ObjectivesTo compare the efficacy of a newly available antiseptic solution (composed of 0.25% chlorhexidine gluconate, 0.025% benzalkonium chloride, and 4% benzyl alcohol), with 10% povidone iodine, on the prevention of central venous or arterial catheter colonization and infection.DesignProspective, randomized clinical trial.SettingSurgical-trauma intensive care unit (ICU) in a university hospital.PatientsAll patients admitted to the ICU and requiring the insertion of a central venous and/or an arterial catheter from July 1, 1992 to October 31, 1993.InterventionsPatients were randomly assigned to one of two groups according to the antiseptic solution used for insertion and catheter care. The same solution was used for skin disinfection from the time of catheter insertion to the time of removal of each catheter.Measurements and Main Resultsor=to103colony-forming units [cfu]/mL by quantitative culture), and catheter-related sepsis (as defined by sepsis abating following catheter removal per 1,000 catheter-days), were significantly lower in the chlorhexidine group (12 vs. 31 [relative risk 0.4, 95% confidence interval 0.1 to 0.9, p < .01] and 6 vs. 16 [relative risk 0.4, 95% confidence interval 0.1 to 1, p = .05], respectively). The rate of central venous catheter colonization and central venous catheter-related sepsis per 1,000 catheter-days were also significantly lower in the chlorhexidine group (8 vs. 31 [relative risk 0.3, 95% confidence interval 0.1 to 1, p = .03] and 5 vs. 19 [relative risk 0.3, 95% confidence interval 0.1 to 1, p = .02], respectively). Finally, the rate of arterial catheter colonization per 1,000 catheter-days was significantly lower in the chlorhexidine group (15 vs. 32 [relative risk 0.5, 95% confidence interval 0.1 to 1, p = .05]), whereas the rate of arterial catheter-related sepsis per 1,000 catheter-days was similar for the two study groups (8 in the chlorhexidine group vs. 10 in the povidone iodine group [relative risk 0.8, 95% confidence interval 0.1 to 2.2, p = .6]). The 0.25% chlorhexidine solution was superior to the 10% povidone iodine solution in preventing catheter colonizations and catheter-related sepsis due to Gram-positive bacteria (5 vs. 20 [p < .001], and 2 vs. 10 [p < .001], respectively), whereas the activity of the 0.25% chlorhexidine solution was nonsignificantly superior in preventing Gram-negative infections (7 vs. 4 [p = .5], and 4 vs. 2 [p = .8], respectively).ConclusionsThe 4% alcohol-based solution of 0.25% chlorhexidine gluconate and 0.025% benzalkonium chloride was more effective than 10% povidone iodine for insertion site care of short-term central venous and arterial catheters. This effect appeared related to a more efficacious prevention of infections with Gram-positive bacteria.(Crit Care Med 1996; 24:1818-1823)