Prevalence and risk factors contributing to antibiotic-resistant Staphylococcus aureus isolates from poultry meat products in South Africa, 2015–2016

Staphylococcus aureus, including methicillin-resistant strains, has been detected in food products of animal origin globally. Limited data have been reported on the factors contributing to antibiotic resistance of food-borne pathogens in South Africa. The primary aim of this study was to determine the prevalence of S. aureus, including antibiotic-resistant strains, in poultry meat products as well as the evaluation of potential risk factors for contamination of poultry meat products with antibiotic-resistant S. aureus isolates. A cross-sectional investigation was conducted in municipalities located across the nine provinces of South Africa, which included abattoirs, meat processing facilities, retail outlets and cold stores at the major ports of entry into South Africa. Staphylococcus aureus isolates obtained from various poultry meat products were tested for susceptibility to 14 antibiotic compounds representing 10 antibiotic classes using the Kirby–Bauer disc diffusion method. Potential risk factors were evaluated using a logistic regression model. Of the 311 samples tested, 34.1% (n = 106) were positive for S. aureus (95% confidence interval [CI], 28.9% – 39.7%). Seventy-two of the 106 isolates were randomly selected for antibiotic sensitivity testing. Twenty-one per cent (n = 15) of the isolates selected for sensitivity testing were methicillin-resistant strains (95% CI, 12.2% – 32.0%). Multi-drug resistance was detected in 22.2% (n = 16) of these isolates tested (95% CI, 13.3% – 33.6%). Origin of the product (p = 0.160), type of meat product (p = 0.962), type of facility (p = 0.115) and facility hygiene practices (p = 0.484) were not significantly associated with contamination of poultry meat products with methicillin-resistant strains. The study provides baseline data for further studies on antibiotic resistance risk assessments for food-borne pathogens, including S. aureus, which should guide the implementation plans of the South African National Antimicrobial Resistance Strategy Framework, 2017–2024.


Introduction
Staphylococcus aureus is a human and animal commensal bacterium with the potential to cause disease in susceptible hosts (Gosbell & Van Hal 2013). It produces enterotoxins, which accumulate in food items during production and processing and which may cause food poisoning (Centers for Disease Control and Prevention 2011; Gosbell & Van Hal 2013;Grumann, Nübel & Bröker 2014). Although the bacteria may be destroyed by heating contaminated food items, their toxins are heat resistant (Centers for Disease Control and Prevention 2011; Grumann et al. 2014).
http://www.jsava.co.za Open Access Antimicrobial resistance (AMR) presents a major threat to global public health and has become a serious challenge in clinical practice and healthcare (Laxminarayan et al. 2013;McAdam et al. 2012;WHO 2016). Because of resistance to existing antibiotics and the lack of development of new antibiotic compounds, higher costs and reduced efficacy of treatment for common infections are experienced in healthcare settings (Centers for Disease Control and Prevention 2011; Laxminarayan et al. 2013;McAdam et al. 2012;WHO 2016).
Previously, MRSA strains were considered to be acquired exclusively through nosocomial transmission (Naber 2009). Methicillin-resistant S. aureus strains are present in communities and have also been detected in a variety of food products of animal origin in countries throughout the world (European Food Safety Authority 2009; Jones et al. 2002;Kluytmans 2010;Newell et al. 2010). The possibility of people acquiring antibiotic-resistant S. aureus infections through contaminated food products has been reported (European Food Safety Authority 2009; Jones et al. 2002;Kluytmans 2010;Newell et al. 2010). The ST398 MRSA strain has been identified in pig, cattle and poultry production systems and zoonotic transmission has been recognised (European Food Safety Authority 2009). The Dutch Food Safety Agency reported that 85% of the 264 MRSA isolates from 2217 retail meat samples were the ST398 strain (European Food Safety Authority 2009). The risk of food poisoning because of S. aureus in food products is not particularly dependent on whether resistant genes are present or not and the development of disease through the ingestion of contaminated food is rare but possible under extreme patient conditions (European Food Safety Authority 2009). However, the risk for colonisation during handling of contaminated products is more relevant and is dependent on various environmental, host and pathogen factors (European Food Safety Authority 2009). Ho, O'Donoghue and Boost (2013) reported that the handling of raw meat posed an occupational risk to food handlers for nasal colonisation by S. aureus. Heysell et al. (2011) reported that 21% of tuberculosis patients in a rural KwaZulu-Natal hospital were found to be nasal carriers of MRSA and 90% of these cultures were found in patients with concurrent HIV infection.
The aim of this study was to determine the prevalence of S. aureus, including antibiotic-resistant strains, in poultry meat products and to identify and evaluate potential risk factors that could contribute to the contamination of poultry meat products with antibiotic-resistant S. aureus isolates in South Africa. We investigated the hypothesis that specific factors related to the origin of meat products, type of meat products, type of facility and facility hygiene contribute to the contamination of poultry meat products with antibioticresistant S. aureus. The findings of the study could assist food safety policy-makers in identifying possible policy gaps and areas of focus for strengthening meat safety, thereby reducing the risk of disease exposure to susceptible people.

