Prevalence of Antibiotic Resistance in Salmonella Serotypes Concurrently Isolated from the Environment, Animals, and Humans in South Africa: A Systematic Review and Meta-Analysis

One of the main global concerns is the usage and spread of antibiotic resistant Salmonella serovars. The animals, humans, and environmental components interact and contribute to the rapid emergence and spread of antimicrobial resistance, directly or indirectly. Therefore, this study aimed to determine antibiotic resistance (AR) profiles of Salmonella serotypes isolated from the environment, animals, and humans in South Africa by a systematic review and meta-analysis. The preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines were followed to search four databases for studies published from 1980 to 2021, that reported the antibiotic resistance profiles of Salmonella serotypes isolated in South Africa. The AR was screened from 2930 Salmonella serotypes which were isolated from 6842 samples. The Western Cape province had high pooled prevalence estimates (PPE) of Salmonella isolates with AR profiles followed by North West, Gauteng, and Eastern Cape with 94.3%, 75.4%, 59.4%, and 46.2%, respectively. The high PPE and heterogeneity were observed from environmental samples [69.6 (95% CI: 41.7−88.3), Q = 303.643, I2 = 98.353, Q-P = 0.045], animals [41.9 (95% CI: 18.5–69.5), Q = 637.355, I2 = 98.745, Q-P = 0.577], as well as animals/environment [95.9 (95% CI: 5.4−100), Q = 55.253, I2 = 96.380, Q-P = 0.300]. The majority of the salmonella isolates were resistant to sulphonamides (92.0%), enrofloxacin and erythromycin (89.3%), oxytetracycline (77.4%), imipenem (72.6%), tetracycline (67.4%), as well as trimethoprim (52.2%), among the environment, animals, and humans. The level of multidrug-resistance recorded for Salmonella isolates was 28.5% in this review. This study has highlighted the occurrence of AR by Salmonella isolates from animals, humans, and environmental samples in South Africa and this calls for a consolidated “One Health” approach for antimicrobial resistance epidemiological research, as well as the formulation of necessary intervention measures to prevent further spread.


Introduction
Salmonella is a genus of Gram-negative rod-shaped bacterium of the Enterobacteriaceae family that includes more than 2700 Salmonella serotypes [1], found in three species, Salmonella bongori, Salmonella subterranean, and S. enterica [2]. Furthermore, S. enterica is divided into six subspecies namely: Indica (subsp. VI), arizonae (subsp. IIIa), houtenae (subsp. IV), enterica (subsp. I), salamae (subsp. II), and diarizonae (subsp. IIIb) [2]. Depending on the serovars and hosts, these bacterial species can cause bacteremia, typhoid fever, gastroenteritis, and non-typhoidal salmonellosis in humans and animals [3,4]. Salmonellosis is one of the most frequent and economically significant zoonotic infections of humans. In addition, it impacts all of the domestic animal species, as well as the wildlife.

Study Characteristics of Eligible Studies
The following are characteristics of the studies that are eligible: Articles published primarily on the quantitative prevalence of AR by Salmonella serovars/isolates in humans, animals (chicken, ducks, cows, pigs, goats, horses, and sheep), and the environment in South Africa; the type of samples and method of diagnostics used; the exact number of samples as well as the number of positives tested; articles reported in English only; and antibiotic resistance. All of the journal papers were published between the years 1980 and 2020.
All of the studies included in this review were derived from all of the provinces in South Africa. Eastern Cape (n = 6) was observed to have the highest number of studies All of the studies included in this review were derived from all of the provinces in South Africa. Eastern Cape (n = 6) was observed to have the highest number of studies followed by North West (n = 4), KwaZulu-Natal (n = 4), Western Cape (n = 3), Gauteng (n = 3), Limpopo (n = 2), and Northern Cape (n = 1) as the least (Figure 2). The most common method for determining AR profiles of Salmonella serotypes isolated from all of the studies included in this systematic and meta-analysis study was disk diffusion. Of the 20 included studies, six were environmental samples, nine were samples from animal sources, and three included both animal and environmental samples. One study examined animals, humans, and environmental samples, and one study included samples from both animals and humans. = 3), Limpopo (n = 2), and Northern Cape (n = 1) as the least (Figure 2). The most common method for determining AR profiles of Salmonella serotypes isolated from all of the studies included in this systematic and meta-analysis study was disk diffusion. Of the 20 included studies, six were environmental samples, nine were samples from animal sources, and three included both animal and environmental samples. One study examined animals, humans, and environmental samples, and one study included samples from both animals and humans.  Table 1 displays a summary of the number of samples processed per subgroup, and the number of those samples that were positive for Salmonella serovars. Overall, the assessment of South African data revealed a pooled prevalence estimate (PPE) of 32.3% of processed samples positive for Salmonella, with a range between 41.9% and 95.9%. At 95% CI, 30.9-33.9 samples in South Africa would be expected to be positive for Salmonella serovars, and the range could be between 5.4% and 100%.

