Are There Effective Intervention Measures in Broiler Production against the ESBL/AmpC Producer Escherichia coli?

Extended-spectrum beta-lactamase (ESBL) and AmpC beta-lactamase (AmpC) producing Enterobacteriaceae occur frequently in livestock animals and the subsequent stages of the meat production chain and are therefore considered a risk for human health. Strict biosecurity measures and optimal farm management should reduce or even prevent poultry flock colonization at farm level. This review summarizes and evaluates published information on the effectiveness of specific intervention measures and farm management factors aiming to reduce the occurrence and spread of ESBL/AmpC producing or commensal or pathogenic E. coli in broiler chicken farms. In this systematic literature review, a total of 643 publications were analyzed, and 14 studies with significant outcome about the effectiveness of specific measures against E. coli and ESBL/AmpC producing E. coli in broiler chicken farms were found. Different feed additives seem to have an impact on the occurrence of those microorganisms. The measures ‘cleaning and disinfection’ and ‘competitive exclusion’ showed strong effects in prevention in some studies. In summary, some intervention measures showed potential to protect against or eliminate ESBL/AmpC-producing, commensal or pathogenic E. coli at farm level. Due to the high variability in the outcome of the studies, more specific, detailed investigations are needed to assess the potential of the individual intervention measures.


Introduction
The World Health Organization (WHO) has declared antibiotic resistance to be one of the biggest threats to global health, food-security, and development [1]. One of the most widespread resistance mechanisms of bacteria is the enzymatic inactivation of antibiotics by beta-lactamases. In that process, the beta-lactam ring of the antibiotics is hydrolyzed, and thus the irreversible blockade of the enzyme essential for the cell wall synthesis of these bacteria is prevented. The occurrence of ESBL-producers and AmpC-producers in Enterobacteriaceae like Escherichia coli has increased. They were frequently detected in livestock, companion, and wildlife animals [2,3], in humans, vegetables and broilers [4,5] but also in water, soil, air, or dust [6][7][8]. In Germany, high detection rates for ESBL/AmpC producing E. coli were identified in broiler houses and broilers [9][10][11]. The possible transmission to humans may occur via direct contact with animals or the environment, or via the consumption of contaminated food [5,12,13].
So-called biosecurity measures shall prevent potentially pathogenic microorganisms from entering a broiler farm and spreading within the farm. External measures shall protect the animals from pathogens in the environment, that enter the farm by vectors such as transport vehicles [21], humans [22][23][24], companion animals [25][26][27], wild animals [28][29][30], rodents [31,32], and water or feed [20,[33][34][35][36]. Internal biosecurity aims to prevent the spread of pathogens within the farm, for example through hygiene measures such as protective clothing, hand washing or cleaning, and disinfection of the pens [37][38][39]. Various management factors can also have an influence on the occurrence of diseases: ventilation and temperature, litter quality, stocking density, breed, and housing conditions or an all-in-all-out-system [40][41][42]. Finally, measures are discussed, which can improve animal health by helping to strengthen the immune system or prevent colonization of the gut by pathogenic bacteria [43][44][45].
Several intervention measures have been successfully applied to control Salmonella in livestock animals [46,47] but they were identified not to be effective against Campylobacter or ESBL/AmpC-producers, as demonstrated by the wide spread and increasing numbers of these microorganisms [48,49].
Therefore, this review aims at evaluating the effectiveness of measures to reduce commensal or pathogenic E. coli as well as ESBL/AmpC producing E. coli on broiler farms.

Literature Search
A total of 643 publications were analyzed (see Figure 1): Half of them resulted from the PubMed search by keywords, one quarter each from experts and by bidirectional literature search. In the first selection step we excluded publications for not meeting our requirements concerning language and peer reviewed publication of primary data. In the next step, we searched for studies investigating ESBL/AmpC-producing Escherichia coli in poultry or broiler or chicken or layer but did not exclude studies about commensal or pathogenic E. coli. With the third step we selected studies investigating the effect of intervention measures in broiler houses or broilers on farm level. Finally, we identified 14 publications. These 14 articles reported results from eight different countries: the seven European countries, Austria, Belgium, Denmark, Finland, Germany, Netherlands, and Poland (13 publications) and one non-European country: China (1).

