**Systematic Review and Meta-Analysis of Integrated Studies on Salmonella and Campylobacter Prevalence, Serovar, and Phenotyping and Genetic of Antimicrobial Resistance in the Middle East—A One Health Perspective**

**Said Abukhattab 1,2,\*, Haneen Taweel <sup>3</sup> , Arein Awad <sup>3</sup> , Lisa Crump 1,2 , Pascale Vonaesch <sup>4</sup> , Jakob Zinsstag 1,2 , Jan Hattendorf 1,2 and Niveen M. E. Abu-Rmeileh <sup>3</sup>**


**Citation:** Abukhattab, S.; Taweel, H.; Awad, A.; Crump, L.; Vonaesch, P.; Zinsstag, J.; Hattendorf, J.; Abu-Rmeileh, N.M.E. Systematic Review and Meta-Analysis of Integrated Studies on Salmonella and Campylobacter Prevalence, Serovar, and Phenotyping and Genetic of Antimicrobial Resistance in the Middle East—A One Health Perspective. *Antibiotics* **2022**, *11*, 536. https://doi.org/10.3390/ antibiotics11050536

Academic Editor: Piera Anna Martino

Received: 30 March 2022 Accepted: 17 April 2022 Published: 19 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Abstract: Background:** *Campylobacter* and *Salmonella* are the leading causes of foodborne diseases worldwide. Recently, antimicrobial resistance (AMR) has become one of the most critical challenges for public health and food safety. To investigate and detect infections commonly transmitted from animals, food, and the environment to humans, a surveillance–response system integrating human and animal health, the environment, and food production components (iSRS), called a One Health approach, would be optimal. **Objective**: We aimed to identify existing integrated One Health studies on foodborne illnesses in the Middle East and to determine the prevalence, serovars, and antimicrobial resistance phenotypes and genotypes of *Salmonella* and *Campylobacter* strains among humans and foodproducing animals. **Methods**: The databases Web of Science, Scopus, and PubMed were searched for literature published from January 2010 until September 2021. Studies meeting inclusion criteria were included and assessed for risk of bias. To assess the temporal and spatial relationship between resistant strains from humans and animals, a statistical random-effects model meta-analysis was performed. **Results**: 41 out of 1610 studies that investigated *Campylobacter* and non-typhoid *Salmonella* (NTS) in the Middle East were included. The NTS prevalence rates among human and food-producing animals were 9% and 13%, respectively. The *Campylobacter* prevalence rates were 22% in humans and 30% in food-producing animals. The most-reported NTS serovars were *Salmonella* Enteritidis and *Salmonella* Typhimurium, while *Campylobacter jejuni* and *Campylobacter coli* were the most prevalent species of *Campylobacter*. NTS isolates were highly resistant to erythromycin, amoxicillin, tetracycline, and ampicillin. *C. jejuni* isolates showed high resistance against amoxicillin, trimethoprim–sulfamethoxazole, nalidixic acid, azithromycin, chloramphenicol, ampicillin, tetracycline, and ciprofloxacin. The most prevalent Antimicrobial Resistance Genes (ARGs) in isolates from humans included tetO (85%), Class 1 Integrons (81%), blaOXA-61 (53%), and cmeB (51%), whereas in food-producing animals, the genes were tetO (77%), Class 1 integrons (69%), blaOXA-61 (35%), and cmeB (35%). The One Health approach was not rigorously applied in the Middle East countries. Furthermore, there was an uneven distribution in the reported data between the countries. **Conclusion**: More studies using a simultaneous approach targeting human, animal health, the environment, and food production components along with a solid epidemiological study design are needed to better understand the drivers for the emergence and spread of foodborne pathogens and AMR in the Middle East.

**Keywords:** Middle East; One Heath; antimicrobial resistance; foodborne pathogens; *Campylobacter* spp.; *Salmonella* spp.; systematic review; meta-analysis

#### **1. Introduction**

*Campylobacter* spp. and *Salmonella* spp. are the leading causes of foodborne diseases worldwide [1,2]. According to a report published by the World Health Organization (WHO) in 2018, the global burden of food-borne illnesses is 1 in 10 individuals each year [3]. Annually, non-typhoid *Salmonella* (NTS) is responsible for more than 155,000 annual deaths and 94 million annual cases worldwide [4]. *Campylobacter* infection is a public health problem, causing about 8% of global diarrheal cases [5]. Since 2005, *Campylobacter* has been the most reported gastrointestinal bacterial pathogen in humans in the European Union (EU) [6,7].

The Middle East region has the third-highest prevalence of foodborne illness, with 100 million people estimated to be ill from foodborne illnesses each year. Norovirus, *Escherichia coli*, *Campylobacter*, and NTS are responsible for 70% of all foodborne diseases in the Middle East region [8]. The incidence rate of NTS among Jordanians was 124 per 100,000 in 2003–2004 and 30 per 100,000 among Israelis in 2009 [9,10]. In addition, *Campylobacter* was identified in 61% of children with dysentery (63/99) in Israel, 33% (76/230) in Iran 4.7% (7/150) in Palestine, and 3.7% (13/356) in Egypt during the period 2005– 2015 [11–14].

Antimicrobial resistance (AMR) is a major public health concern mainly resulting from the use and misuse of antimicrobial agents. AMR occurs when bacteria, fungi, parasites, and viruses change over time and are no longer susceptible to medicines, making infections difficult to treat and increasing the risk of spreading the infection, intensifying the severity of the disease, and raising death rates [15,16]. After the bacteria has acquired resistance, AMR disseminates by clonal spreads of the bacteria and horizontal gene transfer (HGT), that is, by integrons or plasmids, leading to the accumulation of antimicrobial resistance genes (ARGs) in pathogenic and non-pathogenic bacteria within an individual organism [16]. Rising antimicrobial use contributes to the sharing of resistant bacteria and resistance genes between food animals and humans through the food production chain [15]. AMR in *Campylobacter* spp. and *Salmonella* spp. has been shown to be directly associated with antimicrobial use in animal production. Food-borne diseases caused by these resistant bacteria are well documented in humans [15].

Since humans and animals are in close contact and are intricately interconnected, food safety and AMR are fundamental One Health issues [17,18]. However, most of the current research in low- and middle-income countries (LMICs) focuses on human or animal health risks separately and only a few studies have been conducted to understand the problem in an interconnected manner [19,20]. Additional components of human and animal health must be incorporated to make significant progress in reducing many foodborne diseases [19].

The Joint Programming Initiative on Antimicrobial Resistance (JPIAMR) (www.jpiamr. eu, accessed on 19 March 2022) identified several critical knowledge gaps. First, the relative contributions of different sources of antibiotics and antibiotic-resistant bacteria into the environment are unmeasured. Second, the role of the environment, particularly the anthropogenic inputs, on the evolution of resistance is not understood. Third, the overall human and animal health impacts caused by exposure to resistant bacteria from the environment have not been studied. Finally, the efficacy of technological, social, economic, and behavioral interventions to mitigate environmental antibiotic resistance have not been evaluated [21]. A recent review of integrated studies on antimicrobial resistance in Africa concluded that data on AMR from a One Health perspective in Africa are scarce with only 18 studies meeting the minimal standards of addressing simultaneously at least two of the environment–animal–human realms [16].

This systematic review and meta-analysis aims to summarize the scientific literature published between January 2010 and August 2021 on the prevalence, serovars, and antimicrobial resistance phenotypes and genotypes (ARGs) of *Salmonella* and *Campylobacter* strains from integrated studies, studying at the same time humans and food-producing animals and their products in the Middle East region. In addition, it attempts to address the

knowledge gap and summarize the available information about the situation by applying the integrated studies to follow up *Salmonella* spp. and *Campylobacter* spp. as the leading foodborne illnesses in the Middle East.

#### **2. Methodology**

The protocol for this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO ID: CRD42021277400).

#### *2.1. Search Strategy*

We conducted a systematic search on PubMed, Web of Science, and Scopus, limiting the search to the literature published from 2010 until 30 September 2021. Two reviewers performed the initial search, abstract screening, and data extraction, and any discordances were solved by a third reviewer. The exact search strategy used for each database is included in Supplementary Table S1.

#### *2.2. Inclusion and Exclusion Criteria*

We aimed to analyze the available information about prevalence, serovar distribution, and antimicrobial resistance phenotypes and genotypes of *Salmonella* and *Campylobacter* strains among humans and food-producing (terrestrial) animals and their products in the Middle East region. It included all peer-reviewed literature published from 1 January 2010, until 30 September 2021. The search included only studies that were published in English. We excluded publications published before 2010, grey literature, non-peer-reviewed literature, and studies with a different design than cross-sectional, cohort studies, and studies using survey system data (Routine data). In addition to information on Salmonella spp. and Campylobacter spp. isolates originating from companion animals, plant-based food, aquatic products (fish), water sources, and concerning *Salmonella* enterica serotypes Typhi and Paratyphi.

#### *2.3. Study Selection*

Two independent reviewers used Covidence software (www.covidence.org, accessed on 23 September 2021) for the title and abstract screening. Studies that were eligible for fulltext review were further reviewed. Subsequently, risk assessment and data extraction were undertaken. Disagreements between reviewers in the title and abstract screening or full-text review were resolved through consultation with a third reviewer.

#### *2.4. Data Extraction*

Two independent reviewers extracted the data for the included papers, and the required data was entered into an Excel (Microsoft Inc.TM, Redmond, WA, USA) sheet. Data included author, publication year, year of data collection, collection country, study outcomes, study design, the validity and reliability of the study methodology, as well as details available regarding analysis, human and animal sample sizes, sample sources, isolated bacteria source, and prevalence. In addition, data regarding serotype prevalence, AMR gene prevalence, and NTS and *Campylobacter* AMR profiles were collected.

#### *2.5. Risk of Bias Assessment*

We used the risk of bias tool developed by Hoy et al., 2012 [22] to assess the overall quality of the papers. Two independent reviewers performed the risk of bias assessment and disagreements were solved by consensus.

#### *2.6. Data Synthesis*

The number of studies remaining at each stage of the selection process is summarized in the flowchart in Figure 1.

