Fecal Carriage of Staphylococcus aureus in the Hospital and Community Setting: A Systematic Review

Background and rationale: Staphylococcus aureus fecal carriage has been identified as a potential source for nosocomial transmission and a risk factor for disease development. This systematic review determined the overall S. aureus [including methicillin susceptible and resistant S. aureus (MSSA and MRSA)] fecal carriage rates within the community and healthcare settings. Methodology: Peer-reviewed articles indexed in Medline, Scopus, Academic Search Premier, Africa-Wide Information, CINAHL, and Web of Science were identified using applicable and controlled vocabulary through to 11 November 2015. Eligible studies were ascertained by three independent reviewers. Random-effects meta-analyses of proportions were performed to determine S. aureus, MSSA and MRSA fecal carriage rates reported by eligible studies. Results: Twenty six studies were included in this review. The pooled estimates for S. aureus, MSSA and MRSA fecal carriage were 26% (95% confidence interval (CI): 16.8–36.3%), 86% (95% confidence interval (CI): 65.9–97.9%) and 10% (95% CI: 0.7–27.0%), respectively. Fecal S. aureus carriage rates increased on average from 10 to 65% during the first 8 weeks of life, followed by an average carriage rate of 64% at 6 months and 46% at 1 year of life. Genotyping techniques were employed mainly in studies conducted in developed countries and comprised largely of gel-based techniques. Six studies reported on the role of S. aureus fecal strains in diarrhea (n = 2) and the risk for acquiring infections (n = 4). Eight of the 26 studies included in this review performed antibiotic susceptibility testing of S. aureus fecal isolates. Conclusion: This study provides evidence that screening for S. aureus fecal carriage, at least in populations at high risk, could be an effective measure for the prevention of S. aureus transmission and infection in the healthcare and community setting. More well-structured studies need to be conducted and sequence-based genotyping techniques should be employed for the comparison of isolates on a global scale in both developing and developed countries.

The importance of fecal carriage of S. aureus has been recognized more than five decades ago in a study which demonstrated that rectal S. aureus carriage preceded those from the nose and throat in new-borns (Hurst, 1960). Thereafter, several studies have provided evidence on the clinical importance of fecal carriage of S. aureus [in particular methicillin-resistant S. aureus (MRSA)] in the hospital setting (Acton et al., 2009). For example, it has been shown that hospitalized patients with both S. aureus fecal and nasal colonization are significantly more likely to have positive skin cultures compared to patients with nasal carriage only (Bhalla et al., 2007). In addition, S. aureus fecal carriage may serve as an important source for environmental contamination, which can potentially facilitate nosocomial transmission within the healthcare setting (Bhalla et al., 2007). Furthermore, antibioticassociated diarrhea attributed to MRSA has also been reported (Lo and Borchardt, 2009;Sizemore et al., 2012;Avery et al., 2015); and patients with MRSA colonized diarrheal stools impact significantly on environmental contamination (Boyce et al., 2007).
Despite the potential role and significance of the sole fecal carriage of S. aureus (Lee et al., 1997;Squier et al., 2002;Bhalla et al., 2007) and the transmission dynamics of S. aureus in infection, a limited number of studies have focused on fecal S. aureus carriage in the hospital and community setting (Acton et al., 2009). This systematic literature review is therefore aimed to determine the overall rate of S. aureus [including methicillin susceptible and resistant S. aureus (MSSA and MRSA)] fecal carriage amongst individuals in the community and healthcare settings.

METHODOLOGY
This review followed the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines (Moher et al., 2009). The PRISMA check-list for this review is provided in a Supplementary Table (Table S1).

Literature Search Strategy
Peer-reviewed articles (written in English and French) published through to 11 November 2015 on S. aureus fecal carriage within the community and healthcare settings were evaluated using four electronic databases and a combination of keywords ( Table 1). We also explored for additional articles by checking the references cited in the primary eligible studies included in this systematic review.