Materials and methods
Data on registered Food Business Operators (FBOs) were collected from environmental health officers of each municipality throughout South Africa (Table 1) through personal structured interviews conducted by the Department of Agriculture, Forestry and Fisheries (DAFF). Food Business Operators (i.e. abattoirs, meat processing facilities, butcheries, retail outlets and cold stores at the major ports of entry) were the sampling units. Samples were collected from January 2015 to August 2016.
A cross-sectional study was conducted to determine the prevalence of antibiotic resistance among S. aureus isolates and to identify potential risk factors for contamination of poultry meat samples. Samples from various municipalities situated in all nine provinces and the major ports of entry to South Africa were included in this study from January 2015 to August 2016. The municipalities within the provinces that responded to the request from DAFF to participate in the project were included in the sampling frame. Considering the nature of this study, sampling all the FBOs registered by each responding municipality was impractical and therefore a sample was drawn to best represent the total FBOs in South Africa using non-probability sampling.
The number of registered abattoirs in all nine provinces was estimated at 612. High throughput and low throughput abattoirs were targeted in the study with the exclusion of non-slaughtering and infrequently slaughtering abattoirs. Fifty-two municipalities reported the registration of 485 FBOs producing ready-to-eat (RTE) meat products, while municipalities with no registered RTE FBOs and abattoirs were excluded. Selection of FBOs within a municipality was targeted, based on confirmation of production activities on The facilities involved at the major ports of entry, that is, Durban, Cape Town and Port Elizabeth, were cold stores. The selected three ports of entry had less than 40 registered cold stores. The cold stores were randomly selected. At the cold store, products were stratified into red meat and poultry where consignments were randomly selected for sampling from each stratum. Within a consignment, cartons were randomly selected from between 2000 and 2700 cartons constituting a consignment per container for a composite sample from five randomly selected cartons.
The minimum sample size for this study was calculated (Dohoo, Martin & Stryhn 2009) as 385. Information pertaining to sample type and origin was captured on sample submission forms and DAFF VPH officers assessed the facility hygiene using a questionnaire with scoring criteria during sample collection. The information captured pertaining to the products sampled were the country of origin, full product description, other ingredients (for processed meat samples) and the storage temperature. The information captured pertaining to the facility where the product was sampled included location, registered description, units of output and hygiene. Aspects of hygiene evaluated included the quality of General Hygiene Principles delivery in practice, hygiene training of employees, status of maintenance of premises, status of workforce, status of management and status of prerequisite programmes documentation and records. The data captured from laboratory testing involved the detection of S. aureus (by culture, staphylase testing and microscopic examination) and antibiotic sensitivity testing results.  (Magiorakos et al. 2012). Cefoxitin was used as a surrogate marker for the detection of MRSA (Fernandes, Fernandes & Collignon 2005). Oxacillin was included in the testing panel for comparative purposes.
The data from the sample submission forms and facility hygiene questionnaires were digitalised by double entry into Microsoft Excel 2010. These data were exported to STATA (Software for Statistics and Data Science) version 13.0 software programme for statistical analysis using descriptive and analytical methods.
The prevalence of antibiotic-resistant S. aureus was determined from proportional calculations.
The origin of the sample was classified as local or imported and was based on the origin of the animals. The product type was classified as unprocessed for anatomically recognisable meat cuts and processed for all other products. Retail and cold storage facilities were classified as storage facilities, while facilities involved with meat production and processing were classified as production facilities. Scores for facility hygiene practices below 50% were classified as poor, from 50% to 69% were classified as satisfactory and from 70% to 100% were classified as good.

Ethical considerations
Ethics approval for the study was obtained from the University of Pretoria, Faculty of Health Sciences Research Ethics Committee. Written permission was obtained for the use of data from the DAFF project from the project owner, the Directorate: Veterinary Public Health. Anonymity of facilities will be maintained in the public domain. The findings of the study have been notified to the relevant regulatory authority for the investigation of potential legislative non-compliances.