The pooled prevalence rate of *Salmonella* spp. and *Campylobacter* spp. and their main serotypes for human and food-producing animals (live animals and products) were calculated separately based on the following Equation (1): in the flowchart in Figure 1. The pooled prevalence rate of Salmonella spp. and Campylobacter spp. and their main serotypes for human and food-producing animals (live animals and products) were calculated separately based on the following Equation (1):

The number of studies remaining at each stage of the selection process is summarized

Antibiotics 2022, 11, x FOR PEER REVIEW 4 of 19

2.6. Data Synthesis

$$\begin{aligned} \text{Predmediate rate} &= \frac{\text{No. of public batch}}{\text{Total number of collected samples}} \\ \text{MOR} & \text{ profile among NTS and } \odot \text{ joints was calculated using Equation (2)} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ &= \text{MOR} \times \text{Number of students} \\ \end{aligned}$$

#### 2.7. Statistical Analysis *2.7. Statistical Analysis*

Relative risks were assessed based on the total number of samples and the number of NTS, Campylobacter spp., and AMR positive samples (phenotype and genotype). Studies were stratified by bacterial species and sources. A pooled risk ratio (RR) was calculated Relative risks were assessed based on the total number of samples and the number of NTS, *Campylobacter* spp., and AMR positive samples (phenotype and genotype). Studies were stratified by bacterial species and sources. A pooled risk ratio (RR) was calculated separately for each bacterial species. The *I* <sup>2</sup> and *r* 2 statistics assessed heterogeneity. We exclusively used the random-effects model, irrespective of the heterogeneity results. For all statistical analyses, we used the R software environment version 4.0.3 and the "meta-

package" version 4.14-0. We used the function 'metabin' using the Mantel–Haenszel method with inverse variance weighting for pooling [23]. method with inverse variance weighting for pooling [23]. 3. Results

2 and r 2

exclusively used the random-effects model, irrespective of the heterogeneity results. For all statistical analyses, we used the R software environment version 4.0.3 and the "metapackage" version 4.14-0. We used the function 'metabin' using the Mantel–Haenszel

statistics assessed heterogeneity. We

#### **3. Results** 3.1. Studies Identified and Included in the Final Analysis

#### *3.1. Studies Identified and Included in the Final Analysis* Based on the eligibility criteria, a total of 2534 publications were identified. After re-

separately for each bacterial species. The I

Antibiotics 2022, 11, x FOR PEER REVIEW 5 of 19

Based on the eligibility criteria, a total of 2534 publications were identified. After removing duplicates, we screened 1610 abstracts of which 565 were eligible for full-text screening. Out of 565 articles, 41 studies met the inclusion criteria for this meta-analysis (Figure 1). In total, 31 studies used a cross-sectional study design, and 10 studies used routine data (Supplementary Table S2). moving duplicates, we screened 1610 abstracts of which 565 were eligible for full-text screening. Out of 565 articles, 41 studies met the inclusion criteria for this meta-analysis (Figure 1). In total, 31 studies used a cross-sectional study design, and 10 studies used routine data (Supplementary Table S2). The overall result of the risk assessment that was conducted for the included studies

The overall result of the risk assessment that was conducted for the included studies indicated that the majority of studies had an overall low risk of bias, and none of the papers had a high risk of bias. This result was based on the risk of bias assessment using the Hoy et al., 2012 tool [22]. indicated that the majority of studies had an overall low risk of bias, and none of the papers had a high risk of bias. This result was based on the risk of bias assessment using the Hoy et al., 2012 tool [22].

#### *3.2. Overview of the Selected Studies* 3.2. Overview of the Selected Studies A total of 16 countries were included in this literature review: Qatar, United Arab

A total of 16 countries were included in this literature review: Qatar, United Arab Emirates, Bahrain, Saudi Arabia, Kuwait, Israel, Oman, Iran, Jordan, Lebanon, Palestine, Syria, Yemen, Turkey, Iraq, and Egypt. Of these, nine countries had no published literature matching the study inclusion criteria available (Qatar, United Arab Emirates, Bahrain, Saudi Arabia, Kuwait, Oman, Syria, Yemen, and Iraq), while seven countries had at least one article available (Egypt, Iran, Turkey, Lebanon, Israel, Jordan, and Palestine). Emirates, Bahrain, Saudi Arabia, Kuwait, Israel, Oman, Iran, Jordan, Lebanon, Palestine, Syria, Yemen, Turkey, Iraq, and Egypt. Of these, nine countries had no published literature matching the study inclusion criteria available (Qatar, United Arab Emirates, Bahrain, Saudi Arabia, Kuwait, Oman, Syria, Yemen, and Iraq), while seven countries had at least one article available (Egypt, Iran, Turkey, Lebanon, Israel, Jordan, and Palestine). Of the included studies, 26 (63.41%) were conducted in Egypt, 8 in Iran (19.51%), 2 in

Of the included studies, 26 (63.41%) were conducted in Egypt, 8 in Iran (19.51%), 2 in Turkey (4.88%), 2 in Lebanon (4.88%), 1 in Israel (2.44%), 1 in Jordan (2.44%), and 1 in Palestine (2.44%) (Figure 2a). Of these, 17 reports (42%) included data about *Salmonella* spp. (the number of reports used for *Salmonella* spp is the same as that used for NTS), and 8 reports (20%) had data about *Campylobacter* spp. In addition, some articles focused on one of *Salmonella* and *Campylobacter* serovars; *Campylobacter jejuni* (nine reports, 22%), *Campylobacter coli* (one report, 2%), *Salmonella* Enteritidis (four reports, 10%), *Salmonella* Typhimurium (one report, 2%), and *Salmonella* Heidelberg (one report, 2%) (Figure 2b) and (Supplementary Table S3). Turkey (4.88%), 2 in Lebanon (4.88%), 1 in Israel (2.44%), 1 in Jordan (2.44%), and 1 in Palestine (2.44%) (Figure 2a). Of these, 17 reports (42%) included data about Salmonella spp. (the number of reports used for Salmonella spp is the same as that used for NTS), and 8 reports (20%) had data about Campylobacter spp. In addition, some articles focused on one of Salmonella and Campylobacter serovars; Campylobacter jejuni (nine reports, 22%), Campylobacter coli (one report, 2%), Salmonella Enteritidis (four reports, 10%), Salmonella Typhimurium (one report, 2%), and Salmonella Heidelberg (one report, 2%) (Figure 2b) and (Supplementary Table S3).

Figure 2.: Number of studies (a) per country and (b) per pathogen. **Figure 2.** Number of studies (**a**) per country and (**b**) per pathogen.

#### *3.3. Prevalence and Serotype Distribution of Salmonella spp. and Campylobacter spp. among Humans and Food-Producing Animals*

Out of 41 eligible articles, 31 were cross-sectional studies. We used the cross-sectional data to calculate the prevalence rate for each pathogen separately. Of the 1317 human

samples, 167 (13%) were positive for *Salmonella* spp. (14% in diarrhea patients and 9% in high-risk population). In food-producing animals, out of 3520 samples, 585 (17%) were positive for *Salmonella* spp. (31% in poultry and poultry products and 4% in ruminants and ruminant products). Moreover, NTS was reported with a prevalence of 9% (109/1167) in humans (10% in diarrhea patients and 6% in high-risk populations) and 13% (352/2718) in food-producing animals (33% in poultry and poultry products and 4% in ruminants and ruminant products). The two most common NTS serovars were *S.* Typhimurium with a prevalence of 5% (36/780) in humans (4% in diarrhea patients and 6% in highrisk populations) and 3% (91/3038) in food-producing animals (7% in poultry and poultry products and 0.6% in ruminants and ruminant products) and *S.* Enteritidis with a prevalence of 2% (12/585) in humans (2% in diarrhea patients and 2% in high-risk population) and 3% (87/2534) in food-producing animals (9% in poultry and poultry products and 0.3% in ruminants and ruminant products) (Tables 1 and 2).

*Campylobacter* spp. was reported with a prevalence of 22% (435/2008) in humans (23% in diarrhea patients and 14% in high-risk populations) and 30% (1253/4122) in food-producing animals (39% in poultry and poultry products and 10% in ruminants and ruminant products). The two most commonly detected *Campylobacter* spp. serovars were *C. jejuni* with a prevalence of 16% (422/2693) in humans (16% in diarrhea patients and 9% in high-risk populations) and 22% (1182/5472) in food-producing animals (25% in poultry and poultry products and 14% in ruminants and ruminant products) and *Campylobacter coli* with a prevalence of 4% (72/1938) in humans (3% in diarrhea patients and 8% in high-risk populations) and 9% (367/4037) in food-producing animals (13% in poultry and poultry products and 2% in ruminants and ruminant products) (Tables 1 and 2).

#### *3.4. Microbial Resistance Patterns Detected by Phenotypic Screening*

Based on the prevalence rate results and the number of eligible articles included in this review, NTS and *C. jejuni* were the two most prevalent representatives of *Salmonella* spp. and *Campylobacter* spp., respectively. The average resistance of NTS and *C. jejuni* was calculated for each pathogen separately depending on the source of the isolated bacteria (human or food-producing animals) (Supplementary Table S4).

For NTS, information on 13 different antibiotics was available and is summarized in Table 3. NTS isolated from humans showed resistance against erythromycin (100%), amoxicillin (71%), tetracycline (62%), ampicillin (52%), azithromycin (43%), amoxicillin–clavulanic acid (42%), streptomycin (40%), cefotaxime (31%), trimethoprim–sulfamethoxazole (24%), chloramphenicol (15%), ciprofloxacin (9%), imipenem (2%), and ceftriaxone (1%). NTS isolated from food-producing animals showed resistance against erythromycin (100%), amoxicillin (91%), tetracycline (50%), ampicillin (69%), azithromycin (9%), amoxicillin–clavulanic acid (70%), streptomycin (43%), cefotaxime (63%), trimethoprim–sulfamethoxazole (8%), chloramphenicol (12%), ciprofloxacin (17%), and ceftriaxone (7%). Amoxicillin–clavulanic acid was used more frequently in animal isolates (70%), with a pooled risk ratio (RR) of 1.09 (95% confidence interval (CI): 1.01–1.18) and with a heterogeneity of *I* <sup>2</sup> = 34% and *r* <sup>2</sup> <sup>≤</sup> 0.001, while the pooled RR close to 1 in ampicillin and streptomycin suggests a similar probability of occurrence in humans and animals. For the other antibiotics, no clear pattern was detected (Table 3).

For *C. jejuni,* we had data on 11 antibiotics. The phenotypic resistance results are summarized in (Table 4). *C. jejuni* isolated from humans showed resistance against amoxicillin (100%), trimethoprim–sulfamethoxazole (93%), nalidixic acid (89%), azithromycin (88%), chloramphenicol (82%), ampicillin (81%), tetracycline (75%), ciprofloxacin (73%), amoxicillin–clavulanic acid (68%), erythromycin (65%), and streptomycin (39%). *C. jejuni* isolated from food-producing animals showed complete resistance against amoxicillin (100%) and azithromycin (100%) and to a lesser extent resistance against trimethoprim– sulfamethoxazole (83%), nalidixic acid (76%), chloramphenicol (69%), ampicillin (64%), tetracycline (56%), ciprofloxacin (71%), amoxicillin–clavulanic acid (32%), erythromycin (38%), and streptomycin (21%).



**Table 1.** Overall prevalence of *Salmonella* and *Campylobacter* and main serotypes.

**Table 2.** Prevalence of *Salmonella* and *Campylobacter* and main serotypes based on the samples sources.


*Antibiotics* **2022**, *11*, 536



*Antibiotics* **2022**, *11*, 536



Azithromycin was detected more frequently in animal isolates, with a pooled risk ratio (RR) of 1.13 (95% confidence interval (CI): 1.04–1.24) and a heterogeneity of *I* <sup>2</sup> = 72% and *r* <sup>2</sup> = 0.033. Amoxicillin–clavulanic acid was detected more frequently in human isolates, with a pooled risk ratio (RR) of 0.79 (95% confidence interval (CI): 0.67–0.95) and heterogeneity of *I* <sup>2</sup> = 53% and *r* <sup>2</sup> <sup>≤</sup> 0.001. For the other antibiotics, no clear pattern was detected. Most phenotypic resistance had a pooled RR close to 1, suggesting a similar probability of occurrence in humans and animals (Table 4)

#### *3.5. Assessment of Shared Antimicrobial Resistance Genes*

Only six studies reported resistance genes targeted at three serovars of *Salmonella* spp. and *Campylobacter* spp. These serovars were *C. jejuni* (three studies), NTS (two studies), and *Salmonella Heidelberg* (one study). We calculated the average prevalence for every single resistance gene from food-producing animals and human sources separately. For human isolates, *tetO* was the gene with the highest prevalence (85%), followed by Class 1 Integrons (81%), *blaOXA-61 (53%)*, *cmeB (51%)*, *blaCMY-2 (38%),* Class 2 integrons (29%), *tetA (21%), blaOXA (21%), blaSHV (19%), AAC(6')-Ib (16%), blaCTXM-1 (16%), blaAMPc (13%),* and *blaTEM (13%)*. For food-producing animals, tetO was the most prevalent gene (77%), followed by Class 1 Integrons (69%), *blaOXA-61 (35%), cmeB (35%), tetA (30%),* Class 2 integrons (27%)*, blaCTXM-1 (22%), AAC(6')-Ib (22%), blaSHV (20%), blaTEM (15%), blaCMY-2 (11%), blaOXA (11%),* and *blaAMPc (3%)* (Table 5) (Supplementary Table S5a–c).

