Antimicrobial resistance: a global threat with remarkable geographical differences

It is generally accepted that antimicrobial resistance (AMR) is a global threat that requires coordinated action across countries and sectors in order to minimise the emergence and spread of resistant bacteria around the world. However, the impact of AMR on public health varies significantly between countries. This is well exemplified by the great differences in the prevalence of AMR between European countries. The occurrence of methicillin-resistant Staphylococcus aureus (MRSA) among human isolates from bloodstream infections ranges from 56% in Romania to 1% in Norway, Sweden and the Netherlands (Anonymous 2015a). Similarly, the frequency of cephalosporin-resistant Escherichia coli varies from 40% in Bulgaria to 3% in Iceland, and the observed prevalence of high-risk carbapenem-resistant clones of Klebsiella pneumoniae varies between 0–62% (Anonymous 2015a). A remarkable geographical variability is also observed among zoonotic pathogens isolated from human infections. Just to highlight a prominent example, resistance to ciprofloxacin, which is one of the drugs of choice for treatment of severe zoonotic infections caused by foodborne pathogens, ranges between 0–19% and 30–98% in human isolates of non-typhoidal Salmonella spp. and Campylobacter jejuni, respectively (Anonymous 2016). In both humans and animals, positive associations exist between consumption of antimicrobials and the corresponding resistance rates in a given country, and in some cases a positive association has been shown between antimicrobial consumption in animals and resistance in bacteria isolated from humans (Anonymous 2015b). Altogether these data indicate that the geographical variability in the prevalence of AMR mainly reflects national patterns of antimicrobial usage. Accordingly, the type and extent of the interventions needed for prevention and control of AMR are not the same for all countries. Each country should define specific objectives based on the national context, and develop an adequate action plan that is able to accomplish them in line with the resources expected to be available.

This issue of the New Zealand Veterinary Journal provides new insights into antimicrobial use and AMR in livestock and companion animals in New Zealand. The seven articles selected for this issue show that both antimicrobial usage and the occurrence of AMR in animals are relatively low compared to the rest of the world, and resemble those observed in countries with low usage of antimicrobials and low prevalence of AMR, such as the Scandinavian countries (Anonymous 2015b). All the articles provide valuable suggestions on how antimicrobial usage and AMR can be further reduced in accordance with the global action plan on AMR launched by the World Health Organisation (WHO; Anonymous 2015c). Nevertheless, that action plan includes five distinct objectives and, unless unlimited resources are available, priorities should be set in order to develop cost-effective and sustainable interventions. The key factors to be considered in this prioritisation process are firstly the existing data on AMR and antimicrobial consumption, secondly the specificities of national livestock production, and thirdly the human and economic resources available for developing and implementing a national action plan.

The study by Bryan and Hea (2017) provides reassuring data regarding the amount and type of antimicrobials used in dairy cattle, the main livestock industry in New Zealand. Narrow spectrum antimicrobials, such as the penicillins, accounted for over 74% of the total reported antimicrobial consumption, expressed as mg of active ingredient sold. Data on sales of specific antimicrobial agents in cattle are available for Denmark in 2014, where penicillins accounted for approximately 68% of total sales and for 78% of the amount used for intramammary treatment (Anonymous 2015d). As for the consumption of critically important antimicrobials, macrolides and cephalosporins comprised approximately 11% and 8% of the total sales in surveyed dairy herds in New Zealand, respectively, whereas these two antimicrobial classes together represented only 2% of the antimicrobials sold for use in cattle in Denmark (Anonymous 2015d). Comparison between the two countries is difficult as the Danish data are not stratified according to production types (beef vs dairy) but according to four therapeutic groups (i.e. intramammary, cows and bulls, calves younger than 12 months old, and heifers and steers). However, these data suggest that a reduction in the use of macrolides and cephalosporins should be possible in New Zealand. Indeed, a progressive reduction in the use of these antimicrobial drugs may already be taking place according to the study (Bryan and Hea 2017). Further reduction may be achieved through revision and implementation of the existing national practice guidelines for antimicrobial use in dairy cattle veterinary practice.

