In an attempt to decrease the incidence of central venous catheter sepsis in children with cancer, we conducted a study to evaluate the benefit of adding broad-spectrum antibiotics to the catheter "flush solution." In a prospective, placebo-controlled, double-blinded, randomized trial, 69 children with different types of malignancies were studied. The central venous catheters in these children were flushed with either the standard solution (normal saline + 100 U/ml of heparin) or the study solution (25 microgram/ml of both amikacin and vancomycin added to the standard solution). At the conclusion of the study, 64 children with a total of 67 indwelling central venous lines were assessable. The total catheter days on study were 20,700 days, with a median of 323 catheter days per patient. We documented 10 events of catheter-related infections (0.49 events/1,000 catheter days at risk). Five of these events were catheter-related sepsis (0.24 sepses/1,000 catheter days): two were fungal and three were bacterial. Due to the low incidence of catheter-related sepsis in this study, no statement regarding the prophylactic use of antibiotics could be made. The extremely low rate of catheter-related sepsis reported herein may be retrospectively attributed to continuous staff education regarding aseptic techniques in handling these catheters. Staff education is essential, and probably the most effective factor in preventing catheter-related sepsis.