During a 5-year period, 6 patients with clozapine-associated agranulocytosis who received granulocyte colony-stimulating factor (G-CSF) were compared with 5 subjects who did not receive this treatment. Seven patients were asymptomatic, and the weekly leukocyte count alone indicated agranulocytosis. The average duration of agranulocytosis was not significantly different between the treated and untreated groups (6.5 vs. 6.6 days), though the treated group had a significantly shorter average duration of hospitalization (8.2 vs. 13.5 days). G-CSF administration was well tolerated, and no adverse effects were noted. Incidentally noted was the recent addition of ranitidine (with the potential for bone marrow toxicity) to clozapine for 2 patients. Currently, weekly white cell and differential counts remain the main tools for detecting incipient or occurring agranulocytosis. Until efficacy studies prove otherwise, G-CSF administered soon after the diagnosis of clozapine-associated agranulocytosis may shorten the duration of hospitalization and thus prove cost-effective.