Competitive Exclusion
Half of the identified publications (7/14) studied the effect of a probiotic-treatment on the presence of microorganisms in the gastrointestinal tract of broilers and their excretion. Almost all of them (6/7) investigated ESBL [52,55,56], or ESBL and AmpC producing E. coli [51,53,54], one study focused on the poultry pathogenic E. coli O20:K-:H8 and the human pathogenic E. coli O157:H7 [50].
Ceccarelli et al. [52] tested in a seeder-bird model the impact of treating one-day-old broilers with the commercial product Aviguard before challenging the seeder birds the next day with 10 6 CFU/mL E. coli strain E75.01/pE38.27 (ESBL). The average excretion until day 14 for the control group (no Aviguard) was 5.68 log 10 CFU/g feces compared with 1.17 log 10 CFU/g feces in the group where challenged and not challenged chicks received CE flora. They measured 2.22 log 10 CFU/g feces in the group in which only challenged chicks received CE flora and 3.86 log 10 CFU/g feces in the group in which only non-challenged chicks received CE flora, respectively. The resulting differences are 4.5 log 10 , 3.46 log 10 , and 1.82 log 10 CFU/g feces. Methner et al. [53] and Methner, Rösler [54] investigated the effect of Aviguard applied on day 1 of life to the colonization of seven or four ESBL-or AmpC-producing E. coli strains administered the next day in different doses (2 × 10 4 to 2 × 10 8 CFU/bird) in layer birds or broilers. The effect differs for the strains and breeds but can achieve a reduction of 4-5 log 10 CFU/g in layer birds, when exposed to seeder birds (1:5) infected with 1-2 × 10 5 CFU/bird. With Ross 308 broilers a reduction of 2.5-3.5 log 10 CFU/g was achieved in this seeder-bird-experiment. Even when challenging the birds with high doses of ESBL-or AmpC-producing E. coli strains (10 6 to 10 8 CFU/bird), a reduction of about 2 log 10 CFU/g was found in cecal content. Dame-Korevaar et al. published in 2020 [55,56] results of several seeder-bird experiments with prolonged supply of CE-cultures from day 0 until day 7 or even day 14 of life, twice a day. First [55], they investigated time until colonization, excretion and transmission for a CTX-M-1 E. coli (ESBL-producer) and application of the CE-products Aviguard or PoultryStar for fourteen days. Both products delayed the time until colonization (hazard ratio between 3.71 × 10 −3 and 3.11) in all birds, when the seeder-birds (1:1 seeder:contact) were challenged on day 5 with 10 2 CFU/mL, reduced the excretion of CTX-M-1 E. coli and the excretion of total E. coli (up to −1.60 log 10 CFU/g). The concentration in cecal content was slightly lower (up to −2.80 log 10 CFU/g) and a 1.5 to 3-fold reduction in transmission rate was observed. Later in 2020 Dame-Korevaar et al. [56] published experiments where they administered the CE-product Aviguard for seven days and challenged the seeder-birds on day 5 (seeder:contact 1:5) with 10 5 CFU/mL CTX-M-1 E. coli and none of the broilers were positive for the challenging ESBL-producing strain. In the control group, 93.5% of the birds were colonized. They also observed that contact birds were colonized later than the seeder-broilers (Time to event Ratio 3.53, 95% CI 3.14 to 3.93) and the microbiota composition was more diverse in CE-broilers than in control broilers at days 5 and 21.
For Broilact a minimal reduction of 2.0 and a maximum reduction of 5.5 log 10 CFU/g cecal contents of E. coli CK11ctx (ESBL producer) was reported by Nuotio et al. [51]. Moreover, a minimum reduction of 2.2 and a maximum reduction of 5.5 log 10 CFU/g in cecal contents of E. coli O20:K-:H8 was found by Hakkinen et al. [50]. They reported a minimum reduction of 0.9 and a maximum reduction of 6.3 log 10 CFU/g cecal contents for another strain, E. coli O157:H7, in broilers. In both studies, the probiotic product was applied to newly hatched broilers and the challenge was performed the next day.

Cleaning and Disinfection
In our literature search, four of the 14 relevant publications (28.6%) studied the effect of cleaning and disinfection in livestock houses. Luyckx et al. [57] evaluated four different cleaning protocols in broiler houses. They included dry cleaning, wet cleaning, and disinfection. Differences concerned, for example, the duration of the steps and the temperature of the water. From the swab samples taken before cleaning 97% were positive for E. coli, but after disinfection, only 7% were positive, which means 86% of the E. coli were eliminated by cleaning and disinfection. No differences were found between the different cleaning and disinfection protocols. A second study of Luyckx et al. [58] investigated different sampling methods and the significance of different microorganisms as hygiene indicators. They found a total reduction of E. coli by 1.6 log 10 CFU/625 cm 2 from swab samples in broiler houses by cleaning and disinfection. Cleaning, had a decreasing effect of 1.3 log 10 CFU/625 cm 2 , whereas disinfection resulted in a reduction of 0.3 log 10 CFU/625 cm 2 . Gradel et al. [59] investigated a temperature-humidity-time treatment in layer houses by using steam treatment with or without 30 ppm formaldehyde. The most effective method was the application of steam-treatment with 30 ppm formaldehyde for 24 h and it resulted in 100% elimination of naturally occurring E. coli in feces samples. The fourth article is about the use of slightly acidic electrolyzed water (SAEW) as an alternative disinfectant in layer houses, investigated by Hao et al. [60]. When the SAEW was sprayed with a high-pressure sprayer at a rate of 120 mL/m 2 for 5 min in layer houses, the isolation rates of E. coli showed a decrease of 16% compared to samples taken before disinfection.

