Chlorhexidine-Impregnated Dressings and Prevention of Catheter-Associated Bloodstream Infections in a Pediatric Intensive Care Unit

We performed a randomized, controlled experimental study. The data were gathered between December 2012 and January 2014 in the PICU of a university hospital in Istanbul, Turkey. The 6-bed PICU admits an annual mean of 100 to 150 patients who have acute respiratory failure, sepsis, shock, multiple organ failure, or poisoning or who require postoperative follow-up.

This study was performed in accordance with the Declaration of Helsinki. Approval was granted by the academic board of the pediatrics department and the local ethics board of the university hospital before the research began. The families of the children involved in the study were provided information about the study and gave written consent.

The sample consisted of patients admitted to the PICU. We used NASS and PASS statistical software (NCSS, LLC) to perform a power analysis and determine the number of patients required. We found that a minimal sample size of 61 patients would have 80% power to detect a difference of 19% between development and no development of CRBSIs, as reported previously,^3,4,9–11  at α = .05. All patients who met the inclusion criteria were randomized and evenly distributed into chlorhexidine (n = 50) and standard (n = 50) groups (see Figure); the randomization was maintained by using a web-based program (http://www1.assumption.edu/users/avadum/applets/RandAssign/GroupGen.html). In order to avoid selection bias, the person who performed the randomization was not involved in recruiting participants or in obtaining informed consent for the study. The laboratory personnel who interpreted the microbiological results also had no knowledge of the study.

Patients were included in the study if they were 1 month to 18 years old, had no CRBSIs at the time of admission to the hospital, had had a central venous catheter in place for more than 72 hours, were not receiving any neuromuscular blockers, and had given written consent to be part of the study.

Catheters were covered with a sterilized pad in the standard group and with a 2% chlorhexidine-impregnated dressing in the chlorhexidine group. A patient diagnosis form was prepared and completed for each patient.

Catheter care in both groups was carried out as directed in the daily CRBSIs prevention bundle, which had been previously prepared and introduced to the PICU. For both groups, maximum barrier and aseptic precautions were taken during placement of catheters and during catheter care (hand washing, sterile gown with long sleeves, mask, large sterile drapes, and gloves). During placement of catheters, 10% povidone-iodine was used for dermal antisepsis, and the cleansing was maintained for 3 minutes. For catheter care, 10% povidone-iodine was used for dermal antisepsis, and cleansing was maintained for 3 minutes. Then the site was wiped with sterilized physiological saline and, after 30 to 60 seconds to allow for drying, was covered with a sterilized pad (standard group) or a 2% chlorhexidine-impregnated dressing (chlorhexidine group). For both groups, utmost care was taken to avoid getting the catheter wet when the patient’s body was being washed.

In the chlorhexidine group, the 2% chlorhexidine-impregnated dressing remained in situ for 7 days, unless it became wet. The CDC suggests that dressings used on the insertion sites of short-term central venous catheters should be replaced every 2 days for gauze dressings. However, we changed the gauze dressings daily because children’s skin is more sensitive than adults’ skin and frequent exposure of the catheter insertion site allowed earlier recognition of redness or changes. No guidewire exchange was needed because we use 3-lumen silicone catheters in our unit.

A patient diagnosis form with 4 sections was prepared in conformity with the relevant literature.^1,3,4,6  The form consisted of 35 open-ended and closed questions related to patient characteristics, clinical details, physiological parameters, and factors associated with the catheter.

When a catheter was to be removed, 2 samples of blood were obtained for microbial culture, 1 from the central venous catheter and 1 from the peripheral blood. In addition, the skin was cleansed with 10% povidone-iodine to prevent contamination of the catheter’s end by microorganisms on the skin, and then the 2- to 3-cm catheter end (tip) was removed and sent to the laboratory along with the blood samples. Catheters were removed when they were no longer required or they leaked, when the patient had a fever (with no other recognized focus of infection and redness at the catheter site), or when the patient died.

A catheter was considered sterile if both cultures of the catheter end and cultures of the 2 blood samples showed no growth of microorganisms or if cultures of both blood samples showed growth of the same organisms.

A catheter was considered contaminated if a culture of the catheter end had fewer than 15 colony-forming units (cfus) and no growth occurred in cultures of the blood samples or if cultures of the 2 blood samples showed growth of the same microorganisms but the microorganisms differed from the microbes in the culture of the catheter end. Colonization was defined as growth of 15 cfus or more in the catheter-end culture in the absence of local or systemic signs of infection and no growth in the cultures of the 2 blood samples or if cultures of the 2 blood samples showed growth of the same microorganisms but the microorganisms differed from the microbes in the culture of the catheter end.

Local catheter infection was defined as growth of 15 cfus or more in the culture of the catheter end and findings of inflammation at the catheter insertion site in the absence of blood-borne infection. CRBSIs were defined as growth of 15 cfus or more in the catheter-end culture and microorganisms in the 2 blood samples that have the same antibiotic resistance pattern as the microbes in the catheter end.

We used SPSS, version 17.0 for Windows (IBM SPSS), for analysis of the data. We calculated means, medians, number of patients, and percentages. A t test was used to determine if the means of χ² analyses differed from each other and to investigate whether dependence existed between the variables measured by using a 2 or more ordinal (serial) or nominal (categorical) scale.

