This study did not involve live animals or human participants. All E. coli isolates analyzed were obtained from a preserved stock culture collection maintained by the Faculty of Public Health at Thammasat University, Thailand, which were originally collected during routine microbiological surveillance activities conducted in 2017–2018. No animals were handled, restrained, sampled, or subjected to any intervention by the research team. According to Thai national regulations, studies using archived bacterial isolates do not require approval from an Institutional Animal Care and Use Committee or a Human Research Ethics Committee. Institutional exemption was granted by the Ethics Review Committee of the Faculty of Public Health at Thammasat University, confirming that the study was exempt from ethical review because it involved only preserved microorganisms and no identifiable animal, human, or personal data (Exemption ID: 0516.71/706). All laboratory procedures and data handling followed institutional biosafety guidelines, Thailand Ministry of Public Health regulations, and international standards for research involving microbial isolates.

This study used a retrospective laboratory-based design with E. coli isolates collected along the pork production chain, including three slaughterhouses and four fresh markets in a central province of Thailand. Sample collection took place during 2017–2018. All E. coli isolates were stored at −80°C in glycerol stock cultures at the Faculty of Public Health, Thammasat University. Between 2019 and 2020, representative isolates from the stored stocks were selected for laboratory analysis to assess AMR profiles, ESBL determinants, and virulence genes.

E. coli isolates were collected from slaughterhouses supervised by the Department of Livestock Development, Ministry of Agriculture and Cooperatives, and from large fresh markets in a central province of Thailand. All samples were initially identified as E. coli using standard bacterial culture techniques and biochemical tests, then stored at –80°C in glycerol. A total of 509 E. coli isolates were recovered from the preserved cultures. Of these, 498 isolates were successfully revived and included in the study, while 11 isolates could not be cultured. The analyzed isolates included samples from slaughterhouses (117 fecal and 119 carcass isolates) and from fresh markets (116 pork and 146 cutting board isolates).

All 498 E. coli isolates were recovered from preserved stock cultures and verified for purity by streaking onto MacConkey agar (Merck, Darmstadt, Germany), followed by incubation at 37°C for 18–24 h. Typical E. coli colonies appeared as pink colonies with a surrounding precipitation zone. Presumptive isolates were biochemically confirmed using the IMViC tests (indole, methyl red, Voges–Proskauer, and citrate) according to the method described by Feng and Weagant [12].

Antimicrobial susceptibility testing was performed by disk diffusion on Mueller–Hinton agar (MHA; Oxoid Ltd., Basingstoke, United Kingdom) according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [13]. A total of 18 antimicrobial agents were tested, including aminoglycosides (gentamicin, kanamycin, streptomycin), a beta-lactam combination (amoxicillin–clavulanic acid), cephems (cefoxitin, cefotaxime, ceftazidime, cefepime), a folate pathway antagonist (sulfamethoxazole–trimethoprim), a polymyxin (colistin), monobactams (imipenem, meropenem), a nitrofuran (nitrofurantoin), a penicillin (ampicillin), a phenicol (chloramphenicol), quinolones (ciprofloxacin, nalidixic acid), and a tetracycline (tetracycline).

E. coli American Type Culture Collection (ATCC) 25922 was used as a quality control strain. Each isolate was cultured on MHA (4 mm depth) and incubated at 37°C for 24 h. Bacterial suspensions were prepared in sterile normal saline (0.85% NaCl) and adjusted to a 0.5 McFarland standard (approximately 10⁸ CFU/mL). The inoculum was evenly spread across MHA plates using sterile cotton swabs, antimicrobial disks were applied, and the plates were incubated at 35°C ± 2°C for 16–18 h. Inhibition zone diameters were measured with a Vernier caliper and interpreted according to CLSI (2020) criteria [13]. MDR was defined as resistance to three or more antimicrobial classes [14].

ESBL-producing E. coli isolates were first identified by disk diffusion screening, in which inhibition zone diameters of ≤22 mm for ceftazidime and ≤27 mm for cefotaxime indicated potential ESBL production according to CLSI (2020) guidelines [13]. Suspected isolates were confirmed using the combination disk method with cefotaxime (30 μg) and ceftazidime (30 μg), tested alone and with clavulanic acid (10 μg) (BD, Franklin Lakes, NJ, USA). An increase of ≥5 mm in the inhibition zone diameter for the antimicrobial agent with clavulanic acid compared to the agent alone confirmed ESBL production. E. coli ATCC 25922 was used as the quality control strain.

All E. coli isolates were cultured in Luria–Bertani broth (Difco, Detroit, MI, USA) at 37°C for 15–18 h. Genomic DNA was extracted using the QIAamp® DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. DNA concentration and purity were measured with a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and the extracted DNA was stored at −20°C until further use.

