Prevalence and resistance patterns of commensal S. aureus in community-dwelling GP patients and socio-demographic associations. A cross-sectional study in the framework of the APRES-project in Austria

Background The aim of the present study was to assess the prevalence and resistance of commensal S. aureus in the nasal microbiota of community-dwelling persons in Austria, as well as to identify possible associations with socio-demographic factors. Multi-drug resistance in this population was additionally studied. Method This cross-sectional study was conducted within the context of the European APRES project. In nine European countries, nasal swabs were collected from 32,206 general practice patients who received care for non-infectious reasons. In Austria, 20 GPs attempted to recruit 200 consecutive patients without infectious diseases, with each patient completing demographic questionnaires as well as providing a nose swab sample. Isolation, identification, and resistance testing of S. aureus were performed. Statistical analyses included subgroup analyses and logistic regression models. Results 3309 nose swabs and corresponding questionnaires from Austrian subjects were analyzed. S. aureus was identified in 16.6 % (n = 549) of nose swabs, of which 70.1 % were resistant against one or more antibiotics, mainly penicillin. S. aureus carrier status was significantly associated with male sex (OR 1.6; 1.3–2.0), younger age (OR 1.3; 1.0–1.8), living in a rural area (OR 1.4; 1.1–1.7) and working in the healthcare sector (OR 1.5; 1.0–2.1). Multi-drug resistances were identified in 13.7 % (n = 75) of the S. aureus carriers and 1.5 % (n = 8) tested positive for MRSA. The highest resistance rate was observed against penicillin (64.8 %), followed by azithromycin (13.5 %) and erythromycin with 13.3 %. Conclusion This study describes the prevalence and resistance patterns of commensal S. aureus in community-dwelling persons in Austria and shows that differences exist between socio-demographic groups. Demographic associations have been found for S. aureus carriers but not for carriers of resistant S. aureus strains. Only two thirds of S. aureus strains were found to be resistant against small spectrum penicillin. As it is recognized that one of the corner stones for the containment of antibiotic resistance is the appropriate prescription of antibiotics in the outpatient sector, this finding lends support to the avoidance of prescription of broad-spectrum antibiotics to treat S. aureus infections in the community.


Background
One of the major challenges for healthcare systems around the world is the growing prevalence of antibiotic resistance (AR) [1]. The World Health Organization's 2014 report on AR began with an introductory statement that raised concern for, "A post-antibiotic era-in which common infections and minor injuries can kill-far from being an apocalyptic fantasy, is […] a very real possibility for the 21st century.
[1]" AR is known to be associated with an increased morbidity and mortality [1][2][3][4][5][6]. At least four reasons for recent increasing rates in AR have been suggested: 1) the excessive use of antibiotics by humans, including antibiotic overtreatment fostered by over-thecounter selling of antibiotics and inappropriate prescriptions of antibiotics for viral infections [1, [7][8][9][10], 2) influences of travelling [11], 3) the excessive use of antibiotics in livestock breeding [1, [12][13][14][15], and 4) the stagnancy in development of new generations of antibiotics [1,16]. Together, this has led to a large output of resistant bacteria into the environment, where resistance genes can disseminate easily [1, [17][18][19]. These factors for AR imply that community exposures and ambulatory care have a large impact on acquisition and resistance. However, the majority of existing information on antibiotic resistance pattern has been obtained from hospitalized patients, rather than from samples of community-dwelling persons [9,20,21]. Given that community-acquired AR differ from patterns observed in hospitalized patients, identifying the prevalence and resistance in the general population would provide an important source of knowledge on the acquisition of resistance in pathogens [1].
While S. aureus is a common part of the normal human microbiota, it is also often found to be responsible for several severe infections [22][23][24]. S. aureus is of particular interest due to its special capacity to acquire resistance, as exemplified by the recent steep increase in AR levels of S. aureus strains against penicillin [25]. Furthermore, circumstances such as the increasing prevalence of methicillin-resistant S. aureus (MRSA) that result in limited treatment options, continue to press the growing international concern and need for surveillance of the situation in the non-hospitalized setting [1, 26,27].
In the context of the EU-project "APRESthe appropriateness of prescribing antibiotics in primary health care in Europe with respect to antibiotic resistance" [28], it was our aim to gain detailed information about the prevalence and resistance of commensal S. aureus of community-dwelling GP patients in Austria. Specifically, we aimed to evaluate a community population for nasal S. aureus carriage and to observe any associations with socio-demographic data, as well as to assess for multi-drug resistance. Moreover, we performed subgroup-analyses on those patients carrying S. aureus with resistance to any antibiotic aside from penicillin, as this group may be associated with higher risks for acquiring AR and multi-drug resistances.

