Prevalence and Antimicrobial Resistance of Escherichia coli in Chicken Meat and Edible Poultry Organs Collected from Retail Shops and Supermarkets of North Western Province in Sri Lanka

Department of Livestock and Avian Science, Faculty of Livestock, Fisheries and Nutrition, Makandura 60170, Wayamba University of Sri Lanka, Makandura, Gonawila 60170, Sri Lanka Department of Basic Veterinary Sciences, Faculty of Veterinary Medicine and Animal Science, University of Peradeniya, Peradeniya 20400, Sri Lanka Department of Veterinary Pathobiology, Faculty of Veterinary Medicine and Animal Science, University of Peradeniya, Peradeniya 20400, Sri Lanka Faculty of Food Technology, Universiti Sultan Zainal Abidin, Kuala Terengganu 21300, Malaysia Centre of Excellence for Food Safety Research, Faculty of Food Science and Technology, Universiti Putra Malaysia, UPM Serdang 43400, Seri Kembangan, Selangor DE, Malaysia Centre for Southeast Asian Studies, Kyoto University, Kyoto 606-8501, Japan


Introduction
Escherichia coli (E. coli) is a commensal microorganism that is associated with the gut of warm-blooded animals and humans [1]. E. coli is a member of the family Enterobacteriaceae and a commonly identified faecal coliform. Although not all E. coli are considered pathogenic, some strains of E. coli have the ability to acquire genes encoding virulence factors and be pathogenic to humans [2]. Pathogenic E. coli can cause multiple complications to affected humans, such as prolonged diarrheal disease and vomiting, which may further progress into adverse medical conditions such as traveller's diarrhoea and haemolytic uremic syndrome (HUS) [3]. Pathogenic E. coli are commonly identified as verotoxigenic E. coli (VTEC)-or Shiga toxin (ST)-producing E. coli (STEC) [2].
Chicken meat and edible chicken organs are the most commonly consumed meat source in Sri Lanka [4]. e highest poultry production and poultry meat production in Sri Lanka has been reported from the North Western Province [5], where this research was carried out. Chicken meat and its edible organs are considered a rich source of E. coli. e commonly identified E. coli contamination pathway in humans is known as the faecal oral route [6]. Contamination by the faecal oral route can be minimized by establishing proper hygienic practices and effective surveillance systems, and should be launched for large-scale poultry meat production systems [7]. Pathogenic E. coli is considered a major health concern worldwide [8]. Although a significant E. coli outbreak has not been reported from Sri Lanka, the risk level of causing diarrheal illnesses among the public cannot be neglected.
Antibiotics play a vital role in human and veterinary clinical medicine for the prophylaxis and treatment of numerous bacterial infections, thereby reducing morbidity and mortality in a considerable fraction [9]. However, the misuse of antimicrobial agents during clinical medicine, animal husbandry, and agriculture has generated a selective pressure toward generating multiple drug-resistant microbial populations in the world. Multiple antibiotic resistance (MAR) indexing is a method that is widely used to track the bacterial source. MAR indexing is a cost-effective method that is calculated as the ratio between the number of antibiotics that the organism is resistant to and the total number of antibiotics that the organism is exposed to. If the MAR index values are greater than 0.2, it indicates a high risk of contamination by multidrug-resistant bacteria, where antibiotics are often used [10,11]. e magnitude of the antimicrobial resistance problem is potentially higher in developing countries, where the level of infectious diseases is high and this corresponds to a higher use of broad-spectrum antibiotics [12].
is research was conducted to identify the level of E. coli contamination of chicken meat and edible organs collected from supermarkets and retail shops in the North Western Province and to determine the level of antimicrobial resistance among the collected E. coli bacterial samples.

