PURPOSE: The aim of this study is to determine and compare the effects of adjunctive therapy with different doses of recombinant human granulocyte-colony stimulating factor(rhG-CSF) on reversing sepsis-associated neonatal neutropenia, and their survival rate in a group I/II-type trial. METHODS: RhG-CSF was injected subcutaneously to 10 septic-neutropenic neonates with doses of 10 microgram/kg from Oct. 1995 to Sep. 1996, and was administered to another 12 septic-neutropenic neonates with doses of 5 microgram/kg from Oct. 1996 to Sep. 1997. Neutrophilic responses and the outcomes of both groups were compared. RESULTS: In the rhG-CSF 10 microgram/kg treated group and in the 5 microgram/kg treated group, the absolute neutrophil count(ANC) was 1,065+/-89(mean+/-SEM) and 1,053+/-131, respectively. The only difference between the two groups was the peak ANC at 48 hours. Eight patients from the remaining nine of rhG-CSF 10 microgram/kg treated group(88.9%) and ten in 5 microgram/kg treated group(83.3%) survived the sepsis and were discharged without any problems. CONCLUSIONS: RhG-CSF can increase the neutrophil count in critically ill septic neutropenic neonats. The survival rate of both groups were up to 90%. This finding suggests that both doses of rhG-CSF may be effective in a therapeutically useful time frame to treat septic neonates with neonatal neutropenia attributable to bone marrow supression or neutrophil consumption.