This study did not require formal ethical approval from an Institutional Animal Ethics Committee or Institutional Review Board, as it exclusively involved the collection of houseflies (Musca domestica), which are invertebrates and not subject to animal welfare regulations under international guidelines (e.g., those from the World Organization for Animal Health or local Thai regulations). No vertebrate animals, human participants, or procedures causing pain or distress were used. The study was conducted in accordance with general principles of humane insect collection and biosafety protocols at the University of Phayao. This study on M. domestica was reviewed and approved by the Institutional Animal Care and Use Committee of the University of Phayao, Thailand (Ethical Approval No. 640104039).

This cross-sectional surveillance study was conducted from June to November 2023 in urban and peri-urban market settings in Muang District, Phayao Province, Northern Thailand.

Housefly samples were collected from local meat, vegetable, and cooked-food markets selected based on public accessibility, active food-handling activities, and their relevance as high-contact environments with potential environmental contamination and human exposure.

Sampling intensity was adjusted according to market size, with five traps deployed per sampling event in cooked-food markets (small), ten in vegetable markets (medium), and fifteen in meat markets (large). In total, 350 houseflies were collected using the sticky trap method, as previously described [11]. Sterile adhesive strips were placed within market areas, and captured houseflies were transferred individually into sterile containers using forceps. Samples were transported to the laboratory and euthanized at −20°C for 1–2 h. Morphological identification was performed using a stereomicroscope to confirm species identity as M. domestica [12]. A single housefly was then placed in a test tube containing 2 mL of alkaline peptone water (APW) medium (HiMedia Laboratories Pvt. Ltd., Mumbai, India, Cat. No. M618-500G), vortexed, and incubated at 37°C for 16–18 h for whole-fly enrichment without separation of external and internal compartments.

A sterile loopful of enriched APW culture was streaked onto eosin methylene blue agar (HiMedia Laboratories Pvt. Ltd., Cat. No.M317-500G) and incubated at 37°C for 24 h. Presumptive E. coli colonies exhibiting a metallic sheen with a dark center were selected and subjected to biochemical and morphological characterization, including Gram staining, sugar fermentation, methyl red, Voges–Proskauer, indole, and citrate tests, as previously described [13]. Confirmed colonies were cultured and stored on nutrient agar (NA) medium (HiMedia Laboratories Pvt. Ltd., Cat. No. M001-500G) and incubated at 37°C for 24 h for further analyses.

Genomic DNA was extracted using the boiling method, as previously described [14]. A single colony of E. coli grown on NA medium was suspended in 100 μL of sterile water in a 1.5 mL microcentrifuge tube. Samples were heated at 100°C for 10 min, cooled on ice for 5 min, and centrifuged at 10,000 rpm for 10 min. The supernatant was transferred to a new microcentrifuge tube and stored at −20°C until use.

All isolates were molecularly confirmed as E. coli by polymerase chain reaction (PCR) targeting the uidA (β-glucuronidase) gene, following Heijnen and Medema [15] with minor modifications. PCR reactions (25 μL) were prepared using OnePCR™ Ultra (Bio-Helix Co., Ltd., Keelung, Taiwan, Cat. No. 1BHC-MBA01-0100), uidA-F/uidA-R primers (Supplementary Table 1), DNA template, and sterile water. PCR products were analyzed by agarose gel electrophoresis and visualized under UV light using a gel documentation system (BIS 303 PC; DNR Bio-Imaging Systems Ltd., Jerusalem, Israel). A 100 bp DNA marker (OmniMARK 100 RTU., Bio-Helix Co., Ltd., Cat. No. DM101-0100) was used for size estimation. Amplicons were sequenced by Macrogen Inc. (Seoul, South Korea) and analyzed using BLAST in the NCBI GenBank database.

Phylogenetic grouping was performed by multiplex PCR targeting the chuA and yjaA genes and the DNA fragment TSPE4.C2, as described by Clermont et al. [16]. PCR conditions and downstream analyses followed the same procedures as described above.

