Treatment costs and quality of life with granulocyte-macrophage colony-stimulating factor in patients with antineoplastic therapy-related febrile neutropenia. Results of a randomised placebo-controlled trial.

This study examined the costs of treatment of, and quality of life in, patients with antineoplastic therapy-induced neutropenic fever who were treated with antibacterials, with or without granulocyte-macrophage colony-stimulating factor (GM-CSF). Patients with haematological malignancies (n = 47) or solid tumours (n = 87) who had severe neutropenia (neutrophil count < 0.5 x 10(9)/L) and fever (> 38.5 degrees C once, or > 38 degrees C twice, in a 12-hour observation period) were randomised to receive subcutaneous GM-CSF 5 micrograms/kg/day (n = 65) or placebo (n = 69) in conjunction with broad-spectrum antibacterials. GM-CSF enhanced neutrophil recovery compared with placebo. Median neutrophil counts at day 4 were 2.9 (range 0 to 25) x 10(9)/L in the GM-CSF arm and 1.3 (range 0 to 9) x 10(9)/L in the placebo group (p < 0.001). No significant difference was observed with regard to median days with neutrophil count < or = 1.0 x 10(9)/L or in time to resolution of fever. Quality-of-life scores in 90 patients demonstrated significant differences in favour of the placebo group. The results for the oncology and haematology patients were similar to the results for the total group. Patients in the GM-CSF and placebo groups had a mean hospital stay of 7.25 and 8.33 days, respectively. Hospital costs were higher for the GM-CSF-treated patients when GM-CSF was included in the price [mean costs: GM-CSF arm $US 5177 vs placebo arm $US 4178 (p < 0.05; 1992 values)]. The haematology patients stayed longer in hospital than the oncology patients, resulting in higher total costs for the former group. These results indicate that GM-CSF does not affect the number of days required for resolution of fever of the hospitalisation period for this patient group, and does not provide a cost-effective contribution to the treatment of these patients. Sensitivity analyses indicate that GM-CSF would produce savings if the duration of hospitalisation with GM-CSF was < or = 76.5% of that in the placebo group.