Reference/Citation: Salgado CD, Farr BM, Calfee DP. Community-acquired methicillin-resistant Staphylococcus aureus: a meta-analysis of prevalence and risk factors. Clin Infect Dis. 2003;36:131–139..
Clinical Question: What are the prevalence rates and risk factors associated with community-acquired methicillin-resistant Staphylococcus aureus (MRSA)?
Data Sources: Studies were identified by searching MEDLINE (January 1966–February 2002) and abstracts from scientific meetings (1996–2001). Reviews of citations and reference lists were performed to identify additional eligible studies. The search terms included Staphylococcus aureus , infection, colonization, methicillin resistance, community-acquired, community-onset, prevalence, frequency, and risk factors.
Study Selection: The search was limited to English-language investigations identified from the electronic and manual searches. Studies were divided into 2 groups, as follows: group 1, retrospective or prospective studies that reported the prevalence of community-acquired MRSA (CA-MRSA) among hospital patients who were colonized (presence of bacteria without infection) or infected with MRSA; and group 2, studies that reported the prevalence of MRSA colonization in the community. The studies were evaluated independently by 2 authors, and case reports were excluded.
Data Extraction: Data extraction and study quality assessment procedures were not fully explained. The outcome measures for hospital patients were definitions of CA-MRSA used in the study, prevalence of CA-MRSA, sample size, number and type of risk factors assessed, and number of patients with ≥1 health care–associated risk factor. The studies were grouped based on type, retrospective or prospective. The pooled prevalence of CA-MRSA was calculated for each group (retrospective or prospective) and was limited to the prevalence among patients with MRSA. The proportion of patients who reported ≥1 health care–associated risk factor was also calculated. The outcome measures among community members were prevalence of MRSA, sample size, number and type of risk factors assessed, number of members with ≥1 risk factor, and MRSA strain type, when available. The studies were grouped based on the population surveyed (surveillance cultures, contacts with MRSA-colonized individuals, or sport team members or day care contacts). The pooled prevalence of MRSA colonization and the proportion of members with ≥1 reported risk factor were calculated for each of the study populations listed above. The proportion of CA-MRSA strains that represented typical nosocomial (infection that develops in the hospital) strains was also determined. Chi-square analysis was performed to compare proportions and to determine heterogeneity among the studies.
Main Results: Specific search criteria identified 104 studies for review, of which 57 met inclusion and exclusion criteria. Thirty-nine studies focused on CA-MRSA among hospital patients who were colonized or infected with MRSA. Of these, 32 groups (27 retrospective, 5 prospective) reported the prevalence of CA-MRSA using clinical specimens. Seven groups identified risk factors of CA-MRSA among patients previously diagnosed with MRSA. Thirteen different definitions of CA-MRSA were used in 31 of these studies, and 8 groups did not report the definitions used. The isolation of MRSA within 48 hours of hospital admission, with or without recent admission to a hospital or long-term care facility, or previous history of MRSA colonization were the most common definitions in the studies.
The risk factors included recent hospitalization (range, 1–24 months before identification of MRSA infection or colonization), recent outpatient visit (usually within 12 months), recent nursing home admission (usually within 12 months), recent antibiotic exposure (range, 1–12 months), chronic illness (eg, end-stage renal disease, diabetes, or malignancy), injection drug use, and close contact with a person who had risk factor(s) for MRSA acquisition. The presence of health care–associated risk factors was examined in 17 of the retrospective studies, and the median number of factors studied was 2 (range, 1–6). Among 4121 patients in these studies, 86.1% were found to have ≥1 health care–associated risk factor. All authors of prospective studies (5) examined health care–associated risk factors, and the median number of factors studied was 4 (range, 2–4). Among the 636 patients, 86.9% had ≥1 health care–associated risk factor. In the 7 studies with 515 patients previously diagnosed with MRSA, 84.7% had ≥1 health care–associated risk factor. The most common risk factors assessed in the 17 retrospective studies were recent hospitalization and chronic illness requiring health care visits.
