High-dose cytarabine induction is well tolerated and active in patients with de novo acute myeloid leukemia older than 60 years


  • E.W., M.A., and H.J.K. designed the research, reviewed the data, and wrote the article; E.W., K.B., L.T.H., M.M., and A.L. treated patients, reviewed the data, and approved the article; D.H. and J.N. assisted with data collection and reviewed the article; and L.P. and M.T. analyzed the data and approved the article.



High-dose cytarabine (HiDAC) is safe and very effective in younger patients with acute myeloid leukemia (AML), but it generally is not well tolerated in the elderly.


The authors explored the safety and tolerability of a modified HiDAC induction regimen consisting of 6 daily doses of cytarabine at 2 g/m2 in combination with 3 daily doses of daunorubicin at 45 mg/m2 in 59 consecutive patients aged >60 years who had de novo AML diagnosed between July 1996 and February 2005.


The median patient age was 68 years (range, 60-86 years). The regimen was well tolerated. Infections were common and occurred in 39% of patients, but cerebellar toxicities occurred in only 7% of patients and were reversible. The day-30 induction-related mortality rate was 10%. Overall, 69% of patients achieved complete remissions (CR), and 80% received up to 3 consolidations with HiDAC. The median follow-up for surviving patients was 53 months (range, 17-114 months). The median overall survival was 15.3 months (range, 1-114 months), and the relapse-free survival was 13.8 months (range, 1-113 months). Survival for patients who achieved CR was 27 months (range, 2-114 months).


The modified HiDAC regimen was well tolerated in patients aged >60 years with AML and was associated with low induction mortality and high rates of CR. Nevertheless, these high remissions still were associated with poor overall outcomes. Cancer 2011;. © 2011 American Cancer Society.

The incidence of acute myeloid leukemia (AML) increases with age.1 In the United States, the median age of patients with AML is 67 years at diagnosis, and the age-adjusted incidence is 15.9 per 100,000 for those aged >65 years compared with 1.7 per 100,000 for those aged <65 years (Surveillance, Epidemiology, and End Results data, 2003-2007).1 Outcomes in older patients (aged >60 years) are poor because of chemotherapy-related toxicity and inherent, resistant disease.2 Indeed, it is estimated that the median overall survival (OS) of patients with AML aged >65 is months, and the 2-year survival rate is <10%.3

The achievement of complete remission (CR) appears to be a prerequisite to long-term survival.4 However, the degree of toxicity to which patients should be subjected to achieve CR remains a topic of debate. Ongoing clinical trials are incorporating comorbidity scores to assist in this complex decision.5-7

High-dose cytarabine (HiDAC)-containing regimens are effective and well tolerated in younger patients with AML8 but generally are poorly tolerated in patients aged >60 years.9-12 Estey et al reported less neurotoxicity using intermediate dose cytarabine (500 mg/m2 every 12 hours for 9-25 doses).13 In an attempt to reduce HiDAC-related toxicities and preserve antileukemic activity in this patient population, we tested the safety and efficacy of a modified HiDAC regimen in which cytarabine was administered daily at 2 g/m2 per dose for 6 consecutive days. This report focuses on the safety and tolerability of this regimen in patients with de novo AML aged ≥60 years.



The patient population consisted of adults aged ≥60 years with histologically confirmed and previously untreated AML who were referred for induction chemotherapy at Emory University between June 1996 and February 2005. Patients with acute promyelocytic leukemia were excluded. Repeat bone marrow biopsy or institutional review of outside material was performed for all patients who were diagnosed at another institution. The diagnosis of AML initially was defined according to the French-American-British Cooperative Group14 and, later, according to the World Health Organization classification: ie, documentation of ≥20% myeloid blasts in blood or bone marrow.15 Patients with secondary AML, defined by the presence of an antecedent hematologic disorder, or who had received prior chemotherapy or radiation to treat a pre-existing malignant or benign condition received alternative regimens and, thus, are not included in this report. Cytogenetic abnormalities were classified as favorable, intermediate, unfavorable, or unknown according to South West Oncology Group (SWOG) criteria.16 Patients were either enrolled on or treated according to an institutional phase 2 clinical trial. This study was approved by the Emory University Institutional Review Board.


