Comparison of allogeneic stem cell transplantation, high-dose cytarabine, and autologous peripheral stem cell transplantation as postremission treatment in patients with de novo acute myelogenous leukemia


  • The following Hospitals and Physicians participated in the study: First Department of Internal Medicine, National University of Athens, Medical School, “Laiko” General Hospital (D. Loukopoulos, A. M. Tsimberidou, N. Viniou, N. Stavroyianni, J. Meletis, C. Constantopoulos, E. Variami, T. Andreopoulos, Y. Rombos, and X. Yataganas); Department of Clinical Propedeutic National University of Athens Medical School, “Laiko” General Hospital (P. Panagiotidis); Second Department of Internal Medicine, National University of Athens Medical School, Hippokration General Hospital (Th. Kalmantis and M. Belegrati); Department of Internal Medicine, Division of Hematology,University Hospital of Patras, Rion-Patras (N. Zoumbos, T. Matsouka, M. Tiniakou, A. Symeonidis, A. Kourakli, and P. Zikos); First Propedeutic Department of Medicine “AHEPA” University Hospital, Thessaloniki (A. Papadopoulos); Department of Hematology, Regional General Hospital of Athens “G. Gennimatas” (T. Marinakis, A. Galanopoulos, E. Michali, and N. Anagnostopoulos); Department of Hematology, “Benizelio” General Hospital, Heraklion, Crete (P. Iliakis); Department of Hematology, “Theagenion” Cancer Center, Thessaloniki (M. Papaioannou, J. Christakis, J. Korantzis, and A. Lazaridou); Department of Hematology, “401” General Army Hospital, Athens (D. Kolokithopoulos and C. Poziopoulos); Department of Hematology, Demokriteion University Hospital of Alexandroupolis, Thrace (G. Bourikas and K. Tsatalas); Department of Hematology, Crete School of Medicine, University Hospital, Heraklion, Crete (G. D. Eliopoulos and H. A. Papadaki); “Metaxa” Cancer Hospital, Athens (C. Megalakaki, B. Seitanidis, and M. Hatziyanni); University of Ioannina Hospital, Ioannina (K. L. Bourantas); Hematology Clinic, “HYGEIA” Diagnostic and Therapeutic Center, Athens (A. Papagiannis, Barbarousi, and G. Karianakis); Hematology Clinic Transplantation Unit, “Evangelismos” General Hospital, Athens (M. Nikiforakis, A. Skandali, N. Charchalakis, and N. Karakasis); Hematology Clinic, Transplantation Unit, “G. Papanikolaou” General Hospital, Thessaloniki (A. Fassas, A. Anagnostopoulos, and I. Sakellari.); and Laboratory of Immunology and Hematology, “G. Gennimatas” General Hospital (A. Tasiopoulou, E. Griva, E. Goumakou, G. Paterakis, G. Stavroupoulou-Giokas, and G. Androutsos).



Postremission therapy is critical in maintaining complete remission (CR) in patients with de novo acute myelogenous leukemia (AML). The aim of this trial was to compare allogeneic stem cell transplantation (SCT), high-dose cytarabine (ara-C; HiDAC), and autologous SCT as postremission therapy in patients with de novo AML.


One hundred twenty patients age ≤ 60 years with previously untreated AML (non-M3) and a performance status score of ≤ 2 received induction therapy with 3 days of idarubicin and 7 days of ara-C (IA). Patients in CR received one course of HiDAC. Subsequently, patients age ≤ 50 years with available HLA-compatible donors were assigned to receive allogeneic SCT; patients with “favorable” cytogenetics received a second course of HiDAC; and all others were randomized to a second course of HiDAC or autologous SCT.


The IA combination induced CR in 99 patients (82.5%). With a median follow-up of 43 months (range, 18–64 years), the 3-year survival and failure-free survival (FFS) rates were 47% and 45%, respectively. The factors associated with longer survival were those identified for CR (i.e., age and cytogenetics). Forty-nine patients (49%) received the assigned postremission therapy. Fifteen patients underwent allogeneic SCT. Nineteen patients underwent autologous SCT and 15 patients received a second course of HiDAC, after randomization. In the allogeneic SCT group, both the 3-year survival and the FFS rates were 73%. In the autologous SCT and HiDAC groups, the 3-year survival rates were 58% and 46%, respectively (P = 0.80), and the 3-year FFS rates were 42% and 33%, respectively (P = 0.83).


