Results of decitabine (5-aza-2′deoxycytidine) therapy in 130 patients with chronic myelogenous leukemia




General and site-specific DNA methylation is associated with tumor progression and resistance in several cancers, including chronic myelogenous leukemia (CML). Decitabine is a hypomethylating agent that has shown encouraging preliminary anti-CML activity. This study evaluated the activity and toxicity of decitabine in different phases of CML.


One hundred and thirty patients with CML were treated: 123 with Philadelphia chromosome (Ph)-positive CML (64 blastic, 51 accelerated, 8 chronic) and 7 with Ph-negative CML. Decitabine was given at 100 mg/m2 over 6 hours every 12 hours × 5 days (1000 mg/m2 per course) in the first 13 patients, 75 mg/m2 in the subsequent 33 patients, and 50 mg/m2 in the remaining 84 patients.


A total of 552 courses were given to the 130 patients. Only four patients (3%) died during the first course from myelosuppressive complications (three patients) or progressive disease (one patient). Of 64 patients in the CML blastic phase, 18 patients (28%) achieved objective responses. Of these 18 patients, 6 achieved complete hematologic responses (CHR), 2 achieved partial hematologic responses (PHR), 7 achieved hematologic improvements (HI), and 3 returned to the second chronic phase (second CP). Five patients (8%) had cytogenetic responses. Among 51 patients in the accelerated phase, 28 patients (55%) achieved objective responses (12 CHR, 10 PHR, 3 HI, and 3 second CP). Seven patients (14%) had cytogenetic responses. Among eight patients treated in the chronic phase, five (63%) had objective responses. Of seven patients treated for Ph-negative CML, four (57%) had objective responses. There was no evidence of a dose-response effect. The estimated 3-year survival rate was less than 5% in the blastic phase and 27% in the accelerated phase. The only significant toxicity reported was severe myelosuppression, which was delayed, prolonged, and dose dependent. With decitabine 50–75 mg/m2, the median time to granulocyte recovery above 0.5 × 109/L was about 4 weeks. Myelosuppression-associated complications included febrile episodes in 37% and documented infections in 34%.


Decitabine appears to have significant anti-CML activity. Future studies should evaluate lower-dose, longer-exposure decitabine schedules alone in imatinib-resistant CML, as well as combinations of decitabine and imatinib in different CML phases. Cancer 2003;98:522–8. © 2003 American Cancer Society.

DOI 10.1002/cncr.11543

General and site-specific methylation of DNA is a mechanism of resistance and evolution in cancers, including chronic myelogenous leukemia (CML).1–4 Hypomethylating agents may provide new target-oriented antitumor approaches and have included 5-azacytidine and 5-aza-2′-deoxycytidine (decitabine).5–14 Both agents have shown activity against acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). In a randomized study of 5-azacytidine versus supportive care, 5-azacytidine demonstrated significantly better complete (CR; 7% vs. 0%) and objective response rates (60% vs. 5%), longer time to progression (21 months vs. 13 months; P = 0.007), and a trend for longer survival (20 months vs. 14 months; P = 0.1).5 In AML salvage, decitabine therapy was associated with CR rates of 10–30% as a single agent and of 30–60% in combinations.6–13 We had previously reported our preliminary experience with decitabine in CML in transformation.14 This study reports the final results of decitabine given at dose schedules of 500–1000 mg/m2 per course in 130 patients in different phases of CML.


