JSTOR

Materials and Methods

We performed a quasi‐experimental, pre‐post study in the 21‐bed MICU at Rush University Medical Center, a 720‐bed tertiary care teaching hospital in Chicago, Illinois. During the first part of our study (from September 2004 through October 2005 and hereafter referred to as the “soap‐and‐water period”), patients were bathed daily with soap and water. During the second part of our study (from November 2005 through October 2006 and hereafter referred to as the “CHG period”), daily cleaning with no‐rinse, 2% CHG‐impregnated cloths (Sage Products) was substituted for soap‐and‐water bathing. No other infection prevention interventions were introduced during these 2 time periods. Chlorhexidine‐impregnated sponge dressings (Biopatch; Johnson & Johnson, Ethicon 360) were not used, there were no changes in devices or catheter types used, and there were no policy changes made for catheter insertion location. The subsequent introduction of other measures, such as ultrasound‐guided catheter insertion, that have been shown to reduce rates of CVC‐associated BSI prevented us from analyzing the effect of CHG bathing after October 2006. CHG was used for skin disinfection before CVC insertion during both the soap‐and‐water and CHG periods.

Nurses and patient care technicians in the MICU were instructed in the manufacturer's CHG cleansing procedure before the introduction of CHG; details of this procedure have been published elsewhere.^10,11 Products that may interact with CHG and decrease its activity were removed from MICU admission packets and from nursing stockrooms.¹² Bathing was performed daily; additional cleansing with CHG throughout the day was allowed if the patient’s primary nurse felt it was needed. There were no changes in the blood culture method or in phlebotomy procedures during the study period. The study was approved by the institutional review board of Rush University Medical Center; informed consent was waived.

The primary outcome was MICU‐acquired CVC‐associated BSI. Secondary outcomes included blood culture contamination, other MICU‐acquired nosocomial infections (ie, Clostridium difficile infection [CDI], secondary BSI, ventilator‐associated pneumonia [VAP], and urinary tract infection [UTI]), and clinical cultures that grew selected drug‐resistant bacteria (ie, methicillin‐resistant Staphylococcus aureus [MRSA], vancomycin‐resistant enterococci [VRE], and imipenem‐resistant Acinetobacter baumannii). Because our previous study¹¹ observed that the rates of CDI, secondary BSI, VAP, and UTI were unchanged after the use of CHG bathing, these infection rates served as control variables.

Data on patient demographics, number of MICU admissions, and lengths of MICU stay were extracted from hospital administrative databases. Nosocomial infection rates were determined by reviews of infection control practitioner logs and patient medical charts. Clinical culture results were obtained from the microbiologic database of the laboratory information system. ICU‐acquired infections were defined as recommended by the Centers for Disease Control and Prevention.¹³ A blood culture was categorized as contaminated if any of the following organisms were identified in a single blood culture specimen: coagulase‐negative Staphylococcus species, Propionibacterium acnes, Micrococcus species, “viridans”‐group streptococci, Corynebacterium species, or Bacillus species,¹⁴ unless a physician treated the culture result as a pathogen and not a contaminant.

Infection rates were compared using Poisson regression, and continuous variables were evaluated using the t test. Infection rates were calculated using 1,000 CVC‐days, 1,000 patient‐days, or 1,000 ventilator‐days as the denominators. All statistical analyses were done using SAS, version 9.1 (SAS).

Results

From the soap‐and‐water period to the CHG period, there were no significant changes in the average age of patients (59.5 vs 59.3 years; ), average length of stay (3.21 vs 3.01 days; ), or monthly number of admissions to the MICU (155 vs 163 patients; ). The CVC‐associated BSI rate decreased significantly from the soap‐and‐water period to the CHG period (5.31 vs 0.69 cases per 1,000 CVC‐days; relative risk, 0.13 [95% confidence interval, 0.03–0.56]; ) (Table 1 and Figure). There were 19 cases of CVC‐associated BSI during the soap‐and‐water period and 2 cases during the CHG period. The pathogens associated with these 19 cases of CVC‐associated BSI during the soap‐and‐water period were primarily enterococci (6 isolates were recovered), gram‐negative rods (6 isolates), S. aureus (3 isolates), and coagulase‐negative staphylococci (3 isolates). During the CHG period, the 2 cases of CVC‐associated BSI were caused by Enterococcus faecium and Pseudomonas aeruginosa, respectively (Table 2).

Table 1.  Comparison of Nosocomial Infection Rates in the Medical Intensive Care Unit during 2 Study Periods

Open New Window

(27 KB)

Figure.  Rates of central venous catheter (CVC)–associated bloodstream infection (BSI) in the medical intensive care unit during the period when patients were bathed in soap and water (September 2004–October 2005) and during the period when patients were bathed daily with 2% chlorhexidine gluconate (CHG)–impregnated cloths (November 2005–October 2006). The arrow indicates the period of CHG bathing. Data points are missing for months in which nosocomial infection surveillance was not performed.