Design
This cross-sectional study took place in the context of the European APRES project, which included nine European countries and a methodology that has been described in depth in several publications [8,[28][29][30]. Therefore, we describe in this publication only the methods relevant for Austria. [22] The study analysis was designed in accordance with the STROBE statement for cross-sectional studies [31].

Recruitment of study participants in Austria
First, twenty general practitioners (GPs) were recruited via electronic invitation by the Austrian Society of General Practice and the research network of the Department of General Practice at the Medical University of Vienna. In line with the APRES project, those recruited constituted a fair representation of the national GP population with regard to sex, age and federal state [32]. Additionally, GP demographics were included in this study. Between November 2010 and July 2011 these twenty GPs were asked to identify 200 consecutive patients aged four years and older. Detailed inclusion and exclusion criteria for the patients were published elsewhere [28,30]. Most important for this analysis was the exclusion criterion that the patients' consultation had to be due to a non-infectious disease and that the patients did not take antibiotics for a minimum of three months prior to the study. Additional important exclusion criteria included those living in a nursing home, those with any hospitalization in the three month period prior to the study, the immunocompromised, and all those aged less than 4 years. A nose swab was collected and each patient then completed two questionnaires regarding their sociodemographic data. Among this was the APRES patients' questionnaire and an additional questionnaire designed by the Department of General Practice at the Medical University of Vienna, with further questions relating to demographics. The additional questionnaire is described in detail in the publication of Hoffmann et al. [30]. For those patients between four and 14 years, a parent completed the questionnaires to ensure the reliability of the data.

Material
The nose swabs were sent within 24 hours to the national laboratory (Institute of Hygiene, Microbiology and Tropical Medicine, Elisabethinen Hospital, Linz, Austria) for the identification and isolation of S. aureus. Charcoal swabs were used as transport medium (Transystem, 114 C; Copan Italia, Brescia, Italy). The swabs were sent to the laboratory using special envelopes to ensure security and fast delivery of the swabs. All laboratories used standardized protocols for the identification of S. aureus [28]. Those swabs that were found to have isolated S. aureus strains were sent to the APRES central laboratory for further AR testing. All AR testing was performed in accordance with the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines and EUCAST epidemiological cut-offs were used as breakpoints [33]. Antimicrobial susceptibility testing was performed for multiple antibiotics, including, azithromycin (AZITH), ciprofloxacin (CIP), clindamycin (CLIND), daptomycin (DAPT), erythromycin (ERY), gentamicin (GENTA), linezolid (LIN), oxacillin (OXA), penicillin (PEN), tetracyclin (TETRA), trimethoprim-sulfamethoxazole (TRISUL), and vancomycin (VANC). S. aureus ATCC 29213 was used as control strain. All isolates susceptible to CLIN and resistant to ERY were tested for inducible CLIIN resistance by means of the D-test [34]. In the case of a positive D-test, strains were considered resistant to CLIN. Additionally, the presence of MRSA strains was evaluated. The MRSA resistance screening was previously described in detail by den Heijer et al [8].
Multi-drug resistance was defined as a S. aureus strain that was resistant to three or more antibiotics.

