Benedicte Callens

Correlation between veterinary antimicrobial use and antimicrobial resistance in food-producing animals: a report on seven countries

OBJECTIVES To evaluate correlations between antimicrobial use and the prevalence of resistance in commensal Escherichia coli isolates from pigs, poultry and cattle, using data from publicly available national or international reports from seven European countries. METHODS The link between the quantities of different classes of antimicrobials administered to food-producing animals per country (expressed in mg/population correction unit) and the prevalence of resistance to the different antimicrobial classes (interpreted by EUCAST epidemiological cut-off values) in E. coli isolates (4831 isolates in total) was assessed by means of polynomial regression analysis and determination of Spearmans rank correlation coefficient. RESULTS A quadratic regression best fitted the antimicrobial use and antimicrobial resistance data. The coefficient of determination was, in decreasing order, 0.99 for fluoroquinolones and amphenicols, 0.94 for third-generation cephalosporins and sulphonamides, 0.93 for aminopenicillins, 0.86 for fluoroquinolones, 0.81 for streptomycin and 0.80 for gentamicin and tetracycline. Spearmans rank correlation coefficient was 1 for amphenicols, 0.96 for sulphonamides, 0.93 for streptomycin and tetracycline, 0.89 for aminopenicillins, 0.71 for gentamicin and 0.70 for third-generation cephalosporins. CONCLUSIONS These remarkably high coefficients indicate that, at a national level, the level of use of specific antimicrobials strongly correlates to the level of resistance towards these agents in commensal E. coli isolates in pigs, poultry and cattle. However, data restraints reveal the need for further detail in collection and harmonization of antimicrobial resistance and use data in Europe.

Prophylactic and metaphylactic antimicrobial use in Belgian fattening pig herds

The monitoring of antimicrobial use is an essential step to control the selection and spread of antimicrobial resistance. Between January and October 2010 data on prophylactic and metaphylactic antimicrobial use were collected retrospectively on 50 closed or semi-closed pig herds. Ninety-three percent of the group treatments were prophylactic whereas only 7% were methaphylactic. The most frequently used antimicrobials orally applied at group level were colistin (30.7%), amoxicillin (30.0%), trimethoprim-sulfonamides (13.1%), doxycycline (9.9%) and tylosin (8.1%). The most frequently applied injectable antimicrobials were tulathromycin (45.0%), long acting ceftiofur (40.1%) and long acting amoxicillin (8.4%). The treatment incidences (TI) based on the used daily dose pig (UDD(pig) or the actually administered dose per day per kg pig of a drug) for all oral and injectable antimicrobial drugs was on average 200.7 per 1000 pigs at risk per day (min=0, max=699.0), while the TI based on the animal daily dose pig (ADD(pig) or the national defined average maintenance dose per day per kg pig of a drug used for its main indication) was slightly higher (average=235.8, min=0, max=1322.1). This indicates that in reality fewer pigs were treated with the same amount of antimicrobials than theoretically possible. Injectable products were generally overdosed (79.5%), whereas oral treatments were often underdosed (47.3%). In conclusion, this study shows that prophylactic group treatment was applied in 98% of the visited herds and often includes the use of critically important and broad-spectrum antimicrobials. In Belgium, the guidelines for prudent use of antimicrobials are not yet implemented.

Moderate Prevalence of Antimicrobial Resistance in Escherichia coli Isolates from Lettuce, Irrigation Water, and Soil

ABSTRACT Fresh produce is known to carry nonpathogenic epiphytic microorganisms. During agricultural production and harvesting, leafy greens can become contaminated with antibiotic-resistant pathogens or commensals from animal and human sources. As lettuce does not undergo any inactivation or preservation treatment during processing, consumers may be exposed directly to all of the (resistant) bacteria present. In this study, we investigated whether lettuce or its production environment (irrigation water, soil) is able to act as a vector or reservoir of antimicrobial-resistant Escherichia coli. Over a 1-year period, eight lettuce farms were visited multiple times and 738 samples, including lettuce seedlings (leaves and soil), soil, irrigation water, and lettuce leaves were collected. From these samples, 473 isolates of Escherichia coli were obtained and tested for resistance to 14 antimicrobials. Fifty-four isolates (11.4%) were resistant to one or more antimicrobials. The highest resistance rate was observed for ampicillin (7%), followed by cephalothin, amoxicillin-clavulanic acid, tetracycline, trimethoprim, and streptomycin, with resistance rates between 4.4 and 3.6%. No resistance to amikacin, ciprofloxacin, gentamicin, or kanamycin was observed. One isolate was resistant to cefotaxime. Among the multiresistant isolates (n = 37), ampicillin and cephalothin showed the highest resistance rates, at 76 and 52%, respectively. E. coli isolates from lettuce showed higher resistance rates than E. coli isolates obtained from soil or irrigation water samples. When the presence of resistance in E. coli isolates from lettuce production sites and their resistance patterns were compared with the profiles of animal-derived E. coli strains, they were found to be the most comparable with what is found in the cattle reservoir. This may suggest that cattle are a potential reservoir of antimicrobial-resistant E. coli strains in plant primary production.

Clinical resistance and decreased susceptibility in Streptococcus suis isolates from clinically healthy fattening pigs.

