% versus 8% versus 8%, respectively; p = 0.05) [Shimoni et al. 2006]. Long-term follow-up results showed similar OS in all three groups when the HCT was performed at disease remission. However, the 5-year OS was 42%, 19% and 0% for patients with active disease at HCT, conditioned with myeloablative, modified-dose intensity and RIC HCT, respectively (p = 0.01), thereby highlighting that NMA may be a reasonable option in older patients who are in remission [Shimoni et al. 2010].

A large retrospective EBMT analysis compared outcomes in older AML patients with RIC (n = 252; median age 57; range 50–76 years; 31% in CR1, 37% in CR2–3 and 32% with advanced disease) or myeloablative conditioning (n = 182; median age 54; range 50–70 years; 41% in CR1, 27% in CR2–3 and 32% with advanced disease) [Ringden et al. 2009]. Patients receiving RIC were older, had a trend for longer interval between diagnosis and transplant in patients in CR2, received peripheral blood stem cells more frequently, and were treated less often with TBI. The most common indications for selecting RIC in older patients were advanced age (77%), infection (7%), special protocol (7%), and prior transplant or heavy treatment (7%). In patients older than 50 years, RIC was associated with decreased TRM (HR, 0.64; p = 0.04), similar relapse risk (HR, 1.34; p = 0.16) and LFS (HR, 1.04; p = 0.79) as compared with myeloablative conditioning [Ringden et al. 2009].

A prospective trial conducted in Barcelona compared two strategies of allogeneic HCT from HLA-identical siblings in 75 adults with poor-risk AML or MDS (refractory anemia with excess blasts) in CR1 [Martino et al. 2008]. The patients aged 50 years and less received myeloablative conditioning with cyclophosphamide and TBI whereas patients aged over 50 years received RIC with fludarabine and busulfan. The myeloablative group had a significantly higher percentage of AML patients (94% versus 56%, p < 0.01) and a lower probability of HCT the comorbidity index being at least 3 (26% versus 46%; p = 0.002). At a median follow up of 4 years, there was no difference in the TRM between the RIC and myeloablative groups (19% versus 20%; p = 0.80) [Martino et al. 2008]. This suggests that RIC regimens can reduce the risk of TRM in older patients.

The Blood and Marrow Transplant Clinical Trials Network (BMT CTN 0901) recently presented a phase III randomized trial to compare outcomes by conditioning intensity in patients with AML (n = 218; 18–65 years) or MDS (n = 54) who had less than 5% marrow myeloblasts by morphology prior to HCT [Scott et al. 2015]. Among 272 patients enrolled of the planned 356 patients, 135 received myeloablative conditioning and 137 received RIC regimens. The study suspended enrollment based on an interim analysis. After a follow up of 18 months, RIC HCT, compared with myeloablative HCT, was associated with lower TRM (4.4% versus 15.8%; p = 0.02), higher risk of relapse (48.3% versus 13.5%; p < 0.01), and similar OS (67.7% versus 77.4%; p = 0.07) (Table 1) [Scott et al. 2015]. Although this study did not assess RIC HCT specifically in older patients or those with a high comorbidity index, this study indicates a lower risk of relapse with myeloablative HCT, but at the expense of higher TRM in patients who are fit to undergo a myeloablative transplant. In summary, the aforementioned suggest a possibility of lower risk of relapse with myeloablative HCT in fit older patients with lower HCT comorbidity index; however, RIC and NMA HCTs can reduce TRM in less-fit patients and those with significant comorbidities.

Table 1.

Studies comparing intensity of conditioning regimen prior with hematopoietic-cell transplantation.