The three postremission treatment groups had comparable survival. Allogeneic SCT is associated with a prolonged FFS. Cancer 2003;97:1721–31. © 2003 American Cancer Society.

DOI 10.1002/cncr.11240

The combination of cytarabine (ara-C) and anthracyclines has been used extensively as standard therapy in patients with de novo acute myelogenous leukemia (AML). With 3 days of anthracyclines and 7 days of ara-C (“3+7” regimen), average complete remission (CR) rates range between 60% and 70%. Longer treatment with ara-C (“3+10”) does not induce higher remission rates, but the “3+7” and “3+10” regimens reportedly are more effective than less intensive regimens.1–4 In recent years, three large randomized trials, including idarubicin in the induction regimen, produced significantly higher CR rates and prolonged survival compared with daunorubicin (55–60% and 69–80%).5–7 A metaanalysis of such studies suggested that idarubicin is associated with high CR rates and prolonged survival.8

Although the majority of patients age ≤ 60 years achieve CR, the 5-year leukemia-free survival rates (LFS) are < 25% with conventional postremission therapy.9–14 Further intensive treatment is required to prevent recurrence. The treatment options include intensified postremission chemotherapy2, 15–21 and allogeneic or autologous stem cell transplantation (SCT).22 Allogeneic SCT in first CR has been associated with prolonged survival.19, 23, 24 Although some studies have shown that autologous peripheral blood SCT results in longer failure-free survival (FFS) compared with conventional chemotherapy,15, 25–28 other studies failed to show a similar benefit.29

We report the results of a “3+7” idarubicin and ara-C (IA) induction regimen with one course of high-dose ara-C (HiDAC) followed by allogeneic SCT, autologous SCT, or high-dose chemotherapy as postremission therapy in patients with de novo AML.



One hundred twenty consecutive patients age ≤ 60 years with previously untreated de novo AML were entered into the study AML8 of the Hellenic Cooperative Group for the treatment of AML between August 1996 and April 2000. The diagnosis of AML was confirmed with morphologic, cytochemical, and immunophenotypic studies, which were performed centrally, as previously described.30 The French–American–British (FAB) cooperative group criteria were used for classification.31 Other eligibility criteria included non AML-M3 or M3v, a performance status score of ≤ 2 (Zubrod scale), cardiac ejection fraction ≥ 50%, negative human immunodeficiency virus serology, adequate hepatic (transaminases, bilirubin, or alkaline phosphatase level < 2 × the upper reference limit) and renal (creatinine level < 2.5 mg/dL) function, unless these abnormalities were due to leukemia, and no uncontrolled infection. All procedures were approved by the Ethics Committee on Human Experimentation in Greece and were in accordance with the Helsinki Declaration of 1975.

Chromosomal analysis by conventional criteria was performed on bone marrow aspirates at the time of diagnosis.32 Patients with t(8;21) or inv(16) were considered to have a “better” prognosis; patients with normal karyotype (+8 or less than three numerical abnormalities), excluding those involving chromosomes 5 or 7, were considered to have “intermediate” prognosis; and patients with other anomalies were considered to have a “worse” prognosis.

Induction Therapy

Patients received one course of intravenous (i.v.) constant infusion (CI) of ara-C (200 mg/m2 on Days 1–7) and a 30-minute infusion of idarubicin (12 mg/m2 i.v. on Days 1, 3, and 5). A second course with ara-C (200 mg/m2 i.v. CI on Days 1–5) and idarubicin (12 mg/m2 i.v. on Days 1 and 3) was administered on Day 29 or 7 days after the absolute neutrophil count (ANC) was 1.5 × 109/L or higher and the platelet count was 100 × 109/L or higher (Fig. 1). The ara-C dose was reduced by 20% on the second HiDAC course if time to recovery was longer than 28 days or if toxicity of Grade 3 or worse occurred. Patients failing to respond after the second course were removed from the study.

Figure 1.