Study Selection and Data Extraction
Potentially relevant articles (selected based on their titles and abstracts) were assessed for eligibility ( Table 2) by three independent authors. All potentially eligible articles were screened for "predatory journals" using "Beall's list" (Beall, 2015;Shen and Björk, 2015;Siebert et al., 2015). The corresponding authors of potentially relevant articles were contacted to determine the healthcare exposure status of participants so as to assess their eligibility for inclusion in this systematic review (   (Folden et al., 2005;Millar et al., 2007;Otter and French, 2011). • Studies published in either English or French.
• Studies screening for S. aureus or MRSA from samples other than feces/rectal swabs/anal swabs. • Fecal samples studied for parasites or bacteria other than S. aureus.
• Articles reporting on the number of S. aureus or MRSA isolates detected from fecal specimens or on the number of fecal specimens positive for S. aureus or MRSA, but not providing information on the number of participants testing positive for S. aureus or MRSA fecal carriage. • Studies not providing the necessary healthcare exposure data for participants (via the published article or via correspondence with the authors), in order to categorize participants into Healthy participants, Out-patients, In-patients and Healthcare personnel. • Articles published in predatory journals (Beall, 2015).
• Articles not obtainable from the electronic databases, the University of Cape Town (UCT) library or the UCT inter-library loans.
Inclusion criteria for meta-analysis of proportions Exclusion criteria for meta-analysis of proportions • Overall fecal carriage prevalence for S. aureus and/or MRSA must be available.
• Studies providing fecal carriage rates for participants for which fecal carriage rates have previously been reported. • Studies not providing information on the age at which participants were screened. • Studies screening a pre-selected group of participants based on microbiological assessments. • Studies for which MRSA was not confirmed using molecular methods.
• Pregnant women visiting obstetric clinics; • New-borns and mothers at maternity wards during the time of delivery; • Mothers and infants reported as healthy at the time of screening for S. aureus or MRSA fecal carriage, but exposed to the delivery unit or maternity ward during the year preceding screening.

Out-patients
Patients screened for S. aureus or MRSA fecal carriage with ≤48 h of healthcare contact (Folden et al., 2005;Millar et al., 2007;Otter and French, 2011). Patients should not have had contact with healthcare settings in the year preceding the study.

In-patients
Patients screened for S. aureus or MRSA fecal carriage with >48 h of healthcare contact. Patients screened within ≤48 h after admission should be those transferred from another hospital/ward which will allow for >48 h of hospital contact.

Healthcare personnel
Participants screened for S. aureus or MRSA fecal carriage working at a healthcare setting with or without any illness.

Developed and Developing Countries
Countries were categorized as developed or developing countries based on data from the International Monetary Fund (http:// www.imf.org/external/pubs/ft/weo/2015/01/weodata/groups. htm).

Antibiotic Susceptibility Results
The percentage of isolates (obtained from participants with S. aureus or MRSA fecal carriage) resistant to each of the antibiotics assayed was calculated from studies that provided adequate data on antibiotic susceptibility test results.  StatsDirect Ltd. 2016] was also applied to assess the heterogeneity between the studies included in the meta-analyses (Cochran Qtest) (Cochran, 1954) and to determine the inconsistency across the studies included (I 2 statistic) (Higgins et al., 2003). The criterion for statistical significance for the test for heterogeneity was set at alpha = 0.05. The risk of publication bias was assessed and visualized by a Funnel plot (Egger et al., 1997;Sterne et al., 2011).

Study Selection and Characteristics
S. aureus Study Selection Figure 1 outlines the study selection process and the broad reasons for exclusion. The search strategy identified 2522 records. An additional record was identified from the reference list of one of the eligible articles included in the review. A total of 124 potentially eligible reports were identified, of which 69 fulfilled the primary inclusion criteria (Figure 1). The vast majority (80%; 55/69) of these potentially eligible articles did not provide information on healthcare exposure during the year preceding screening and/or did not indicate the duration for which patients were admitted prior to the time of screening. Following correspondence with authors, seven articles were excluded as these reports did not fulfill our inclusion criteria. Moreover, 36 articles were excluded due to lack of required information from corresponding authors or as a result of unavailable author contact information. Consequently, only 26 (11 and 15 reports based on their full texts and information obtained from the authors, respectively) of the 69 studies could be included in our systematic review. The main findings reported by each of the 26 eligible studies are summarized in detail in Tables 3, 4. Select studies that screened for S. aureus fecal carriage from both community and healthcare settings are also reported accordingly in Tables 3, 4.