Resistance in *Campylobacter* spp. was exclusively reported as data for *C. jejuni* isolates. The three studies reporting data on resistance compromised 274 isolates (232 human isolates and 42 food-producing animals and their products). The most frequent genes were Class 1 Integrons (96%), *tetO* (85%), *blaOXA-61* (53%), *cmeB* (51%), and *tetA* (17%). For foodproducing animals and their product isolates, the most frequently detected genes were Class 1 Integrons (100%), *tetO* (77%), *blaOXA-61* (35%), *cmeB* (35%), and *tetA* (30%). There was no evidence for a significant difference in the occurrence of the genes between human and food-producing animals and their products (Table 5) except for Class 1 integrons, which were detected more frequently in food-producing animals, with a risk ratio (RR) of 1.04 (95% confidence interval (CI): 1.01; 1.08) (Table 5). No clear pattern was detected for the other genes, with most of the genes having a pooled RR close to 1, suggesting a similar probability of occurrence in humans and animals (Table 5) (Supplementary Table S5a).

The two studies on NTS compromised 197 isolates (125 human isolate and 72 foodproducing animals and their products). The most frequent genes were Class 1 Integrons (51%), Class 2 Integrons (29%), *blaSHV* (16%), *blaCTXM-1* (16%), *AAC(6')-Ib* (16%), *blaTEM* (10%), and blaAMPc (1%). For food-producing animal isolates, the most frequently detected genes were Class 1 Integrons (41%), Class 2 Integrons (27%), *blaSHV* (22%), *blaCTXM-1* (22%), *AAC(6')-Ib* (22%), and *blaTEM* (15%). No clear pattern emerged for the majority of the genes in the random effect models comparing frequencies in humans and animals (Table 5), suggesting there was no evidence for a significant difference in the occurrence of the genes between humans and food-producing animals (Table 5) (Supplementary Table S5b).

The single study including *Salmonella* Enterica Serovar Heidelberg compromised 33 isolates (24 human isolates and 9 food-producing animals and their products). In isolates from human sources, the most frequent genes were *blaAMPc* (50%), *blaCMY-2* (38%), *blaTEM* (29%), blaSHV (25%), and blaOXA (20%). For food-producing animal isolates, the most frequently detected genes were *blaAMPc* (11%), *blaCMY-2* (11%), *blaTEM* (11%), *blaSHV* (11%), and *blaOXA* (11%). There was no evidence of a significant difference in the occurrence of the genes between humans and food-producing animals (Supplementary Table S5c).



HN: Number of human isolates: HI: human isolates that have this gene; prevalance\_H: prevalence among human isolates; AN: number of animal isolates; AI: animal isolates that have this gene; prevalance\_A: prevalence among animal isolates; RR: relative risk; NTS: non-typhoidal *Salmonella*; S.H: *Salmonella* Heidelberg; *C. jejuni*: *Campylobacter jejuni.*

#### **4. Discussion**

Although 41 articles were eligible for inclusion in this systematic review and metaanalysis, there is an uneven distribution of the sources of the studies included. The majority (63%) of eligible studies were from Egypt and 20% from Iran. On the other hand, there are no published papers on applying a comprehensive One Health approach to study one of the two major foodborne diseases (*Salmonella* and *Campylobacter*) in 9 counties from the 16 Middle Eastern countries, and none came from those high-income countries members of the Gulf Cooperation Council.

Of the 41 studies included in this review, 31 were cross-sectional, and 10 were routine data studies. Studies allow a comparison between human and animal sources; they do not evaluate actual transmission methods because the few studies eligible for inclusion in this review suffered from insufficient statistical data on foodborne pathogens and AMR and assess only selected sections of the social ecosystem.

Furthermore, our systematic review and meta-analysis showed the prevalence of *Salmonella* spp. and *Campylobacter* spp., resistance rates, and antimicrobial resistance genes circulating in the Middle East region by using the random-effects model. The model showed high heterogeneity results, which indicate variability in the study data. This might be due to the study design (epidemiological study vs. routine data) or due to diverse sample types. The human isolates used in the studies were from different sources (symptomatic and asymptomatic participants), and the animal isolates used were from various sources (live animals and products). Finally, the small sample size in each study and, in particular, the human sample size could influence the results when measuring the prevalence and the relationship between the humans and animal settings. The heterogeneity might explain the insignificant relationship between animals and humans.

The low quantity (low sample size), uneven distribution in the reported data, and weak epidemiological study designs from a One Health methodological perspective [24] in the studies that targeted foodborne illness and antimicrobial resistance in the Middle East can be explained by the food safety system's challenges in this region. These challenges are the lack of epidemiological and disease ecological capacity, diagnostic tools, and laboratory facilities. Moreover, there is a lack of quality control and standardization of microbiological identification and susceptibility testing techniques [15].

Our review demonstrates the prevalence of NTS and *Campylobacter* spp. and their serovars circulating in the Middle East. The pooled prevalence of *Campylobacter* spp. among humans was close to the higher estimate for the ranges reported in Sub-Saharan Africa: and Northern Africa (2–27.5%) and more than the ranges reported in Southeast Asia (8%) [25–27]. Additionally, the results show the prevalence of *Campylobacter* spp. in foodproducing animals and their products (30%). For *Campylobacter* spp., the prevalence rate is similar to the systematic review and meta-analysis results that targeted *Campylobacter* spp. globally, with approximately 30% of animal food products analyzed reporting *Campylobacter* spp. [28]. Additionally, we looked at which *Campylobacter* serovars are circulating in the Middle East and found *C. jejuni* and *C. coli* to be the predominant serovars, similar to results that targeted *Campylobacter* in Africa as the *C. jejuni* and *C. coli* predominates in Sub-Saharan Africa [29].

We identified two systematic reviews conducted by Al-Rifai and his colleagues that targeted the Middle East and South African countries in 2019 and 2020; we will compare our results with these relevant studies. In this review, the pooled prevalence rates of NTS were 9% and 13% among humans and animals and their products, respectively. The pooled prevalence in humans is higher than the results in the Al-Rifai study (2019) which was 7% [30]. In addition, the prevalence of the food-producing animals in this review is more than the results of the Al-Rifai (2020) study, which was 9% [31]. Our findings are similar to Al-Rifai's studies of NTS serovars, in which *S.* Typhimurium and *S.* Enteritidis were the main NTS serovars reported in this region [30,31].

Furthermore, this systematic review showed *Campylobacter* and *Salmonella* serovars are highly prevalent in poultry and poultry products in the Middle East. The Campylobacter prevalence in animals was less than the prevalence reported in broiler meat in Poland, Slovenia, Spain, and Austria. Conversely, more than reported in Denmark and Finland [32]. The *Salmonella* prevalence in animals showed results less than the prevalence reported in raw chicken at retail markets in China and more than reported in chicken carcasses in Spain [33,34]. This endemic *Campylobacter* and *Salmonella* bacteria in animal food products can be explained, at least partially, by the changes in animal production systems that have tended to be more intense over the past decades [28]. These findings are essential because transmission along the production chain is generally established as the most common pathway used by *Campylobacter* and *Salmonella* to generate human infection [29].

AMR is a transboundary public health problem. New types of AMR strains can expand worldwide following initial endemic emergence, as demonstrated by several resistant pathogens that spread globally [35]. Our meta-analysis revealed a high NTS resistance against erythromycin, amoxicillin, tetracycline, and ampicillin for isolates from humans and food-producing animals. The isolates have similar resistance rates between humans and animals in erythromycin but are higher in isolates from animal sources for amoxicillin and ampicillin and higher in isolates from human sources in tetracycline. These results are close to those reported by Alsayeqh's systematic review in the Middle East region [15]. In addition, the most recent report on AMR in the EU in 2019–2020 found that resistance of NTS to sulfonamides, ampicillin, and tetracycline was high in human isolates, while it ranged from moderate to very high in animal isolates [36].

AMR phenotypic results for *C. jejuni* isolates (human and food-producing animals) showed high resistance against amoxicillin, trimethoprim–sulfamethoxazole, nalidixic acid, azithromycin, chloramphenicol, ampicillin, tetracycline, and ciprofloxacin. These findings were close to Alsayeqh's systematic review for trimethoprim–sulfamethoxazole, nalidixic acid, and tetracycline. In comparison, it has a lower resistance rate for amoxicillin, chloramphenicol, ampicillin, and ciprofloxacin [15]. Our results show that *C. jejuni* isolated from humans has a phenotypical resistance rate against nalidixic acid and tetracycline more than that reported in Italy and less against ciprofloxacin based on the same study results [37]. At the same time, our results show that *C. jejuni* isolated from animals has a phenotypical resistance rate against nalidixic acid, ciprofloxacin, and tetracycline more than that reported in broiler chicken in Belgium [38]. This systematic review demonstrated moderate to high resistance of *C. jejuni* to erythromycin. Conversely, the recent EU report on AMR found that *C. jejuni* resistance to erythromycin was either undetected or detected at very low levels in *C. jejuni* from food-producing animals and humans [36].

The WHO, Food and Agriculture Organization of the United Nations (FAO), and World Organization for Animal Health (OIE) recommend reducing antibiotic use in animal husbandry, particularly for those known to cause cross-resistance [39–41]. However, some antimicrobials traditionally used in animal production as growth promoters and/or for treating gastrointestinal infections are also used to control human infectious diseases (e.g., tetracycline and quinolones) [42]. The misuse and overuse of antimicrobials in clinical and veterinary medicine and agriculture have increased antimicrobial resistance pathogens, including *Campylobacter* and *Salmonella* [43]. For instance, the Quesada study showed that *Salmonella* isolated from animal food has significant antibiotic resistance in Latin American countries [44]. The therapeutic and prophylactic use of antibiotics in animal production for long periods is likely contributing to the widespread resistance against antibiotics [43]. More integrated environmental–animal–human studies are needed in the region to ascertain its effect on public health. This way, microbiological and clinical evidence on the transmission of AMR between animals and humans can be ascertained in Middle Eastern countries [43–45].

Data on antimicrobial resistance genes (ARGs) among *Salmonella* spp. and *Campylobacter* spp. in the Middle East is limited. However, based on the reported information, we can argue that food-producing animals and their products in the Middle East are not the main drivers for the emergence of ARGs.

Based on our eligibility criteria, six studies targeted ARGs among *Salmonella* and *Campylobacter* as foodborne illnesses in the Middle East region [46–51]. Besharati's study and Youssef's study were two studies that reported the ARGs among NTS. Besharati's study shows an association between the AMR phenotype results and ARGs results in Integron 1 and 2 classes and trimethoprim/sulfamethoxazole in Iran. Conversely, in Youssef's study, results from Egypt revealed no association between AMR phenotype results and ARGS.

Three studies reported the ARGs among*C. jejuni*(Abd-El-Aziz, Divsalar, andGhoneim) [46–48]. The results in Divsalar and Ghoneim could not show a significant association between the targeted ARGs and the AMR phenotype results. In turn, Abd-El-Aziz found an association between Class 1 integrons and aminoglycoside resistance.

