Staphylococcus aureus and CA-MRSA Carriage among Brazilian Indians Living in Peri-Urban Areas and Remote Communities

The emergence of Community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) infections among indigenous populations has been reported. Usually, indigenous communities live in extreme poverty and are at risk of acquiring infections. In Brazil, healthcare inequality is observed in this population. To date, there are no reports of CA-MRSA infections, and no active search for asymptomatic S. aureus carriage has been conducted among Brazilian Indians. The aim of this study was to investigate the prevalence of colonization with S. aureus and CA-MRSA among Brazilian Indians. We screened 400 Indians (from near urban areas and remote hamlets) for S. aureus and CA-MRSA colonization. The isolates were submitted to clonal profiling by pulsed-field gel electrophoresis (PFGE), and selected isolates were submitted to multilocus sequence typing (MLST). Among 931 specimens (nasal and oral) from different indigenous individuals in remote hamlets, S. aureus was cultured in 190 (47.6%). Furthermore, CA-MRSA was found in three isolates (0.7%), all SCCmec type IV. PFGE analysis identified 21 clusters among the S. aureus isolates, and MLST analysis showed a predominance of sequence type 5 among these isolates. Our study revealed a higher prevalence of S. aureus carriage among Shanenawa ethnicity individuals (41.1%). Therefore, ethnicity appears to be associated with the prevalence of S. aureus in these populations.


Introduction
Infections caused by methicillin-resistant Staphylococcus aureus (MRSA) are of particular concern because they are harder to treat due to resistance to some antibiotics and because they are no longer limited to healthcare settings [1,2]. On the other hand, globally, strains of methicillin-sensitive S. aureus (MSSA) have been associated with a considerable burden of invasive disease, especially among indigenous communities [2][3][4][5]. Colonization by S. aureus represents a risk factor for the occurrence of autogenous infections and cross-transmission to other individuals [6]. Historically, indigenous populations have been affected by high rates of infectious diseases [2,7,8]. Outbreaks of Community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) were first described among indigenous communities. A remarkable observation was made in Australia in 1980 when a strain of MRSA arose spontaneously in remote Aboriginal communities [9,10]. There is a global trend of considerably higher rates of invasive S. aureus disease among indigenous populations [11][12][13][14].
According to the most recent Brazilian Census [15], 305 indigenous ethnic groups are found across the Brazilian territory, with a population of 896,917 indigenous people living in the country: 324,834 in urban areas and 572,083 in rural hamlets. Ethnic groups and tribes vary widely in their customs and living conditions, including food, personal hygiene, religious rituals, and housing.
To date, there are no reports of CA-MRSA infections among Brazilian Indigenous, and no active search for asymptomatic S. aureus carriage and virulence profile has been conducted. To improve our understanding of S. aureus epidemiology among Brazilian Indians, we investigated the contribution of individual, household, and pathogen-related factors associated with S. aureus carriage. The objectives of the study were to investigate the prevalence of colonization with overall S. aureus and CA-MRSA and virulence factors among Brazilian Indians. We were especially interested in comparing populations living close to urban centers with those from remote areas.

Results
A total of 190 S. aureus isolates were recovered from the nasal and oral mucosa of 400 individuals (116 from the São Paulo [SP] state and 284 from the Acre [AC] state). The overall prevalence of S. aureus colonization was 47.6% (95% confidence interval [CI], 42.6-52.6%) and did not differ among the two study groups (49.5% in AC, 43.1% in SP). Only three subjects (all from AC) carried CA-MRSA, with a prevalence of 0.7% (95% CI: 0. 19-2.37). In addition, the CA-MRSA isolates were tested for susceptibility to penicillin, ceftaroline, quinupristin, sulfamethoxazole, clindamycin, erythromycin, and levofloxacin, showing resistance only to penicillin. All CA-MRSA strains harbored SCCmec type IV.
In the analysis of risk factors associated with the carriage of S. aureus, the univariate Poisson regression model revealed a positive association with age and the number of baths (Table 1). When the factor ethnicity was observed in which the Shanenawa group (the largest group among the indigenous populations studied) was used as a reference, there was a negative association between the Teregua and Kaxinawa ethnicities and the outcome ( Table 2). The Shanenawa ethnicity showed a positive association with S. aureus carriage when analyzed as a dichotomous variable and was therefore used as the reference in relation to the other ethnicities.
However, in the multivariate analysis, ethnicity was the only independent factor associated with S. aureus carriage. The age variable showed a marginally significant p-value (p = 0.08). Although age and the number of baths did not remain in the multivariate model, they seem to contribute to the outcome studied.
Considering the power of the association of ethnicity with the outcome, univariate analyses were performed comparing the prevalence of virulence factors associated with the ethnic types in order to identify differences in pathogenicity between the S. aureus isolates that colonize the indigenous groups studied. From the same perspective, risk factors (habits and customs) were analyzed, and demographic variables (income and the number of household members) were associated with S. aureus carriage according to ethnic group. The results show that the Shanenawa ethnicity stands out in relation to the prevalence of the virulence genes, as well as the variables related to habits, customs, and demographics (Tables 2-4).