Streptococcus suis (S. suis) has often been reported as an important swine pathogen and is considered as a new emerging zoonotic agent. Consequently, it is important to be informed on its susceptibility to antimicrobial agents. In the current study, the Minimum Inhibitory Concentration (MIC) population distribution of nine antimicrobial agents has been determined for nasal S. suis strains, isolated from healthy pigs at the end of the fattening period from 50 closed or semiclosed pig herds. The aim of the study was to report resistance based on both clinical breakpoints (clinical resistance percentage) and epidemiological cutoff values (non-wild-type percentage). Non-wild-type percentages were high for tetracycline (98%), lincomycin (92%), tilmicosin (72%), erythromycin (70%), tylosin (66%), and low for florfenicol (0%) and enrofloxacin (0.3%). Clinical resistance percentages were high for tetracycline (95%), erythromycin (66%), tylosin (66%), and low for florfenicol (0.3%) and enrofloxacin (0.3%). For tiamulin, for which no clinical breakpoint is available, 57% of the isolates did not belong to the wild-type population. Clinical resistance and non-wild-type percentages differed substantially for penicillin. Only 1% of the tested S. suis strains was considered as clinically resistant, whereas 47% of the strains showed acquired resistance when epidemiological cutoff values were used. In conclusion, MIC values for penicillin are gradually increasing, compared to previous reports, although pigs infected with strains showing higher MICs may still respond to treatment with penicillin. The high rate of acquired resistance against tiamulin has not been reported before. Results from this study clearly demonstrate that the use of different interpretive criteria contributes to the extent of differences in reported antimicrobial resistance results. The early detection of small changes in the MIC population distribution of isolates, while clinical failure may not yet be observed, provides the opportunity to implement appropriate risk management steps.

Effect of Antimicrobial Consumption and Production Type on Antibacterial Resistance in the Bovine Respiratory and Digestive Tract.

The aim of this study was to investigate the relationship between antimicrobial use and the occurrence of antimicrobial resistance in the digestive and respiratory tract in three different production systems of food producing animals. A longitudinal study was set up in 25 Belgian bovine herds (10 dairy, 10 beef, and 5 veal herds) for a 2 year monitoring of antimicrobial susceptibilities in E. coli and Pasteurellaceae retrieved from the rectum and the nasal cavity, respectively. During the first year of observation, the antimicrobial use was prospectively recorded on 15 of these farms (5 of each production type) and transformed into the treatment incidences according to the (animal) defined daily dose (TIADD) and (actually) used daily dose (TIUDD). Antimicrobial resistance rates of 4,174 E. coli (all herds) and 474 Pasteurellaceae (beef and veal herds only) isolates for 12 antimicrobial agents demonstrated large differences between intensively reared veal calves (abundant and inconstant) and more extensively reared dairy and beef cattle (sparse and relatively stable). Using linear mixed effect models, a strong relation was found between antimicrobial treatment incidences and resistance profiles of 1,639 E. coli strains (p<0.0001) and 309 Pasteurellaceae (p≤0.012). These results indicate that a high antimicrobial selection pressure, here found to be represented by low dosages of oral prophylactic and therapeutic group medication, converts not only the commensal microbiota from the digestive tract but also the opportunistic pathogenic bacteria in the respiratory tract into reservoirs of multi-resistance.

Associations between a decreased veterinary antimicrobial use and resistance in commensal Escherichia coli from Belgian livestock species (2011–2015)

In this study the possible association between antibiotic use and resistance was explored, focusing on commensal Escherichia coli from livestock (veal calves, young beef cattle, pigs and broiler chickens) in Belgium between 2011 and 2015. A continuous decreasing trend in antibiotic use was observed for all classes, except for the phenicols. Antibiotic resistance of commensal E. coli significantly decreased for several of the tested antibiotics in all livestock species. A more rapidly reverted resistance was seen to 3th/4th generation cephalosporins and fluoroquinolones. Moderate to strong correlations between antibiotic use and resistance were found, except for antibiotic resistance to chloramphenicol and gentamicin and the use of the corresponding antibiotic class. Yet, total antibiotic use was positively correlated with chloramphenicol resistance, showing the potential importance of co-selection for chloramphenicol resistance. These results suggest that national antimicrobial usage reduction campaigns have beneficial effects on the overall resistance levels. Analyses were performed on small datasets, though, and care must be taken while making inference. For more detailed analysis, antibiotic use data at an animal species level are required.

Antimicrobial resistance surveillance in Escherichia coli by using normalized resistance interpretation

OBJECTIVES To improve antimicrobial surveillance accuracy for results obtained by disk diffusion for porcine Escherichia coli, by comparing traditional clinical breakpoint interpretation with the Normalized Resistance Interpretation (NRI) method. METHODS The susceptibilities of 921 E. coli isolates from clinically healthy pigs at slaughter age was determined for 15 antimicrobials by the Kirby Bauer disk diffusion technique. NRI with previously established optimal controlled parameters for E. coli ATCC25922 was used to reconstruct the fully susceptible population of the tested E. coli isolates. Based on a lower limit for susceptibility, set at 2.5 standard deviations below the mean of the reconstructed susceptible population, the non-wild type percentage isolates was compared with the clinical resistance percentage. RESULTS The NRI method was applicable for 11 out of the 15 antimicrobials tested. Antimicrobials for which no normal distribution of inhibition zones for the population of susceptible isolates was seen, could not be used to reconstruct the susceptible population. Clinical breakpoints much lower than the epidemiological cut-off values resulted into presumptively identifying isolates as clinically susceptible, but likely carrying acquired resistance determinants. Otherwise, clinical breakpoints did cut through the WT population for several antibiotics tested, categorizing isolates from the WT population as not susceptible. CONCLUSIONS NRI was shown to be a valid method to define the WT population for disk diffusion outcomes, provided a normal distribution of the susceptible bacterial species population is present. Until international harmonization of breakpoints is achieved, it might give rise to a wide application in monitoring antimicrobial resistance in veterinary medicine.