https://orcid.org/0000-0002-8370-7911
Figures
Abstract
Extended-spectrum beta-lactamase (ESBL)-producing Escherichia (E.) coli have been widely described as the cause of treatment failures in humans around the world. The origin of human infections with these microorganisms is discussed controversially and in most cases hard to identify. Since they pose a relevant risk to human health, it becomes crucial to understand their sources and the transmission pathways. In this study, we analyzed data from different studies in Germany and grouped ESBL-producing E. coli from different sources and human cases into subtypes based on their phenotypic and genotypic characteristics (ESBL-genotype, E. coli phylogenetic group and phenotypic antimicrobial resistance pattern). Then, a source attribution model was developed in order to attribute the human cases to the considered sources. The sources were from different animal species (cattle, pig, chicken, dog and horse) and also from patients with nosocomial infections. The human isolates were gathered from community cases which showed to be colonized with ESBL-producing E. coli. We used the attribution model first with only the animal sources (Approach A) and then additionally with the nosocomial infections (Approach B). We observed that all sources contributed to the human cases, nevertheless, isolates from nosocomial infections were more related to those from human cases than any of the other sources. We identified subtypes that were only detected in the considered animal species and others that were observed only in the human population. Some subtypes from the human cases could not be allocated to any of the sources from this study and were attributed to an unknown source. Our study emphasizes the importance of human-to-human transmission of ESBL-producing E. coli and the different role that pets, livestock and healthcare facilities may play in the transmission of these resistant bacteria. The developed source attribution model can be further used to monitor future trends. A One Health approach is necessary to develop source attribution models further to integrate also wildlife, environmental as well as food sources in addition to human and animal data.
Citation: Perestrelo S, Correia Carreira G, Valentin L, Fischer J, Pfeifer Y, Werner G, et al. (2022) Comparison of approaches for source attribution of ESBL-producing Escherichia coli in Germany. PLoS ONE 17(7): e0271317. https://doi.org/10.1371/journal.pone.0271317
Editor: Jan Kluytmans, Amphia Ziekenhuis, NETHERLANDS
Received: November 18, 2021; Accepted: June 28, 2022; Published: July 15, 2022
Copyright: © 2022 Perestrelo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the article and its Supporting Information files.
Funding: This work was supported by funding from the European Union’s Horizon 2020 Research and Innovation program under grant agreement No 773830: One Health European Joint Program and is a contribution to the DiSCoVeR Joint Research Project; and by the RESET Consortium which is financially supported by the German Federal Ministry of Education and Research (BMBF) through the German Aerospace Center (DLR) grant numbers 01KI1013A-H The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Extended-spectrum beta-lactamase (ESBL)-producing Escherichia (E.) coli were first mentioned in 1983 [1] and have been widely known for several years as the cause of treatment failures around the world. ESBLs are enzymes hampering the use of beta-lactam antibiotics such as penicillins and early cephalosporins, but also cephalosporins of the 3rd and 4th generation and monobactams. This broad activity is the reason, why they are designated as beta-lactamases with an extended spectrum. Given that, they pose an additional risk for patients suffering from a serious infection of ESBL-producing E. coli as treatment may be compromised. The ESBL encoding genes may be further transmitted to other bacterial clones, or species through horizontal gene transfer due to their location on mobile genetic elements (MGE) such as plasmids [2]. Apart from this, E. coli harboring ESBLs are frequently multidrug-resistant (MDR), to not only β-lactams but also to other antimicrobial classes, such as quinolones, aminoglycosides, tetracyclines and sulfonamides [3], with responsible genes being co-located on the same MGE as the ESBL genes or other MDR plasmids in the same host, or in the hosts chromosome.
The origin of colonization with ESBL-producing E. coli amongst humans is controversial and the role of the environment, foodstuffs, pets, wildlife, livestock animals and healthcare facilities is still under discussion. Furthermore, the routes and direction of transmission between animals and humans are not clear and it is more likely to be rather bilateral than unilateral, meaning that humans may also contribute to the increased presence of ESBL-producing bacteria amongst animals and the environment. In addition, the human-to-human transmission has been further mentioned with more studies referring to the increasing role of travelers, clinical patients and farmers as risky vehicles of ESBL-positive bacteria into the general community [4, 5].
In Germany, data on the prevalence of ESBL-producing E. coli was gathered in 2011–2013 by the RESET-consortium (http://www.reset-verbund.de), which has broadened up the understanding of the current situation in the country. Many studies have described and characterized the presence of ESBL-carrying E.coli in Germany in different environments and hosts; on animal farms [6, 7], in the environment [8], in hospitals [9] and in the general human population [10].
Source attribution modeling based on microbial subtyping has been used to identify potential animal and non-animal sources of different zoonotic bacteria in various countries, such as Denmark (for Salmonella) [11], Netherlands (for Shiga toxin-producing E. coli (STEC)) [12] and New Zealand (for Campylobacter) [13]. These mathematical models use subtyping data from molecular as well as phenotypical techniques.
In 2019, Jabin et al. [14] used two different source attribution models to predict the origin of Salmonella amongst humans from different food sources (broilers, laying hens, pigs and turkeys) in Germany. Nevertheless, there was still lacking a similar study in Germany regarding ESBL-producing E. coli. In this study, we aim to analyze the available data from the RESET project in Germany and to apply a Bayesian source attribution model to gain more knowledge about the origin of ESBL-producing E. coli cases in the general population. The sources considered in this study are cattle, chicken, pigs, horses, dogs and nosocomial cases.
This kind of analysis can add useful information for the important public health question of at which point(s) interventions against the rise of antibiotic resistance may have the largest effect. If the largest part of resistant bacteria in humans are attributable to livestock (and to the corresponding foods), this might suggest that focusing on animals in order to reduce resistances might be the most effective thing to do. This could include studying mitigating strategies against the spread of resistant bacteria in livestock like for example veterinarian antimicrobial (also called antibiotic) stewardship [15–17] or husbandry practices [18, 19]. If, on the other hand, the largest part of resistances circulates within human populations then this might suggest that it is more effective focusing on the human side in order to mitigate the rise of resistant strains of bacteria (again for example through antibiotic stewardship [20]). If finally both, human and animal sources play a considerable role a broader one health approach seems more appropriate [21, 22].
Material and methods
Basis for this analysis were E. coli isolates from several studies on the prevalence and phenotypic and genotypic characteristics of ESBL-producing E. coli in Germany conducted firstly within the research consortium RESET and then recovered for the present study within the DiSCoVer project (One Health EJP consortium).
The available data is from healthy livestock animals (broilers, fattened pigs, dairy cattle and beef cattle), horses and dogs from Germany. Clinical isolates from livestock animals were excluded, nevertheless, all isolates from pets (dogs and horses) originated from clinical samples. As regards to humans, the focus was set on isolates from the healthy community and patients with nosocomial infections.
Based on phenotypic and genotypic characteristics, the isolated E. coli isolates were grouped into “subtypes”. Different approaches were used here in order to study how source attribution results change with the characteristics considered for defining the subtypes. In the following, we will use the terms type, subtype and ESBL-type interchangeably.
Origin of non-human isolates
Three studies collected faecal, dust and environmental samples from farms with healthy livestock animals. In the first study, seven conventional broiler and seven pig-fattening farms from different regions in Germany were sampled. Each farm was visited at the beginning, middle and end of the fattening period [7, 23, 24]. To avoid overrepresentation of isolates with specific characteristics which might occur through repeated sampling of the same group of birds, only one isolate of each unique type (as defined above) collected within the three visits of the same farm were included to simulate a cross-sectional study.
The other two studies were cross-sectional studies. One investigated 34 broiler, 48 fattening pig and 42 cattle farms in several regions of Germany, except Bavaria [25, 26] and the other 30 mixed dairy and beef cattle farms and 15 beef cattle farms in Bavaria [27].
All livestock samples were tested for the presence of ESBL-producing E. coli using selective MacConkey agar containing 1 mg/L cefotaxime [23] by study partners. Isolates from positive samples were sent to the National Reference Laboratory for Antimicrobial Resistance at the Federal Institute for Risk Assessment (BfR) in Berlin, Germany for confirmation and further molecular characterization [28, 29].
Regarding the dog and horse E. coli isolates, these originated from a collection of 600 Enterobacteriaceae strains from the Institute of Hygiene and Infectious Diseases of Animals (IHIT) in Giessen that were selected on MacConkey agar containing 1 mg/L cefotaxime [30]. The isolates were PCR screened for the presence of ESBL genes and genes contributing to (fluoro)quinolone resistance by the Institute of Medical Microbiology in Giessen. In the present study, we selected them based on the presence of ESBL genes [30].
Origin of human isolates
Isolates from one study of the general population and one study on nosocomial infections were included. In the first study, faeces from the healthy Bavarian community showing no symptoms of gastroenteritis were inoculated on MacConkey agar supplemented with 1 mg/L cefotaxime. Positive samples were considered as “community associated cases” colonized with ESBL-producing E. coli [10] and in the following we will refer to them as cases for short.
In the second study, a laboratory network operating throughout Germany collected ESBL-producing E. coli isolates from nosocomial infections (isolated in hospitals at least 48 hours after submission) on the basis of their phenotypic resistance to 3rd generation cephalosporins as identified by commercial selective media [9]. The isolates were then forwarded for further characterization to the National Reference Centre at the Robert Koch Institute (RKI) in Wernigerode, Germany [31, 32]. The nosocomial isolates included in our study were both tested for cefotaxime and ceftadizime. As we only used archived data, which was fully anonymous and did not contain any information relating their participants, no ethics committee was consulted in the present study.
Phenotypic and genotypic isolate characterization
All ESBL-producing E. coli were further analyzed in order to determine phenotypic and genotypic information related to their resistance-associated traits.
A large panel of antibiotics (amikacin, ampicillin, ceftazidime, cefotaxime, cefepime, chloramphenicol, ciprofloxacin, gentamicin, ertapenem and imipenem) was tested using minimum inhibitory concentration (MIC) or disk diffusion (DD) and evaluated according to EUCAST clinical breakpoints (www.eucast.org). For the present analysis, intermediate results were subsequently categorized as resistant.