Discussion
In this study, we found that ethnicity was the only predictor associated with S. aureus colonization in Brazilian Indians. Furthermore, we identified that income was associated with the prevalence of S. aureus carriage among almost all ethnicities. Previous studies also identified a possible association between race, ethnicity, and socioeconomic status, suggesting that these factors seem to contribute to the phenomenon of S. aureus colonization [15,16]. Ethnicity is related to the collectivity of individuals, and it is distinguished by their sociocultural specificity, mainly reflected in the language, religion, habits, and living conditions of populations. Considering these facts, it is important to emphasize that most Indian populations live in agglomerations under poor hygiene and

Discussion
In this study, we found that ethnicity was the only predictor associated with S. aureus colonization in Brazilian Indians. Furthermore, we identified that income was associated with the prevalence of S. aureus carriage among almost all ethnicities. Previous studies also identified a possible association between race, ethnicity, and socioeconomic status, suggesting that these factors seem to contribute to the phenomenon of S. aureus colonization [15,16]. Ethnicity is related to the collectivity of individuals, and it is distinguished by their sociocultural specificity, mainly reflected in the language, religion, habits, and living conditions of populations. Considering these facts, it is important to emphasize that most Indian populations live in agglomerations under poor hygiene and sanitation conditions, with large socioeconomic disparities that can contribute to S. aureus carriage [17]. The traditional epidemiological factors for S. aureus colonization, such as young age, male sex, underlying comorbidities, smoking, and previous hospitalization, were not relevant in the population studied, in agreement with a study investigating an Aboriginal community in Canada [18].
Despite widespread knowledge of the process of S. aureus carriage as a predictor of S. aureus infections, in Brazil, little is known about the dynamics of colonization and infection with S. aureus among indigenous populations. Our study identified a prevalence of S. aureus and CA-MRSA carriage of 47.6% and 1.0%, respectively. CA-MRSA strains were restricted to indigenous people belonging to the Amazon region of Brazil, all of them harboring SCCmec IV. A study conducted on the Amerindian population of Wayampi in the village of Trois Sauts, an isolated region in the Amazon forest of French Guiana, demonstrated rates of nasal carriage and persistent colonization with S. aureus of 57.8% and 26%, respectively. None of the isolates showed resistance to methicillin [19].
Our findings revealed that CA-MRSA was rare and, as also reported in a populationbased survey conducted in Brazil, all CA-MRSA harbored SCCmec IV [20]. On the other hand, the prevalence of S. aureus was higher than that reported for the general population [20]. Antimicrobial susceptibility testing of the MRSA isolates using disks impregnated with penicillin, ceftaroline, quinupristin, sulfamethoxazole, clindamycin, erythromycin, and levofloxacin revealed sensitivity to all antimicrobials, except for penicillin. All isolates were also resistant to penicillin. According to Sader et al. [21], sensitivity rates of CA-MRSA strains are generally higher when compared to hospital-associated MRSA, especially for clindamycin and levofloxacin. The same was observed for ceftaroline, a drug that has recently been introduced on the market as an option for the treatment of infections caused by MRSA.
The results obtained regarding the detection of virulence genes show diversity in the pathogenicity profile of S. aureus isolated from indigenous populations, similar to those already identified in non-indigenous populations [22]. Among the most prevalent virulence factors were genes coding for hemolysins hla and hld and the icaA and icaD genes involved in biofilm formation. According to Bride et al. [23], these genes are often produced by a large number of S. aureus strains. The findings obtained in the present study showed 13.5% of isolates carrying the PVL gene, none of them associated with CA-MRSA. According to Boan et al. [24], infections with PVL-positive S. aureus strains disproportionately affect young indigenous people or subjects with fewer healthcare-related risk factors. In a previous study involving a non-indigenous population, only five isolates were found to harbor the gene coding for PVL, corresponding to 2.2% of all isolates [22]. Considering the specific living conditions of the indigenous populations mentioned above, it is possible to infer that such factors may contribute to S. aureus carriage and the occurrence of skin infections associated with the PVL gene (lukS-PV) [24].
The univariate analyses compared the prevalence of virulence genes associated with ethnicity type and also analyzed the risk factors (habits, customs, and demographic factors) associated with S. aureus carriage according to ethnicity; significant results were obtained for the sec, hld, tst, icaA, and icaD genes. Regarding habits, customs, and demographics, all variables tested were statistically significant: skin pigmentation, number of baths, use of medicinal herbs, baths in the river, monthly income, and number of household members. It is worth noting that individuals of the Shanenawa ethnicity, which was used as a reference and showed a positive association with S. aureus when tested as a dichotomous variable, carried the most virulent and resistant S. aureus strains. Commonly, this ethnicity more frequently exhibits habits, customs, and demographic factors that may be associated with a higher prevalence of colonization by S. aureus.
Regarding PFGE analysis, 21 clusters circulating among the indigenous populations from southeastern and northern Brazil were identified. Of these, five major clusters (1, 4, 6, 10, and 14) contained a larger number of isolates that were more prevalent among the regions studied. Interesting data were observed for these clusters in which isolates from indigenous communities in the southeast formed clusters with isolates identified in the Amazon region. Another interesting finding was that, although not one of the most prevalent, cluster 7 contained sensitive isolates that grouped with one resistant isolate belonging to ST 5. According to Robinson and Enright [25], the dissemination of genotypes occurs not only among individuals but also through the ability to transmit the mobile genetic element SCCmec IV through MSSA strains. The remaining CA-MRSA isolates formed clusters with clone USA800. This clone represents a lineage of pandemic MRSA such as CC5 [26].
The sequence type (ST) number 5 was the most frequent. This ST is one of the most prevalent clonal complexes in both hospital and community settings [27,28]. Moreover, the other STs identified in the indigenous populations studied resemble those commonly disseminated in other types of populations [29,30]. ST 5 associates with different types of SCCmec (I, II, and IV) and represents the ancestor of the clone described as a Cordobes/Chilean clone and of the MRSA pediatric epidemic clone detected in 1999 [26,31].

