Clinical effectiveness of mupirocin for preventing S. aureus infections in non-surgical settings: A Meta-analysis
Objective: To identify the optimal setting and patient population to implement mupirocin decolonization for prevention of S. aureus infections using meta-analytic methods.
Methods: We conducted systematic searches in PubMed, Cochrane Library Databases, Scopus, Web of Science, and ClinicalTrials.gov to identify papers published until 2013 on effectiveness of mupirocin in healthcare settings. Studies were included if they provided data on incidence of S. aureus infections. Two investigators independently assessed eligibility of studies and abstracted data with a pilot-tested form. Risk of bias was assessed using the Cochrane tool. The crude odds ratios were pooled (cpOR) using a random-effects model. Heterogeneity was evaluated using the Woolf’s test for homogeneity and I2 statistics.
Results: Of the 12,644 studies identified, 8 randomized controlled trials and 19 quasi-experimental studies met the study inclusion criteria. Mupirocin was observed to reduce the odds of S. aureus infections by 70% (cpOR=0.30, 95%CI 0.23, 0.39) and 60% (cpOR=0.40, 95%CI 0.27, 0.62) in both dialysis and non-dialysis settings, respectively. Nevertheless, there was highly significant (p=0.0009) and moderate heterogeneity (I2=46%) among studies. Studies were homogeneous (p>0.1) when stratified analyses were performed by specific clinical settings. Among the 6 studies that took place in adult intensive care units (ICUs), mupirocin decolonization was associated with a 56% reduction in the odds of S. aureus infection (cpOR=0.44, 95%CI 0.26, 0.73). There was also a protective effect of mupirocin against S. aureus exit site infections among patients undergoing peritoneal dialysis (cpOR=0.23, 95%CI 0.15, 0.36) and against bacteremia among hemodialysis patients (cpOR=0.15, 95%CI 0.06, 0.36).
Conclusion: Mupirocin decolonization is protective against S. aureus infections among both dialysis and adult ICU patient populations. Future studies should target other patient settings such as long-term care facilities.
A. Blevins, None
M. Schweizer, None