Optimal timing (Preemptive versus supportive) of granulocyte colony-stimulating factor administration following high-dose cyclophosphamide.

PURPOSE: The aim of this study is to compare various time schedules of granulocyte colony-stimulating factor (G-CSF) treatment in a clinical model of patients who received high-dose cyclophosphamide (HDCY 45 g/m2) for the treatment of an underlying malignancy in order to investigate the optimal time (preemptive vs. supportive) of G-CSF initiation upon the incidence and duration of cytopenias and related parameters, such as incidence of febrile episodes, antibiotic use, duration and cost of G-CSF administration and overall clinical benefit and cost effectiveness of various schedules used.

PATIENTS AND METHODS: Seventy-two courses were given in a sequential cohort study. G-CSF was administered either 24, 48, 72, 96 h after chemotherapy (preemptive treatment) or upon the onset of leukopenia (WBC </=1,000/ microl) (supportive treatment). Study parameters were compared among the various groups as well as to a control group who received HDCY without G-CSF support.

RESULTS: (1) Patients who received G-CSF early (24, 48 h) had a shorter duration of leukopenia (WBC </=1,000/ microl) compared to those who received G-CSF at a later stage (72, 96 h) or as supportive treatment (p < 0. 05). However the duration of neutropenia (ANC </=500/ microl) or thrombocytopenia (platelets </=20,000/ microl) was not affected by the different time schedules of treatment. (2) Patients who received G-CSF early (up to 72 h) had less febrile days with neutropenia in comparison to late treatment (>96 h), supportive and control groups (p < 0.05). The cost of antibiotics was also in favor of the early treatment group. The median duration of febrile days of the delayed (>72 h) treatment groups and antibiotic cost was similar to those in patients who did not receive G-CSF at all. (3) When G-CSF was given preemptively a shorter time was required to reach normal WBC (5,000/ microl) in comparison to the sup- portive and control group. This was due to a prolonged WBC recovery rather than to an early onset of leukopenia (tail effect). A delayed leukopenia recovery occurs as administration of G-CSF is delayed. (4) As a result the required length of G-CSF treatment to reach normal WBC (5,000/ microl) was shorter in the early treatment group and the cost from G-CSF use was less in that group in comparison to the late (>72 h) and supportive groups which indicated an increased cost without clinical benefit over controls.

CONCLUSIONS: G-CSF administration after HDCY has a similar effect upon the incidence and duration of severe leukopenia and thrombocytopenia. However, severe leukopenia is shorter when G-CSF starts up to 72 h after HDCY. The length of G-CSF administration and its cost is also in favor of early initiation of treatment as well as the number of febrile days and antibiotic use. Delayed (>72 h) or supportive treatment indicate more febrile episodes, antibiotic use and higher cost when compared to the early groups. Late (>72 h) or supportive G-CSF administration in this study indicates no benefit versus no treatment in relation to length of leukopenia, febrile days, antibiotic use and overall treatment cost.