Patients' information
Patient information in the two questionnaires used for this study included sex, age, educational level, country of origin, location of residence, and profession. Age was clustered into four groups: age 4-18 years (children), 19-39 years (young adults), 40-64 years (older adults), and 65 years and older (elderly). The highest educational level attained was surveyed in three groups: primary, secondary and tertiary education. Tertiary education was defined as completion of university or any further or post graduate education. Country of origin was assessed with the question "What is your country of origin?" This variable was grouped into four clusters: 1) Austria, 2) European Union (EU) 15 countries including European Free Trade Association (EFTA) countries (EU15+), 3) new EU 28 countries (EU28), and 4) all other countries. Location of residence was dichotomized into urban areas (big and intermediate cities) and rural areas (small cities, villages and countryside). Classification of the area was applied by using the European DEGURBA (degree of urbanization) classification [35]. Profession was assessed by "Do you work in any of the following occupational fields?" with the answer categories "healthcare", "livestock farming", "kindergarten teacher/ (day) nanny", or "others". All sociodemographics were defined as independent variables.

Data analyses
The initial analysis was carried out for the distribution of socio-demographic data for all patients, S. aureus carriers, S. aureus carriers with antibiotic resistance to any antibiotic as well as any antibiotic besides PEN, and those carriers of MRSA. This was performed using descriptive statistical methods (absolute and relative frequencies and cross-tabs) and the Chi-Square Independency or Fisher's Exact test for categorical variables. If independency could not be proven, the two-proportion z-test to compare two or more column proportions including the Bonferroni method for multiple testing was used to determine which particular categories were not independent. Next, multivariable logistic regression models were conducted [36]. In the first regression model, S. aureus carrier status was the dependent variable, while those S. aureus carriers with AR were recognized as the dependent variable for the second regression model and those with AR besides PEN for the third model. All socio-demographic variables were included simultaneously in the model and a backward regression model was performed. Variables were excluded at a level of 0.20. We chose a p-value cut-off point of 0.2 for the backward selection as more traditional levels, such as 0.05, can fail to identify variables known to be important [37,38]. This was followed by purposeful testing of eventual missed effects of the excluded variables. This step can be helpful in identifying variables that, by themselves, are not significantly related to the outcome, yet make an important contribution in the presence of other variables. At the end of this final step, the preliminary main effects model could be analyzed [37]. Additionally, the regression models were adjusted for GP practice code to account for any possible inter-practice effect. This was performed by building a dichotomous dummy variable for each GP practice, which was included in the models.
Then, the number of resistances of S. aureus against AB was assessed and the distribution of the socio-demographic variables in relation to the number of antibiotics resistant to S. aureus calculated by means of descriptive statistical methods.
The significance level for all calculations was p < 0.05 and the confidence interval was 95 %. SPSS Statistics version 22.0 was used for the statistical analyses.

Ethical considerations
The study was approved by the Ethics Committee of the Medical University of Vienna (EC # 568/2010).
Each participant had to complete a written informed consent form prior to participation. If the patient was younger than 18 years, both a parent and the child each completed written informed consent forms. Those aged 4 to 13 completed a special consent form for children, while those aged 14 to 18 completed a form for adolescents.
Prevalence of S. aureus, resistance, MRSA and socio-demographic factors Table 1 shows the distribution of the socio-demographic factors within the S. aureus carrier group, the group of carriers with AR, the group of carriers with AR besides PEN, and the MRSA group. Males, younger adults and those living in the countryside showed a statistically The subscript letters represent a subset of the variable category which is not significantly different at a significance level of p < 0.05 if it is the same subscript for the same sub-variable Significance at significance level of p < 0.005 significant increased likelihood to be S. aureus carriers. Additionally, Table 1 shows that 64.8 % of the S. aureus carriers had resistances aside from PEN. These cases were more frequent in persons living in rural areas.
The number of MRSA carriers was too small to detect serious differences in socio-demographics. Also for S. aureus carriers with resistance to one or more antibiotics, no differences in the socio-demographics could be found.
In the first adjusted multivariable logistic regression model, positive associations with S. aureus carrier status could be found for those of the male sex, between the age of 19 and 39 years, living in an rural area, and working in the health care sector (Table 2). Additionally, in the third adjusted model living in a rural area was found to increase the probability of being a S. aureus carrier with resistances besides PEN ( Table 2).