Among the 100 patients, 63% were 1 month to 1 year old, 60% were male; 33% had respiratory system diseases; and 18% had additional diseases (Table 1). Catheters were in place in the 100 patients in the study group for a mean of 14.01 days (SD, 6.98 days), and the patients’ mean length of stay in the PICU was 15.35 days (SD, 10.10 days).

The standard group and the chlorhexidine groups were similar to each other in sex, age, diagnosis, existence of additional disease, nutritional status, and the vein used for catheterization (Table 1).

The results of the comparison of groups for the characteristics of the sterile catheter, colonization, contamination, local catheter infection, and CRBSI are shown in Table 2. Colonization, contamination, local catheter infection, and CRBSI did not differ significantly between the 2 groups (all P > .05).

Compared with the standard group, the chlorhexidine group had a shorter PICU length of stay, shorter duration of mechanical ventilation, and shorter duration of catheterization (Table 3). However, the differences between the 2 groups were not significant (all P > .05).

Vancomycin-resistant enterococci were isolated from 1 patient in the chlorhexidine group who had a CRBSI and from 2 patients in the standard group who had a CRBSI (Table 4). Among patients in the standard group who had a CRBSI, Pseudomonas aeruginosa was isolated from 1 patient, methicillin-resistant Staphylococcus aureus from 1 patient, and Acinetobacter from 1 patient. The pathogens isolated from the standard group did not differ from the pathogens isolated from the chlorhexidine group (P = .20).

Catheter care with an appropriate antiseptic solution and suitable dressings is efficient in preventing CRBSIs. In a meta-analysis of the solutions used in catheter care, Chaiyakunapruk et al¹  found that use of chlorhexidine rather than povidone-iodine for catheter cleansing reduced the rate of CRBSIs by 50% in patients whose catheter was in place for a short time. Conversely, Humar et al²  compared use of chlorhexidine with use of povidone-iodine for antisepsis during insertion of a central venous catheter and found no differences between the 2 products in the rate of CRBSIs.

The CDC recommends use of either chlorhexidine or povidone-iodine for catheter care. We used povidone-iodine for both groups in our study. Both the dressing used to cover the catheter and the solution used for antisepsis during catheter care have an impact on the rate of infection.³  Levy et al¹⁰  found that a dressing impregnated with chlorhexidine remarkably reduced colonization of central venous catheters. In a randomized, controlled study with 1636 patients, Timsit et al⁸  found that compared with a standard dressing, use of a chlorhexidine-impregnated dressing reduced the rate of CRBSIs. In a study in infants, Garland et al¹¹  compared skin cleansing with povidone-iodine with use of a chlorhexidine-impregnated dressing at the insertion site of central venous catheters. Colonization was remarkably reduced in the chlorhexidine group, but the 2 groups did not differ in the rates of CRBSIs. Arvaniti et al¹²  reported that silver-impregnated catheters and chlorhexidine-impregnated sponges with standard multilumen central venous catheters did not reduce colonization or infection rates. Similarly, we found no significant difference between the 2 groups in our study in colonization, contamination, local catheter infection, and CRBSIs.

However, we think that even though the lower values in the chlorhexidine group were not statistically significant, they are clinically important. A meta-analysis and systematic review by Rivas Ruiz et al⁹  indicated that dressings made with chlorhexidine gluconate pads were effective in reducing CRBSIs in infants and children and rarely were associated with serious adverse effects.

CRBSIs extend the length of hospital stay.^13,14  In our study, although the differences between the standard group and the chlorhexidine group were not statistically significant, the chlorhexidine group had a shorter ICU length of stay, shorter duration of mechanical ventilation, and shorter duration of catheterization. Moreover, significantly fewer colonies were found in the microbial cultures of the catheter ends in the chlorhexidine group.

During the study, the PICU in which the research took place was relocated to a smaller area because not enough units were available for care of patients with chronic neurological illnesses. Thus, patients stayed in the PICU longer than was necessary, a situation that limited the number of new patients admitted. The nurses and physicians responsible for care of our patients could not be blinded to the allocation of dressings because of the physical appearance of the chlorhexidine-impregnated dressings. This situation could be a source of possible bias.

Use of 2% chlorhexidine-impregnated sterilized dressings in the PICU reduced the rate of CRBSIs, contaminations, and colonization. In addition, compared with the standard group, the chlorhexidine group had a shorter ICU length of stay, shorter duration of mechanical ventilation, and shorter duration of catheterization. Although the group with chlorhexidine-impregnated dressings had reduced rates of CRBSIs, contamination, colonization, and local catheter infection, the differences between the 2 groups were not statistically significant. Our results might have clinical usefulness and may merit further evaluation. We recommend the use of 2% chlorhexidine-impregnated sterile dressings in catheter care in PICUs to prevent CRBSIs or reduce the rate of the infections.

This research was presented in part at the 10th International Congress of the World Federation of Critical Care Nurses and the 6th National Critical Care Nursing Congress, Antalya, Turkey, November 12–15, 2014. We gratefully thank all the nurses and physicians working in the PICU. We also thank David F. Chapman for his assistance with the language in this article.