Study Design, Subjects, and Procedures
A cross-sectional, population-based study was conducted. Individuals of indigenous populations from two different Brazilian states were included. One state was São Paulo, the largest urban center of the country with an area of 1,521,110 km 2 and 11,253,503 inhabitants, and the other state was Acre, located in the remote region of the Amazon forest, which comprises a large area of 8,835,520 km 2 and has 377,057 inhabitants according to IBGE [14].
Our sample size was established by calculating proportions, assuming a level of significance of 5% and accuracy of 5%, according to the number of individuals in each village/hamlet. A total of 400 subjects were included in the study. Of those, 116 lived in a peri-urban village in the city of Bauru, São Paulo State (SP), southeastern Brazil, and 284 lived in small hamlets in Acre State (AC), Brazilian Amazon region. Nasal and oropharyngeal swabs were collected from September to November 2014. Data on demographics (including ethnicity), habits and customs, comorbidities, and clinical variables were collected in interviews with the subjects. Table 5 lists the variables analyzed.

Specimen Collection, Culture, and Antimicrobial Susceptibility Tests
Nasopharyngeal and oral swabs were collected, transported in Stuart medium, and cultured on a selective medium (Baird-Parker Agar) for up to 48 h. Staphylococcus aureus was identified based on colony morphology and standard biochemical tests [32]. Susceptibility to methicillin was tested by disk diffusion using cefoxitin (30 µg) disks according to the recommendations of the Clinical Laboratory Standards Institute [33]. For the MRSA isolates, susceptibility to penicillin, clindamycin, levofloxacin, erythromycin, sulfamethoxazoletrimethoprim, quinupristin-dalfopristin, and ceftaroline was tested by disk diffusion and to oxacillin and vancomycin by the E test (BioMerieux) [32].

Molecular Methods
Methicillin resistance was also assessed through the detection of the mecA gene by real-time PCR performed in a LightCycler system (Roche) [39]. The staphylococcal cassette chromosome (SCCmec) was characterized using the protocol described by Milheiriço et al. [40]. Pulsed-field gel electrophoresis (PFGE) was performed on SmaI-digested chromosomal DNA according to the protocol described by McDougal et al. [41]. Band patterns were analyzed with BioNumerics 7.6 (Applied Maths), and a dendrogram was generated using the unweighted pair group method with arithmetic mean (UPGMA). Clusters were defined as any group of more than four isolates with a Dice similarity coefficient ≥ 80%, assuming tolerance and optimization of 0.5 and 1.25%, respectively [41].
Representative isolates of the PFGE clusters were submitted to multilocus sequence typing (MLST) [42]. The arcC, aroE, glpF, gmk, pta, tpi, and yqIL genes were amplified separately with the specific primers described by Enright et al. [26], and both strands were sequenced. Sequence types (ST) were assigned using the BioNumerics 7.6 software [26].

Epidemiological Analysis
Data were collected and analyzed using Epi Info for Windows, version 3.5.1 (© Centers for Disease Control and Prevention, Atlanta, GA, USA) and SPSS 20.0 (IBM, Armonk, NY, USA). The base of the study was formed by subjects (indigenous people) with a culture positive for S. aureus. Univariate and multivariate Poisson regression models were built. A stepwise forward strategy was used to select variables for the multivariate models, with a p-value of 0.05 as a limit for inclusion/removal [42,43]. Figure A1 (Appendix A) shows the workflow of the methodological steps.

Conclusions
In conclusion, ethnicity appears to be associated with a higher prevalence of S. aureus and virulence in special populations, even though the prevalence of CA-MRSA was low. These results might be related to specific habits and customs since poor housing conditions, hygiene, and sanitation are features present in most ethnical groups and can influence the carriage and dissemination of S. aureus among populations. It is, therefore, extremely important that these factors be considered for the development and implementation of strategies designed to control the spread of microorganisms among different human populations, especially considering the health inequality of indigenous populations in Brazil.