All ESBL-producing E. coli isolates were screened for the presence of β-lactamase genes bla_TEM, bla_CTX-M, and bla_SHV using multiplex polymerase chain reaction (PCR) assays. Primer sequences are provided in Table 1 [15–17]. Each PCR reaction (20 μL) contained Green GoTaq reaction buffer, deoxynucleoside triphosphates, forward and reverse primers, GoTaq DNA polymerase (Promega, Fitchburg, WI, USA), and template DNA. Thermocycling conditions included an initial denaturation at 94°C for 5 min, followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min 30 s, with a final extension at 72°C for 10 min. PCR products were separated on 2.0% agarose gels stained with a red stain (Thermo Fisher Scientific, Waltham, MA, USA) and visualized under blue LED illumination using the PrepOne system (Embi Tec, San Diego, CA, USA). Reference strains carrying bla_TEM, bla_CTX-M, and bla_SHV served as positive controls, while E. coli JM109 was used as the negative control.

All 498 E. coli isolates were tested for 10 virulence genes using multiplex PCR. Primers were arranged into two sets. Primer set A targeted ETEC (lt, sth, and stp), enterohemorrhagic E. coli (EHEC) (stx1A and stx2A), and enteroaggregative E. coli (EAEC) (aggR and pCVD432) [18]. Primer set B targeted EPEC (bfpA and eaeA) and enteroinvasive E. coli (EIEC) (ipaH) [19]. Each PCR reaction (25 μL) included Green GoTaq reaction buffer, deoxynucleoside triphosphates, primers, GoTaq DNA polymerase, and template DNA. Amplification conditions featured an initial denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 95°C for 1 min, annealing at 52°C for 1 min, and extension at 72°C for 1 min, ending with a final extension at 72°C for 5 min. Amplicons were analyzed as described above. Reference strains for each virulence gene served as positive controls, and E. coli JM109 was used as a negative control.

Descriptive statistics were used to summarize the frequency and percentage of AMR, ESBL production, and virulence genes among E. coli isolates from slaughterhouses and fresh markets. Pearson’s chi-square test and Fisher’s exact test (for expected cell counts <5) were applied to compare detection rates between sample sources, and 95% confidence intervals (CIs) were calculated. Statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS version 18.0; SPSS Inc., Chicago, IL, USA). A p-value <0.05 was considered statistically significant.

The antimicrobial susceptibility of all E. coli isolates was assessed across 11 antimicrobial classes using the disk diffusion method with 18 antimicrobial agents. Overall, 97.4% (485/498) of the isolates showed resistance to at least one antimicrobial agent (Table 2). Resistance was detected in 98.3% (232/236) of isolates from slaughterhouses and 96.6% (253/262) from fresh markets. No statistically significant difference in resistance rates was observed between slaughterhouse and fresh-market isolates (p > 0.05; 95% CI: 0.6–6.8).

Over 50% of E. coli isolates from slaughterhouse samples (feces and carcasses) showed resistance to ampicillin, tetracycline, chloramphenicol, streptomycin, and sulfamethoxazole–trimethoprim. Likewise, isolates from pork and cutting boards in fresh markets exhibited resistance rates over 50% for ampicillin, tetracycline, streptomycin, and chloramphenicol. Conversely, most isolates from both slaughterhouses and fresh markets remained susceptible to imipenem, meropenem, cefoxitin, and colistin.

All 498 E. coli isolates were evaluated for MDR based on resistance to various antimicrobial classes. A high rate of MDR was observed, with 87.3% (435/498) of isolates identified as MDR (Table 2). MDR rates were high in both slaughterhouse (90.3%; 213/236) and fresh-market (84.7%; 222/262) isolates. There was no statistically significant difference in MDR prevalence between the two sources (p > 0.05; 95% CI: 0.9–2.8).

A total of 130 unique AMR patterns were identified among all isolates (Supplementary Table 1). The most common patterns were AMI–PHE–PEN–FOL–TET (50 isolates), AMI–PEN–TET (24 isolates), and AMI–PHE–PEN–QUI–FOL–TET (24 isolates).

ESBL-producing E. coli accounted for 23.5% (117/498) of all isolates. Of these, 14.3% (71/498) came from slaughterhouses, while 9.2% (46/498) were recovered from fresh markets. The prevalence of ESBL-producing isolates was significantly higher in slaughterhouses than in fresh markets (p < 0.05; 95% CI: 1.3–3.1).

Among slaughterhouse samples, ESBL-producing E. coli were found in 34.2% (40/117) of fecal isolates and 26.1% (31/119) of carcass isolates. In fresh markets, ESBL-producing isolates appeared in 18.9% (23/116) of pork samples and 15.8% (23/146) of cutting board samples (Table 3).

