To evaluate the preventive effects of topical skin disinfection with chlorhexidine on bloodstream infection (BSI) associated with totally implantable venous port (Port-A).Two consecutive cohorts of solid cancer patients were prospectively followed for the occurrence of Port-A associated BSI (PABSI). The first cohort used povidone-iodine as topical skin disinfection and the second cohort used chlorhexidine. The primary endpoint was the time to first PABSI. Propensity score analysis was applied. The preventive effects of chlorhexidine were analyzed by Cox proportional hazards models.There were 396 patients (81,752 catheter-days) in the iodine cohort and 497 (99,977 catheter-days) in the chlorhexidine cohort. Gram-negative bacteria were the most common pathogens to cause first episode of PABSI (iodine cohort (I) vs chlorhexidine cohort (C) and 0.404 vs 0.450 per 1,000 catheter-day), followed by Gram-positive bacteria (I vs C and 0.269 vs 0.110 per 1,000 catheter-day), and fungi (I vs C and 0.098 vs 0.070 per 1,000 catheter-day). Three hundred forty-three patients were selected from each cohort by propensity score match analysis. Chlorhexidine use was associated with a significant improvement on time to first PABSI caused by Gram-positive bacteria (log-rank test, p=0.00175; HR=0.35, 95 % CI, 0.14-0.85, p=0.02). No significant preventive effects of chlorhexidine on time to first PABSI caused by Gram-negative bacteria or fungi was found.Chlorhexidine topical skin disinfection may prevent PABSI caused by Gram-positive bacteria in patients with solid cancers. The nonsignificant effect on preventing overall PABSI may be attributed to the high incidence of Gram-negative bacteria related PABSI.
