Granulocyte-macrophage colony-stimulating factor (GM-CSF) was administered to 40 pediatric patients undergoing partially matched related, or closely matched unrelated, allogeneic marrow transplants. This trial was set up in a prospective, randomized, double-masked, placebo-controlled manner to establish if the administration of GM-CSF to such patients enhanced neutrophil recovery in this allogeneic transplant setting. The GM-CSF group had a significantly shorter time to neutrophil recovery to > 500 x 10(9) cells/L (15 days) than the placebo group (p = 0.0036). In addition, the GM-CSF group had a significantly shorter neutrophil recovery time to > 1,000 x 10(9) cells/L (18 days) than the placebo group (p = 0.0053). The primary objective of this study was met by showing that GM-CSF enhanced neutrophil recovery in this allogeneic setting. However, within the study group of patients, there was no effect of GM-CSF on the incidence or severity of graft-vs.-host disease (GvHD), one of the secondary end-points of the study. With regard to the other secondary end-points, there was no effect of GM-CSF on marrow cellularity, duration of systemic antibiotics given for real infections or as prophylaxis to prevent infections, risk of significant infections (as defined by systemic culture of virus, fungus, or bacteria), and duration or cost of hospitalization, platelet recovery, and nutritional support. With the secondary end-points, it will be necessary to study larger numbers of pediatric patients to identify differences that are small in this study group. In conclusion, GM-CSF can be safely administered to children with few, if any, significant side-effects. Additional work remains to facilitate earlier discharge of patients and decreased antibiotic usage, to offset the cost of using a neutrophil growth factor.