Simplified schema of induction and postremission therapy in patients with de novo acute myelogenous leukemia. HiDAC: high-dose cytarabine; ara-C: cytarabineallo-BMT: allogeneic bone marrow transplantation; Auto-SCT: autologous peripheral blood stem cell transplantation.

Postremission Therapy

Patients in CR received one course of HiDAC 3 g/m2 i.v. over 2 hours every 12 hours on Days 1, 3, and 5, followed by peripheral blood stem cell (PBSC) collection on leukocyte count recovery. After the first course of HiDAC, patients age < 50 years were assigned to receive an allogeneic SCT if an HLA-identical donor was available. Patients were assigned to receive further therapy according to the cytogenetic risk group. For example, patients with “better prognosis” cytogenetics were to receive a second course of HiDAC. In addition, patients with “intermediate” or “worse” prognosis cytogenetics were randomized centrally to autologous peripheral blood SCT without purging or a second postremission course of HiDAC.

Conditioning Regimen for Autologous Peripheral Blood SCT

Patients received busulfan 1 mg/kg orally every 6 hours on Days −8 through −6, etoposide 250 mg/m2 i.v. with a maximum infusion rate of 100 mg/kg/hr on Days −5 through −4, 3-hour infusion cyclophosphamide 60 mg/kg/d every 12 hours i.v. on Days −3 through −2, and 12-hour infusion sodium mercaptoethanesulfonate 1.4 × 60 mg/kg/d on Days −3 through −1, starting 4 hours before the administration of cyclophosphamide.

Supportive Care

Administration of granulocyte–colony-stimulating factor (G-CSF) or granulocyte-macrophage–colony-stimulating factor (GM-CSF) in neutropenic patients was left to the discretion of the treating physician and was administered at least 24 hours after the completion of chemotherapy. A dose of G-CSF or GM-CSF 5 μg/kg/d subcutaneously (s.c.) was administered to all patients who received HiDAC, starting on Day 8 until completion of PBSC collection or until the ANC was > 0.5 × 109/L. In patients who underwent autologous SCT, G-CSF or GM-CSF 5 μg/kg/d s.c. was administered on Day +1 until the ANC was > 1 × 109/L.

End Points and Statistical Methods

Response to induction treatment was defined according to standard criteria. Bone marrow examinations were performed on Day 21 of induction chemotherapy and then every 7 days until the response was assessed and before the start of every course of chemotherapy. After the completion of therapy, bone marrow examinations were performed every 2–4 months.

Early death was defined as any death within 14 days from the start of chemotherapy. Death in aplasia was defined as any death during treatment-induced bone marrow hypoplasia. Resistance was defined as failure to achieve a CR with two courses of induction chemotherapy or early recurrences within 6 months following the achievement of CR. Toxicity criteria were those recommended by the National Cancer Institute.33 Comparisons were made on an intention-to-treat basis. Survival was measured from the time of entry into the study until death or last follow-up. FFS was defined from the time of CR until recurrence, death, or last contact. In the comparisons of the postremission groups, survival was measured from the time of CR until death or last contact. Survival curves were calculated according to the method of Kaplan and Meier and compared using the logrank test.34, 35 Comparisons of CR rates according to the patients' characteristics were evaluated by the chi-square test36 or by the Fisher exact test. Data from patients undergoing allogeneic or autologous SCT were not censored at the time of transplantation and were evaluated according to eventual outcome.



One hundred twenty patients were evaluable. Five patients were inevaluable because of delay of the second course of induction chemotherapy (n = 1), refusal of follow-up (n = 1), and ineligibility (previous myelodysplastic syndrome; n = 3).

The clinical characteristics of the 120 patients are summarized in Table 1. Cytogenetic analysis was performed in 112 patients (93%). Nineteen patients (17%) had “better” prognosis chromosomal abnormalities, 65 patients (58%) had “intermediate” prognosis, 15 patients (13%) had “worse” prognosis cytogenetics, and 13 of the 112 patients (12%) had insufficient metaphases for analysis. In 7 of 15 patients (47%) with t(8;21), blast cells expressed CD56.

