Granulocyte colony-stimulating factor has differing effects comparing intravascular versus extravascular models of sepsis

BACKGROUND: Previously, neutrophil stimulation with granulocyte colony-stimulating factor (G-CSF) pretreatment increased survival rates in canines challenged with intraperitoneal or intrabronchial Escherichia coli and in rats challenged with intrabronchial Staphylococcus aureus. We investigated whether G-CSF pretreatment would be beneficial with intravascular challenge in these models. METHODS: Animals were randomized to G-CSF or placebo pretreatment followed by intravenous E. coli challenge in canines (n = 24) or intravenous or intrabronchial S. aureus challenge in rats (n = 273). All animals were treated with antibiotics. RESULTS: In canines, G-CSF before intravenous E. coli did not decrease mortality rates (7 of 12 [58%] G-CSF vs. 5 of 12 [42%] controls), which contrasted with prior reductions during extravascular infection (10 of 35 [29%] G-CSF vs. 37 of 65 [57%] controls). Consistent with the present and previously published studies in canines, in rats, G-CSF decreased mortality rates with intrabronchial S. aureus (22 of 90 [24%] G-CSF vs. 26 of 51 [51%] controls, p = 0.009) but did not decrease them with intravenous infection (34 of 67 [50%] G-CSF vs. 27 of 65 [42%] controls, p = 0.2) in patterns that were very different (p = 0.005 for the effects of G-CSF with intravascular vs. intrabronchial S. aureus). CONCLUSION: In contrast to extravascular infection, sepsis with intravascular E. coli in canines and S. aureus in rats may not provide a compartmentalized nidus of bacteria on which G-CSF-stimulated neutrophils can exert a beneficial antimicrobial effect. Extrapolated clinically, a proinflammatory agent like G-CSF may be most beneficial with sepsis related primarily to a compartmentalized extravascular site of infection.