Emergence and Spread of Extended Spectrum β-Lactamase Producing Enterobacteriaceae (ESBL-PE) in Pigs and Exposed Workers: A Multicentre Comparative Study between Cameroon and South Africa

Extended spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE) represent a significant public health concern globally and are recognized by the World Health Organization as pathogens of critical priority. However, the prevalence of ESBL-PE in food animals and humans across the farm-to-plate continuum is yet to be elucidated in Sub-Saharan countries including Cameroon and South Africa. This work sought to determine the risk factors, carriage, antimicrobial resistance profiles and genetic relatedness of extended spectrum β-lactamase producing Enterobacteriaceae (ESBL-PE) amid pigs and abattoir workers in Cameroon and South Africa. ESBL-PE from pooled samples of 432 pigs and nasal and hand swabs of 82 humans were confirmed with VITEK 2 system. Genomic fingerprinting was performed by ERIC-PCR. Logistic regression (univariate and multivariate) analyses were carried out to identify risk factors for human ESBL-PE carriage using a questionnaire survey amongst abattoir workers. ESBL-PE prevalence in animal samples from Cameroon were higher than for South Africa and ESBL-PE carriage was observed in Cameroonian workers only. Nasal ESBL-PE colonization was statistically significantly associated with hand ESBL-PE (21.95% vs. 91.67%; p = 0.000; OR = 39.11; 95% CI 2.02–755.72; p = 0.015). Low level of education, lesser monthly income, previous hospitalization, recent antibiotic use, inadequate handwashing, lack of training and contact with poultry were the risk factors identified. The study highlights the threat posed by ESBL-PE in the food chain and recommends the implementation of effective strategies for antibiotic resistance containment in both countries.


Introduction
Enterobacteriaceae are rod-shaped, Gram-negative bacteria, fermenting glucose, usually motile and facultative anaerobes, with the majority of genera being natural residents of gastrointestinal tract of animals, humans and some of them can be found in the environment [1,2].The extensive use of third and fourth generation cephalosporins in human and animal health, has led to the emergence of extended spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE).ESBL-PE represent a significant public health concern globally and have recently been classified by the World Health Organization as pathogens of critical priority in research [3].
Several studies have detected ESBL-PE in food animals, especially pigs, poultry and cattle and food products throughout the world and their transmission from livestock to humans in the farm-to-plate continuum has been evidenced [4,5].However, the prevalence of ESBL-PE in food animals and humans across the farm-to-plate continuum is yet to be elucidated in Sub-Saharan countries including Cameroon and South Africa.It is therefore imperative to understand the epidemiology and determine the burden of ESBL-PE in food animals in order to highlight the threat posed by these resistant bacteria and provide evidence for decision-makers to implement effective prevention and containment measures of antibiotic resistance (ABR) in Cameroon and South Africa.The objectives of this study were thus to assess and compare the colonization, antibiotic resistance profiles and genetic relatedness of ESBL-PE among pigs and exposed workers and delineate risk factors of ESBL-PE carriage in humans in these countries.

Demographic Characteristics
Altogether, 114 people were contacted in the five selected slaughterhouses and 83 (73%) workers agreed to participate in the study, with the response rate being higher in Cameroon (71%) than in South Africa (59%).Table 1 describes nasal and hand ESBL-PE carriage of workers in relation to individual, medical/clinical history and slaughterhouse-related characteristics.
Table 1.Nasal and hand ESBL-PE carriage of workers in relation to personal, medical/clinical and slaughterhouse-related characteristics.Of 84 workers enrolled, one withdrew prior to the sample collection and six refused the nasal sampling, yielding a total of 83 hand and 77 nasal samples.A few workers could not recall precise information whilst other refused to answer some questions leading to missing information that was not considered in the analysis.

Epidemiological Background of ESBL-PE in Pigs
Overall, the prevalence of ESBL-PE in the pooled nasal and rectal samples was 75% (108/144) and 71% (102/144), respectively (Table 4).At country-level, 42% (30/72) and 50% (36/72) ESBL-PE were detected in rectal and nasal pooled samples in South Africa respectively, whereas a 100% ESBL-PE prevalence was isolated in both specimen types in Cameroon (Table 1).In Cameroon, the main species identified were E. coli (61%) and Klebsiella pneumoniae (25%) whereas in South Africa, E. coli was the sole Enterobacteriaceae species isolated in both types of samples (Supplementary Table S2).ESBL-PE isolated from pigs exhibited high resistance to ampicillin, cefuroxime, cefuroxime-acetyl, cefotaxime, ceftazidime and trimethoprim-sulfamethoxazole in both countries (Table 2).One South African isolate expressed high resistance to colistin (8 mg/L) and no resistance to ertapenem, meropenem, imipenem and tigecycline was observed.The majority of ESBL-producing E. coli isolated from pigs in both countries showed the resistance profile AMP.TMP/SXT.CXM.CXM-A.CTX in both type of samples (Table 5).