In the present modelling approach, four different antibiotics were considered: gentamicin (S≤2mg/L | S≥17mm with 10μg disc; aminoglycoside), ertapenem (S≤0.5mg/L | DD not used; carbapenem), ciprofloxacin (S≤0.5mg/L | S≥22mm with 5μg disc; fluoroquinolone) and chloramphenicol (S≤8mg/L | S≥17mm with 30μg disc; phenicol). The selection of the antibiotics was a compromise of the test panels used in the routine antibiotic susceptibility testing of the involved institutions. In addition, this selection was made to assure a heterogeneous antibiotic resistance pattern.
The focus of this study was solely on ESBL genes (blaCTX-M, blaSHV, blaTEM), therefore, all the subtypes linked to AmpC β-lactamases were excluded from our dataset. ESBL genes were detected by PCR and when necessary confirmed by sequence analysis of the PCR product (as for blaSHV, blaTEM) [29]. Usually, the blaTEM gene was not sequenced if an ESBL gene (e.g. blaCTX-M and blaSHV) was detected in the same isolate; in this case the blaTEM gene is denominated as neg (as in “negative”). If the blaTEM gene was sequenced, either the gene denomination is given (e.g. blaTEM-52) or we denominated it as neg in case blaTEM was not identified. Furthermore, a PCR assay was performed to differentiate E. coli isolates into the four major phylogenetic groups (A, B1, B2, D) [28, 33].
Only isolates of an ESBL-type with a complete typing dataset were considered in this modelling approach. Table 1 summarizes the final number of ESBL-producing E. coli isolates per group and study.
The model
The model used in this approach is a variation [34] of the well-established Bayesian source attribution model originally developed for Salmonella [11, 35]. It compares the subtype distribution of a bacterial species in animals with the subtype distribution in humans achieved by the same subtyping methods. The model calculates the number of human cases that can be attributed to each of the potential sources (animal species and nosocomial infections) based on the assumption that there is a pathway from these sources into the general population.
The aim of the present modelling approach was to use microbial and molecular information on ESBL-producing E. coli to estimate the contribution of different potential sources to colonize or infect humans. The following equations were used to define the model:
(1)
(2)
(3)
where aj and qi are prior distributions.
The parameters used to feed the model are explained in Table 2 and in line with those already previously described by Hald and co-workers [11].
The amount of food consumed (Mj) and the prevalence (pij) had been introduced into the model by Hald et al. [11] to weight the model results. Due to the interest to use pets and humans as a potential source, the observation of high prevalence rates in all animal species and the expected low impact of the amount of food consumed, it was decided to use fixed values for these parameters (Mj = 1 and pij = 1).
The model was implemented by using the software R (version 4.0.3) [36] and JAGS which uses Markov Chain Monte Carlo methods to sample from the posterior distribution. In order to use JAGS with R we used the R package rjags [37]. Five Markov chains were used. JAGS features a process called adaptation (also called tuning). In this process JAGS tunes internal parameters of the Gibbs sampler in order to optimize its efficiency for sampling the posterior distribution. Users can control the extension of this tuning process via the n.adapt parameter. We set the n.adapt parameter to 1000. After that, we ran a burn-in period of 400 iterations followed by 500 sampling iterations. We applied thinning, which in our case means that of the 500 sampling values only every tenth was kept for each chain. The R code is available upon request.
Defining subtypes and excluding data
We generated three different subtype sets of the same set of isolates, which vary by the incorporation of the information: i) (Set 1) the ESBL gene pattern [blaCTX-M | blaSHV | blaTEM], ii) (Set 2) additionally to Set 1 the phylogenetic groups [A, B1, B2 or D] and iii) (Set 3) additionally to Set 2 the antibiotic resistance pattern (ARP) of four selected antibiotics [gentamicin, ertapenem, ciprofloxacin, and chloramphenicol] with ‘R’ coding for resistant and ‘S’ for susceptible.
The character string 1.neg.1.D.RRSR is an example for describing subtypes in Set 3. In the first position (positions are separated by a dot), “1” stands for the presence of blaCTX-M-1, “neg” in the second position stands for the absence of blaSHV and “1” in the third position stands for the presence of blaTEM-1. Since blaTEM-1 and blaTEM-135 don´t confer resistance to 3rd generation cephalosporins (non-ESBL), the subtypes only containing these blaTEM genes (e.g. neg.neg.1) were excluded from the data that went into the model calculations. In the fourth position, “D” reflects the phylogenetic group and in the last position, the resistance pattern consisting of resistance (R) to gentamicin, ertapenem, chloramphenicol and sensitivity (S) to ciprofloxacin is depicted.
The model requires that for each single bacterial type at least one human case. In a first step, the model looks whether at the same time at least one single bacterial type was isolated in one of the sources considered. We will call these types from now on matching types because human cases can in principle be matched to potential sources.
Consequently, there were two kinds of bacterial types in which this matching is not possible and had therefore to be excluded from the data used to run the model. First, bacterial types never found in a human case had to be excluded before running the model. For descriptive reasons, we denominate these types as other since they cannot explain any human cases. The second kind of type we excluded for running the model are types only found in human cases but never in a source. This means that we have no clue from which source these types originated. Therefore, we use the term unknown alone or in concert with “types” or “sources”, which is a shorthand for “types or isolates associated to unknown sources”.
We did not include other and unknown types in the data used to run the model, since the model cannot deal with them as they make it impossible to associate human cases to possible sources. Nevertheless, after running the model to complete the source attribution process, we manually included the number of cases related to unknown sources in our analysis, since these give an important information about the magnitude of attributed (matched) cases amongst the different sets. This means that the number of human cases allocated to unknown sources were not an outcome of the mathematical model, but were instead integrated on a later stage in the data analysis.
Modelling scenarios
Two different scenarios attributing isolates from several sources to isolates from healthy humans colonized with ESBL-producing E. coli were analyzed. The first scenario (Approach A) considers only the potential animal sources broilers, cattle, fattening pig, horses and dogs. The second scenario (Approach B) additionally includes as potential source humans infected with hospital acquired ESBL-producing E. coli. The latter approach stands for the potential spread of bacteria from hospitals (nosocomial infections) into the community.
Modifications of considered antibiotics (modified Approach B)
The third scenario reflects modifications of including each antibiotic on the model. The Approach B was modelled only using three antibiotics instead of four in order to check if there were major changes on the allocations when one of the antibiotics was absent. For this scenario, only Set B3 was used. All four antibiotic combinations, each with one of the considered antibiotics removed, were used in this scenario.
Results
Approaches considering animal sources
Set A1.
In Approach A, overall 723 isolates were included. Set A1 incorporated only information on the presence of the ESBL genes blaCTX-M, blaSHV and blaTEM, which led to the definition of 28 different subtypes. Among these, 11 were categorized as other types (with 38 isolates) and 6 were unknown types (with 22 isolates). The remaining 663 isolates allocated to 11 matching types (Table 3) were used for running the model.
After running the model, the majority of the community cases were attributed to the sources dog (35.0%) and cattle (24.3%). Pig was attributed to 13.1% of the cases followed by horse (12.1%). Broiler accounted for only 5.1% of the human cases. The remaining 10.3% of human cases could not be attributed to any of the considered sources (see Set 1 in Approach A of Fig 1 and S1 Table).
Approach A–sources considered are chicken, cattle, pig, horse and dog. Approach B–sources considered are chicken, cattle, pig, horse, dog and hospitalized patients (nosocomial infections). Set 1, considers only ESBL-types; Set 2, considers ESBL-types and phylogroup of E. coli; Set 3, considers ESBL-types, phylogroup of E. coli and resistance pattern of four antimicrobials.
Set A2.
Adding to the gene type (Set A1) the information about the E. coli phylogenetic group (A, B1, B2 or D), the Set A2 comprised 64 different subtypes. From these subtypes, 12 were unknown types (with 32 isolates) and 24 were considered other types (with 48 isolates). The remaining 667 isolates, which belonged to 28 matching types (Table 4), were used to run the model.
Using the ESBL gene and phylogenetic group information in the model (Set A2), dogs and pigs explained the majority of the human cases. Nevertheless, when compared to Set A1, the attributed proportion of human cases to the source dog (30.0% vs. 35%) and cattle (11.3% vs. 24.3%) decreased. In contrast, the attribution to the source pig (28.6% vs. 13.1%) increased considerably compared to Set A1. The attributed cases to chicken (3.8%) and to horse (11.3%) were quite similar to the previous model (Set A1). The number of cases attributed to an unknown source (15.0% vs. 10.3%) increased compared to Set A1 (see Set 2 of Approach A in Fig 1 and S1 Table).
Set A3.
Considering additionally the antibiotic resistance pattern (ARP), Set A3 contained 186 different subtypes. Among these, 62 were unknown types (with 119 isolates) and 87 were other types (with 244 isolates). The remaining 360 isolates, which consisted of 37 matching types (Table 5), were used to run the model.
Using ESBL genes, phylogenetic groups and antimicrobial resistance pattern (Set A3) most human cases could not be attributed to any of the animal sources. The proportion of attributed cases to the animal sources was 15.9% for pig, 3.3% for horse and 1.9% for dog, which was lower than in Set A2. In contrast, the cases attributed to cattle (19.2%) increased compared to the previous set. The attributed cases to chicken (4.2%) remained quite similar and the majority of the cases (55.6%) was attributed to an unknown source (see Set 3 of Approach A in Fig 1 and S1 and S2 Tables).
Approaches considering nosocomial cases as a source
Set B1.
For Approach B, 935 isolates were considered. Based on the information on the presence of the ESBL genes blaCTX-M, blaSHV and blaTEM (Set B1) 34 different subtypes were identified. Among these, 17 types (comprising 52 isolates) were categorized as other types and 2 types were unknown types (comprising 7 isolates). The remaining 876 isolates allocated to 15 matching types (Table 5) were used for running the model.
The introduction of nosocomial cases as new a “source” to the model redistributed a considerable number of human community cases to that new source (44.9%). Hence, the number of attributed cases to the other sources decreased. The model attributed less cases to dog (20.4%), cattle (9.3%), pig (8.8%), horse (8.8%), chicken (4.6%) and the unknown source (3.2%) compared to Set A1 (see Set 1 of Approach B in Fig 1 and S1 Table).
Set B2.
Adding the information on the phylogenetic group (A, B1, B2 or D) of E. coli to the type definition (Set B2) resulted in 79 different subtypes with 8 unknown types (with 12 isolates) and 39 other types (with 68 isolates). The remaining 32 matching types (Table 4) were used to run the model.
