A One Health Review of Community-Acquired Antimicrobial-Resistant Escherichia coli in India

Antimicrobial resistance (AMR) threatens to undermine nearly a century of progress since the first use of antimicrobial compounds. There is an increasing recognition of the links between antimicrobial use and AMR in humans, animals, and the environment (i.e., One Health) and the spread of AMR between these domains and around the globe. This systematic review applies a One Health approach—including humans, animals, and the environment—to characterize AMR in Escherichia coli in India. E. coli is an ideal species because it is readily shared between humans and animals, its transmission can be tracked more easily than anaerobes, it can survive and grow outside of the host environment, and it can mobilize AMR genes more easily than other intestinal bacteria. This review synthesized evidence from 38 studies examining antimicrobial-resistant E. coli (AR-E) across India. Studies of AR-E came from 18 states, isolated from different sample sources: Humans (n = 7), animals (n = 7), the environment (n = 20), and combinations of these categories, defined as interdisciplinary (n = 4). Several studies measured the prevalence of AMR in relation to last-line antimicrobials, including carbapenems (n = 11), third-generation cephalosporins (n = 18), and colistin (n = 4). Most studies included only one dimension of the One Health framework, highlighting the need for more studies that aim to characterize the relationship of AMR across different reservoirs of E. coli.


Introduction
Antimicrobial resistance (AMR) has been recognized as a major global health threat. Recent predictions estimate that deaths caused by AMR could reach 10 million by 2050 [1]. Recognizing the scale and importance of this global issue, the World Health Organization (WHO) has initiated a Global Action Plan to mitigate the effects of AMR. A key component of the Global Action Plan is to improve AMR surveillance capacity, especially in lowand middle-income countries (LMICs), with a One Health approach [2]. One Health is an approach that recognizes the interconnections between people, animals, and the environment [3]. It is applicable to understanding Escherichia coli (E. coli) because of its prevalence and high transmission rates between humans, animals, and environmental interfaces. We focus on the role of AMR in E. coli because it is perhaps the most studied indicator organism, and its transmission can be tracked more easily (among animal hosts) than anaerobes [4][5][6][7]. Furthermore, E. coli can survive and even grow in the environment outside of the host, and may mobilize AMR genes more easily than other intestinal bacteria (such as Bacteroides) [8][9][10][11][12][13][14]. Many strains of E. coli have acquired the virulence genes necessary to cause a wide spectrum of intestinal and extra-intestinal infections such as diarrhea, urinary tract infections, and both community-and hospital-acquired bacteremia.
This review was conducted between May to September 2020 and followed the preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines [24]. The study focused on the prevalence of AR-E in environmental samples of water obtained from lakes and rivers, of food animal sources (e.g., fish, chicken, vegetables, and bovine milk), and of community-acquired AMR in humans, and it targeted samples collected in community or outpatient settings (we excluded nosocomial AMR). We defined interdisciplinary studies as those that investigated the occurrence of AR-E in a combination of the two or three categories-environmental samples, animal food sources, and/or human samples.
PubMed was searched and the query terms used are outlined in the Appendix. No limit on publication dates was set. The database was queried on 19 June 2020. Results were imported into Covidence (www.covidence.org accessed on 20 June 2020), a systematic review management software, and duplicates were removed. All included studies focused on AR-E. Studies that included human isolates only assessed community-acquired infections in the human population. We searched PubMed (https://pubmed.ncbi.nlm.nih.gov, accessed on 19 June 2020) using the following terms pertaining to antimicrobial resistance: ("Enterobacteriaceae" OR "Gram negative bacteria" OR "E. coli" OR "Escherischia coli") AND ("Drug Resistance" OR "Extended-spectrum beta-lactamase" OR "ESBL" OR "antibiotic" OR "antimicrobial" OR "AMR") AND ("resistance" OR "resistant") AND ("human" OR "community acquired" OR "community-acquired" OR "livestock" OR "poultry" OR "cattle" OR "cows" OR "pets" OR "chickens" OR "Environment" OR "water") AND ("India"). We also searched the resulting reference lists to identify additional articles. The search terms are described in Table A1.
After the initial search, the research focus was narrowed further to include only E. coli. If studies examined other bacterial species along with E. coli, they were included. However, only information relevant to findings related to E. coli were included to improve comparability. Of the 747 non-duplicate results, 613 studies were excluded based on the title and abstract screening process. Common reasons for exclusion included: (1) Focus was on human clinical AMR; (2) predominantly reported AMR in other species within Enterobacteriaceae and other Gram-negative bacteria; (3) study conducted outside of India; (4) study focused on isolation and molecular characterization of phages as opposed to AMR. Of the 134 selected articles, 96 articles were excluded after reviewing the full manuscript (see Figure 1). After a full-text review, 38 full texts were included for data extraction. Publications that described human or animal populations or environmental samples, bacteria isolates, and specific laboratory methods such as disk diffusion, antimicrobial susceptibility patterns, interpretation of resistance profiles of E. coli, and multidrug resistance (MDR) were considered and included in the study. After the initial search, the research focus was narrowed further to include only E. coli. If studies examined other bacterial species along with E. coli, they were included. However, only information relevant to findings related to E. coli were included to improve comparability.
Of the 747 non-duplicate results, 613 studies were excluded based on the title and abstract screening process. Common reasons for exclusion included: (1) Focus was on human clinical AMR; (2) predominantly reported AMR in other species within Enterobacteriaceae and other Gram-negative bacteria; (3) study conducted outside of India; (4) study focused on isolation and molecular characterization of phages as opposed to AMR. Of the 134 selected articles, 96 articles were excluded after reviewing the full manuscript (see Figure 1). After a full-text review, 38 full texts were included for data extraction. Publications that described human or animal populations or environmental samples, bacteria isolates, and specific laboratory methods such as disk diffusion, antimicrobial susceptibility patterns, interpretation of resistance profiles of E. coli, and multidrug resistance (MDR) were considered and included in the study.

