Granulocyte–colony-stimulating factor (G-CSF) is effective in accelerating neutrophil recovery after intensive chemotherapy for acute myeloid leukemia (AML). However, the optimal G-CSF dosage for patients with AML has not been determined. To the authors' knowledge, G-CSF dosages have not been compared in a randomized AML study.
Patients who were enrolled on the St. Jude AML97 protocol and remained on study after window therapy were eligible to participate. The effect of the dosage of G-CSF given after induction chemotherapy Courses 1 and 2 was analyzed in 46 patients who were assigned randomly in a double-blinded manner to receive either 5 μg/kg daily or 10 μg/kg daily of G-CSF. The number of days of G-CSF treatment, neutropenia (an absolute neutrophil count <0.5 × 109/L), and hospitalization; the number of episodes of febrile neutropenia, grade 2 through 4 infection, and antimicrobial therapy; transfusion requirements; the cost of supportive care; and survival were compared between the 2 study arms.
No statistically significant differences were observed between the 2 arms in any of the endpoints measured.
Intensive chemotherapy1-4 and advances in supportive care5, 6 have improved the rates of remission and long-term survival in children with acute myeloid leukemia (AML). Granulocyte–colony-stimulating factor (G-CSF) is effective in accelerating neutrophil recovery in patients who are receiving intensive chemotherapy for AML.7, 8 In several adult AML studies, prophylactic use of 5 μg/kg daily of G-CSF after induction and consolidation therapy reduced the duration of neutropenia, infection, and antibiotic use.9-11 However, the duration of febrile episodes, the frequency of documented infection, the duration of hospitalization, and overall survival were not affected. It is unknown whether higher doses can extend the benefits of G-CSF beyond reduction of the duration of neutropenia.
G-CSF generally has been used at doses of 10 μg/kg daily or more in the setting of autologous bone marrow stem cell collection12,13; however, to our knowledge, no studies to date have compared different dosages of G-CSF in patients who are receiving chemotherapy for AML. Herein, we report the results from a randomized, double-blind trial that compared the effects of 2 different doses (5 μg/kg daily and 10 μg/kg daily) of G-CSF after induction treatment in children with newly diagnosed AML.
MATERIALS AND METHODS
Between March 1997 and June 2002, 102 children with previously untreated AML or myelodysplastic syndrome (MDS) were enrolled on the single-institution AML97 protocol at St. Jude Children's Research Hospital.3 Patients with acute promyelocytic leukemia were not eligible. The protocol was amended in May 1999 to compare the effects of 2 dosages of G-CSF given after remission induction chemotherapy. Patients who were enrolled on AML97 and remained on study after window therapy were eligible.
The study design specified double-blind, random assignment of at least 36 patients to receive G-CSF (filgrastim) at 5 μg/kg daily or 10 μg/kg daily intravenously after induction Courses 1 and 2. This number would provide 90% power to detect a 5-day difference in the number of neutropenic days (primary outcome measure) at an alpha level of .05 (1-sided test). Secondary outcomes that were compared between treatment arms were the number of days of G-CSF treatment and hospitalization; the cumulative number of febrile neutropenia episodes, episodes of grade 2 through 4 infection, antibiotic therapy courses, intravenous antibiotic therapy courses, and antifungal therapy courses; the number of erythrocyte and platelet transfusions; the cost of supportive care; and estimates of event-free survival and survival. The study was approved by the St. Jude Institutional Review Board, and signed informed consent was obtained from patients, parents, or guardians, as appropriate.