Study Group

Adults (age 15 years or older) with a diagnosis of Philadelphia chromosome (Ph)-positive or Ph-negative CML entered the study if they met the eligibility criteria. Informed consent was obtained according to institutional guidelines. Patients were treated on two consecutive institutional review board-approved protocols, DM92-83 (94 patients; January 1993 through February 1998) and DM98-25 (36 patients; February 1998 through November 2000). Eligibility criteria included a patient age of 15 years or older, adequate hepatic (bilirubin level < 2 mg/dL) and renal functions (creatinine level < 2 mg/dL), a Zubrod performance score of 0–3, and normal cardiac status. Patients were treated in all CML phases according to previously established criteria.14–18


Decitabine was given initially as 100 mg/m2 over 6 hours every 12 hours for 5 days (1000 mg/m2 per course) every 4–8 weeks, depending on the recovery of blood counts and serial bone marrow studies. Subsequent courses were initiated when granulocyte counts were above 109/L and platelet counts were above 80–100 × 109/L, unless myelosuppression was attributed to persistent disease. For Grade 3–4 extramedullary toxicity or for prolonged myelosuppression (delayed recovery beyond 6 weeks from the start of therapy without evidence of disease), subsequent courses were reduced by 25% to 50% and rounded to the following: 75, 50, 37.5, 30, 22.5, 15, 10, and 7.5 mg/m2. For severe life-threatening side effects, dose reductions of 50% were allowed if judged appropriate by the treating physician. Because of severe prolonged myelosuppression after the first 13 patients, the dose was reduced to 75 mg/m2 in the next 33 patients and to 50 mg/m2 in the remaining 84 patients. The number of patients treated with each dose schedule by CML phase is shown in Table 1.

Table 1. Number of Patients Treated at Each Dose Schedule of Decitabine by Chronic Myelogenous Leukemia Phase (130 patients)
Decitabine (mg/m2 IV over 6 hrs every 12 hrs for 5 days)No. treated
  1. IV: intravenous; Ph: Philadelphia chromosome.

10094 13

Response Criteria and Statistical Considerations

Response criteria have been described previously.14–18 Briefly, a complete hematologic response (CHR) required normalization of peripheral blood counts and differentials (a granulocyte count of ≥ 109/L, a leukocyte count of 2–10 × 109/L, and a platelet count of ≥ 100 × 109/L) with 5% or less blasts in the bone marrow for at least 4 weeks. Hematologic improvement (HI) was based on the same criteria as CHR except for allowing persistent thrombocytopenia (< 100 × 109/L) and a few immature cells (no blasts or promyelocytes; ≤ 5% myelocytes and metamyelocytes) to be present in the peripheral blood. Patients who achieved CHR or HI were categorized further by their cytogenetic response as follows: complete (no Ph-positive metaphases), partial (1–34% Ph-positive metaphases), and minor (35–90% Ph-positive metaphases). A partial hematologic response (PHR) was based on the same criteria used for CHR but allowed persistent palpable splenomegaly (although spleen size had to have been reduced by 50% or more), thrombocytosis (platelet count > 450 × 109/L), and/or a few immature cells (no blasts or promyelocytes, ≤ 5% myelocytes and metamyelocytes) to be present in the peripheral blood. A return to the second chronic phase (second CP) was defined by the disappearance of blast-phase features and a return to chronic-phase features (i.e., peripheral blasts < 15%, peripheral blasts and promyelocytes < 30%, peripheral basophils < 20%, and platelet count >100 × 109/L).

All other responses were treatment failures and were categorized as early death if death occurred within 2 weeks from the start of therapy, aplastic death if the patient died during remission induction with a hypocellular bone marrow and a marrow leukemic infiltrate (MLI; i.e., the percentage of marrow blasts × marrow cellularity) of less than 20%, and resistant disease if the bone marrow MLI was persistently greater than 20% (primary) or if it decreased to less than 20% with subsequent regrowth of leukemia (secondary). Response criteria for patients with extramedullary disease were the complete disappearance of disease for a CR and a disease reduction of 50% or more for a partial response.

Survival analysis was based on the Kaplan–Meier method. Survival was calculated from the initiation of treatment for the particular CML phase until death or last follow-up. Univariate analyses for differences in response rates by patient and disease characteristics were evaluated using the chi-square test and the log-rank test was used for differences in survivals. Multivariate analyses used the logistic regression model for response and the Cox proportional hazards model for survival. Toxicity was graded according to the National Cancer Institute common toxicity criteria, version 2.0.


