Investigation of Escherichia coli isolates from pigs and humans for colistin resistance in Lao PDR- a cross-sectional study

Background In Laos, colistin is not currently registered for use in humans. This One Health study aimed to estimate the prevalence of meat-producing pigs carrying colistin-resistant Escherichia coli, and investigate if E. coli causing invasive human infections were colistin-resistant. Methods Between September 2022 and March 2023, rectal swabs were collected from 895 pigs from abattoirs in 9/17 Lao provinces. Pig rectal swabs and stored E. coli isolates from human blood cultures, submitted to Mahosot Hospital Microbiology laboratory between 2005 and 2022, were screened for colistin resistance on selective chromogenic agar with organism identification confirmed using MALDI-TOF MS. Suspected colistin-resistant isolates underwent colistin susceptibility testing by broth microdilution following European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines. Isolates with MIC values of ≥2 μg/ml were tested for plasmid-mediated colistin resistance genes (mcr-1, mcr-2, and mcr-3) by multiplex SYBR Green PCR. Results A total of 15/620 (2.41%) invasive human E. coli isolates were phenotypically colistin-resistant by broth microdilution (MIC values 4 to 8 μg/ml). The earliest isolate was from 2015 in a patient from Phongsaly province in Northern Laos. A total of 582/895 (65.02%) pig rectal swab samples contained colistin-resistant E. coli. The detected colistin resistance genes were predominantly mcr-1 (57.8%, 346/598), followed by mcr-3 (20.23%,121/598), and 22.24% (133/598) were found to co-harbour mcr-1 and mcr-3. Among the 15 human isolates with colistin MIC values of ≥4 μg/ml, 12/15 were mcr-1. Conclusions We found that colistin resistant E. coli is causing invasive infection in humans in Laos despite the fact it is not available for human use. Use in animals seems to be widespread, confirmed by high carriage rates of colistin-resistant E. coli in pigs. It is probable that food-producing animals are the source of colistin-resistant E. coli bloodstream infection in Laos, although these have been infrequent to date. This is a serious public health concern in the region that needs to be addressed by appropriate enforceable legislation.


Introduction
The World Health Organization (WHO) regards antimicrobial resistance (AMR) as one of the top ten global public health threats facing humanity [1].Colistin (Polymyxin E) is one of the WHO Reserve (AWaRe) group of antibiotics used as a last resort option for the treatment of carbapenem-resistant Gram negative infections, especially in low-income settings where newer antibiotics for multidrug-resistant infections are either not registered or not affordable [2][3][4].
Colistin is used in food-producing animals, including pigs, as a treatment and as a growth promoter [5].In 2015, the first plasmid-borne mobile colistin resistance gene (mcr-1) was detected in Escherichia coli (E.coli) from 78/523 (14.91%) samples of raw pork, 166/804 (20.64%) pigs, and in 16/1322 (1.21%) patients with infection in China [6].Since then, an additional nine mcr genes have been described (mcr-2 to mcr-10) [7][8][9].Currently, colistin resistant E. coli has been reported from 54 countries on five continents worldwide.Asia has the highest reported burden, with most reports from China [10].Asia and Europe have confirmed a wide diversity of mcr variants from various sample types.[10].The most common plasmid-mediated colistin resistance genes in Enterobacteriaceae such as E. coli, Salmonella enterica serovar Typhi, and Klebsiella pneumoniae are mcr-1, mcr-2, and mcr-3, with high mcr-1 prevalence in E. coli in pigs.There have been no reports of mcr-2 in E. coli in pigs in the Asia region [11].
The World Organization for Animal Health (WOAH) and WHO have stated that colistin cannot be used to prevent infections in animals, or as a first-or second-line treatment, unless no alternatives are available [12].WOAH called for an urgent ban on the use of colistin, fluoroquinolones, and third and fourth-generation cephalosporins as growth-promoters in animals since these are essential drugs for the treatment of serious infections in humans [13,14].
Data on colistin resistance in bacterial isolates from human infections is sparse in many countries since standard disk-diffusion, or gradient minimum inhibitory concentration (MIC) methods are unreliable for antimicrobial susceptibility testing, related to poor solubility of colistin in solid agar [15].Broth microdilution is preferred.
The Emergency Center for Transboundary Animal Diseases (ECTAD) of the Food and Agriculture Organization of the United Nations (FAO) has been working with the Department of Livestock and Fisheries (DLF) of the Ministry of Agriculture and Forestry (MAF) in Laos to develop a multisectoral strategy to tackle AMR which includes monitoring AMU in animals.To date, there have only been a few studies on colistin resistance in Laos in humans and animals and the full extent of the geographical distribution in various animal species is not known.
The aim of this study was to estimate the prevalence of colistin-resistant E. coli colonization in pigs being slaughtered for meat in multiple provinces in Laos, as well as the proportion of invasive human E. coli isolates in Laos that are colistin-resistant, and the earliest year of detection.Secondary objectives were to characterize the antimicrobial susceptibility of isolates from human invasive infections, and describe the associated colistin resistance genes (mcr-1, mcr-2, and mcr-3).