Reports on S. aureus fecal carriage
A total of 19 reports investigated fecal S. aureus carriage within the community setting, of which five and 14 studies reported on fecal carriage rates from outpatients and healthy participants, respectively (Table 3). Moreover, the majority (64%; 9/14) of reports on fecal S. aureus carriage rates from healthy participants were of longitudinal design and investigated infants up until one year of age ( Table 3). Of the five reports on fecal carriage rates from outpatients, a single study performed a longitudinal analysis of S. aureus fecal carriage (Efuntoye and Adetosoye, 2003) and another investigated infants during the first year of life (Shehabi et al., 2013). Study sizes for the community setting ranged between 21 and 1761 participants ( Table 3).
Fecal S. aureus carriage within the healthcare setting was noted in 12 reports ( Table 4). Of these, 10 were from inpatients and two from healthcare personnel. All reports on inpatients were of cross-sectional design and the majority (60%; 6/10) did not provide information on the age of the participants. In addition, the two studies on healthcare personnel were crosssectional in design and carried out in the United States of America (USA) (Carmeli et al., 1998;Andrews et al., 2009). Study sizes for healthcare-based reports ranged between 37 and 2727 participants ( Table 4).

Reports on methicillin susceptible and resistant S. aureus fecal carriage
Six of the 19 reports on S. aureus fecal carriage from the community setting provided MRSA fecal carriage rates confirmed by molecular methods ( Table 3). Five of these studies (conducted in developed countries) reported both S. aureus and MRSA fecal carriage rates which allowed for the calculation of MSSA fecal carriage rates. Only one study within the healthcare setting (conducted in the USA) confirmed fecal MRSA carriage by screening specimens using a molecular approach (Andrews et al., 2009).

Pooled Estimates of S. aureus Fecal Carriage Rates Assessed by Meta-Analyses
Studies included in all of the proportional meta-analyses were heterogeneous, as determined by the Cochrane Q test and I 2 statistic (Figures 2-4). We could not determine pooled MSSA or MRSA fecal carriage rates within the healthcare setting as only a single study was considered eligible for this analysis.

S. aureus Fecal Carriage Rates According to the Age of Participants
The report on this section is not based on meta-analysis. S. aureus fecal carriage rates within the community setting were higher during the first year of life (Figure 8). On average, reports from longitudinal studies revealed an increase in S. aureus fecal carriage rates from approximately 10-65% during the first 8 weeks of life (Figure 8). At 6 months of age, the average fecal carriage rate was 64%, thereafter it decreased to approximately 46% at 1 year of life. A longitudinal investigation of fecal MRSA carriage rates from healthy participants from the USA showed an increase in fecal MRSA carriage from 0 to 9% during the first 2 weeks of life (Gries et al., 2009). The highest MRSA fecal carriage rate (23%) reported was from Spanish infants screened at ≤1 year of life .

Assessment of Antibiotic Susceptibility of Fecal S. aureus Isolates
Eight of the 26 eligible studies (31%) included in this review assayed for antibiotic susceptibility of fecal S. aureus or MRSA isolates (Table 5). Overall, S. aureus or MRSA isolates were screened with 32 different antibiotics across the respective studies using disk diffusion, agar dilution, or the Vitek Legacy System. The use of published guidelines for susceptibility testing were reported by six of the eight studies (Table 5). Susceptibility testing to erythromycin was performed most frequently (88%; 7/8), followed by chloramphenicol, clindamycin, ciprofloxacin, gentamicin, penicillin and vancomycin (75%; 6/8) ( Table 5).
Vancomycin intermediate or resistant S. aureus (VISA/VRSA) were not identified in five of the six studies that screened for vancomycin resistance (Table 5). Only the study by Onanuga and Temedie (2011) reported fecal VRSA carriage of 37% (14/38).

