Multi-drug resistant extended-spectrum beta-lactamase producing E. coli and Salmonella on raw vegetable salads served at hotels and restaurants in Bharatpur, Nepal

Objective Antimicrobial resistance among the bacteria present in ready-to-eat foods like vegetable salads is an emerging concern today. The current study was undertaken to investigate the presence of multi-drug resistant extended-spectrum β-lactamase (ESBL) producing E. coli and Salmonella spp. in raw vegetable salads served at hotels and restaurants in Bharatpur. A total of 216 salad samples were collected from three different grades of hotels and restaurants and examined for the presence of E. coli and Salmonella spp. in Microbiology laboratory of Birendra Multiple Campus by conventional microbiological techniques. Results Out of 216 samples, 66 samples (35.2%) showed the presence of Salmonella spp. whereas E. coli was recovered from 29 (13.4%) samples of which 3 samples harbored E. coli O157: H7. Antibiotic susceptibility testing revealed that 9 (13.6%) Salmonella and 4 (13.8%) E. coli isolates were detected as multi-drug resistant. Total ESBL producers reported were 5 (7.57%) Salmonella and 4 (13.8%) E. coli. The study also assessed a significant association between occurrence of E. coli and Salmonella with different grades of hotels and restaurants, personal hygiene and literacy rate of chefs and with the type of cleaning materials used to wash knives and chopping boards (p < 0.05). The findings suggest an immediate need of attention by the concerned authorities to prevent the emergence and transmission of food-borne pathogens and infections antimicrobial resistance among them.


Introduction
Salad is any food preparations made up of a mixture of chopped or sliced vegetables that provides a good source of minerals and dietary fiber of low fat and calories to the consumer [1]. Salad vegetables are consumed without any form of heat treatment, sometimes without washing, thus, has the possibility of causing food-borne diseases [2]. Some previous studies have revealed the presence of E. coli, E. coli O157: H7, Salmonella spp., Enterobacter aerogenes, Pseudomonas spp., Klebsiella spp., Providencia spp., Listeria monocytogenes and Proteus spp. from vegetable salads [3][4][5][6].
Drug resistance is one of nature's not ever ending processes which may be due to a pre-existing factor in the organisms or results from the acquired factors [7]. The enteric bacteria in fecal flora are often stated to be highly resistant and E. coli is reported to be the main transporter of antimicrobial resistance [8]. Beta-lactamase production is the most important mechanism of resistance to penicillins and cephalosporins [9]. Bacterial species that hold extended-spectrum beta-lactamase (ESBL) genes reside in the gastrointestinal tract and food is the key route of them [10]. To the best of our knowledge, no any study has yet been made in Nepal that endeavors to examine the bacterial quality of green salads. Hence this study was carried out to screen multi-drug resistant extended spectrum beta-lactamase producing E. coli and Salmonella on raw vegetable salads served at hotels and restaurants in Bharatpur.

Study design and setting
A descriptive cross-sectional study was carried out for 5 months from November 2017 to March 2018 in Bharatpur metropolitan city. The salads samples were collected using random sampling methods without repetition from different grades of hotels and restaurants in the city. The sample size was determined in accordance with the prevalence rate based on the previous study [4]. A total of 216 salad samples were included in this study.

