Prevalence of extended-spectrum and AmpC β-lactamase-producing Escherichia coli in young calves on Dutch dairy farms

In young calves on dairy farms the animal prevalence of extended-spectrum and AmpC β-lactamase-producing Escherichia coli (ESBL/AmpC-EC) is significantly higher compared with the animal prevalence in young stock and dairy cows. Hitherto it was unknown at what age antimicrobial resistant bacteria appear for the first time in the gut of calves on dairy farms, and how long these infections persist. The aim of this study was to examine the prevalence of ESBL/AmpC-EC, the number of excreted ESBL/AmpC-EC (in cfu/g of feces), as well as the ESBL/AmpC genotypes in young dairy calves (0–21 d of age) and the variation of these parameters between calves of different ages. Next to this, the course of shedding ESBL/AmpC-EC during the first year in dairy calves was studied. In a cross-sectional study, fecal samples from 748 calves, from 0 to 88 d of age, on 188 Dutch dairy farms were collected. The prevalence of calves testing positive for ESBL/ AmpC-EC in a phenotypic assay was determined for different age categories (per 2 d of age). Positive samples were subjected to a semiquantitative test to determine the numbers of ESBL/AmpC-EC per gram of feces and for a selection of ESBL/AmpC-EC isolates the ESBL/AmpC genotype was determined. Ten of the 188 farms were selected for a longitudinal study based on the presence of at least 1 female calf with ESBL/ Amp-EC in the cross-sectional study. These farms were additionally visited 3 times with a 4-mo interval. All calves that were sampled in the cross-sectional study were, if still present, resampled during the follow-up visits. Results show that from the day of birth ESBL/ AmpC-EC can be present in the gut of calves. The phenotypic prevalence of ESBL/AmpC-EC was 33.3% in 0-to 21-d-old calves and 28.4% in 22-to 88-d-old calves. The prevalence of ESBL/AmpC-EC positive calves varied per age category among calves up to 21 d of age: significant increases and decreases at an early age were shown. Results of the longitudinal study show that after 4, 8, and 12 mo the prevalence of ESBL/ AmpC-EC positive calves dropped to 3.8% (2/53), 5.8% (3/52), and 2.0% (1/49), respectively. This indicates that early gut colonization in young calves with ESBL/ AmpC-EC is transient and does not lead to long-term shedding of these bacteria.


INTRODUCTION
The emergence of extended-spectrum and AmpC β-lactamase-producing Escherichia coli (ESBL/ AmpC-EC) in the last decades has been of great concern for both human and animal health organizations (WHO, 2011).These bacteria pose the ability to break down β-lactam antibiotics, resulting in resistance to third-and fourth-generation extended-spectrum cephalosporins and penicillins, which are classified by the World Health Organization as the highest priority critically important antimicrobials for human medicine.In fact, for serious Salmonella and E. coli infections these antimicrobials are one of few available treatment options in humans (WHO, 2018).The ESBL/AmpC-EC are present in humans, the environment, and (foodproducing) animals, including cattle (Korzeniewska et al., 2013;Wu et al., 2013;EFSA, 2014;Laube et al., 2014;von Salviati et al., 2015;Gonggrijp et al., 2016;Hordijk et al., 2019).
In 2013, the animal prevalence of ESBL/AmpC-EC among dairy cattle, with a minimum shedding level of 10 3 cfu/g of feces, in different age categories in dairy herds was established and showed a significant higher animal prevalence in calves (19.3%) compared with young stock (0.9%) and dairy cows (0.8%) (Heuvelink et al., 2019).Moreover, the level of shedding ESBL/ AmpC-EC in feces (cfu/g) was also higher in calves compared with older cattle.Earlier studies described the prevalence and dynamics of ESBL/AmpC-EC in dairy and veal calves.Results of these studies show that the prevalence of calves shedding ESBL/AmpC-EC is high during the first weeks of their lives and decreases with increasing age.The colonization of the gut of young calves with antimicrobial resistant E. coli appears to be transient (Hoyle et al., 2004;Hordijk et al., 2013a;Bastard et al., 2021;Homeier-Bachmann et al., 2022).The high prevalence of calves shedding ESBL/AmpC-EC indicates that calves could play an important role in the dynamics of ESBL/AmpC-EC on dairy farms.However, only limited information is available about the colonization of these antibiotic resistant bacteria in the gut of young calves at a dayto-day level during the first weeks of their lives, raising the questions about the age at which colonization occurs and how long this colonization persists in the gut.
The aim of this study was to examine the prevalence of ESBL/AmpC-EC, the number of excreted ESBL/ AmpC-EC (in cfu/g of feces), as well as the ESBL/ AmpC genotypes in young dairy calves (0-21 d of age) and the variation of these parameters between calves of different ages.Next to this, the course of shedding ESBL/AmpC-EC during the first year in dairy calves was studied.