Study Group

The characteristics of the study group are shown in Tables 1–3. One hundred thirty patients were treated (123 with Ph-positive CML and 7 with Ph-negative CML). Their median age was 55 years (range, 16–78 years). Of the study group, 40 patients (31%) were 60 years or older, and 50 (38%) were females. A total of 552 courses were given to the 130 patients. Patients received a median of two courses (range, 1–42 courses). Forty-six patients received only 1 course for the following reasons: progressive disease (n = 28), referral for allogeneic stem cell transplantation (SCT; n = 8), prolonged myelosuppression and patient wish to change therapy (n = 6), and death during the first course (n = 4; three aplastic deaths and one death from the progressive blastic phase). With a median follow-up time of 42 months, 121 patients have already died. Of the eight patients referred for allogeneic SCT, three were in the blastic phase and five were in the accelerated phase. Two of the eight patients (both in the accelerated phase) were alive without evidence of disease at the time of last follow-up for 39 + and 26 + months post-SCT, respectively. The causes of death in the 121 patients were blastic phase (n = 95), infections (n = 8), bleeding (n = 1), post-SCT complications (n = 8), and death outside the institution without definite documentation of cause of death (but most likely disease progression; n = 9).

Table 2. Characteristics Associated with Objective Response and Survival with Decitabine in the Blastic Phase (64 Patients)
ParameterNo.No. of objective responses (%)P value18-mo survival (%)P value
  1. CML: chronic myelogenous leukemia.

Age (yrs)     
 < 604815 (31) 8 
 ≥ 6016 3 (19)0.5260.70
 No3911 (28) 10 
 Yes25 7 (28)1.040.97
Hemoglobin (g/dL)     
 < 104112 (29) 10 
 ≥ 1023 6 (26)1.040.80
Leukocyte count (×109/L)     
 < 504614 (30) 9 
 ≥ 5018 4 (22)0.7660.94
Platelets (× 109/L)     
 < 5020 6 (30) 5 
 ≥ 504412 (27)1.090.29
Bone marrow blasts (%)     
 < 5041/6215 (37) 12 
 ≥ 5021/62 2 (10)0.0300.04
Peripheral blasts (%)     
 < 302913 (45) 17 
 ≥ 3035 5 (14)0.0100.11
Albumin (g/dL)     
 ≤ 3.54110 (24) 7 
 > 3.523 8 (35)0.4090.33
Cytogenetic clonal evolution     
 No21 7 (33) 14 
 Yes4311 (26)0.5650.54
CML duration (mos)     
 < 129 2 (22) 11 
 12–3520 7 (35)0.70100.41
 ≥ 3635 9 (26) 6 
Salvage no.     
 First4413 (30) 7 
 Second14 3 (21)0.8070.93
 Third or more6 2 (33) 17 
Decitabine dose (mg/m2)     
 75–10022 7 (32) 9 
 504211 (26)0.7770.62
Table 3. Response to Decitabine in Chronic Myelogenous Leukemia Blastic Phase (64 Patients)
ResponseNo. (%)
Complete hematologic response 6 (9)
 Cytogenetic response 
  Partial 2 (3)
  Minor 3 (5)
Partial hematologic response 2 (3)
Hematologic improvement 7 (11)
Second chronic phase 3 (5)
Induction death 3 (5)
Primary resistance20 (31)
Secondary resistance23 (36)

Blastic Phase

Sixty-four patients were treated for blastic-phase disease (Table 2). Blastic morphology was myeloid in 51 patients (80%), lymphoid in 2 (3%), undifferentiated in 6 (9%), and megakaryocytic in 5 (8%). Overall, 18 patients (28%) achieved an objective response, including 6 CHR, 7 HI, 2 PHR, and 3 second CP (Table 3). Five patients (8%) achieved a cytogenetic response (two partial, three minor). Response and survival by different patient characteristics are shown in Table 2. Significantly worse response rates (P < 0.05) were noted with high peripheral blasts (≥ 30%) and bone marrow blasts (≥ 50%). Significantly worse survival was noted with high bone marrow blasts (P = 0.04). There was no difference in the response rate or survival by the dose schedule of decitabine. Multivariate analyses for response and for survival identified only bone marrow blasts to be associated independently with survival (P = 0.02). The median survival was 5 months and the 1-year survival rate was 17% (Fig. 1).