Sample collection
The study was conducted between September 2022, and March 2023.All available stored human E. coli isolates from all blood cultures from patients presenting with febrile illness submitted to Mahosot Hospital Microbiology laboratory in Vientiane (2005-2022) were included in the study, and their antimicrobial susceptibility data were extracted from the laboratory information management system.We performed a cross-sectional survey to collect rectal swabs from pigs before slaughter in 15 urban and rural abattoirs in nine provinces in Laos.These provinces were selected to represent the geographical range of Laos, and the abattoirs were selected after a pre-survey to determine which abattoirs slaughtered the highest number of pigs.One rectal swab was collected per pig, using a sterile Amies transport swab with charcoal (TRANSWAB®), by trained provincial veterinarians under the supervision of the National Animal Health Laboratory.Samples were stored at 4 • C and transferred to the Microbiology Laboratory, Mahosot Hospital for further processing, with a maximum transport time of 72 h.Data on pig demography (i.e., sex, age, breed), sampling site, sampling date, type of farm, and origin of pig were collected (Supplementary Material).Our target sample size for the pig survey was 100 per site which would enable us to detect a prevalence of 10% colistin resistance with 6% precision and 95% confidence.

Isolation and identification of colistin-resistant E. coli
Frozen stored human E. coli isolates from blood cultures were thawed and sub-cultured onto nutrient agar to ensure isolates were still viable before inoculating on Chromatic Colistin agar (Liofilchem®) to screen for resistance.Pig rectal swabs were placed in a 9 ml tube of Brain Heart Infusion broth with a colistin sulfate 10 μg disc (colistin sulfate CS 10 μg, Liofilchem®), and incubated at 35 ± 2 • C for five hours.After incubation, 25 μl were placed on the Chromatic Colistin agar, and streaked out, then incubated aerobically at 35 ± 2 • C for approximately 18-24 h.E. coli ATCC 25922 strain was used for quality control.
Colistin-resistant E. coli was identified by morphology and appearance with pink-reddish-mauve colonies on the selective Chromatic Colistin agar.Isolates were confirmed as E. coli by MALDI-TOF MS (VITEK MS, bioMérieux).

Colistin susceptibility testing
Confirmed E. coli isolates that grew on the Chromatic Colistin agar were tested for colistin susceptibility using the broth microdilution (BMD) method (ComASP™ Colistin, Liofilchem®) according to the manufacturer's instructions.A suspension of 0.5 MacFarland was added to Muller Hinton II broth and 100 μl was dispensed into the control well, and following wells (0.25-16 μg/ml).After incubating at 35 ± 2 • C for 16-20 h in ambient air, the lowest concentration of colistin that inhibited visible growth was read by the naked eye.The MICs were interpreted according to EUCAST guidelines version 12.0 (2022).The clinical breakpoint for defining colistin resistance is >2 μg/ml, but all isolates with an MIC of 2 μg/ml or greater were submitted for further testing.

