The use of azacitidine in clinical practice is based primarily on the findings of 2 phase 3 randomized clinical trials of azacitidine (75 mg/m² per day for 7 days every 28 days) conducted in patients with MDS in the United States [6] and Europe [5]. Patients in both trials had a median age of 68 or 69 years, and most had higher-risk MDS as defined using the IPSS or French-American-British (FAB) classification system. In the Cancer and Leukemia Group B (CALGB) 9221 trial, 191 patients were randomized to treatment with either azacitidine or best supportive care; patients who received supportive care could switch to azacitidine if their disease worsened [6]. Of the evaluable patients, 60% (60/99) treated with azacitidine responded (CALGB criteria) to treatment compared with 5% (5/92) who received supportive care. Progression to AML or death was delayed significantly in patients who were treated with azacitidine versus best supportive care (21 months vs. 12 months, P=0.007), and the median survival in patients randomized to treatment with azacitidine was significantly longer than that in patients who received supportive care and never switched to azacitidine or who switched after 6 months (18 months vs. 11 months, P=0.03) [6]. In addition, findings from a pooled analysis of 3 CALGB clinical trials showed that between 10% and 17% of patients treated with azacitidine (N=268) achieved complete remission [19].

In the AZA-001 trial, 358 patients were randomized to treatment with either azacitidine or conventional care with 3 options for active therapy (best supportive care, intensive chemotherapy, or low-dose cytarabine) and the median time of follow-up was 21 months [5]. Treatment with azacitidine, compared with conventional care, resulted in significantly improved overall survival (24.5 months versus 15.0 months, log-rank P=0.0001) and significant delays in progression to AML (17.8 months vs. 11.5 months, P<0.0001). The improvement in overall survival was consistent across all IPSS cytogenic subgroups, including those with unfavorable cytogenetic abnormalities. Of the 111 patients who were transfusion dependent at baseline, 45% became transfusion independent during treatment with azacitidine.

One prospective, noncomparative clinical trial examined the pharmacokinetics, efficacy, and safety of azacitidine in 53 patients with MDS from Japan [20]. Patients in this trial had intermediate- or high-risk MDS (defined using the IPSS), were a median age of 65 years, and were treated with azacitidine 75 mg/m² per day for 7 days every 28 days by either subcutaneous or intravenous administration. There were no apparent differences in the pharmacokinetics of azacitidine between patients from Japan and Western countries. One-third (10/30) of patients with higher-risk MDS had a hematologic response (28.3% [15/53] of all patients), and of the 27 patients who were transfusion dependent at baseline, 55.6% became transfusion independent during treatment. There were no differences in efficacy or safety outcomes for the different routes of administration.

Two retrospective studies were conducted in the Republic of Korea to assess the factors affecting survival in patients with MDS who were treated with azacitidine [21, 22]. Patients in these studies had a median age of 59 [21] and 64 [22] years, and approximately one-third were classified as having higher-risk MDS (defined using the IPSS). In general, the findings from these studies were comparable with those from the CALGB and AZA-001 trials. Approximately 17% of patients achieved a complete response [21, 22], and the median overall survival was 20 months [22].

The standard dosing schedule for azacitidine is subcutaneous or intravenous administration of 75 mg/m² per day for 7 consecutive days every 28 days. In practice, 7 consecutive days of treatment may be problematic for outpatients who cannot attend weekend clinic visits or for outpatient clinics that do not offer weekend treatment. Alternative dosing schedules that eliminate the need for weekend treatment have been assessed in an open-label randomized trial [23]. In this study, patients were assigned to (i) a 5-2-2 schedule (5 days of azacitidine at 75 mg/m² per day followed by a 2-day break and an additional 2 days of azacitidine, total cumulative dose=525 mg/m²), (ii) a 5-2-5 schedule (5 days of azacitidine at 50 mg/m² per day followed by a 2-day break and an additional 5 days of azacitidine, total cumulative dose=500 mg/m²), or (iii) a 5-day schedule (5 days of azacitidine at 75 mg/m² per day, total cumulative dose=375 mg/m²). Findings from this study showed similar hematologic improvement across all groups (44% to 56%) and similar rates of transfusion independence (50% to 64%). As expected, the lowest total cumulative dose (5-day schedule) was better tolerated, with more patients completing 6 cycles of treatment than patients who received the higher total cumulative doses (5-2-2 or 5-2-5 schedules). However, the findings of this study should be interpreted with caution, particularly for patients with higher-risk MDS, because the standard dosing schedule was not included in this study as a control arm and almost two-thirds of patients had lower-risk MDS (defined using the FAB system). In addition, because overall survival or progression to AML was not assessed in this study, it is not possible to determine the effect of a modified dosing schedule on the survival benefit of azacitidine.