Table 1. Clinical and Laboratory Characteristics of Patients with Primary AML (n = 120)
CharacteristicsNo. (%)
  1. AML: acute myelogenous leukemia; FAB: French–American–British.

Age (yrs) 
 15–3025 (21)
 31–4019 (16)
 41–5041 (34)
 51–6035 (29)
Male/female Leukocyte count (× 109/L)55/65 (46/52)
 ≤ 424 (20)
 > 4 and ≤ 1014 (12)
 > 10 and ≤ 3034 (28)
 > 3048 (40)
Hemoglobin level (g/dL) 
 ≥ 1037 (31)
 < 10 and ≥ 836 (30)
 < 847 (39)
Platelet count (× 109/L) 
 ≥ 10024 (20)
 < 100 and ≥ 5035 (29)
 < 50 and ≥ 2037 (31)
 < 2024 (20)
FAB subtypes 
 M08 (7)
 M111 (9)
 M248 (40)
 M4/M4 Eo33/3 (27/3)
 M510 (8)
 M64 (3)
 Biphenotypic3 (3)
 Favorable19 (16)
  t (8;21)16 (13)
  inv (16)3 (3)
 Intermediate risk65 (54)
  Diploid49 (41)
 Poor risk15 (12)
 Insufficient metaphases13 (11)
 Not done8 (7)


Treatment induced CR in 99 of 120 patients (82.5%). The median time to CR was 36 days. Eighty-three patients (84%) achieved CR with the first course of induction chemotherapy, whereas 16 patients (16%) required 2 courses. As expected, the CR rate decreased with increasing age and was lower in patients with “worse” prognosis cytogenetics in univariate (Table 2) and multivariate analyses (P = 0.01 for age and P = 0.03 for cytogenetics).

Table 2. Response and Survival of 120 Patients with Primary AML by Pretreatment Characteristics
CharacteristicsNo. of CRs/totalPercentageP value5-yr survival rate (%)P value
  • AML: acute myelogenous leukemia; CR: complete remission.

  • a

    Estimates are based on fewer than 10 patients.

All patients99/12082.5   
Age (yrs)     
 31–4016/1984 41a 
 41–5032/4178 36 
 51–6026/3574 38 
Leukocyte count (× 109/L)     
 ≤ 418/24750.5734a0.07
 > 4 and ≤ 1013/1493 71 
 > 10 and ≤ 3028/3482  
 > 3040/4883 51 
Hemoglobin level (g/dL)     
 ≥ 1031/37840.12540.36
 < 10 and ≥ 833/3692 39 
 < 835/4774 37 
Platelet count (× 109/L)     
 ≥ 10019/24790.88560.37
 < 100 and ≥ 5029/3583 39 
 < 50 and ≥ 2030/3781 40 
 < 2021/2488 42 
 Intermediate risk55/6585 54 
 Poor risk8/1553 8a 
 Insufficient metaphases10/1377 39a 


The median follow-up of surviving patients was 43 months (range, 18–64 months) and the estimated 5-year survival rate was 43% (95% confidence interval [CI] 34–52%) (Fig. 2A). Sixty-five patients had died at the time of last follow-up. Three patients (2.5%) had early death. Fifteen patients (12.5%) died with aplasia (i.e., 9 patients [7.5%] died during induction chemotherapy and 6 patients [5%] died during postremission therapy). Nine patients (7.5%) had resistant disease. Thirty-eight patients (32%) died after leukemia recurrence. The factors associated with longer survival were essentially those associated with achievement of CR, i.e., an age of ≤ 30 years and “intermediate” or “better” prognosis cytogenetics in univariate (Table 2) (Fig. 2B, C) and multivariate analyses (P = 0.006 for age and P = 0.02 for cytogenetics). The number of courses required to achieve CR (one vs. two courses) was not found to affect survival (P = 0.80).

Figure 2.

(A) Overall survival in 120 patients with de novo acute myelogenous leukemia treated with idarubicin/cytarabine and postremission therapy by age (B) and by cytogenetic risk group (C).