Detection of mcr genes by the multiplex PCR assay
DNA was extracted from all E. coli isolates with a colistin MIC ≥2 μg/ ml using the GeneJET Genomic DNA Purification Kit (Thermo Fisher Scientific Inc.), following the manufacturer's recommendations.PCR analysis for the presence of mcr-1, − 2 and − 3 genes was carried out using a multiplex real-time PCR, based on a previously published assay [22].Briefly, a 25 μl reaction mixture consisted of 1.25 U High Sensitivity Taq DNA Polymerase (PCRBIO), 0.3 μM each primer, and 1× SYBR Green.Reactions were run on Gentier 96E with the following thermal conditions: 95 • C for 2 min, followed by 40 cycles of 98 • C for 15 s, 60 • C for 15 s, 72 • C for 15 s.

Data analysis
Patient, and pig demographic details including age, sex, and province were included in the analysis.Farms were separated into large commercial, and small-scale traditional farms.Other antimicrobial susceptibility testing (AST) results were extracted from the laboratory database (disc diffusion method following contemporaneous Clinical and Laboratory Standard Institute (CLSI) until 2018, or EUCAST from 2018 to 2022 guidelines).All AST results were re-interpreted using EUCAST breakpoint tables version 12 (2022).Extended-spectrum β-lactamase (ESBL) confirmation test was done on isolates resistant to third-generation cephalosporins or cefpodoxime using the double-disc diffusion test (cefotaxime and ceftazidime with and without clavulanic acid).Statistical analyses were conducted using R Version 4.2.3 (2023/ 03/15).Data are reported in accordance with the STROBE checklist for cross-sectional studies (Supplementary data).

Results
From the Mahosot Microbiology laboratory records of 1000 stored E. coli isolates from blood cultures, we were able to find 620 isolates collected from 2005 to 2022.Patients came from 16 provinces in Laos with the majority from Vientiane Capital (457/620, 73.70%)where Mahosot Hospital is located.There were 403/620 (65%) female patients, and ages ranged from 1 day to 98 years old.A total of 895 rectal swabs from pigs were collected from 15 abattoirs in nine provinces.Data indicated that the province of origin of some pigs was different from the abattoir sites in provinces due to pigs being transported to abattoirs in provinces where the meat was to be sold so it remained fresh.There were seven provinces that transported pigs to slaughter in different provinces.

