Antibiotic resistance through the lens of One Health: A study from an urban and a rural area in Sri Lanka

This study aimed to investigate and compare the proportion of AMR Escherichia coli (E. coli) between urban (Dompe in the Western province) and rural (Dambana in the Sabaragamuwa province) areas in Sri Lanka. The overall hypothesis of the study is that there is a difference in the proportion of AMR E. coli between the urban and the rural areas. Faecal samples were collected from healthy humans (n = 109), dairy animals (n = 103), poultry (n = 35), wild mammals (n = 81), wild birds (n = 76), soil (n = 80) and water (n = 80) from both areas. A total of 908 E. coli isolates were tested for susceptibility to 12 antimicrobials. Overall, E. coli isolated from urban area was significantly more likely to be resistant than those isolated from rural area. The human domain of the area had a significantly higher prevalence of AMR E. coli, but it was not significantly different in urban (98%) and rural (97%) areas. AMR E. coli isolated from dairy animals, wild animals and water was significantly higher in the urban area compared with the rural area. There was no significant difference in the proportion of multidrug resistance (MDR) E. coli isolated from humans, wild animals and water between the two study sites. Resistant isolates found from water and wild animals suggest contamination of the environment. A multi‐sectorial One Health approach is urgently needed to control the spread of AMR and prevent the occurrences of AMR in Sri Lanka.


| INTRODUC TI ON
Antimicrobial resistance (AMR) is rising worldwide, with increasing morbidity and mortality in humans and animals (Meier et al., 2022;Murray et al., 2022).
AMR burden in Sri Lanka is at an alarming level (Dhanapala et al., 2021;Jayatilleke, 2014;Jayaweera & Kumbukgolla, 2017;Nakkawita et al., 2021;Perera et al., 2022aPerera et al., , 2022b)).An example to illustrate the burden in human domains is that 59.3% of E. coli isolated from the blood cultures of hospitalized patients in tertiary care hospitals in Sri Lanka showed resistance to cefotaxime while 7.5% of isolates were resistant to meropenem (Jayatilleke, 2014).Not only in the human domain but the AMR burden in the animal domain is also at a critical level in Sri Lanka (Jayaweera & Kumbukgolla, 2017;Liyanage & Pathmalal, 2017).A recent study investigated the proportions of Aeromonas species in samples collected from the aquatic fish farming environment and ornamental fish showed high frequencies of AMR among these species, specifically for amoxicillin (92.5%), enrofloxacin (67.1%) and nalidixic acid (63.4%) (Dhanapala et al., 2021).
Another study revealed the prevalence of AMR among wild animals in Sri Lanka and it was reported that 35% of isolated E. coli were resistant to ampicillin and 25% to streptomycin (Bamunusinghage et al., 2019).According to Kumar et al., the Kelaniya River and Gin River in Sri Lanka exhibited a significant prevalence of AMR E. coli.
and their proportion changed based on seasonality.These studies highlighted the current AMR situation within the environmental domain (Kumar, Chaminda, & Honda, 2020;Kumar, Chaminda, Patel, et al., 2020).
There are multiple pathways for the transmission of AMR bacteria among the One Health domains.Humans can be exposed to animal-originated AMR bacteria through direct contact, animal-based food products, drinking water and activities such as swimming, bathing and fishing in-contaminated water (Jayaweera et al., 2020;Kottawatta et al., 2017;Kulasooriya et al., 2019).Similarly, both farm animals and wild animals are exposed to AMR bacteria present in humans through environmental contamination (Iskandar et al., 2020;Rousham et al., 2018).While there have been independent studies assessing the prevalence of AMR within distinct domains (human, animal and environmental) in Sri Lanka, a crucial research gap exists in investigating the occurrence of AMR bacteria across these domains within a specific geographical area.
Understanding the difference between urban and rural areas is crucial in the context of AMR for several reasons.Firstly, the prevalence and patterns of AMR can vary significantly between urban and rural areas due to differences in population density, healthcare access and antibiotic usage.Some reasons for this is urban areas typically have higher population densities and more extensive healthcare facilities, which can lead to increased transmission of resistant bacteria and greater antibiotic consumption (Neiderud, 2015).In contrast, rural areas may face challenges in healthcare access, limited diagnostic capabilities and inadequate surveillance systems, potentially resulting in delayed AMR detection and management (Gandra et al., 2020).Secondly, the socio-economic factors associated with urban and rural settings can influence AMR dynamics.Urban areas often have higher levels of industrialization, international travel and exposure to diverse microbial populations, leading to potential hotspots for the emergence and spread of resistant strains.
Conversely, rural areas may experience unique challenges such as agricultural practices involving antibiotic use in livestock, which can contribute to the development of AMR.
The human-domestic animal-wildlife-environment interface plays a crucial role in the dynamics of AMR because, AMR is not confined to a single sector but involves complex interactions between humans, domestic animals, wildlife and the environment.
Human activities such as agriculture, livestock farming and aquaculture often involve the use of antibiotics.In these settings, resistant bacteria can emerge and spread among animals, potentially reaching humans through direct contact, consumption of contaminated food or environmental contamination.This interface also includes the environment, where antibiotic residues, resistant bacteria and genetic material can persist and be disseminated through water, soil and air, further facilitating the spread of AMR.Additionally, wildlife can serve as reservoirs of resistant bacteria, where AMR bacteria or genetic elements persist and can serve as a potential source for their transmission to humans, animals or the environment.
Unlike farm animals and humans, wild animals are rarely subjected to antimicrobial treatments.Therefore, the proportion of AMR bacteria in wild animals serves as an indicator of environmental contamination with AMR (Swift et al., 2019).Moreover, wild animals who live in or around the farmlands have access to consume animal waste such as dead carcasses.Also, their drinking water sources might be contaminated with resistant organisms through wastewater from farmlands (Ishibashi et al., 2019).Rats and shrews are common animal species that are found around the farms and a study reported that the AMR pattern of rats and shrews were closely associated with the AMR pattern of livestock animals in nearby/surrounding farms (Nhung et al., 2015).Similarly, wild animals can also have exposure to human-originated resistant organisms (Lagerstrom & Hadly, 2021).
The current study was based on the One Health approach to investigate and compare the AMR profiles in E. coli isolated from humans, animals and the environment and between the urban and rural areas.
The overall hypothesis of the study is that 'there is a difference in the proportion of AMR E. coli in domain between urban and rural areas'.The overall hypothesis can further expand into two sub-hypotheses: firstly, a larger proportion of people living in urban areas had AMR bacteria compared to people living in rural areas, and secondly wildlife can act as a reservoir and transmit AMR bacteria to livestock and human through the environment (Figure 1).

