Few studies have evaluated granulocyte colony-stimulating factor (G-CSF) priming in elderly patients with intensively treated acute myeloid leukemia (AML), and no data are available for genetically defined AML subgroups. We provide long-term results (median follow-up 7.6 years) of a randomized trial in which 183 patients (median age 67 years) received G-CSF prior to (G-CSF priming) or after two cycles of induction chemotherapy. CR rates with G-CSF priming and G-CSF post-chemotherapy were comparable (57 vs. 67 %, p = 0.153), with overall survival (OS) probabilities of 14 vs. 17 % at 10 years. Induction mortality was significantly higher with G-CSF priming (23 vs. 10 %, p = 0.015), primarily in normal karyotype (NK) AML. In this subgroup, a trend for better relapse-free survival (RFS) was observed with G-CSF priming (44 vs. 22 % at 10 years, p = 0.074) but did not translate into an OS benefit. G-CSF priming had no impact on AML with FLT3-ITD and NPM mutations and did not improve outcome in patients with adverse cytogenetics. In a landmark analysis, late consolidation with autologous stem cell transplantation or a second consolidation cycle significantly improved RFS compared with one consolidation cycle (21.0 vs. 12.8 months, p = 0.046). Future studies on G-CSF priming should be restricted to NK AML and used only in post-remission therapy.