Genotypic Relatedness
ERIC-PCR allowed the differentiation of the 93 E. coli into 14 clusters named alphabetically from A-N (Figure 1).A batch of isolates in cluster M (PR210, PR212E *, PR209E2, PR246B1C and PN254E), collected from pigs of abattoir SH004 and SH005 in South Africa was considered to be closely related.Moreover, great interest was observed in cluster I, where one pair of animal strains, PR085E3 and PR209E1 isolated in abattoirs SH002 and SH004 in Cameroon and South Africa, respectively, showed 100% similarity and were closely related with a human strain (HN503E2II) detected in abattoir SH001 in Cameroon (Figure 1).

Risk Factors of Human ESBL-PE Carriage
Table 6 shows the relationship between ESBL-PE carriage in workers and the foremost putative risk factors.Nasal and hand ESBL-PE colonization were univariately associated with an odds ratio (OR) of 39.11 (95% CI 2.02-755.72;p = 0.015).Other determinants, univariately associated with nasal and hand ESBL-PE carriage were previous hospitalization, recent antibiotic use, inadequate handwashing, occupation of relatives and year in the employment.The multivariate analysis reveals that nasal and hand ESBL-PE carriage in humans were associated with contact with other animals, particularly poultry with high statistical significance for both sample types (OR = 5.83, 95% CI 1.58-21.48,p = 0.008; vs. OR = 8.41, 95% CI 2.27-31.11,p = 0.001).