MATERIALS AND METHODS
From October until December 2013, a crosssectional study was performed that was followed by a longitudinal study on a subset of the herds in the cross-sectional study.The cross-sectional study was part of a larger study that was described earlier by Heuvelink et al. (2019).Calves sampled in this study were housed on Dutch nonorganic dairy farms.During sampling the calves were restrained as little as possible by the farmer to limit the amount of possible discomfort.

Cross-Sectional Study
In the cross-sectional study, 1,000 randomly selected nonorganic dairy farmers were asked by letter to submit fecal samples from all calves in the age category 0 to 21 d present in their herd.Of these, 188 farmers submitted 748 fecal samples of calves, who were 0 to 88 d old.Samples were collected through rectal palpation by the private veterinary practitioner.
The samples were transported to the laboratory at refrigerator temperature (2-8°C) where bacteriological examination started within 24 h after collection of the samples.

Longitudinal Study
Ten herds were randomly selected from 33 herds with an ESBL/AmpC-EC positive fecal sample from at least 1 female calf and with at least 5 female calves being sampled in the cross-sectional study, and included in the longitudinal study.These 10 herds were visited by the private veterinary practitioner every 4 mo for a period of 1 yr, leading to 3 samplings in addition to the first sampling.Calves present on these 10 herds were sampled during the cross-sectional study (first sampling) and resampled 4 (second sampling), 8 (third sampling), and 12 mo later (fourth sampling) and are referred to as resampled calves.When the number of resampled calves dropped below 5 at the second or third sampling of the longitudinal study, due to early removal of calves from the farm, new calves between 0 and 21 d of age were sampled.These calves are referred to as additional calves, and they were also sampled at the subsequent samplings.