Multidrug-resistance of S. aureus and associations with socio-demographic factors
If S. aureus strains were resistant to three or more antibiotics, nearly all were resistant to PEN, AZITH, ERY, and CLIND (Table 3). S. aureus strains that were resistant against PEN most commonly demonstrated isolated resistant against PEN, but did show secondary resistances against three other AB, mainly AZITH, ERY and CLIND ( Table 4). Out of the MRSA strains, five were resistant against other AB besides OXA and PEN (Table 5). Of all S. aureus carriers, with the exception of MRSA carriers, 13.7 % (n = 72) were identified as multi-drug resistant strains. Persons living in urban areas were found to have significantly more frequent multi-drug resistant strains (Table 6).

Discussion
The present study evaluated the prevalence and resistance patterns of commensal nasal S. aureus strains from urban and rural GP patients in Austria. The results subsequently show differences associated with sociodemographic factors in these populations.
Males, young adults, persons living in the countryside and those persons working in the healthcare sector were significantly more likely to be S. aureus carriers. Rural areas demonstrated a clear association with S. aureus carrier status, even when adjusting for GP practice. This is likely a consequence of members of rural communities having more contact with S. aureus. Rural areas include those with farmland, which may suggest a higher exposure to the influence of livestock breeding. This has previously been described in literature as a risk factor for S. aureus carrier and resistance carrier status [12][13][14][15]. However, interestingly, livestock workers did not demonstrate a higher prevalence of carrier status in our analysis, though this could be due to the small sample size of this subgroup. It was additionally recognized that males are more likely to be S. aureus carriers, with an OR of 1.60 compared to women. This result is supported by other studies that have also observed higher S. aureus carriage in men [22]. Possible reasons for this finding include host genetics or human innate immune factors, smoking status, or vitamin D levels [39][40][41]. Compared to other professions, healthcare workers were found to have a greater likelihood to be S. aureus carriers (OR of 1.47). This finding is not surprising, given the regular exposure and proximity of healthcare workers to variable and resistant strains of microbials in hospital and ambulatory settings. However, it does offer important implications for healthcare workers and the health care system. In the consultation setting, healthcare workers should be considered to have a higher likelihood to carry S. aureus strains and an increased chance to transfer these to their patients. Therefore, hand-hygiene following patient contact is essential to hinder the transfer [42].
Nearly 30 % of S. aureus strains identified were susceptible to all antibiotics tested. The vast majority (64.8 %) of commensal S. aureus strains was resistant to small-spectrum PEN, followed by the macrolide antibiotics AZITH and ERY as well as CLIND, with a resistance rate around 12 %. Comparing these findings to those in other European countries, it becomes clear that Austria is nearly average in terms of prevalence rates when compared to other European nations [8]. As it is recognized that one of the corner stones for the containment of AR is the appropriate prescription of antibiotics in the outpatient sector [28], this finding could support to the avoidance of prescription of broad-spectrum antibiotics to treat S. aureus infections in the community. This is because of low levels of resistance in these settings, aside from PEN resistances in ambulatory populations.
With regard to the overall resistances of the S. aureus strains carried, no associations with socio-demographics were directly identified. However, for the S. aureus strains that were resistant to antibiotics besides PEN, associations with living in rural areas could be found. This goes in line with the S. aureus carriers that were more frequently found in rural areas. In contrast, multi-drug resistances were observed more often in individuals living in urban areas. The higher rates of multi-drug resistant S. aureus strains in individuals living in urban areas could be a consequence of higher travel activity compared to those living in rural areas [11]. It could additionally be surmised that persons living in urban areas have easier access to ambulatory or hospital sectors, and thus have a higher likelihood to present for even simple conditions, such as a common cold. Due to the fact that Austria has no gatekeeping system in primary care, patients may access care at any point and as often as they wish. Thus, these patients could have a higher probability of being exposed to antibiotics [43]. In line with this, Hoffmann et al. found that in Eastern Austria, approximately one third of persons that do not get an antibiotic prescription from the GP will subsequently present to a specialist to receive one [44].
In this study with participants without infectious diseases, we observed low rates of MRSA in Austria, with only 1.5 % of S. aureus carriers having MRSA. An important finding of    our study is that there exists a difference among the MRSA rates previously examined in hospital samples and those samples from sick patients presenting to the ambulatory sector with a bacterial infectious disease. For example, in Austria in 2013, the Austrian resistance report stated that hospitalized patients were found to have invasive (blood and cerebrospinal fluid) MRSA strains with a prevalence of 9.1 %, and non-invasive MRSA strains in 6.7 % of cases studied. Meanwhile, non-invasive MRSA strains were detected from sick persons in the outpatient sector with a prevalence of 4.0 % [21]. If S. aureus strains were resistant to three or more antibiotics, we observed that nearly all of these were resistant to PEN, AZITH, ERY, and CLIND (Table 3, Table 4 and Table 5). This has profound and practical implication for care of patients in Austria, particularly as prescribers can look to utilize other antibiotics in patient populations with high likelihood for multiple resistances.

