To assess the feasibility and efficacy of rhGM-CSF in ameliorating chemotherapy-induced leukopenia in patients with advanced non-small-cell lung cancer, we conducted a double-blind placebo controlled phase III study in a multicenter setting. Patients were eligible if they had cytologically or histologically proven cancer, no prior chemotherapy, stage IIIB or IV disease, an Eastern Cooperative Oncology Group (ECOG) performance status of 0-2, an age of less than 76 years, and no symptomatic brain metastasis, disseminated bone metastasis, or previous vertebral/pelvic irradiation. The chemotherapy regimen consisted of mitomycin given at 8 mg/m2 on day 1, cisplatin given at 100 mg/m2 on day 1, and vindesine given at 3 mg/m2 i.v. on days 1 and 8 (MVP). If the granulocyte nadir count recorded after the first cycle of MVP was less than 1,000/mm3, patients were randomly assigned to receive recombinant human granulocyte-macrophage colony-stimulating factor (rhGM-CSF) or placebo during the second cycle of MVP. The dose of rhGM-CSF was 125 micrograms/m2 given daily s.c. for 14 consecutive days starting on day 2. Of the 52 patients enrolled, 45 were evaluable. The nadir of granulocytes was significantly lower in the placebo group (P = 0.007). The period during which the granulocyte count was less than 1,000/mm3 was significantly longer in the placebo group (median, 6 vs 10 days; P = 0.04). The incidence of adverse effects related to rhGM-CSF, such as fever (> or = 38 degrees C) and skin rash, was significantly higher in the rhGM-CSF group (P = 0.011). The rate of response to chemotherapy did not significantly differ between the two groups. In conclusion, rhGM-CSF reduced the duration of chemotherapy-induced granulocytopenia. The clinical usefulness of this agent may be deminished because of the adverse effects encountered when it is used in combination with a moderately myelotoxic chemotherapy regimen.