Study [year]	n	Median age in years (range)	Type of conditioning	Regimen	Donor source	TRM	Relapse	RFS/DFS/PFS	OS
Aoudjhane et al. [2005]	407	54 (50–64)	MA	Various	HLA identical sibling	32%	24%	44%	46%
	315	57 (50–73)	RIC			18% (p < 0.0001)	41% (p < 0.0001)	40% (p = 0.8)	47% (p = 0.43; 2 years)
Alyea et al. [2005]	81	54 (51–66)	MA	Mainly Cy/TBI	HLA matched related/unrelated	50%	30%	25%	29%
	71	58 (51–70)	NMA	Bu/Flu		32% (p = 0.01)	46% (p = 0.05)	27% (p = 0.24)	39% (p = 0.06; 2 years)
Scott et al. [2006]	112	53 (40–65)	MA	Bu/Cy	HLA matched related/unrelated	34%	23%	44%	48%
	38	62 (40–72)	NMA	TBI 2Gy ± Flu		41% (p = 0.94)	31% (p = 0.43)	28% (p = 0.60)	27% (p = 0.56; 3 years)
Shimoniet al. [2006] *	45	42 (22–58)	MA	Bu/Cy	HLA matched related/unrelated	22%	33%	45%	50%
	41	57 (18–70)	RIC	Bu/Flu		8%	49%	43%	47%
	26	51 (18–64)	Modified MA	Bu/Flu		8% (p = 0.05)	43% (p > 0.05)	49% (p > 0.05)	49% (p > 0.05; 2 years)
Scott et al. [2015] *	135	18–65	MA	Bu/Cy, Bu/Flu, Cy/TBI	HLA matched related/unrelated	15%	13%	67%	77%
	137	18–65	RIC	Flu/Bu or Mel		4% (p = 0.02)	48% (p < 0.01)	47% (p < 0.01)	67% (p = 0.07; 18 months)

Bu, busulfan; Cy, cyclophosphamide; DFS, disease-free survival; Flu, fludarabine; Mel, melphalan; n, number of patients; OS, overall survival; RFS, relapse-free survival; TBI, total body irradiation; TRM, transplant-related mortality; HLA, human leukocyte antigen; RIC, reduced-intensity conditioning; PFS, progression-free survival; MA, myeloablative; NMA, nonmyeloablative.

^*The Shimoni study is a prospective study, and the Scott study is a prospective randomized trial. All other studies are retrospective designs.

Impact of donor age

Donor age influences the outcomes of allo-HCT. Younger age of donors is associated with higher stem-cell yield and improved outcomes, particularly in RIC HCT [Mehta et al. 2006]. The question then becomes whether the older-matched sibling donor or the younger-matched unrelated donor is the preferred donor. A retrospective French study analyzed outcomes of 714 patients aged over 55 years with AML in CR1 after RIC HCT using matched-unrelated donors (median age 35 years) and matched-sibling donors (median age 61 years) [Peffault de Latour et al. 2015]. There was no significant difference in 3-year TRM (17% versus 23%; p = 0.17), relapse rate (37% versus 30%; p = 0.12), and OS (49% in both) between older sibling versus younger unrelated donor HCT. However, another retrospective study of AML (n = 117) and MDS (n = 62) patients from Spain who underwent allo-HCT from matched donors demonstrated conflicting results. The median age of patients was 51 years (range 18–69 years), and that of donors was 47 years (range 12–75 years), with 81 donors aged over 50 years [Bastida et al. 2015]. A total of 112 patients received RIC and 67 patients received myeloablative conditioning regimens. The patients who received HCT from donors aged over 50 years had worse OS (51% versus 73%; p = 0.01), TRM (20% versus 8%; p = 0.038) relapse rate (28% versus 39%; p = 0.03) as compared with those who received HCT from donors aged less than 50 years. On a subgroup analysis, a 3-year OS, using matched-sibling donors aged over 50 years was worse in comparison with matched-unrelated donors aged less than 50 years (54% versus 72%; p < 0.001).

Role of minimal residual disease monitoring

The success of NMA or RIC HCT has been primarily demonstrated in patients who were transplanted in CR. Recently, multiparameter flow cytometry has allowed detection of minimal residual disease (MRD), and hence a more stringent remission status. In a retrospective study from Seattle, bone marrow aspirate from 241 adults receiving NMA (n = 86, median age ~62 years) or myeloablative HCT (n = 155, median age ~50 years) for AML in CR1 was analyzed for the presence of MRD [Walter et al. 2015]. The presence of MRD was associated with a higher risk of relapse rate in both NMA (HR, 2.83; 95% CI, 1.38–5.81) and myeloablative arms (HR, 7.22; 95% CI, 3.85–13.54). Another study from University of Minnesota analyzed the impact of the presence of MRD on outcomes in 203 AML patients who received RIC (n = 123, median age 61 years) or myeloablative (n = 80, median age 26 years) conditioning prior to HCT [Ustun et al. 2015]. On a retrospective analysis, MRD was present in 10 patients who received RIC and 15 patients who received myeloablative HCT. Among the RIC arm, the presence of MRD was associated with higher relapse rate (HR, 1.38; 95% CI, 1.7–8.7), DFS (HR, 2.9; 95% CI, 1.4–5.9), and OS (HR, 3.4; 95% CI, 1.7–7.0). A similar association was not seen in the patients who received myeloablative conditioning. This suggests a potential role for monitoring MRD to ascertain MRD-negative status prior to NMA or RIC allo-HCT.