Laboratory Diagnostics
In the cross-sectional as well as the longitudinal study, all fecal samples were individually examined.On all of these samples a phenotypic assay was performed, followed by a semiquantitative assay on samples testing positive in the phenotypic assay and with sufficient material left.After the semiquantitative assay, a random selection of the ESBL/AmpC-EC positive samples was generated by statistical software (STATA 15.0;StataCorp, 2017) and subsequently screened for ESBL/pAmpC genotypes (Figure 1).The phenotypic assay, semiquantitative assay, and genotypic assay are described in brief below and in detail by Heuvelink et al. (2019).
Phenotypic Assay: Isolation of ESBL/AmpC-EC.Fecal samples were streaked either directly or after selective enrichment in Luria-Bertani broth (Becton Dickinson) supplemented with 1 mg/L cefotaxime (Sigma-Aldrich), onto MacConkey agar (Oxoid Ltd.) supplemented with 1 mg/L cefotaxime.After overnight aerobic incubation at 37°C, presumptive E. coli colonies were selected for confirmation by using the MALDI-Biotyper (Bruker Daltonics GmbH).Combination disc diffusion tests using cefotaxime and ceftazidime with and without clavulanic acid (Becton Dickinson) were used to examine the confirmed E. coli isolates for ESBL or AmpC production, according to Clinical and Labo- To detect AmpC phenotypes, a cefoxitin disc (30 μg, Becton Dickinson) was included in the test.Escherichia coli isolates from the longitudinal study were not examined for differentiation between ESBL and AmpC production using the disc diffusion test, due to limited resources.
Quantitative Assay: Quantification of ESBL/ AmpC-EC.All ESBL/AmpC-EC positive samples were subjected to a (semi)quantitative determination of ESBL/AmpC-EC and total E. coli, using the trackdilution technique (Jett et al., 1997).The number of ESBL/AmpC-EC and total E. coli were calculated based on the highest 10-fold dilution with typical E. coli colonies.To distinguish between calves shedding ESBL/AmpC-EC due to amplification of these bacteria during passage through the gut, calves with ESBL/ AmpC-EC colonization in the gut and calves shedding extremely high numbers of ESBL/AmpC-EC due to colonization in the gut, the samples were categorized as containing low (<10 3 cfu/g feces), moderate (10 3 to 10 5 cfu/g feces), and high (≥10 6 cfu/g feces) numbers of ESBL/AmpC-EC and E. coli isolates (Chase-Topping et al., 2008).
Genotypic Assay: Identification of ESBL/ AmpC Genes.Per ESBL/AmpC-pos calf, 1 to 3 isolates were screened for the presence of different β-lactamase genes and plasmid-mediated AmpC genes (pAmpC) by using the E. coli Genotyping Combined microarray (Alere Technologies GmbH).Further identification of the ESBL/AmpC gene types was performed by PCR amplification followed by sequencing using different primers, as described previously (Gonggrijp et al., 2016).
Cross-Sectional Study.The data of the qualitative and quantitative assays of the cross-sectional study were combined in one data set.Birthdates of calves were added to the data set to classify the calves in the following age categories: 0 to 4 d, 5 to 9 d, 10 to 14 d, 15 to 20 d, and >20 d.Descriptive statistics were used to determine the prevalence of ESBL/AmpC-EC positive calves, ESBL-EC positive calves, and AmpC-EC positive calves, in total and per age category.The variable containing the number of excreted ESBL/ AmpC-EC (in cfu/g of feces) was transformed to 3 binary variables containing the prevalence of calves shedding low, moderate, or high levels of ESBL/ AmpC-EC (as described above).These prevalences and the identification of ESBL/pAmpC genes were examined per age category of calves using descriptive statistics.To determine if the prevalence of ESBL/AmpC-EC positive calves differed between age categories, multilevel mixed-effects logistic regression analyses were conducted (Stata command: melogit).Data were analyzed on animal level taking herd of origin into account as a random effect.In this way, the pairwise comparisons between age categories were corrected for possible correlations in the prevalence of ESBL/AmpC-EC positive calves between calves originating from the same herd.The same models were used to determine if the prevalence of ESBL-EC positive calves, AmpC-EC positive calves, and levels of shedding ESBL/AmpC-EC differed between age categories.Differences between age categories in the carriage of ESBL and AmpC genes in E. coli isolates from calves were studied using proportion tests (Stata command: prtesti).All of the results of pairwise comparisons between age categories were corrected for the use of multiple comparisons according to Bonferroni's method [Stata command: pwcompare, mcompare(bonferroni)].
Longitudinal Study.To describe the results of the phenotypic and genotypic assays of the samples of calves from the longitudinal study, only descriptive statistics were used.The results of all included calves were analyzed and at each moment in time (age <22 d, at 4, 8, or 12 mo of age) it was determined whether a calf shed ESBL/AmpC-EC and, if so, which genes were identified.These results could give an indication of the duration of shedding ESBL/AmpC-EC with a certain genotype and the possible selection of ESBL/AmpC-EC with another genotype in time.Due to the limited number of calves that were included in the longitudinal study, no regression analyses were conducted on these data.