Among the 117 ESBL-producing E. coli isolates, 97.4% (114/117) were classified as MDR. The prevalence of MDR was similarly high in isolates from slaughterhouses (97.2%; 69/71) and fresh markets (97.8%; 45/46). Among slaughterhouse-derived ESBL producers, MDR was found in 34.2% (40/117) of fecal isolates and 24.8% (29/117) of carcass isolates. In fresh markets, MDR ESBL-producing isolates were present in 19.7% (23/117) of pork samples and 18.8% (22/117) of cutting board samples. Three ESBL-producing isolates were not classified as MDR.

A total of 44 unique AMR patterns were identified among ESBL-producing isolates. The most common patterns were AMI–PHE–CEP–PEN–FOL–TET (16 isolates) and AMI–PHE–CEP–PEN–QUI–FOL–TET (16 isolates) (Figure 1).

Among the 117 ESBL-producing E. coli isolates, 89.7% (105/117) carried at least one of the tested β-lactamase genes (bla_TEM, bla_CTX-M, or bla_SHV). The bla_TEM gene was the most commonly detected, found in 66.7% (78/117) of isolates, including 69.0% (49/71) from slaughterhouses and 63.0% (29/46) from fresh markets. The bla_CTX-M gene was present in 59.0% (69/117) of isolates, with a significantly higher prevalence in slaughterhouse isolates (76.0%; 54/71) compared to those from fresh markets (32.6%; 15/46).

The bla_SHV gene was detected at a low frequency (4.3%; 2/46) and was found only in isolates from fresh markets. Co-occurrence of bla_TEM and bla_CTX-M was observed in 35.9% (42/117) of isolates, including 47.9% (34/71) from slaughterhouses and 17.4% (8/46) from fresh markets. The prevalence of bla_TEM, bla_CTX-M, and their combination was significantly higher in slaughterhouse isolates than in fresh-market isolates (p < 0.05) (Table 3).

Virulence genes were found in 3.2% (16/498) of E. coli isolates, indicating the presence of pathogenic strains. Three virulence genes, eaeA, lt, and stp, were identified (Table 4). The eaeA gene, which encodes intimin and is linked to EPEC and EHEC, was detected in 4.7% (11/236) of isolates from slaughterhouses and 1.1% (3/262) from fresh markets. The lt and stp genes were each found at a low frequency of 0.8% (2/262) in fresh-market samples, with lt identified in one cutting board isolate and stp in one pork isolate.

No statistically significant difference was observed in the prevalence of pathogenic E. coli between slaughterhouse and fresh-market isolates (p > 0.05; 95% CI: 0.8–7.3). Of the 16 virulence gene–positive isolates, 62.5% (10/16) were MDR and carried β-lactamase genes. The virulence genes bfpA, ipaH, sth, stx1A, stx2A, aggR, and pCVD432 were not detected in any isolate.

Antimicrobial agents are commonly used in the pork production chain, especially at the farm-level for therapeutic, growth-promoting, and disease-preventing purposes [20]. However, such practices promote the emergence and persistence of AMR at various stages of pork production. In this study, more than 50% of E. coli isolates from slaughterhouse samples (feces and carcasses) and fresh-market samples (pork and cutting boards) showed resistance to ampicillin, tetracycline, chloramphenicol, and streptomycin. These antimicrobials are among the ones most frequently linked with resistance in E. coli isolates from pigs [21].

The observed resistance patterns align with earlier reports from pig farms in Thailand [4, 22] and slaughterhouse-based studies in the Mekong Delta, Vietnam [23], Brazil [24], and Australia [25]. Notably, studies from Brazil and Australia reported high resistance rates to ampicillin (81.1% and 60.2%, respectively) and tetracycline (97.8% and 68.2%, respectively), along with significant resistance to chloramphenicol [24, 25]. High volumes of tetracyclines, penicillins, and sulfonamides sold for veterinary use have been documented [26], and the improper or excessive use of these antimicrobials in pig production likely drives the elevated resistance rates observed [27]. The detection of chloramphenicol-resistant isolates is especially concerning, as this antimicrobial has been banned for use in food-producing animals [28]. Such resistance may persist due to co-selection or cross-resistance caused by ongoing use of other antimicrobials [29], emphasizing the complex and multifactorial nature of AMR development [28, 30].

Colistin resistance was found in 4.6% of isolates from both slaughterhouse and fresh-market samples. Since colistin is considered a last-resort antibiotic for treating MDR gram-negative infections, this finding raises serious public health concerns. The plasmid-mediated mcr-1 gene is key in spreading colistin resistance among animal, environmental, and human reservoirs, enabling rapid horizontal transfer [31].