Including the phylogenetic group into the subtype definition did not affect too much the number of attributed cases in the model. Compared to Set B1, there was a slight increase in the cases attributed to the nosocomial cases (46.9% versus 44.9%). In Set B2, the attributed cases to dog (17.8%), horse (5.6%) and pig (4.6%) also showed a slight decrease compared to Set B1. Surprisingly, a minor increase in the attributed cases was seen in cattle (12.7%) and chicken (6.6%) compared to Set B1. Only 5.6% of the human cases could not be explained by any source considered (see Set 2 of Approach B in Fig 1 and S1 Table).
Set B3.
Adding the ARP to ESBL genes and phylogenetic group in the typing scheme (Set B3) lead to 238 subtypes, including 44 unknown types (with 76 isolates) and 139 other types (with 369 isolates). Overall, 55 matching types (Table 5) were used to run the model.
The introduction of the nosocomial source in Set B3 resulted in a significant reduction of the number of cases attributed to an unknown source (34.9%) compared to results in Set A3 (55.8%) (S1 and S2 Tables). When compared to Set B2, the proportion of unknown cases increased significantly (from 5.6% to 34.9%) and allocation to some sources decreased considerably, especially when it comes to dog (0.9%) and nosocomial (32.1%). Still, the nosocomial source accounted for almost as much cases as the sum of the animal sources (33.8%). Similar to Approach A, in Set B3, there were some increases in cases attributed to cattle (16.5%) and chicken (5.0%) compared to Set B2 (see Set 3 of Approach B in Fig 1 and S1 Table).
Modifications of considered antibiotics
As expected, when only three antimicrobials are considered in our modified Approach B in defining the subtypes, the total number of subtypes decreased compared to Set B3 (with 238 subtypes). The Set B3 without chloramphenicol has 175 subtypes, without ciprofloxacin 188, without ertapenem 227 and without gentamicin 181. When it comes to matching types, the number of subtypes was the same between Set B3 and the Set B3 without chloramphenicol and the Set B3 without ciprofloxacin with 55 matching types each. The Set B3 without ertapenem had 57 matching types and the Set B3 without gentamicin had 49 matching types. Even though the number of matching types was similar, the composition of subtypes changed compared to the full Set B3.
Set B3 without chloramphenicol.
Based on this dataset, nosocomial (38.7%), cattle and horse (12.0% each) were the sources that contributed the most to the human cases. Thus allocation of cases to nosocomial, horse and dogs increased whereas it decreased for cattle. Whereas still a considerable portion (20.7%) of cases weren´t attributed to any of the sources this was clearly lower than for the full dataset with 34.9% for the unknown source (see Fig 2 and S1 Table).
Modified Approach B–each set considers the ESBL-type, phylogenetic group of E. coli and resistance pattern of three antimicrobials. Sources included are chicken, cattle, pig, horse, dog and hospitalized patients (nosocomial infections). GenErtCip, gentamicin, ertapenem, ciprofloxacin (without chloramphenicol); GenErtChl, gentamicin, ertapenem, chloramphenicol (without ciprofloxacin); GenCipChl, gentamicin, ciprofloxacin and chloramphenicol (without ertapenem); ErtCipChl, ertapenem, ciprofloxacin, chloramphenicol (without gentamicin).
Set B3 without ciprofloxacin.
Most community cases were attributed to nosocomial (38.8%) and to cattle (32.7%). The other sources were of minor relevance. Not considering ciprofloxacin in categorizing the isolates thus led to a shift towards cases attributed to cattle. The unknown cases (17.8%) were also considerably lower than in the full B3 dataset (34.9%) (see Fig 2 and S1 Table).
Set B3 without ertapenem.
When not considering ertapenem, most of the matching types belonged to the phylogroup A (42.2%), followed by D (21.6%), B1 (20.3%) and B2 (16.0%). The changed categorization approach had an impact on the composition of the blaCTX-M genes in the matching types. The proportion of blaCTX-M-15 carrying isolates increased slightly whereas it decreased for isolates carrying blaCTX-M-1. Several sources (dog, cattle, pig, nosocomial) contributed to this change. Most matching types were susceptible to all three antimicrobials (55.0%), 20.1% of isolates were resistant to at least one antimicrobial, 16.0% were resistant to two of the three substances and 8.9% of the isolates were resistant to all three substances.
The number of cases attributed to nosocomial (35.5%) was the highest, followed by cattle (18.4%), pig (6.9%), chicken (3.7%), horse (3.2%) and dog (0.9%). The cases not attributed to any of the sources (unknown) represented 31.3% of cases in the model (see Fig 2 and S1 Table). Compared to the full B3 model, there were no major changes in the allocated proportions to the individual sources.
Set B3 without gentamicin.
When gentamicin was not considered in the grouping approach, most isolates belonged to the phylogroup A (43.1%), followed by B1 (23.3%), D (21.1%) and B2 (12.6%). Compared to the full Set B3, there was a shift towards isolates with phylogroup B1 in the matching types. Isolates from dogs, horses and nosocomial cases contributed to this. Compared to Set B3, the proportion of isolates carrying blaCTX-M-1 decreased. This was due to a higher frequency of isolates from horses, pigs and nosocomial cases. The majority of isolates in the matching types was susceptible to all three antimicrobials (51.8%). Isolates showing resistance to one antimicrobial were quite high (33.8%), resistance to two antimicrobials was present in 13.9% and full resistance was shown in 0.5% of the matching types.
Most of the cases were attributed to nosocomial subtypes (45.8%), followed by cattle (18.1%), chicken (5.1%), pig (3.2%), horse (1.9%) and dog (0.9%). A high portion of cases were attributed to a unknown source (25.0%) (see Fig 2 and S1 Table). Compared to the full B3 model, there was mainly a shift from unknown sources towards the nosocomial source.
Discussion
This study focused on assessing the possible role of major food producing animals, companion animals and humans that have acquired a nosocomial infection as a source for colonization of humans with ESBL producing E. coli. The sources included in our study are considered at the first stage of the food chain, at the point of reservoir, since the samples collected in our study were taken directly from the animals or their direct environment and our approach in based on microbial subtyping [38]. Routes of transmission including slaughter, processing and retail were not included and the point of exposure in which the transmission happened (e.g. by direct contact or consumption) is not covered in our study [39].
An important finding of this study is that all considered animal sources (broiler, cattle, pig, dog and horse) and nosocomial isolates share most subtypes of ESBL-positive bacteria with the general population. This suggests that animal sources play an important role in the transmission of resistant bacteria. At the same time, we found out that human cases were more often attributed to nosocomial isolates than to any of the animal sources, implying a tendency for the human-to-human transmission and contribution in the transmission of ESBL-producing bacteria.
Another interesting point is that the estimated magnitude of this contribution largely depends on the information that is considered when defining subtypes. The more information is incorporated in defining the subtypes, the fewer is the number of community isolates that can be explained by the sources considered. Based on ESBL genes only, nearly 90% of the human cases can be explained by any of the sources included, which reflects the focus on the transferable element, which might emerge in several different bacterial backgrounds. When ESBL gene, phylogenetic group and resistance pattern are combined, this proportion decreases considerably, leaving around half of the cases without any matching source. This suggests that animals as a whole do not provide a straightforward route for the resistant bacteria found in humans. Rather it might be that smaller pockets in the animal population harbour the resistant bacteria found in humans. In this case it would be interesting to study where these pockets lie. Alternatively, it might be that complex process of bacterial evolution happens when bacteria move from animals to humans. Humans might serve as a melting pot for bacteria bringing together bacteria with different mobile elements from different animal populations.
Role of animal sources in the model
In our study differences were observed in the contribution of the individual animal species as reservoir for community cases which was additionally depending on the information included in the subtype definition.
Pets (dogs and horses).
This observation is especially visible for dogs as a potential reservoir. When the ESBL type only (Set 1) or additionally the phylogenetic group (Set 2) are considered, dogs may contribute considerably to the human cases as a reservoir. This is linked with the fact that in our study sample, interestingly, isolates from dogs carried most frequently blaCTX-M-15, which was also the most common ESBL gene in our studied human populations. Contrary to our results, studies in the Netherlands [40, 41] identified blaCTX-M-1 as the most common blaCTX-M gene amongst dogs. In line with our data, a study in Mexico [41, 42] identified mainly blaCTX-M-15 genes in healthy dogs and considered this ESBL-type to be emerging in animal species. Furthermore, other European studies highlighted the increase of blaCTX-M-15 in companion animals (including horses) in the last years [43, 44]. This increase in blaCTX-M-15 in pets follows the same line as in the human population, in which blaCTX-M-15 has been the dominating ESBL gene in the last years in Europe and in most parts of the world [45]. Moreover, when it comes to horses, in our study, blaCTX-M-15 was the second most common ESBL gene identified, preceded by blaCTX-M-1 and followed by blaCTX-M-2. Even though that horses contributed less than dogs to the human cases, the high proportion of blaCTX-M-15 in horses is similar with the results from dogs and differs from the livestock isolates in our study, which rarely carried blaCTX-M-15 genes. A study in the UK [46] described a change in the prevalence of genes from the blaCTX-M group in horses over a ten-year period (from 2008 to 2017) in the same equine hospital. In 2008, none of the equine isolates carried the gene blaCTX-M-15, while in 2017 this ESBL gene was present in 60.8% of the analysed samples. Regarding blaCTX-M-2, this ESBL gene is also described in other studies [47–49] as an emerging ESBL gene in horses and mostly associated with clonal and nosocomial spread in equine clinics [47]. In our study, CTX-M-2-producing E. coli were present in both human populations but with a higher prevalence in the nosocomial isolates. Even though that the blaCTX-M-2 gene is not widely described in humans in Europe, other studies found it hospital facilities in Israel [50] and mostly in South America [51], where this gene from the CTX-M group is one of the most common amongst humans. Even though horses were not such a relevant source compared to dogs in our study, our findings shed some light on the similarities that exist between the human population and these two species in Germany and which ESBL genes may be most relevant when evaluating the potential role of pets in the transmission of ESBL-positive bacteria.