Pooling and Heterogeneity Estimates of Salmonella Species
Prevalence Based on Provinces, Study Years, and Diagnostic Techniques Table 1 displays a summary of the number of samples processed per subgroup, and the number of those samples that were positive for Salmonella serovars. Overall, the assessment of South African data revealed a pooled prevalence estimate (PPE) of 32.3% of processed samples positive for Salmonella, with a range between 41.9% and 95.9%. At 95% CI, 30.9-33.9 samples in South Africa would be expected to be positive for Salmonella serovars, and the range could be between 5.4% and 100%.
The analyses of the PPE in animals, environment, and animal/environment analyses all revealed a significant heterogeneity. For the environment, pooled heterogeneity was observed

Antibiotic Resistance Profile of Salmonella spp. Isolates to Antibiotics
The PPE of antibiotic resistance for Salmonella serovars included in this meta-analysis (Table 1) Table 2. Ampicillin was the most tested antibiotic with 13 studies. Overall, the MDR recorded for Salmonella spp. was 28.5% [95% CI: 11.2-55.7]. The overall effect estimates and their accompanying 95% confidence intervals (CI) did not overlap in any of the meta-analyses performed ( Figure 3).  The AR was screened from a total of 8843 Salmonella serovars, collected from 20 studies. A total of 35 antibiotics were tested on these isolates. The majority of the isolates tested for AR vary by provinces (Table 3), which included Eastern Cape (n = 2439), followed by KwaZulu-Natal (n = 2094), Limpopo (n = 1673), Northern Cape (n = 1069), Gauteng (n = 793), Western Cape (n = 606), and finally by North West (n = 169). In terms of the antibiotic profile tested, ampicillin was the most tested antibiotic in all of the provinces. A total of five studies reported multidrug resistance, which is defined as resistance to at least three classes of antibiotics. The AR was screened from a total of 8843 Salmonella serovars, collected from 20 studies. A total of 35 antibiotics were tested on these isolates. The majority of the isolates tested for AR vary by provinces (Table 3), which included Eastern Cape (n = 2439), followed by KwaZulu-Natal (n = 2094), Limpopo (n = 1673), Northern Cape (n = 1069), Gauteng (n = 793), Western Cape (n = 606), and finally by North West (n = 169). In terms of the Antibiotics 2021, 10, 1435 7 of 14 antibiotic profile tested, ampicillin was the most tested antibiotic in all of the provinces. A total of five studies reported multidrug resistance, which is defined as resistance to at least three classes of antibiotics.