Among the 55 patients who were alive at the time of last follow-up, 44 patients were in continuous CR and 11 patients achieved a second CR with salvage therapy after first recurrence. The median FFS was 17 months (range, 15–61 months) and the 3-year FFS rate was 45% (95% CI 33–53%; Table 3, Fig. 3). In univariate analysis, the 3-year FFS rates were higher in patients with intermediate risk cytogenetics and a leukocyte count between 4 and 10 × 109/L. The number of courses (one vs. two courses) required to achieve CR did not affect the duration of FFS (P = 0.76).

Table 3. FFS of 99 Patients with Primary AML by Patient Characteristics
CharacteristicsNo.3-yr FFS rate (%)5-yr FFS rate (%)P value
  • AML: acute myelogenous leukemia; FFS: failure-free survival.

  • a

    Estimates are based on fewer than 10 patients.

All patients994543 
Age (yrs)    
Leukocyte count (× 109/L)    
 ≤ 418370.03
 > 4 and ≤ 10136969a 
 > 10 and ≤ 302829 
 > 30405252 
Hemoglobin level (g/dL)    
 ≥ 103158520.42
 < 10 and ≥ 8333636 
 < 83542 
Platelet count (× 109/L)    
 ≥ 1001968570.23
 < 100 and ≥ 50293636 
 < 50 and ≥ 20304747 
 < 202133 
 Intermediate risk555656 
 Poor risk819 
 Insufficient metaphases102020a 
Figure 3.

Failure-free survival in 99 patients with de novo acute myelogenous leukemia.

Postremission Therapy

Twenty-one patients who were age ≤50 years had HLA-matched donors, 15 of whom underwent allogeneic SCT as postremission therapy. The median follow-up of surviving patients was 55 months (range, 33–64 months). The median survival had not been reached and the 3-year survival rate was 73% (95% CI 51–96%; Table 4). Four patients died 178 days (range, 63–658 days) after allogeneic SCT. The causes of death were aspergillosis in two patients whose disease was in remission and progressive leukemia in two patients.

Table 4. Results of Postremission Therapy in 49 Patients with primary AMLa
CharacteristicsAllogeneic BMT (n = 15)Randomized
HiDAC (n = 15)Autologous SCT (n = 19)
  • AML: acute myelogenous leukemia; BMT: bone marrow transplantation; HiDAC: high-dose cytarabine; SCT: stem cell transplantation; FFS: failure-free survival.

  • a

    Patients were assigned to allogeneic transplantation based on the availability of an HLA-compatible donor.

 Median (mos)Not reached37Not reached
 Range (mos)9–6112–599–59
 3-yr survival (%)734658
 Median (mos)Not reached1515
 Range (mos)5–614–583–57
 3-yr FFS (%)733342

Nineteen patients were randomized to receive autologous SCT after 1 course of HiDAC and 15 patients were randomized to an additional course of HiDAC (Table 5). The randomized groups were similar with regard to age, cytogenetic group, FAB subtypes, or the number of courses of induction therapy required to achieve CR. The 3-year survival rates between patients randomized to receive autologous SCT and HiDAC as postremission therapy did not appear to differ significantly (58% and 46%, respectively; P = 0.80) (Table 4). Similarly, the 3-year FFS rates were 42% and 33%, respectively (P = 0.83).

Table 5. Clinical and Laboratory Characteristics of Patients with de novo AML Randomized to Autologous Peripheral Blood SCT or HiDAC
CharacteristicsAutologous SCT (n = 19)HiDAC (n = 15)P value
  1. AML: acute myelogenous leukemia; HiDAC: high-dose cytarabine; SCT: stem cell transplantation; CR: complete remission.

Median age45440.81
Leukocyte count (× 109/L)50530.6
Hemoglobin level (g/dL)
Platelet count (× 109/L)78560.88
M0, M1, M2 vs. M4, M5, M610/79/50.99
 Intermediate prognosis11100.75
 Worse prognosis11 
No. of induction courses for CR   

The three postremission groups did not differ with respect to leukocyte count, hemoglobin level, or platelet counts. Age did not differ between patients in the autologous SCT and HiDAC groups (P = 0.76). However, in the “intermediate” prognosis cytogenetic risk group, patients who underwent allogeneic SCT were younger than those who underwent autologous SCT or HiDAC (mean age, 28 years vs. 44 years and 46 years; P = 0.004 and 0.006, respectively). No significant difference was noted among the three postremission groups in terms of survival (P = 0.41) (Fig. 4A) or FFS (P = 0.09) (Fig. 4B).