We identified small-scale studies with a small sample size for the ARGs in the NTS and *C. jejuni*. The small sample size in the eligible studies might be responsible for the insignificant difference in the occurrence of the genes between humans and food-producing animals. Our results agreed with Escher's systematic review that targeted ARGs in Africa and found eligible studies characterized by small-scale studies and with a small sample size [16]. Therefore, future studies should have an integrated approach to assess the ARGs and should have a suitable sample size.

Partial sequencing of *C. jejuni* and NTS were performed using conventional PCR to extract the ARGs. Therefore, there is a lack of laboratory techniques that determine the order of bases in an organism's genome in one process such as with Whole-genome sequencing (WGS), to follow the foodborne illnesses and ARGs. Undertaking WGS of isolates, especially those with high-level antibiotic resistance, is strongly encouraged to demonstrate the involved ARGs and their genetic localization (plasmid, chromosome, genomic islands, integrative and conjugative element, and transposon) as well as to detect the most prevalent resistant serovars [36,52,53], detail their potential of horizontal transmission, and evaluate the different sources and comparison of human and animal isolates [54].

#### **5. Conclusions**

To the best of our knowledge, this is the first systematic review assessing integrated environment–animal–human studies using a One Health approach in the Middle East to pursue foodborne illnesses and antimicrobial resistance. The One Health approach was not rigorously applied in the Middle East countries. In addition to weak epidemiological study designs from a One Health methodological perspective, there is an uneven distribution in the reported data with about 60% of Middle Eastern countries having no published papers included in this review. More research on foodborne illnesses and AMR in the Middle East is urgently needed. The AMR phenotype results showed a high prevalence of resistance rate for the isolated bacteria that highlights the importance of antimicrobial stewardship in humans and animals in tandem. Furthermore, introducing new laboratory techniques that determine the order of bases in an organism's genome is essential to follow up the foodborne illness outbreak and ARGs.

A simultaneous approach that targets human and animal health in tandem with a solid epidemiological study design has a high potential to provide evidence for understanding the drivers for the emergence and spread of foodborne pathogens and AMR. A comprehensive One Health approach, integrating by a sound epidemiological design the spatio-temporal relationship of humans, animals, and their environment, will allow us to identify key transmission pathways, which are essential for designing more efficient food safety systems and AMR control policies.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antibiotics11050536/s1, Supplementary Table S1: The search strategy used for each database., Supplementary Table S2: Overview of the selected studies, Supplementary Table S3: Summary of the selected studies showing the country and pathogens together, Supplementary Table S4: Phenotypic resistance to antibiotics for all isolated serovars, Supplementary Table S5. a, b, and c: Genotypic resistance to antibiotics for all isolated serovars.

**Author Contributions:** S.A., H.T., L.C., J.Z., P.V. and N.M.E.A.-R., designed the search strategy and data-extraction form; S.A. and H.T. screened studies for selection; S.A., H.T. and A.A. extracted the data; S.A. and J.H. managed and analyzed the data; P.V. provided expert microbiological and antimicrobial resistance advice; J.Z. provided expert One Health and epidemiology advice; N.M.E.A.-R. provided expert epidemiology advice; S.A. and H.T. wrote the first draft of the manuscript; A.A., L.C., J.Z., P.V., J.H. and N.M.E.A.-R. critically revised and finalized the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The datasets generated during the current meta-analysis are available from the corresponding author upon reasonable request. All data analyzed for meta-analysis are included in the corresponding published articles, as reported in Supplementary Table S2.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article Salmonella* **in Pig Farms and on Pig Meat in Suriname**

**Patrick Butaye 1,2,\*, Iona Halliday-Simmonds <sup>1</sup> and Astrid Van Sauers <sup>3</sup>**


**Abstract:** *Salmonella* is one of the most important food borne zoonotic pathogens. While mainly associated with poultry, it has also been associated with pigs. Compared to the high-income countries, there is much less known on the prevalence of *Salmonella* in low- and middle-income countries, especially in the Caribbean area. Therefore, we investigated the prevalence of *Salmonella* in pigs and pig meat in Suriname. A total of 53 farms and 53 meat samples were included, and *Salmonella* was isolated using standard protocols. Strains were subjected to whole genome sequencing. No *Salmonella* was found on pig meat. Five farms were found to be positive for *Salmonella*, and a total of eight different strains were obtained. Serotypes were *S.* Anatum (*n* = 1), *S.* Ohio (*n* = 2), a monophasic variant of *S.* Typhimurium (*n* = 3), one *S.* Brandenburg, and one *S.* Javaniana. The monophasic variant of *S.* Typhimurium belonged to the ST34 pandemic clone, and the three strains were very similar. A few resistance genes, located on mobile genetic elements, were found. Several plasmids were detected, though only one was carrying resistance genes. This is the first study on the prevalence of *Salmonella* in pigs in the Caribbean and that used whole genome sequencing for characterization. The strains were rather susceptible. Local comparison of similar serotypes showed a mainly clonal spread of certain serotypes.

**Citation:** Butaye, P.;

Halliday-Simmonds, I.; Van Sauers, A. *Salmonella* in Pig Farms and on Pig Meat in Suriname. *Antibiotics* **2021**, *10*, 1495. https://doi.org/10.3390/ antibiotics10121495

Academic Editor: Piera Anna Martino

Received: 11 November 2021 Accepted: 3 December 2021 Published: 6 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Keywords:** *Salmonella enterica*; pigs; Suriname; whole genome sequencing

#### **1. Introduction**

Few studies have been performed on *Salmonella* in the Caribbean region, and those mainly dealt with the prevalence in poultry. Prevalences were very variable depending on the study. Moreover, most studies were performed in Trinidad and were mainly on food products, both fresh and ready to eat. The prevalence on those products varied between 0% (mainly ready to eat) and 7%. Several serotypes were isolated, including *S.* Agona *S.* Kiambu, *S.* Kentucky, *S.* Derby, and *S.* Mbandaka as the most reported. The few studies in the Caribbean that have been performed on live animals dealt with layer chickens, and apart from one study, all were on eggs. Eggs have been a major focus for the Caribbean. Prevalence on eggs varied from 1 to 13%, and a multicountry study in 2014 showed that 3% of the layer farms were positive. Prevalence in Barbados on layer farms was the highest, with 73% of the farms being positive. It should be noted that for the different studies, different sampling and isolation methods were used, and as such, comparisons are difficult [1,2].

Several studies were performed on pet animals and mainly wildlife. In dogs, eight *Salmonella* strains were recovered [3]; in iguanas, mongooses, tree boa, leatherback turtles, blue land crabs, and toads, a few strains were found, though it should be noted that the serotypes were very different, and that the iguanas had to be regarded as wildlife. These strains in general carried few resistances [4–10].

There are very few studies on pork or pigs in the Caribbean region. An old study in Trinidad and Tobago showed that about 18% of the swine carcasses sampled were positive

for *Salmonella* [11], while a small-scale baseline study in the Bahamas showed that 8/42 of the retail meat samples were positive for *Salmonella*. At a pig slaughter plant, 44.1% (15/34) and 2.9% (1/34) were found positive at the beginning and the end of the slaughter process, respectively [12]. In both diarrheic and non-diarrheic pigs of different ages from Trinidad, a prevalence between 3 and 4.5% was recorded. A more recent study in Cuba demonstrated a prevalence of 2.2% in weaned piglets [13]. Apart from those studies, we found no other studies in live pigs.

In humans, there is no systematic surveillance of nontyphoid *Salmonella* infections, and as such, the incidence is unknown. However, the incidence has been estimated at an annual rate of over five cases per 100,000 persons in Trinidad and Tobago [6,14].

Pig production is important in the food supply in Suriname. In 2018, 208 farms were active, with a total of 21,362 pigs divided over the different farms. Since pig meat can also be a major source of *Salmonella* infections in humans, we investigated the prevalence, as well as the types of *Salmonella* in pigs in Suriname.

#### **2. Materials and Methods**

#### *2.1. Sampling*

Samples from pigs were taken at the only pork slaughterhouse in Paramaribo, Suriname in 2018. Farms were located in the districts of Paramaribo, Wanica Saramacca, and Coronie. Pigs were mixed breed pigs, with breeds involved being the Dutch Landrace, Yorkshire, Pietrain, and Duroc. Pigs originated from 53 farms in the central part of Suriname that were delivering pork to the slaughterhouse. One pig per farm was sampled except for two farms; from one of these, two pigs were sampled on the same day; and from another farm, two pigs were sampled, but on different sampling days. A sample was taken from the cecum and ileocecal lymph node of each pig.

Meat samples were obtained from 53 retail stores in Paramaribo and Wanica. Most retail stores are in Paramaribo, which contains about half of the population of Suriname. Per retail store, one piece of fresh pork chop sample was obtained. There is no food tracing system in Surname, and as such, the slaughter date nor origin of the meats could be traced.

#### *2.2. Isolation and Identification*

*Salmonella* was isolated using the ISO 6579 annex D. Briefly, after homogenization, the sample (25 g) was inoculated in Buffered Peptone Water (BPW, Bio-Rad, Hercules, CA, USA) (1:10, W/V) and incubated for 16–20 h at 37 ◦C. Of this, 0.1 mL was inoculated on a Modified Semisolid Rappaport Vassiliadis agar plate (MSRV, Bio-Rad, Hercules, CA, USA) and incubated for 21–27 h at 41 ◦C. Positive samples were inoculated on Xylose Lysine Deoxycholaat agar plate (XLD, Bio-Rad, Hercules, CA, USA) and a nalidixine-BGA plate and incubated for 21–27 h at 37 ◦C. Suspected colonies were inoculated on Triple Sugar Iron agar (TSI, Bio-Rad, Hercules, CA, USA) and lysine decarboxylase broth (Thermo Fisher Scientific, Waltham, MA, USA) and incubated for 18–24 h at 37 ◦C for presumptive identification. Further identification was performed using the Gram-Negative ID Plate (Sensititre GNID, TREK Diagnostic Systems, Cleveland, OH, USA), using the protocols provided by the supplier, and using a Sensititre OptiRead analyzer (TREK Diagnostic Systems, Cleveland, OH, USA). *Salmonella* isolates were then serogrouped using the Wellcolex Color *Salmonella* Rapid Latex Agglutination Test Kit (Thermo Fisher Scientific, Waltham, MA, USA). Strains that could serogrouped were selected for whole genome sequencing.

#### *2.3. Whole Genome Sequencing*

Purified strains were sent to Macrogen (Seoul, Korea) for DNA extraction and sequencing on an Illumina platform using a TruSeq Babi DNA kit and 151 bp long pairedend sequencing.

Raw sequences (fastq files) were submitted to the NCBI database under PRJNA751882 with SAMN20584910, SAMN20584911, SAMN20584912, SAMN20584913. SAMN20584914, SAMN20584915, SAMN20584916, and SAMN20584917.

Sequences were trimmed and assembled using SKESA [15]. Annotation was done with PROKKA [16] and RAST [17]. The serotype was determined with Sistr [18]. The MLST profile was determined using 'mlst' [19]. Phylogenetic analysis of the isolates of a same serotype was done with NASP [20]. Further analysis of the strains was done using ARIBA against the plasmid finder [21], ARg-ANNOT [22], and ResFinder [23,24]. Plasmids were confirmed with Platon [25]. Using RAST analysis, we located the different associated genes.

#### *2.4. Statistical Analysis*

Confidence intervals of the prevalence of *Salmonella* on farms in Suriname were calculated using exact binomials in an Excel file.

