The administration of G-CSF post transplant has been shown to accelerate the time to neutrophil engraftment. However, this does not necessarily translate into a meaningful clinical benefit to the patient. This randomized study was designed to determine the role of G-CSF following transplantation in patients with breast cancer (BC). A total of 241 evaluable patients with BC were included. There were 200 patients with high-risk BC, and 41 had disseminated BC in complete remission. All patients received conventional dose chemotherapy prior to transplantation. Patients were mobilized with G-CSF, received the STAMP V regimen, were transplanted with > or = 2.5 x 10(6) of CD34(+) cells/kg and were then randomized to receive 5 microg/kg of G-CSF starting on the day of infusion (arm A), five days later (arm B), or no G-CSF (arm C). The need for transfusion support, infectious complications and length of hospitalization were the variables chosen to demonstrate clinical benefit. Patients receiving G-CSF reached 500 and 1000 neutrophils significantly faster (P = 0.001) than patients with no G-CSF. This translated into a significantly (P < 0.05) shorter hospitalization time for patients receiving G-CSF. Arm C was closed and, after recruiting 110 patients in arm A, and 106 in arm B, the significant difference in neutrophil recovery persisted with no difference in the time of hospitalization between arms A and B. Therefore, G-CSF significantly accelerates the time to neutrophil engraftment. This translates into a shorter time of hospitalization. There is no difference in this variable regarding the time of administering the G-CSF: day 0 vs day +5. Therefore, G-CSF on day +5 should be the standard in this setting.