Current state and future directions

In the assessment of potential HCT recipients, clinicians should take into account the treatment strategy as a whole, including the induction chemotherapy used, and the degree of myeloablation achieved by the RIC regimen itself [Blaise et al. 2007]. A use of lower-intensity induction therapy such as 10-day decitabine, which is associated with a CR rate of up to 40–50% even in AML patients with complex karyotype, may preserve the performance status and organ function of patients as they proceed to HCT [Ritchie et al. 2013]. Early incorporation of multidisciplinary care, including attention to physical therapy, nutrition and psychosocial health may improve a patient’s fitness to undergo and tolerate HCT in general. Additionally, in the selection of conditioning regimens, clinicians should take into consideration multiple factors, such as performance status and overall health status of an individual patient, a patient’s preferences of accepted level of mortality risk, status of the disease and availability of clinical trials of novel agents. In 2005, an HCT comorbidity index was introduced as a tool to assess comorbidities in HCT recipients [Sorror et al. 2005]. This was further modified to form a composite comorbidity or age index where anyone aged 40 years or over was assigned a value of 1 in addition to the other parameters in the HCT comorbidity index [Sorror et al. 2014]. This tool demonstrated a higher predictive value for TRM and OS as compared with age alone. Similarly, comprehensive geriatric assessment can be utilized to determine the fitness of older patients [Stuck et al. 1993; Muffly et al. 2013]. Recently, a modified European LeukemiaNet classification for patients with AML undergoing HCT in CR1 has also been evaluated that utilizes molecular markers to stratify patients into four prognostic groups [Oran et al. 2015]. Utilization of such tools can guide better selection of conditioning intensity in an individual.

At the present time, the aforementioned studies indicate a lower risk of TRM with RIC or NMA HCT in older AML patients with a possibility of higher risk of relapse, compared with myeloablative HCT. Hence, NMA or RIC HCT may be utilized in older AML patients who achieve CR1 with induction chemotherapy, are less fit, or have a high HCT comorbidity index and hence are not candidates for myeloablative HCT. NMA or RIC HCT may also be the preferred approach in patients who are older than 70 years old. Conversely, myeloablative HCT may be considered in fit patients with a low HCT comorbidity index, aged up to 65–70 years, who have achieved remission and accept a higher risk of TRM. Frail older patients, or those with active disease should be encouraged to enroll in well-designed clinical trials of novel therapies. The current National Comprehensive Cancer Network (NCCN) Guidelines (available at http://www.nccn.org) recommend considering RIC allogeneic HCT at CR1 in patients aged 60 years and older with minimal comorbidities who have an available donor, and patients with low-volume disease who fail induction therapy (preferably in the context of a clinical trial).

A number of strategies, such as early tapering of immunosuppression, monitoring of MRD after HCT, pre-emptive therapy of impending clinical relapse and post-transplant maintenance with chemotherapy, immunotherapy or cellular therapy are being explored to reduce the risk of relapse after HCT [Wayne et al. 2013; Gress et al. 2013]. Whether such strategies can improve the outcomes of RIC and NMA HCT will require prospective trials. Recently, ‘microtransplantation’ has been reported as a novel HCT strategy in older patients without a matched-related donor. In this strategy, the infusion of granulocyte colony-stimulating factor–mobilized donor peripheral-blood stem cells following induction and consolidation chemotherapy in elderly patients resulted in higher CR rates and DFS without GVHD [Guo et al. 2011]. This emerging HCT technique should be further investigated. Finally, survivorship programs are integral to enhance long-term outcomes and quality of life of older transplant recipients.