| MATERIAL S AND ME THODS
Figure 2 illustrate the model of the ecosystem in Sri Lanka.And also, Figure 2 shows the predictions made by this study.

Impacts
• Antimicrobial resistance (AMR) is a growing health problem worldwide and we evaluate the AMR profile of Escherichia coli (E.coli) isolated from all domains (humans, animals and environment) from urban and rural areas of Sri Lanka.
• The findings of the present study confirm that AMR E. coli are present in all domains of the area in Sri Lanka and a multi-sectorial One Health approach is necessary to combat the public health crisis due to AMR.

| Site selection
One km 2 area from an urban area and a rural area was selected, and areas were mapped using GPS to determine the exact sampling area.

| Urban area
The urban area (Dompe) is an area located in the Gampaha district in the Western province of Sri Lanka (Figure 3b).Gampaha is the district with the second-highest population density at 1539 persons per square km.The selected study site consisted of organized livestock farms mainly poultry and with urban wild animals.Furthermore, it consists of approximately 1200 households, one medium-scale (~6000 birds) poultry farm, two small scale (~1000 birds) broiler farms, around 15 backyard poultry farms, two medium-scale dairy farms (~60 animals), one government district hospital which has 102 beds, three private dispensaries and one medium-scale poultry processing plant.

| Rural area
The rural area (Dambana) is in Badulla District of Uva province, Sri Lanka where the population density is around 279 persons per square km (Figure 3c).This study site is home to the indigenous tribal community in Sri Lanka.There are no organized livestock farming or any other potential AMR sources in Dambana and wild animals reside in the selected study site.Furthermore, Indigenous medicine museum and Dambana Indigenous tourist centre is located in the study area that suggest higher visitor numbers to the study area.

