Allogeneic transplant: does age still matter?

Tsila Zuckerman

In this issue of Blood, on behalf of the Center for International Blood and Marrow Transplant Research (CIBMTR), Muffly et al report on the incidence and outcome of allogeneic stem cell transplantations (alloSCTs) performed between 2000 and 2013 in 1106 patients aged ≥70 years.1

The incidence of alloSCT in this age group accounted for 0.1% of all stem cell transplantations (SCTs) reported to the CIBMTR in 2001, and it rose to 3.85% by 2013. Comparison of 2 time periods of performing SCTs, 2000 to 2007 and 2008 to 2013, revealed significant improvement in overall survival (OS) and progression-free survival (PFS) rates in the later period (26% vs 39%; P < .001 and 22% versus 32%; P = .003, respectively). At the same time, the 2-year treatment-related mortality (TRM) remained unchanged, varying between 33% and 35% (P = .54). The upper age limit of alloSCT for acute myeloid leukemia (AML), used as a curative modality, has steadily increased over the past 4 decades from 40 to 45 to 70 to 75 years. However, improvements in the survival curves of older adults over the same period, modest as they are, do not seem to reflect this quantum leap in the availability of the most potent antileukemic strategy.

The current retrospective registry study, despite its inherent patient selection bias, encompasses the scope of SCT outcomes in the elderly patient population.

Because the incidence of almost all hematological malignancies increases with age, particularly peaking in those aged >65 years,2 and given the fact that, at least in acute leukemia, older age is associated with high-risk genetic markers and poor prognosis,3 there arises the question of why older patients, who are mostly in need of alloSCT, are under-referred to this treatment modality. A simple explanation could be the scarcity of SCT data in this age group and the lack of a good comparator enabling the evaluation of SCT efficacy relative to other treatments, because the majority of older patients are not included in clinical trials and are often referred to supportive or low-dose therapies. A more complex issue, which is difficult to assess, is the personal beliefs of physicians and patients and reservations toward alloSCT, which is widely regarded as too aggressive for the elderly population, despite the availability of such treatment modalities as truly nonmyeloablative SCTs. However, because the world population is aging and life expectancy in the Western world is approaching 79.3 years, we cannot avoid confronting this issue. Moreover, the economic burden on the health care system associated with the selection of costly therapies, like alloSCT, needs to be considered.

Because the long-term outcome of patients in the SCT setting is a composite of TRM on the one hand and the disease risk index (DRI) on the other, these 2 parameters should be evaluated in every individual patient. Over the years, TRM has been reported to have significantly decreased, mainly due to better supportive care, use of reduced or nonmyeloablative regimens, and improved HLA matching.4,5 Despite an increase in unrelated donor availability, the CIBMTR has estimated that only 26% of eligible patients actually undergo alloSCT, and in AML patients in the age group of 65 to 74 years, it is only 10% as compared with the anticipated 40%.6

Although TRM is associated with chronological age, biological age may be even more clinically meaningful. The hematopoietic cell transplantation comorbidity index (HCT-CI) introduced by Sorror et al7 encompasses the variables influencing TRM on a biological basis. The composite comorbidity/age score is found to improve TRM prediction.8 Patients with scores <3 are considered to have acceptable rates of nonrelapse mortality (<20%), which is in line with the results of the current study, where decreased OS is associated with an HCT-CI score ≥3.

The DRI, which is estimated based on disease type and remission status at transplantation, is found to reliably segregate patients into 4 risk groups and predict posttransplant OS and PFS. This risk score is independent of the HCT-CI, implying that the 2 scoring systems should be employed concomitantly.9 However, some caution is in order, because the DRI score is based on a patient population with a median age of 49 years, and its applicability to the older population should be evaluated.

In AML, the 2017 recommendations of the European LeukemiaNet (ELN) suggest performing alloSCT if the expected relapse rate without the procedure is >35% to 40%.3 The integrated risk approach proposed by ELN, which is based on relapse risk and TRM risk, recommends alloSCT in first complete remission if the anticipated disease-free survival with alloSCT is >10% compared with the non-alloSCT approach.10

Because relapse is the primary cause of death in the current study as well as in other studies, one should aim at reducing the relapse rate by earlier referral to SCT and by incorporating innovative methods, such as the use of minimal residual disease evaluation, to monitor the depth of remission at various time points followed by rapid interventions and posttransplant immune modulation or targeted therapy. Achievement of these goals requires generation of quality data, preferably in the setting of controlled trials. With the use of new statistical methods, such as data mining, the prediction of SCT outcomes could become more precise and aid in alloSCT candidate selection.

Overall, this excellent study by Muffly et al, reporting a 2-year post-SCT OS rate of 40% for patients in the 7th and 8th decades of life, provides ample evidence for the safety and feasibility of alloSCT for older adults. This study makes a strong case that this patient population, where hematological malignancies reach their peak, deserves special attention and prospective reevaluation of applicable treatment modalities. This is the only way to move forward.


  • Conflict-of-interest disclosure: The author declares no competing financial interests.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.