Antimicrobial resistance (AMR) is one of the most important threats to human and animal health. There are indications of a pending “post-antibiotic era” where AMR will cause 10 million human deaths per year globally by 2050 if no serious action is taken (O’Neill, 2016). As AMR is linked to antimicrobial use (AMU), there is increasing pressure to optimize AMU in both animals and humans.
Fortunately, researchers from around the world continually strive to derive knowledge that will assist dairy farmers, veterinarians and policy makers to implement effective strategies to curb AMR without negative impacts on animal welfare. In Calgary, for the past 15 years, we have worked diligently to identify what is driving AMR on dairy farms and to determine best management practices that will optimize AMU while ensuring that animal health and welfare will not be compromised. Because most AMU on dairy farms is used for treatment or prevention of mastitis, the Calgary team has focused primarily on udder health.
Non-aureus staphylococci: antimicrobial resistance and mastitis
During my PhD studies in Calgary, I was first tasked with investigating AMR in a specific group of bacteria causing mastitis in dairy cows, namely non-aureus staphylococci (NAS) (Figure 1).
The NAS (formally called coagulase-negative staphylococci) have emerged as the most prevalent group of bacteria causing mastitis worldwide (Condas et al., 2017). Due to their high prevalence in the udder, if AMR develops on virulent NAS strains, there is potential for devastating impacts on the dairy industry globally. My first job was to investigate how often AMR was present in NAS isolated from cows in Canadian dairy herds. We had a large collection of bacteria obtained from 89 herds across Canada, and we tested those isolates for AMR. Resistance to critically important antimicrobials (CIAs) such as vancomycin, fluoroquinolones, linezolid and daptomycin was absent or extremely uncommon. The so called “CIAs” are very important in human medicine, as they are used to treat specific diseases in humans for which limited options are available, including some infections acquired from animals. Conversely, resistance against non-CIAs commonly used in dairy herds, including narrow-spectrum penicillins and tetracyclines was present at relatively low levels (< 20% of isolates).
The next step was to investigate why some NAS were drug resistant. We were interested in identifying genetic elements that conferred drug resistance. Exploring AMR at the molecular level could increase understanding on how AMR develops and ultimately lead to development of effective prevention strategies. To that end, we sequenced the whole genome of 405 NAS, and looked for presence of antimicrobial resistance genes (ARGs). In NAS, the following 8 genes or group of genes were involved in AMR: blaZ, mecA, fexA, erm, mphC, msrA, and tet genes. When any of these elements were present, bacteria were genetically resistant to antimicrobials. Conversely, these genes were rarely present in susceptible isolates. Furthermore, we were also able to first explain the genetic basis of AMR in a number of NAS species.
Finally, we were interested in identifying what was driving AMR in NAS from dairy cows. Specifically, we wanted to detect antimicrobials that were more likely to induce AMR. We enrolled the same 89 herds across Canada, and instructed farmers and staff to dispose all empty drug containers into 40-L receptacles. Concurrently, milk samples were obtained from cows and NAS were stored and tested for AMR. The observation period lasted approximately 2 years.
In spite of an intuitive notion that more AMU will necessarily result in more AMR, we determined that the major risk factor for AMR in NAS was the route of administration of antimicrobials. Herds with more systemic (intramuscular, subcutaneous or intravenous) administration of antimicrobials had increased rates of AMR. In contrast, we did not detect a clear linkage between the number of local therapies (intramammary) and AMR (Fig. 2). We believe that associations between AMU and AMR in NAS for antimicrobials given systemically were due to prolonged bacterial exposure to very low antimicrobial concentrations in the udder and other tissues following antimicrobial injections not necessarily used as mastitis treatments. From a genetic perspective, prevalence of some genes strongly associated with resistance in NAS were also lower in herds that recorded fewer systemic therapies.
We also demonstrated that resistance was associated with the systemic use of 2 CIAs, namely 3rd generation cephalosporins and macrolides. Herds with greater use of these antimicrobials had increased rates of AMR in NAS.
In conclusion, we demonstrated that AMR in NAS was relatively uncommon in Canadian dairy herds, which was very positive. Nevertheless, we also demonstrated the potential for AMR in NAS to increase, particularly if a subset of CIAs was used as systemic therapies on dairy farms.
Use of critically important antimicrobials as mastitis treatment
Our work on NAS shed light on the importance of limiting the use of CIAs in dairy herds. Resistance against CIAs can increase rapidly in dairy herds if AMU is left unchecked. A critical pillar of an effective antimicrobial stewardship program is successful treatment of clinical mastitis during lactation. The next manuscript that was part of my thesis work in Calgary was a systematic review and network meta-analysis of studies that reported data on efficacy of various antimicrobial approaches to treating mastitis during lactation. We explored effects of route of administration (intramammary versus systemic), dose, antimicrobial classes (including whether antimicrobials were classified as CIAs) treatment interval, and bacteria involved (NAS, Staphylococcus aureus, Streptococcus agalactiae, environmental streptococci and coliforms).