Modified HiDAC induction therapy consisted of 2 planned courses of intravenous cytarabine given at a dose of 2 g/m2 daily over 4 hours for 6 consecutive days in combination with 3 days of daunorubicin (45 mg/m2) administered on days 2 through 4. Patients without blast clearance based on the day-14 bone marrow biopsy (>5% blasts in bone marrow with >15% cellularity) received the second induction course beginning between days 15 and 22 or after resolution of nonhematologic toxicities from the first induction. Patients who achieved CR after 1 cycle received the second course upon recovery of counts and resolution of toxicities from the first course of therapy. Patients who did not achieve CR after 2 inductions were offered alternative therapies. All patients who achieved CR after induction received a first HiDAC and daunorubicin consolidation followed by risk-adapted consolidation. Patients who were considered unfit for transplantation and those who had favorable cytogenetics received 2 courses of HiDAC at 2 g/m2 intravenously daily for 6 consecutive days. Fit patients with intermediate-risk cytogenetics were offered an autograft; those with unfavorable cytogenetics and a suitable donor underwent allogeneic hematopoietic stem cell transplantation. Patients received supportive care according to institutional standards. Patients' demographic data (age, sex), disease-related information (histology, white blood cell [WBC] count at diagnosis, cytogenetics), response to therapy, bone marrow biopsy results at days 14 and 30, relapse, and survival information were collected.

Response Criteria and Statistical Methods

Disease status was assessed using criteria developed by a workshop sponsored by the National Cancer Institute and revised by the International Working Group.17 CR was defined as the recovery of counts (absolute neutrophil count ≥1000/mcL and platelet count ≥100,000/mcL) with a normocellular bone marrow and <5% blasts. Relapse-free survival (RFS) was calculated from the date of remission to the date of relapse, last contact, or death. OS was calculated from the time of diagnosis to the time of death or last contact. RFS and OS were estimated by using the Kaplan-Meier method. A Cox proportional-hazards model was used to estimate the unadjusted and adjusted effects of prognostic factors (age, cytogenetics, presenting leukocyte count, and the number of comorbidities) on RFS and OS. The significance level was set at P = .05 for all analyses. The SAS statistical package (version 9.2; (SAS Institute, Inc., Cary, NC) was used for data management and all statistical analyses.


Patient Characteristics

Between July 1996 and February 2005, 59 consecutive patients aged ≥60 years with newly diagnosed, previously untreated, de novo AML received modified HiDAC induction chemotherapy as described above. The median patients age was 68 years (range, 60-86 years), and 20 patients (34%) were aged >70 years. Cytogenetic abnormalities were favorable in 7% of patients, intermediate in 61%, unfavorable in 25%, and unknown in 7% (Table 1). The median leukocyte (WBC) count at presentation was 20,200/μL (range, 700-278,500/μL). Comorbidities18 before the initiation of induction were present in 55 of 59 patients (93%) and are listed in Table 2.

Table 1. Patient Characteristics (N = 59)
Patient CharacteristicNo. of Patients (%)
Age, y 
 Median [range]68 [60-86]
 60-7039 (66)
 70-8019 (32)
 80-861 (2)
 Women29 (49)
 Men30 (51)
Cytogenetic risk group 
 Favorable4 (7)
 Intermediate36 (61)
 Unfavorable15 (25)
 Unknown4 (7)
Table 2. Comorbidities at Baseline
Patient Comorbidities, n = 55No. of Patients (%)
  • Abbreviations: EF, ventricular ejection fraction

  • a

    Renal toxicity was defined as a serum creatinine level up to twice the upper limit of normal.

  • b

    These were solid tumors that were treated with surgery or hormone therapy only.

Pulmonary24 (44)
Cardiac25 (45)
 Arrhythmia6 (11)
 Heart valve disease3 (5)
 EF ≤50%/coronary artery disease16 (29)
Renala22 (40)
Diabetes9 (16)
Cerebrovascular disease9 (16)
Psychiatric disturbance5 (9)
Infection, active12 (22)
Rheumatologic6 (11)
Prior solid tumorb9 (16)

Safety and Tolerability

Regimen-related toxicities are listed in Table 3. Infection was the most common toxicity (39%) and led to the death of 4 patients during induction and 4 during consolidation. Central nervous system toxicity, including cerebellar toxicity (n = 4) and bleed or stroke (n = 2), occurred during induction in 6 of 59 patients (10%). Cerebellar ataxia was reversible in all patients. Six patients (10%) died by day 30, and 5 patients died between days 30 and 60, for a 60-day mortality rate of 19%. Day-30 and day-60 mortality rates were not affected by age; specifically, no difference was observed between patients aged <70 years or >70 years. Causes of death by day 60 included sepsis (4 patients), pneumonia (2 patients), intracranial hemorrhage or stroke (2 patients), and progressive disease (3 patients).