Sampling Procedure.
A total of 250 chicken meat (n = 144) and edible chicken organs (n = 106) were collected from retail shops (n = 181) and supermarkets (n = 69) from 25 divisional secretariat divisions of the North Western Province of Sri Lanka during the year 2018. A 100 g portion of chicken meat samples, preferably drumstick, chicken thigh or leg pieces, and approximately 100 g of edible chicken organs such as liver, gizzard, and heart, were collected during the sampling process. One sample of chicken meat and/or edible poultry organs was collected from a single shop based on its availability. e sample size for this study was calculated using the following equation described in Naing et al., [13]: (1) where n indicates the sample size; and Z, P, and d indicate Z statistic for a level of confidence (95% for this study), the expected prevalence or proportion (0.8 for this study, according to the pilot study conducted by Anwarama et al., [14]), and precision (0.05 at 95% confidence level), respectively.
Samples for this study were collected by using the random sampling method. Ten random retail shops and supermarkets were selected for the sample collection from each divisional secretariat division, to collect a total of 250 poultry samples.

Bacterial Isolation and Identification.
All samples were collected aseptically and transported isothermally using refrigerated conditions to the laboratory within 4-5 hours of collection. Sampling methods were collected using Bacteriological Analytical Manual standard methods (1998); 10.0 g of the sample was mixed with 90.0 mL of peptone water (HiMedia, Mumbai, India); and the mixture was vortexed for homogenization. Serial dilutions were prepared up to 10 −3 by mixing 1 mL from the original dilution with 9 mL of MacConkey broth (HiMedia, Mumbai, India) in culture tubes. A Durham tube was added to the bottom of the culture tube for the observation of positive gas production. Serial dilutions were then incubated at 37°C for 24 hours. e level of contamination was categorized based on the Bacteriological Analytical Manual (BAM) table standards to understand the risk level of chicken meat and edible organ consumption. Samples with positive E. coli growth were cultured on MacConkey agar (HiMedia, Mumbai, India) plates and incubated at 37°C for 24 hours. Colony morphology and standard biochemical tests were conducted to identify the presumptive colonies of E. coli.

Molecular Biological
Identification. E. coli colonies that were positive on biochemical tests were then subjected to molecular biological identification by polymerase chain reaction (PCR).
e genomic E. coli DNA was extracted using the boiling method as described by Al Gallas et al. [15] with a slight modification. An Eppendorf tube containing one millilitre of an overnight E. coli bacterial culture in peptone water broth (HiMedia, Mumbai, India) was first centrifuged to harvest a bacterial pellet. e resulted bacterial pellet was then washed in 1 mL of distilled water to eliminate the residual media and possible inhibitors present. e bacterial pellet was resuspended in 500 µL of sterile water and boiled at 100°C for 10-12 minutes to perform cell lysis to allow nucleic acids to be released to the lysate. e resulted lysate was centrifuged, and a 100 µL sample of the supernatant (an aliquot) of each bacterial culture was stored at −20°C as a template DNA stock [15]. Quantification of the bacterial DNA concentration was obtained using a Nanodrop spectrophotometer ( ermo Fisher Scientific, USA) at the wavelength of 260 nm for double standard DNA. After measuring the absorbance of DNA sample through a spectrophotometer at 260 nm and 320 nm, the DNA quantity was measured using an equation mentioned in Ref. [16].
2.4. PCR Amplification. PCR amplification was done by using two oligonucleotide primers lacZ3 and YaiO, according to the protocol described in Ref. [17]. e multiplex PCR assay was performed as follows: each 25 µL of reaction mixture contained 2.5 mM deoxynucleoside triphosphates, 10 pmol of each oligonucleotide primers, 2.5 mM of Mg 2+ , 2.5 µL of 10X Taq reaction buffer (50 mM Tris-HCl, pH 7.9, 50 mM KCl, 0.1 mM EDTA, 1 mM DTT, 0.5 mM PMSF, 50% glycerol), 0.2 µL of Taq DNA polymerase (AmpONE ™ Taq DNA polymerase), and 1 µL of template DNA. Samples were amplified for 35 cycles, with each cycle consisting of 30 s at 95°C for denaturation, 30 s at 58°C for primer annealing, and 30 s at 72°C for strand elongation. PCR products were visualized following electrophoresis (80 V, 120 mA current, 40 min) through 1.5% agarose gels (W/V) stained with ethidium bromide, and the amplicons were identified by a 100-bp ladder (Vivantis, Malaysia) based only on the size of the amplified product.