Methodology
A total of 216 ready-to-eat salad samples (72 radish, 72 carrots and 72 cucumbers) were collected in sterile ziplocked plastic bags from three different grades of hotels and restaurants in Bharatpur city and transported aseptically to the Microbiology laboratory of Birendra Multiple Campus within 1 h for further analysis. The hotels and restaurants were categorized into three grades: A, B and C based on the cost, economic class of the customers and number of staffs involved in serving [11]. Twenty-five gram of salad sample was mixed with 225 ml sterile water and left covered for 30 min. Then, 1 ml of the sample was inoculated in buffered peptone water and Selenite F broth and incubated at 37 °C for 24 h. One loopful of the culture was streaked on Eosine Methylene Blue agar and Xylose Lysine Deoxycholate agar for the identification of E. coli and Salmonella spp. respectively and incubated at 37 °C for 24 h. The suspected colonies after incubation were sub-cultured on Nutrient agar. Identification of bacterial isolates was carried out based on their cultural, morphological and biochemical characteristics [12]. For the preliminary identification of E. coli O157: H7, E. coli isolates were streaked on the MacConkey Sorbitol agar and incubated at 37 °C for 24 h and colorless colonies were suspected as E. coli O157: H7. Later, the slide agglutination test using anti-O157 and flagellar H7 serum was done (Defico, USA) to confirm E. coli 0157:H7 strains.
Antibiotic susceptibility tests (AST) were performed by modified Kirby Bauer's disc-diffusion method [13]. Antimicrobial discs amoxicillin (10 µg), cotrimoxazole (25 µg), ciprofloxacin (5 µg), gentamicin (10 µg) and azithromycin (15 µg) for Salmonella and in case of E. coli, gentamicin (10 µg), ciprofloxacin (5 µg), cotrimoxazole (25 µg), ampicillin (10 µg) and chloramphenicol (30 µg) were used in the study. Resistance shown to at least three or more antibiotics of different structural classes was considered MDR as described elsewhere [14,15]. Screening of ESBL was performed by using ceftazidime (30 µg) and cefotaxime (30 µg) discs. The zone of inhibition (ZOI) ≤ 22 mm for ceftazidime and ≤ 27 mm for cefotaxime was considered as potential ESBL producer as recommended by Clinical and Laboratory Standards Institute (CLSI). Possible ESBL producers were subjected to combined disc test for phenotypic confirmation as recommended by CLSI [16]. The combination of ceftazidime and cefotaxime alone and in combination with clavulanic acid (CA) (10 µg) were used for the confirmation of ESBL producing isolates. An increase ZOI of ≥ 5 mm for either antimicrobial agent tested in combination with CA versus its zone when tested alone was interpreted as positive for ESBL production. E. coli ATCC 25922 and Salmonella Typhimurium ATCC 14028 were used as reference strains for quality control.
Raw data obtained in this study were tabulated in SPSS V.20 and Chi square test was performed. p ≤ 0.05 was assigned as significant.

Distribution of E. coli and Salmonella in salads
A total number of 216 salad samples were collected and analyzed for the presence of Salmonella and E. coli. Salmonella was isolated from 66 (30.56%) samples while E. coli was isolated from 29 (13.43%) samples, out of which 3 (10.34%) were found to be E. coli O157: H7.

Association of various variables with the rate of bacterial contamination
A total of 216 salad samples were analyzed for the presence of E. coli and Salmonella. Salads served in Grade C restaurants showed higher contamination of E. coli (20.83%) and Salmonella (41.64%) as compared to Grade A and B types (p ≤ 0.05). Similarly, salads prepared by the illiterate chefs were found more prone to contamination by E. coli (44%) and Salmonella (66.7%) (p ≤ 0.05). The high incidence of contamination by E. coli (16.67%) and Salmonella (32.8%) were observed in the salads prepared by the non-gloved chefs (p ≤ 0.05). Chopping boards and knives washed by water showed higher contaminations of Salmonella (44.4%) in comparison to those washed by detergents and soaps (p ≤ 0.05) ( Table 1).

Antibiotic susceptibility pattern
AST was also performed for all the E. coli and Salmonella isolates against different antibiotics like ampicillin, cotrimoxazole, chloramphenicol, ciprofloxacin, azithromycin, amoxicillin and gentamicin. More than half of the E. coli isolates (58.60%) were found to be resistant to ampicillin. Gentamicin was the most effective antibiotic as 96.60% of the isolates were found to be sensitive to it followed by cotrimoxazole killing 93.20% of E. coli isolates. The least effective antibiotic was ampicillin which inhibited the growth of 41.40% of E. coli isolates. Out of 29 isolates of E. coli, 4 (13.80%) were MDR. Likewise, all Salmonella isolates were found to be resistant to amoxicillin. Cotrimoxazole was the most effective antibiotics as 97% of Salmonella isolates were found to be sensitive to it followed by gentamicin killing 95.5% of Salmonella isolates. The least effective antibiotics was ciprofloxacin which inhibited the growth of only 4% Salmonella isolates followed by amoxicillin to which all the isolates were resistant. Out of 66 isolates of Salmonella subjected to AST, 19 (28.78%) were MDR ( Table 2).

Frequency of ESBL producing E. coli and Salmonella isolates
Out of 95 isolates, 12 isolates gave ESBL screening test positive. Four isolates of E. coli and 5 isolates of Salmonella accounting a total of 9 (9.47%) were further confirmed as ESBL producers (Table 3).