| Sample collection
The sample collection was conducted in two different sites (urban and rural) in Sri Lanka between July 2020 and December 2021.Due to the COVID-19 pandemic situation, a non-probability convenience sampling technique was conducted to collect the samples.Moreover, the most readily available and easily accessible samples were collected.During the study period, faecal samples were collected from several components of each area namely humans, farm and wild animals.Furthermore, water and soil samples were collected as environmental samples (Table 1).

| Faecal samples of humans
Human faecal samples were collected from healthy individuals who had not taken antibiotics for at least 2 weeks period before sampling.
Since this study was conducted during the COVID-19 pandemic, a primary inclusion criterion for participant selection was individuals who were not infected with COVID-19 or had no close contact with COVID-19-infected patients.Researcher YDG visited households within the selected 1 km 2 area.Their consent was obtained, and they were advised on the sampling procedure.Each individual who gave consent to participate in the study was provided with a sterile faecal collection swab and advised on the sample collection procedure.
Briefly participants were instructed to insert the swab up to 1 cm into the anal orifice and touch the anal wall, peri-rectal area and anal wall, after passing stools and before washing.

| Faecal samples of animals
Voided fresh faecal samples were collected into a sterile sample collection bag from farm animals.In the urban area, faecal samples were collected from poultry and dairy animals, including cattle, buffalo and goats.In the rural ecosystem, only faecal samples from dairy cattle were collected since no other farm animals were being reared.
Wild animals in the study sites were also included.Freshly voided faecal samples were collected from wild birds and wild mammals, primarily through direct observation.When animals were not observed during defecation, fresh faecal samples were collected instead.Animal identification was then conducted based on the unique characteristics present in the collected faecal samples.Citizen scientists, involved in a parallel project, assisted in collecting data on wild animals within the studied 1 km 2 area.To aid in this data collection, This information was utilized to identify the locations where animals resided within the study areas.Faecal samples were collected from bats, monkeys and porcupines as representative wild mammals in the urban ecosystem, while samples from elephants, foxes, rabbits and porcupines were collected in the rural ecosystem.

| Water samples
Water samples were collected from water channels and stagnant water bodies that had a higher potential of exposure to main effluent water outflow tanks, farms, livestock processing plants, hospitals and household drainages.
Water was collected into sterile 250 mL glass bottles up to half of the container volume.The sampling depth for surface water samples was 6-12 inches below the water surface.

| Soil samples
Soil samples were collected by using sterile scooper from the topsoil in 0-10 cm depth in the north, south, east and west directions in four concentric circles at a distance of 0, 1, 2 and 4 feet away from the farm animals' lying areas/livestock sheds from the demarcated area within each ecosystem.The samples collected in each direction at each distance were pooled separately to form uniform samples and packed separately in sterile bags.

| Transportation of samples
All the samples were properly labelled and transported in a cool box at 4°C within 24 h of collection to the Food Microbiology laboratory of the Department of Veterinary Public Health and Pharmacology, Faculty of Veterinary Medicine and Animal Science, University of Peradeniya, Sri Lanka for further processing.on nutrient agar (NA) (Oxoid) and incubated at 37°C for 24 h.Pure isolates grown on NA were tested with a series of biochemical tests (Indole, Triple Sugar Iron Agar, Citrate and Urease) for confirmatory identification of E. coli.

| Isolation of E. coli from the water/ soil samples
Twenty-five mL/grams from each water/soil sample were homogenized with 225 mL buffered peptone water (BPW, Oxoid) and incubated overnight at 37°C for pre-enrichment.The following day, selective enrichment was done by transferring 1 mL of pre-enrichment broth to 9 mL of MacConkey broth (Oxoid) and overnight incubation at 37°C.A loopful from enriched MacConkey broth was streaked on Eosin Methylene Blue (EMB, Oxoid) agar on the next day and incubated overnight at 37°C.Three typical E. coli colonies with a green metallic sheen were picked from each plate and subcultured on NA.The same experiments described under faecal sample processing were carried out for identification of E. coli.

| Antimicrobial susceptibility test
Two E. coli isolates from each sample were subjected to antimicrobial susceptibility test (AST) using disk diffusion method for 12 clinically important antimicrobials following the guidelines given by the Clinical and Laboratory Standards Institute (CLSI).