Although the minimum duration of treatment needed with azacitidine to discriminate between responsive and unresponsive patients is unknown, between 4 and 6 cycles are in general sufficient to demonstrate a response to treatment for most patients [19, 24]. Findings from the CALGB and AZA-001 trials showed that the median time to a first response was 2 to 3 cycles, with approximately 90% of responses occurring within the first 6 cycles of treatment [19, 24]; the overall survival benefit in the AZA-001 trial was shown in a median of 9 cycles [5]. Indeed, findings from a retrospective study of clinical practice in the Republic of Korea showed that 95% of responses to azacitidine in patients with MDS (33% with higher-risk MDS and 57% with intermediate-1 MDS defined using the IPSS) occurred within 4 cycles of treatment, with only 3 additional patients showing a response after 5 cycles [22].

More recently, a subgroup analysis of the AZA-001 trial data showed that although most patients (91%) responded within the first 6 cycles, the best response was achieved by 92% of responders after treatment with azacitidine was continued for an additional 6 cycles [24]. This finding is supported by a recent study in the Republic of Korea comparing outcomes of patients with MDS (65% with lower-risk MDS) who received between 4 and 8 cycles (N=106) with those who received more than 8 cycles (median=11 cycles, N=34) (Moon et al. 17th Congress of the European Hematology Association. June 14-17, 2012. Amsterdam, the Netherlands). Patients who were treated with more than 8 cycles of azacitidine had higher response rates (67.6% vs. 29.2%) and prolonged overall survival (890 days vs. 612 days).

The adverse events that are commonly encountered during treatment with azacitidine include hematologic adverse events (neutropenia, thrombocytopenia, and anemia), gastrointestinal adverse events (primarily nausea, vomiting, diarrhea, and constipation), injection-site reactions, fatigue, and pyrexia [5, 6, 20, 25]. However, most adverse events are transient, occur primarily within the first 2 cycles of treatment, and decrease in frequency with ongoing treatment [25].

Infection and bleeding episodes can worsen with azacitidine treatment.In the patients who were randomized to azacitidine in the CALGB trial (N=150), the number of grade 3 or 4 infections and bleeding episodes was estimated to be 0.21 and 0.14 events per patient-year of exposure, respectively [25]. In the subgroup of patients in the AZA-001 trial who were preselected to receive best supportive care and who were randomized to azacitidine (N=114), the number of grade 3 or 4 infections and bleeding episodes was estimated to be 0.51 and 0.34 events per patient-year of exposure, respectively [25]. The number of infections requiring treatment with intravenous antimicrobial drugs in all patients randomized to azacitidine (N=175) in the AZA-001 trial was 0.6 per patient-year [5]. Although most deaths during the first 3 months of treatment in the AZA-001 trial (11% of patients on azacitidine) occurred because of underlying disease, there were 2 cases of sepsis and 2 cases of bleeding that were thought to be related to treatment and resulted in death [5].

Most adverse events can be managed effectively without the need for treatment discontinuation, dose delays, or dose reductions [25]. Findings from the AZA-001 trial showed that 86% of patients treated with azacitidine were able to remain on the standard dose without the need for a dose adjustment and, of the 32 patients who received 6 or more cycles, 62.5% did not require a dose adjustment [5, 25]. These findings are supported by a retrospective analysis of patients in clinical practice in the Republic of Korea showing that 76% of patients were able to receive 80% of their dose per cycle [21].

Insufficient data are available to support the prophylactic use of antiviral or antifungal agents or granulocyte colony-stimulating factor (G-CSF) in patients with MDS. However, some evidence is available to support the use of prophylactic antibiotics in the reduction of the incidence of fever. Findings from a retrospective study in the Republic of Korea showed that oral administration of prophylactic antibiotics reduced the incidence of febrile episodes in patients with MDS [26].

Traditionally, the goals of treatment for patients with lower-risk MDS have been to improve hematologic outcomes, improve quality of life, and reduce the risk of progression to higher-risk MDS or AML. The choice of treatment is currently based on the individual patient's transfusion needs and the presence of clinically significant cytopenias [1]. In general, treatment with a hypomethylating agent is recommended for patients with thrombocytopenia or neutropenia or in patients with anemia who are unresponsive to other therapies [1]. However, the prognosis of patients with lower-risk MDS is highly variable, with survival rates that vary from several months to many years [31] and with most patients dying because of secondary infections and systemic hemorrhage before progression to AML [32, 33]. As such, patients with lower-risk MDS and a poor prognosis may benefit from early and more aggressive therapeutic interventions, irrespective of their transfusion needs [34].

