Objective: To estimate the prevalence and significance of nasal methicillin-resistant Staphylococcus aureus colonization in the ICU and its predictive value for development of methicillin-resistant S. aureus infection.
Data Sources: MEDLINE and EMBASE and reference lists of all eligible articles.
Study Selection: Studies providing raw data on nasal methicillin-resistant S. aureus colonization at ICU admission, published up to February 2013. Analyses were restricted in the general ICU setting. Medical, surgical, and interdisciplinary ICUs were eligible. ICU studies referring solely on highly specialized ICUs populations and reports on methicillin-resistant S. aureus outbreaks were excluded.
Data Extraction: Two authors independently assessed study eligibility and extrapolated data in a blinded fashion. The two outcomes of interest were the prevalence estimate of methicillin-resistant S. aureus nasal colonization at admission in the ICU and the sensitivity/specificity of colonization in predicting methicillin-resistant S. aureus–associated infections.
Data Synthesis: Meta-analysis, using a random-effect model, and meta-regression were performed. Pooled data extracted from 63,740 evaluable ICU patients provided an estimated prevalence of methicillin-resistant S. aureus nasal colonization at admission of 7.0% (95% CI, 5.8–8.3). Prevalence was higher for North American studies (8.9%; 95% CI, 7.1–10.7) and for patients screened using polymerase chain reaction (14.0%; 95% CI, 9.6–19). A significant per year increase in methicillin-resistant S. aureus colonization was also noted. In 17,738 evaluable patients, methicillin-resistant S. aureus infections (4.1%; 95% CI, 2.0–6.8) developed in 589 patients. The relative risk for colonized patients was 8.33 (95% CI, 3.61–19.20). Methicillin-resistant S. aureus nasal carriage had a high specificity (0.96; 95% CI, 0.90–0.98) but low sensitivity (0.32; 95% CI, 0.20–0.48) to predict methicillin-resistant S. aureus–associated infections, with corresponding positive and negative predictive values at 0.25 (95% CI, 0.11–0.39) and 0.97 (95% CI, 0.83–1.00), respectively.
Conclusions: Among ICU patients, 5.8–8.3% of patients are colonized by methicillin-resistant S. aureus at admission, with a significant upward trend. Methicillin-resistant S. aureus colonization is associated with a more than eight-fold increase in the risk of associated infections during ICU stay, and methicillin-resistant S. aureus infection develops in one fourth of patients who are colonized with methicillin-resistant S. aureus at admission to the ICU.