| Storage of isolates for further analysis
All confirmed E. coli isolates were transferred to nutrient broth (NB, Oxoid) with 15% glycerol and stored at −80°C for further analysis.

| Data analysis
Statistical data analysis was carried out by SPSS version 27.In all these analyses, statistical significance was taken as p < 0.05.To analyse AST data, intermediate resistant readings were classified as resistant.Univariable binomial logistic regression analysis was used to examine the effects of the area and the effects of the domains of One Health on the proportion of AMR and MDR E. coli (Collett, 2002).

| Ethics approval
This study was approved by the ethics review committee, Faculty of Medicine, University of Peradeniya, Sri Lanka.Ethics grant No. 2020/EC/28.

| RE SULTS
In total, 574 samples were collected including 278 from the urban ecosystem and 296 from the rural ecosystem.During the study period, a total of 119 human faecal samples; 42 from urban and 77 from rural ecosystems were collected.From the urban ecosystem, 86 farm animal faecal samples were collected including 35 poultry and 51 livestock animals (28 cattle, 16 buffalo and, 7 goats) whereas it was 52 in the rural ecosystem.All samples from the rural ecosystem were from dairy cattle as no other farm animals were reared in that ecosystem.A total of 70 faecal samples from wild animals were collected from the urban ecosystem.Among these wild animal samples, 31 were from birds (mainly crows and pigeons) and 39 from mammalians (22 monkeys, 16 bats and 1 porcupine).There were 45 bird samples (more than 10 species of birds) and 42 mammalian samples (31 elephants, 4 foxes, 3 porcupines, 2 rabbits and 2 bats) bringing a total of 87 wild animal faecal samples in the rural ecosystem.
A total of 80 water samples, 40 from each ecosystem were enrolled for the study.In total, 80 soil samples were collected.Further, 40 soil samples were collected from 10 farm environments (2 dairy cattle farms and 8 poultry farms) from the urban ecosystem.40 soil samples were collected from the rural ecosystem, it was from the soil around the free-grassing cattle (a total of 7 places) and wild animals lying areas (a total of 3 places).1).

Urban area
Among all the isolates, 81.6% of E. coli isolates were found to be resistant isolates in the urban area.Nearly all E. coli isolates (98.0%) from the human samples were found to be resistant to at least one antimicrobial.Furthermore, the results of the univariable binomial  2a).For the comparison between the dairy and poultry domains, the dairy domain had a significantly low AMR E. coli proportion compared to the poultry domain.Similarly, in the comparison between the wild mammals and wild birds domains, the wild mammals domain had a significant p-Value ≤0.001, suggesting a reduced likelihood of AMR proportion in the wild mammals domain compared to the wild birds domain (Table 2b).

Rural area
There were 66.8% of E. coli isolates from the rural area that were found to be resistant for at least one antimicrobial agent tested.The highest percentage of AMR E. coli was found in human isolates (97%), and the lowest was in dairy animals (36.7%).Among wild animals and water, the percentages were 50.8% and 74.4% respectively.When comparing the domains of the rural area, E. coli isolates from humans were significantly more likely to be resistant than those isolated from the dairy, water and wild animal domain of the area (Table 3a).
In the rural ecosystem, the proportion of AMR E. coli in wild animals was significantly higher than that of dairy animals.There was no obvious difference in the proportion of AMR between E. coli isolated from faecal samples of wild mammals and wild birds (Table 3b).

E. coli resistant pattern between urban area and rural area
Figure 3a indicates the percentage of AMR profiles of the E. coli for 12 antimicrobials tested.E. coli shows the highest resistance against Streptomycin that was isolated from all the compartment except poultry.In poultry, E. coli shows the highest resistance against ampicillin (77.1%) followed by tetracycline (67.6%) nalidixic acid (66.2%) and sulpha trimethoprim (58.1%).Furthermore, 36.5% E. coli that was isolated from poultry shows resistance against ciprofloxacin and it was comparatively higher than in other domains of the area (dairy 11%, water 6.9%, humans 6% and wild animals 5.6%) (Figure 4a).
As shown in the figure , E. coli from all other domains except water in rural area showed the highest resistance to streptomycin whereas ampicillin had the highest resistance percentage among isolates from water.Comparatively the highest resistance was found for nalidixic acid among isolates from human samples (Figure 4b).