The pooled CA-MRSA prevalence was 30.2% (range, 1.9%– 96%) among 5932 patients from the 27 retrospective studies and 37.3% (range, 18.2%–51.2%) among 636 patients from the 5 prospective studies. Eighteen groups reported the prevalence of MRSA colonization in the community. Ten of these reported MRSA prevalence using surveillance cultures, 4 examined colonization status of household contacts with discharged hospital patients with nosocomial MRSA colonization, and 4 reported colonization status of sports team members or day care contacts of persons colonized with MRSA. In the 10 surveillance studies, the pooled MRSA colonization prevalence was 1.3% (95% confidence interval [CI], 1.04%–1.53%; range, 0.2%– 7.4%) among 8350 community members. Nine of these studies were stratified based on culture samples taken before the assessment of risk factors, and among 4825 people, the pooled MRSA colonization prevalence was 2.1%. When examining health care–associated risk factors, the median number of factors studied was 5 (range, 1–10), and 47.5% with MRSA had ≥1 health care–associated risk factor. The risk factors included those previously identified. In the remaining surveillance study, the MRSA colonization prevalence was 0.20% among 3525 people without prior health care contact. Compared with subjects in the 9 stratified studies with a health care contact, subjects in this study were 90% less likely to have MRSA (relative risk, 0.10; 95% CI, 0.05–0.21). Cultures for 3898 subjects in 7 of the 10 surveillance studies were obtained at the time of a hospital admission, an outpatient clinic visit, or an emergency department visit, and the pooled prevalence of MRSA colonization was 1.8%. In 3 studies in which cultures were obtained outside of a health care facility (schools, day care centers, homeless shelters, or military bases), the pooled MRSA colonization prevalence among 4452 subjects was reported to be 0.76%. Therefore, subjects in a health care facility were 2.35 times more likely to carry MRSA than were subjects outside of a health care facility (95% CI, 1.56–3.53). In one study examining 94 subjects in a semiclosed community, the prevalence of MRSA colonization was 7.4%. These subjects were 36 times more likely to carry MRSA than were subjects who were not in a semiclosed community (95% CI, 13.7–94.7).
The studies also identified 70 MRSA isolates (pure form of an organism in a microbial culture) from subjects who reported no health care–associated risk factors. Strain typing was performed with 32 isolates, and 29 (91%) isolates were similar to strains identified in hospitals. The colonization status of 191 household contacts of 93 patients with nosocomial MRSA colonization discharged from the hospital was examined in 4 studies. The results demonstrated that 17.8% of the contact subjects were colonized with a strain of MRSA having the same antibiogram (record of the susceptibility of bacteria to antibiotics) as the index case (initial individual with the strain). The authors reported that subjects who had household contacts with MRSA-colonized patients were 14 times more likely to be colonized than were community subjects without a known MRSA contact (95% CI, 9.8–20.1). In 4 studies examining 517 sports team members or day care contacts of persons known to be colonized with MRSA, 5.4% demonstrated colonization of MRSA with the same strain as the index case.
Conclusions: Based on the available data, the prevalence of MRSA among community members without health care–associated risk factors was relatively low. However, 85% of hospital patients diagnosed with CA-MRSA and 47.5% of healthy community members colonized with MRSA were found to have ≥1 health care–associated risk factor. The risk factors identified were recent hospitalization, outpatient visit, nursing home admission, antibiotic exposure, chronic illness, injection drug use, and close contact with a person with risk factor(s). Most MRSA colonization occurred among community members who had health care–associated risk factors or contact with persons with risk factors. The evidence indicated that control of MRSA in the community may require control of MRSA in the health care setting (hospital, health care office, and nursing home). The absence of a standardized definition for CA-MRSA and questions regarding the actual site of colonization versus acquisition should be considered in the interpretation of these findings.