|Home | About | Journals | Submit | Contact Us | Français|
Reduced-intensity-conditioning (RIC) hematopoietic-stem-cell-transplantation (HSCT) is markedly underutilized in the elderly, in part because the impact of advanced-age on outcomes is poorly understood. We retrospectively analyzed outcomes in 158 consecutive hematologic-malignancy patients aged ≥60 years (median, 63; range, 60–71) undergoing fludarabine/busulfan-based RIC, with a median-follow-up of 34 months (range, 12.0–85.7). Multivariate analysis was undertaken for factors impacting outcome. For the patients aged ≥60 years, 2-year non-relapse-mortality (NRM) and relapse was 10% and 54.6% respectively. 2-year overall and progression-free survival (OS, PFS) was 46% and 35% respectively. Grade II–IV acute and chronic GVHD incidence was 19.6% and 45.9% respectively. Comparing 110 patients aged 60–64 years vs. 48 patients aged ≥65 years, 2-year NRM and relapse was 10.5% vs. 8.3% (p=0.84) and 53.5% vs. 56.3% (p=0.31) respectively. Grade II–IV acute and chronic GVHD incidence was 19.1% vs. 22.9% (p=0.52) and 51.8% vs. 32.5% (p=0.01) respectively. 2-year OS and PFS was 49% vs. 41% (p=0.11) and 36% vs. 35% (p=0.24) respectively. In a multivariate Cox-model, high-risk disease associated with poorer PFS (HR=2.1, p=0.01) and OS (HR=1.84, p=0.03); AML/MDS diagnosis (HR=1.66, p=0.03) and matched-related donor (HR=1.62, p=0.03) associated with poorer PFS. RIC HSCT is well-tolerated with reasonable survival in elderly patients. Age is not associated with impaired outcomes. HSCT should not be excluded solely based on advanced patient age.
The incidence of hematologic malignancies rises with age. For instance, acute myeloid leukemia (AML) has a ten-fold greater incidence in adults aged ≥65 years compared to adults aged 20–44 years. 1,2 Survival is also markedly impaired in elderly AML patients aged ≥60 years, even for relatively healthy older patients enrolled in acute leukemia cooperative group chemotherapy trials. 1,3 Impaired survival in older patients is due to increased chemotherapy-related toxicity and adverse disease biology. 3 Similarly, with regards primary myelodysplastic syndromes (MDS), 75% of patients are aged ≥60 years. Stratified by International Prognostic Scoring System (IPSS) risk, survival for older patients with low and intermediate-1 IPSS-risk MDS is markedly impaired compared to those aged under 60 years. 4 Elderly patients aged ≥60 years also have disproportionately impaired outcomes in acute lymphoblastic leukemia, Hodgkin and non-Hodgkin lymphomas. 5–10
Reduced-intensity-conditioning (RIC) hematopoietic stem cell transplantation (HSCT) is a curative therapeutic option for older hematologic malignancy patients as well as for younger patients with significant comorbidities who are not candidates for high-intensity myeloablative HSCT. However, there is limited direct information regarding the impact of age greater than 60 years on RIC HSCT outcomes. Studies reporting on RIC HSCT outcomes typically include younger patients aged below 60 years in their analyses. 11–13 Hematologic malignancy patients younger than 60 years are often preferential candidates for myeloablative HSCT, with RIC HSCT reserved for those with impaired performance status and/or significant comorbidities. Therefore, the inclusion of patients aged under 60 years in these reports is a source of clinical heterogeneity, increases bias, and hinders assessment of the impact of age on RIC HSCT outcomes.
There is even less information on RIC HSCT outcomes in patients aged ≥65 years, the typical age for Medicare coverage eligibility in the USA. A-priori it is possible that patients of such advanced age have more aggressive disease biology, further impaired physiologic reserve, increased treatment-related toxicity and worse RIC HSCT outcomes even compared to those aged below 65 years. Assessing the clinical effectiveness of RIC HSCT in this patient subgroup is therefore a priority, both to determine its clinical utility and to justify the expenditure of limited health care resources.
We undertook a retrospective study of consecutive hematologic malignancy patients aged ≥60 years treated with a consistent fludarabine/low-dose busulfan RIC HSCT regimen at the Dana-Farber Cancer Institute. In additional analyses, we compared outcomes in patient cohorts aged 60–64 years with those aged ≥65 years, to directly assess for the impact of advanced age on RIC HSCT outcomes.