Discussion
This One Health study described the prevalence of colistin-resistant E. coli in human invasive isolates, and pig carriage isolates in Laos.Antimicrobial resistance is increasing in the country with >50% of E. coli isolates from bloodstream infections diagnosed in a central hospital confirmed as ESBL-producers in 2023 [18].This is fuelling the use of meropenem and carbapenem resistance has emerged in recent years [23] which will necessitate using colistin for treatment in the near future.Although colistin is not currently available in Laos in the human health sector, 2.42% (15/620) of invasive human E. coli isolates were found to be colistin-resistant in this study.Two earlier studies in Laos found a prevalence of 15-45% of colistin-resistant E. coli from human rectal swabs but the prevalence in invasive infections has not been described previously [20,21].The first colistin-resistant isolate from this study was from 2015 from a patient from Northern Laos, which borders China.Since 2015 the number of colistin-resistant E. coli isolates has increased but remains relatively low in our centre.Most patients (11/15, 73.33%) came from Vientiane Capital, which is the main catchment area for Mahosot Hospital, with no positive results from 12 patients from the southern provinces of Laos, despite very high carriage rates in pigs.
Colistin-resistant E. coli was found in 65.02% (582/895) of pig rectal samples.The highest percentage (100%, 100/100) of colistin resistance in E. coli from pigs was seen in samples from Savannakhet in the south of Laos as shown in Fig. 1.
The mcr-1 gene was the most common gene detected in this study (57.86%, 346/598), which has been reported from most studies around the world [10].Isolates co-harboring mcr-1 and mcr-3 genes were seen in 22.24% (133/598) of isolates which has also been found in several studies [17].There were no mcr-2 genes detected which is similar to other studies from the region.The majority of reports of mcr-2 come from Europe [24], Egypt [25], and China [26].Interestingly, one of the two isolates with a MIC of 2 μg/ml from this study was positive for mcr-1.
In a previous study in Vietnam, E. coli isolates from humans, animals and the environment were screened for mcr, irrespective of the colistin MIC results.They found 60.63% (97/160) of mcr-1 carrying E. coli isolates were phenotypically susceptible to colistin with a MIC of ≤2 μg/ml [27].How the presence of mcr with MIC <2 μg/ml correlates with clinical response to treatment is unclear but suggests clinical breakpoints may need to be reviewed.
We do not know whether the higher percentage of colistin-resistant E. coli from pig rectal swabs from small and large farms in this study (65.02%, 582/895) compared to neighboring countries with 41.64% (112/269) colistin-resistant E. coli reported from pigs in small-scale farms in Thailand [28], and 45.69% (53/116) small and large pig farms in Vietnam is as a result of different antimicrobial usage in the different countries or different sampling strategies [29].
Ten human colistin-resistant isolates were ESBL-positive, and six were also resistant to ciprofloxacin and gentamicin, although all retained susceptibility to meropenem.This multidrug class resistant phenotype is typical of most ESBL-producing E. coli bloodstream infections in Laos, necessitating that almost all affected patients are treated with meropenem.Carbapenem resistance in Enterobacterales is starting to emerge and colistin is the last line antibiotic needed to treat infections caused by these organisms.Newer antibiotics are largely unaffordable and inaccessible in most low resource settings currently.
There is increasing demand for pork and poultry meat in Lao PDR (Laos), with pork supply of around 15.23 kg per capita per year since 2015, according to FAOSTAT data.A high-level cross-sectoral collaboration involving the WHO, the Food and Agriculture Organization, the Ministry of Health, the Ministry of Agriculture and Forestry, and other key stakeholders in Laos has been developing the national strategic plan to tackle the issue of AMR with a key priority action to improve the awareness and understanding of AMR in the country [30].More data on colistin consumption and usage in the animal sector would be useful.There also needs to be better engagement with farmers and the public to increase understanding of AMR, and the role of antimicrobial use as a major driver of AMR.
Limitations of this study include our inability to prove animal-tohuman transmission of colistin-resistant E. coli to confirm the likely role of the food chain, and in particular pigs or pig meat, in causing infection.Whole-genome sequencing may have provided stronger evidence of pig to human transmission but we did not have the resources to perform this.The number of samples collected was either capped at the target number, or was all pigs that were slaughtered which did not always meet our target.Therefore, our sample may not be representative of pigs across the whole country.Although there were human isolates from 16 provinces, the majority came from Vientiane Capital.As mcr genes 1-3 are the most commonly detected in the region, we only tested for those, but it is possible that other genes may have been in the samples.Finally, we were not able to determine what food the pigs were given and if it included colistin.

Conclusions
In this study we described colistin resistance in E. coli in pigs and humans in Laos.We confirmed that colistin-resistant E. coli is a cause of invasive infections in humans in Lao PDR (2.42%, 15/620), and carriage is widespread in more than half (65.02%) of pigs from both small and large farms.The most likely driver is the use of colistin as a growthpromoter in the animal sector.Further research could help confirm the main reservoirs of resistance in humans, and routes of transmission.The findings of this study can be used to advocate for a robust antimicrobial stewardship program to minimize unnecessary use of colistin within animal sector, and to reserve colistin as a treatment for critically ill patients with carbapenem-resistant infections.

Fig. 1 .
Fig. 1.Map showing percentage of colistin-resistant (MIC ≥ 2 μg/ml) E. coli isolates from pig rectal swab samples by province of origin, and location of human invasive colistin-resistant E. coli bloodstream infections with detected mcr gene in Laos.

Table 1
Number of pig rectal swab samples collected, and percentage with colistinresistant E. coli as shown by growth on colistin chromogenic agar and MIC >2 μg/mL, by site and farm type.