Feed Additives
The impact of feed additives on the gut microbiota was tested in three studies. The microorganisms analyzed were ESBL/AmpC producing E. coli (1 publication) and E. coli (2), the livestock animals investigated were broilers in all cases.
Jamroz et al. [62] achieved a reduction of E. coli by 0.84 log 10 CFU/g intestinal digesta with a combination of carvacrol, cinnamaldehyde and capsaicin in a diet based on maize and a reduction of even 1.6 log 10 CFU/g intestinal digesta with this combination of plant extracts in a diet based on wheat and barley.
Goodarzi Boroojeni et al. [61] investigated the thermal processing of feed and the inclusion of organic acids in broiler diets. They supplemented the feed with three different levels of a commercial product containing formic acid and propionic acid. Furthermore, it was treated with four different types of thermal processes. Neither for the thermal treatment nor for the organic acid supplementation significant results were observed. However, the measured cell number of E. coli/Hafina/Shigella was the lowest for treatment of the feed with 0.75% acid in the crop, in the ileum and in the caecum.
Roth et al. [63] investigated in their study the effect of a commercial product containing organic acids (formic, acetic, and propionic acids) and, additionally, essential oils (cinnamaldehyde) on the prevalence of resistant E. coli. This feed additive based on formic (20%) and acetic (10%) and propionic acids (5%) as well as 2.5% cinnamaldehyde, was applied at a dosage of 2 kg/t of feed for 38 days to 480 broiler chickens. This product showed a about 1.84 log 10 CFU/g lower count of ESBL-E. coli in broiler cecum on day 38 in the group receiving feed-additives compared to the control group, but this effect was not significant.

Discussion
The occurrence of ESBL/AmpC producing E. coli in broiler meat is possessed as a risk for human health, as the transmission to humans is largely confirmed [64,65] and may lead to difficulties in the treatment of diseases [66]. To combat this danger, the broiler production chain is under observation and intervention measures are investigated to prevent the entry into the animals and the spread beneath the animals, the flocks, the pen, or even the environment.
In this review, we summarize the results from intervention studies aiming at reducing E. coli, commensal, pathogenic and ESBL-/AmpC-producing, in broiler and broiler chicken farms.
We used different approaches and search strategies and identified 14 articles (2.2%) as matches for our research question. This yield is quite in line with other systematic reviews [66][67][68]. Nevertheless, we have tried to find the balance between specific and general search parameters. For example, we have been simultaneously searching for all intervention methods with general key words like "risk" or "effect", instead of searching all known intervention measures like "disinfection" or "feed additives separately. This led to a greater yield but also to unspecific results. However, it allowed us to find measures that were unknown to us. Furthermore, we did not restrict the literature search to title and abstract, which unfortunately also led to results in the references of the publications. We therefore decided to complement the database search with the references and studies citing relevant articles, starting with already identified publications. This approach, the so-called bidirectional search, promises numerous results in finding relevant literature [69]. Nevertheless, 14 articles is a small number and only three intervention measures have been investigated. The effectiveness of control measures against ESBL-AmpC-producing E. coli in broiler fattening farms seems to be an underreported topic. However, we expected the intervention measures to aim on combating these microorganisms in the broiler farm like cleaning and disinfection or in the broiler itself, as competitive exclusion and feed additives do. The largest number of publications were on competitive exclusion (50%), a method known since 1973, were Nurmi and Rantala [70] proved the effectiveness for Salmonella. In the last decades it has been further developed and different microorgan-isms and commercial products have been tested [50,71,72]. The principle is as complex as the microbiota of broilers. Nevertheless, it seems, as if there is a significant effect, when applying CE-cultures preventively [50][51][52][53][54][55][56]. Although various animal trials were conducted with various CE-products and challenging strains of E. coli, they all discovered a maximum reduction of around 5 log 10 CFU/g cecal content or even complete prevention of colonization. Competitive exclusion seems to be a potential intervention method to suppress colonization of pathogenic or ESBL-or AmpC-producing E. coli, as applied, in broiler chickens if used very early in life and depending on the strains and amounts used for protection and challenge.
Cleaning and disinfection is the standard approach to eliminating pathogens from hands, boots, equipment, transport vehicles and the stables between the production cycles. It has been known to be an important part of the reduction strategy against pathogenic bacteria for a long time [73]. Unfortunately, the cleaning protocols and the detergents and disinfection substances are variable and lead to various results, possibly, to transmission to subsequent flocks, as Daehre et al. have proven recently [11]. Therefore, the general effect for cleaning and disinfection differed from small (16% and 1.6 log 10 reduction) to big (86% or even 100% elimination). In addition to the cleaning and disinfection substance, the number of substances used and the method as well as the temperature of the water, further factors might have an impact on the effectiveness as exposure time or handling differences due to different persons involved [39,74,75]. Not to forget that the long-term use of disinfectants can have opposite effects, such as the development of cross-resistances or the increased formation of biofilm [76,77].
Feed additives like essential oils, organic acids but also enzymes or phytochemicals can improve feed intake or nutrient digestibility [78][79][80][81]. Due to the variety of the active ingredients, the same great variability in the effect is not surprising. However, if one considers the results of the studies found for this review, their efficacy seems to be rather small, regardless of whether the essential oils and organic acids were applied alone or in combination [61][62][63]. In contrast, a more recent study found a synergistic effect for organic acids and essential oils in the reduction of E. coli in broilers [82]. It seems as if feed additives provide again a broad range of substances and application methods and therefore their effectiveness for reducing ESBL/AmpC-producers in the broiler production is still under investigation.
In summary, we have found few and very different data for the effectiveness of intervention measures against ESBL/AmpC-producing, commensal or pathogenic E. coli in the broiler production. On the one hand other authors concluded that there is no such thing as a totally biosecure farm [83], and there is no route, vehicle, or vector which can be clearly identified as a unique target for intervention [84]. Moreover, Dame-Korevaar et al. [49] found limited to no causal evidence of transmission along the transmission routes investigated. On the other hand, the issue of biosecurity is omnipresent [85]. In this respect, the validation of biosecurity protocols needs to be strengthened, otherwise the farmer may not be using time and money effectively [86]. Fortunately, some of the intervention measures can have a great impact, as can be seen from the results presented. Additional publications present interesting results for other livestock animals, microorganisms for intervention measures such as a prolonged vacancy period [87], different litter materials [88,89] or acidified litter [90]. The results are also variable, ranging from no effect [87,89] to a reduction of 5 log 10 CFU/g [88]. These studies should be carried out for poultry and ESBL/AmpC-producing E. coli as well. Thus, further and more detailed studies are needed to investigate these practices and other risk factors.

