Antimicrobial resistance in bacteria has emerged as a problem in both human and veterinary medicine. One of the currently most important resistance mechanisms in Enterbacteriaceae, which reduces the efficacy even of modern expanded-spectrum cephalosporins (except cephamycins and carbapenems) and monobactams is based on plasmid-mediated production of enzymes that inactivate these compounds by hydrolyzing their beta-lactam ring. Such resistance is encoded by an increasing number of different point-mutational variants, called extended spectrum beta-lactamases (ESBL), of classical broad-spectrum beta-lactamases (BSBL): most are derivates of TEM and SHV beta-lactamase families, whereas other groups, such as CTX-M, OXA, PER and VEB beta-lactamases have been described more recently [1]. The phenotypical difference between BSBLs and ESBLs is that the latter efficiently hydrolyze 3^rd- and 4^th-generation cephalosporins, in addition to penicillins and 1^st and 2^nd generation cephalosporins as the BSBLs are capable of. ESBLs are inhibited by clavulanic acid, sulbactam and tazobactam [2], a feature that is used (i) as a criterion for classification of β-lactamases and (ii) for diagnostic ESBL detection purposes. As a matter of growing concern, resistance caused by ESBLs is often associated with resistance to other classes of antibiotics like fluoroquinolones, aminoglycosides and trimethoprim-sulfamethoxazole [3], [4].

Since the first description of ESBL producing Enterobacteriaceae isolated from hospitalized humans [5], many nosocomial outbreaks have been reported. However, since a few years, there is an increase in the detection of ESBL producing strains in the community [6], [7]. More recently, some reports have alerted about the dissemination of ESBL producing E. coli in healthy cattle in several countries in Europe and USA [8], [9], [10] or in cattle derived food products like meat and raw milk [11], [12], [13]. Therefore, the impact of healthy farm animals as a reservoir for an input of ESBL producing E. coli in the food processing chain has to be assessed.

The aim of the present study was to assess the prevalence of ESBL producing Enterobacteriaceae in the Swiss cattle population younger than two years, and to assess risk factors for shedding ESBL producing Enterobacteriaceae based on a slaughterhouse monitoring approach that is aiming at achieving a geographically representative sample. Whilst for monitoring purposes a geographically representative set is easily achieved for samples taken on a farm level, where farms to be sampled can be determined in advance by applying a stratified randomization scheme, this is typically more difficult for samples taken at the slaughterhouse.

The sampling regimen resulted in a geographically representative sample population that was uniformly spread over the entire territory of Switzerland. Farms of origin of cattle shedding ESBL-producing Enterobacteriaceae are as well uniformly spread over the densely populated Swiss midlands as well as over the canton of Ticino in the south east of Switzerland. There were no cattle shedding ESBL-producing Enterobacteriaceae found originating from the canton of Valais in the south west of Switzerland.

ESBL-producing Enterobacteriaceae were detected in 48 (8.4%; 95% C.I. 6.3–11.1%) of the 571 tested animals. ESBL prevalence in different age classes, different production types and different animal movement activities in the farm of origin are summarized in Table 2. Results of the final multivariate logistic regression models applied for risk factor analysis are shown in Tables 3 and 4. Animals from within the age class ≤180 days were at a significantly higher risk of shedding ESBL-producing Enterobacteriaceae than animals from age classes 180 days to 1 year or 1 year to 2 years. Animals originating from farms with primary production type „dairy“ were at a 5.95 times greater risk of shedding ESBL-producing Enterobacteriaceae than animals originating from farms with primary production type „beef“. Finally, animals originating from farms with more than one animal movement per day per 100 animals were at a 2.37 times higher risk of shedding ESBL-producing Enterobacteriaceae than animals originating form farms with less than one animal movement per day100 animals.

In total, 48 isolates from 48 different animals showing the ESBL phenotype were further characterized. Species identification revealed 46 E. coli, one Enterobacter cloacae and one Citrobacter youngae (Figure 1). In view of β-lactam antibiotics, all 48 isolates were resistant to ampicillin (AM), the 1^st-generation cephalosporin, cephalothin (CF) and the 3^rd-generation cephalosporin, cefpodoxime (CPD) (Figure 1). Forty-five (93.8%) isolates were resistant to the 2^nd-generation cephalosporin cefuroxime (CXM); one (2.1%) isolate to cefoxitin (FOX), 28 (58.3%) isolates to cefotaxime (CTX), 2 (4.2%) isolates to ceftazidime (CAZ), and 2 (4.2%) isolates to the 4^th-generation cephalosporin cefepime (FEP). Seventeen strains (35.4%) showed resistance to amoxicillin-clavulanic acid (AMC). Besides β-lactam resistance, susceptibility to other classes of antibiotics was tested. Hereby, 26 (54.2%) isolates were resistant to gentamicin (GM), 27 (56.25%) to streptomycin (S), 21 (43.8%) to nalidixic acid (NA), 20 (41.7%) to ciprofloxacin (CIP), 36 (75.0%) to tetracycline (TE), 22 (45.8%) to chloramphenicol (C), and 30 (62.5%) to trimethoprim-sulfamethoxazole (SXT).