Discussion
Enterobacteriaceae and especially ESBL-PE, were acknowledged as critical priority antibiotic-resistant bacteria (ARB) by the WHO and their emergence at the animal-human-environment interface presents a to serious and multifaceted public health concern globally [3].This study investigated the carriage, risk factors, antibiotic resistance profiles and genetic relatedness of ESBL-PE isolated from apparently healthy pigs and occupationally exposed workers in Cameroon and South Africa.
The overall prevalence of human ESBL-PE carriage was 50% in hand and 45.75% in nasal samples.Comparable data was reported by Magoue et al. (2013) in Cameroon, where the prevalence of ESBL-PE faecal carriage was 45% in outpatients in the region of Adamaoua [6].Our findings are nevertheless higher than that described by Dohmen et al. (2015) where a 27% prevalence of ESBL-PE carriage in faecal samples of people with daily exposure to pigs in Netherlands was described [7].
Our results are in contrast to a study of Fisher et al., (2016), where none of the 66.7% Enterobacteriaceae detected in the nares of participants were ESBL producers and where the authors concluded that nares were a negligible reservoir for colonization of ESBL-PE in pig's exposed workers [8].Our finding shows that the prevalence of ESBL-PE carriage in nasal samples substantially increased (8.33 vs. 91.67%;p < 0.001) and was statistically significantly correlated with their carriage on hand (OR 39.11; 95% CI 2.02-755.72;p = 0.015).In addition, nasal ESBL-PE carriage was associated with inappropriate handwashing with high statistical significance (OR 4.71; 95% CI 2.28-9.70;p < 0.001).This suggests, that nares might likely become reservoir of ESBL-PE when limited hygienic conditions prevail and biosecurity measures are not adequately implemented.It further reveals that, as with the transmission of nosocomial infections in hospital settings, hands constitute important vectors of ABR transmission in the food production industry and may not only drive the transfer from person-to-person but also the contamination of food products intended for the end consumer.Nasal ESBL-PE carriage reported herein might also be ascribed to airborne contamination as recently reported by Dohmen et al., (2017) who revealed that human CTX-M-gr1 carriage was statistically associated with presence of CTX-M-gr1 in dust (OR = 3.5, 95% CI = 0.6-20.9)and that inhalation of air might constitute another transmission route of ESBL-PE in the food chain [9].
The difference in the prevalence of ESBL-PE carriage in humans in both countries could be explained by the fact that South Africa has existing abattoir regulations in place and South African abattoirs were compliant with international food safety standard ISO 22000 and Hazard Analysis Critical Control Points (HACCP) plans.In Cameroon, slaughterhouse/markets were principally low-grade, lacking in basic amenities, with sub-optimal sanitary conditions and limited or non-existent biosecurity measures.The Food and Agriculture Organization for the United Nations (FAO) report on abattoir facilities in Central African countries including Cameroon, already underlined the gaps in term of biosecurity measures in these settings [10].Our findings, therefore, reinforce the importance of and the need to implement strict biosecurity procedures as when effective prevention and containment measures are implemented, the risk of ABR dissemination is reduced.
The overall prevalence of ESBL-PE in pigs was 71% and 75% in rectal and nasal pooled samples, respectively.The results are consistent with that reported by Le et al., (2015) in food animals and products in Vietnam where a 68.4% prevalence of ESBL-producing E. coli was described [11].They are however lower than that reported in pig farms in Germany, where 88.2% of ESBL-producing E. coli was detected [12] and higher than that reported in two other studies with prevalence ranging from 8.6 to 63.4% in food animals and food products in Netherlands, [13] and 8.4% in cattle in Switzerland [2].
The high rate of ESBL-PE carriage detected in both nasal and rectal samples in Cameroon may suggest that ESBL-PE are consistently widespread in food animals in Cameroon, disseminate in the farm-to-plate continuum and represent a grave public health threat in the country.Similarly, the ESBL-PE prevalence detected in pigs in South Africa is not surprising, especially because the use of antibiotics as growth promoter agents is legally allowed in the country [14].These findings reveal gaps in the current state of knowledge about antibiotic use and ABR in food animals and suggests that the debate about ABR-related consequences in the farm-to-plate continuum is neglected in Cameroon and South Africa and should be more seriously considered in these countries.Additionally, our study revealed a high frequency (95%) of ESBL-producing E. coli, emphasizing the relevance of this indicator bacteria as a serious public health issue.
ERIC analysis demonstrated relative associations amongst human and animal isolates within and across countries.Some strains isolated in humans were highly related to those detected from pigs at similar or dissimilar abattoirs suggesting that the occurrence of ESBL-PE in humans may have an animal origin or vice-versa and that these bacteria may spread to humans via the food chain, allowing their dissemination to the global population.Although not providing evidence on the transmission dynamics of ESBL-PE, our results nevertheless show an epidemiological link amongst isolates from humans and animals.
Hospitalization, antibiotic use and contact with (food) animals are known risk factors for human ESBL-PE carriage [15].Twenty-one abattoir workers or their family members had been admitted to a hospital within the year of the sampling.Of these, 39.29% evidenced nasal ESBL-PE carriage and 71.43% hand ESBL-PE colonization (Table 1).Likewise, the majority of workers who had used antibiotics within the month of the sampling were colonized by ESBL-PE in nares (55.26%) and hands (71.05%) (Table 1).
There are certain limitations to consider in this cross-sectional study.First, the duration of ESBL-PE carriage was not investigated and there was no apparent relationship between human ESBL-PE carriage and contact with ESBL-PE colonized pigs (Table 6).Secondly, in contrast, a clear association was established between contact with other (food) animals, mainly poultry and human ESBL-PE colonization, with high statistical significance (Table 6), suggesting that further work should be undertaken in high risk populations and other food animals such as poultry in order to expand our understanding on the public health impact of the likely zoonotic transmission of ESBL-PE through the farm-to-plate continuum.Thirdly, the small human sample size precluded any direct conclusions on the prevalence of antibiotic resistance among abattoir workers.Finally, the molecular analyses were only carried out on a representative sub-sample and not all isolates due to financial constraints.Comprehensive molecular analysis would have certainly allowed better understanding of the genetic exchanges and evolution that are likely to occur within and between bacteria in this continuum.
To the best of our knowledge, this is the first report of ESBL-PE in animals and humans in both Cameroon and South Africa taking food safety perspective.The high prevalence of ESBL-PE found in pigs in both countries as well as in humans in Cameroon highlights the food safety issue associated with their presence in the farm-to-plate continuum.It demonstrates the urgent need to implement multi-sectorial, multi-faceted and sustainable collaboration and activities among all stakeholders involved in this continuum in order to reduce the prevalence and contain the dissemination of ESBL-PE and ABR in these countries.