Discussion
A total of 216 salad samples were analyzed in the present study, out of which 13.45% of samples were contaminated by E. coli. A similar study was performed in Kashan city of Iran which reported E. coli in 85% salad samples [17]. In a study from India, E. coli was found in 38.3% of different vegetable salads [18]. These two studies show a higher level of contamination by E. coli than our study. Similarly, the prevalence of Salmonella reported in this study was 30.6%. It is lower than the study done in India, which   reported only 3.33% of vegetables salads contaminated with Salmonella spp. [19]. A similar study was conducted in Nigeria which reported the presence of 25.0% salad samples contaminated by Salmonella [20]. A higher incidence Salmonella might be associated with the increased thermal resistance of Salmonella in chicken litter applied in the field [21]. In this study, E. coli O157: H7 have been reported in 1.4% salad samples. A similar study reported E. coli O157: H7 in 1.3% (6/480) of the raw salad vegetables [22]. Vegetables grown in soil fertilized by animal manure have a greater chance to be contaminated with E. coli O157: H7 [23]. A higher number of salad samples harboring Salmonella (41.64%) and E. coli (20.83%) were isolated from the Grade C restaurants compared to Grade A and B in the current study (p ≤ 0.05). This might be due to the fact that chefs were unaware of health as well as personal hygiene, and water used for washing salad samples might not be disinfected properly in Grade C types. Distribution of E. coli and Salmonella also vary based on the literacy rate of the chefs [24]. Our study revealed a higher incidence of E. coli (44%) and Salmonella (66.7%) in salads made by illiterate chefs (p ≤ 0.05). Illiterate chefs might lack the knowledge of consumer's health and they might cross-contaminate the salads in the course of preparation. Personal hygiene of the chefs may aid in the variation in the distribution of Salmonella isolates [25]. Salads prepared by the non-gloved chefs showed a high rate of contamination-Salmonella (32.8%) and E. coli (16.67%) (p ≤ 0.05). This finding is in tune with a work which revealed that non-gloved vendors served highly contaminated ready to eat foods than gloved vendors (p < 0.01) [26]. Contaminated hands can easily transmit the pathogens like Salmonella and E. coli O157H7 in foods [27]. Obviously, non-gloved chefs might carry dirt and dust particles which include the higher number of microbes that can contaminate the salads while preparing. Wide fluctuation in the prevalence of Salmonella in salads can be observed in the cleaning trend of chopping board and knife [28]. More isolates of Salmonella (44.4%) was found in salads cut in the chopping boards and knives which are regularly washed by soaps and detergents (p ≤ 0.05). A study showed a significant coliform reduction (p ≤ 0.05) for all the methods of disinfection. Disinfection processes can actively reduce the bacterial loads from inanimate surfaces, so a significant coliform reduction can be achieved [29,30].
On microbial testing, cotrimoxazole and gentamicin were the most effective antibiotics for both E. coli and Salmonella isolates. Amoxicillin was the most ineffective antibiotic as 100% of the Salmonella isolates were found to be resistant to it. A similar study conducted in Nigeria found all the Salmonella isolates from salads were resistant to amoxicillin [20]. This could due to the easy hydrolysis of the β-lactam ring by Salmonella as well as the frequent usage of the drug due to its low cost leading to the development of resistance by most bacteria.
In this study, 9 (13.6%) Salmonella isolates and 4 (13.8%) E. coli were found to be MDR. The prevalence of MDR in Salmonella and E. coli has been reported comparatively higher in some other studies. A study in major markets in Chittagong of Bangladesh reported that all the Salmonella and E. coli isolates from green salads were MDR [31]. Moreover, another similar work conducted in the same city, Chittagong, found that 48.2% of isolates were MDR [11]. The problem of antibiotics resistance has been increasing day by day and it has become a serious burden in the medical world. The frequency of ESBL producers for Salmonella was 5 (7.57%) and E. coli was 4 (13.8%) in the present study. A study conducted by Raphael et al. observed an ESBL incidence rate of 2.3% among bacterial isolates from spinach [32]. A similar study was conducted by Bezanson et al. who detected 1.9% of ESBL producers from lettuce [33]. These two studies revealed a low prevalence of ESBL producers than our work. Overconsumption and haphazard use of beta-lactam antibiotics might have contributed to the emergence of ESBL producers.

Conclusions
The results of this study reveal that the safety aspect of fresh vegetable salads served at the hotels and restaurants in Bharatpur city is unacceptable from microbiological point of view. Poor personal hygiene of the chefs, improper handling and lack of knowledge on consumer health are some of the factors the concerned authorities should pay immediate attention to control the transmission of food-borne pathogens and emergence of antimicrobial resistance among them.

Limitations
The study only determines the prevalence of E. coli and Salmonella species and their MDR pattern including ESBL production. The source of contamination of these items was not assessed however. The present study also lacks the molecular characterization of the isolates. Other isolates besides E. coli and Salmonella were not included in the study. Food safety knowledge of the chefs was not assessed. Similarly, the bacterial quality of water used for the preparation of salads was not tested.