In this cross-sectional study, fecal samples of a geographically representative cattle subpopulation for Switzerland were sampled at slaughterhouse level to determine the occurrence of ESBL producing Enterobacteriaceae in the Swiss cattle population younger than 2 years, and to assess risk factors for shedding such organisms.

To assess the population prevalence of a pathogen or indicator organism on a national level, a geographically representative set of samples needs to be selected. Whilst this is easily achieved for samples taken on a farm level, where farms to be sampled can be determined in advance by applying a stratified randomization scheme, this is typically more difficult for samples taken at the slaughterhouse. This study demonstrated that the goal of a geographically representative survey could be achieved with a carefully planned sampling scheme involving the major cattle slaughterhouses and additional small abattoirs in areas disconnected from the main animal traffic.

To our knowledge, this is the first study on risk factors for shedding ESBL producing Enterobacteriaceae working with risk factors derived from animal movement data stored in a national AMD. Performing risk factor analysis with such data offered several advantages compared to risk factors derived form surveys, especially when the samples were collected at the abattoir and not on the farm. Getting information on animal movements and mortalities from 571 farms of origin of slaughtered animals via telephone or email surveys would be very time consuming. Furthermore, such information would be very imprecise compared to data from the Swiss AMD whose content is very accurate and complete.

The overall prevalence of cattle hosting ESBL producing Enterobacteriaceae found in this study (8.4%; 95% C.I. 6.3–11.1%) is slightly lower than found in previous, smaller scaled Swiss studies that reported prevalences of 13.7% [16] and 17.1% [10]. Recent studies from other countries reported a very low (0.2%) prevalence in cattle in Korea [17] or even zero prevalence in cattle in Tunisia [18]. Studies from European countries reported 35.4% (22.2–50.5% C.I. 95%) in North West England and North Wales [19] and 4.8% in France [20].

One of the significant risk factors for shedding ESBL producing Enterobacteriaceae obtained in the final multivariate model was age, with cattle over 6 months being at a considerably lower risk than calves (younger than 6 months) This is in accordance with a previous study of Geser et al. [16] who reported a overall prevalence of 13.7% in cattle as opposed to a prevalence of 25.3% among calves.

Animals from dairy farms were at a 5.95 times higher risk for hosting ESBL producing Enterobacteriaceae than animals from beef or fattening farms. This might be explained by differences between production types with respect to farming practices and antimicrobial compounds applied. In Switzerland, sales of cephalosporins have increased over the past years, especially the sales of 3^rd and 4^th generation cephalosporins for intramammary application [21]. Snow et al. [19] reported that in North West England and North Wales farms that had used 3^rd or 4^th generation cephalosporins in livestock during the previous 12 months were nearly 4 times more likely to host ESBL E. coli. On Swiss dairy farms, calves not in consideration for breeding are either fattened on their farm of birth or they are sold to fattening farms at a very young age. Those fattened on the dairy farms are primarily fed with milk. For economic reasons, milk that cannot be put on the market because of elevated cell counts or because of recent antimicrobial treatments is often fed to calves. Moreover, ESBL hosting Enterobacteriaceae present on dairy farms might be transmitted to calves by the fecal-oral route. Calves later to be fattened on fattening farms leave their dairy farms very few weeks after birth and are thus at a lower risk for acquiring ESBL producing Enterobacteriaceae than animals staying on the dairy farms until slaughter. In comparison to dairy farms, beta-lactam antibiotics, especially 3^rd and 4^th generation cephalosporins, do not represent the predominantly used antimicrobial compounds on beef and fattening farms. The fact that among 40 animals originating from the canton of Valais no ESBL producers were found, might also be explained by the farming type, as meat production is predominant over milk production in this area.

The number of animal movements per farm per day per 100 animals is a factor related to introduction of new stock, and it was associated with presence of ESBL Enterobacteriaceae. Animals from farms with a high number of animal incoming movements in relation to total farm size were at a 2.37 times higher risk of hosting ESBL producing Enterobacteriaceae than animals from farms with a lower level of animal traffic. This is in agreement with Snow et al. [19] who reported that operating a closed farm policy reduced the risk of the farm having ESBL E. coli compared to farms that were open and did not quarantine new cattle.

The confirmation of relatively high rates of ESBL producers in cattle and the high diversity among the isolates are worrisome and indicate an established reservoir, especially in dairy farms. A prudent use of antibiotics, especially of 3^rd and 4^th generation cephalosporins, restrictions in feeding milk of treated cows to calves and management improvements to facilitate the operation of closed herd policies could represent modes of action towards reduction of ESBL prevalence in cattle. Experimental studies would be needed to assess the effectiveness of such measures. Moreover, this study showed that with a carefully planned sampling scheme involving the major cattle slaughterhouses and – in addition – small abattoirs in areas disconnected from the main animal traffic, geographically representative surveys can be achieved by taking random samples at the abattoir.