A better understanding of the natural course of disease and prognostic variables that influence survival is fundamental to improving outcomes for patients with MDS. However, a major challenge in improving the survival of patients with lower-risk MDS is identifying which patients have a poor prognosis for survival. Although simple to use, the IPSS does not accurately predict the prognosis of patients with lower-risk MDS and does not take into account the more recently identified prognostic variables, including the considerable influence of specific cytogenic abnormalities on prognosis [31, 35-37]. The World Health Organization classification-based Prognostic Scoring System incorporates variables known to affect prognosis that are not included in the IPSS (e.g., transfusion dependency) [38, 39] and can identify some subgroups of patients at risk for a poor prognosis in the lower-risk IPSS groups [40]. In addition, the recently published revised-IPSS [37] stratifies patients into 5 rather than 4 prognostic risk groups and has incorporated revised low marrow blast categories and cutoff points for cytopenias, 5 (vs. 3) cytogenetic subgroups and 16 (vs. 6) specific cytogenic abnormalities, and additional prognostic variables for survival (age, performance status, serum ferritin level, lactate dehydrogenase level, beta2-microglobulin level, complex [>3] cytogenic abnormalities). However, external validation of the revised-IPSS is now required to confirm whether this system allows better prediction of clinical outcomes in patients with MDS, particularly lower-risk MDS, in clinical practice.

Currently, there is limited clinical information to support the use of azacitidine for the treatment of patients with lower-risk MDS. Most studies assessing the use of azacitidine in patients with lower-risk MDS are small, noncomparative studies that focus primarily on treatment response (Table 1). Findings from these studies show that azacitidine is a feasible option for patients with lower-risk MDS who are elderly, transfusion dependent, have severe thrombocytopenia or neutropenia, and are refractory to previous treatment. In addition, findings from these studies have shown that the adverse event profile of azacitidine in patients with lower-risk MDS is acceptable [41] (and Bally et al. ASH Annual Meeting. Abstract 2786. Blood 2011;118; Filì et al. ASH Annual Meeting. Abstract 4029. Blood 2010;116; Garcia et al. ASH Annual Meeting. Abstract 3801. Blood 2009;114; Grinblatt et al. ASH Annual Meeting. Abstract 1646. Blood 2008;112; Tobiasson et al. ASH Annual Meeting. Abstract 3798. Blood 2011;118). Consistent with findings from the AZA-001 and CALGB clinical trials, the most frequently reported adverse events in these studies were myelosuppression and gastrointestinal conditions.

Two prospective studies have assessed the use of shorter dosing schedules of azacitidine for patients with lower-risk MDS [23] (and Filì et al. ASH Annual Meeting. Abstract 4029. Blood 2010;116). Similar hematologic improvement was demonstrated in response to azacitidine with a 5-day (50%) and a 7-day interrupted (49%) dosing schedule in patients with lower-risk MDS (defined using the FAB system) [23]. In comparison, a 5-day dosing schedule in elderly patients with lower-risk MDS (defined using the IPSS) who were transfusion dependent elicited a lower hematologic response (39%) (Filì et al. ASH Annual Meeting. Abstract 4029. Blood 2010;116). The use of a shorter dosing schedule has yet to be validated. At least one study is under way in the Republic of Korea (NCT01652781, clinicaltrials.gov) to directly compare the administration of azacitidine using a 5-day dosing schedule with the standard approved dosing schedule in patients with lower-risk MDS.

Allogeneic HSCT is the only treatment available with the potential to cure patients with higher-risk MDS [34]. Findings from a retrospective analysis of 178 patients aged between 60 and 70 years with MDS suggests that there may be a survival benefit for allogeneic HSCT compared with azacitidine [43]. However, the therapeutic options for patients who experience a relapse after HSCT are limited [44] and many issues remain unresolved with regard to the timing of transplantation and the need to reduce the risk of relapse.

Pretransplantation therapy with a hypomethylating agent is currently recommended for patients who are eligible for HSCT but who are awaiting availability of a donor or who require treatment to reduce the risk of relapse after transplantation [1]. However, because of poor accrual rates, no prospective controlled trials have been conducted to evaluate the effect of pretreatment on outcomes after transplantation. Findings from retrospective studies (Table 2) that have been conducted to investigate the feasibility of pretransplantation therapy with azacitidine before HSCT [45-49] suggest that pretransplantation therapy with azacitidine is feasible and there does not appear to be an increased incidence of post-transplantation toxicities compared with no treatment or compared with standard induction chemotherapy [45-48]. In addition, a good response to azacitidine appears to be a predictor of better post-transplantation outcomes [45, 49]. Although favorable effects on survival and the risk of relapse after transplantation have been reported in patients receiving pretransplantation therapy with azacitidine, the retrospective nature of the studies precludes any firm conclusions.

