Investigation of Persistence of Antimicrobial Resistant Organisms in Livestock Production

Final Report: The aims of this project were to investigate the impact of farm management and medication strategies on the presence, levels and nature of antimicrobial resistant commensal organisms on pig and poultry farms with a view to informing the design of farm management plans that incorporate resistance control strategies. Conventional farming systems were contrasted with organic systems wherein the use of antimicrobial drugs is restricted. Two commensal species of bacteria were investigated, Escherichia coli (E. coli) and Enterococcus faecium (Ent. faecium). These species are adept at acquiring and transferring resistance genes and act as markers of resistances that are present in a given environment. Resistance to 5 antimicrobials was investigated; ampicillin, gentamicin and ciprofloxacin for the E. coli, and erythromycin and vancomycin for the Ent. faecium.
<P>

25 pig and broiler poultry farms were recruited. Each farm was visited at least twice over a two-year period resulting in a total of 57 farm visits and 24,000 bacterial isolations. Pooled faecal samples were collected from the house floor on each visit and data regarding drug use, farm production figures and farm management were obtained. The presence of antimicrobial resistant E. coli and Ent. faecium was detected within a sample using direct-plating onto selective agar incorporating the antimicrobial drugs at designated breakpoint concentrations. Resistant bacterial isolates were stored for further characterisation studies.
<P>
The use of antimicrobial drugs on the 13 organic farms was very low, with many of the farms never using antimicrobials. The 12 conventional farms showed a wide range of levels of drug use. The farms administering the highest quantity of drugs were the large poultry units, however those farms administering the highest intensity of drugs (measured in µg active drug per kg live-weight finished product) were conventional pig units. On both pig and poultry farms, 50-60% of the total quantity of drugs used was due to the incorporation of antimicrobial growth promoting agents (AGPs), predominantly avilamycin, in the animals’ feed.
<P>
Resistant bacteria were detected on all of the 25 study farms, including those where antimicrobials had not been used for many years if at all. The most commonly detected resistances were ampicillin-resistant E. coli and erythromycin-resistant Ent. faecium. The detection of ciprofloxacin-resistant E. coli was positively associated with the use of fluoroquinolone drugs on pig units (P=0.0007), whilst the broiler farms using lincomycin-spectinomycin were positively associated with the detection of high levels of vancomycin-resistant Ent. faecium (P=0.0001). However, for ampicillin-resistant E. coli for both livestock species, gentamicin-resistant E. coli on pig farms and erythromycin-resistant Ent. faecium on pig farms, no direct relationships were found between levels of detection and either the administration of a specific drug or the quantity of therapeutic agents that had been administered on a farm over the preceding 12-months. On the other hand, when the amount of AGPs (predominantly avilamycin) used was taken into account, those farms using the highest total quantity of antimicrobial drugs were positively associated with the highest levels of detection of resistant bacteria. As avilamycin does not exert a direct effect upon the Gram negative gut flora, the basis of this observation merits further attention.
<P>
In the second part of the study, quantitative studies were undertaken. Two broiler farms were studied in detail, one conventional and one organic and more than 10,000 plates of E. coli were enumerated during this work. An initial five-day study on the organic broiler farm observed an increased proportion of ampicillin-resistant E. coli within the faecal E. coli population in the older, heavier birds compared with the younger birds. This reveals that some resistant E. coli not only persist in the absence of antimicrobial use, but are also capable of proliferating, thus suggesting that they are not constrained by fitness costs due to the carriage of resistance and may even be fitter than sensitive counterparts. The quantitative data from the two broiler farms, was mathematically modelled using deterministic infectious disease models in order to study the persistence and potential control strategies for ampicillin-resistant E. coli (AREC). This model found that 100% birds, both conventional and organic, were colonised with AREC at an early age, however the numbers of AREC per gram of faeces was generally 10-fold lower for the organic birds than that seen for the conventional birds. The models determined that in order to eliminate AREC from broiler production, the poultry houses would need to be completely devoid of AREC after cleaning and no chicks should arrive carrying AREC. Thus it seems likely that AREC will remain highly prevalent. In order to prevent other resistance genes from becoming as common as ampicillin-resistant E. coli the reasons behind the ubiquitous presence of AREC should be further elucidated.