Cross-Sectional Study
Prevalence and Age Distribution of ESBL/ AmpC-EC Positive Calves.A total of 748 calves originating from 188 farms were sampled: median 3 individual calves per herd (mean: 4, minimum: 1, maximum: 15 calves).The median age of all calves was 10 d (mean: 11, minimum: 0, maximum: 88 d).Seven calves were sampled on their day of birth and 651 calves were between 1 and 20 d of age.Per age category a median of 72 calves were sampled (mean: 66, minimum: 39, maximum: 85 calves).The remaining 90 calves were between 21 and 88 d of age (Figure 2).This group of older calves, which was sampled despite the protocol, were included in the study but not further divided by age.
The prevalences of calves testing positive for ESBL-EC (ESBL-pos) and calves testing positive for AmpC- EC (AmpC-pos) in the phenotypic assay were established next to a combination of these 2 prevalences: the prevalence of calves testing positive for ESBL-or AmpC-EC (or both) in the phenotypic assay (ESBL/ AmpC-pos).This last prevalence (ESBL/AmpC-pos) was 32.9% (95% CI: 29.5-36.4%)among all sampled calves (Heuvelink et al., 2019).The prevalence of ESBL/AmpC-pos calves varied per age category [i.e., between 26.7% (d 21-88) and 36.2%(d 10-14); Figure 2].One ESBL/AmpC-pos calf already tested positive for AmpC-EC in the phenotypic assay (AmpC-pos) on the day of birth (14.3%, 95% CI: 0.4-57.9%)and 3 out of 26 calves (11.5%, 95% CI: 2.4-30.2%)tested positive in the phenotypic assay for ESBL/AmpC-EC 1 d after birth: 2 for ESBL-EC (ESBL-pos) and 1 for AmpC-EC.Among the thirty-nine 2-d-old calves, on the final day of colostrum feeding, 11 calves (28.2%, 95% CI: 15.0-44.9%)tested positive in the phenotypic assay for ESBL/AmpC-EC (6 for ESBL-EC and 5 for AmpC-EC).The results of the multilevel mixed models showed that no significant differences were found in the prevalence of ESBL/AmpC-pos calves between calves of different age categories.On the other hand, herd of origin did have a significant (random) effect: the prevalence of ESBL/AmpC-pos calves was significantly more correlated among calves originating from the same herd compared with calves originating from different herds.
Among calves 0 to 4 d of age, the prevalence of ESBL-pos and the prevalence of AmpC-pos calves were comparable (17.9% and 13.9%, respectively).However, after d 4 of age these prevalences divided and showed different patterns among the different age categories (Figure 2).The prevalence of ESBL-pos calves dropped after d 4 of age, resulting in significantly lower odds for excreting ESBL-EC among calves 5 to 9 d of age (P < 0.02) compared with calves 0 to 4 d of age (Table 1).After this drop, the prevalence of ESBL-pos calves slowly increased again among calves of older age and was not significantly different anymore compared with the prevalence of ESBL-pos calves among the youngest calves (Figure 2).
In contrast to the prevalence of ESBL-pos calves, the prevalence of AmpC-pos calves did not decrease after 4 d of age but instead it increased, resulting in significantly higher odds for excreting AmpC-EC among calves 5 to 9 d of age (P < 0.01) compared with calves younger than 5 d of age (Table 2).After this peak the prevalence of AmpC-pos calves decreased again, resulting in significantly lower odds for excreting AmpC-EC among calves older than 20 d compared with calves 5 to 9 d of age (P < 0.02).However, the prevalence of AmpC-pos calves among calves older than 20 d was not significantly different from the same prevalence among the youngest calves (0-4 d of age) and also comparable to the prevalence of ESBL-pos calves among calves older than 20 d (Figure 2).
Since the total number of excreted E. coli showed only little variation between calves and did not vary with age, the results of the analyses of the ratio between the total number of excreted E. coli and the excretion of ESBL/AmpC-EC did not differ from the above described results and are therefore not shown.
Table 3. Results of the multivariable multilevel logistic regression models of the association between calves testing positive for extendedspectrum and AmpC β-lactamase-producing Escherichia coli (ESBL/ AmpC-EC) in the phenotypic assay, shedding moderate (10 3 to 10 5 cfu/g of feces) levels of ESBL/AmpC-EC, and age of the calves 1 Parameter Shedding moderate levels of ESBL/AmpC-EC