In this study, 23.5% of E. coli isolates were identified as ESBL producers, a prevalence similar to that previously reported in healthy pigs in Thailand (19.2%) [4]. The detection rate of ESBL-producing E. coli was significantly higher in slaughterhouses (14.3%) than in fresh markets (9.2%) (p < 0.05), indicating greater contamination pressure at earlier stages of the pork production process rather than later stages.

Nevertheless, ESBL prevalence varied greatly when compared to other studies conducted in Thailand and other countries. For example, Boonyasiri et al. [32] reported ESBL prevalence of 33.3% in slaughterhouses and 61.5% in markets in eastern and northern Thailand, while Sornsenee et al. [33] documented ESBL-producing E. coli in minced chicken (79.17%), pork (43.75%), and beef (22.73%) in southern Thailand. Studies from other nations have also shown significant ESBL contamination in food animals and retail meat, including Singapore [34], Vietnam [35], South Korea [36], and Cameroon [37]. Such differences likely result from variations in sample types, sampling strategies, timing, farming practices, slaughterhouse hygiene, and antimicrobial use patterns.

The detection of ESBL-producing E. coli in both slaughterhouse and fresh-market samples indicates potential risks of transmission to consumers and environmental spread. Maintaining strict hygiene practices during and after slaughter is therefore essential to reduce fecal contamination and limit ESBL dissemination [33]. The presence of ESBL-producing E. coli on pork and cutting boards further suggests cross-contamination, likely from fecal sources, although molecular typing is needed to confirm transmission pathways.

An extremely high proportion (97.4%) of ESBL-producing E. coli isolates showed MDR, highlighting the serious clinical and public health concerns of ESBL phenotypes in pork production systems. Similar MDR rates have been reported in Thailand [33, 38], Vietnam [35], and Cameroon [37]. The high MDR frequency may result from extensive antimicrobial use in industrial pig farming, especially β-lactam antimicrobials used for disease prevention and growth promotion [39]. Such practices exert strong selective pressure, fostering the emergence and persistence of MDR organisms capable of spreading through food production systems to humans and the environment [40].

β-lactamase genes were found at significantly higher rates in ESBL-producing E. coli isolates from slaughterhouses compared to those from fresh markets (p < 0.05), indicating that slaughterhouses are a key control point for ESBL spread. Overall, 89.7% of ESBL-producing isolates carried at least one bla gene, with bla_TEM being the most common (66.7%). The TEM enzyme is the most widespread β-lactamase among gram-negative bacteria, providing resistance to penicillins and early-generation cephalosporins commonly used in pig farming [41, 42]. Similar high prevalence of bla_TEM has been reported in Thailand among healthy pigs and minced meat [33, 43].

The bla_CTX-M gene was found in 59% of ESBL-producing isolates, aligning with previous reports from Thailand [44] and other countries, including the United Kingdom, Germany, Tunisia, and Switzerland [45]. Notably, 35.9% of isolates carried more than one bla gene, providing evidence that ESBL-associated plasmids often contain multiple resistance determinants [46, 47]. These genetic configurations likely help maintain and spread β-lactam resistance throughout the pork production chain.

Virulence genes were detected in only 3.2% of E. coli isolates, indicating a low prevalence of pathogenic E. coli within the study group. The identified virulence genes included eaeA, lt, and stp. The eaeA gene, which encodes intimin and is linked to EPEC and EHEC, was detected in both slaughterhouse and fresh-market samples and is associated with diarrheal illness in humans. The lt and stp genes, responsible for producing heat-labile and heat-stable enterotoxins in ETEC, respectively, were detected at very low levels [8, 48].

The low detection rate of virulence genes may be due to the age of the slaughtered pigs or to the lack of active infection at the time of sampling. Notably, the contrasting pattern seen in this study, high prevalence of AMR and ESBL production but low virulence gene detection, indicates that the selective pressures promoting AMR are different from those affecting the distribution of genes associated with pathogenicity.

This study has several limitations. First, using archived isolates collected during 2017–2018 limited the ability to directly trace or confirm cross-contamination pathways between slaughterhouses and individual fresh markets. Second, only a limited set of virulence genes was examined, which may underestimate the overall pathogenic potential of circulating E. coli strains. Third, molecular typing methods, such as whole-genome sequencing, were not performed due to financial constraints, thereby restricting the assessment of genetic relatedness and transmission dynamics along the pork production chain. Finally, environmental samples were not included, preventing the identification of potential AMR reservoirs within slaughterhouses and fresh markets. Despite these limitations, the study provides valuable baseline data to inform future research and guide targeted interventions to reduce AMR dissemination in pork production systems.