Comparison of four skin preparation strategies to prevent catheter-related infection in intensive care unit (CLEAN trial): a study protocol for a randomized controlled trial.

Background: Catheter-related infection is the third cause of infections in intensive care units (ICU), increasing the length of stay in ICU and hospital, mortality, and costs. Skin antisepsis is one of the most prevalent preventive measures. In this respect, it would appear preferable to recommend the use of alcoholic povidone iodine or chlorhexidine rather than aqueous povidone iodine. However, the data comparing chlorhexidine to povidone-iodine, both of them in alcoholic solutions, remain limited. Moreover, the benefits of enhanced cleaning prior to disinfection of skin that is not visibly soiled have yet to be confirmed in a randomized study.

Methods: A prospective multicenter, 2 × 2 factorial, randomized-controlled, assessor-blind trial will be conducted in 11 intensive care units in six French hospitals. All adult patients aged over 18 years requiring the insertion of at least one peripheral arterial catheter and/or a non-tunneled central venous catheter and/or a hemodialysis catheter and/or an arterial pulmonary catheter will be randomly assigned to have all their catheters cared with one of four skin preparation strategies (2% chlorhexidine/70% isopropyl alcohol or 5% povidone iodine/69% ethanol with or without prior skin scrubbing). At catheter removal, catheter tips will be quantitatively cultured. Sets of aerobic and anaerobic blood cultures will be routinely obtained when a patient has fever, hypothermia, or other indications. In case of suspected catheter-related infection the patient's form will be reviewed by an independent adjudication committee. We plan to enroll 2,400 patients (4,800 catheters). The main objective is to demonstrate that use of 2% alcoholic chlorhexidine compared to 5% alcoholic povidone iodine in skin preparation lowers the rate of catheter-related infection. The second endpoint is to demonstrate that enhanced skin cleaning prior to disinfection of skin that is not visibly soiled does not reduce catheter colonization. Other outcomes include comparison of skin colonization at catheter insertion site, comparison of catheter colonization and catheter-related bacteremia taking place during implementation of the four strategies of skin preparation, and cutaneous tolerance, length of hospitalization, mortality, and costs.

Discussion: This study will help to update recommendations on the choice of an antiseptic agent to use in skin preparation prior to insertion of a vascular catheter and, by extension, of an epidural catheter and it will likewise help to update recommendations on the usefulness of skin scrubbing prior to disinfection when the skin is not visibly soiled.

Trial Registration: Clinicaltrials.gov number NCT01629550.