Antimicrobial Susceptibility Test.
Antimicrobial susceptibilities of the isolates to nine common antimicrobial agents were determined by the Kirby-Bauer disc diffusion method for E. coli samples using Mueller Hinton agar (HiMedia, Mumbai, India) by the guidelines of the Clinical Laboratory Standards Institute [18]. Isolates were inoculated onto nutrient agar and incubated at 37°C for 24 hours. A single colony was picked from the nutrient agar plate, and it was suspended in 0.9% saline water and adjusted to give a reading of 0.5 McFarland turbidity standard. A 0.1 mL volume of the 0.5 McFarland suspension was swabbed evenly in at least three directions on Mueller Hinton agar plates. Each plate was left to dry, and the antimicrobial discs for each antimicrobial were placed at a specific place on the agar plates. A sterilized filter paper disc dipped in sterilized distilled water was placed on the centre of each Petri dish as the control disc. Petri plates were placed with antimicrobial discs, and the control was incubated lid side up at 37°C for 24 hours. E. coli ATCC 25922 was used as quality control organism in antimicrobial susceptibility determination. e tested antibiotics (HiMedia, Mumbai, India) were ampicillin (AMP; 25 µg/mL), amoxicillin (AMX; 30 µg/mL), chloramphenicol (C; 30 µg/mL), ceftazidime/clavulanic acid (CAC; 30/10 µg/mL), ciprofloxacin (CIP; 30 µg/mL), cephalexin (CN; 30 µg/mL), erythromycin (E; 15 µg/mL), gentamicin (GEN; 30 µg/mL), and tetracycline (TE; 30 µg/mL). e diameter of the inhibition zone was used as a measurement of the effectiveness of antibiotics against each sample. Isolates were classified as resistant, intermediate, or sensitive based on the Clinical and Laboratory Standards Institute guidelines (CLSI M100-ED30: 2020 Performance Standards for Antimicrobial Susceptibility Testing, 30th Edition (2020)).

Statistical Analysis.
Obtained data related to E. coli presence in chicken meat and edible chicken organs collected from retail shops and supermarkets were transferred to Microsoft Excel spreadsheet (Microsoft Corp., Redmond, WA) for the analysis of E. coli prevalence. Prevalence levels were calculated based on E. coli presence in the collected samples using Microsoft Excel software, and the obtained prevalence data were subjected to a two-proportion Z test using SPSS 16.0 software (SPSS Inc., Chicago, IL). P-values were calculated, and statistical analysis was used to find any significant difference of E. coli prevalence levels between retail shops and supermarkets of North Western Province, Sri Lanka. e P-value less than 0.05 was considered statistically significant.

Results
Among the 250 collected samples, 167 (66.80%) chicken meat and edible organ samples were contaminated with E. coli, and isolated E. coli were confirmed by PCR identification ( Figure 1). e prevalence of E. coli in retail shops and in supermarkets was calculated as 66.85% and 66.67%, respectively. Statistical analysis at 95% confidence level shows the MPN value for E. coli contamination is not significant between the supermarkets and retail shops. 23% of the positive chicken meat samples had higher contamination, which is higher than 1100 MPN/g. According to the overall prevalence records of this research, the highest E. coli contamination was recorded from the chicken meat samples collected from retail shops. Based on the MPN value, edible chicken organs collected from retail shops showed higher contamination (84%) compared to the meat and organ samples collected from supermarkets. Overall prevalence data for chicken meat and edible poultry organs from retail shops and supermarkets are indicated in Tables 1 and 2. e majority of the isolates were resistant to at least one or more antibiotics (Table 3).
Antimicrobial resistance patterns of 166 E. coli isolates showed high rates of resistance to erythromycin (80.84%) and amoxicillin (76.05%). However, the lowest rate of resistance Journal of Food Quality was observed for gentamicin (4.79%) and cephalexin (25.75%). e MAR index of Escherichia coli ranges from 0.1 to 0.8. e current study indicates 82.63% multiple antimicrobial resistance of E. coli, indicating resistance to at least two or more antimicrobial agents. Only 17.37% of the isolates were resistant to two or fewer antimicrobial agents.