Table 3. Grade ≥3 Toxicity
 No. of Patients (%)
  1. Abbreviations: CNS, central nervous system.

Fever and neutropenia13 (22)6 (10)
Pneumonia10 (17)10 (17)
Arrhythmia10 (17)0 (0)
Acute renal failure5 (8)1 (2)
Cerebellar3 (5)1 (2)
Other CNS2 (3)0 (0)
Congestive heart failure2 (3)0 (0)
Hepatic dysfunction1 (2)0 (0)
Septic death4 (7)4 (7)


Overall, 41 of 59 patients (69%) achieved CR after 1 induction (n = 36; 88%) or 2 inductions (n = 5; 12%). Eighteen patients with primary resistant disease received alternative therapies and died. Of the patients who achieved first CR postinduction, 33 of 41 (80%) received consolidations with 1 course (n = 6), 2 courses (n = 11), or 3 courses (n = 10) of HiDAC, and 6 underwent hematopoietic stem cell transplantation in first CR (5 autologous and 1 allogeneic, matched-sibling donor transplantation). The remaining 8 patients in first CR did not receive the planned consolidations because of cerebellar toxicity (n = 1), refusal to continue therapy (n = 1), physician preference (n = 5), or loss to follow-up (n = 1). Alternative consolidations were administered in 5 patients and included mitoxantrone and etoposide (n = 3), fludarabine with topotecan and gemtuzumab ozogamicin (n = 1), and 5-azacytidine (n = 1).


The median follow-up for surviving patients was 53 months (range, 17-114 months). Overall, the median survival for the entire cohort was 15.3 months (range, 1-114 months) (Fig. 1), and the median OS for patients who achieved CR was 27 months (range, 2-114 months). It is noteworthy that all 4 patients who had favorable cytogenetics achieved remission, and their median survival was 39 months (range, 4-74 months).

Figure 1.

This Kaplan-Meier curve illustrates overall survival in patients aged ≥60 years with acute myeloid leukemia after high-dose cytarabine induction.

RFS for the entire cohort was 13.8 months (range, 1-113 months) (Fig. 2). Among the 23 patients who relapsed, 15 patients (65%) received salvage therapy with modified HiDAC (n = 5); azacitidine (n = 2); gemtuzumab ozogamicin (n = 5); etoposide/mitoxantrone (n = 1); and combined cytarabine, topotecan, and gemtuzumab ozogamicin (n = 2); and 5 of 15 patients (33%) achieved a second CR, including 1 patient w ho underwent allogeneic hematopoietic stem cell transplantation in second CR.

Figure 2.

This Kaplan-Meier curve illustrates recurrence-free survival in patients aged ≥60 years with acute myeloid leukemia after high-dose cytarabine induction.

Prognostic Factors

Univariate analysis indicated that neither age, nor WBC count, nor cytogenetics was associated with CR; whereas age influenced RFS, and cytogenetics influenced both RFS and OS. Cytogenetics independently predicted RFS and OS in multivariate analyses. The number of comorbidities (0, 1, 2, or >2) did not affect outcomes (Table 4).

Table 4. Factors Impacting Outcomes of Patients With de Novo Acute Myeloid Leukemia Aged ≥60 Years
  1. Abbreviations: CI, confidence interval; HR, hazard ratio; OR, odds ratio; WBC, white blood cells.

VariableOR (95% CI)POR (95% CI)P
Complete Remission    
 WBC0.999 (0.990-1.008).78220.998 (0.988-1.008).6681
 Age0.991 (0.901-1.090).85760.958 (0.856-1.072).4558
 Cytogenetic risk: Unfavorable vs favorable plus intermediate2.625 (0.759-9.084).12763.149 (0.832-11.919).0912
 Comorbidities1.025 (0.716-1.468).89201.264 (0.780-2.047).3413
VariableHR (95% CI)PHR (95% CI)P
Overall survival    
 WBC0.998 (0.992-1.003).34370.998 (0.992-1.004).5030
 Age1.046 (0.989-1.108).11771.059 (0.994-1.128).0740
 Cytogenetic risk: Unfavorable vs favorable plus intermediate0.414 (0.209-0.821).01160.435 (0.213-0.888).0220
 Comorbidities1.114 (0.931-1.333).23800.954 (0.773-1.177).6610
Relapse-free survival    
 WBC0.995 (0.987-1.002).16260.996 (0.989-1.003).2410
 Age1.092 (1.011-1.180).02541.109 (1.009-1.218).0322
 Cytogenetic risk: Unfavorable vs favorable plus intermediate0.229 (0.089-0.594).00240.267 (0.097-0.734).0105
 Comorbidities1.126 (0.922-1.375).24400.855 (0.658-1.109).2380