Literature Search
A systematic literature review was conducted. The literature search was performed by using online databases and additional bidirectional searching for references and publications citing relevant articles. First, we searched the online database PubMed using terms reflecting possible risk factors like biosafety, internal and external biosecurity, and management factors with impact on the occurrence of pathogens as well as non-pathogenic microorganisms in livestock farming ( Table 2). The transmission routes into a farm and from a farm into a slaughterhouse were considered as well. Therefore, these keywords were included. The starting point for the bidirectional search (bidirectional citation searching to completion, BCSC-Method; [69] were articles already identified by experts as potentially relevant (project proposal of EsRAM (German acronym for "Development of interventions to reduce antimicrobial resistance in pathogens in the broiler production")). Their references and the publications citing them were investigated.

Data Selection
Publications in international journals were included in the evaluation. We did not exclude any study due to its origin country or publication date.

Inclusion or Exclusion of Publications
Within the first selection step we excluded articles in other language than English or German and selected publications with primary data, peer-reviewed from international journals. In the next step these studies were analyzed for the livestock animal and the investigated microorganism. The animal production groups broiler, layer, chicken, turkey and duck, poultry in general, and the microorganisms ESBL/AmpC producing E. coli but additionally commensal/non-resistant E. coli in general were included. All other sources of samples, e.g., cattle, pigs, humans, rodents, environment or meat, and investigated bacteria (Enterobacteriaceae, Campylobacter, Salmonella) but also viruses and combinations like total mesophilic aerobic bacteria as well as other resistances were excluded. In the third step the articles were analyzed by the aim of the study. Only articles about specific measures to reduce bacterial occurrence in livestock animals or livestock farms were kept.

Calculating the Effectiveness
The published results for the reduction of relevant microorganisms were converted, if necessary, into a log 10 -reduction. For example: 2.04 × 10 7 = log 10 (20400000) = 7.3 log 10 .

Definition of Intervention Measures
General biosecurity measures are expected to prevent potentially pathogenic microorganisms entering a broiler farm and spreading within the farm. Those measures are divided into external and internal biosecurity measures. External actions are supposed to protect the animals from pathogens in the environment, entering the farm by vectors like transport vehicles, farm workers, visitors, equipment, feed, water, rodents, or pests. The internal biosecurity should prevent the spread of pathogens within the farm [91,92]. These extensive biosecurity methods and recommendations are complex and their effectiveness on the whole is hard to evaluate.
For our review we looked for focused intervention methods which were applied once or at least for a short time to the chickens or the chickens houses/pens and are expected to lead to measurable effects on the prevalence of ESBL/AmpC producing E. coli or approximately E. coli as closest relatives within the species.