TA B L E 2
Univariable binomial logistic regression model outputs explaining the proportion of AMR in the domains of the urban area.
(a) The coefficient for the domains of the areas compared to the human domain of the ecosystem.(b) The coefficient for the dairy was compared with the poultry and wild mammals were compared with the wild birds.

Comparison of AMR between urban area and rural area
Overall, E. coli isolated from urban areas was significantly more likely to be resistant than those isolated from rural area (p-Value = 0.001).
Furthermore, the odds of having a proportion of AMR E. coli in samples from the urban area were increased by a factor of 2.203 (95% CI 1.567-3.096)compared with the rural are.
The proportion of E. coli isolated from dairy animals, wild birds and water was significantly higher in the urban area compared with rural area.There was no obvious difference in the proportion of AMR of E. coli isolated from humans and wild mammals between the different sites tested.
However, there was a tendency for wild mammalian faecal samples from rural area to have a higher proportion of resistant E. coli than wild mammalian faecal samples from the urban area (Table 4).

| Proportion of multidrug resistance in E. coli
The proportion of MDR E. coli was calculated among examined AMR E. coli isolated from the One Health domain of both areas and is shown in Table 5.

Urban area
Among AMR E. coli isolates, 41.7% of isolates of the urban area were found to be MDR isolates.

Comparison of MDR between areas
There was no obvious difference in the proportion of MDR between E. coli isolated from humans, wild animals and water between the two sites enrolled in the study (Table 6).The proportion of MDR in E. coli isolated from a faecal sample of dairy animals in an urban area was significantly higher than that from the rural area.As there were no organized poultry farms in the rural area it was not possible to compare this sector.

| AMR profile of E. coli isolated from soil collected from animal associated environment
In total, 80 soil samples were collected.Furthermore, 40 soil samples were collected from 10 farm environments (2 dairy cattle farms and 8 poultry farms) from the urban ecosystem.Another 40 soil samples were collected from the rural ecosystem, it was from the soil around the free-grazing cattle (a total of 7 places) and wild animal lying areas (a total of 3 places).E. coli isolation rate was 92.7% and it was similar among soil samples collected from both areas.Altogether, 207 E. coli colonies were isolated, and 141 colonies were subjected to the AST which was respectively 67 and 74 from the urban and rural ecosystem.Among tested isolates, 94% of isolated E. coli from urban area were found to be AMR, while it was 32.4% in the rural ecosystem.In the urban ecosystem, the highest resistance was observed to streptomycin (94.0%) followed by ampicillin (77.6%), tetracycline (53.7%) and similar resistance (40.3%) was observed in ciprofloxacin and sulpha trimethoprim.
A comparatively low resistance level was observed among E. coli isolated from the rural area where the highest resistance was observed in streptomycin (20.3%) followed by ampicillin (9.5%) and