<P>
Resistance-profiles and minimum inhibitory concentrations were obtained for a total of 2800 isolates using 90-well micro-broth dilution Sensititre® techniques. This work found that the E. coli isolated from 3 of the conventional farms, 2 pig and 1 poultry, were of higher multiple-resistance than those isolated from 2 organic farms (median number of resistances per isolate of 5 and 1 respectively on the broiler farms). Gentamicin-resistant and ciprofloxacin-resistant E. coli isolated on the conventional pig farms showed an average number of resistances per isolate of 8 and 10 respectively. The levels of multiple-resistance shown by bacteria persisting in cleaned houses on a conventional and an organic broiler farm were similar to those shed by the previous flocks, whilst the E. coli brought on to the farms by incoming chicks were largely sensitive. Studies of the E. coli shed by the birds on a single conventional broiler farm showed that the bacteria shed by the birds went from fully sensitive to a median number of 5 resistances within the first few days of life and that this level of multiple resistance then persisted stably throughout the life of the birds. This is likely to be related to the practice of the prophylactic dosing of day-old chicks with lincomycin-spectinomycin which is acting to abolish an important source of more susceptible bacteria to the house and flock. In fact, the farm studied actually reported a decrease in clinical response to amoxicillin treatments given to flocks showing signs of dysbacteriosis after 21 days of age over the course of this work.
<P>
Field studies of emptied and cleaned houses on the same 2 broiler farms found that ampicillin- and chloramphenicol-resistant E. coli were more commonly isolated from cleaned sheds on the organic farm compared with the conventional farm. This suggests that the vigorous cleaning regimes implemented by the large conventional broiler unit are effective at decreasing levels of between-flock carry-over of bacteria. Vancomycin-resistant Ent. faecium (VREF) were seen to decrease in prevalence within the cleaned houses on the conventional farm when a phenolic disinfecting step was added to the cleaning regime.
<P>
Deterministic infectious disease models of VREF using data from these studies predicted that on large broiler farms administering lincomycin-spectinomycin treatment to day-old chicks and further amoxicillin treatment after day 20, 12.2% of the flock would be positive for VREF at slaughter. This model was sensitive to the dose of VREF required to colonise a bird and to the level of VREF remaining in the house after cleaning. Thus the results of the both the field work and the modelling carried out in this project indicate that attention should be directed to determining the most effective methods of decreasing carry-over of resistant bacteria between consecutive groups of animals on farms.
<P>
The VREF were seen to persist on 3/6 conventional broiler farms. VREF from these farms showed high multiple-resistance, with 85% of isolates resistant to 10 or more antibiotics. Furthermore, 50% of these VREF were additionally resistant to quinupristin-dalfopristin, an important human drug for the treatment of VREF infections in hospitals. The VREF were not found to be clonal in nature as a wide-range of genetic fingerprints was seen using PFGE techniques. However characterisation at the level of the transposon showed that a panel of 127 VREF, collected from 18 broiler units, only contained 5 transposon types, thus providing strong evidence for the ability of this bacterial species to pass genes horizontally between genetically distinct E. faecium. Indeed, conjugation studies found that 8 of 45 possible vancomycin-resistance gene donors were able to jointly pass vancomycin resistance and erythromycin resistance to a recipient strain. This supported the finding that the use of lincomycin-spectinomycin is associated with the persistence of VREF on broiler farms as erythromycin resistance and lincomycin resistance are encoded by the same gene, ermB. Transposon characterisation of the VREF also demonstrated the probable contamination of an organic broiler unit with VREF from a spatially distant conventional unit that concurrently operated an organic brooding facility and supplied the organic farm with 21-day old birds. After the organic farm started brooding its own chicks, VREF were no longer isolated on the farm. These observations highlight the ease with which resistant organisms can move through integrated livestock systems.
<P>
A review of published literature on the persistence of resistant bacteria on livestock farms was produced. Decreasing the use of antimicrobials generally appears to act to decrease levels of resistance at a farm and country level. However resistant organisms can persist in the farm environment and livestock at low levels for years after the original antimicrobial selective agent has been withdrawn. In fact, once a resistant organism is present the elimination of resistance is rarely reported. Therefore if the appropriate selective pressures arise these resistances may become a predominant part of the farm flora once more.

<P>