Results
The systematic review included studies from 18 states spread across the entire geography of the country. Thirty-eight studies were included in total, with samples collected from humans (n = 7), the environment (n = 20), animals (n = 7), and interdisciplinary sources (n = 4).
All AMR research studies (n = 38) applied culture-based antimicrobial susceptibility testing (AST) using the disk diffusion method. Studies (n = 29) also applied polymerase chain reaction (PCR) for AMR gene identification and assessing AMR mobile genetic elements. Studies that employed PCR provided estimates of resistance genes for the isolates identified.
In Figure 2, we tabulated the prevalence of antimicrobial resistance in E. coli for six drugs that have high clinical relevance: Ampicillin, amikacin, cephalosporins, ceftazidime, imipenem, and polypeptides that span five different classes of antimicrobials. In Figure 3, we map the location of the studies included in this review.

Results
The systematic review included studies from 18 states spread across the entire geography of the country. Thirty-eight studies were included in total, with samples collected from humans (n = 7), the environment (n = 20), animals (n = 7), and interdisciplinary sources (n = 4).
All AMR research studies (n = 38) applied culture-based antimicrobial susceptibility testing (AST) using the disk diffusion method. Studies (n = 29) also applied polymerase chain reaction (PCR) for AMR gene identification and assessing AMR mobile genetic elements. Studies that employed PCR provided estimates of resistance genes for the isolates identified.
In Figure 2, we tabulated the prevalence of antimicrobial resistance in E. coli for six drugs that have high clinical relevance: Ampicillin, amikacin, cephalosporins, ceftazidime, imipenem, and polypeptides that span five different classes of antimicrobials. In Figure 3, we map the location of the studies included in this review. The AR-E burden has serious implications for human health, owing to the potential transmission of bacteria from animals and the environment to humans, thereby impairing the efficacy of antimicrobial treatment and compromising public health. All included studies were in the English language and the key study findings are described in Table  A2. Below, we summarize the study results by the source from which the AR-E were derived.

AR-E in Humans
Seven studies examined AR-E in human isolates, three of which used human stool samples and the other four used urine samples. These studies reported the prevalence of community-acquired AMR (i.e., no hospital-acquired infections were included). Of these, three studies reported the highest susceptibility of E. coli to amikacin [25][26][27]. In New Delhi, one study examined the prevalence of AMR to β-lactams in neonates (less than 60 days old) with no history of antimicrobial use. Nearly 87% of E. coli from neonates were resistant to ampicillin [28]. This study was the first to demonstrate the load of communityacquired beta-lactamase-producing E. coli and suggested that the babies likely acquired these strains from the maternal flora.

AR-E in Animals
Seven studies included in this research examined the presence of E. coli in samples derived from animals that are mainly consumed as food and constitute an important part of the human diet. The detection of ESBL-producing E. coli was conducted in three of these The AR-E burden has serious implications for human health, owing to the potential transmission of bacteria from animals and the environment to humans, thereby impairing the efficacy of antimicrobial treatment and compromising public health. All included studies were in the English language and the key study findings are described in Table A2.
Below, we summarize the study results by the source from which the AR-E were derived.