By adding the antimicrobial resistance pattern (Set 3) in our model, the similarity of isolates from humans and pets changes, especially in dogs which become almost irrelevant as a source to the human cases. This change is due to a considerable proportion of human cases being susceptible to all four considered antimicrobials while the subtype with the highest proportion in dogs was resistant to three antimicrobials. This may be explained since the dog isolates included in our study were clinical samples, while human cases were isolates from the healthy community. Thus, the high prevalence of multidrug-resistant E. coli in dogs may be due to previous or ongoing antibiotic therapy in the dogs included in our study [52].
Livestock (cattle, pigs and chicken).
When it comes to food producing animals, in our study, cattle and pigs may act as a reservoir for human exposure. The magnitude of this contribution depends on the range of possible sources included, as reflected in the difference between the outcome of Approach A, when only animal sources are considered and Approach B (discussed later). Most pig and cattle isolates carried blaCTX-M-1, which is the second most common ESBL gene in the human cases and similarly to the human cases, a high portion of cattle and pig isolates showed to be susceptible to the four tested antimicrobials. Nevertheless, when nosocomial isolates are considered in the model (Approach B), cattle and pig become less relevant as sources, highlighting the similarity between the nosocomial and the general community isolates in our study.
Chicken played in all approaches a minor role as reservoir for human cases. This can be explained by the predominance of ESBL genes, especially blaSHV-12 and blaTEM-52 in chicken isolates, but rarely in human cases. Our results are similar to recent studies by Mughini-Gras et al. and Dorado-Garcia et al. that showed that farm animals aren´t considered the most important sources when it comes to ESBL-producing E. coli and that human-to-human transmission should be considered as an important route instead [5, 40]. Yet, Mughini-Gras et al., Dorado-Garcia et al. and de Been et al., pointed out that farmers and farming communities may be more vulnerable to clonal transfer of ESBL genes due to having direct contact with livestock [5, 40, 53]. The same may apply for pets and pet owners.
Transmission of ESBL genes between humans and animals
Our results show that the incorporation of ESBL genes (blaCTX-M, blaSHV and blaTEM) and the phylogenetic group allows attribution of a broad number of human cases to animal sources. In contrast, the incorporation of phenotypic resistances made the source attribution less broad which was due to a growing dissimilarity between human and animal isolates in terms of phenotypic resistance patterns.
There is still lack of consensus about the direction of transmission of resistance determinants between pets and food animals to humans [54] and the mechanisms involved. In the last years, source attribution of resistance genes has been in the spotlight amongst the scientific community; nevertheless, most studies [54] have not been able to show evidence of the directionality of transmission of resistance genes between animals and humans. While some studies [55, 56] consider that direct contact with animals pose a risk to humans in the transmission of zoonotic bacteria such as ESBL-positive E.coli, other studies [57, 58] point out that the transmission pathway is most probable from humans to animals and not vice-versa. Better genomic data analysis (e.g. whole genome sequencing) combined with epidemiological details (e.g. spatiotemporal coherence of data in the populations; antibiotic usage in animals and humans) are needed to reconstruct the direction of transmission of resistance determinants between animals and people [54].
Importance of human to human transmission
When incorporating human nosocomial cases as a potential source to get colonized with ESBL-producing E. coli, we observed that more cases were accounted to this source than by any animal source. The similarity of isolates in the community and nosocomial cases may reflect the close link between hospitalized patients and the general population, in that the general population might bring resistant bacteria into the hospital or that resistant bacteria are “bread” in the hospitals from which they come into the general population highlighting the relevance of human-to-human transmission [59]. Another possibility is that there is a common source for both populations.
In our study, we observed that CTX-M-15-producing E. coli in combination with phylogroup B2 was the most common subtype amongst nosocomial isolates and the second most common among the general population, while being very rare amongst animal sources. The epidemic lineage E. coli B2 025:H4-ST131, have been observed worldwide and is associated with the blaCTX-M-15 gene [45]. Although the ways of transmission and the mechanisms for the successful spread of E. coli-ST131 are still unclear, this lineage has shown to be associated with high levels of antimicrobial use and high prevalence in healthcare facilities [60]. Even though we didn´t consider sequence type (ST) information in our study, a previous study by Pietsch et al. which used the same human isolates as our study, identified E. coli 025:H4-ST131 in 32.3% of the nosocomial isolates [9]. The horizontal spread, specifically through IncF plasmids plays a major role for dissemination of blaCTX-M-15 in ST131 strains [61].
The similarity of isolates from the nosocomial and general population is characterized by subtypes carrying ESBL genes in both human populations that were absent in animals, such as blaCTX-M-27, blaCTX-M-3, blaCTX-M-32 and blaCTX-M-55. Even though all of these ESBL genes have been already identified in animal sources in other studies [62, 63], most of them were seldom reported and mostly outside of Europe [64–67]. An exception is observed in blaCTX-M-55, which was first identified as the second most common ESBL gene in food producing animals in China and later, also, as the second most common ESBL gene in community-associated infections in Chinese hospitals [68].
When it comes to blaCTX-M-3 in the human population, a report by Bevan et al. showed that, nowadays, CTX-M-3-producing E.coli is not common in Europe, even though its first identification happened in a Polish hospital [45, 69]. After its first appearance, this ESBL-type has been occasionally identified in humans in Asia [70, 71] and in Africa [72].
Regarding blaCTX-M-27, this ESBL gene was identified in humans in various regions of the world [73]. In Japan, a study by Matsumura et al. analysed ESBL-producing isolates in hospitals from 2001 to 2012 and identified CTX-M-27-producing E. coli as being associated with an ST131 clone epidemic in the country [74]. Environmental studies have also identified blaCTX-M-27 in Switzerland in aquatic samples [75], in well water in China [76] and in wastewater in Tunisia [77]. In our study, the blaCTX-M-27 gene was more common in the general community than in the nosocomial isolates and it was the most frequent ESBL gene amongst the unique human subtypes.
Lastly, blaCTX-M-32 was first isolated in a Spanish hospital [78]; nevertheless, there are few reports [79] that show the presence of this ESBL gene in humans in Europe. Other reports that identified blaCTX-M-32 were located in the American continent [80, 81] and in Asia [82]. When identified, this ESBL gene showed to be very rare amongst the isolates and was typically found in healthcare facilities.
Variations due to subtype definition
By using different subtype definitions, we attempted to get a better understanding of the potential role of both, horizontal gene transfer and clonal spread of strains through different source attribution models.
On Set 1, when only the ESBL genes were considered, we assessed the transmission of genetic information related to the ESBL gene, which may be transferred through plasmids or other genetic elements. This showed that most of cases could be explained by the sources, while the number of cases not explainable (allocated to unknown sources) was the lowest. The similarity of ESBL genes between the sources and the cases in Set 1 may be related to the existence of epidemic plasmids that are successful in carrying resistant traits such as CTX-M-1 and CTX-M-15 in different hosts [83]. Epidemic plasmids are well adapted to different bacterial hosts, they have a tendency to acquire resistance genes and support the rapid dissemination of Enterobacteriaceae [84].
When including information present on the chromosome (e.g. phylogenetic group; Set 2) into the subtype definition, we noticed that this combination of ESBL genes with phylogenetic group lead to a lower number of attributed cases and an increase in cases with unknown sources. This reflects that the same ESBL genes are located in E. coli strains of different phylogenetic groups.
Modifications in the considered antibiotics
Resistances to fluoroquinolones, including ciprofloxacin, have been described in E. coli of human origin especially associated with in urinary infections [85, 86]. A study by Arslan et al., found out that patients who received ciprofloxacin as a treatment were twice more susceptible to carrying ESBL-producing E. coli compared to the ones who did not [87]. When it comes to livestock, specifically cattle, resistance to fluoroquinolones, including ciprofloxacin have been also widely described in Germany [88, 89]. In veterinary medicine, fluoroquinolones (predominantly enrofloxacin) are used as a treatment for varied conditions, which show cross-resistance to ciprofloxacin [88]. Due to their importance for treating severe infections, fluoroquinolones are considered by the WHO [90] and OIE [91] as “Critically Important Antimicrobials” (CIA) for human and veterinary medicine, respectively. Hence, due to the concomitant clinical use of fluoroquinolones in humans and animals, other studies [92, 93], have considered that the surge in resistances in fluoroquinolones in humans in the past decades have emerged mostly because of the use and misuse of these substances in animals.
Interestingly, in our study, we observed that when ciprofloxacin was not considered in the subtype definition, there was a higher number of all susceptible isolates in the general population compared to the other sets. It was also observed; when ciprofloxacin was present in the modified sets, there were higher number of isolates with resistance linked to this substance in the human cases than resistance to any of the others antimicrobials. That means that significant amounts of human isolates showed ciprofloxacin resistance.
At the same time, when ciprofloxacin was removed from the considered group of antibiotics, we observed that the number of matching isolates between cattle and humans increased considerably which elevated the number of attributed cases to cattle. This reflects that ciprofloxacin resistance plays only a smaller role among the phenotypes in our cattle data.
Other studies in Germany [94, 95] from the RESET-consortium, have demonstrated that calves and dairy cattle have a higher chance to carry resistances than beef cattle. Also, another study in Germany [96] showed that, in general, clinical isolates in cattle have a higher prevalence of antimicrobial resistances than non-clinical isolates. When it comes specifically to ciprofloxacin, Tenhagen et al., observed that the highest levels of resistance were seen in calves with enteritis, followed by non-clinical veal calves and young cattle at slaughter, cows with mastitis, milk from bulk tanks and lastly in beef cattle [96].
In our study, we have excluded clinical isolates from cattle, which can partly explain the low levels of isolates with resistance to ciprofloxacin. Nevertheless, by allowing only samples from healthy animals in our model, we obtained a more realistic scenario since the probability of sick animals getting in contact and/or being consumed by the general population is rather scarce.
We haven´t divided the cattle population into dairy and beef, leaving us with no conclusions about the role of these two production types.