Studies were eligible to be included if they met all the following criteria: 1) original studies that reported data on lactating cows with clinical mastitis; 2) affected cattle were given antimicrobials; 3) presence of a group that was either left untreated or given another antimicrobial; and 4) reported data on bacteriological cure. The electronic databases CAB Abstracts, NCBI PubMed, Web of Science, Scopus and MEDLINE were searched for papers published in the last 40 years. Grey literature included conference proceedings, professional associations, health agencies and other specialized databases. We did not place constraints regarding publication language nor date of publication.
We have used standardized forms for data extraction, which contained fields such as author and year, study design, clinical presentation, disease definition, bacteria, method to identify bacteria, outcome definition, treatment protocol, number of cows or quarters tested, and number of infections cured. We have also assessed each individual study with respect to study quality.
At the analysis stage, we created categories of antimicrobial treatment protocols, which included whether any kind of supportive therapy was administered, route of administration, duration of treatment as well as antimicrobial categories as defined by the WHO. The WHO categories included newer generation (third, fourth, and fifth) cephalosporins, macrolides, quinolones, penicillins (natural and anti-staphylococcal), amphenicols, and older generation (first and second) cephalosporins. Some protocols used in studies included the administration of > 1 antimicrobial. In such cases, the WHO prioritization category of each molecule was initially determined and used thereafter. Five network analyses were carried out for the most commonly isolated bacteria causing mastitis worldwide: S. aureus, Escherichia coli, NAS, non-agalactiae streptococci, and Klebsiella spp.
For E. coli, 14 studies were included in the analysis. Interestingly, no protocol based on the use of CIAs was superior to all protocols relying on non-CIAs. More importantly, there was very limited evidence to suggest that antimicrobials were necessary to treat non-severe clinical mastitis caused by E. coli, as cows left untreated had similar cure rates when compared to treated cows, regardless of the antimicrobial used. Similar findings were observed for Klebsiella spp., although the number of studies analyzed was too low to confidently make specific conclusions.
With respect to Gram-positive bacteria, 15 studies reported sufficient data to be included in our NAS network. None of the CIA-based protocols had increased cure rates when compared to protocols using non-CIAs. Likewise, conclusions involving S. aureus and non-agalactiae streptococci were all similar.
Our study demonstrated that CIAs and non-CIAs are equally effective to treat mild and moderate mastitis in dairy cattle caused by bacteria. At the farm-level, choice of antimicrobials to treat mastitis will involve several factors, e.g., cost, duration of therapy, milk discard, etc. Each element will have an important role in selection of antimicrobials to be used as mastitis treatment and will be weighted differently on each individual farm. Nevertheless, it is extremely important to emphasize that, in terms of efficacy, there is no evidence to justify the use of CIAs as a first choice to treat non-severe mastitis of dairy cattle. As the use of CIAs in food-producing animals has recently come under scrutiny (Apostolakos and Piccirillo, 2018), veterinarians should choose non-CIA of comparable or superior efficacy for treating infections in animals whenever possible.
The WHO recently published recommendations with respect to the use of CIAs in livestock, which were underpinned by independent reviews (Tang et al., 2017, Scott et al., 2018). Points of action included a limited use of CIAs to treat infections of food-producing animals. Our work in NAS and our systematic review on mastitis treatment supported a restricted use of CIAs for treating mastitis in dairy cattle. Nevertheless, many unknowns remained with respect to potential effects of such restrictions.
Restricting AMU in livestock and its potential impacts on AMR genes
The last research chapter of my thesis involved a very successful collaboration among faculties at the University of Calgary. In brief, in a series of systematic reviews, we investigated effects of restricted antimicrobial use in food animals towards AMR in bacteria isolated from animals and humans. Our work, commissioned by the WHO, received international and national media coverage, and had significant impacts on food animal production worldwide by supporting WHO’s guidelines on the appropriate use of antimicrobials in livestock.
I led the investigation of antimicrobial effects on the presence of AMR genes. The goal of this systematic review was to determine whether restricting antimicrobial use in food-producing animals will decrease prevalence of AMR genes in bacteria isolated from animals and humans. The electronic databases Agricola, AGRIS, BIOSIS Previews, CAB Abstracts, MEDLINE, EMBASE, Global Index Medicus, (AIM, LILACS, IMEMR, IMSEAR, WPRIM, WHOLIS and SciELO), ProQuest and Science Citation Index were searched for potential original studies, which were subsequently included in the review if they met the following criteria: 1) population studied included food animals and/or humans; 2) intervention was a restriction in use of antimicrobials in food animals; 3) a comparator group from same population that did not undergo such an intervention; and 4) outcome defined as presence of AMR genes.