Discussion
Faecal contamination of fresh produce is a significant public hazard in the majority of developing countries [19]. Chicken meat and edible chicken organs are the most popular meat types consumed among Sri Lankans [4,20]. ey are mostly consumed fully cooked or sometimes half-cooked (in salads), processed (sausage, luncheon meats, and hot dogs), cured (bacon and ham), and as further processed meat products (nuggets, meat fingers, drumsticks) based on the application [21]. e primary objective of this study was to identify the E. coli prevalence in chicken meat and edible organs available in the North Western Province, Sri Lanka.
is study indicates an alarming risk of high E. coli prevalence in poultry meat and edible organs collected from retail shops and supermarkets, and multiple antimicrobial resistance in E. coli isolated from the North Western Province. ese results are significant as this province has the highest poultry and poultry meat production in Sri Lanka [4].
Although E. coli prevalence in edible chicken organs shows a greater value when compared to E. coli prevalence in chicken meat, the difference between E. coli prevalence in chicken meat and edible organs is not statistically significant. Maintenance of poor sanitation is the major cause of high prevalence of gut microflora such as E. coli in food commodities. It was clear that meat handlers, especially at retail levels were not maintaining an adequate levels of hand washing. A majority of the meat handlers were not using   gloves or meat handlers to handover meat to the customers. Moreover, they were using bear hands to handle meat, and instead of washing their hands, the common practice they had was to wipe their hands with a piece of cloth. Meat handling and selling shops did not maintain adequate cold chains, especially in retail shops. Although cold chains were maintained properly and were using meat handlers or gloves to handle meat in regard to supermarkets, E. coli prevalence in samples collected from supermarkets was still considerably high. However, the sanitation of meat sold in both retail and supermarkets was unclear after the slaughtering step until the vendor level in Sri Lanka. High E. coli contamination in the Sri Lankan scenario from both retail shops and supermarkets may be due to improper food handling [22], poor hygienic practices [22,23], and not maintaining refrigeration or appropriate cold chains [6]. e establishment of good hygienic practices in food handling, maintenance of continuous cold chains for raw material storage at retail levels, and the application of adequate heat treatments to inactivate live pathogens, as well as harmful toxins while cooking are effective remedies against high E. coli prevalence [24]. Ruminants such as cattle and sheep act as active carriers of pathogenic Shiga toxin-producing E. coli (STEC); faecal contamination of water bodies and raw food items by these animals' faeces can lead to a potential E. coli risk to the general public [25,26]. e need for various treatment options to the affected public includes the use of antimicrobial drugs and the high cost of hospitalization. ese issues significantly affect the medical sector of a country.
When E. coli prevalence data are compared with literature findings, it is clear that the prevalence data can be affected due to numerous reasons. Although E. coli is considered a heat-sensitive bacterium that is destroyed with proper cooking methods, enterotoxigenic E. coli (ETEC), enterotoxins, and verotoxins produced by E. coli O157 : H7 can remain in food commodities causing food-borne diseases, a potential public health risk [27,28]. Moreover, the survival of E. coli in raw or undercooked food products such as salads or hamburgers can lead to diarrheal diseases and related complications since they are directly consumed without preliminary heat treatment [29]. Although the presence of E. coli in food commodities does not necessarily emphasize the risk of E. coli leading to an outbreak, the public health risk of contaminated food items cannot be readily ignored until the pathogenicity of the present E. coli is completely ruled out [30]. Outbreaks related to generic  E. coli contamination have also been reported worldwide apart from pathogenic E. coli strains [31]. e results of the current research study show a similar pattern of E. coli significance in raw meat with similar studies conducted worldwide. For example, prevalence study conducted using raw meat collected from retail shops in the United States indicated 72% of E. coli contamination [32] and a second study in Washington D.C. indicated 70.7% [33] prevalence rate. Similar studies conducted in the Czech Republic and Eastern Turkey using poultry meat indicate E. coli contaminations of 63% [34] and 70% [35], respectively. e prevalence percentages described in the current study show similar patterns to the E. coli contaminations in the South Asian studies. Neighbouring countries to Sri Lanka such as India and Bangladesh have conducted similar studies with similar ranges of prevalence scores. An Indian study suggests 77% of E. coli prevalence [36], while two studies conducted in Bangladesh suggest the overall E. coli contamination values of 63.5% in broiler and layer chicken meat [37], and 76.1% prevalence using frozen chicken meat [38].
However, the prevalence of generic E. coli values varies between the ranges of very high to very low based on the differences of countries, regents, and the associated sample type. For example, a study conducted in Mexico City indicates a very high E. coli prevalence of 85% in ready-to-eat (RTE) salads [39]. Conversely, a study associated with E. coli contamination in beef and lamb samples denotes the prevalence as 17.8% and 16.7%, respectively [40]. e MAR value for E. coli in the current study elaborates a greater value of 82.63%.