Furthermore, the low number of ciprofloxacin-resistant cattle isolates in our study may be due to the composition of our data set or a positive consequence of the nationwide measures in Germany to minimize the use of antimicrobials in animals [97]. Between the years 2011 and 2017 a fell by 57% of the amount of antimicrobials supplied to veterinarians and a decrease of resistance to ciprofloxacin was registered in E. coli from calves [97]. Also, other countries in Europe, such as Denmark, have already restricted the use of fluoroquinolones in food animals and managed to reduce significantly the use of these substances in livestock in the past years [98]. Decreases in fluoroquinolones’ consumption in the primary healthcare sector has been also registered in Denmark for the last years and their use is recommended only in specific situations [99]. Nevertheless, resistances against fluoroquinolones are still observed in Denmark but mostly related to travelers that come from abroad; resistances in food production animals are also existent in Denmark but in lower amounts when compared to humans [99]. In Australia, fluoroquinolones are forbidden to be administered to animals [98] and levels of resistances in the community are lower than in other countries but still present [100].
Most studies [101] concerning the origin of ciprofloxacin-resistant E. coli in the general population are related to community and hospital-acquired urinary tract infections (UTI). This supports our results, given that most ciprofloxacin-resistant isolates in the community were attributed to nosocomial isolates. Other European studies reported that the strongest predictors for acquiring ciprofloxacin-resistant E. coli in patients suffering with urinary tract infections were previous hospitalization in the last 3–6 months, the age of 51–65 and previous antibiotic treatment during the last 12 months [102, 103]. Another study in the Netherlands found that community-acquired urinary tract infections due to ciprofloxacin-resistant E. coli were associated with a high intake of pork and chicken, but not with beef [104].
Herein, we have observed that isolates from human cases and cattle showed to be phenotypically different when it comes to resistance to ciprofloxacin. For that reason, we conclude that cattle is probably an overestimated source to the human population when ciprofloxacin is not considered in the subtype definition. More studies that evaluate and compare the level of resistances to fluoroquinolones between the cattle and human populations and their potential catalysts (e.g. farm/slaughterhouse management and hygiene, antimicrobial use and administration method) and transmission routes (e.g. direct contact, meat consumption, cross-contamination and environmental sources) are necessary.
Limitations
This study reflects the benefits of harmonized laboratory protocols for typing of isolates from different sources and across disciplines. The results achieved with the model can be just as good as the information included.
The dataset available has its limitations as regards representability of community cases for Germany, being Bavaria the only region covered, and the total number of isolates with complete information. Some animal species had more representation than others (e.g. dogs had a very low number of isolates compared to the other sources). Certainly, there is the risk to miss rare subtypes in a population due to the limited number of isolates included.
Furthermore, miss-classifications may have happened as resistance genes may have gone lost or phenotypic expression of resistance may have been triggered differently. Also, one laboratory analysed the human isolates while several others analysed the animal isolates; this may have led to different methods of antimicrobial susceptibility testing, even though the resistance breakpoints applied for interpretation of the data were the same.
Moreover, other sources were not included in this study, such as insects, that are considered a good indicator for environmental contamination [105], turkey, seafood, wildlife or vegetables, which may explain some of the unknown subtypes. As regards humans, there might be additional populations not included in the study e.g. travellers and people working in close contact with animals.
Our findings raised many questions, which cannot be answered by the present data. In further studies, it has to be investigated whether the prevalence of subtypes uniquely identified in humans continues to be the same, and if subtypes often found in animals continue to be rare in humans and vice versa. Furthermore, it needs to be addressed how important other sources are, such as vegetables, the environment and other animals, which could not be considered in this study. It remains unclear if and how often ESBL-producing bacteria are transmitted between animals and humans and how much they are subject to genetic modifications. There is also a lack of knowledge about the quantitative horizontal and vertical transmission rates of resistant bacteria and its mobile elements and the probability of a colonisation or an infection with such an organism. The existence and prevalence of epidemic lineages, such as E. coli 025:H4-ST131 amongst animals and humans also needs further investigation.
We could not make any inferences about the direction of transmission nor the complexity of the transmission cycles happening between the sources and humans, since we didn’t have additional epidemiological data and also not time series were available.
Conclusion
In our work, we showed the potential impact of some animal sources on the transmission of ESBL E. coli to humans. We additionally explored a possible path for human-to-human transmission of ESBL-producing bacteria. By using different subtype definitions, we attempted to get a better understanding of the role of both, horizontal gene transfer and clonal spread of strains through source attribution models. At the same time different subtypes definitions lead to varying attribution results which highlights the complexity of the transmission pathways of ESBL-positive bacteria. This complexity calls for a One Health approach to study further the development and transmission of antimicrobial resistance.
Supporting information
S1 Table. Number of human cases attributed to each source by the source attribution model and their proportions (%).
Approach A (Set A1, A2, A3) considers only animals (chicken, cattle, pig, horse and dog) as potential sources for the human cases; Approach B (Set B1, B2, B3) considers all the sources from Approach A and isolates from nosocomial infections. Set 1, considers only ESBL-types; Set 2, considers ESBL-types and phylogenetic group of E.coli; Set 3, considers ESBL-types, phylogenetic group of E.coli and resistance pattern of four antimicrobials. Set GenErtCip includes gentamicin, ertapenem, ciprofloxacin (without chloramphenicol); Set GenErtChl, includes gentamicin, ertapenem, chloramphenicol (without ciprofloxacin); Set GenCipChl, includes gentamicin, ciprofloxacin and chloramphenicol (without ertapenem); Set ErtCipChl, includes ertapenem, ciprofloxacin, chloramphenicol (without gentamicin). The isolates attributed to the unknown source represent the human cases which could not be attributed to any of the sources in the study.
https://doi.org/10.1371/journal.pone.0271317.s001
(DOCX)
S2 Table. Unknown types and respective number of cases that could not be attributed to any of the sources.
The nosocomial isolates represent the number of isolates present in this source on Set B3. The grey shaded subtypes represent the isolates that on Set A3 were unknown but then on Set B3 were attributed to the nosocomial source. These subtypes shifted from unknown to matching types when the nosocomial was introduced as a source.
https://doi.org/10.1371/journal.pone.0271317.s002
(DOCX)
S1 File. Excel workbook with unique subtypes in each set and respective number of isolates per source and human cases.
The data about subtypes present in both human cases and sources in this excel workbook were used in the source attribution model.
https://doi.org/10.1371/journal.pone.0271317.s003
(XLSX)
Acknowledgments
We thank all the people involved in the RESET Consortium and in the DiSCoVeR Joint Research Project from the One Health European Joint Program (OHEJP).
Special thanks also to Alexandra Irrgang for supporting us in the writing phase.
References
- 1. Knothe H, Shah P, Krcmery V, Antal M, Mitsuhashi S. Transferable resistance to cefotaxime, cefoxitin, cefamandole and cefuroxime in clinical isolates of Klebsiella pneumoniae and Serratia marcescens. Infection. 1983;11(6):315–7. Epub 1983/11/01. pmid:6321357.
- 2.
ECDC/EFSA/EMA. ECDC/EFSA/EMA first joint report on the integrated analysis of the consumption of antimicrobial agents and occurrence of antimicrobial resistance in bacteria from humans and food—producing animals. Stockholm/Parma/London: ECDC/EFSA/EMA, 2015.
- 3. Winokur PL, Canton R, Casellas JM, Legakis N. Variations in the prevalence of strains expressing an extended-spectrum beta-lactamase phenotype and characterization of isolates from Europe, the Americas, and the Western Pacific region. Clin Infect Dis. 2001;32 Suppl 2:S94–103. Epub 2001/04/26. pmid:11320450.
- 4. Arcilla MS, van Hattem JM, Haverkate MR, Bootsma MCJ, van Genderen PJJ, Goorhuis A, et al. Import and spread of extended-spectrum beta-lactamase-producing Enterobacteriaceae by international travellers (COMBAT study): a prospective, multicentre cohort study. Lancet Infect Dis. 2017;17(1):78–85. Epub 2016/10/19. pmid:27751772.
- 5. Mughini-Gras L, Dorado-Garcia A, van Duijkeren E, van den Bunt G, Dierikx CM, Bonten MJM, et al. Attributable sources of community-acquired carriage of Escherichia coli containing beta-lactam antibiotic resistance genes: a population-based modelling study. Lancet Planet Health. 2019;3(8):e357–e69. Epub 2019/08/24. pmid:31439317.
- 6. Fischer J, Rodriguez I, Baumann B, Guiral E, Beutin L, Schroeter A, et al. blaCTX-M-(1)(5)-carrying Escherichia coli and Salmonella isolates from livestock and food in Germany. J Antimicrob Chemother. 2014;69(11):2951–8. Epub 2014/07/31. pmid:25074857.
- 7. Von Salviati C, Friese A, Roschanski N, Laube H, Guerra B, Kasbohrer A, et al. Extended-spectrum beta-lactamases (ESBL)/AmpC beta-lactamases-producing Escherichia coli in German fattening pig farms: a longitudinal study. Berl Munch Tierarztl Wochenschr. 2014;127(9–10):412–9. Epub 2015/04/15. pmid:25868169.
- 8. Savin M, Bierbaum G, Hammerl JA, Heinemann C, Parcina M, Sib E, et al. ESKAPE Bacteria and Extended-Spectrum-beta-Lactamase-Producing Escherichia coli Isolated from Wastewater and Process Water from German Poultry Slaughterhouses. Appl Environ Microbiol. 2020;86(8). Epub 2020/02/09. pmid:32033950; PubMed Central PMCID: PMC7117925.
- 9. Pietsch M, Eller C, Wendt C, Holfelder M, Falgenhauer L, Fruth A, et al. Molecular characterisation of extended-spectrum beta-lactamase (ESBL)-producing Escherichia coli isolates from hospital and ambulatory patients in Germany. Vet Microbiol. 2017;200:130–7. Epub 2015/12/15. pmid:26654217.
- 10. Valenza G, Nickel S, Pfeifer Y, Eller C, Krupa E, Lehner-Reindl V, et al. Extended-spectrum-beta-lactamase-producing Escherichia coli as intestinal colonizers in the German community. Antimicrob Agents Chemother. 2014;58(2):1228–30. Epub 2013/12/04. pmid:24295972; PubMed Central PMCID: PMC3910888.
- 11. Hald T, Vose D, Wegener HC, Koupeev T. A Bayesian approach to quantify the contribution of animal-food sources to human salmonellosis. Risk Anal. 2004;24(1):255–69. Epub 2004/03/19. pmid:15028016.