The AML97 treatment protocol has been described previously.3 Briefly, patients who agreed to participate in “upfront window” therapy were assigned randomly to receive either 5 daily, short infusions (2-hour) of cytarabine (500 mg/m2 daily; Arm A) or a 5-day, continuous infusion of cytarabine (500 mg/m2 daily; Arm B); both arms received 5 daily 30-minute infusions of cladribine (9 mg/m2). Induction chemotherapy Courses 1 and 2 comprised daunorubicin (30 mg/m2 daily as a continuous infusion on Days 1-3), cytarabine (250 mg/m2 daily as a continuous infusion on Days 1-5), and etoposide (200 mg/m2 daily as a continuous infusion on Days 4 and 5) (DAV1 and DAV2, respectively). Patients with high-risk AML (megakaryoblastic leukemia; refractory anemia with excess blasts in transformation; secondary AML; AML with -7, 5q-, or t(9,22) translocation; or persistent leukemia after DAV1) were eligible for allogeneic hematopoietic stem cell transplantation (HSCT) after DAV2. Patients with the t(8;21) or the inv(16) erase were classified with low-risk AML and were not eligible for allogeneic HSCT. All other patients were classified with standard-risk AML and were eligible for HSCT if a matched sibling donor was available. Patients who did not undergo allogeneic HSCT received 2 courses of consolidation chemotherapy, which consisted of cytarabine (3 g/m2 daily every 12 hours on Days 1, 2, 8, and 9) and L-asparaginase (6000 U/m2 after the fourth and eighth doses of cytarabine) followed by mitoxantrone (10 mg/m2 daily on Days 1-5) and cytarabine (1 g/m2/ daily every 12 hours on Days 1-3).
All patients received 1 intrathecal treatment with cytarabine at diagnosis. Patients without central nervous system (CNS) disease received 4 triple-intrathecal chemotherapy treatments with methotrexate, hydrocortisone, and cytarabine (with doses adjusted based on age) beginning during DAV1. Patients with CNS leukemia received triple-intrathecal therapy weekly until the cerebrospinal fluid was clear of leukemia cells (minimum, 4 doses), and then they received 4 additional doses.
Random Assignment to G-CSF Dosage
After May 1999, patients were assigned randomly before the start of DAV1 to receive 5 μg/kg daily or 10 μg/kg daily of intravenous G-CSF after DAV1 and DAV2. Daily 30-minute G-CSF intravenous infusions began 24 hours after the last day of each chemotherapy cycle and continued until the absolute neutrophil count remained ≥0.5 × 109/L for 2 days. The next chemotherapy cycle was started at least 24 hours after discontinuation of G-CSF. G-CSF was not administered to patients who were scheduled to undergo HSCT after DAV2 or to patients who had a poor response to DAV1 and, thus, were taken off of the AML97 protocol.
Patient features were compared between G-CSF treatment arms by using the exact chi-square test. Outcome variables were measured during the period beginning with the end of each DAV course and ending with the start of the subsequent chemotherapy course. The median number of days of G-CSF treatment in the 2 arms was compared separately for each induction cycle by using the Wilcoxon rank-sum test. A repeated-measures, mixed-effects model based on normal distribution was used to analyze the effect of G-CSF dosage on the number of days of neutropenia (absolute neutrophil count >0.5 × 109/L) and hospitalization and on the cost of supportive care, adjusting for chemotherapy course effect and modeling the correlation with an autoregressive structure.14 The supportive care charges in DAV1 and DAV2 were divided into 6 categories: antimicrobial agents, laboratory tests, diagnostic imaging tests, room/procedure charges, transfusion charges, and other general supportive care. The charges during induction therapy were calculated and log-transformed. Proportional means models were used to compare the cumulative number of febrile neutropenia episodes, episodes of grade 2 through 4 infection (according to National Cancer Institute Common Toxicity Criteria, version 2.0), antibiotic therapy courses, intravenous antibiotic therapy courses, antifungal therapy courses, and erythrocyte and platelet transfusions with G-CSF treatment as fixed covariate.15
Event-free survival was defined as the time between G-CSF randomization and disease recurrence, death, secondary malignancy, or last follow-up. Remission induction failure was treated as an event at Time 0. The Kaplan-Meier method was used to estimate the probability of event-free survival and survival; standard errors were determined according to the method of Peto et al.16 Survival distributions were compared by using the Mantel-Haenszel log-rank test.17 All analyses were performed using SAS software for Windows version 9.1 (SAS Institute, Cary, NC), and StatXact for Windows version 7.1 (Cytel Corporation, Cambridge, Mass).