The treatment options for patients who experience a relapse after HSCT include a second HSCT, salvage chemotherapy, and/or donor lymphocyte infusion [44]. Despite these options, the long-term survival for many patients with relapsed AML is poor [50, 51] and alternative treatment options are needed. As azacitidine is relatively well tolerated and can confer a survival benefit in patients with higher-risk MDS [6], several small clinical studies have investigated the use of azacitidine as pre-emptive therapy in patients with minimal residual disease [51] and as maintenance [52] or salvage therapy [52, 53] for recurrent MDS or AML after HSCT. Findings from these preliminary studies are encouraging; however, prospective randomized controlled trials in larger patient populations are needed to confirm the role of azacitidine after HSCT.

The therapeutic options for patients with higher-risk MDS who are not eligible for HSCT include high-intensive induction chemotherapy, hypomethylating agents, or supportive care [1]. Before the availability of hypomethylating agents, high-intensive chemotherapy was used to treat relatively younger patients with MDS. However, the response rate for these patients is lower than for patients with de novo AML and results in short duration of remission, particularly in patients with unfavorable cytogenetics [54]. Recent findings from a phase 2 clinical trial have demonstrated the efficacy of the thrombopoietic compound romiplostim for increasing platelet levels and decreasing thrombocytopenic adverse events [55] in patients with lower-risk MDS. However, the potential benefit of romiplostim and other stimulating agents, such as erythropoietin or G-CSF, or iron chelation for patients with higher-risk MDS is unclear.

In the AZA-001 trial, patients who received azacitidine had significantly longer overall survival than patients who received conventional care [5]. In this trial, conventional care comprised best supportive care, low-dose cytarabine, or intensive chemotherapy as recommended by the treating physician. Subgroup analyses showed that treatment with azacitidine resulted in significantly longer overall survival than best supportive care (21.1 months vs. 15.5 months, P=0.0045) or low-dose cytarabine (24.5 months vs. 15.3 months, P=0.0006). Although overall survival was numerically longer with azacitidine than intensive chemotherapy (25.1 months vs. 15.7 months, P=0.51), the trial was not designed or powered to demonstrate a statistically significant difference in overall survival between treatment with azacitidine and intensive chemotherapy.

Despite the favorable outcomes that can be achieved in patients with MDS who are treated with hypomethylating agents, up to 50% do not respond and most responders will experience disease progression within 2 years of a response [5]. Because of the limited treatment options available after treatment failure, patients with higher-risk MDS who fail to respond to treatment with hypomethylating agents have a very poor prognosis for survival [56, 57]. Retrospective analysis of patients with higher-risk MDS who failed treatment with azacitidine suggests that the factors associated with worse outcomes are older age, male sex, unfavorable cytogenetics, higher bone marrow blasts, and no previous hematologic response [57].

Of the limited options available for patients who fail treatment with azacitidine, allogeneic HSCT or treatment with an investigational agent is associated with better outcomes than best supportive care, low-dose chemotherapy, or intensive chemotherapy [57] (Lee J. 10th Korean Society of Hematology AML/MDS Working Party Spring Symposium. April 14, 2012. Seoul, Republic of Korea). A retrospective analysis of 435 patients who failed treatment with azacitidine showed that the median overall survival was 19.5 months for those receiving HSCT, 13.2 months for those receiving investigational therapy, 7 to 9 months for those receiving chemotherapy, and 4 months for those receiving best supportive care [57]. In addition, retrospective analysis of patients from the Republic of Korea showed that of the patients who fail hypomethylating treatment and undergo HSCT (N=23), better outcomes can be achieved if transplantation occurs before progression to AML (Lee J. 10th Korean Society of Hematology AML/MDS Working Party Spring Symposium. April 14, 2012. Seoul, Korea). The probability of 2-year overall survival in this study was 78.6% for patients who underwent HSCT during MDS versus 33.3% for patients who underwent HSCT after progression to AML. Limited information is available on the treatment options for patients who fail hypomethylating treatment and progress to AML. Findings from a recent retrospective analysis of these patients (N=46) in Korea suggests that active treatment with HSCT following intensive chemotherapy can confer a survival benefit over HSCT or intensive chemotherapy alone [58]. In this study, the median overall survival was 15 months for HSCT following intensive chemotherapy, 3.8 months for HSCT alone, 6.3 months for intensive chemotherapy alone, and 1.4 months for supportive care.