AR-E in Humans
Seven studies examined AR-E in human isolates, three of which used human stool samples and the other four used urine samples. These studies reported the prevalence of community-acquired AMR (i.e., no hospital-acquired infections were included). Of these, three studies reported the highest susceptibility of E. coli to amikacin [25][26][27]. In New Delhi, one study examined the prevalence of AMR to β-lactams in neonates (less than 60 days old) with no history of antimicrobial use. Nearly 87% of E. coli from neonates were resistant to ampicillin [28]. This study was the first to demonstrate the load of community-acquired beta-lactamase-producing E. coli and suggested that the babies likely acquired these strains from the maternal flora.

AR-E in Animals
Seven studies included in this research examined the presence of E. coli in samples derived from animals that are mainly consumed as food and constitute an important part of the human diet. The detection of ESBL-producing E. coli was conducted in three of these studies, and the results confirmed ESBL production as follows: 54.5% E. coli isolated from milk (n = 22 isolates) [29]; 25.4% E. coli isolated from pig feces (n = 867 isolates) [30]; 71.6% E. coli isolated from fresh seafood samples (n = 475 isolates) [31]. In a study conducted by Vinayananda et al., the researchers sampled 840 eggs available in markets across southern India. The samples included table eggs from three sources: Processed commercial layer farms, unprocessed commercial layer farms, and free-range eggs collected from household backyards. E. coli was present in 28.5% of the overall samples with an occurrence rate of 22.9, 29.2 and 50.0% from processed, unprocessed and free-range table eggs, respectively.
The study recovered 24 isolates, in which 100% were resistant to cloxacillin, irrespective of the source from where the eggs where obtained. The multiple antimicrobial resistance (MAR) index was calculated and interpreted according to Krumperman (1983) using the formula: a/b, where "a" represents the number of antimicrobials to which a particular isolate was resistant and "b" the total number of antimicrobials tested. The pattern of antimicrobial resistance was similar among the three sources, with an MAR index of 0.19 for free-range table eggs, 0.23 for unprocessed eggs, and 0.28 for processed eggs. The authors stated that the high MAR index demonstrated the need to control the use of antimicrobials in animal farming to prevent cross-transfer of ESBL-producing E. coli to humans, as eggs function as a staple in the Indian diet [32].

AR-E in the Environment
The majority of studies included in this systematic review focused on the presence of AR-E in the environment (n = 20). Sources were predominantly water samples from lakes, reservoirs, and river water, while some other studies included sources such as vegetables and fruits. Several studies focused on pathogenic forms of E. coli and examined the presence of Shiga toxin genes stx1 and stx2 [33][34][35][36][37][38]. The presence of virulence genes specific to certain E. coli pathotypes may cause severe health outcomes such as diarrhea, urinary tract infections, hemolytic colitis, neonatal meningitis, nosocomial septicemia, hemolytic uremic syndrome, and surgical site infections [39]. In a study that focused on determining the prevalence of resistant coliforms in the Yamuna river, 123 (86.6%) isolates showed resistance to three or more drug classes and were considered MDR. High resistance rates were observed for cefazolin (88.7%), followed by vancomycin (74.6%), cefuroxime and cefotaxime (60%), and gentamicin (52.8%). One fourth of the total isolates (n = 141) were resistant to 10 or more drugs [40].

AR-E in Interdisciplinary Studies
Four studies included in this review were classified as interdisciplinary. Only two studies, by Sahoo et al. and Puii et al., analyzed E. coli from all three dimensions of the One Health approach (i.e., environment, animals, and humans). Sahoo et al. investigated AR-E isolated from child stool samples, cow dung, and drinking water. Ninety percent of the isolates were resistant to at least one antimicrobial [41]. E. coli isolates from non-coastal regions exhibited significant resistance to second-and third-generation cephalosporins and nalidixic acid [41]. Rasheed et al. examined resistance in isolates of E. coli from vegetables and animal food sources (i.e., meat, eggs, and milk). The overall prevalence of drug-resistant E. coli was 14.7%. Pathogenic E. coli cycling through food is of particular concern, as it poses a significant health risk to humans [11].

