OBJECTIVE To characterize the effects of granulocyte colony-stimulating factor (G-CSF) or granulocyte-macrophage colony-stimulating factor (GM-CSF) on clinical outcomes in neutropenic HIV-infected patients, by means of a retrospective cohort study at an urban teaching hospital. METHODS Data were reviewed from all patients discharged between January 1, 1996, and August 31, 1997, with human immunodeficiency virus and neutropenia (absolute neutrophil count (ANC) <1000 cells/mL), with outcome measures of length of stay, infectious complications, and survival to discharge. RESULTS Of the 228 discharged patients who met selection criteria, 71 had received G-CSF or GM-CSF; 157 controls had not. Cases had lower CD4+ cell counts (30 vs. 54 cells/mL; P = 0. 017) and lower nadir ANCs (372 vs. 579 cells/mL; P < 0.001). Granulocyte-CSF or GM-CSF usage was not associated with the frequency of site-related infections, fever, or sepsis (all P > 0. 20). No difference was found in duration of hospitalization (23 vs. 21 days; P > 0.20). In a logistic regression model for survival to discharge, higher nadir ANC and CSF use were independently associated with improved survival (P = 0.034 and P = 0.026, respectively). CONCLUSIONS Use of G-CSF or GM-CSF was associated with improved survival to discharge among hospitalized HIV-infected patients with neutropenia.