The purpose of the study was to compare the efficacy of 10% povidone-iodine with that of 0.5% chlorhexidine gluconate in 70% isopropyl alcohol for the prevention of peripheral intravenous catheter colonization in neonates. This was a multicenter, nonrandomized prospective study in a tertiary neonatal intensive care setting in which povidone-iodine and chlorhexidine gluconate were each used as antiseptic skin preparations over sequential 6-month periods. During the first 6 months of the study when povidone-iodine was in use 9.3% (38 of 408) of catheters were colonized. During the second 6 months of the study when chlorhexidine gluconate was in use, catheter colonization occurred in 4.7% (20 of 418, P = 0.01). Catheter-related bacteremia occurred during only 0.2% (2 of 826) of all catheterizations. Heavy skin colonization before catheter insertion (relative risk, 3.6; 95% confidence interval, 1.9, 7.0), catheterization > or = 72 hours (relative risk. 2.0; 95% confidence interval, 1.01, 3.8) and gestational age < or = 32 weeks (relative risk, 1.8; 95% confidence interval, 1.02, 3.3) increased colonization risk. Ampicillin infusion (relative risk, 0.4; 95% confidence interval, 0.2, 0.7) and 0.5% chlorhexidine gluconate cutaneous antisepsis (relative risk, 0.4; 95% confidence interval, 0.2, 0.8) were factors associated with decreased colonization risk. We conclude that 0.5% chlorhexidine gluconate in 70% isopropyl alcohol appears to be more efficacious than 10% povidone-iodine for the prevention of peripheral intravenous catheter colonization in neonates.