The treatment of older patients with AML is among the most challenging dilemmas for hematologists. In addition, currently, there is no commonly accepted standard of care for this patient population, and there is no reliable comorbidity scoring system to help identify patients who may tolerate induction chemotherapy. Low-dose arabinosylcytosine is an accepted standard treatment for older patients with AML in Europe based on the results from a randomized trial that demonstrated higher remissions (18% vs 1%) and improved survival (median, 80 weeks vs 10 weeks) with low-dose arabinosylcytosine compared with hydroxyurea.4 Previously, Menzin et al reported a median survival of 7 months for patients who received any type of “chemotherapy” compared with 1 month for those Medicare recipients who received supportive care alone.3

In patients aged ≥60 years, standard antileukemic induction chemotherapy (anthracycline plus cytarabine 100-200 mg/m2 daily for 7-10 days) yielded CR rates ranging between 38% and 48% with a 30% average early (30-day) mortality rate (Table 5). Given the rising incidence of AML concurrent with the aging population, recent efforts have focused on developing single agents (hypomethylating agents, clofarabine, cloretazine) and reduced-intensity combination regimens.19, 20 It is noteworthy that these single-agent regimens have yielded CR rates and survival comparable to those produced with combination therapies19 but with lower induction-related mortality and regimen-related toxicities in this vulnerable population.

Table 5. Summary of Outcomes of Induction Therapy in Patients Aged >60 Years
Study GroupRegimenPatientsCR (%)Early Death (%)
  1. Abbreviations: TAD9, 9-day combination of 6-thioguanine with cytosine arabinoside and daunorubicin; DAT, daunorubicin, cytarabine, and 6-thioguanine followed by 2 or 6 DAT consolidations; LDAC, low-dose Ara-C.

AMLCG23 TAD9 induction and consolidation ± maintenance5115127-34
BMRC24 DAT induction and consolidation63646-4830-52
CALGB25 Standard or intensified 7+3 ± maintenance55641-4731-54
SECSG26 Idarubicin vs Daunorubicin + Cytarabine1115320
EORTC-LCG-HOVON27 Mitoxantrone vs Daunorubicin ± LDAC3 maintenance48938-476-15
MDAC11 Cytarabine-based intensive induction43045%36
Clofarabine28 Clofarabine monotherapy1123810
Current reportModified HiDAC induction/consolidation5969%10

Our single-institution experience with the use of a modified HiDAC induction for patients with de novo AML aged ≥60 years had a high rate of CR with an early mortality rate of 10% by day 30 and <20% by day 60. Patients aged >60 years were offered an aggressive regimen based on the subjective impression in the treating physician that these patients could tolerate intensive induction. These high remission rates were not affected by age or cytogenetics and compare favorably with previous reports (Table 5). A recent report described the outcomes of patients with AML aged ≥70 years (median age, 74 years) who received various intensive cytarabine-based regimens (up to 2 g/m2 daily). In that study, 45% of patients achieved CR, and OS was 4.6 months. The early mortality rates were 26% and 36% at 4 weeks and 8 weeks, respectively.11

Despite the presence of comorbidities in 93% of our patients, the induction mortality was surprisingly low (10% by day 30 and 19% by day 60) compared with previous reports in older patients with AML who received induction chemotherapy. Cerebellar toxicity was a main concern when this regimen was conceived given a reported rate of up to 32% by the Cancer and Leukemia Group B in patients aged >60 years who received HiDAC consolidation.9 In our cohort, 7% of patients developed reversible cerebellar ataxia, which led to discontinuation of HiDAC in 1 patient. This relatively low incidence may have been related the to lower cumulative daily doses of cytarabine in our modified regimen.

In the current study, physician's subjective impression of the tolerability of this regimen, based on overall performance status, was the main criteria for enrollment. Therefore, despite the presence of comorbidities in 93% of patients, patient selection biases may have influenced the results of this analysis. In addition, modern molecular prognostic factors, such as fms-related tyrosine kinase 3 and nucleophosmin 1, were not available or known at the time these patients were treated. These factors may be useful in determining the prognosis for older patients with AML.21, 22

In summary, the modified HiDAC regimen described in this report was well tolerated in patients aged >60 years with de novo AML despite the presence of comorbidities and was associated with high CR rates (69%). However, these high CR rates were short-lived, suggesting that alternative consolidations are worth exploring.


No specific funding was disclosed.


The authors made no disclosures.