Discussion
Among the studies reviewed, extensive resistance to ampicillin and third-generation cephalosporins (3GC) were observed, while carbapenem resistance was less common. Many studies discussed the importance of unrestricted access to over-the-counter antimicrobials as a factor amplifying the problem of AMR [28]. Fundamental change is required in the way that antimicrobials are prescribed, distributed, and consumed. Risk assessments conducted by the WHO show that the Southeast Asia region is likely to be the most at-risk part of the world [2]. In light of the recent global COVID-19 pandemic, experiences and consequences need to be translated into policy and on the ground action to prevent the more silent but rising AMR pandemic. The unprecedented speed at which new strains and "superbugs" develop and are transmitted globally needs to be met with an equally urgent response by the global community. AMR needs to be addressed as a serious and potentially catastrophic issue with global significance.
The presence of ESBL genes isolated from various sample sources is a great concern for human beings [28,29,[42][43][44]. Drug resistance can be transferred between bacteria in the gut of humans and animals and this can greatly complicate treatment. The antimicrobial susceptibility profiles of E. coli isolates included in this review indicate that most of them have acquired multidrug resistance against a broad set of antimicrobial classes. In the studies we reviewed, the presence/absence of a resistance gene does not necessarily mean that the genes are expressed.
The emergence of multidrug-resistant E. coli isolates, involving co-resistance to four or more unrelated families of antimicrobials, is a serious concern, and many of the studies in this review demonstrated a high prevalence of MDR E. coli [45]. Verma et al. highlighted that E. coli strains from vegetables and fruits exhibited resistance to almost all classes of drugs, including quinolones, carbapenems, penicillin, and aminoglycosides. Furthermore, the review highlighted that many researchers apply different definitions in their AMR studies. For example, the definition of MDR varied across the studies. In some studies, MDR E. coli was defined as isolates resistant to five or more antimicrobials. For example, Sukumaran et al. reported 53.3% of E. coli serotypes showing resistance to five or more antimicrobials and classified them as MDR [46]. However, studies conducted by Abhirosh et al. and Mohanta and Goel defined MDR as those resistant to three or more antimicrobials [47,48]. This review demonstrated that more needs to be done in order to create standard definitions and protocols for studying AMR in India.
There is a lack of research on exposure assessment and exposure patterns to AR-E. There is a need for more quantitative microbial risk assessment in order to assess and objectively quantify the risk of human exposure to AR-E. There is a significant need to identify the highest risk sources (e.g., food or drinking water), geographic locations, and human-animal or human-environment interfaces that result in the greatest risk of humans to be exposed to AMR.
The prevalence of antimicrobial-resistant infections caused by extended spectrum β-lactamase (ESBL)-producing E. coli has increased over the last decade across every geographic region of the world.This study highlights the presence of diverse reservoirs of drug-resistant E. coli. Most studies are not able to confirm whether the increase in the number of resistant strains isolated in one medium such as water necessarily causes an increase in the number of strains isolated from animals consuming this water.

Conclusions
The prevalence of resistance to last-resort antimicrobials such as colistin and imipenem increases treatment complications. This should be viewed seriously from a population outcomes perspective, as these classes of drugs are categorized as lifesaving and used in treating serious infections. In India, countrywide measures for the surveillance of AMR needs to be strengthened and expanded. Surveillance should incorporate both environmental samples and clinical isolates from humans and animals in order to better determine the sources and drivers of AMR.
The prevalence of AR-E across various interfaces in the community is a reminder of the intimate and delicate relationship between humans, animals, and the environment. The studies included in this review individually highlight the prevalence of AMR across different areas of India and across multiple sources. There was difficulty, however, in making comparisons between the results of included studies given the diverse settings, methods used, and sources of samples. More research should apply a multidisciplinary methodology focusing on the interfaces between AMR in humans, the environment, food and animals, and/or social-ecological systems. We also need more microbial risk assessment research that aims to define the root causes of AMR and the factors that increase the risk of human exposure to AMR. Finally, there is a need to develop feasible and sustainable public health interventions that encompass the One Health approach and focus on developing a holistic solution to AMR.
Funding: This research received no external funding.

Informed Consent Statement: Not applicable.
Data Availability Statement: All articles and reports used for this review are available to the public.

Conflicts of Interest:
The authors declare no conflict of interest, and that they have no commercial associations that might create a conflict of interest in connection with the submitted manuscript.    Sukumaran et al. [46] Cochin Estuary, Vembanadu Lake Environmental Water samples from five water stations 75

Appendix A
More than half of the isolates were MDR. Two E. coli isolates were resistant to more than seven antimicrobials and both contained class 1 integrons.
Abhirosh et al. [47] Vembanadu Lake Environmental Lake water samples from 10 locations 33 Multidrug resistance (i.e., resistance to three or more drugs) in E. coli was 100%.