Virulence genes associated with diarrheagenic E. coli were detected by PCR using primers and conditions described previously [7]. The target genes included aggR, stx1, stx2, astA, estp, esth, elt, bfpA, eae, and invE (Supplementary Table 1).

Antimicrobial susceptibility was evaluated using the disk diffusion method in accordance with Clinical and Laboratory Standards Institute guidelines (CLSI M100-S23, 2018). Twelve antibiotics representing eight antimicrobial classes were tested: aminoglycosides (gentamicin, streptomycin), penicillins/β-lactams (penicillin G, ampicillin), carbapenems (imipenem, meropenem), tetracyclines (tetracycline, tigecycline), phenicols (chloramphenicol), fluoroquinolones (ciprofloxacin), macrolides (erythromycin), and sulfonamides (trimethoprim–sulfamethoxazole) (HiMedia Laboratories Pvt. Ltd., Cat. No. SD170, SD031, SD028, SD002, SD073, SD727, SD037, SD278, SD006, SD060, SD013, and SD010, respectively). Mueller–Hinton agar plates (HiMedia Laboratories Pvt. Ltd., Cat. No. M173) were inoculated with a 0.5 McFarland standardized suspension and incubated at 37°C for 18–24 h. Inhibition zones were measured in millimeters. E. coli TISTR 073 served as the quality control strain. Isolates resistant to one or more antibiotics in three or more antimicrobial classes were classified as MDR [17, 18].

Phenotypically resistant E. coli isolates were screened for resistance-associated genes by PCR. Target genes included acc(3)-IV, aac(6)-Ib-cr, aadA1, CITM, bla_SHV, catA1, ereA, sul1, and tetA. Primer sequences and expected amplicon sizes are provided in Supplementary Table 1. PCR conditions followed previously described protocols [7, 10, 15, 16, 19, 20].

Descriptive data are presented as frequencies and percentages. Differences in categorical variables, including E. coli prevalence, virulence gene positivity, MDR phenotypes, and AMR gene distribution among market types, were assessed using the chi-square test. Associations between phylogenetic groups and virulence status, as well as between phylogenetic groups and MDR status, were also evaluated. Statistical analyses were performed using GraphPad Prism version 5.00 for Windows (GraphPad Software Inc., San Diego, CA, USA), and p ≤ 0.05 was considered statistically significant.

In this study, 106 (30.3%) E. coli strains were isolated from 350 housefly samples and confirmed by PCR analysis. The highest prevalence was observed in the meat market (39.5%), followed by the cooked-food market (26.3%) and the vegetable market (14.0%) (Table 1). A statistically significant difference in E. coli prevalence was detected among market types (p < 0.0001). Phylogenetic analysis revealed that phylogroup A was predominant (52 isolates; 49.1%), followed by phylogroups D (21 isolates; 19.8%), B1 (20 isolates; 18.9%), and B2 (13 isolates; 12.3%) (Table 1).

All E. coli isolates were screened for 10 virulence genes associated with diarrheagenic pathotypes. Overall, 74 isolates (69.8%) were positive for at least one virulence gene (Table 2). The most prevalent gene was bfpA (26.4%; 28/106), followed by astA (12.3%; 13/106), stx1 (7.5%; 8/106), esth (4.7%; 5/106), eae (2.8%; 3/106), and estp (0.9%; 1/106). In contrast, elt, invE, and stx2 were not detected. No significant difference was observed in the distribution of virulence gene-positive isolates among market types (p = 0.057). Isolates carrying two virulence genes accounted for 15.1% (16/106), with four observed combinations, the most common being aggR–astA (9.4%; 10/106), followed by aggR–bfpA (3.8%; 4/106), astA–esth (0.9%; 1/106), and bfpA–stx (0.9%; 1/106).