Genotyping of S. aureus Isolated from Fecal Specimens
Techniques used to genotype S. aureus isolated from fecal specimens included multiple-locus variable-number tandem repeat analysis (MLVA), pulsed-field gel electrophoresis (PFGE), random amplified polymorphic DNA (RAPD) analysis, staphylococcal cassette chromosome mec (SCCmec), accessory gene regulator (agr) and Staphylococcus aureus protein A (spa) typing (Tables 3, 4). Genotyping was performed in slightly more reports from the healthcare setting (67%; 8/12) compared to the community (58%; 11/19). Gel-based methods (PFGE, RAPD and MLVA) were employed in 58% (7/12) and 42% (8/19) of studies in the healthcare and community settings, respectively. In addition, similar rates (26% vs. 25%) in the use of sequence-based methods (spa typing, SCCmec typing and MLST) for genotyping of S. aureus strains were reported from community and healthcare settings. Only a single study conducted in a developing country (Jordan) performed genotyping of the S. aureus strains (Shehabi et al., 2013).

S. aureus and MRSA Fecal Carriage as Risk Factors for Disease Development
Two studies included in this review identified enterotoxin producing S. aureus strains from fecal specimens of patients with diarrhea (Efuntoye and Adetosoye, 2003;Flemming and Ackermann, 2007). Another study reported that all patients colonized with MRSA in both the nares and rectum (8/8) developed an infection (Srinivasan et al., 2010). In addition, two of the nine patients, colonized with MRSA in the rectum only, were concurrently or subsequently infected. Spa typing on a  subset of colonizing isolates from the nares and rectum noted that the majority (69%; 9/13) were clonally related to infecting isolates (Srinivasan et al., 2010). In support of the potential of fecal carriage for infection, it has also been shown that S. aureus detection occurs more frequently from rectal specimens of children with skin and soft tissue abscesses (47%; 28/60) compared with the control group (1%; 1/90) (P = 0.0001) (Faden et al., 2010).