Figure 4.

(A) Overall survival and (B) failure-free survival in 49 patients with de novo acute myelogenous leukemia according to postremission therapy. Allo-BMT: allogeneic bone marrow transplantation; Auto-SCT: autologous stem cell transplantation.

In the “intermediate” prognosis cytogenetic group, survival and FFS were longer in patients who underwent allogeneic SCT (10 patients, including 8 patients with normal karyotype) than in patients who underwent autologous SCT or received HiDAC after randomization (P = 0.05 and P = 0.01, respectively). A similar comparison could not be performed in the other cytogenetic groups because of the limited number of patients.

In Table 6, the 34 patients randomized to autologous peripheral blood SCT (n = 19) or HiDAC (n = 15) are compared with the 37 patients who were not randomized. The two groups differ in cytogenetics and leukocyte counts. However, there were no significant differences in the 3-year survival and FFS rates between the two groups.

Table 6. Comparison of the Clinical and Laboratory Characteristics between Randomized and Nonrandomized Patients
CharacteristicsRandomized (n = 34)Nonrandomized (n = 37)P value
  • CR: complete remission; FFS: failure-free survival.

  • a

    Median survival not reached in either group.

Median age43420.96
Leukocyte count (× 109/L)51240.002
Hemoglobin level (g/dL)
Platelet count (× 109/L)68870.85
M0, M1, M2 vs. M4, M5, M6, M719/1223/130.99
 Intermediate risk2116 
 Poor risk24 
No. of induction courses to CR   
 3-yr survival (%)52630.76
 FFS Median (mos)15170.82
 3-yr FFS (%)3848 

Feasibility of Postremission Therapy

Forty-nine patients (49%) received the assigned postremission therapy. The median time from CR to postremission therapy was 4 months (range 2–7 months) for patients who underwent allogeneic SCT, 4 months (range, 2–6 months) for those who underwent autologous peripheral blood SCT, and 2 months (range, 1–4 months) for patients who received HiDAC.

A total of 52 patients (60%) received HiDAC as postremission therapy (randomized, n = 15; nonrandomized, n = 37). Twenty patients received one and 32 patients received two courses of HiDAC. Excluding patients who underwent allogeneic or autologous SCT, patients who received two courses of HiDAC had longer survival and FFS compared with those who received one course of HiDAC (P = 0.009 and P = 0.0004, respectively; Fig. 5). When this analysis was limited to surviving patients, FFS duration was significantly longer in patients who received two rather than one course of HiDAC (P = 0.05).

Figure 5.

(A) Overall survival and (B) failure-free survival according to high-dose cytarabine (HiDAC), excluding the allogeneic and autologous stem cell transplantation groups.


In this study, “3+7” IA induced CR in 82.5% of the patients. This rate is at the highest range of other induction regimens (50–90%).18, 37, 38 Complete remission rates in the better prognosis (100%), intermediate (85%), and worse prognosis cytogenetic groups (53%) are in accordance with those of the Medical Research Council (MRC), Eastern Cooperative Oncology Group/Southwest Oncology Group (ECOG/SWOG), and The Groupe Ouest Est Leucemies Aigues Myeloblastiques studies: 84–90%, 76–84%, and 55–58%, respectively.29, 39, 40 The number of courses of induction chemotherapy required for the achievement of CR did not affect the clinical outcome with this induction regimen. Using HiDAC as induction prolongs disease-free survival (DFS).37 Data from the MRC AML 10 trial, which used two cycles of daunorubicin, ara-C, and etoposide for induction, emphasize the importance of response after one course in the clinical outcome of patients age < 55 years. In this trial response after one course combined with cytogenetics were two of the three components that provided a simple and robust index, predicting for survival and recurrence rates.41 The results of the current study confirmed the importance of previously well established factors predicting achievement of CR and duration of survival, such as younger age and poor-risk cytogenetics.39–42 The survival rates of patients who are 30 years old or younger are comparable with those in children with AML,43 suggesting that the age of 30 may be a useful cutoff point to predict clinical outcome.

