The role of glycosylated recombinant human granulocyte colony-stimulating factor (G-CSF) in the induction treatment of older adults with acute myeloid leukemia (AML) is still uncertain. In this trial, a total of 722 patients with newly diagnosed AML, median age 68 years, were randomized into 4 treatment arms: (A) no G-CSF; (B) G-CSF during chemotherapy; (C) G-CSF after chemotherapy until day 28 or recovery of polymorphonuclear leukocytes; and (D) G-CSF during and after chemotherapy. The complete remission (CR) rate was 48.9% in group A, 52.2% in group B, 48.3% in group C, and 64.4% in group D. Analysis according to the 2 x 2 factorial design indicated that the CR rate was significantly higher in patients who received G-CSF during chemotherapy (58.3% for groups B + D vs 48.6% for groups A + C; P = .009), whereas no significant difference was observed between groups A + B and C + D (50.6% vs 56.4%, P = .12). In terms of overall survival, no significant differences were observed between the various groups. Patients who received G-CSF after chemotherapy had a shorter time to neutrophil recovery (median, 20 vs 25 days; P < .001) and a shorter hospitalization (mean, 27.2 vs 29.7 days; P < .001). We conclude that although priming with G-CSF can improve the CR rate, the use of G-CSF during and/or after chemotherapy has no effect on the long-term outcome of AML in older patients.