Figure 1.

Survival of patients with decitabine therapy by chronic myelogenous leukemia phase.

Accelerated Phase

Fifty-one patients were treated in the accelerated phase (Table 4). Seven patients were categorized to have accelerated phase disease based on clonal evolution only. Overall, 28 patients (55%) achieved objective responses (Table 5), including 12 CHR, 10 PHR, 3 HI, and 3 second CP. Seven patients (14%) had cytogenetic responses. The median survival was 17 months and the 3-year survival rate was 27% (Fig. 1). Response and survival by patient characteristics are shown in Table 5. Poor prognostic factors for an objective response were anemia (hemoglobin level < 10 g/dL) and cytogenetic clonal evolution (P < 0.01). Both remained prognostically independent factors by multivariate analysis (P = 0.02 and 0.04, respectively). Poor prognostic factors for survival were anemia, leukocytosis, and a higher dose schedule of decitabine (P ≤ 0.05). None remained independently significant by multivariate analysis. Although cytogenetic clonal evolution was an adverse factor in general (with or without other accelerated features), the seven patients with cytogenetic clonal evolution as the only accelerated-phase feature had a response rate of 29% (one CHR, one partial cytogenetic response). Their median survival was 13 months and two were alive at 3 + and 42 + months, respectively.

Table 4. Characteristics Associated with Objective Response and Survival with Decitabine in Accelerated Phase (51 Patients)
ParameterNo.No. (%)P value2-yr survival (%)P value
  1. CML: chronic myelogenous leukemia.

Age (yrs)     
 < 603521 (60) 41 
 ≥ 6016 7 (44)0.37360.52
 No2716 (59) 40 
 Yes2412 (50)0.70390.48
Hemoglobin (g/dL)     
 < 102910 (34) 25 
 ≥ 102218 (82)< 0.0158< 0.01
Leukocyte count (×109/L)     
 < 504024 (60) 46 
 ≥ 5011 4 (36)0.29180.05
Platelets (×109/L)     
 < 1002012 (60) 44 
 100–45021 9 (43)0.31300.55
 > 45010 7 (70) 50 
Bone marrow blasts (%)     
 < 59/49 6 (67) 45 
 5–1420/4910 (50)0.70250.41
 ≥ 1520/4911 (55) 50 
Peripheral blasts (%)     
 0–42313 (57) 38 
 5–1418 9 (50)0.86360.88
 ≥ 1510 6 (60) 50 
Albumin (g/dL)     
 < 3.518 7 (39) 33 
 ≥ 3.53321 (64)0.14430.14
Cytogenetic clonal evolution     
 No2117 (81) 48 
 Yes3011 (37)< 0.01350.41
CML duration (mos)     
 < 126 4 (67) 50 
 12–351510 (67)0.37470.96
 ≥ 363014 (47) 34 
Salvage no.     
 First2916 (55) 40 
 Second10 4 (40)0.46400.98
 Third or more12 8 (67) 40 
Decitabine dose (mg/m2)     
 75–1002110 (48) 29 
 503018 (60)0.41480.05
Table 5. Response to Decitabine in Chronic Myelogenous Leukemia Accelerated Phase (51 Patients)
ResponseNo. (%)
Complete hematologic response12 (24)
 Cytogenetic response 
  Complete 3 (6)
  Partial 3 (6)
  Minor 1 (2)
Partial hematologic response10 (20)
Hematologic improvement 3 (6)
Second chronic phase 3 (6)
Induction death 0 (0)
Primary resistance 7 (14)
Secondary resistance16 (31)