#### **3. Results**

#### *3.1. Prevalence, Serotyes, and Epidemiology*

*Salmonella* could not be isolated from any of the food samples. Of the 53 farms investigated, a total of eight strains were isolated. Strains originated from five different farms (9%, of the farms were positive, confidence interval 3.1–21%) and seven different animals. Five different serotypes were isolated, including one *S.* Anatum, *S.* Ohio (*n* = 2), a monophasic variant of *S.* Typhimurium (*n* = 3), one *S.* Brandenburg, and one *S.* Javaniana (Table 1).


**Table 1.** Isolation and serotype results and origin of the strains.

\* Strains from the same animal.

#### *3.2. Antimicrobial Resistance Genes*

All strains had at least one resistance gene (Table 2), though it should be taken into account that all *Salmonella* have the *aac(6*0 *)* gene in their chromosome [26]. As such, three strains did not show any additional resistance gene. A total of five out of eight strains were resistant to penicillins, mediated by the *bla*TEM-1 gene. All strains had at least one resistance gene for aminoglycosides, and some strains had up to four different aminoglycoside resistance genes. Four strains carried the *sul2* gene, indicating sulfonamide resistance. Only two strains carried the tetracycline resistance gene, *tet*(B), and one strain carried a trimethoprim resistance gene, *dfrAB*. However, the latter strain was susceptible to sulfonamide.


**Table 2.** T Antimicrobial resistance and plasmids associated with the different isolates.

#### *3.3. Plasmids*

Two strains did not carry any plasmids (strain 174, *S.* Ohio; and 250, *S.* Javaniana). Five different Inc plasmids were detected: IncQ1\_1, IncI1\_1\_Alpha, IncFIA(HI1)\_1\_HI1, IncH1A/B, and IncHI1B(R27)\_1\_R27.

Three different colicin plasmids were found. The full sequence of a small plasmid ColpVC\_1 was found by Platon in strain 173, and this plasmid was nearly similar to earlier described plasmids in *E. coli* and *Salmonella* (NCBI BLAST result). The Col440I\_1 plasmid from the *S.* Brandenburg and *S.* Ohio strain were identical. It was a small circular plasmid of 1748 bp and contained four genes. The ColRNAI\_1 plasmid from the *S.* Brandenburg strain was 4597 bp and had seven genes.

#### *3.4. Association of Resistance Genes with Mobile Genetic Elements*

The *aph(6)-Id (strB), aph(3")-Ib (strA)* genes in the three monophasic *S.* Typhimurium strains were located on a same contig together with a *sul2*. However, there were no other genes on that contig, making its location speculative. In the *S.* Brandenburg strain, this same gene cluster was located on a 63,539 pb contig that also contained the *bla*TEM-1 gene, downstream of the *sul2* gene. In the *S.* Ohio strain 543, the *aph(6)-Id, aph(3")-Ib*, *bla*TEM-1, and *sul2* genes were located as one cluster with a similar structure to the *S.* Typhimurium and *S.* Brandenburg strains, and associated with the IS*1* gene *InsB*, though the downstream part associated with Tn*7* was missing on the contig. In addition, copper and zinc resistance genes were present in this cluster. This gene cluster most likely was associated with a mobile Tn*7*-like transposon.

The *aph(3*0 *)-Ia* gene was typically located on the InQ1\_1 plasmid in the three monophasic *S.* Typhimurium strains.

The tetracycline resistance genes found in the strains of this collection were located on the same structure; however, the contig containing the *tet*(B) gene was too small to determine the exact location Blast analysis of the whole contig showed that it was similar to a chromosomal location in the bacterial species *Glasseralla parasuis*, *E. coli*, and *Shigella flexneri*.

#### **4. Discussion**

This is one of the few more systematic investigations on *Salmonella* in pigs in the Caribbean. Though few samples per farm could be tested, when lowering the sensitivity for detecting positive farms, we found 5/53 farms positive, which was close to 10% of the farms being positive. This was higher than what has been found in Cuba, though it should be noted that this study was conducted on weaned pigs, while our study was conducted on pigs at slaughter [13]. Our findings were lower than what was found on pig

carcasses in Trinidad in 1978, while higher than what was found in the same country in fecal material of pigs in 1993 (4.1% of the 294 pigs) and in 1994, with 4.5% of the diarrheic pigs and 3.4% (*n* = 179) of the non-diarrheic pigs (*n* = 117) originating from 25 farms sampled at different ages being positive [1,27]. The variations can also be explained by the age or carcass contamination, as it is known that older pigs are in general more prone to be positive than piglets [28], while contamination at the slaughter line may increase the apparent prevalence [29].

In this study, we found five different serotypes. The most commonly found in the European Union and the US are *S.* Typhimurium, *S.* 1,4,[5],12:i:-, *S.* Derby, and *S.* Rissen, though this varies in time and space [29,30]. While the monophasic variant of *S.* Typhimurium is one of the most commonly found serotypes in pigs worldwide, and *S.* Brandenburg is less commonly found, the other serotypes, *S.* Ohio, *S.* Anatum, and *S.* Javaniana, are more rarely reported. This may indicate that there may be some differences in the epidemiology of Salmonella in the Caribbean; however, due to the low numbers of strains recovered, this should be interpreted with care. Two of the monophasic *S.* Typhimurium strains were the same based on the NASP SNP analysis, as they had no SNPs in the core genome, while the other strain had only 235 SNPs different. It was striking that the two equal monophasic *S.* Typhimurium strains were from two different farms. Unfortunately, we could not determine whether those two farms had epidemiological connections. The two different monophasic *S.* Typhimurium strains came from the same farm, but from different pigs, indicating several different clones were circulating in that farm. All the monophasic Typhimurium strains were ST 34, an epidemic clone in Europe [31] and other parts of the world that is to be regarded as a pandemic strain [32–34]. Compared to other *S.* Typhimurium ST34 strains, the Caribbean clone was remarkably susceptible. This may be due to the limited availability of antimicrobials in Suriname. *S.* Anatum ST64 has been reported as one of the most prevalent strains on retail pork in Juiangsu China [35]; however, few data are available on this clone. The two *S.* Ohio strains originated from different farms and had 716 SNPs different, indicating that they were unrelated strains.

We could not find a single study on antimicrobial resistance genes in *Salmonella* from pigs in the Caribbean, and only few on *Salmonella* as a whole in the Caribbean. In poultry strains in Trinidad, a quite higher prevalence of resistance was found compared to our study, and this included plasmid-mediated colistin resistance [36–38].

All monophasic *S.* Typhimurium strains carried the IncQ1\_1 plasmid. The full sequence of 6644 bp of this mobilizable plasmid was obtained. This plasmid also has been found in other *Salmonella* and *E. coli* (BLAST search 2/2021), as well as in *Aeromonas hydrophila* isolated from swine [39]. The IncI1 plasmid, found in one of the monophasic *S.* Typimurium strains, is a conjugative plasmid frequently associated with antimicrobial resistance genes, though here we could not identify any. However, since the sequence was not closed or complete (though it consisted of 68,689 bp), we could not confirm this. Nevertheless, we found genes associated with heavy metal resistance: there was a silver and copper resistance gene cluster on this plasmid. Copper is used in swine rearing, and this use might have selected for this resistance.

The two IncHI1A\_1 and IncHI1B\_1 sequences were found on the same contig, indicating this was a hybrid plasmid with two different replicons. Most likely, the Inc-FIA(HI1)\_1\_HI1 replicon was also on this plasmid, as this rep protein was on a single gene contig, and has been reported before on the same plasmid in other different serotypes of *Salmonella* worldwide [40]. This combination of replicons, together with an IncN replicon, has also been found in *E. coli*. This plasmid carried an *mcr*-gene encoding colistin resistance [41]. A plasmid with exactly the same sequences was found in *S.* Brandenburg and *S.* Ohio from two different farms, indicating its spread amongst *Salmonella* in Suriname. Using BLAST on the NCBI database, a similar plasmid, though with lower coverage and similarity, was found in a *Klebsiella* strain. Similarly, the ColpVC\_1 plasmid was found worldwide in different bacterial species [42]. It is clear that all the col plasmids spread between the different Salmonella serotypes in Suriname.

It was somehow difficult to locate all resistance genes, as Illumina sequencing, also depending on the sequence depth, created several contigs that cannot be linked. Nevertheless, on bigger contigs as well, it is not always evident to link to structures. We could link the cluster encoding aminoglycoside, sulphomamide, and β-lactam (*aph(6)-Id* (*strB*), *aph(3")-Ib* (*strA*), *sul2*, *bla*TEM-1) resistance to an Tn7-like structure, also indicating its mobility, as it was present in several strains and moving around as coresistances. This means when one of the antimicrobials was used, it selected for all the resistance genes, and thus created a larger problem. Moreover, we also had to take into account the use of heavy metals such as zinc and copper in veterinary medicine. Resistance genes against these metals are also located on this mobile genetic element (data not shown). The *aph(3*0 *)-Ia* gene could not be linked to any mobile structure. The location of the *tet*(B) gene could not really be determined; however, through BLAST analysis, the same genes in the contig were also present in *Glasseralla parasuis*, *E. coli*, and *Shigella flexneri*, which indicated the mobility or preferential location of this gene.

#### **5. Conclusions**

In conclusion, this was the first study on the prevalence of *Salmonella* in pigs in the Caribbean. About 25 of all the farms in Suriname were sampled, and prevalence was 9% (CI 3.1–21%). Few strains were isolated, although they included the pandemic monophasic *S.* Typhimurium ST34. Fewer resistances were found compared to other monophasic *S.* Typhimurium ST34 strains isolated in other countries. In general, the strains were rather susceptible, except for aminoglycoside resistance. Most resistances could be located on mobile genetic elements, with a multi-resistant, Tn7-like element spreading. Local comparison of similar serotypes showed a rather clonal spread of certain serotypes. More strains should be analyzed to determine the local epidemiology of *Salmonella* in pigs in Suriname. The fact that no strains could be isolated from meat samples showed that food safety is not hampered very much by the presence of *Salmonella* in pigs.

**Author Contributions:** Conceptualization, P.B. and A.V.S.; methodology, P.B. and A.V.S.; software, P.B.; validation, P.B., A.V.S. and I.H.-S.; formal analysis, P.B.; investigation, P.B.; resources, P.B. and A.V.S.; data curation, P.B. and A.V.S.; writing—original draft preparation, P.B.; writing—review and editing, P.B., A.V.S. and I.H.-S.; supervision, P.B.; project administration, P.B.; funding acquisition, P.B. and A.V.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by WHO AGISAR, under grant number "Focused project, Determination of the transfer of Salmonella from pigs to food in Suriname" 2017/725664-0.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Raw sequences (fastq files) were submitted to the NCBI database under PRJNA751882 with SAMN20584910, SAMN20584911, SAMN20584912, SAMN20584913. SAMN20584914, SAMN20584915, SAMN20584916, and SAMN20584917.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


### *Article* **Multi-Drug Resistance to** *Salmonella* **spp. When Isolated from Raw Meat Products**

**Joanna Pławi ´nska-Czarnak 1,\* , Karolina Wódz <sup>2</sup> , Magdalena Kizerwetter-Swida ´ <sup>3</sup> , Janusz Bogdan <sup>1</sup> , Piotr Kwieci ´nski <sup>2</sup> , Tomasz Nowak <sup>2</sup> , Zuzanna Strzałkowska <sup>1</sup> and Krzysztof Anusz <sup>1</sup>**


**Abstract:** *Salmonella* spp. is the most frequent cause of foodborne diseases, and the increasing occurrence of MDR strains is an additional and increasing problem. We collected *Salmonella* spp. strains isolated from meat (poultry and pork) and analysed their antibiotic susceptibility profiles and the occurrence of resistance genes. To determine the susceptibility profiles and identify MDR strains, we used two MIC methods (MICRONAUT and VITEC2 Compact) and 25 antibiotics. Phenotypic tests showed that 53.84% strains were MDR. Finally, molecular analysis strains revealed the presence of *bla*SHV, *bla*PSE-1, *bla*TEM, but not *bla*CTX-M genes. Moreover, several genes were associated with resistance to aminoglycosides, cephalosporins, fluorochinolones, sulfonamides, and tetracyclines. This suggests that further research on the prevalence of antibiotic resistance genes (ARGs) in foodborne strains is needed, especially from a One Health perspective.