Discussion
In this review, the meta-analysis was used to evaluate the prevalence of AR in Salmonella serovars isolated from the environment, animals, and humans. This study examined 20 published articles from South Africa on AR in Salmonella serovars.
The majority (95%) of the studies included in this review were conducted between 2010 to 2021. Based on our meta-analysis, the PPE of AR against sulfonamide was 92%. Our findings correlate with other studies, which reported an increased development of AR against sulfonamide in Spain, Ghana, Central African Republic, Morocco, and Italy with 38.1% 72.4%, 29%, 25%, and 69%, respectively [14][15][16][17][18]. Sulfonamides resistance can arise from chromosomal dihydropteroate synthase (DHPS) mutations or from the acquisition of DHPS drug resistance genes, whose products have lower affinity for sulfonamides [19,20]. The presence of sul genes, which encode dihydropteroate synthase in a form that is not inhibited by the drug, is linked to sulfonamide resistance in Salmonella serotypes [20]. Generally, sulfonamide is encoded by three sul genes which are sul1, sul2, and sul3 [20]. The sul1 and sul2 genes are the most frequently reported genes found among sulfonamideresistant isolates, and also found in plasmids of other Salmonella species which are still common in Gram-negative bacterial plasmids [19][20][21].
In this meta-analysis, the PPE of tetracycline was 67.4%. Our results are consistent with the findings of the studies conducted in Iran, United States, Saudi Arabia, and China, which reported an increasing development of AR against tetracycline of 66.9%, 63%, 90.71%, and 43%, respectively by Salmonella isolates [22][23][24][25]. In South Africa, over 70% of the antibiotics used in livestock production are available over the counter [26], and this contributes to the increased prevalence of AR in the country. Tetracyclines have been an important class of antibiotics in the health and production of food animals for decades, as they are used in veterinary medicine for illness prevention and control [27]. They have a broad spectrum and have good activity against Gram-positive and Gram-negative bacteria that cause acute infections [28].
Tetracyclines act as protein synthesis inhibitors by binding to the small ribosomal subunit and preventing aminoacyl-tRNA from attaching to the protein synthesis complex [22]. The complicated syntheses of several systems are capable of conferring tetracycline resistance, including efflux pumps, enzymatic inactivation, and mutations [22,27,29,30]. In South Africa, tetracyclines are the most often used or over-used antibiotics in the livestock production [31]. This is due to the fact that they are relatively affordable and widely available as over-the-counter veterinary medications [22,31,32].
According to a survey conducted by Eagar et al. [33], 16.7% and 12.4% of tetracyclines and sulphonamides were commonly used antibiotics in animals in South Africa between the years 2002 and 2004. The Stock Remedies Act of 1947 in South Africa consents tetracycline to be purchased over the counter without a prescription from a veterinarian [26].
The PPE of AR for streptomycin by Salmonella isolates was 37.7%. Our results are lower than the observations from other studies for resistance against streptomycin in Saudi Arabia, Italy, and Egypt with prevalence of 80%, 95%, 65.5%, and 20.1% [12,[34][35][36]. Among the different varieties of aminoglycoside adenylyltransferase coding genes, aadA provides resistance to streptomycin in Salmonella [37,38]. Streptomycin is one of the most commonly antibiotic used by farmers and veterinarians in the country [31]. The widespread use of antibiotics in the treatment of bacterial infections in plants and animals could explain the high prevalence of streptomycin resistance [26].
The current review also revealed that most of the Salmonella isolates have high PPE for AR against erythromycin (89.3%). Similar findings were observed previously in studies from Nigeria, Saudi Arabia, Ghana, and South Africa [Gauteng] with 100%, 100%, 86.0%, and 94.9% resistance to erythromycin [31,34,39,40].
Amoxicillin and ampicillins are two of the most common antibiotics used globally to treat salmonellosis [41,42]. In this study, the PPE for AR by Salmonella isolates for amoxicillin and ampicillins were lower as compared to the other antibiotics with 19.2% and 38.6%, respectively. Our results were much higher than 0.43% [43] and 0% [44], which were recorded earlier. Antimicrobials, such as ampicillin and others, are still widely used in livestock production in South Africa for growth promotion, prophylaxis, and treatment [26], and in hospitals [43].
The antibiotic imipenem combined with cilastatin is commonly used to prevent its degradation by renal tubular dehydropeptidase. Resistance to this medicine has been reported in a number of bacterial species, adding to the worldwide burden of antibiotic resistance in humans. The PPE of imipenem 72.6% in this study was found to be much higher than the one conducted by Ng and Rivera [45], whereby none of the Salmonella enterica isolates were resistant to imipenem.
A total of five studies reported multidrug resistance for the current study. In addition, our results are in line with previous study [46] where multidrug resistance (MDR) was exhibited.
This is the first comprehensive study addressing AR for Salmonella serovars isolated from the environment, humans, and animals in South Africa. This study selected peer reviewed and high-quality papers to provide a summary of unbiased findings on the prevalence of AR in Salmonella species isolated from the environment, humans, and animals in South Africa. In addition, it reviews the current knowledge and aids in identifying areas where appropriate evidence is lacking, resulting in new research questions/topics. With respect to provinces, there are some provinces where few or no Salmonella AR studies were conducted. Between the study periods, there was a greater disparity in study outputs, from one publication in 2000-2010 to 19 articles published between 2010-2021. In terms of overall studies, only two AR studies were conducted on Salmonella infecting humans.
Despite the fact that we have systematized data on the occurrence of AR of Salmonella species, this study has the following limitations: The PPE of animals/humans and animals/humans/environment were not calculated since there is a single published article on each. The number of studies from some provinces were unusually high, which may have influenced the overall estimate. Moreover, few reports for Salmonella AR from humans were observed, whereby only two studies were undertaken in two provinces, namely, Western Cape and KwaZulu-Natal. Furthermore, the AR genes were not included from this meta-analysis due to the few number of studies conducted.

Study Design
The current systematic review and meta-analysis were conducted in accordance with the preferred reporting items for systematic reviews and meta-analyses (PRISMA 2020) guidelines for selection criteria, literature search, statistical analysis, and data extraction [47].

