Granulocyte colony-stimulating factor (G-CSF) increases the number of circulating granulocytes and decreases TNF production while improving survival in sepsis models. To study the effects of G-CSF administration on sepsis and rejection, 37 primary liver allograft recipients received intravenous recombinant human G-CSF (rhG-CSF; 5-10 micrograms/kg/day) for the first 7-10 days following transplantation, targeting a blood absolute granulocyte count of between 10,000 and 20,000 cells/mm3. These recipients were monitored prospectively for sepsis and rejection, as were the previous 49 primary liver allograft recipients who did not receive G-CSF. Both groups utilized identical protocol immunosuppression and standardized diagnosis and treatment of sepsis and rejection. Univariate and logistic regression analysis of risk factors for sepsis and rejection revealed no difference between the two patient groups. G-CSF-treated patients developed an increased absolute granulocyte count over time (P < 0.0001, repeated-measures analysis of variance). G-CSF-treated patients had a decreased number of sepsis episodes per patient (0.92 +/- 1.5 vs. 2.18 +/- 2.8, P < 0.02, t test), and a lower percentage of sepsis-related deaths (8% vs. 22%, P < 0.04, chi-square test). The incidence of acute rejection was decreased in the G-CSF-treated group (22% vs. 51%, P < 0.01, chi-square test). These pilot data support further investigation into G-CSF's favorable effects on sepsis and rejection.