PURPOSE: To review the literature on prevention of intravascular catheter-related infections. DATA SOURCES: The MEDLINE database, conference proceedings, and bibliographies of review articles and book chapters were searched for relevant articles. Primary authors were contacted directly if data were incomplete. STUDY SELECTION: Studies met the following criteria unless otherwise stated: Trials were prospective and randomized; catheters were inserted into new sites, not into old sites over guidewires; catheter cultures were done by using semi-quantitative or quantitative methods; and, for prospective studies, catheter-related bloodstream infection was confirmed by microbial growth from percutaneously drawn blood cultures that matched catheter cultures. DATA EXTRACTION: Data on population, methods, preventive strategy, and outcome (measured as catheter-related bloodstream infections) were gathered. The quality of the data was graded by using preestablished criteria. DATA SYNTHESIS: The recommended preventive strategies with the strongest supportive evidence are full barrier precautions during central venous catheter insertion; subcutaneous tunneling short-term catheters inserted in the internal jugular or femoral veins when catheters are not used for drawing blood; contamination shields for pulmonary artery catheters; povidone-iodine ointment applied to insertion sites of hemodialysis catheters; specialized nursing teams caring for patients with short-term peripheral venous catheters, especially at institutions with a high incidence of catheter-related infection; no routine replacement of central venous catheters; antiseptic chamberfilled hub or hub-protective antiseptic sponge for central venous catheters; and use of chlorhexidine-silver sulfadiazine-impregnated or minocycline-rifampin-impregnated short-term central venous catheters if the rate of infection is high despite adherence to other strategies that do not incorporate antimicrobial agents (for example, maximal barrier precautions). CONCLUSIONS: Simple interventions can reduce the risk for serious catheter-related infection. Adequately powered randomized trials are needed.