DISCUSSION
Our results clearly showed that fecal S. aureus carriage from healthy infants is high during the first year of life. Specifically, S. aureus fecal carriage rates increased during the first 8 weeks of life followed by a gradual decrease towards 1 year of life. The reasons for this abrupt increase in fecal carriage very early in life (especially from healthy infants) is not yet clear, however a potential explanation may be early life care-giving practices, particularly breastfeeding. For example, colostrum contains the highest levels of human milk oligosaccharides (HMOs) (Bode, 2012), which have been suggested to stimulate S. aureus growth (Hunt et al., 2012). Moreover, S. aureus strains may be transmitted from parents via skin contact (Lindberg et al., 2004a) or from the mother via breastfeeding (Kawada et al., 2003;Lindberg et al., 2004a;Benito et al., 2015). Furthermore, staphylococci from the maternal GIT or skin surrounding the areola may be transferred to breast milk during lactation (Thum et al., 2012;Fernández et al., 2013). Higher S. aureus fecal carriage rates have also been noted from breast-fed in comparison to formula-fed or mixed-fed infants (González et al., 2013;Salminen et al., 2015). The observed change in the dynamics of S. aureus fecal carriage after 8 weeks of life may be explained by the increase in anaerobic bacteria from around 1 week of life (Bezirtzoglou, 1997;Adlerberth et al., 2006;Adlerberth and Wold, 2009;Jost et al., 2012), as well as the introduction of formula feeding (González et al., 2013) and solid foods (Bergström et al., 2014;Voreades et al., 2014). Infant fecal bacterial profiles have also been shown to change during the course of the lactation period (Cabrera-Rubio et al., 2012;González et al., 2013). This systematic review does not only provide insight into the dynamics of fecal S. aureus carriage rates during the first year of life; but also highlights that S. aureus and MRSA fecal carriage is a potential risk factor for subsequent infections. Vancomycin is  Frontiers in Microbiology | www.frontiersin.org regarded as one of the drugs of choice for MRSA infections (Tarai et al., 2013); however the emergence of vancomycin resistant S. aureus (VRSA) poses yet another threat to infection control (Hiramatsu, 1998;Spagnolo et al., 2014). The intestinal tract, in particular, may be a key potential reservoir for the emergence and transmission of VRSA isolates due to the intestinal coexistence (Ray et al., 2003), and potential transfer of the vanA gene from VRE to MRSA (Courvalin, 2006). Although, 23% of the studies included in this review screened for fecal carriage of VRSA within community and healthcare settings (Domínguez et al., 2002;Lindberg et al., 2004b;Srinivasan et al., 2010;Onanuga and Temedie, 2011;Benito et al., 2013Benito et al., , 2015; only a single study, performed in Nigeria, reported VRSA fecal carriage (Onanuga and Temedie, 2011). It is noteworthy, however, that this finding should be interpreted with caution as the disk diffusion method was used to screen for vancomycin resistance at 30 µg/ml, which is not recommended by the CLSI guidelines (Clinical Laboratory Standards Institute, 2012). Healthcare associated fecal screening for S. aureus and MRSA is of key importance in infection control (Campillo et al., 2001;Ray et al., 2003;Bhalla et al., 2007). For example, it has been shown that select staphylococcal enterotoxins (SEs) may contribute to the colonizing success of S. aureus strains in the GIT (Nowrouzian et al., 2011), which could potentially facilitate in its transmission. Moreover, S. aureus and MRSA fecal carriage may complicate de-colonization, with a potential to contribute to infections within the healthcare setting (Campillo et al., 2001;Dupeyron et al., 2002;Ray et al., 2003;Srinivasan et al., 2010). To prevent nosocomial transmission and infection, two recent studies (Roth et al., 2016;Senn et al., 2016) have also highlighted the importance of screening for S. aureus fecal carriage on admission in the following risk groups: patients admitted to surgery or intensive care units with a history of MRSA colonization or infection; hospitalization during the past year; or direct transfer from another healthcare facility. Only a single study was considered eligible for inclusion in our metaanalyses of the proportions on MSSA and MRSA fecal carriage within the healthcare setting. Therefore we could not determine the fecal carriage rate for MSSA or MRSA within this setting.
A major limitation in this systematic review is the poor study design and limited data available from studies assessing the fecal carriage rates of S. aureus and MRSA. For example, a large proportion of potentially eligible articles were excluded due to the lack of information regarding participants' contact with healthcare facilities as well as the duration of hospital admission prior to S. aureus and MRSA screening. This information is essential in comparing fecal carriage rates from community and healthcare settings. Furthermore, a number of studies could not be included in calculating the pooled estimates for MSSA and MRSA fecal carriage (from both community and healthcare settings) due to the lack of molecular techniques incorporated to confirm MRSA carriage. On the other hand, the extent in which our observations could have changed if unavailable articles were included is unclear. However, based on the rigorous appraisal of various studies in this systematic review, we conclude that the excluded articles are not likely to impact significantly on observations presented in the manuscript. In addition, more studies from both developed and developing countries are needed in order to determine S. aureus and MRSA fecal carriage and transmission within and between the community and healthcare settings. In support of this, rural areas and low socioeconomic status have been shown to contribute to higher fecal transmission rates of S. aureus and MRSA (Vale and Vítor, 2010). Finally, there is the need for more sequence-based genotyping data on S. aureus and MRSA fecal carriage as the majority of studies from developed countries made use of gel-based methods which are not ideal when comparing isolates on a global level.

CONCLUSION
S. aureus, MSSA and MRSA fecal carriage rates within both the community and healthcare setting are not negligible and estimated at 26, 86, and 10%, respectively. Therefore, preventative strategies which include fecal S. aureus screening of high risk patients are necessary for infection control within these settings. More studies are needed to determine the role of fecal S. aureus carriage as a risk factor for disease development; as well as fecal carriage rates of MSSA, MRSA, and VRSA from both community and healthcare settings. Furthermore, wellstructured research should be conducted and sequence-based genotyping techniques should be employed. The latter will allow for comparison of isolates on a global level in both developing and developed countries.