is indicates a considerable adverse effect in common antibiotic usage. e high prevalence of erythromycin and amoxicillin resistance in E. coli isolates from poultry meat shows the frequent usage of common antibiotics. Antibiotics are used in farm animals and in veterinary medicine as growth promoters [41] to increase feed efficacy, to decrease waste production [42], and to prevent diseases [43]. e highest frequency of resistant pattern (12/167) in all MAR indices is denoted by AMP, AMX, C, CIP, E, TE, with a MAR value of 0.7. e highest MAR frequency of 0.8 was showed by the pattern of AMP, AMX, C, CIP, CN, E, and TE (8/167), indicating possible drug interactions between the above resistant antimicrobial agents.
A MAR value greater than 0.2 indicates a high contamination risk where frequent antibiotics were used. Frequent and unmonitored application of antibiotics gives rise to multidrug resistance in associated bacteria, acquiring antibiotic-resistant genes. On the other hand, antimicrobial residues could affect the persistent normal microflora of host animals, which eventually grow and replicate into antibioticresistant bacteria [44]. E. coli especially has the ability to transfer resistant genes to other bacteria once antimicrobial resistance is acquired [45]. Multidrug-resistant bacteria may transfer antimicrobial resistant genes to other bacteria in human intestinal microflora resulting resistance acquired zoonoses [46]. E. coli shows a specific mechanism of MAR through the production of extended-spectrum β-lactamases (ESBL) [9,47], which makes E. coli infections even more difficult to treat with fluoroquinolones, aminoglycosides, and trimethoprim-sulfamethoxazole. ESBL-producing E. coli are capable of inactivating such antibiotics due to the resistant mechanisms present [48]. e global burden of MAR has reached unacceptable levels due to the wide consumption of inexpensive and outdated antimicrobial agents [12,49]. If necessary actions have not been taken in the near future, this usage pattern will show an increasing trend. e weight of the MAR problem is considerably high in low-income countries, where inexpensive antimicrobials are often used to treat a wide range of nosocomial diseases when compared to many other developed countries [50]. Several studies in the literature suggest the presence of MAR and a relatively similar level of MAR levels to support the current research findings in E. coli isolated from poultry meat. Reference [37] shows the presence of 78.06% of MDR in Bangladesh, with ampicillin, erythromycin, and tetracycline resistance (98.95%, 89.5%, and 85.3%, respectively). Amoxicillin and ampicillin resistance among humans is found in most of the cases, and it is due to the presence of plasmid-encoded β-lactamases, such as TEM-1, TEM-2, or SHV-1, where they can hydrolyse and inactivate amoxicillin and ampicillin drugs [51]. Previous studies suggest antimicrobial resistance by extendedspectrum cephalosporin (ESC)-resistant E. coli showing coresistance to ampicillin, amoxicillin/clavulanic acid, chloramphenicol, and tetracycline [52]. Another study suggests that nearly 75% of ampicillin-resistant E. coli were also resistant to tetracycline [53]. Higher resistance to amoxicillin [54], ampicillin [54], tetracycline [55], and chloramphenicol has also been reported from E. coli specimens isolated from poultry samples, due to increased usage of the above drugs in commercial poultry [56]. High resistance to ciprofloxacin around the world may be due to the increased use of quinolones in poultry farms, and this result supports the possible reasons for the high resistance to ciprofloxacin in human specimens to be due to the colonization of ciprofloxacin-resistant E. coli in the human gut [56]. Ciprofloxacin resistance of the family Enterobacteriaceae has become a growing concern in clinical settings around the world [54]. Antimicrobial resistance to gentamicin may be due to the use of apramycin, a commonly used veterinary medicine that is structurally similar to gentamycin. ese apramycin-resistant E. coli may be resistant to gentamycin as well [56]. Resistance to cephalosporin antibiotics is mainly associated with extensive usage of cephalosporins (cephalexin, ceftazidime) in clinical practice [57]. E. coli isolated from faecal-contaminated water bodies show significantly higher resistance to erythromycin [54], which indicates a high risk of erythromycin-resistant E. coli infections after the consumption of contaminated water from such water bodies. However, previous studies suggest that the level of resistance observed by E. coli isolates from human samples indicates comparatively low resistance compared to the E. coli isolates collected from poultry and animal samples [56,58].
Although the usage of antibiotics as growth promoters in animal feed has been banned in many countries around the world (EU since 2016), some countries still use common antimicrobial agents to promote growth in farm animals [59]. Many of the commonly used antibiotics are freely accessible in several countries and can be purchased over the counter without a prescription [60]. e convenience of obtaining antimicrobial drugs without a prescription over the counter makes it easy for the general public to misuse antimicrobial drugs to a greater extent.
As MDR problem is an alarming issue that needs immediate effective approaches to prevent further generation of MDR genes, several initiatives could be taken to reduce the acceleration of the AMR problem. e usage of narrowspectrum antibiotics [61], prudent use of antibiotics in veterinary medicine and animal feed [62], use of antimicrobial agents in a rotation pattern [63], use of combined antimicrobial therapy in clinical medicine [63], administration of proper veterinary practices to avoid the usage of antimicrobials [61], and antibiotic regulatory approaches and monitoring systems [62] are few such actions to be launched immediately. e prevention of misusing antibiotics by the general public and enabling antimicrobial surveillance programmes, especially in nosocomial infections, are mandatory actions that need to be taken to reduce the spread of antibiotic-resistant bacteria [44]. Multiple antimicrobial resistance in E. coli isolated from poultry meat and edible poultry organs may be due to the linkage of resistant genes in their plasmids [30,63]. Further studies are required to identify the genetic basis of coresistant phenotypes.