- 12. Mughini-Gras L, van Pelt W, van der Voort M, Heck M, Friesema I, Franz E. Attribution of human infections with Shiga toxin-producing Escherichia coli (STEC) to livestock sources and identification of source-specific risk factors, The Netherlands (2010–2014). Zoonoses Public Health. 2018;65(1):e8–e22. Epub 2017/09/19. pmid:28921940.
- 13. Mullner P, Spencer SE, Wilson DJ, Jones G, Noble AD, Midwinter AC, et al. Assigning the source of human campylobacteriosis in New Zealand: a comparative genetic and epidemiological approach. Infect Genet Evol. 2009;9(6):1311–9. Epub 2009/09/26. pmid:19778636.
- 14. Jabin H, Correia Carreira G, Valentin L, Kasbohrer A. The role of parameterization in comparing source attribution models based on microbial subtyping for salmonellosis. Zoonoses Public Health. 2019;66(8):943–60. Epub 2019/09/04. pmid:31478354.
- 15. Lloyd DH, Page SW. Antimicrobial Stewardship in Veterinary Medicine. Microbiol Spectr. 2018;6(3). Epub 2018/06/20. pmid:29916349.
- 16. Patel SJ, Wellington M, Shah RM, Ferreira MJ. Antibiotic Stewardship in Food-producing Animals: Challenges, Progress, and Opportunities. Clin Ther. 2020;42(9):1649–58. Epub 2020/08/21. pmid:32819723; PubMed Central PMCID: PMC7434449.
- 17. Tang KL, Caffrey NP, Nobrega DB, Cork SC, Ronksley PE, Barkema HW, et al. Restricting the use of antibiotics in food-producing animals and its associations with antibiotic resistance in food-producing animals and human beings: a systematic review and meta-analysis. Lancet Planetary Health. 2017;1(8):E316–E27. WOS:000525863000013. pmid:29387833
- 18. Projahn M, Sachsenroeder J, Correia-Carreira G, Becker E, Martin A, Thomas C, et al. Impact of On-Farm Interventions against CTX-Resistant Escherichia coli on the Contamination of Carcasses before and during an Experimental Slaughter. Antibiotics-Basel. 2021;10(3). WOS:000633303700001. pmid:33668337
- 19. Robe C, Daehre K, Merle R, Friese A, Guenther S, Roesler U. Impact of different management measures on the colonization of broiler chickens with ESBL- and pAmpC- producing Escherichia coli in an experimental seeder-bird model. PLoS One. 2021;16(1):e0245224. Epub 2021/01/08. pmid:33411808; PubMed Central PMCID: PMC7790425.
- 20. Charani E, Holmes A. Antibiotic Stewardship-Twenty Years in the Making. Antibiotics-Basel. 2019;8(1). WOS:000464141600001. pmid:30678365
- 21. Torjesen I. Doctors and vets working together for antibiotic stewardship. Bmj-Brit Med J. 2018;362. WOS:000439739900001.
- 22. McEwen SA, Collignon PJ. Antimicrobial Resistance: a One Health Perspective. Microbiology Spectrum. 2018;6(2). WOS:000433084400005. pmid:29600770
- 23. Laube H, Friese A, von Salviati C, Guerra B, Kasbohrer A, Kreienbrock L, et al. Longitudinal monitoring of extended-spectrum-beta-lactamase/AmpC-producing Escherichia coli at German broiler chicken fattening farms. Appl Environ Microbiol. 2013;79(16):4815–20. Epub 2013/06/12. pmid:23747697; PubMed Central PMCID: PMC3754693.
- 24. Friese A, Schulz J, Laube H, von Salviati C, Hartung J, Roesler U. Faecal occurrence and emissions of livestock-associated methicillin-resistant Staphylococcus aureus (laMRSA) and ESbl/AmpC-producing E. coli from animal farms in Germany. Berl Munch Tierarztl Wochenschr. 2013;126(3–4):175–80. Epub 2013/04/02. pmid:23540202.
- 25. Hering J, Hille K, Fromke C, von Munchhausen C, Hartmann M, Schneider B, et al. Prevalence and potential risk factors for the occurrence of cefotaxime resistant Escherichia coli in German fattening pig farms—a cross-sectional study. Prev Vet Med. 2014;116(1–2):129–37. Epub 2014/07/22. pmid:25042772.
- 26. Hering J, Fromke C, von Munchhausen C, Hartmann M, Schneider B, Friese A, et al. Cefotaxime-resistant Escherichia coli in broiler farms-A cross-sectional investigation in Germany. Prev Vet Med. 2016;125:154–7. Epub 2016/01/20. pmid:26783199.
- 27. Schmid A, Hormansdorfer S, Messelhausser U, Kasbohrer A, Sauter-Louis C, Mansfeld R. Prevalence of extended-spectrum beta-lactamase-producing Escherichia coli on Bavarian dairy and beef cattle farms. Appl Environ Microbiol. 2013;79(9):3027–32. Epub 2013/03/05. pmid:23455336; PubMed Central PMCID: PMC3623142.
- 28. Doumith M, Day MJ, Hope R, Wain J, Woodford N. Improved Multiplex PCR Strategy for Rapid Assignment of the Four Major Escherichia coli Phylogenetic Groups. Journal of Clinical Microbiology. 2012;50(9):3108–10. WOS:000307941900047. pmid:22785193
- 29. Rodriguez I, Barownick W, Helmuth R, Mendoza MC, Rodicio MR, Schroeter A, et al. Extended-spectrum beta-lactamases and AmpC beta-lactamases in ceftiofur-resistant Salmonella enterica isolates from food and livestock obtained in Germany during 2003–07. J Antimicrob Chemoth. 2009;64(2):301–9. WOS:000267888800015. pmid:19474065
- 30. Schmiedel J, Falgenhauer L, Domann E, Bauerfeind R, Prenger-Berninghoff E, Imirzalioglu C, et al. Multiresistant extended-spectrum beta-lactamase-producing Enterobacteriaceae from humans, companion animals and horses in central Hesse, Germany. BMC Microbiol. 2014;14:187. Epub 2014/07/13. pmid:25014994; PubMed Central PMCID: PMC4105247.
- 31.
Pfeifer Y, Eller C, Leistner R, Valenza G, Nickel S, Guerra B, et al. ESBL-Bildner als Infektionserreger beim Menschen und die Frage nach dem zoonotischen Reservoir. Robert Koch-Institut, Infektionskrankheiten / Erreger; 2013.
- 32.
Pfeifer Y, Eller C. Aktuelle Daten und Trends zur β-Lactam-Resistenz bei gramnegativen Infektionserregern. Robert Koch-Institut, Infektionskrankheiten / Erreger; 2012.
- 33. Clermont O, Lavollay M, Vimont S, Deschamps C, Forestier C, Branger C, et al. The CTX-M-15-producing Escherichia coli diffusing clone belongs to a highly virulent B2 phylogenetic subgroup. J Antimicrob Chemother. 2008;61(5):1024–8. Epub 2008/03/13. pmid:18334490.
- 34. Sharp H, Valentin L, Fischer J, Guerra B, Appel B, Kasbohrer A. [Estimation of the transfer of ESBL-producing Escherichia coli to humans in Germany]. Berl Munch Tierarztl Wochenschr. 2014;127(11–12):464–77. Epub 2015/04/16. pmid:25872256.
- 35. Hald T, Lo Fo Wong DM, Aarestrup FM. The attribution of human infections with antimicrobial resistant Salmonella bacteria in Denmark to sources of animal origin. Foodborne Pathog Dis. 2007;4(3):313–26. Epub 2007/09/22. pmid:17883315.
- 36.
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2020.
- 37. Plummer M. rjags: Bayesian Graphical Models using MCMC. 2019.
- 38. Pires SM, Duarte AS, Hald T. Source Attribution and Risk Assessment of Antimicrobial Resistance. Microbiol Spectr. 2018;6(3). Epub 2018/06/20. pmid:29916343.
- 39. Pires SM, Evers EG, van Pelt W, Ayers T, Scallan E, Angulo FJ, et al. Attributing the human disease burden of foodborne infections to specific sources. Foodborne Pathog Dis. 2009;6(4):417–24. Epub 2009/05/07. pmid:19415971.
- 40. Dorado-Garcia A, Smid JH, van Pelt W, Bonten MJM, Fluit AC, van den Bunt G, et al. Molecular relatedness of ESBL/AmpC-producing Escherichia coli from humans, animals, food and the environment: a pooled analysis. J Antimicrob Chemother. 2018;73(2):339–47. Epub 2017/11/23. pmid:29165596.
- 41. van den Bunt G, Fluit AC, Spaninks MP, Timmerman AJ, Geurts Y, Kant A, et al. Faecal carriage, risk factors, acquisition and persistence of ESBL-producing Enterobacteriaceae in dogs and cats and co-carriage with humans belonging to the same household. J Antimicrob Chemother. 2020;75(2):342–50. Epub 2019/11/12. pmid:31711228; PubMed Central PMCID: PMC6966097.
- 42. Rocha-Gracia RC, Cortes-Cortes G, Lozano-Zarain P, Bello F, Martinez-Laguna Y, Torres C. Faecal Escherichia coli isolates from healthy dogs harbour CTX-M-15 and CMY-2 beta-lactamases. Vet J. 2015;203(3):315–9. Epub 2015/01/28. pmid:25624187.
- 43. Ewers C, Bethe A, Stamm I, Grobbel M, Kopp PA, Guerra B, et al. CTX-M-15-D-ST648 Escherichia coli from companion animals and horses: another pandemic clone combining multiresistance and extraintestinal virulence? J Antimicrob Chemother. 2014;69(5):1224–30. Epub 2014/01/09. pmid:24398338.
- 44. Ewers C, Grobbel M, Stamm I, Kopp PA, Diehl I, Semmler T, et al. Emergence of human pandemic O25:H4-ST131 CTX-M-15 extended-spectrum-beta-lactamase-producing Escherichia coli among companion animals. J Antimicrob Chemother. 2010;65(4):651–60. Epub 2010/02/02. pmid:20118165.
- 45. Bevan ER, Jones AM, Hawkey PM. Global epidemiology of CTX-M beta-lactamases: temporal and geographical shifts in genotype. J Antimicrob Chemother. 2017;72(8):2145–55. Epub 2017/05/26. pmid:28541467.