| DISCUSSION
To best of our knowledge, this was the first 'One Health' study in Sri Lanka that aimed to investigate the AMR profiles in E. coli that were isolated from One Health domains of an ecosystem.
The findings of this study demonstrated the occurrence of AMR or MDR E. coli in Sri Lanka, regardless of the ecosystem's urban or rural context.However, the AMR E. coli prevalence was particularly critical in urban area except for the human population.A recently published study from India concluded that urban informal settlements in India as hotspots of AMR (Nadimpalli et al., 2020).
A study from Nepal, revealed that in Kaski, a rural area of Nepal, there was a high AMR prevalence among subsistence farming communities, their livestock and the environment (Subramanya et al., 2021).However, neither of these studies compared urban versus rural unlike the current study.
The human domain of the study areas had a significantly higher proportion of AMR, which was not altered irrespective of whether it was urban or rural.This is an unexpected finding as it was hypothesized that humans who may have a higher chance of exposure to antibiotics and AMR bacteria in the urban setting have a considerably higher risk of gut colonization of AMR E. coli than those who resided in the rural setting.Therefore, our findings contradicted the main hypothesis with regard to human population.This finding is consistent with the research that was conducted to evaluate AMR at the community level in an urban and rural setting in Karnataka, India where it was revealed that AMR was high in community members in both rural and urban settings (Balachandra et al., 2021).In parallel with this study, two separate studies were conducted at the same study sites and results of these studies were already published as journal articles (Gunasekara et al., 2022;Gunasekera et al., 2022); a quantitative survey that was conducted to determine the knowledge, attitudes and perceptions (KAP) on antibiotics and AMR among the general public.According to the KAP survey, misconceptions about antibiotics were highlighted among general public.For example, paracetamol, which is a painkiller, was thought to be an antibiotic by more than 50% of both urban and rural respondents.In addition, 18.5% urban and 35.4% rural participants admitted they would keep and re-use what they perceived as leftover antibiotics.The misconception regarding antibiotics can lead to their misuse, which in turn can contribute to the emergence of AMR within the respective population (Gunasekera et al., 2022).Another study based on qualitative methodology was conducted to investigate the KAP on antibiotics and AMR among healthcare professionals.According to the results, healthcare professionals showed poor awareness regarding the spread of AMR.The poor KAP regarding AMR among health professionals residing in both study sites contribute to the misuse of antibiotics among community members (Gunasekara et al., 2022).
Finally, it can be concluded that misuse of antibiotics and their poor KAP on antibiotics and AMR as a potential explanation for the high prevalence of AMR E. coli in the human domain of both ecosystems.
Direct links to associate human gut colonization by AMR bacteria and livestock has been attempted in Kenya (Muloi et al., 2019) and Bangladesh previously (Rousham et al., 2021).In these studies, human density and the proportion of livestock manure increased the AMR but keeping livestock (Muloi et al., 2019), and high (poultry seller, slaughter, poultry farm worker and poultry owners) or low (non-poultry seller, non-farm worker, non-poultry owner) exposure to poultry (Rousham et al., 2021) did not increase AMR bacteria in humans.In the current study, an association between livestock (dairy/poultry) and human colonization cannot be established as humans had similar numbers of AMR bacteria despite different exposure to livestock in urban area and rural ecosystem.
In the case of the dairy animals domain, the wide confidence in- Although it is necessary to understand the resistant profiles of AMR bacteria determining the most prevalent drug resistance combination remained complicated.In the human domain of the current study, the highest resistance to AMR was observed with streptomycin followed by ampicillin and nalidixic acid.17 distinct forms of MDR patterns were identified in the urban area but the most prevalent MDR pattern in the rural area was aminoglycosides, penicillin, quinolones and cephalosporins.A likely explanation can be the extensive use of these antibiotics in the community leading to differ-  (Kulasooriya et al., 2014).It is not surprising, that AMR in tertiary care hospitals in Sri Lanka was increased (Nakkawita et al., 2021).Furthermore, among tested E. coli isolated from blood cultures, only 9.5% were sensitive to ampicillin whereas 33.9% were to amoxicillin-clavulanic acid.Also, only 36.1% and 25.5% of isolated E. coli were susceptible to piperacillin-tazobactam and ciprofloxacin respectively.It is noteworthy to mention that, 7.5% of isolates were resistant to meropenem.
AMR E. coli were found in faecal samples collected from both the dairy and poultry sectors because, antibiotics are broadly used in animals to treat infectious diseases ranging from simple topical infections to severe systemic illnesses, as prophylactic agents in disease prevention and as growth promoters in sub-therapeutic levels in healthy animals (Rico et al., 2013).According to the results of the current study, AMR and MDR in poultry farming was higher when compared to dairy farming samples.Our study findings are consistent with the previous studies that were conducted worldwide (Aworh et al., 2021;Faridah et al., 2020;Subramanya et al., 2021).The similarity observed between our study findings and the above-mentioned studies may be due to similarities in poultry farming practices because the majority of poultry are reared under intensive management systems where birds are kept in high densities, making them susceptible to infectious diseases.
Lack of proper biosecurity measures such as vaccination and hygienic practices in such scenarios further aggravates the situation and increases the proportion of infectious diseases leading to the usage of antimicrobials and the development of AMR (Hedman et al., 2020).Another possible explanation for the high proportion of AMR/MDR in poultry could be due to the differences in antibiotic administration routes because oral antibiotic prescriptions are at a negligible level in the dairy sector.For example, a study conducted in Tanzania revealed that 98% of poultry farms used oral antibiotics while all studied dairy farms used parenteral administration of antibiotics (Azabo et al., 2022).In the end, gut colonization of AMR E. coli will be lower in dairy compared with the poultry sector where oral antibiotics are used vastly.Farmers and veterinarians who had poor KAP regarding antibiotics usage and AMR play a significant role in the rise of AMR within livestock worldwide (Adekanye et al., 2020;Dankar et al., 2022;Hossain et al., 2022;Sharma et al., 2020;Vijay et al., 2021;Wangmo et al., 2021).The implementation of effective antibiotic stewardship practices holds significant importance in addressing the prevalent issue of antibiotic overuse within the livestock sector.This overuse specifically pertains to antibiotics that are vital for treating human diseases.
It is imperative to acknowledge and actively address this concern due to the potential consequences it poses for both human health and the environment.But the poor implementation of AMR prevention policy and laws, undue influence by pharmaceutical companies and over-the-counter availability of veterinary medical products in low-and middle-income countries including Sri Lanka are considered major barriers to antibiotic stewardship within the animal domain (Gozdzielewska et al., 2020).
AMR isolates were found in E. coli isolated from faecal samples collected from wild animals in both areas.Because wild animals who live in or around the farmlands have access to and consume animal waste such as dead carcasses and their drinking water sources might be contaminated with resistant organisms through wastewater that arises from farmlands and communities.This is an important finding and builds towards the third hypothesis of the current study that wildlife can act as a reservoir and transmit AMR bacteria to livestock and human through the environment.Furthermore, it is essential to expand the research to investigate the role of wild animals in relation to their potential as reservoirs of AMR.Our finding is comparable with other studies conducted worldwide (Abdullahi et al., 2021;Plaza-Rodríguez et al., 2021;Swift et al., 2019;Torres et al., 2020).
While wild animals can serve as environmental sentinels for AMR contamination, it will be challenging to utilize them to assess the extent of AMR prevalence within the ecosystem and compare the prevalence across ecosystems using different animal species, because the proportion of AMR among wild animals is influenced by both the level of AMR contamination in the environment and the feeding habits of wild animals.Moreover, compared to herbivorous wild animals, omnivorous and carnivorous wild animals have a high potential to get infected by a resistant organism through human and animal waste and wastewater (Bamunusinghage et al., 2022).Also, their extensive home range and migratory behaviour of wild animals make them vulnerable to exposure to AMR from another area (Elsohaby et al., 2021).
The urban environmental domain showed higher contamination levels with AMR E. coli compared to rural environments.This is supported by the significantly higher proportion of AMR E. coli isolated from samples of urban water sources.This finding can be attributed to multiple sources, including livestock, hospitals, industries and human activities, which release antimicrobials into urban environments in varying quantities.This explanation is supported by the finding that AMR organisms are present in hospital wastewater in Sri Lanka (Guruge et al., 2021).A study conducted in Sri Lanka reported the occurrence of AMR organisms in municipal sewage treatment plants (Samaraweera et al., 2019).Similar to our finding, environmental contamination of AMR organisms around the farmlands were reported worldwide previously (Macedo et al., 2021;Peng et al., 2021;Smith et al., 2019).
This study has some limitations.It was conducted within two areas of one square km area each and this small area may not represent other urban and rural areas in Sri Lanka.Therefore, to generalize the findings, specific aspects of the urban and rural contexts that were used as defining criteria must be taken in to consideration.
Also, this study only investigates the phenotypic resistance of E. coli, but genotypic resistance may be different and that was one of the major limitations of this study.The genotypic resistance profiles can prove direct links between the One Health domains, but due to lack of resources genotypic profiles were not