Conclusion
Findings from this research project emphasize the need for adequate heat treatment prior to consuming raw chicken meat and edible chicken organs available in the Sri Lankan market and the importance of prudent antibiotic usage during veterinary medicine and farm animals to stop the further spread of the present unacceptable levels of antimicrobial resistance further.
A limitation of this study was that it has only focussed on the prevalence and antimicrobial resistance in generic E. coli, and the pathogenicity of the collected E. coli was not tested. As the North Western Province has the highest poultry production in Sri Lanka and operates a large number of commercial poultry farms compared to other provinces, there is a possibility that the current study results may have slightly exaggerated the prevalence and MAR values in comparison with the wider poultry farming sector in Sri Lanka.

Further Approaches.
Pathogenicity of the collected E. coli isolates needed to be tested for pathogenicity identification. PCR confirmed the samples needed to be tested for phylogenetic analysis. e prevalence of pathogenic E. coli strains should also be tested for other types of meat as well as fresh produce associated with the wider food industry in Sri Lanka.

Data Availability
e data collected to support this study are included within the article.

Journal of Food Quality
Disclosure A small part of this manuscript has been published as conference abstracts in "7th Conference on Sri Lanka-Japan Collaborative Research-SLJCR 2019," also funded by National Research Council grant no. 17-008.

Conflicts of Interest
e authors declare that there are no conflicts of interest in this study.