Between January 2002–June 2008, 158 consecutive hematologic malignancy patients aged ≥60 years underwent adult donor RIC HSCT at the Dana-Farber Cancer Institute. Three patients underwent multiple RIC transplants after relapsing; in these cases, only the first transplant was used in the analysis. The patients provided IRB approved informed consent for data analysis.
All patients received a consistent RIC regimen comprising fludarabine (30mg/m2 IV) and once- or twice-daily busulfan (0.8 mg/kg IV) × 4 days. 152 patients (96%) received peripheral-blood stem-cell infusions. Graft-versus-host-disease (GVHD) prophylaxis regimens included: i) tacrolimus/sirolimus ± mini-methotrexate (93); ii) tacrolimus/mini-methotrexate (50); iii) CD8 T-cell depletion + tacrolimus/mini-methotrexate (3); and other (12). Tacrolimus and sirolimus were started on day −3, with a goal to taper-off by day +180 in the absence of GVHD. All patients received filgrastim at 5 μg/kg SC daily, from day +1 until an absolute neutrophil count (ANC) >1000 cells/μl; and a minimum of 12 months of Pneumocystis jiroveci and HSV/VZV prophylaxis. A pre-emptive treatment strategy with ganciclovir or valganciclovir was used if CMV reactivation was detected on routine monitoring in the first 100 days after SCT. No pre-emptive or prophylactic donor lymphocyte infusions were given after HSCT.
Neutrophil and platelet engraftment was assessed by the number of days to ANC≥500/μl and platelets≥20,000/μl respectively, in the absence of transfusions. Unfractionated donor chimerism was assessed from bone marrow aspirates and/or peripheral blood at approximately day +30–45, and 3–4 months after transplant. Genotype of donor and recipient were determined using DNA extracted from pre transplant samples, and percent donor chimerism was determined by analyses of informative short tandem repeat (STR) loci using the ABI Profiler-Plus Kit (Applied Biosystems Inc.) and the ABI 310 Genetic Analyzer. Acute GVHD was assessed per consensus grading. 13
Descriptive statistics was provided for patient baseline characteristics. Two-sided Fisher’s exact test was used to compare categorical variables between age groups, and two-sided Wilcoxon-Rank-Sum test was used to compare continuous variables between age groups.
Cumulative incidence curves for grade II–IV acute GVHD and chronic GVHD were constructed reflecting early death and death or relapse as a competing risk, respectively. Cumulative incidence curves for treatment-related death and relapse with or without death were constructed reflecting time to relapse and time to treatment related death as competing risks. The difference between cumulative incidence curves in the presence of a competing risk was tested using the Gray method. 14
Time-to-relapse and time-to-non-relapse-death were measured from the date of stem cell infusion. Patients who were alive without relapse were censored at the time last seen alive. Overall survival (OS) and progression-free survival (PFS) were calculated using the Kaplan-Meier method. Overall survival was defined as the time from stem cell infusion to death from any cause. Progression-free survival was defined as the time from stem cell infusion to relapse, disease-progression or death from any cause. The Log-rank test was used for the comparisons of Kaplan-Meier curves, whereas the Gray test was used for the comparisons of cumulative incidences of NRM and relapse. Prognostic factors for overall and progression-free survival were examined in Cox proportional hazard models, whereas relapse and non-relapse mortality were examined in competing risks regression model. 15 Interactions between covariates were examined in the Cox model and none was significant.
Peri-transplant characteristics of the 158 patients included in this study are shown in Table 1. The median patient age was 63 years (range, 60–71). Median follow-up time among survivors was 34.0 months (range, 12.0–85.7) post HSCT. There were 106 male and 52 female patients. The principal diseases were myeloid in 70% and lymphoid in 27%. Seventy-six percent had high-risk disease (i.e. acute leukemia in relapse or ≥ CR2, MDS RAEB or secondary MDS, CML beyond CP1, lymphoma beyond first remission) and 12% had received prior autologous transplantation. Matched-unrelated (MUD), matched-related (MRD) and 1–2 HLA locus (-A, -B, -C, -DRB1) mismatched (MM) adult donors were used in 56%, 34%, and 10% of patients respectively. The median stem cell dose was 8.26 (range, 1.66–47.67) × 106 CD34+ cells/kg; and was not different between MUD/MM versus MRD HSCT, at 8.39 (range, 2.35–47.67) × 106 CD34+ cells/kg versus 8.05 (range, 1.66–23.18) × 106 CD34+ cells/kg, respectively (p=0.13).