Of 55 patients who were approached for enrollment, 5 patients declined, and 3 patients did not remain on the protocol after window therapy. Forty-seven patients were assigned randomly to receive G-CSF 5 μg/kg daily or 10 μg/kg daily. One patient was excluded from analysis because of treatment with granulocyte-macrophage–colony-stimulating factor at the physician's discretion; 46 patients were analyzed for DAV1, and 36 patients were analyzed for DAV2. The median patient age was 9.03 years (range, 0.05-21 years), and the median white blood cell count was 12.3 × 109/L (range, 1.2-166.8 × 109/L). Patient characteristics did not differ significantly according to G-CSF arm (Table 1).
There were no significant differences between the 2 G-CSF treatment arms in the duration of G-CSF treatment after DAV1 or DAV2 (Table 2). The number of neutropenic days also did not differ significantly in the 2 treatment arms (Table 2).
Table 2. Comparison of the 2 Granulocyte–Colony-Stimulating Factor Arms
Duration of Treatment, Neutropenia, and Hospitalization
G-CSF 5 μg/kg/d
G-CSF 10 μg/kg/d
G-CSF indicates granulocyte-colony-stimulating factor; DAV, daunorubicin, cytarabine, and etoposide.
Episodes of Febrile Neutropenia and Infection and Days of Hospitalization
There was no evidence that the number of febrile neutropenic episodes or episodes of grade 2 through 4 infection differed significantly between the 5 μg/kg daily and 10 μg/kg daily G-CSF arms (Table 2, Fig. 1A,B). The number of hospitalization days also did not differ significantly between the 2 arms (Table 2).
Episodes of Antibiotic and Antifungal Therapy
There was no evidence that the 5 μg/kg daily G-CSF arm differed significantly from the 10 μg/kg daily G-CSF arm in the number of antibiotic therapy courses, intravenous antibiotic therapy courses, or antifungal therapy courses (Table 2).
The number of erythrocyte and platelet transfusions did not differ significantly between the 2 G-CSF arms (Table 2).
Cost of Supportive Care
We observed no evidence that patients who received 5 μg/kg daily G-CSF differed significantly from those who received 10 μg/kg daily G-CSF with regard to total supportive care costs or any of the 6 categories of supportive care costs (Table 2).
Event-Free and Overall Survival
We observed no significant differences between the 5 μg/kg daily and 10 μg/kg daily G-CSF arms in the proportion of complete responses (Table 1) or in estimates of event-free survival or overall survival. The 6-year event-free survival estimate was 52.2% ± 10% for patients who received 5 μg/kg daily G-CSF and 39.1% ± 9.7% for patients who received 10 μg/kg daily G-CSF (P = .43) (Fig. 2). The 6-year survival estimate was 65.2% ± 9.6% for patients who received 5 μg/kg daily G-CSF and 52.2% ± 11.4% for patients who received 10 μg/kg daily G-CSF (P = .45).
Our prospective, randomized trial of 2 dosages of prophylactic G-CSF (5 μg/kg daily vs 10 μg/kg daily) in children who were receiving intensive induction chemotherapy for AML indicated that there was no significant difference in any of the outcomes measured. We acknowledge that statistical power for secondary outcomes may have been limited. However, to our knowledge, the current trial is the first randomized study comparing different dose of G-CSF in AML.
Several adult studies have demonstrated that G-CSF can modestly reduce (by 5-6 days) the duration of neutropenia when it is started shortly after induction chemotherapy.9, 10, 18, 19 However, it is not clear whether this acceleration of neutrophil recovery is clinically meaningful. Endpoints such as duration of hospitalization and frequency of severe infection have been reduced only occasionally and modestly. Furthermore, in the majority of studies, growth factor therapy did not alter the likelihood of complete remission, disease-free survival, or overall survival.
