Literature Update - August, 2007
Compiled by Dr Leon Worth, ID Physician
Value of Whole-body Washing with Chlorhexidine for MRSA Colonisation
Wendt C et al (Infect Control Hosp Epidemiol. 2007; 28:1036-1043) assessed the effectiveness of whole-body chlorhexidine washes for
eradication of MRSA colonisation in general hospital and nursing home patients. 4% chlorhexidine was used daily for 5 days, and compared
with placebo. The study population was co-administered intranasal mupirocin ointment and was subject to environmental cleaning.
482 patients were excluded, and the majority of these (215) were discharged or were outpatients who were not referred to the study.
In total, 114 participants with MRSA colonisation were studied (56 in treatment arm). At day 30 following the treatment, the rate of
MRSA eradication was 7% in the treatment group and 13% in the placebo group (p=0.47). One patient from each treatment group developed
MRSA infection during follow-up.
Adverse effects were significantly more common in the treatment group (skin fissures, itching, burning of skin), leading to 4 patients
in the treatment arm discontinuing the daily regime. Failure to eradicate MRSA was associated with groin or perineal colonisation at
baseline, and >1 positive site of colonisation at baseline.
Compared to previous reports of efficacy of chlorhexidine body-washes, which have been predominantly of ICU patients, this study focuses
upon a general patient population. With colonisation and infection rates not significantly different in the treatment and placebo groups,
and the possible difficulty in completing the study protocol (a large number of patients were excluded because of hospital discharge) with
a high rate of adverse effects (71%), the utility and efficacy of body washes with 4% chlorhexidine is not supported by this study. Remaining
questions include the impact of this intervention upon colonisation with other multi-resistant organisms (e.g. VRE), and the efficacy of this
intervention in high-risk populations.
Automated Strategies for Surveillance of Nosocomial Bacteremia
The CDC recommends surveillance for nosocomial BSI, but traditional methods based on manual review of clinical and microbiological data are
resource-intensive. Bellini C et al (Infect Control hosp Epidemiol. 2007; 28:1030-1035) report a comparison of different strategies for
automated surveillance of nosocomial BSI. The study was conducted at a single European hospital over a 3 year period. An algorithm was
developed to distinguish contamination from true bacteremia, new episodes of bacteremia from duplication, community-acquisition from
nosocomial events, and catheter-related from other sources of BSI. Microbiological data and administrative data were used as input to the
algorithm. For 100 randomly selected episodes of positive blood culture, manual collection of data from medical records with review by an
infectious diseases specialist or study investigator was used as ‘gold standard’ for comparison.
The studied population (669 episodes of positive blood culture) included medical, surgical and intensive care units. The automated
classification of hospital-acquired, community-acquired or culture contamination agreed with manual review in 95% cases. This was
consistent with a sensitivity of 96% and specificity of 94% for detection of hospital-acquired BSI; sensitivity and specificity of 100%
for diagnosis of community-acquired BSI; sensitivity of 96% and specificity of 97% for diagnosis of contaminated blood cultures; and,
sensitivity of 68% and specificity of 94% for diagnosis of catheter-related BSI.
Two components of the proposed algorithm were identified as amenable to improvement (taken together, accounted for 81% of discrepancies
between automated and manual review): (i) extending the definition of an episode of Candida fungemia to the entire duration of hospital
stay, rather than 72 hours, (ii) all episodes of coagulase-negative staphylococcal bacteremia were attributed to a catheter-related infection.
By incorporating these modifications into the algorithm, sensitivity for detection of catheter-related BSI was increased (to 78%),
with specificity of 93%.
These results suggest that if linkage of appropriate microbiological and administrative data can be performed, community-acquired and
nosocomial-acquired BSI may be distinguished. The classification of events as catheter-associated or ‘other’ is less reliable, and this
is consistent with previous studies which have frequently relied upon combined clinical and automated data to achieve reliable classification.
It may be possible to restrict an automated strategy to particular organisms (e.g. remove coagulase-negative staphylococci from the strategy),
in order to improve sensitivity. However, such an approach to analysis was not reported as part of this study.