Study Design and Study Sites
From March to October 2016, a multicentre study was conducted in five abattoirs in Cameroon (n = 3) and South Africa (n = 2).All abattoirs were coded for ethical reasons as SH001, SH002, SH003, SH004 and SH005.They were visited thrice at different time points to allow a representative sample.Apparently healthy and freshly slaughtered/stunned pigs were randomly sampled in both Cameroon and South Africa.The interior cavity of both anterior nares were swabbed and rectal swabs of pigs were obtained using sterile Amies swabs (Copan Italia Spa, Brescia, Italia).Altogether, 432 nasal and rectal pigs were collected in both countries, with the number of samples from each slaughterhouse (SH001, n = 129; SH002, n = 57; SH003, n = 30; SH004, n = 120; SH005, n = 96) proportional to the annual pig production per site.

Human Sampling Procedure
Total sampling was employed where all exposed workers (≥21 years old) willing to participate were recruited in the study upon oral and written informed consent.Participants were requested to answer a questionnaire describing socio-demographic and medical/clinical history, as well as probable risk factors associated with ESBL-PE emergence/colonization and spread.Amies swab was used to collect both anterior nares and hand (between fingers for each right and left hand) samples which were processed within 4 h of collection.

Bacteriological Analysis
For the bacteriological analysis, three individual pig samples were pooled per abattoir, gender, specimen and area of breeding leading to 288 pools (144 nasal and 144 rectal) representing 432 original specimens collected from 432 pigs.Pooled pig samples and human swabs were cultured onto an in-house selective MacConkey agar supplemented with 2 mg/L cefotaxime (MCA+CTX) and incubated for 18-24 h at 37 • C for ESBL-PE screening.Presumptive ESBL-PE were phenotypically confirmed with Vitek ® 2 System (BioMérieux, Marcy l'Etoile, France).

ESBL Detection, Species Identification and Antimicrobial Susceptibility Testing
Each colony with a unique morphotype growing on MCA+CTX was screened for ESBL production through the standard double disk synergy test (DDST) as recommended by the Clinical Laboratory and Standards Institute (CLSI) [16].

Genotypic Relatedness Determination of ESBL-Producing Escherichia coli
The Thermo Scientific ® GeneJet Genomic DNA purification kit (Thermo Fisher Scientific, Johannesburg, South Africa) was used for the genomic DNA extraction following the manufacturer's instructions.ERIC-PCR was carried out using the primers ERIC 1 5 -ATG TAA GCT CCT GGG GAT TCA C-3 and ERIC 2 5 -AAG TAA GTG ACT GGG GTG AGC G-3 [18].Reactions were performed in a 10 µL final solution containing 5 µL DreamTaq Green Polymerase Master Mix 2× (Thermo Fisher Scientific, South Africa), 2.8 µL nuclease free water, 0.1 µL of each primer (100 µM) and 2 µL DNA show their gratitude to Serge Assiene, Arthur Tchapet and Zamabhele Kubone, for their assistance with the sample collection and preliminary screening of samples in Cameroon and South Africa, respectively.We are very thankful to the abattoir owners/coordinators in South Africa, abattoirs' leaders, supervisors, food safety inspectors, veterinarians, workers and study participants, for their support and assistance during the sample collection in both Cameroon and South Africa.

Figure 1 .
Figure 1.Genotypic relationship of ESBL-E.coli strains (n = 93) detected from exposed workers and pigs in Cameroon and South Africa.Dendogram generated by Bionumerics using UPGMA method and the Dice similarity coefficient.

Figure 1 .
Figure 1.Genotypic relationship of ESBL-E.coli strains (n = 93) detected from exposed workers and pigs in Cameroon and South Africa.Dendogram generated by Bionumerics using UPGMA method and the Dice similarity coefficient.

Table 6 .
Predictors of nasal and hand ESBL-PE carriage among humans.Univariate and multivariate analysis (logistic regression).
a : Relative working with food animals, food products or at hospital, b : Within one year prior the sampling date; c: Within one month prior the sampling date.
Prior to the implementation of the study, ethical approvals were obtained from the National Ethics Committee for Research in Human Health of Cameroon (Ref.2016/01/684/CE/CNERSH/SP), Biomedical Research Ethics Committee (Ref.BE365/15) and Animal Research Ethics Committee (Ref.AREC/091/015D) of the University of KwaZulu-Natal.In addition, ministerial approvals from the Cameroonian Ministry of Scientific Research and Innovation (Ref.015/MINRESI/B00/C00/C10/C14) and Ministry of Livestock, Fisheries and Animal Industries (Ref.061/L/MINEPIA/SG/DREPIA/CE) were also granted.This study was further placed on record with the South African National Department of Agriculture, Forestry and Fisheries [Reference: 12/11/1/5 (878)].