The Berlin-Frankfurt-Munich (BFM) 98 study randomly assigned children with AML to receive G-CSF (5 μg/kg daily) or no G-CSF starting on Day 15 of the first and second induction therapy courses.20 The duration of neutropenia in their trial was significantly less in the G-CSF arm after the first induction course (cytarabine, idarubicin, and etoposide; median, 18 days vs 23 days; P = .02) and after the second induction course (cytarabine and mitoxantrone; 11 days vs 16 days; P = .0005), but the duration of thrombocytopenia, the frequency of grade 3 and 4 infection, and 5-year event-free survival did not differ significantly between the arms. In another pediatric AML study, Children's Cancer Group (CCG) 2891, patients who received intensive chemotherapy and 5 μg/kg daily of G-CSF were compared with historic controls who received similar treatment but without G-CSF.21 Although the duration of neutropenia, number of hospital days, and delays in planned therapy in that study were reduced significantly, the frequency of severe infection and estimates of event-free survival and survival were not affected.
Studies in healthy hematopoietic stem cell donors who received G-CSF suggest a correlation between the G-CSF dose and the level of circulating CD34 positive progenitor cells 4 to 7 days after administration.12, 13 All volunteers who received 10 μg/kg daily and underwent a single leukapheresis mobilized more than the minimum target number of stem cells, whereas volunteers who received lower daily doses of G-CSF did not consistently mobilize as many cells.12 However, the higher G-CSF dose did not have a comparable effect in patients who were undergoing autologous stem cell harvest after cytotoxic chemotherapy, probably because of their limited number of progenitor cells.22 Similarly, higher daily doses of G-CSF (range, 2-16 μg/kg) did not induce a greater improvement in neutrophil recovery than lower doses in adults and children with solid malignancies and lymphoma who had received chemotherapy or undergone autologous HSCT.23-25 These findings are consistent with our results, which failed to demonstrate any benefit provided by the higher G-CSF dose of 10 μg/kg daily.
Although we observed no statistically significant difference between the study arms in event-free survival or survival, both estimates were lower in patients who received 10 μg/kg daily. Increased relative expression of the Class IV G-CSF receptor isoform in AML uncouples the proliferative and maturational G-CSF receptor signaling pathways, and higher doses of G-CSF may enhance this effect.26, 27 In the BFM 98 study, patients with standard-risk AML that overexpressed the maturation-defective G-CSF receptor isoform IV had a significantly higher incidence of recurrence with G-CSF treatment.28 Further studies of the function and expression patterns of G-CSF receptor and its isoforms in AML cells are warranted.
Improved supportive care is likely to enhance treatment outcome in patients who are receiving AML therapy.1, 5, 6 The lower rate of early death in recent BFM-AML trials than in previous trials probably reflects both experience and improved supportive care.5 However, infectious mortality, particularly that caused by invasive bacterial and fungal infections, remains unacceptably high.29, 30 We recently reported that the routine prophylactic use of 1) vancomycin, oral ciprofloxacin or cephalosporin, and voriconazole or 2) cefepime and voriconazole reduced morbidity and dramatically decreased the incidence of septicemia and the number of hospital days.14 Improved 3-year event-free survival (63% ± 4.1%) and survival (71.1% ± 3.8%) in our recent multi-institutional AML02 study may be attributed in part to better supportive care.31 Likewise, the excellent survival rates in the Japanese Childhood AML Cooperative Study Group (5-year event-free survival, 61.6%; survival, 75.6%) were associated with routine inpatient care throughout the treatment course, allowing close patient monitoring and ready access to antibiotics.4
In conclusion, we suggest that first priority be given to the use of prophylactic antibiotics and close patient monitoring (by inpatient care or frequent outpatient visits) and that growth factors (eg, G-CSF 5 μg/kg daily) should be reserved for patients with extremely prolonged neutropenia and/or clinically complicated infection, ideally after confirming the expression pattern of G-CSF receptor isoform IV in AML cells. A recent survey of Children's Oncology Group and BFM group institutions revealed systematic differences in infection-related supportive care practices for pediatric AML patients.6 To improve supportive care in AML, harmonized use and/or randomized studies of antibacterial and antifungal prophylaxis and criteria for growth factor use and for discharge after chemotherapy are needed.
We acknowledge the expertise of Ms. Sharon Naron in editorial review of the article.
CONFLICT OF INTEREST DISCLOSURES
Supported in part by Cancer Center Support Grant CA 21765 from the National Institutes of Health and by the American Lebanese Syrian Associated Charities.