F
I G U R E 1 Figure illustrates the predictions made by this study, suggesting that wild animals serve as both reservoirs and transmission routes for AMR bacteria across different domains of the One Health framework.Wildlife can act as indicators of environmental contamination with AMR. a mobile application called 'Nature Citizen' was developed, allowing the capture of photographs of wild animals along with GPS data.
Isolation of E. coli from the faecal samples Whole faecal samples from humans, farm animals and wild animals were directly streaked with a sterile wire loop onto MacConkey agar (Oxoid) and incubated at 37°C for 24 h.The next day, three lactosefermenting colonies with typical E. coli morphology (Bright pink colonies) were randomly selected from each plate and sub-cultured F I G U R E 2 Figure represents a hypothetical environment in Sri Lanka, illustrating the spread of AMR bacteria within the area from a One Health perspective.Humans and animals can share AMR bacteria through direct contact, food and water.Soil and water contamination through refuse and sewage disposal can cause the AMR bacteria to spread between humans, animals and the environment through multiple pathways.
logistic regression model explaining the proportion of AMR E. coli in the domains of the urban area are presented in Table 2.The poultry domain showed a non-significant different compared to the human domain.The dairy domain had a significant p-Value (≤0.001) indicating a lower likelihood of AMR proportion compared to the human domain.The wild animals domain also showed a significant p-Value (0.01) indicating a reduced likelihood of AMR proportion compared to the human domain (Table indicates that there is substantial uncertainty in quantifying the true magnitude of the association between this domain and the proportion of AMR.The wide confidence interval could be due to various factors, such as limited sample size, variability in the data or potential confounding variables.It indicates the need for further research with larger sample sizes or more comprehensive data to obtain a more precise estimate of the relationship between the dairy animals domain and AMR proportion.
Sample number, proportion of E. coli positive samples, proportion AMR E. coli and MDR E. coli in each domain.poultry.In the rural area, 247 samples out of 296 (70.9%) were positive for E. coli and diary animals had the highest E. coli positive rate (96.2%), whereas 66.7% of samples from wild mammals in the urban area reported the lowest positivity rate.The detailed results are shown in Table1.3.1.1| Proportion of AMR E. coli in each domain of One Health in urban area and rural area A total of 908 E. coli isolates were subjected to AST from both areas, including 149 humans, 245 farm animals, 236 wild animals, 137 water and 141 soil, respectively.The proportion of phenotypic AMR is expressed as the percentage of resistant E. coli to the isolated E. coli (Table In the urban ecosystem, 235 (84.53%) samples were positive for E. coli and the highest positive rate (94.26%) of E. coli was obtained TA B L E 1