Chronic Phase

Eight patients were treated in the chronic phase. Their median age was 55 years and one patient (13%) was 60 years or older. Splenomegaly was present in 2 patients (25%), anemia (hemoglobin level < 10 × 109/L) in 2 patients (25%), leukocytosis greater than or equal to 50 × 109/L in 4 patients (50%), and chronic phase duration of 36 months or longer in 7 patients (88%). All had received interferon-based therapy, but none had received imatinib. Two patients received decitabine as first salvage and six received decitabine as a second or subsequent salvage. Overall, one patient (13%) achieved CHR and a minor cytogenetic response, and four (50%) had PHR for an overall response rate of 63%.

Philadelphia-Negative Chronic Myelogenous Leukemia

Seven patients with Ph-negative CML were treated. They were also negative for BCR-ABL rearrangement. Their median age was 67 years and all were 60 years or older. Splenomegaly was present in 3 patients (43%), anemia (hemoglobin level < 10 g/dL) in 5 patients (71%), thrombocytopenia less than 100 × 109/L in 5 patients (71%), leukocytosis greater than 50 × 109/L in 1 patient (14%), and bone marrow blasts greater than 5% in 4 patients (57%). Overall, four patients responded (two CHR, one PHR, one second CP). Their median survival was 13 months and the 2-year survival rate was 14% (Fig. 1).

Side Effects

Nonmyelosuppressive side effects were evaluated in all 130 patients, 81 of whom had adequate pretreatment counts (granulocyte counts 109/L or more and platelets 50 × 109/L or more; Table 6). Nonhematologic side effects with decitabine were acceptable and included drug fever in 27 patients (21%), diarrhea (Grade ≥ 2) in 6 patients (5%), nausea or emesis (Grade ≥ 2) in 5 patients (4%), fatigue (Grade ≥ 2) in 3 patients (2%), cardiac events in 3 patients (2%), and alopecia in 2 patients (2%). Myelosuppression- associated complications included febrile episodes of unknown origin in 48 patients (37%) and documented infections in 44 patients (34%): pneumonia (n = 16), fungal infections (n = 9), mixed infections (n = 13), and others (n = 6).

Table 6. Myelosuppression Profiles with the First Course of Decitabine in 81 Patients with Adequate Counts at the Start of Therapy
ParametersNo. (%) at dose schedules (mg/m2 IV over 6 hrs every 12 hrs × 5 days)
  1. IV: intravenous.

No. evaluable52236
No. (%) with lowest granulocyte count < 0.5 × 109/L45 (87)20 (87)5 (83)
No. (%) with lowest platelet count < 30 × 109/L38 (73)21 (91)4 (67)
Median days to granulocyte recovery above 0.5 × 109/L and above 109/L45/4746/5155/55
Median days to platelet recovery above 30 × 109/L272860

Myelosuppression was significant and dose related. Table 6 shows the degree of myelosuppression by dose schedule among the 81 patients who started with adequate granulocyte (> 109/L) and platelet counts (> 50 × 109/L) during their first course of therapy.

Response Patterns

Of the 55 patients who achieved an objective response, 10 patients responded after 1 course, 14 patients after 2 courses, and 31 patients after 3 courses or more. A median of three courses was required to achieve the best hematologic response and a median of three courses was required to achieve the best cytogenetic response among the 13 cytogenetic responders. There was no significant difference in the response rate by the dose of decitabine (50, 75, or 100 mg/m2) in the accelerated or blastic phase, nor was there a difference in patient survival by the three dose schedules in the blastic phases. However, higher decitabine doses were associated with worse survival in the accelerated phase (P = 0.05; Table 4), probably because of worse myelosuppression-associated mortality.