In the current study, no differences were noted between patients randomized to autologous SCT or HiDAC in terms of survival or FFS. Previously published randomized studies with more patients have also demonstrated that autologous SCT does not add any substantial benefit in survival compared with chemotherapy. Two of these studies included at least one course of postremission chemotherapy before randomization but total body irradiation therapy was used instead of busulfan in the conditioning regimen.15, 28 However, in the intergroup study, survival after CR at 4 years was similar in patients treated with high-dose chemotherapy, autologous SCT, or allogeneic SCT.22 In two large cooperative studies (European Organization for Research and Treatment of Cancer/Gruppo Italiano Malattie Ematologiche Maligne dell'Adulto and MRC10), DFS is significantly longer in patients treated with autologous SCT compared with chemotherapy (48% vs. 30% at 4 years and 53% vs 40% at 7 years, respectively).15, 28 In contrast, no difference was found in DFS in the intergroup study (35% at 4 years in both groups).22 In agreement with the results of the current study, recurrences were least common in patients treated with allogeneic SCT, less common in those assigned to autologous BMT, and more frequent in patients assigned to high-dose chemotherapy.22

The role of postremission chemotherapy before SCT for AML patients in first CR remains unclear. Tallman et al.44 suggested that postremission ara-C before allogeneic BMT for AML in first CR does not improve survival and LFS.44 The European Group for Blood and Marrow Transplantation found that the dose of ara-C at induction or after disease remission had no influence on the outcome of autologous SCT or allogeneic bone marrow transplantation in terms of LFS or transplant-related mortality.45 One course of HiDAC was administered before allografting to prevent early recurrence during the interval from CR to SCT; this, in fact, was longer in the current study compared with other studies.

In the current study, patients with “better-risk” cytogenetics had a shorter survival compared with those with “intermediate-risk” cytogenetics. Considering the relatively small number of patients included in the different cytogenetic risk groups, which precludes robust statistical results, some caution is warranted in explaining this rather paradoxical finding. It is noteworthy that a high proportion of patients with t(8;21) expressed CD56, a marker associated with short CR duration and survival.46 Data published after the initiation of our study demonstrated that patients in the “favorable-risk” cytogenetic group who were treated with one course of HiDAC intensification had a significantly shorter FFS and survival than those treated with three or more cycles.47, 48 The need of intensified postremission therapy to ensure a favorable outcome is also supported by the results of the SWOG/ECOG study. Patients age < 56 years with favorable cytogenetics did significantly better after autologous and allogeneic SCT than with chemotherapy alone.40 In another study with a large number of patients with t(8;21), survival was similar between patients treated with SCT or chemotherapy and the leukocyte count was the only identified prognostic factor.49 In our study, blasts of 47% of the patients with t(8;21) expressed CD56 and most had a high leukocyte count. Finally, accumulating evidence suggests that additional genetic abnormalities not identified by standard cytogenetics, such as flt-3 gene mutations, might confer an adverse prognosis for patients with AML.50

Allogeneic SCT in patients with intermediate-risk cytogenetics has been associated with superior outcome compared with the no-allogeneic SCT.51 Visani et al.51 showed that patients with “intermediate” or “better-risk” cytogenetics who underwent allogeneic SCT had a significantly longer DFS than those who did not receive allogeneic SCT. In our study, all 10 patients with intermediate-risk cytogenetics who underwent allogeneic SCT remain alive and disease-free.

The difficulty of a significant proportion of patients to receive the assigned postremission therapy has been addressed in previous reports, in which 50–69% of the patients received the assigned therapy.15, 22, 28, 29, 52 Our results are in line with previous reports, showing that the proportion of patients receiving postremission therapy decreases with the worsening-risk cytogenetic group.40

In conclusion, “3+7” IA induction chemotherapy was an effective regimen in patients with newly diagnosed AML. Allogeneic SCT, autologous SCT, and HiDAC were comparable in terms of survival, but allogeneic SCT was associated with prolonged FFS. Prospective, randomized clinical trials with a large population of patients are needed to optimize the approach of postremission therapy in AML patients.