Discussion
The 14 antibiotics were selected on the basis of use in animal production systems (Andreasen 2013;Eagar, Swan & Van Vuuren 2012) and treatment of infections caused by bacteria in humans (Auwaerter 2016). Although only 72 out of the 106 isolates were tested for antibiotic sensitivity, the distribution of the 72 isolates in terms of product origin and description was similar to that of the 106 isolates because the 72 isolates were randomly selected for sensitivity testing. From this study, the prevalence of S. aureus in poultry meat samples on the South African market between 2015 and 2016 was estimated to be 34.1%. This is below the estimation of 44% made by Oguttu et al. (2014) for RTE chicken in Tshwane, Gauteng Province, South Africa. This variation in observation might be because of the sourcing of the RTE chicken products from an informal market where regulated hygiene measures are not likely to be implemented and enforced.
The prevalence of MRSA among S. aureus isolates from poultry meat products considered in this study was 20.8%.    International studies reported this prevalence to be 7.89% in Thailand (Akbar & Anal 2013), 37.2% in Germany (Feßler et al. 2011), 1.6% in Italy (Normanno et al. 2007), 24.8% in the Netherlands (De Boer et al. 2009) and 26% in a collective study of five major US cities (Waters et al. 2011). The variation observed might be attributed to differences in the regulation of antibiotics used and hygiene management in production systems that were included in these studies.
Multi-drug resistance was observed in 22.2% of isolates. The highest resistance of S. aureus isolates observed was against oxytetracycline (43.1%) and penicillin (23.8%). Resistance to oxytetracycline was also reported to be the highest among antibiotics tested in Thailand (Akbar & Anal 2013). Staphylococcus aureus is well known to express the highest resistance to penicillin among the beta-lactam antibiotic class, and penicillin resistance by Gram positive bacteria has been reported since 1940 (Laxminarayan et al. 2013). The resistance to oxytetracycline is not surprising because this is one of the most commonly used antibiotics in South African livestock production systems. Oxytetracycline is registered for use in animal feed under the Fertilizers, Farm Feeds and Agricultural Remedies Act (Act No. 36 of 1947) and tetracyclines were reported to be the second most consumed class of antibiotic by the South African animal production industry (Eagar et al. 2012).
Previous studies have linked the detection of S. aureus and MRSA isolates in meat products to food handlers and poor hygiene practices during production processes. However, this study failed to demonstrate significant associations between the presence of MRSA in poultry meat products and facility factors, including processing and hygiene practices. This finding could indicate that contamination of products with the bacteria was not associated with slaughter and processing, but rather that the birds already carried the bacteria before presentation for slaughter and that the bacteria are most likely acquired either during rearing on farms or during handling before transport and slaughter. This inference is plausible because chickens, as well as other livestock species such as pigs and cattle, are known to be carriers of S. aureus, which is supported by the study conducted by Mkhize, Zishiri and Mukaratirwa (2017), which found that faecal and caecal samples tested from commercial broiler birds in KwaZulu-Natal, South Africa, contained antibiotic-resistant S. aureus bacteria. However, the lack of statistically significant findings could also be because of the study lacking sufficient power to detect significant associations as a result of the inadequate sample size.
This study had several limitations. With regard to the sampling, the number of samples analysed per province was not proportional to production or consumption in each province. Samples may have been misclassified by country of origin because packaging at retail outlets does not always carry the original labelling from the country of origin. All 553 samples and 106 S. aureus isolates could not be tested for antibiotic sensitivity because of limitations of laboratory resources and time constraints, which could have introduced sample bias; however, a comparison of the distribution of the 553 samples collected with those 311 included in the study (Table 1) did not reflect major differences between the groups. Convenience sampling of municipalities and establishments was conducted and therefore not necessarily representative of the entire country. The data were incomplete for facility hygiene, sample temperatures and enumeration of S. aureus from samples. In general, risk factors relating to animal health and environmental factors such as feed and farm hygiene were not accounted for in this study.
Despite the limitations of the study, the findings nevertheless present pertinent questions, including the effects of exposure to antimicrobial-resistant organisms from products of animal origin on people and how prevention of exposure to antimicrobial-resistant organisms in products of animal origin can be achieved. The association between the presence of antibiotic-resistant organisms in products of animal origin and infections with these organisms is not clear. This must prompt further studies on such a relationship, particularly where these organisms were thought to be exclusively encountered through nosocomial routes.
The findings of this study highlight the need to identify the source of antibiotic-resistant pathogens in food products of animal origin. This will support the identification of existing policy gaps and weaknesses. The resolution on AMR, adopted at the 39th Food and Agriculture Organization of the United Nations (FAO) conference in June 2016, urged member countries to develop or strengthen national plans, strategies and international collaboration for the surveillance, monitoring and containment of AMR in food, agriculture and the environment, in close coordination with related plans for human health. The management of AMR and food safety must be tackled using the One Health approach, as advocated by the FAO, World Organisation for Animal Health (OIE) (2010) and WHO.