In this study, decitabine, a hypomethylating agent, has shown significant activity in CML. Objective response rates were 28% in the blastic phase, 55% in the accelerated phase, and 63% in the chronic phase. Cytogenetic responses were also noted in 7–12% of patients. These results are encouraging, considering the characteristics of the study group and the extent of previous therapy. Ongoing studies are evaluating the activity of decitabine in patients in all three CML phases following failure on imatinib mesylate therapy, the present standard of care.

The only significant side effect of decitabine is severe and prolonged myelosuppression. In the early European studies, decitabine was developed as a classical antitumor agent at the maximum tolerated dose (MTD), judged then to be 1500–2500 mg/m2 per course. Responses were noted in 30–60% of patients treated with single-agent or decitabine combinations.7–13 However, further studies were not pursued, perhaps because of the excessive myelosuppressive toxicity noted, as well as the unusual (and yet unknown) profile of delayed and prolonged myelosuppression for decitabine.7–13 Subsequent studies in MDS in Europe19, 20 used low-dose decitabine 45–50 mg/m2 daily for 3 days every 4–8 weeks (135–150 mg/m2 per course), demonstrating a CR rate of 20%, an objective response rate of 50%, and good tolerance in relation to myelosuppression and induction mortality rates (5–8%).

Our studies in CML, which started in 1993 and were based on the European experience, used decitabine 100 mg/m2 over 6 hours every 12 hours for 5 days (i.e., 1000 mg/m2 per course). This dose was associated with excessive and delayed myelosuppression. Recognizing this problem, the dose schedule was reduced to 750 and then to 500 mg/m2 per course. With this schedule in CML, using decitabine in a classic antileukemic agent mode (at the MTD), myelosuppression was significant, but the induction mortality rate was acceptably low (4 induction deaths among 130 patients; 3%). Subsequent basic knowledge regarding the hypomethylating properties of decitabine led to an important question: Would a better therapeutic approach target the molecular endpoint of hypomethylation induction (biologic dose), rather than deliver the drug at the clinical MTD? In a Phase I dose escalation biologic study, we treated patients with decitabine 5 mg/m2 daily for 10 days (to cover 2 cell cycles required for decitabine incorporation into 1, then the other DNA strands), after which we escalated the dose to 10, 15, and 20 mg/m2 daily for 10 days, and finally attempted to increase the duration of days to 15 then 20 days (unpublished data).21 Overall, approximately 40% of patients achieved objective responses. In AML salvage in particular, the objective response rate was 35% with low-dose decitabine versus only 7% with higher dose schedules and the mortality rate was lower, suggesting that the lower-dose, longer-exposure schedules of decitabine were more effective and less toxic. The dose schedule of decitabine 15 mg/m2 daily for 10 days was optimal for future Phase II studies.

It is interesting to note the time to best response with decitabine. Similar to the MDS experience,19, 20 a median of three courses were required to achieve the best response. This suggests that, mechanistically, decitabine exerts a slow but continuous effect on leukemia cells to achieve the best response. The results show that decitabine therapy should be continued in patients who manifest a suboptimal response after one or two courses of therapy.

Previous studies have shown the benefit of decitabine compared with intensive chemotherapy in blastic-phase CML, particularly in older patients.22, 23 Compared with interferon-α with or without low-dose cytarabine, decitabine was associated with better response rates in blastic-phase (28% vs.10% or less) and accelerated-phase CML (55% vs.30–40%).24

Decitabine has demonstrated significant antileukemic efficacy in all phases of CML. Future studies will evaluate lower-dose, longer-exposure schedules of decitabine in patients whose CML phase becomes resistant to imatinib. Because progressive disease was the most common cause of treatment failure and because prolonged myelosuppression was the main issue with the current decitabine dose schedules, combinations of lower-dose, longer-exposure schedules of decitabine plus imatinib may improve the results in CML transformation.