Value Exp(B) 95% CI for Exp(B) Lower Upper
TA B L E 4Univariate binomial logistic regression model outputs explaining the proportion of AMR in the domains of One Health.p-and it was significantly higher than those were observed in all other domains.The second-highest MDR proportion was observed in AMR E. coli from humans followed by wild animals (Table4).The MDR E. coli isolated from both humans and poultry exhibited 17 different types of MDR patterns and among them, 22 (40%) MDR isolates from poultry and 7 (35%) MDR isolates from human showed resistance to at least 5 antibiotic classes that were tested in the study.The most common (32.7%)MDR patterns observed in E. coli isolated from poultry were resistant to aminoglycosides, penicillin, tetracycline, quinolones and sulphonamides.When considering the environmental domain, 14 different types of MDR patterns were observed in E. coli isolated from water.Furthermore, among these patterns, 10 patterns were observed in either human or farm animal domains.The most common MDR pattern 16.7% (4/24) in E. coli isolated from wild animals were aminoglycosides, penicillin, tetracycline and amphenicols.Furthermore, these 4 isolates were found in faecal samples collected from wild birds.However, this resistant combination was not found in either E. coli isolated from humans, farm animals or water.sistance pattern.When considering the environmental component of the rural area, seven different types of MDR patterns were observed in E. coli isolated from water samples whereas it was 14 in E. coli isolated from faecal samples of wild animals.Among these 14 MDR patterns, 7 patterns were not found in either E. coli isolated from human or water samples.These 7 unique patterns comprised 51.6% (16 /31) of all MDR isolates obtained from wild animals.The majority (14/16) of these patterns were observed in E. coli isolated from faecal samples of wild birds.

Prevalence of MDR among examined AMR E. coli (95% CI a ) Total MDR isolates Prevalence of MDR among examined AMR E. coli (95% CI a )
Prevalence of MDR E. coli among examined AMR E. coli isolate from the two areas.
TA B L E 5 Comparison of proportion of MDR E. coli between two areas.current study, human gut colonization of MDR E. coli among the Sri Lankan community has been reported in a study conducted to investigate the AMR pattern of human faecal E. coli among citizens in Kandy.Nearly 60% of isolated E. coli were resistant to more than four antibiotics that were tested the