PURPOSE Postoperative infections are a frequent source of preventable morbidity and mortality in the oncologic population. Granulocyte-macrophage colony-stimulating factor (GM-CSF) is a potent modulator of immune effector cells in vitro and in vivo. This study was conducted to determine whether GM-CSF, when administered perioperatively, could reduce the incidence of surgical infections in cancer patients. METHODS This was a prospective, randomized, placebo-controlled, multicenter study. Cancer patients at high risk of infectious surgical morbidity were randomized to receive GM-CSF 125 microg/m2 per day or placebo subcutaneously for 8 days beginning 3 days preoperatively. Routine antibiotic prophylaxis was administered to all patients. RESULTS Three hundred ninety-nine patients were enrolled, with 198 randomized to receive GM-CSF. Twenty-one percent of patients experienced infections during the first 2 weeks postoperatively, and there was no difference in infection rate between the study groups. The most common sites of infection were respiratory tract (53%) and surgical wound (25%). The duration of operation and American Society of Anesthesiology (ASA) physical status classification were the most significant predictors of infection in multivariate analysis. GM-CSF was well tolerated and was not associated with fever. CONCLUSION The eligibility criteria for this study were successful at defining a patient subgroup at high risk for postoperative infections. At an immunomodulatory dose of 125 microg/m2 per day, GM-CSF was safe and well tolerated, but did not reduce the incidence of postoperative infections in this high-risk oncologic population. Infectious morbidity in surgical oncology remains an important subject for continued clinical investigation.