Longitudinal Study
In total, 73 calves of 10 farms were selected from the cross-sectional study (sampling 1) to be sampled again during the longitudinal study, the resampled calves.The number of resampled calves per farm varied between 5 and 11 in the cross-sectional sampling.Not all 73 calves were sampled 4 times because at 1 sampling moment 4 calves were on pasture, and 24 calves left the farm early to go to veal calf operations or calf rearing facilities.At 4 farms 1 calf and on 1 farm 2 calves were added (additional calves) at sampling 2 (after 4 mo).Of these 6 additional calves, 4 were sampled 3 times (at an age of <22 d, 4 mo, and 8 mo) and 2 only 2 times (at an age of <22 d and 8 mo) because they were on pasture on 1 sampling moment and therefore could not be sampled.This resulted in 79 calves that were sampled at an age of <22 d, 53 calves at an age of 4 mo, 52 calves at an age of 8 mo, and 49 calves that were sampled at an age of 12 mo (Table 5).
At an age of <22 d, 46 of the 79 calves (58.2%) tested positive for ESBL-or AmpC-EC (or both) in the phenotypic assay.After 4, 8, and 12 mo the prevalence of ESBL/AmpC-pos calves dropped to 3.8% (2/53), 5.8% (3/52), and 2.0% (1/49), respectively.Of the 46 ESBL/ AmpC-pos calves of <22 d old only 5 tested positive a second time at older ages: 1 at an age of 4 mo, 3 at an age of 8 mo, and 1 after 12 mo (Table 6).The ESBL or pAmpC genotypes (or both) detected in E. coli isolates from samples of 1 calf were the same at an age of <22 d and 8 mo (bla CTX-M-2 ).In isolates of 2 other calves that tested positive twice, different ESBL/pAmpC genes were detected when the calves grew older.From the remaining 2 calves, not enough material was left for the identification of ESBL/pAmpC genes.