**Keywords:** *Salmonella enterica*; Enteritidis; multidrug-resistant; Derby; foodborne pathogens

#### **1. Introduction**

The annual report on trends and sources of zoonoses published in December 2021 by the European Food Safety Authority (EFSA) and the European Centre for Disease Prevention and Control (ECDC) shows that nearly one in four foodborne outbreaks in the European Union (EU) in 2020 were caused by *Salmonella* spp., which makes this bacteria the most frequently reported causative agent for foodborne outbreaks (694 foodborne outbreaks in 2020) [1].

In the EU 52,702 confirmed cases of salmonellosis in humans were reported and salmonellosis remains the second most commonly reported zoonosis in humans after campylobacteriosis. The three most commonly reported *Salmonella enterica* subsp. *Enterica* serovars in 2020 were *S. enteritidis*, *S. typhimurium*, and monophasic *S. typhimurium*, representing 72.2% of confirmed human cases with known serovar in 2020. Most of the reported salmonellosis foodborne outbreaks were caused by *S. enteritidis* serovar (57.9%). *S. enteritidis* was the predominant serovar in both human salmonellosis cases and reported foodborne outbreaks. Due to the COVID-19 pandemic, total numbers of reported salmonellosis cases as well as foodborne outbreaks are lower compared to previous years' data. Increased use of hygiene equipment, reduced exposure to food served in restaurants and canteens, and more frequent cleaning during domestic food preparations might have had an impact on reported data on salmonellosis. Despite the facts above, trends in salmonellosis occurrence since 2016 data did not reveal statistically significant changes (EFSA December 2021) [1].

**Citation:** Pławi ´nska-Czarnak, J.; Wódz, K.; Kizerwetter-Swida, M.; ´ Bogdan, J.; Kwieci ´nski, P.; Nowak, T.; Strzałkowska, Z.; Anusz, K. Multi-Drug Resistance to *Salmonella* spp. When Isolated from Raw Meat Products. *Antibiotics* **2022**, *11*, 876. https://doi.org/10.3390/ antibiotics11070876

Academic Editor: Piera Anna Martino

Received: 1 June 2022 Accepted: 27 June 2022 Published: 29 June 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Bacteria of the genus *Salmonella* are gram-negative, mostly motile rods, belonging to the *Enterobacteriaceae* family. *Salmonella* spp. is well-established as a pathogen causing gastrointestinal diseases in humans and animals all over the world. Two species are included in the genus *Salmonella*: *Salmonella enterica* spp. and *Salmonella bongori* spp. Almost 99% of the *Salmonella* strains that cause infections in humans or other warm-blooded animals belong to the species *S. enterica*, which includes six subspecies and >2587 serovars [2].

*Salmonella enterica* subsp. *Enterica* includes approximately 1547 serotypes which can cause infections in animals and humans [2]. *Salmonella* infections in humans are usually caused by eating food of animal origin, mostly eggs, poultry meat, or pork [3,4]. The analysis by Gutema et al. (2019) shows that beef and veal can also be a source of *Salmonella* spp. infection due to these animals being potential asymptomatic carriers [3].

Currently, one of the most important health problems in the world is the antimicrobial resistance of *Salmonella* spp. [4,5]. Data from the EU show that the occurrence of resistance in *Salmonella* from pigs, cattle, and broiler chickens largely resembles the appearance of resistance reported for *Salmonella* in various foodstuffs and in people (EFSA [4]).

Multi-drug resistant *Salmonella* constitutes a serious threat to public health through food-borne infections [6–8]. Currently, such multi-drug resistant strains are increasingly isolated from beef and pork [9,10] poultry [11].

Because the problem of antimicrobial resistance became a global problem, in 2003 WHO, together with the Food and Agriculture Organization of the United Nations (FAO) and the World Organization for Animal Health (OIE), began work on creating a List of Critically Important Antimicrobials for Human Medicine (WHO CIA List) [12]. Tacconelli et al. in 2018, pointed out that global research and development strategies should also include antibiotics active against more common community bacteria, such as *Salmonella* spp., *Campylobacter* spp. and *H. pylori*, which are resistant to antibiotics [13]. Therefore, the scope of the new edition of the WHO CIA List, published in 2019, is limited to antibacterial drugs of which most are also used in veterinary medicine. It is very important to use critically important antimicrobials the most prudently in human and veterinary medicine. With accordance monitoring of antimicrobial resistance in food and food-producing bacteria, as defined in Commission Implementing Decision 2013/652/EU, *Salmonella* antibiotics resistance, isolated from food and food-producing animals, should be targeted at broilers, fattening pigs, calves less than 1 year old, and their meat (CID 2013/652/EU).

The aim of our research is to determine the antibiotic resistance of *Salmonella* spp. isolated from raw meat products from beef, pork, and poultry production plants.

#### **2. Results**

Of the 170 meat samples tested, no *Salmonella* spp. were found in beef samples; but, three *Citrobacter braakii* were isolated from them. Only one of the pork samples was positive for *Salmonella* spp. and three *Citrobacter braakii* were isolated from them. Details of any identification difficulties during the isolation of *Salmonella* spp. from meat samples tested were presented by Pławi´nska-Czarnak in 2021 [14]. From the poultry samples, 38 were positive for *Salmonella* spp. All *Salmonella* strains of the isolated species belong to *Salmonella enterica* subsp. *enterica* and represented seven serotypes which shown in Table 1.

The most common serovars from all positive samples were: *S*. Enteritidis (58.97%); *S*. Derby (12.82%) and *S*. Newport (12.82%), which were less frequently isolated; *S*. Infantis (5.13%); *S*. Kentucky (5.13%); *S*. Indiana (2.56%); and *S*. Mbandaka (2.56%) (the details of the results are presented in Table 1).


**Table 1.** The *Salmonella enterica subsp. enterica* variously identified serovars isolated from meat samples of pork and poultry.

Annotation: Antigenic formula according to White-Kauffmann-Le Minor scheme somatic; somatic antigen O (1,9,12 group O9, 1,4,12; 4,12 group O4, 6,8,20; 8,20 group O8, 6,7 group O8, flagellar antigen H phase I and II.

#### *2.1. Antibiotic Susceptibility*

Antibiotic susceptibility testing conducted on the 39 *Salmonella* strains shows that only one strain (*S. enteritidis*) has resistance to two classes of antibiotics (CPH-GEN-STR) whereas 38 strains (64%) were resistant to one or more of the tested antibiotics. However, no resistance against imipenem or colistin was detected. Surprisingly, we detected that 100% of *Salmonella* strains were phenotypically resistant to streptomycin and gentamycin. *Salmonella* strains had intermediate resistance to: amoxicillin (5.13%, *S. Kentucky*, *S.* Newport), cephalexin (30.77%, *S.* Infantis, *S. enteritidis*), ceftiofur (2.56%, *S*. Infantis), neomycin (7.96%, *S.* Newport), enrofloxacin (23.08%, *S.* Infantis, *S.* Mbandaka, *S.* Newport, *S. enteritidis*), norfloxacin (15.8%, *S. derby*, *S*. Indiana, *S*. Enteritidis), doxycycline and oxytetracycline (5.13%, *S*. Derby, *S*. Enteritidis), florfenicol (56.41%, *S*. Mbandaka, *S*. Kentucky, *S*. Newport, *S*. Enteritidis), and trimethoprim-sulfamethoxazole (2.26%, *S*. Derby). In total, 35.9% (14/39) of the strains were resistant to ampicillin, 38.46% (15/39) to amoxicillin, and 7.69% (3/39) to amoxicillin and clavulanic acid. In the case of cephalosporins 46.15% (18/39) of the strains were resistant to cephalexin, 38.46% (14/39) to cefalotin, 97.43% (38/39) to cefapirin, 17.95% (7/39) to cefoperazone, 23.08% (9/39) to ceftiofur, and 12.82% (5/39) to cefquinome. In the case of aminoglycosides, 10.25% (4/39) were resistant to neomycin. In the case of fluoroquinolones, 28.2% (11/39) were resistant to enrofloxacin, 82.05% (32/39) to flumequine, 33.33% (13/39) to marbofloxacin, and 10.25% (4/39) to norfloxacin. A total of 25.64% (10/39) were resistant to tetracyclines, 38.46% (14/39) to florfenicol, 56.41% (22/39) to lincomycin/spectinomycin, and 7.69% (3/39) to trimethoprim/sulfamethoxazole.

#### *2.2. Prevalence of Multiple Drug Resistance*

In our study, most of *S*. Enteritidis showed an MAR index lower than 0.3, whereas one (*S*. Newport) showed an MAR index above 0.5. We observed a high prevalence of multiple antibiotic resistance amongst the isolates where 53.84% of the isolates were MDR strains, with resistance from three to six different classes of antibiotics.

#### *2.3. Antimicrobial Resistance Profile*

All *Salmonella* strains of the isolated species belongs to *Salmonella enterica subsp*. *enterica* and represented seven serotypes (Derby, Indiana, Infantis, Mbandaka, Kentucky, Newport, and Enteritidis). All isolated *Salmonella* were sensitive to imipenem (IMP) and colistin (COL)/polymixin B (PB).

A total of 53.84% *Salmonella* spp. strains isolated from meat were classified as MDR strains that were resistant to the six antibiotic classes: penicillins, cephalosporins, aminoglycosides, fluorochinolones, sulfonamides, and tetracyclines. *S*. Newport (sample 1) presented the most extensive resistance profiles to 17 antibiotics (AMP-AMX-AMX/CL-CFX-CFT-CPH-GEN-NEO-STR-ENR-UB-MRB-NOR-DOX-OXY-TET-LIN/SP), belonging

to 5 classes of antibiotics (β-lactams, aminoglycoside, fluorochinolones, tetracyclines and lincosamides with spectinomycin. In one of *S*. Derby (AMP-AMX-CFX-CFT-CPH-CFP-CFTI-CFQ-GEN-STR-ENR-UB-MRB-FLR-LIN/SP-TR/SMX) and *S*. Newport (AMP-AMX-AMX/CL-CFX-CFT-CPH-GEN-STR-ENR-UB-MRB-NOR-DOX-OXY-TET-LIN/SP), extensive resistance profiles to 16 antibiotics were present. In *S. indiana* (AMX-AMX/CL-CTX-CPH-CFTI-GEN-NEO-STR-DOX-OXY-TET-FLR-LIN/SP-TR/SMX), extensive resistance profiles to 14 antibiotics were present.

The classes to which it presented the highest resistance were β-lactams (AMP, AMX) and beta-lactam/beta-lactamase inhibitor combination (AMX/CL), I generation cephalosporin (CFX-CFT-CPH), III generation cephalosporin (CFTI, CFP), aminoglycosides (GEN-NEO-STR), fluorochinolones (ENR-UB-MRB-NOR), and tetracyclines (DOX-OXY-TET). The most diverse serotype in terms of antimicrobial resistance turned out to be *S*. Enteritidis, in which 13 patterns of resistance were observed. Serovar *S.* Mbandaka showed complete resistance to 9 antibiotics (AMP-AMX-CFX-CFT-CPH-GEN-STR-UB-LIN/SP), and *S.* Infantis showed resistance to 10 antibiotics to varying degrees. The least resistant strain of *S*. Enteritidis was strain from pork meat resistant to 3 antibacterial substances (CPH-GEN-STR), and the most resistance to *S*. Enteritidis was strain 11 from poultry meat (AMP-CFX-CFT-CPH-CFTI-GEN-STR-UB-MRB-FLR-LIN/SP).