Search Strategy
The literature search was performed on four databases, namely, (https://www. ncbi.nlm.nih.gov/pubmed, accessed on 27-28 May 2021), ScienceDirect (https://www. sciencedirect.com, accessed on 5-7 June 2021), African Journal Online (https://www.ajol. info/index.php/ajol, accessed on 11-12 June 2021), and Scopus (https://www.scopus.com, accessed on 22-24 June 2021) from June to August 2021. The following keywords were used to search for the articles: Antibiotic resistance AND Antibiotic AND drug resistance AND bacteria resistance AND multi-drug resistance AND Salmonella species AND Human OR animal AND Environment AND salmonellosis AND South Africa. We conducted our last search on 24 June 2021.

Inclusion and Exclusion Criteria
The following inclusion criteria had to be met for all of the eligible studies included in the review: Articles published primarily on the quantitative prevalence of Salmonella spp. in the environment, animals, and humans in South Africa; the type of samples and method of diagnostics used; the exact number of samples as well as the number of positives tested; and articles reported in English only on antibiotic resistance, published between January 1980 and August 2021. The studies were excluded if they were not undertaken in South Africa (Supplementary Materials). In addition, books and book chapters were excluded. Moreover, a review, a smaller sample size, and articles not reported in English, were discarded, as well as articles not published between January 1980 and August 2021.

Data extraction and Data Collection
The titles and abstracts of journal articles were examined and downloaded, and the full versions of potentially relevant articles were obtained to determine eligibility. Data including names of authors, publication year, location, total sample size, and standard methods to detect the antibiotic resistant were collected from each publication independently. Then, they were entered into a spreadsheet (Microsoft Excel ® ), tables, and a word document template. Only Salmonella species/isolates/serotypes specific journal articles were included in the meta-analysis.
The studies with insufficient data were excluded. In addition, review articles and studies with an abstract, but without a full text, were excluded. The papers that included the first author, year of publication, year of study, sampling, region, sample size, target population, method, drug susceptibility test, sample type, number of isolates, and number of multidrug resistance strains were considered and included in this study. For inclusion in the study, we considered all of the standard guidelines, although only the clinical and laboratory standards institute (CLSI) guidelines were used in all of the included studies (Table 4).

Data Analysis and Assessment of Risk of Bias
The Begg's and Egger's tests were used to investigate the possibility of propagation bias. For each study, the prevalence, effect size, and 95% CIs were calculated as a point estimate.
A comprehensive meta-analysis was used for all of the statistical analyses, version 3.0 (CMA) program (https://www.meta-analysis.com/, accessed on 18 August 2021). The software was used to generate the pooled estimates, Cochran's Q, p-values, and forest plots. I 2 (level of inconsistency) was used to assess the study heterogeneity (Cochran's Q). The I 2 values above 75% were regarded to have a high degree of heterogeneity [64]. For each study, the prevalence, effect size, and 95% CIs were calculated as a point estimate. Exploring the funnel plots was used to assess the publication bias.

Conclusions
This systematic review and meta-analysis provided an overview of published Salmonella serovar AR scientific data from humans, animals, and environmental samples in South Africa. The higher PPE rate of Salmonella isolates were tested. In addition, expressing AR was observed in the Western Cape province. Disk diffusion was the most prevalent method for identifying the antibiotic resistance profiles of Salmonella serotypes isolated from all of the studies. The MDR has been identified as a serious and growing issue, according to the findings of this study. Our results revealed the highest PPE of Salmonella AR to sulphonamides, followed by enrofloxacin and erythromycin. In addition, this finding calls for the restricted usage of this group of antibiotics. According to our data analysis, the most tested antibiotics against Salmonella isolates are tetracycline, ciprofloxacin, chloramphenicol, ampicillin, streptomycin, gentamicin, erythromycin, and kanamycin. This study highlights the lack of published Salmonella serovar AR scientific data from humans. The development of AR to commonly prescribed antibiotics is very common, whilst it appears that surveillance has a lot of gaps whereby there has been no studies on AR in two provinces, namely, Free State and Mpumalanga. To stop the spread of Salmonella AR, control strategies should be strengthened. To describe the epidemiology of the serotypes across the country, large-scale investigations are required. Control practices should be strengthened to slow the spread of AR in South Africa and there is a need for a "One Health" collaborative research between the animal and human health sector, as well as the environmental sector on the epidemiology of AR development and necessary interventions.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10.3 390/antibiotics10121435/s1, Table S1: Checklist of items to include when reporting a systematic review or meta-analysis, Figure S1: Forest plot of the meta-analysis of overall studies on the prevalence of Salmonella isolates in South Africa. Funding: This review received no external funding.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

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