Background: Neutropenia occurs in up to 17% of human immunodeficiency virus (HIV)-infected individuals. Although granulocyte colony-stimulating factor (G-CSF) can reverse HIV-related neutropenia, it is not established that this therapy can reduce bacterial infections and affect survival.

Methods: A retrospective cohort study of 152 neutropenic, HIV-infected patients was performed to determine the therapeutic utility of G-CSF. Medical records of 71 patients who received G-CSF and 81 patients who never received G-CSF, during the years of 1991 to 1994 at Parkland Memorial Hospital, were reviewed for the incidence of bacteremia, G-CSF use, antiretroviral therapy (AR), Pneumocystis carinii pneumonia prophylaxis (PCPP), and opportunistic infections (OI).

Results: The two patient groups had similar baseline characteristics including CD4 count (37 cells/mm³ versus 40 cells/mm³, P = 0.7). Univariate analysis revealed and trend toward decreased rates of all bacteremias in the G-CSF-treated group compared with the controls (0.54 bacteremias/100 patient months versus 2.2 bacteremias/100 patient months, P = 0.064) and a marked decrease in the rates of gram-negative rod bacteremias in the G-CSF-treated group compared with the untreated group (0.09 gram-negative rod bacteremias/100 patient months versus 1.7 gram-negative rod bacteremias/100 patient months, P = 0.002). In a multivariate analysis, significant decreased risk for bacteremia was found with G-CSF use (odds ratio [OR] = 0.15, P = 0.02). Survival was longer in patients treated with G-CSF than in the untreated group (median: 397 days versus 165 days). Multivariate analysis using Cox Proportional Hazards Model showed decreased risk of death in patients treated with G-CSF, ARs, PCPP.

Conclusions: We conclude that G-CSF use is associated with decreased bacteremias and is associated with prolonged survival in neutropenic, HIV-infected patients.