Strength and limitations
A strength of the present study was the community setting in which patients were drawn from a GP population who visited the practice for a non-infectious reason. This is particularly important as no previous study of depth has been carried out in this population. Another strength was the large sample size and the similarity of the sample with the Austrian population with regard to sex, age and educational level [45]. However, as the study population has been recruited in general practices we lacked participants among those that otherwise did not present to a GP within this time frame. Other limitations were the non-randomized recruitment strategy of GPs and patients and the fact that the questionnaire was available exclusively in German. It may be speculated that more GPs and patients interested in the topic of AR participated in the study, which would have led to an underestimation of the real resistance pattern of antibiotics, especially, regarding non-German speaking migrants. Finally, socio-demographic factors explain only about 5 % of variance of the S. aureus prevalence and resistance rates, suggesting that other factors such as host genetics or human innate immune factors, smoking status, vitamin D levels, travel activities, or exposure to antibiotics may play a role as well. As this study is crosssectional there is limitation to the explanatory power. The subscript letters represent a subset of the variable category which is not significantly different at a significance level of p < 0.05 if it is the same subscript for the same sub-variable Significance at significance level of p < 0.005

Conclusion
Emergence of increasingly resistant strains of microbials has been recognized by the WHO and others as a global health threat. AR is heavily influenced by dynamics that occur outside of hospital settings, such as prescribing patterns and livestock breeding. Despite this, the majority of studies of AR in Europe have occurred in hospitalized patients or sick ambulatory patients presenting with bacterial infections. This study, in contrast, describes the prevalence and resistance pattern of commensal S. aureus in community-dwelling GP patients in Austria and shows differences between socio-demographic groups. Particularly, differences were identified in the prevalence and multi-drug resistances of commensal S. aureus between the sexes and individuals living in rural and urban areas in Austria. More than two thirds of all S. aureus carriers had resistant strains, though the vast majority was resistant against small-spectrum PEN only. Physicians should consider these findings when determining therapeutic options with antibiotics. Particularly, this finding could lend support to the avoidance of prescription of broad-spectrum antibiotics to treat S. aureus infections in rural community settings.
As we continue to witness a global rise in AR, it is imperative that acknowledged patterns of resistance are accurate and reflect true prevalence at the community level. Responding to outbreaks in resistant strains will require an understanding of social and demographic factors as well. We urge the further study of communities and, otherwise healthy, populations in order to continue to elucidate representative prevalence and resistance patterns. Ideally, continual surveillance of resistance patterns and antibiotic consumption in the outpatient setting should be carried out to detect changes early and to inform evidencebased decisions in a timely manner. First of all, we would like to thank the APRES consortium as well as the APRES steering committee for providing the frame for all the research activities. In addition, we would like to thank the 20 GPs for their participation and the recruitment of the participants, Lukas Heschl and Dominik Stelzer for the acquisition of the data, and Paulina Dabrowska for the data management support.

Funding
This study was conducted within the context of the European APRES project. APRES is financially supported by the Seventh EU Framework Programme "APRES -The appropriateness of prescribing antibiotics in primary health care in Europe with respect to antibiotic resistance" (grant agreement number 223083).