For the cohort analysis, there were 110 patients aged 60–64 years (median, 62; range, 60–64) and 48 patients aged ≥65 years (median, 67; range, 65–71). The median follow-up time among survivors for the cohorts was 35.0 months (range, 12–85.7) and 32.4 months (range, 12.7–72.8) respectively (p=0.40). The cohorts did not differ significantly with regards covariates of patient-sex, patient-donor-sex-match, prior autologous transplant, donor stem-cell source, disease-risk at transplant, donor-type (matched-related, matched-unrelated, mismatched), ECOG performance-status at transplant, GVHD prophylaxis regimen, year of transplant, donor CMV status and donor-age (Table 1).
In this patient population aged ≥60 years, matched-unrelated and mismatched donors (MUD/MM) were more common than matched-related donor (MRD) transplants, at 104 and 54 respectively; and as expected the MUD/MM donors were significantly younger with median age of 31 years (range, 18–59) versus MRD median age of 59 years (range, 30–73) (p=<0.0001). However, as noted above, the median donor age was not different for patient cohorts of ≥65 years versus those aged 60–64 years. In addition, there was no survival difference when OS and PFS were examined by donor age categories of less than or equal to 30 years; 31–40 years; and greater than 40 years.
The median time to neutrophil engraftment was 13 days (range, 2–70) in the 81 patients whose ANC nadired below 500 cells/μl. The median time to platelet engraftment was 20 days (range, 11–78) in the 79 patients whose platelet-count nadired below 20,000 cells/μl. The median donor chimerism between 20–50 days after RIC HSCT was 94% (range, 3–100), and between 90–120 days after RIC HSCT was 94% (range, 1–100). There was no clinically relevant difference in time to neutrophil and platelet engraftment between patients aged 60–64 years and those aged ≥65 years. Median time to neutrophil engraftment was 13 days (range, 2–70) in the 59 patients aged 60–64 years who experienced a nadir, and 15 days (range, 3–33) in the 22 patients aged ≥65 years who experienced a nadir. The median time to platelet engraftment was 20 days (range, 11–78) in the 58 patients aged 60–64 years who experienced a nadir and 19 days (range, 12–60) in the 21 patients aged ≥65 years who experienced a nadir.
The cumulative incidence of relapse (or progression) and NRM at 2 years was 54.6% and 10% respectively (Figure 1a). OS and PFS at 2 years was 46% (95% CI, 38–54) and 35% (95% CI, 28–43) respectively (Figure 1b). When we compared patients aged 60–64 years with those aged ≥ 65 years, the cumulative incidence of relapse (or progression) and NRM at 2 years was 53.5% versus 56.3% (p=0.31), and 10.5% versus 8.3% (p=0.84) respectively (Figure 2a). OS and PFS at 2-years were 49% (95% CI, 39–58) versus 41% (95% CI, 26–54) (p=0.11), and 36% (95% CI, 27–45) versus 35% (95% CI, 22–49) (p=0.24) respectively (Figure 2b). In both cohorts, early relapse was the primary cause of treatment-failure and death, with 100-day relapse/progression-rate of 27% versus 42% (p=0.09) and 100-day mortality rate of 11% versus 27%, p=0.02) respectively.
The maximum cumulative incidence of grade II–IV acute GVHD was 20.3%. Cumulative incidence of chronic GVHD at 2 years was 45.9%. Comparing patients aged 60–64 yrs with those aged ≥ 65 yrs, the maximum cumulative incidence of grade II–IV acute GVHD was 19.1% versus 22.9% (p=0.52) respectively, and chronic GVHD at 2 years was 51.8% versus 32.5% (p=0.01) (Figure 2c, d). The higher early mortality rate in patients aged ≥65 years may partly account for the lower chronic GVHD rate in this cohort.
Importantly, outcomes were not impaired in those lacking a matched-related donor. 2-year OS and PFS of MUD/MM versus MRD HSCT was 52% (95% CI, 42–62) versus 35% (95% CI, 22–47) (p=0.10), and 43% (95% CI, 33–53) versus 20% (95% CI, 11–32) (p=0.006) respectively. With regards advanced age and donor outcomes, MUD/MM recipients aged ≥65 years did not experience poorer outcomes compared to those aged 60–64 years, with 2-year OS and PFS of 50% (95% CI 31–67) versus 53% (95% CI, 41–64) (p=0.69) and 45% (95% CI, 27–61) versus 43% (95% CI, 32–55) (p=0.65) respectively. However, MRD recipients aged ≥65 years experienced poorer 2-year OS, but not PFS, compared to those aged 60–64 years, at 24% (95% CI, 7–45) versus 40% (95% CI, 24–55) (p=0.04), and 18% (95% CI, 4–38) versus 21% (95% CI, 10–36) (p=0.19) respectively.