For the particular serotypes of *Salmonella enterica* spp. *enterica*, all individual patterns of resistance to multiple antibiotics are presented in Table 2.

The isolates were subjected to antibiotic susceptibility tests against 33 antibiotics belonging to ten different classes using the MIC method Merlin MICRONAUT (MERLIN Diagnostika GmbH, Niemcy) and AST-GN96 CARD and VITEK2 system (Biomerieux, Marcy-l'Étoile, France). The AST card is essentially a miniaturised and abbreviated version of the doubling dilution technique for MICs determined by the microdilution [15]. The multiple antibiotics resistance index (MAR) was performed for isolates showing resistance to more than two antibiotics and is presented in the Table 2 [16].


**Table 2.** Multiple Antibiotic Resistance Index and phenotype pattern of *Salmonella enterica* spp. *enterica* all identified serovars isolates from meat samples of pork and poultry.


**Table 2.** *Cont.*

Letter abbreviations correspond to the individual antibiotics according to list: ampicilln (AMP), amoxicillin (AMX), amoxicillin and clavulanic acid (AMX/CL), cephalexin (CFX), cefalotin (CFT), cefapirin (CPH), cefoperazone (CFP), ceftiofur (CFTI), cefquinome (CFQ), imipenem (IPM), gentamicin (GEN), neomycin (NEO), streptomycin (STR), enrofloxacin (ENR), flumequine (UB), marbofloxacin (MRB), norfloxacin (NOR), docycycline (DOX), oxytetracycline (OXY), tetracycline (TET), florfenicol (FLR), lincomycin/spectinomycin (LIN/SP), trimethoprimsulfamethoxazole (TR/SMX).

#### *2.4. Genotypic Resistance*

The gene *bla*CMY-2 that confers resistance to cefoperazone/ceftiofur was detected in 41.02%, and *bla*SHV in 35.9%. of strains. However, some *Salmonella* spp. strains did not exhibit phenotypic resistance to III generation cephalosporins. In addition, 30.77% of the strains demonstrated the presence of the genes *bla*PSE-1 and 48.72% *bla*TEM that conferred resistance to ampicillin. Most of ampicillin-resistant strains (85.71%) contained *bla*PSE-1 and *bla*TEM, and 14.28% harboured only *bla*TEM gene. The gene *aadB* was detected in eight strains, mainly in *S*. Derby. However, all *Salmonella* spp. strains were phenotypically resistant to gentamicin. The genes *aadA*, *strA*/*strB* that confers resistance to streptomycin was detected in all strains. All of neomycin resistant strains carried *aphA1* and *aphA2* genes. The *tetA* and *tetB* genes were detected in all strains resistant to doxycycline and oxytetracycline. Sulphonamide-resistant strains contained at least one *sul* (1, 2, 3) and *adfR* gene, of which the *sul2* and *adfR1* were the most frequently detected genes. The gene *floR,* that confers resistance to florfenicol, was detected in all strains resistant to florfenicol.

Distribution of the various resistance genes and the prevalence of the corresponding serovars are shown in Table 3.


**Table 3.** Distribution of resistance genes in relation to antimicrobial resistance patterns.


**Table 3.** *Cont.*

Letter abbreviations correspond to the individual antibiotics according to list: ampicilln (AMP), amoxicillin (AMX), amoxicillin and clavulanic acid (AMX/CL), cephalexin (CFX), cefalotin (CFT), cefapirin (CPH), cefoperazone (CFP), ceftiofur (CFTI), cefquinome (CFQ), imipenem (IPM), gentamicin (GEN), neomycin (NEO), streptomycin (STR), enrofloxacin (ENR), flumequine (UB), marbofloxacin (MRB), norfloxacin (NOR), docycycline (DOX), oxytetracycline (OXY), tetracycline (TET), florfenicol (FLR), lincomycin/spectinomycin (LIN/SP), trimethoprimsulfamethoxazole (TR/SMX).

#### **3. Materials and Methods**

#### *3.1. Sampling*

A total number of 190 raw meat samples (60 beef, 60 pork, and 70 poultry) were obtained from three sources within the meat industry, such as cuttings of beef, pork and poultry carcasses in central Poland. All samples were obtained from carcass parts of animals recognised as healthy: the tissues and organs of which were classified by the veterinary inspection as fit for human consumption. All samples were considered a single sample, weighing at least 200 g for each type of meat. The meat samples were collected randomly, using an aseptic technique and packed into sterile bags, which were labeled. All samples were transported to the laboratory in refrigerated containers at a temperature 4 ◦C and processed within five hours.

#### *3.2. Salmonella spp. Isolation and Identification*

*Salmonella* spp. from all samples were isolated in accordance with PN-EN ISO 6579- 1:2017-04 Microbiology of the food chain—Horizontal method for the detection, enumeration and serotyping of *Salmonella*—Part 1: Detection of *Salmonella* spp. (ISO 6579-1:2017). Samples were pre-enriched: for pork and beef samples, the 10 g of each sample was mixed with 90 mL Buffered Pepton Water (GRASO, Gdansk, Poland), and the 25 g of each poultry meat sample was mixed with 225 mL BPW with a temperature of 25 ◦C (±3 ◦C) in a sterile stomacher bag (Whirl-Pak, Nasco, Madison, WI, USA), and crushed for 2 min. After that, they were incubated at 37 ◦C for 18 h. Selective proliferation of *Salmonella* spp. was carried out using the MSRV agar (Modified semi-solid Rappaport-Vassiliadis—MSRV agar, GRASO, Poland) with 0.1 mL of the pre-enriched culture as three equally spaced spots on the surface of the MSRV agar were incubated at 41.5 ◦C for 24 h and 1 mL of the culture obtained was put to a tube containing 10 mL of Muller-Kauffmann tetrathionate-novobiocin (MKTTn) broth (GRASO, Gdansk, Poland) and incubated at 37 ◦C for 24 h. From the positive growth obtained on the MSRV agar, it was chosen as the furthest point of opaque growth from the inoculation points, and picked up a 1 µL loop and was inoculated on two selective agars: XLD (Xylose Lysine Deoxycholate agar, GRASO, Gdansk, Poland) and BGA (Brilliant Green agar, OXOID, Hampshire, UK). From the liquid culture obtained in the MKTTn, broth was picked up of a 10 µL loop and spread on XLD agar and BGA agar to obtain well-isolated colonies. All selective agars were incubated at 37 ◦C for 24 h (±3 h). *Salmonella*-suspect colonies were transferred to Nutrient agar (GRASO, Gdansk, Poland) to obtain the pure culture for further testing.

#### 3.2.1. DNA Preparation and Presumptive Salmonella Confirmation

The Real-time PCR method, and an amplification based on detection gene specific for *Salmonella*, was used to confirm presumptive identification. DNA for real-time PCR was extracted from bacterial cells, using commercial Kylt® DNA Extraction-Mix II (Anicon, Emstek, Germany). For the detection of *Salmonella* spp. *commercial* Kylt® *Salmonella* spp. (Anicon, Germany) was used, and for the simultaneous detection of *Salmonella* Enteritidis, the Typhimurium commercial Spp-Se-St PCR (BioChek, Reeuwijk, The Netherland) kit was used. The Real Time PCR method to detect *Salmonella* was performed according to the manufacturer's instructions with using Applied Biosystems 7500 Fast Real-Time PCR System (Thermo, Waltham, MA, USA).

#### 3.2.2. Biochemical Strain Identification

For identification of the strains, two commercially available biochemical tests were used according to the manufacturer's instructions: Api20E (BioMérieux, Marcy-l'Étoile, France) and the VITEK® 2 GN cards (Biomerieux, Marcy-l'Étoile, France).

#### 3.2.3. Serological Testing

Serotyping was performed according to the White-Kauffmann-Le Minor scheme. Serological testing was carried out by slide agglutination with commercial H poly antisera

to verify the genus of *Salmonella enterica* (IBSS Biomed, Lublin, Poland), O group antisera to determine the O group, (IBSS Biomed, Poland), and H phase and H factor antisera to determine the H phase and H factor (IBSS Biomed, Lublin, Poland, Bio-Rad, Chercules, CA, USA), as described in Pławi ´nska-Czarnak [17].

#### *3.3. Antimicrobial Sensitivity Testing*

Each *Salmonella* strain was first subcultured as described previously. From an 18–24 h culture, a DensiCHEK Plus (Biomerieux, Marcy-l'Étoile, France) instrument was used to perform a suspension with a 0.5 McFarland range. Then, 145 µL of this inoculum was transferred to another VITEK® tube containing 3 mL 0.45% saline. The card was automatically filled by a vacuum device and automatically sealed. It was manually inserted in the VITEK2 Compact reader-incubator module, and every card was automatically subjected to a kinetic fluorescence measurement every 15 min. This is an automated test methodology based on the MIC technique reported by MacLowry and Marsh [18], and Gerlach [19]. A loop of the suspension was also inoculated onto blood agar (GRASO, Poland) for the purity check.

Antimicrobial susceptibility was assessed by determining the MIC values using a 96 well MICRONAUT Special Plates with antimicrobials: β-lactams/aminopenicillin (amoxicillin—AMX, amoxicillin and clavulanic acid—AMX/CL), β-lactams/I generation cephalosporins (cephalexin—CFX, cephapirin—CPH), β-lactams/III generation cephalospo rins (ceftiofur—CFTI), β-lactams/IV generation cephalosporins (cefquinome—CFQ), βlactams/penicillin cloxacillin—CLO, penicillin G—PG, nafcillin—NAF), aminoglycoside (gentamicin—GEN, neomycin—NEO, streptomycin—STR), polymyxins (colistin—COL), fluorochinolones (enrofloxacin—ENR, norfloxacin—NOR), tetracyclines (doxycycline— DOX, oxytetracycline—OXY), macrolides erythromycin—ERY, tylosin—TYL), florfenicol— FLR), lincosamides (lincomycin—LIN, lincomycin/spectinomycin—LIN/SP), trimethoprimsulfamethoxazole—TR/SMX, tiamulin—TIA, tylvalosin—TYLV (MERLIN Diagnostika GmbH, Bremen, Niemcy). Simultaneously, antimicrobial susceptibility was assessed by determining the MIC values using a VITEK® 2 System and AST-GN96 cards for Gramnegative bacteria (BioMérieux). The AST card is essentially a miniaturised and abbreviated version of the doubling dilution technique for MICs determined by the microdilution method [c].

The MERLIN antibiotics concentration (µg/mL) is as follows: amoxicillin—0.25, 2, 4, 8, 16; amoxicillin and clavulanic acid—4/2, 8/4, 16/8; cephalexin—8, 16; cephapirin—8, ceftiofur—2; cefquinome—2, 4; cloxacillin—2; penicillin 0.0625, 0.125, 2, 8; nafcillin— 2; gentamicin—4, 8; neomycin—8; streptomycin—8; colistin—2; enrofloxacin—0.5, 2; norfloxacin—1, 2; doxycycline—2, 4, 8; oxytetracycline—2, 4, 8; erythromycin—0.25; 0.5, tylosin—TYL; florfenicol—2, 4; lincomycin—2, 8; lincomycin/specinicin—8, 32; trimethopri m-sulfamethoxazole—2/38; tiamulin—16; and tylvalosin—2, 4.

