It is generally believed that elderly patients are less able to tolerate aggressive cancer chemotherapy than their younger counterparts. Bone marrow cellularity diminishes with age and elderly patients may have decreased tolerance to myelosuppressive agents. Between November 1995 and October 1999, 68 chemotherapy-naive elderly (70 or more years old) patients with histologically or cytologically proven lung cancer who were to receive platinum-based chemotherapy were enrolled in this study. All patients had adequate cardiac, hematological, liver and renal function to receive chemotherapy. Patients were randomized into 3 groups. Patients in groups 1 and 2 received 2 microg/kg and 4 microg/kg granulocyte colony-stimulating factor (G-CSF, lenograstim), respectively, when grade 3 leukopenia (<2,000/microl) or neutropenia (<1,000/microl) appeared after chemotherapy. Patients in group 3 received 2 microg/kg G-CSF when grade 2 leukopenia (<3,000/microl) or neutropenia (<1,500/microl) appeared after chemotherapy. G-CSF was stopped in all groups when the leukocyte count increased to over 10,000/microl or the neutrophil count exceeded 5,000/microl. Full blood cell counts were examined 3 times a week after chemotherapy. All patients received platinum-based chemotherapy. Eighteen, 16 and 22 patients (78%, 73% and 96%) in groups 1, 2 and 3, respectively, received G-CSF when leukopenia or neutropenia appeared. The durations of G-CSF treatment required by groups 1 and 3 (5.7+/-3.6 and 6.6+/-3.2 days, respectively) did not differ significantly, but the duration of treatment required by group 2 (3.7+/-2.8 days) was significantly shorter than that of group 1 (p=0.048). The duration of grade 4 neutropenia in group 2 (0.7+/-1.1 days) was marginally shorter than that in group 1 (1.6+/-2.1 days, p=0.076). The neutrophil nadir of group 2 (949+/-757/microl) was marginally higher than that of group 1 (592+/-438/microl, p=0.058). No patients in group 2 experienced grade 4 neutropenia for 4 days or more or a neutrophil nadir less than 100/microl a significant difference from group 1, where 22% and 17% of patients experienced these events (p=0.02 and p=0.04, respectively). Similarly, no infections requiring antibiotics after chemotherapy occurred in patients in group 2, a significant difference from group 1 (26%, p=0.01). The rates of neutropenia and infection in groups 1 and 3 did not differ significantly. The peak plasma concentration of G-CSF in group 2 was significantly higher than in group 1 (p=0.0018), but did not differ significantly between groups 1 and 3. Doubling the dose of G-CSF could help to decrease neutropenia and prevent infection after chemotherapy in elderly patients with lung cancer.