Virulence gene profiles varied across phylogenetic groups (Supplementary Table 2). Phylogroup A exhibited the greatest diversity of virulence-associated genes, with bfpA being the most prevalent profile (36.4%; 12/33), whereas rare profiles such as bfpA–stx were least frequent (3.0%; 1/33). In phylogroup B1, bfpA predominated (35.7%; 5/14), while stx1 occurred at the lowest frequency (7.1%; 1/14). Phylogroup B2 showed marked dominance of bfpA (63.6%; 7/11), representing the highest single-gene prevalence among all groups, whereas estp was detected infrequently and exclusively in this group (9.1%; 1/11). Phylogroup D displayed a more even distribution of virulence profiles, with bfpA, astA, stx1, and aggR–astA representing the most common patterns (18.8%–25.0%; 3–4/16), while eae was the least prevalent (6.3%; 1/16). No significant association was found between phylogenetic group and overall virulence gene positivity (p = 0.4314).

Phenotypic antimicrobial resistance was widely observed among E. coli isolates from all markets. All 106 isolates were tested against 12 antibiotics representing eight antimicrobial classes. Complete resistance to penicillin G (100%) and erythromycin (100%) was observed in all isolates (Table 3). High resistance rates were recorded for ampicillin (66.0%; 70/106), tetracycline (35.8%; 38/106), and chloramphenicol (17.0%; 18/106). Lower resistance rates were noted for gentamicin (11.3%; 12/106), streptomycin (11.3%; 12/106), sulfonamide (11.3%; 12/106), and ciprofloxacin (2.8%; 3/106). No resistance was detected to imipenem, meropenem, or tigecycline.

Multidrug resistance was identified in 46.2% (49/106) of isolates (Table 4). The prevalence of MDR isolates was significantly higher in the meat market than in the vegetable and cooked-food markets (p = 0.0225). A total of 23 distinct AMR patterns were observed, with ampicillin–erythromycin–penicillin–tetracycline being the most frequent MDR profile (11.3%; 12/106). Only two isolates, both from the meat market, exhibited resistance to seven tested antibiotics.

Among the 106 E. coli isolates, MDR positivity was highest in phylogroup D (57.1%; 12/21), followed by B1 (50.0%; 10/20) and A (48.1%; 25/52), and lowest in phylogroup B2 (15.4%; 2/13) (Supplementary Table 3). Most MDR-positive isolates originated from the meat market (85.7%; 42/49), followed by the cooked-food market (10.2%; 5/49) and the vegetable market (4.1%; 2/49). By phylogenetic group, MDR-positive isolates in groups A and D were almost exclusively associated with the meat market, whereas group B1 included isolates from all three market types and group B2 from meat and cooked-food markets only. No significant association was detected between phylogenetic group and overall MDR positivity (p = 0.1037).

The distribution of AMR genes among resistant E. coli isolates by market type is shown in Figure 1 and Supplementary Table 4. The most prevalent gene was bla_SHV (20.8%), followed by ereA (19.8%), tetA (17.9%), CITM (15.1%), acc(3)-IV (9.4%), catA1 (3.8%), aadA1 (3.8%), sul1 (1.9%), and aac(6)-Ib-cr (0.9%). The majority of AMR gene–harboring isolates originated from meat markets.

Overall, 67 (63.2%) of the 106 isolates carried at least one AMR gene (Table 5). The most frequently detected single gene was bla_SHV (12.3%), followed by ereA (11.3%) and CITM (8.5%). Among isolates carrying multiple resistance genes, the most common combinations were acc(3)-IV–tetA (4.7%) and acc(3)-IV–ereA–tetA (4.7%). A significant difference was observed in the distribution of AMR-positive isolates among market types (p = 0.0394).

Houseflies (M. domestica) are among the most common synanthropic insects coexisting with human communities, where they cause nuisance and contribute to public health problems. High housefly densities are often indicative of poor household and environmental hygiene [21]. Houseflies possess a strong capacity to acquire and disseminate pathogens from septic sources to other environments through contamination of their body surfaces, including feet, wings, and mouthparts, thereby acting as mechanical vectors for pathogenic agents and AMR dissemination [22].