- 46. Isgren CM, Edwards T, Pinchbeck GL, Winward E, Adams ER, Norton P, et al. Emergence of carriage of CTX-M-15 in faecal Escherichia coli in horses at an equine hospital in the UK; increasing prevalence over a decade (2008–2017). BMC Vet Res. 2019;15(1):268. Epub 2019/07/31. pmid:31357996; PubMed Central PMCID: PMC6664770.
- 47. Apostolakos I, Franz E, van Hoek A, Florijn A, Veenman C, Sloet-van Oldruitenborgh-Oosterbaan MM, et al. Occurrence and molecular characteristics of ESBL/AmpC-producing Escherichia coli in faecal samples from horses in an equine clinic. J Antimicrob Chemother. 2017;72(7):1915–21. Epub 2017/03/24. pmid:28333298.
- 48. Smet A, Boyen F, Flahou B, Doublet B, Praud K, Martens A, et al. Emergence of CTX-M-2-producing Escherichia coli in diseased horses: evidence of genetic exchanges of bla(CTX-M-2) linked to ISCR1. J Antimicrob Chemother. 2012;67(5):1289–91. Epub 2012/02/14. pmid:22328640.
- 49. Sukmawinata E, Sato W, Mitoma S, Kanda T, Kusano K, Kambayashi Y, et al. Extended-spectrum beta-lactamase-producing Escherichia coli isolated from healthy Thoroughbred racehorses in Japan. J Equine Sci. 2019;30(3):47–53. Epub 2019/10/09. pmid:31592106; PubMed Central PMCID: PMC6773619.
- 50. Chmelnitsky I, Carmeli Y, Leavitt A, Schwaber MJ, Navon-Venezia S. CTX-M-2 and a new CTX-M-39 enzyme are the major extended-spectrum beta-lactamases in multiple Escherichia coli clones isolated in Tel Aviv, Israel. Antimicrob Agents Chemother. 2005;49(11):4745–50. Epub 2005/10/28. pmid:16251320; PubMed Central PMCID: PMC1280129.
- 51. Villegas MV, Kattan JN, Quinteros MG, Casellas JM. Prevalence of extended-spectrum beta-lactamases in South America. Clin Microbiol Infect. 2008;14 Suppl 1:154–8. Epub 2007/12/25. pmid:18154539.
- 52. Schmidt VM, Pinchbeck G, McIntyre KM, Nuttall T, McEwan N, Dawson S, et al. Routine antibiotic therapy in dogs increases the detection of antimicrobial-resistant faecal Escherichia coli. J Antimicrob Chemother. 2018;73(12):3305–16. Epub 2018/09/15. pmid:30215725.
- 53. de Been M, Lanza VF, de Toro M, Scharringa J, Dohmen W, Du Y, et al. Dissemination of cephalosporin resistance genes between Escherichia coli strains from farm animals and humans by specific plasmid lineages. PLoS Genet. 2014;10(12):e1004776. Epub 2014/12/19. pmid:25522320; PubMed Central PMCID: PMC4270446.
- 54. Muloi D, Ward MJ, Pedersen AB, Fevre EM, Woolhouse MEJ, van Bunnik BAD. Are Food Animals Responsible for Transfer of Antimicrobial-Resistant Escherichia coli or Their Resistance Determinants to Human Populations? A Systematic Review. Foodborne Pathogens and Disease. 2018;15(8):467–74. WOS:000442257800002. pmid:29708778
- 55. Garcia A, Fox JG, Besser TE. Zoonotic enterohemorrhagic Escherichia coli: A One Health perspective. ILAR J. 2010;51(3):221–32. Epub 2010/12/07. pmid:21131723.
- 56. Crump JA, Sulka AC, Langer AJ, Schaben C, Crielly AS, Gage R, et al. An outbreak of Escherichia coli O157:H7 infections among visitors to a dairy farm. N Engl J Med. 2002;347(8):555–60. Epub 2002/08/23. pmid:12192014.
- 57. Ljungquist O, Ljungquist D, Myrenas M, Ryden C, Finn M, Bengtsson B. Evidence of household transfer of ESBL-/pAmpC-producing Enterobacteriaceae between humans and dogs—a pilot study. Infect Ecol Epidemiol. 2016;6:31514. Epub 2016/06/23. pmid:27330043; PubMed Central PMCID: PMC4916256.
- 58. Gronthal T, Osterblad M, Eklund M, Jalava J, Nykasenoja S, Pekkanen K, et al. Sharing more than friendship—transmission of NDM-5 ST167 and CTX-M-9 ST69 Escherichia coli between dogs and humans in a family, Finland, 2015. Euro Surveill. 2018;23(27). Epub 2018/07/12. pmid:29991384; PubMed Central PMCID: PMC6152158.
- 59. Doi Y, Adams-Haduch JM, Peleg AY, D’Agata EMC. The role of horizontal gene transfer in the dissemination of extended-spectrum beta-lactamase–producing Escherichia coli and Klebsiella pneumoniae isolates in an endemic setting. Diagnostic Microbiology and Infectious Disease. 2012;74(1):34–8. https://doi.org/10.1016/j.diagmicrobio.2012.05.020. pmid:22722012
- 60. Banerjee R, Johnson JR. A new clone sweeps clean: the enigmatic emergence of Escherichia coli sequence type 131. Antimicrob Agents Chemother. 2014;58(9):4997–5004. Epub 2014/05/29. pmid:24867985; PubMed Central PMCID: PMC4135879.
- 61. Jamborova I, Johnston BD, Papousek I, Kachlikova K, Micenkova L, Clabots C, et al. Extensive Genetic Commonality among Wildlife, Wastewater, Community, and Nosocomial Isolates of Escherichia coli Sequence Type 131 (H30R1 and H30Rx Subclones) That Carry blaCTX-M-27 or blaCTX-M-15. Antimicrob Agents Chemother. 2018;62(10). Epub 2018/08/01. pmid:30061277; PubMed Central PMCID: PMC6153832.
- 62. Melo LC, Haenni M, Saras E, Duprilot M, Nicolas-Chanoine MH, Madec JY. Emergence of the C1-M27 cluster in ST131 Escherichia coli from companion animals in France. J Antimicrob Chemother. 2019;74(10):3111–3. Epub 2019/07/13. pmid:31299071.
- 63. Lupo A, Saras E, Madec JY, Haenni M. Emergence of blaCTX-M-55 associated with fosA, rmtB and mcr gene variants in Escherichia coli from various animal species in France. J Antimicrob Chemother. 2018;73(4):867–72. Epub 2018/01/18. pmid:29340602.
- 64. Cottell JL, Kanwar N, Castillo-Courtade L, Chalmers G, Scott HM, Norby B, et al. blaCTX-M-32 on an IncN plasmid in Escherichia coli from beef cattle in the United States. Antimicrob Agents Chemother. 2013;57(2):1096–7. Epub 2012/11/21. pmid:23165469; PubMed Central PMCID: PMC3553738.
- 65. Song J, Oh SS, Kim J, Park S, Shin J. Clinically Relevant Extended-Spectrum beta-Lactamase-Producing Escherichia coli Isolates From Food Animals in South Korea. Front Microbiol. 2020;11:604. Epub 2020/05/12. pmid:32390965; PubMed Central PMCID: PMC7188773.
- 66. Sun Y, Zeng Z, Chen S, Ma J, He L, Liu Y, et al. High prevalence of bla(CTX-M) extended-spectrum beta-lactamase genes in Escherichia coli isolates from pets and emergence of CTX-M-64 in China. Clin Microbiol Infec. 2010;16(9):1475–81. WOS:000281310900029. pmid:21681998
- 67. Zhang CZ, Ding XM, Lin XL, Sun RY, Lu YW, Cai RM, et al. The Emergence of Chromosomally Located bla CTX-M-55 in Salmonella From Foodborne Animals in China. Front Microbiol. 2019;10:1268. Epub 2019/06/25. pmid:31231347; PubMed Central PMCID: PMC6560199.
- 68. Zhang J, Zheng B, Zhao L, Wei Z, Ji J, Li L, et al. Nationwide high prevalence of CTX-M and an increase of CTX-M-55 in Escherichia coli isolated from patients with community-onset infections in Chinese county hospitals. BMC Infect Dis. 2014;14:659. Epub 2014/12/04. pmid:25466590; PubMed Central PMCID: PMC4265337.
- 69. Gniadkowski M, Schneider I, Palucha A, Jungwirth R, Mikiewicz B, Bauernfeind A. Cefotaxime-resistant Enterobacteriaceae isolates from a hospital in Warsaw, Poland: identification of a new CTX-M-3 cefotaxime-hydrolyzing beta-lactamase that is closely related to the CTX-M-1/MEN-1 enzyme. Antimicrob Agents Chemother. 1998;42(4):827–32. Epub 1998/04/29. PubMed Central PMCID: PMC105550. pmid:9559791
- 70. Yan JJ, Ko WC, Tsai SH, Wu HM, Jin YT, Wu JJ. Dissemination of CTX-M-3 and CMY-2 beta-lactamases among clinical isolates of Escherichia coli in southern Taiwan. J Clin Microbiol. 2000;38(12):4320–5. Epub 2000/12/02. pmid:11101558; PubMed Central PMCID: PMC87599.
- 71. Karim A, Poirel L, Nagarajan S, Nordmann P. Plasmid-mediated extended-spectrum beta-lactamase (CTX-M-3 like) from India and gene association with insertion sequence ISEcp1. FEMS Microbiol Lett. 2001;201(2):237–41. Epub 2001/07/27. pmid:11470367.
- 72. Hawkey PM, Jones AM. The changing epidemiology of resistance. J Antimicrob Chemother. 2009;64 Suppl 1:i3–10. Epub 2009/08/19. pmid:19675017.
- 73. Matsumura Y, Pitout JD, Gomi R, Matsuda T, Noguchi T, Yamamoto M, et al. Global Escherichia coli Sequence Type 131 Clade with blaCTX-M-27 Gene. Emerg Infect Dis. 2016;22(11):1900–7. Epub 2016/10/22. pmid:27767006; PubMed Central PMCID: PMC5088012.
- 74. Matsumura Y, Johnson JR, Yamamoto M, Nagao M, Tanaka M, Takakura S, et al. CTX-M-27- and CTX-M-14-producing, ciprofloxacin-resistant Escherichia coli of the H30 subclonal group within ST131 drive a Japanese regional ESBL epidemic. J Antimicrob Chemother. 2015;70(6):1639–49. Epub 2015/02/18. pmid:25687644.