With using AST-GN96 susceptibility for β-lactams/aminopenicillin (ampicillin—AMP, amoxicillin and clavulanic acid—AMX/CL), β-lactams/I generation cephalosporins (cefalexin -CFX), β-lactams/III generation cephalosporins (cefalotin—CFT, cefoperazone CFP), β-lactams/III generation cephalosporins (ceftiofur—CFTI), β-lactams/IV generation cephalo sporins (cefquinome—CFQ), carbapenems (imipenem—IPM), polymyxin (polymixin B -PB), aminoglycoside (gentamicin—GEN, neomycin—NEO), fluorochinolones (enrofloxacin— ENR), flumequine—UB), marbofloxacin—MRB), tetracycline -TET, florfenicol—FLR, and trimethoprim/sulfamethoxazole (TR/SMX), were assessed.

The AST-GN96 antibiotics concentration (µg/mL) is as follows: ampicillin—4, 8, 32; amoxicillin and clavulanic acid—4/2, 16/8, 32/16; cephalexin—8, 16, 32; efalotin—2, 8, 32; cefoperazone 4, 8, 32; cefquinome—0.5, 1.5, 4; imipenem 1, 2, 6, 12; polymixin B 0.25, 1, 4, 16; gentamicin—4, 16, 32; neomycin—8, 16, 64; enrofloxacin—0.25, 1, 4; flumequine—2, 4, 8; marbofloxacin—1, 2; tetracycline—2, 4, 8; florfenicol—1, 4, 8; trimethoprim/sulfamethoxazo le—1/19, 4/76, 16/304.

The MICs were interpreted according to Clinical and Laboratory Standards Institute (CLSI) and FDA breakpoints (CLSI M100-ED28, 2018). The AST card is essentially a miniaturised and abbreviated version of the doubling dilution technique for MICs determined by the microdilution method.

#### *3.4. Determination of Antibiotics Resistance Profile of Salmonella spp. Isolates*

In order to calculate multiple antibiotics resistance, we used the formula according to the Akinola 2019, MAR index [16]:

> MAR = Number of resistance to antibiotics Total number of antibiotics tested

Detection of Antimicrobial Resistance Genes by PCR

Mueller–Hinton agar was used to culture the bacterial isolates overnight at 35 ◦C. Bacterial DNA isolation was performed using a standard bacterial DNA isolation Kylt® DNA Extraction-Mix II (Anicon, Emstek, Germany). Eighteen resistance genes (*aadA*, *strA/strB*, *aphA1*, *aphA2*, *aadB*, *tetA*, *tetB*, *sul1*, *sul2*, *sul3*, *dfrA1*, *dfrA10*, *dfrA12*, *floR*, *bla*TEM, *bla*SHV, *bla*CMY-2, *bla*PSE-1 and *bla*CTX-M) were analysed by conventional PCR, using specific primer pairs in multiplex or a single PCR reaction. The primer sequences predicted PCR product sizes and references shown in Table 4.

**Table 4.** Description of primer sets, annealing temperature and product size for the molecular gene identification [20–22].


#### *3.5. Statistical Assessment*

Statistical testing was performed with Statistica software, version 13.1. Descriptive statistics were computed to determine the proportions of isolates resistant to different antimicrobial agents. Chi square tests were adopted for the determination of statistical significance of differences between the proportions.

#### **4. Discussion**

Our data show that poultry meat is a relevant source of *Salmonella*, and the prevalent serovar was Enteritidis (56.41%). We estimate the antibiotic susceptibility profiles of *Salmonella* strains, and we found a high rate of strains showing at least one phenotypic resistance. In our study, sensitivity to 25 antibiotics were assessed. Penicillins (cloxacillin, penicillin G, nafcillin), macrolides (erythromycin, tylvalosin), lincomycin, tiamulin, and tylvalosin were excluded from analysis, due to a natural lack of activity against *Salmonella*.

The results of the antibiotic resistance indicate that the *Salmonella* spp. strains isolated from meat can be categorized as resistant to MDR: that is, bacteria exhibiting resistance to one or more antibiotics from three or more classes of antibiotics. These bacteria are resistant to β-lactams, aminoglycosides, cephalosporins, fluorochinolones, sulfonamides, and tetracyclines. Resistance to third generation cephalosporins exhibited by the strains isolated from meats represents a concern, because these antibiotics are used for salmonellosis treatment in human, thus rendering the transmission of resistant bacteria a public health problem. All strains isolated from meat were resistant to gentamycin, which is one of the major antibiotics used in the treatment of urinary infections in humans, and were resistant to streptomycin used to treat tuberculosis and *Burkholderia* infection. Although streptomycin is an aminoglycoside and not used for *Salmonella* treatment, streptomycin resistance has been widely used as an epidemiological marker. Resistance to streptomycin is analogous to the phenotypic characteristics observed in multi-drug resistance to ampicillin, chloramphenicol, streptomycin, sulfonamides, and tetracyclines [23,24]. Regarding the resistance to ampicillin (35.89%), previous studies from different countries report highest resistance rates [25].

Moreover, *Salmonella* Derby from meat shows resistance to cefequinome, fourth generation cephalosporins, and antibiotics used in the treatment of mastitis and bovine pneumonia. In *Salmonella* Derby and Indiana (both in the BO4 group), we found resistance against sulphonamides, a class of antibiotics used in severe *Salmonella* infections. We also observed resistance to third generation cephalosporins (cefoperazone and ceftiofur) in four *Salmonella* Derby strains isolated from poultry meat. In addition, a high percentage of strains (Indiana, Infantis, Kentucky, and Newport) showed resistance to tetracyclines (24.64%), despite the fact that, in 2006, the European Union, imposed a ban on the non-therapeutic use of antibiotics important to humans, such as tetracyclines, in animal treatment. A total of 53.84% of tested strains showed an MDR profile with resistance to one or more antibiotics from three or more classes of antibiotics. On the other hand, all the *Salmonella* spp. strains were susceptible to imipenem, which is similar to the result reported previously [26]. Carbapenems are the final choice of antibiotics used in the treatment of salmonellosis when the bacteria exhibit resistance to antibiotics, such as ciprofloxacin and third generation cephalosporins.

These data are alarming for consumers because of the real possibility of an infection with an MDR strain in food, but also because these strains showed resistance to antibiotic classes crucial in human medicine, such as beta-lactamases.

Finally, because these antibiotic phenotypes can be conferred by several ARGs, the detection of resistance genes was performed in order to confirm phenotypic pattern.

In *Salmonella*, the main mechanism of resistance to β-lactams is the acquisition *bla* gene encodes beta-lactamase hydrolytic enzymes, which inactivate the antibiotic [27]. Extended-spectrum beta-lactamases (ESBLs), which inactivates first-, second-, and thirdgeneration cephalosporins and penicillins, and are encoded multi-variant *bla*TEM, *bla*SHV and *bla*CTX-M genes [28]. The *bla*CTX-M genes encode for the extended-spectrum of βlactamases (ESBLs) were not present in analysed strains. These types of β-lactamases are

active against cephalosporins and monobactams (but not carbapenems), and are currently of great epidemiological and clinical interest. The *bla*SHV gene was found to be the most prevalent gene amongst our isolates, mainly in *S. enteritidis*. The *bla*SHV gene is associated with *Enterobacteriaceae* in causing nosocomial infections, but also in isolates from different sources (human, animal, and environment). The gene *bla*CMY-2 encodes an extendedspectrum beta-lactamase that is responsible for hydrolyzing the β-lactam ring that was detected in 35.89% of strains. However, some *Salmonella* spp. strains did not expose phenotypic resistance to this antibiotic. This gene confers resistance to ampicillin, ceftiofur, cefoperazone and is associated with mobile elements, thus increasing the probability of transmission between bacteria [29]. In our study, 28.21% of the strains demonstrate the presence of the genes blaPSE-1 and *bla*TEM that encode β-lactamases that confer resistance to ampicilin. In a study conducted in Colombia, 69.4% of the strains isolated from broiler farms had both genes; thus, a frequency was higher than that found in the present study [30]. Five *S. derby*, one *S. enteritidis*, and all *S. Kentucky* that were phenotypically resistant to ampicillin and third generation cephalosporins, showed the presence of the genes *bla*PSE-1, *bla*TEM, *bla*CMY-2, but not and *bla*CTX-M. The streptomycin resistance gene *aadA* and *strA/strB* were detected in all of the strains. Interestingly, White et al. [31] showed that *Salmonella* strains isolated from meat that had the *aadA* genes but were susceptible to streptomycin, probably due to gene silencing. The gene *sul2* encodes DHPS (dihydropteroate synthase) was found in 7.69% of the strains (*S. derby* and *S. indiana*). In a previous study, the gene *sul1* is reported to be the most prevalent (57.1%) [24], whereas in the present study, it was found in only 5.13% of the strains. Trimethoprim resistance is mediated by the expression of the enzyme DHFR (dihydrofolate reductase) and is encoded by the *dfrA1* gene that was detected in 7.69% of the strains. In general, the strains that were resistant to trimethoprimsulfamethoxazole showed the *sul* (*sul1*, *sul2* or *sul3*) and *dfrA* (*dfrA1*, *dfrA12*) resistance genes, mainly in *S*. Derby. However, all strains were resistant to this antibiotic. This resistance may be mediated by other resistance genes, which are not assessed in this study. In *S. derby*, *S*. *indiana*, *S*. *newport*, and in two *S*. Enteritidis, the *floR* gene was detected. This gene encodes an efflux pump that confers resistance to amphenicols, which has been reported in the genomic island of *Salmonella* (SGI1) [32].

Our data are very alarming, since all of our strains came from food samples, mainly poultry meat for human consumption. Thermal processing of these products may reduce the risk of foodborne disease, but ARGs can be transferred to the gut microbiota and transfer resistance to other bacteria [33]. Therefore, our data are in line with recommendations, which confirm how important it is in the monitoring and control of antibiotic resistance to assess the presence or absence of ARGs in foodborne strains, especially in a One Health approach that recognises the circularity of human, animal, and environmental health.

#### **5. Conclusions**

The *Salmonella* spp. strains exhibited resistance to multiple antibiotics, as well as multiple genes associated with them. A high resistance rate to multiple antibiotics combined with multiple ARGs in isolates from raw meat, as revealed in this study, suggests that the situation is alarming in where irrational use of antibiotics is combined with inadequate surveillance and facilities to detect MDR. Continued monitoring of antimicrobial resistance in *Salmonella* strain collection along the food chain is required so that comparisons of antimicrobial resistance from the different origins can be effectively performed.

**Author Contributions:** Conceptualization, J.P.-C. and M.K.-S.; methodology, J.P.-C., K.W. and M.K.- ´ S.; validation, J.P.-C., K.W., M.K.- ´ S. and J.B.; formal analysis, J.P.-C., K.W., M.K.- ´ S., T.N. and Z.S.; ´ investigation, J.P.-C., P.K. and K.A.; resources, J.P.-C.; data curation, J.P.-C., K.W. and M.K.-S.; writing— ´ original draft preparation, J.P.-C., K.W., M.K.-S., T.N., Z.S. and J.B.; writing—review and editing, ´ J.P.-C., K.W., M.K.-S., K.A., Z.S. and P.K.; visualization, J.P.-C., K.W. and M.K.- ´ S.; supervision, J.P.-C. ´ and K.A.; project administration, J.P.-C.; funding acquisition, K.A. and P.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to their containing information that could compromise the image of the meat processing plants.

**Acknowledgments:** Special thanks to Jolanta Przybylska for help with the laboratory work and Maria Górka for help with editing the text.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**

