The use of granulocyte colony-stimulating factor (G-CSF) has been established to improve hematological recovery after allogeneic bone marrow transplantation (BMT). The optimal timing to start with G-CSF has not been determined. This study investigates whether delayed use of G-CSF starting on day 6 is as safe and efficient as starting treatment with G-CSF immediately after BMT. Thirty-eight patients undergoing allogeneic BMT were randomized to either receive post-transplant G-CSF treatment starting at day 1 or at day 6. The time to hematological recovery was monitored and the groups were compared with respect to peritransplant morbidity and mortality. Recovery of the neutrophil granulocyte counts (PMN) to >100/microl, >500/microl and >1000/microl was comparable in both groups. The nadir of the PMN counts after stopping G-CSF was also similar. There was no difference in the recovery of red blood cells and platelet counts and no difference between the two groups with respect to the number of febrile episodes, number of days with antibiotics or number of documented bacterial, fungal or viral infections. Delayed treatment with G-CSF resulted in a reduction of G-CSF treatment from 19 days to 14 days (P = 0.0017). Reducing the length of treatment by 5 days lowered G-CSF treatment costs by 26.3%. Therefore, postponing treatment with G-CSF has no influence on the hematological recovery after allogeneic BMT. There is an economical benefit of postponing G-CSF use without any clinical disadvantages.