We evaluated a competing risk regression model for relapse and NRM incidence, including covariates of diagnosis (AML/MDS versus non-AML/MDS), disease-risk (high versus low), patient-donor-sex-match (gender-matched versus mismatched), donor-type (matched-related versus matched-unrelated or mismatched), patient-age (dichotomized at age 65 years), prior autologous transplant (yes versus no), year of transplant (continuous variable) and GVHD prophylaxis (sirolimus containing versus not) (Table 2). Use of sirolimus-based acute GVHD prophylaxis was associated with reduced NRM (HR=0.29, 95% CI 0.10–0.80, p=0.02), and AML/MDS diagnosis at transplant was borderline associated with relapse (HR=1.60, 95% CI 0.96–2.67, p=0.07). Age ≥65 years was not associated with increased NRM (HR=0.72, 95% CI 0.21–2.47, p=0.60) nor increased relapse incidence (HR=1.14, 95% CI 0.70–1.87, p=0.60).
We also evaluated a multivariate Cox model of survival outcomes (PFS, OS), including covariates of diagnosis (AML/MDS versus non-AML/MDS), disease-risk (high versus low), patient-donor-sex-match (gender-matched versus mismatched), donor-type (MRD versus MUD/MM), patient age (continuous variable, or dichotomized at age 65 years), prior autologous transplant (yes versus no), year of transplant (continuous variable) and GVHD prophylaxis (sirolimus containing versus not) (Table 3). Donor age and interactions between covariates were examined, but none were significant and were thus not included in the final model. High-risk disease at transplant was associated with poorer PFS (HR=2.10, 95% CI 1.23–3.58, p=0.01) and with poorer OS (HR=1.84, 95% CI 1.05–3.22, p=0.03). AML/MDS diagnosis at transplant (HR=1.66, 95% CI 1.04–2.66, p=0.03) and matched-related donor (HR=1.62, 95% CI 1.05–2.50, p=0.03) were associated with poorer PFS. However, age ≥65 years was not associated with poorer PFS (HR=1.11, 95% CI 0.71–1.75, p=0.64) nor poorer OS (HR 1.32, 95% CI 0.83–2.10, p=0.25).
Hematologic malignancies are more frequent in patients aged ≥60 years, and they have a worse prognosis with chemotherapy as compared with younger adults. Most patients are not considered for RIC HSCT, despite its tolerability and curative potential in older patients with otherwise fatal hematologic cancers. For instance, at a major US cancer center, despite a protocol mandating that all AML patients aged ≥50 years be evaluated by a HSCT physician with a default plan for RIC HSCT in first complete remission (CR) if they had a matched related or unrelated donor, only 53 of 99 patients (54%) in CR actually underwent such an evaluation. Of these, 26 patients had an HLA-matched donor and 14 patients (14% of CR patients) underwent RIC HSCT. Even within the cohort of patients aged >50 years, the younger patients (<60 years) were significantly more likely to be referred for HSCT evaluation (p=0.01). 16 It is clear that for hematologic malignancy patients aged ≥60 years, the likelihood of receiving potentially curative RIC HSCT is markedly low, despite their known dismal prognosis with non-transplant therapy.
The under-utilization of curative RIC HSCT may be due, in part, to the limited data on transplant outcomes in older patients. Very few analyses restricted to patients aged ≥60 years have been reported. Two small retrospective analyses of 17 and 24 patients, with limited median follow-ups of 13 months and 21 months respectively, reported NRM of 33% and 25% respectively, and OS of 29% at 1-year and 44% at 2-year, respectively. 17,18 The limited sample sizes precluded any assessment of the impact of advance age ≥65 years.