DISCUSSION
The aim of this study was to examine the prevalence of ESBL/AmpC-EC, the number of excreted ESBL/ AmpC-EC (in cfu/g of feces), as well as the ESBL/  4. Extended-spectrum β-lactamase and plasmid-mediated AmpC genes identified in phenotypically confirmed extended-spectrum/AmpC β-lactamase-producing Escherichia coli isolates from 155 individual calves on 63 dairy farms, aged 0-4, 5-9, 10-14, 15-20, and d 10-14).One calf already tested positive for ESBL/AmpC-EC in the phenotypic assay on the first day of life, which raises the question whether the gut was colonized with resistant bacteria before or directly after birth or whether this result was due to contamination of the fecal sample.The prevalence of ESBL-pos and AmpC-pos calves increased rapidly in the first 4 d of age, followed by a significant drop of the prevalence of ESBL-pos calves among calves 5 to 9 d old.On the contrary, the prevalence of AmpC-pos calves among these calves significantly increased, followed by a decrease in prevalence of AmpC-pos calves among older calves.Apparently, the gut of calves in the first days of life is highly susceptible to colonization with these resistant bacteria, which the calves either pick up from their environment or receive from contaminated colostrum.Selection of resistant bacteria in the gut due to antimicrobial treatment of the calves can also result in colonization of the gut with ESBL/AmpC-EC (Ewers et al., 2012).Selection in the gut due to the presence of residues of antimicrobials in colostrum and (waste) milk is, in this study, not likely because in the Netherlands farmers are not allowed to provide waste milk (milk from cows treated with antimicrobials that cannot be used for human consumption) and the dry-cow therapy used in the Netherlands mainly consists of penicillins, which do not select for ESBL/AmpC-EC.Furthermore, in several studies (Schmid et al., 2013;Hordijk et al., 2019) no significant association was found between the prevalence of ESBL/AmpC-EC in calves and feeding colostrum from cows treated with dry-cow antimicrobials.Unfortunately, information about antimicrobial treatments, the feeding of calves, or dry-cow treatment of the dams was not available in our study to evaluate these risk factors.
The above-mentioned variation in prevalence of ESBL/AmpC-pos calves per age category also gives an indication of an instable microbiota of the gut of young calves, where E. coli isolates, carrying ESBL or pAmpC genes (or both) are only to a certain extent able to colonize the gut for a longer period of time.Earlier studies already demonstrated the highly variable fecal microbiota of calves in the first days of their lives (Alipour et al., 2018;Klein-Jöbstl et al., 2019).However, other studies also showed the transient character of the shedding of ESBL/AmpC-EC, not only in calves but also in older cattle (Hordijk et al., 2013a(Hordijk et al., , 2019)).The variability of the microbiota and the transient character of ESBL/AmpC-EC probably both led to our finding that a specific ESBL genotype (bla CTX-M-1 ) was found to be significantly more frequent in calves 0 to 4 d of age than in calves 5 d of age or older.In calves 5 d of age or older, ESBL/AmpC-EC isolates relatively more often carried other genotypes.In total, the most common ESBL genotypes in calves in this study were bla CTX-M-1 and bla CTX-M-2 .This result is in accordance with an earlier study in dairy herds (Gonggrijp et al., 2016) and partly corresponding with results from other studies in dairy herds (Hordijk et al., 2019) and veal herds (Hordijk et al., 2013b,c;Ceccarelli et al., 2019) in which E. coli isolates most frequently carried bla CTX-M-1 followed by bla CTX-M-14 and bla CTX-M-15 .
The longitudinal study revealed a low prevalence of ESBL/AmpC-pos calves at an age of 4, 8, and 12 mo in calves ESBL/AmpC-pos in the first 21 d of their lives.The same results were found in another longitudinal study in 3 veal calf herds (Hordijk et al., 2013a) in which the ESBL/AmpC-EC prevalence in veal calves also declined from 18 to 26% at the moment the calves arrived at the veal farm to 0% in 2 farms and to 1.4% in 1 farm, 10 wk after arrival.The results of both studies indicate that in the end, the carriage of ESBL or pAmpC genes (or both) does not have a competitive advantage for E. coli isolates and the gut of calves is overgrown by nonresistant E. coli isolates.
This study shows that dairy calves at a young age can carry ESBL/AmpC-EC, but the colonization in their guts is temporary.More research is needed to find the source of the ESBL/AmpC-EC that colonize the gut.

CONCLUSIONS
It has been demonstrated that calves can be colonized by ESLB/AmpC-EC in the first days of their lives.In addition, the prevalence of ESBL-and AmpC-pos calves significantly varies per age category.The levels of shedding ESBL/AmpC-EC and the ESBL/pAmpC genotypes also vary per age category.These results confirm the variability of the microbiota of young calves and the transient colonization of the gut with ESBL/ AmpC-EC, which prevents long-term shedding of these isolates.

Figure 1 .
Figure 1.Overview of the laboratory assays applied on the fecal samples.*Eleven samples were not tested in the quantitative assay due to an insufficient amount of feces.ESBL/AmpC-EC = extended-spectrum and AmpC β-lactamase-producing Escherichia coli.

Table 1 .
Gonggrijp et al.: ESCHERICHIA COLI IN YOUNG CALVES ON DUTCH FARMS Results of the multivariable multilevel logistic regression model of the association between calves testing positive for extendedspectrum β-lactamase-producing Escherichia coli in the phenotypic assay (ESBL-pos) and age of the calves 1 1 Odds ratios (OR), 95% CI, and the P-values of the comparisons between age categories are corrected according to Bonferroni's method.

Table 2 .
Results of the multivariable multilevel logistic regression models of the association between calves testing positive for AmpCproducing Escherichia coli in the phenotypic assay (AmpC-pos) and age of the calves 1 1Odds ratios (OR), 95% CI, and the P-values of the comparisons between age categories are corrected according to Bonferroni's method.
Gonggrijp et al.: ESCHERICHIA COLI IN YOUNG CALVES ON DUTCH FARMS