A total of 430 manuscripts were reviewed, of which 48 were included. Restricted use of avoparcin in food-producing animals was strongly associated with a reduced prevalence of the vanA gene in bacteria isolated from animals and humans. Likewise, for the majority of AMR genes reported from studies, restricted antimicrobial use was associated with a decreased prevalence of AMR genes in bacteria isolated from animals. In dairy cattle specifically, there was a lower prevalence of the mecA gene in bacteria isolated from systems where the use of antimicrobials was restricted.
In summary, the available body of scientific evidence supported the idea that restricting use of antimicrobials in food animals was associated with a lower presence of AMR genes in bacteria. Moving forward, it is expected that our results will be valuable to guide informed decisions regarding future interventions in the use of antimicrobials in livestock.
Wrapping it up and future directions
My work in Calgary suggested that AMR in dairy herds could be reduced by restricting systemic use of CIAs to treat non-severe clinical mastitis. This approach would safeguard existing antimicrobials and reduce environmental load of AMR genes, with no significant impact on animal welfare, given the comparable efficacy of CIAs and non-CIAs.
Regarding future directions, it has become clear in the last decade that AMU can be reduced without substantial adverse effects on farm profitability (Rojo-Gimeno et al., 2016). Much progress can be made with respect to the adoption of biosecurity protocols, increased communication between farmers and veterinarians (Speksnijder et al., 2015) and genetic selection of cows with reduced susceptibility to diseases (Martin et al., 2018). As such, to reduce AMR on dairy farms, we proposed the following: 1) enhance adoption of biosecurity protocols and increase uptake of new infectious diseases surveillance tools; 2) increase understanding of host genetics, including identification of genetic variants associated with increased resistance against diseases that account for most AMU; and 3) develop efficient strategies to communicate results among multiple stakeholders, ultimately increasing producer adherence to protocols promoting decreased AMU.
Text and picture: Diego B. Nobrego
- Apostolakos, I. and A. Piccirillo. 2018. A review on the current situation and challenges of colistin resistance in poultry production. Avian Pathol 47:546-558. https://doi.org/10.1080/03079457.2018.1524573
- Condas, L. A. Z., J. De Buck, D. B. Nobrega, D. A. Carson, S. Naushad, S. De Vliegher, R. N. Zadoks, J. R. Middleton, S. Dufour, J. P. Kastelic, and H. W. Barkema. 2017. Prevalence of non-aureus staphylococci species causing intramammary infections in Canadian dairy herds. J Dairy Sci 100:5592-5612. https://doi.org/10.3168/jds.2016-12478
- Martin, P., H. W. Barkema, L. F. Brito, S. G. Narayana, and F. Miglior. 2018. Symposium review: Novel strategies to genetically improve mastitis resistance in dairy cattle. Journal of dairy science 101(3):2724-2736. https://doi.org/10.3168/jds.2017-13554
- O’Neill, J. 2016. Tackling drug-resistant infections globally: Final report and recommendations. The Review on Antimicrobial Resistance.
- Rojo-Gimeno, C., M. Postma, J. Dewulf, H. Hogeveen, L. Lauwers, and E. Wauters. 2016. Farm-economic analysis of reducing antimicrobial use whilst adopting improved management strategies on farrow-to-finish pig farms. Prev Vet Med 129:74-87. https://doi.org/10.1016/j.prevetmed.2016.05.001
- Scott, A. M., E. Beller, P. Glasziou, J. Clark, R. W. Ranakusuma, O. Byambasuren, M. Bakhit, S. W. Page, D. Trott, and C. D. Mar. 2018. Is antimicrobial administration to food animals a direct threat to human health? A rapid systematic review. Int J Antimicrob Agents 52:316-323. https://doi.org/10.1016/j.ijantimicag.2018.04.005
- Speksnijder, D. C., A. D. C. Jaarsma, A. C. van der Gugten, T. J. M. Verheij, and J. A. Wagenaar. 2015. Determinants associated with veterinary antimicrobial prescribing in farm animals in the Netherlands: A qualitative study. Zoonoses Pub Health 62:39-51. https://doi.org/10.1111/zph.12168
- Tang, K. L., N. P. Caffrey, D. B. Nóbrega, S. C. Cork, P. E. Ronksley, H. W. Barkema, A. J. Polachek, H. Ganshorn, N. Sharma, J. D. Kellner, and W. A. Ghali. 2017. Restricting the use of antibiotics in food-producing animals and its associations with antibiotic resistance in food-producing animals and human beings: a systematic review and meta-analysis. The Lancet Planetary Health 1(8):e316-e327. https://doi.org/10.1016/S2542-5196(17)30141-9