Prospective RIC HSCT data for patients aged ≥60 years are also limited. A small prospective trial utilizing fludarabine-based RIC involving 19 AML/MDS patients aged ≥60 years (median, 64 years; range 60–70) and 27-month median follow-up has been reported. 19 It indicated 1-year OS and PFS of 68% and 61% respectively. Another small trial with fludarabine/low-dose total-body-irradiation RIC involving 32 hematologic malignancy patients aged ≥60 years (median, 62 years; range 60–70) and a short 17-month median follow-up reported 1-year NRM and relapse rates of 10% and 50% respectively, with OS and PFS of 53% and 43%, respectively. 20 The 95% confidence-intervals of these estimates remain wide, owing to limited sample size. An assessment of the impact of advanced age ≥65 years on RIC HSCT outcomes remains unavailable, also likely due to limited sample size. It therefore remains uncertain whether patients aged ≥65years experienced outcomes similar to those of patients aged 60–64 years.
A multi-center CIBMTR retrospective analysis evaluating RIC HSCT outcomes in AML/MDS patients aged ≥40 years has been reported in abstract form. 21 Comparing patient cohorts aged 40–54 years, 55–60 years, 60–64 years and ≥65 years, the authors report that age was not significantly associated with survival outcomes. As previously discussed, the comparison of younger patients with those aged ≥60 years is problematic and represents a source of bias, as elderly patients aged ≥60 years likely have more biologically aggressive disease than younger patients. There is also further increased selection bias since younger RIC HSCT patients (who would typically be candidates for myeloablative HSCT) likely have more comorbidities compared to the elderly patient cohorts, that can confound survival outcomes as well. 22 Additionally, inclusion of patients from multiple centers, with inter-center variability with-regards patient referral, evaluation, selection, conditioning regimens, GVHD prophylaxis, post-transplant care and data collection and reporting are additional sources of heterogeneity and bias. It is therefore reassuring that we also found no significant survival differences between the 37 AML/MDS patients ≥65 years versus the 61 patients aged 60–64 years, with 2-year OS and PFS of 39% versus 45% (p=0.46) and 38% versus 27% (p=0.82), respectively.
Our analysis also has important inherent limitations. As a retrospective single-institution analysis, it is susceptible to bias in patient selection, and the generalizability of our results to RIC HSCT outcomes at other centers remains a question. We could not assess the impact of comorbidities directly as routine HCT-comorbidity index scoring was only instituted recently and is therefore unavailable for many patients in this analysis. However, for all patients ≥60 years selected for RIC HSCT, significant comorbidities were an exclusionary factor per our consistent institutional paradigm. We also cannot comment on RIC HSCT outcomes in patients of more extreme age (i.e. those in their 70s and older), where additional data is required. The strength of this analysis is its primary focus on elderly patients aged ≥60 years, all transplanted at a single center, with more consistent patient referral, evaluation and selection, as well as a consistent RIC regimen, GVHD prophylaxis and post-transplant monitoring. Such consistency also improves the comparability between patients aged 60–64 years and those aged ≥65 years.
Our data indicate that for appropriately selected patients aged ≥60 years, RIC HSCT is well tolerated with a 2-year cumulative NRM incidence of 10%, and with acute grade II-IV and chronic GVHD rates of 19.6% and 45.9% at 200-days and 2-years respectively. In this elderly hematologic malignancy population, survival after RIC HSCT was reasonable, especially for these otherwise fatal illnesses, with a 2-year OS and PFS of 46% and 35% respectively. Importantly, RIC HSCT outcomes for patients lacking matched-related donors (as is common in older patients) were not impaired. As in younger patients, early relapse is a major cause of death. Novel approaches to reduce relapse (e.g., vaccine-based strategies to boost immunologic graft-versus-tumor effect23) without increasing GVHD remain a priority.
Specifically with regards the impact of advanced age, patients aged ≥65 years did not experience impaired engraftment or greater treatment-related toxicity (NRM, GVHD), and their survival was comparable to patients aged 60–64 years. These findings are also concordant with the preliminary results, reported in abstract form, of a multi-center retrospective analysis of hematologic malignancy patients aged ≥60 years, wherein advanced age was not associated with poorer outcomes after RIC HSCT. 24
In summary, our analysis indicates that RIC SCT is a viable treatment option for appropriately selected hematologic malignancy patients aged ≥60 years. While early disease relapse remains problematic, GVHD and NRM were acceptable and did not appear to increase with advancing age. Indeed, advanced disease at time of transplant and a diagnosis of AML/MDS, but not advanced age, were associated with poorer survival. Advanced age should not be the basis for excluding hematologic malignancy patients from potentially curative RIC HSCT.
Supported in part by NIH grant CA142106
Conflicts of Interest: None
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.