- 75. Muller A, Stephan R, Nuesch-Inderbinen M. Distribution of virulence factors in ESBL-producing Escherichia coli isolated from the environment, livestock, food and humans. Sci Total Environ. 2016;541:667–72. Epub 2015/10/06. pmid:26437344.
- 76. Zhang H, Zhou Y, Guo S, Chang W. Prevalence and characteristics of extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae isolated from rural well water in Taian, China, 2014. Environ Sci Pollut Res Int. 2015;22(15):11488–92. Epub 2015/03/31. pmid:25821088.
- 77. Sghaier S, Abbassi MS, Pascual A, Serrano L, Diaz-De-Alba P, Said MB, et al. Extended-spectrum beta-lactamase-producing Enterobacteriaceae from animal origin and wastewater in Tunisia: first detection of O25b-B23-CTX-M-27-ST131 Escherichia coli and CTX-M-15/OXA-204-producing Citrobacter freundii from wastewater. J Glob Antimicrob Resist. 2019;17:189–94. Epub 2019/01/15. pmid:30639890.
- 78. Cartelle M, del Mar Tomas M, Molina F, Moure R, Villanueva R, Bou G. High-level resistance to ceftazidime conferred by a novel enzyme, CTX-M-32, derived from CTX-M-1 through a single Asp240-Gly substitution. Antimicrob Agents Chemother. 2004;48(6):2308–13. Epub 2004/05/25. pmid:15155242; PubMed Central PMCID: PMC415568.
- 79. Vinue L, Saenz Y, Martinez S, Somalo S, Moreno MA, Torres C, et al. Prevalence and diversity of extended-spectrum beta-lactamases in faecal Escherichia coli isolates from healthy humans in Spain. Clin Microbiol Infect. 2009;15(10):954–7. Epub 2009/06/13. pmid:19519849.
- 80. Araque M, Millan B, Mendoza E, Salas M. CTX-M-32 beta-lactamase-producing uropathogenic Escherichia coli isolated in Latin America. J Infect Dev Ctries. 2013;7(5):432–5. Epub 2013/05/15. pmid:23669435.
- 81. Hasan B, Laurell K, Rakib MM, Ahlstedt E, Hernandez J, Caceres M, et al. Fecal Carriage of Extended-Spectrum beta-Lactamases in Healthy Humans, Poultry, and Wild Birds in Leon, Nicaragua-A Shared Pool of blaCTX-M Genes and Possible Interspecies Clonal Spread of Extended-Spectrum beta-Lactamases-Producing Escherichia coli. Microb Drug Resist. 2016;22(8):682–7. Epub 2016/03/24. pmid:27007258.
- 82. Lee SG, Jeong SH, Lee H, Kim CK, Lee Y, Koh E, et al. Spread of CTX-M-type extended-spectrum beta-lactamases among bloodstream isolates of Escherichia coli and Klebsiella pneumoniae from a Korean hospital. Diagn Microbiol Infect Dis. 2009;63(1):76–80. Epub 2008/12/17. pmid:19073302.
- 83. Carattoli A. Plasmids and the spread of resistance. Int J Med Microbiol. 2013;303(6–7):298–304. Epub 2013/03/19. pmid:23499304.
- 84. Mathers AJ, Peirano G, Pitout JD. The role of epidemic resistance plasmids and international high-risk clones in the spread of multidrug-resistant Enterobacteriaceae. Clin Microbiol Rev. 2015;28(3):565–91. Epub 2015/05/01. pmid:25926236; PubMed Central PMCID: PMC4405625.
- 85. Aguiar JM, Chacon J, Canton R, Baquero F. The emergence of highly fluoroquinolone-resistant Escherichia coli in community-acquired urinary tract infections. J Antimicrob Chemother. 1992;29(3):349–50. Epub 1992/03/01. pmid:1592705
- 86. Goettsch W, van Pelt W, Nagelkerke N, Hendrix MG, Buiting AG, Petit PL, et al. Increasing resistance to fluoroquinolones in escherichia coli from urinary tract infections in the netherlands. J Antimicrob Chemother. 2000;46(2):223–8. Epub 2000/08/10. pmid:10933644.
- 87. Arslan H, Azap OK, Ergonul O, Timurkaynak F, Urinary Tract Infection Study G. Risk factors for ciprofloxacin resistance among Escherichia coli strains isolated from community-acquired urinary tract infections in Turkey. J Antimicrob Chemother. 2005;56(5):914–8. Epub 2005/09/22. pmid:16174685.
- 88. Schulz J, Kemper N, Hartung J, Janusch F, Mohring SAI, Hamscher G. Analysis of fluoroquinolones in dusts from intensive livestock farming and the co-occurrence of fluoroquinolone-resistant Escherichia coli. Sci Rep. 2019;9(1):5117. Epub 2019/03/28. pmid:30914675; PubMed Central PMCID: PMC6435704.
- 89. Kaesbohrer A, Schroeter A, Tenhagen BA, Alt K, Guerra B, Appel B. Emerging antimicrobial resistance in commensal Escherichia coli with public health relevance. Zoonoses Public Health. 2012;59 Suppl 2:158–65. Epub 2013/03/19. pmid:22958260.
- 90.
WHO. Critically important antimicrobials for human medicine, 6th revision Geneva: World Health Organization; 2019 [March 9, 2021]. Available from: https://apps.who.int/iris/bitstream/handle/10665/312266/9789241515528-eng.pdf?ua=1.
- 91.
OIE. OIE List of antimicrobial agents of veterinary importance Paris: World Organization for Animal Health; 2019 [09.03.2021]. Available from: https://www.oie.int/app/uploads/2021/03/a-oie-list-antimicrobials-july2019.pdf.
- 92. Notario R, Borda N, Gambande T, Bermejo J, Ponessa A, Toledo V. [Quinolone resistant Campylobacter jejuni strains isolated from humans and from poultry]. Medicina (B Aires). 2011;71(4):331–5. Epub 2011/09/07. pmid:21893445.
- 93. Agabou A, Lezzar N, Ouchenane Z, Khemissi S, Satta D, Sotto A, et al. Clonal relationship between human and avian ciprofloxacin-resistant Escherichia coli isolates in North-Eastern Algeria. Eur J Clin Microbiol Infect Dis. 2016;35(2):227–34. Epub 2015/12/05. pmid:26634353.
- 94. Hille K, Ruddat I, Schmid A, Hering J, Hartmann M, von Munchhausen C, et al. Cefotaxime-resistant E. coli in dairy and beef cattle farms-Joint analyses of two cross-sectional investigations in Germany. Prev Vet Med. 2017;142:39–45. Epub 2017/06/14. pmid:28606364.
- 95. Hille K, Felski M, Ruddat I, Woydt J, Schmid A, Friese A, et al. Association of farm-related factors with characteristics profiles of extended-spectrum beta-lactamase- / plasmid-mediated AmpC beta-lactamase-producing Escherichia coli isolates from German livestock farms. Vet Microbiol. 2018;223:93–9. Epub 2018/09/04. pmid:30173759.
- 96. Tenhagen BA, Kasbohrer A, Grobbel M, Hammerl J, Kaspar H. [Antimicrobial resistance in E. coli from different cattle populations in Germany]. Tierarztl Prax Ausg G Grosstiere Nutztiere. 2020;48(4):218–27. Epub 2020/08/22. pmid:32823326; PubMed Central PMCID: PMC7655263.
- 97.
BMEL. Bericht des Bundesministeriums für Ernährung und Landwirtschaft über die Evaluierung des Antibiotikaminimierungskonzepts der 16. AMG-Novelle Berlin: Bundesministerium für Ernährung und Landwirtschaft; 2019 [March 9, 2021]. Available from: https://www.bmel.de/SharedDocs/Downloads/DE/_Tiere/Tiergesundheit/Tierarzneimittel/16-AMG-Novelle-Bericht.pdf?__blob=publicationFile&v=2.
- 98.
Wegener HC. Antibiotic resistance—linking human and animal health. Improving food safety through a One Health approach. Washington DC: National Academies Press; 2012.
- 99.
DANMAP. Use of antimicrobial agents and occurrence of antimicrobial resistance in bacteria from food animals, food and humans in Denmark. Denmark: 2019.
- 100. Kennedy KJ, Roberts JL, Collignon PJ. Escherichia coli bacteraemia in Canberra: incidence and clinical features. Med J Aust. 2008;188(4):209–13. Epub 2008/02/19. pmid:18279126.
- 101. Fasugba O, Gardner A, Mitchell BG, Mnatzaganian G. Ciprofloxacin resistance in community- and hospital-acquired Escherichia coli urinary tract infections: a systematic review and meta-analysis of observational studies. BMC Infect Dis. 2015;15:545. Epub 2015/11/27. pmid:26607324; PubMed Central PMCID: PMC4660780.
- 102. Hitzenbichler F, Simon M, Holzmann T, Iberer M, Zimmermann M, Salzberger B, et al. Antibiotic resistance in E. coli isolates from patients with urinary tract infections presenting to the emergency department. Infection. 2018;46(3):325–31. Epub 2018/01/26. pmid:29368165.
- 103. Ny S, Edquist P, Dumpis U, Grondahl-Yli-Hannuksela K, Hermes J, Kling AM, et al. Antimicrobial resistance of Escherichia coli isolates from outpatient urinary tract infections in women in six European countries including Russia. J Glob Antimicrob Resist. 2019;17:25–34. Epub 2018/11/18. pmid:30447337.
- 104. Mulder M, Kiefte-de Jong JC, Goessens WH, de Visser H, Hofman A, Stricker BH, et al. Risk factors for resistance to ciprofloxacin in community-acquired urinary tract infections due to Escherichia coli in an elderly population. J Antimicrob Chemother. 2017;72(1):281–9. Epub 2016/09/23. pmid:27655855.
- 105. Wetzker W, Pfeifer Y, Wolke S, Haselbeck A, Leistner R, Kola A, et al. Extended-Spectrum Beta-Lactamase (ESBL)-Producing Escherichia coli Isolated from Flies in the Urban Center of Berlin, Germany. Int J Environ Res Public Health. 2019;16(9). Epub 2019/05/06. pmid:31052188; PubMed Central PMCID: PMC6539871.