Open New Window

Table 2.  Pathogens Isolated from Cases of Central Venous Catheter–Associated Bloodstream Infection in the Medical Intensive Care Unit During 2 Study Periods

Open New Window

For secondary outcomes, the blood culture contamination rate declined significantly from the soap‐and‐water period to the CHG period (6.99 vs. 4.1 cases per 1,000 patient‐days; relative risk, 0.59 [95% confidence interval, 0.36–0.97]; ). Secondary BSI, CDI, VAP, and UTI rates did not change from the soap‐and‐water period to the CHG period (Table 1). During the soap‐and‐water period, 3 cases of secondary BSI were attributed to UTIs. Of the 4 cases of secondary BSI during the CHG period, 1 case was attributed to a UTI, 1 case to a surgical site infection, and 2 cases to VAP. The number of clinical cultures that grew imipenem‐resistant A. baumannii, MRSA, and VRE decreased from the soap‐and‐water period to the CHG period, but the decrease was not statistically significant (Table 1). There were 2 cases of CVC‐associated MRSA BSI during the soap‐and‐water period and no cases of CVC‐associated MRSA BSI during the CHG period (Table 2).

Discussion

Routine daily cleansing of MICU patients with CHG‐impregnated cloths reduced the rate of CVC‐associated BSI to zero during the last 6 months of the study. The findings of our real‐world trial are concordant with the results of a controlled study¹¹ of CHG skin cleansing that we performed previously in the MICU of an affiliated institution where we demonstrated a reduction in the rate of CVC‐associated BSI similar to that seen in other ICUs where bundled interventions were introduced.⁵ Also consistent with results of our previous study¹¹ was a significant decrease in the blood culture contamination rate. Control variables—infection rates that were shown to be unaffected by CHG bathing in our previous study—were unchanged from the soap‐and‐water period to the CHG period, supporting the causal effect of CHG bathing on reducing the rates of CVC‐associated BSI and blood culture contamination. The microorganisms that were responsible for the CVC‐associated BSIs and that were eliminated by CHG included predominantly pathogenic bacteria (ie, Enterococcus species and gram‐negative rods); only 3 of 19 cases of CVC‐associated BSI were attributable to coagulase‐negative staphylococci. This finding supports the clinical relevance of the intervention. The study by Bleasdale et al¹¹ reported a higher rate of CVC‐associated BSIs caused by coagulase‐negative staphylococci. We hypothesize that the patient population in the present study setting, a tertiary care center with an active solid‐organ transplant service and a significant proportion of patients receiving chemotherapy and bone marrow transplants, compared with the patient population in the study by Bleasdale et al,¹¹ accounts for the different pathogens causing CVC‐associated BSIs in each study. In addition, surveillance for CVC‐associated BSIs was performed by infection control practitioners in our present study but by research personnel in the study by Bleasdale et al.¹¹ Differences in the application of definitions of blood culture contamination¹⁵ may be partly responsible for lower rates of BSI due to coagulase‐negative staphylococci in the present study.

Cleansing with CHG, which has activity against a broad range of bacterial and fungal pathogens,¹⁶ results in the maintenance of a low level of microbial colonization on patients’ skin.^10,17 This depletes a primary source of direct CVC contamination and contamination of healthcare workers' hands, which can result in the transfer of pathogens to a catheter or its hub. Furthermore, daily bathing with CHG‐impregnated cloths is well tolerated by patients.¹⁰ So far, we have examined the effect of CHG cleansing on the rates of CVC‐associated BSI and blood culture contamination in the MICUs of 2 hospitals during 3 different time periods^11,18 and therefore believe that our results are likely generalizable to other MICUs. However, we do not know whether our results are applicable to ICUs with different patient populations in which other clinical factors may influence BSI rates. In a parallel study,¹⁹ we were unable to demonstrate a reduction in the rate of CVC‐associated BSI among surgical ICU patients, for whom the beneficial reduction in pathogenic skin flora achieved by CHG bathing may be trumped by the potential for large contaminated surgical wounds to act as a source of translocation of bacteria in a patient's bloodstream or for contamination of healthcare workers' hands.

Our study has other limitations. It was an analysis of a real‐world intervention, rather than a randomized trial, and we were limited to a pre‐post study design with inclusion of control variables.²⁰ We do not have data on catheter insertion sites; however, there were no changes in infection prevention policy during the 2 study periods. Furthermore, although no new infection prevention measures were introduced in the MICU during our study, we cannot be certain that provider‐level changes in infection control practice might not have occurred. The finding that rates of secondary outcome (ie, control variables), such as CDI, did not change significantly between the soap‐and‐water period and the CHG period supports our conclusion that compliance with general infection control measures such as hand hygiene remained stable.

CHG has been used for many infection control applications, including oropharyngeal antisepsis to prevent VAP, presurgical skin preparation, impregnation of devices and dressings, and healthcare worker hand disinfection.²¹ Controlled studies in MICUs have demonstrated CHG’s efficacy in reducing rates of colonization of patients’ skin with potential pathogens and in reducing rates of CVC‐associated BSI.^10,11,18 The present study, a real‐world evaluation of daily skin cleansing with CHG‐impregnated cloths, extends the work of these previous trials by demonstrating a significant reduction in cases of CVC‐associated BSI among MICU patients. Controlled trials are needed to determine whether the beneficial effect of CHG skin cleansing extends to populations outside of the MICU.