Impaired neutrophil responses contribute to the neonate's increased susceptibility to infection. Because granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage colony-stimulating factor (GM-CSF) enhance granulocyte and macrophage number and function, their use in the management of neonatal sepsis may be beneficial. Little is known about the endogenous levels of G-CSF and GM-CSF. In adults, raised values for G-CSF, but not GM-CSF, have been demonstrated in patients with infection, and conflicting data has emerged regarding CSF levels in neonates. We have used an ELISA to measure maternal and cord serum G-CSF and GM-CSF at the time of delivery, with gestational age between 25 and 42 wk. In mothers, an inverse linear relationship between gestational age and GM-CSF levels (p= 0.049) was found, but no association with G-CSF levels was observed. In neonates, a quadratic association was found between GM-CSF levels and gestational age (p= 0.019), whereas G-CSF levels showed an inverse linear association (p= 0.015). In addition, an association was found between maternal and cord GM-CSF (p= 0.007) but not G-CSF levels in paired samples. The effect of gestational age on the cytokine levels could not be explained by the white cell count, the absolute neutrophil count, pregnancy-induced hypertension, or the presence of infection. We suggest that 1) GM-CSF levels in mothers vary throughout gestation and may have a role in the maintenance of normal pregnancy; 2) G-CSF and GM-CSF levels in neonates vary with gestational age and may have a developmental role; and 3) G-CSF has theoretical benefit in the management of neonatal neutropenia and sepsis; clinical trials are now needed to establish its optimal use.