In this study, E. coli was isolated from houseflies collected from meat, vegetable, and cooked-food markets in Muang District, Phayao Province, Northern Thailand. The overall prevalence of E. coli was 30.3%, with the highest prevalence observed in meat markets (39.5%). These findings are consistent with reports from India, where a comparable prevalence of 30% was observed [23], but higher than that reported from Bangladesh (19%) [24]. In contrast, substantially higher contamination rates ranging from 51.4% to 77% have been documented in other geographic regions [8, 10, 25]. The elevated prevalence observed in meat markets may be attributed to inadequate sanitation and frequent exposure of flies to animal feces and contaminated materials during meat handling and processing [26, 27].

Phylogenetic analysis revealed that phylogroup A predominated among the E. coli isolates (49.1%), consistent with previous studies identifying phylogroup A as the most prevalent group across diverse sources [28]. This phylogroup is commonly associated with commensal as well as pathogenic E. coli strains in humans and animals. However, the proportion observed in the present study was slightly higher than the approximately 40% prevalence reported in similar investigations conducted in other regions [29].

Virulence gene detection is essential for assessing the pathogenic potential of E. coli, particularly diarrheagenic strains [30]. In the present study, 69.8% of isolates carried at least one virulence gene, indicating substantial pathogenic potential among housefly-associated E. coli. The most frequently detected gene was bfpA (26.4%), which occurred predominantly in isolates from meat markets. This prevalence exceeds that reported from surface water in Australia (24%) [31]. The bfpA gene is a hallmark of EPEC and encodes bundle-forming pili that mediate adherence to intestinal epithelial cells, leading to diarrheal disease [32].

The stx1 gene, associated with EHEC, was detected in 7.5% of isolates, a prevalence slightly lower than that reported by Sobur et al. [33]. The absence of invE and elt, which are associated with EIEC and ETEC, respectively [34], may reflect the low circulation of these pathotypes in the studied setting. The lack of significant differences in virulence gene distribution among market types (p = 0.057) highlights the need for continued surveillance of DEC.

The detection of multiple DEC-associated virulence markers (eae, bfpA, aggR, astA, stx1, esth, and estp) across different phylogenetic groups, with bfpA predominating, suggests that these virulence traits are not restricted to specific phylogroups. This distribution aligns with previous observations of genetic heterogeneity among atypical EPEC strains and supports the concept that DEC virulence determinants can spread via horizontal gene transfer, thereby weakening strict phylogroup–virulence associations [34–36].

The detection of antibiotic-resistant E. coli in houseflies represents a significant public health concern. All isolates exhibited complete resistance to penicillin G and erythromycin, consistent with previous reports documenting high resistance rates to these antimicrobials among housefly-associated E. coli [8, 10, 37, 38]. Elevated resistance to ampicillin (66.0%), tetracycline (35.8%), and chloramphenicol (17.0%) was observed, exceeding rates reported from Bangladesh, Iran, and Nigeria [10, 22, 39]. Conversely, the absence of resistance to imipenem and meropenem corroborates findings from China, indicating sustained efficacy of carbapenems against E. coli [25].

MDR was detected in 46.2% of isolates, with a significantly higher prevalence in meat markets than in other market types, reflecting environmental contamination and antimicrobial exposure [40]. Comparable MDR rates have been reported in urban environments [38], whereas lower rates were observed in dairy farm settings [41]. In contrast, higher MDR prevalence has been documented in hospital-adjacent environments [35]. The most frequent MDR pattern, ampicillin–erythromycin–penicillin–tetracycline, underscores the adaptability of E. coli to commonly used antibiotics [41]. Although no significant association was found between phylogenetic group and MDR phenotype (p = 0.1037), MDR is largely driven by mobile genetic elements rather than phylogenetic background alone [42–44].

AMR gene carriage was detected in 63.2% of isolates, highlighting the role of houseflies as environmental reservoirs of resistance determinants. The most prevalent gene was bla_SHV (20.8%), although this rate was lower than those reported in food centers and fish markets [10, 40]. Moderate prevalence of ereA, tetA, and CITM further supports the linkage between phenotypic resistance and underlying genetic mechanisms. Variability in gene prevalence across studies likely reflects geographic and environmental differences [41, 45]. These findings emphasize the importance of sustained surveillance and targeted interventions to limit environmental dissemination of AMR genes.