The article by Hillerton et al. (2017) describes the patterns of antimicrobial sales in New Zealand in the period 2012–2015 using a methodological approach similar to that used in the European Surveillance of Veterinary Antimicrobial Consumption (Anonymous 2015e). This study allows comparison between New Zealand and 26 European countries, Australia, Canada and the United States of America. The amount of antimicrobials sold in New Zealand for veterinary use was estimated to be the third lowest after Norway and Island, and was approximately 13 times lower than in human medicine, where New Zealand was ranked as number 16 among the 30 countries examined. These data unequivocally indicate that resources should primarily be directed to reduce antimicrobial consumption in humans. As pointed out by the authors, further reductions in antimicrobial consumption in animals would require benchmarking, which in turn requires a significant public investment to implement more sophisticated integrated systems for data collection, such as those available in Scandinavian countries and in The Netherlands. I wonder whether this type of investment should be prioritised in New Zealand, where antimicrobial consumption is even lower than in some of the countries where One Health integrated surveillance systems are established. The lesson from the land of the Kiwi is that antimicrobial consumption in animals can be controlled without having such sophisticated (and expensive) systems in place. The low antimicrobial consumption observed in livestock in New Zealand is likely linked to the pastoral production systems that are widely dispersed in this country.

While designing a national action plan for control of AMR in animals, decisions must be taken on whether surveillance of antimicrobial sales should be prioritised over surveillance of AMR, or vice versa. Ideally the risks of AMR should be managed based on assessment of the actual risk, which is AMR, not antimicrobial consumption. Although there is a clear correlation between antimicrobial usage and AMR, it should be noted that the spread of certain resistant bacteria of zoonotic interest within a given livestock production system is not clearly attributable to antimicrobial usage. This phenomenon is well illustrated by the recent spread of cephalosporin-resistant E. coli producing CMY-2 β-lactamase in broiler meat products in Europe. These resistant E. coli and the CMY-2-encoding plasmids they carry are introduced into parent and broiler flocks by vertical transmission though import of contaminated 1-day old chicks from the top of the production pyramid and their spread is not driven by use of cephalosporins (Borjesson et al. 2016). Spread of AMR is often favoured by complex mechanisms of co-selection, such as co-selection of MRSA in pigs by use of zinc oxide and tetracyclines (Guardabassi et al. 2013). Notably, despite the low consumption of antimicrobials in Denmark, up to 70% of all pig farms were reported to be MRSA-positive (Anonymous 2015d). These data suggest that, for some resistant bacteria of high zoonotic risk, interventions aimed at prevention of transmission within and between farms may be more effective than restrictions on antimicrobial use, especially in countries where antimicrobial consumption is low. Such interventions require significant investments for nationwide mapping of the occurrence of AMR at farm level.

In New Zealand, the prevalence of AMR in slaughter animals has not yet been determined, as noted by Toombs-Ruane et al. (2017). Due to the multitude of antimicrobials and bacterial species, a One Health integrated surveillance system should preferably focus on AMR types for which either resistant bacteria or resistance genes may be transmitted to humans from animal reservoirs. For example, in New Zealand it would be of particular interest to know the frequency of extended spectrum β-lactamase (ESBL)-producing E. coli in cattle. The vast majority (78%) of human ESBL-producing E. coli infections in the country in 2006 was associated with the CTX-M-15 type of ESBL, and the second most common type was CTX-M-14, accounting for approximately 14% of infections (Heffernan et al. 2009). CTX-M-15 is virtually absent in livestock, but CTX-M-14 has previously been associated with cattle in other countries such as the United Kingdom (Cottell et al. 2011; Dhanji et al. 2012). Thus, one of the priorities of the Kiwi national action plan could be to investigate the prevalence of CTX-M-14-producing E. coli in cattle and beef, and compare the strains and plasmids associated with this ESBL type in human, animal and food isolates using high-resolution whole genome sequencing techniques. Ideally, this One Health investigation should include companion animals as CTX-M-14 appeared to be the most common ESBL type found in Enterobacteriaceae isolated from infections in dogs and cats (Karkaba et al. 2017a).

Three of the articles included in this special issue focus on the occurrence of AMR in bacteria isolated from companion animals. The first study provides evidence that dogs and cats in New Zealand are rarely infected or colonised with MRSA (Karkaba et al. 2017b), and that the clone involved in the sporadic infections (E-MRSA-15) was also predominant in human MRSA infections (Heffernan et al. 2014), suggesting likely human-to-animal transmission. Similarly, the second study by Karkaba et al. (2017a) shows that the two main ESBL types isolated from infections in companion animals are shared with humans, although, in contrast to humans infections (Heffernan et al. 2009), CTX-M-14 appeared to be more common than CTX-M-15 in this group of animals. Lastly, antimicrobial susceptibility data are reported from a large number of bacterial isolates collected between 2005–2012 from canine urinary tract infections (McMeekin et al. 2017). If the results of this study are compared to those from a recent European multicentre study (Marques et al. 2016), New Zealand falls in the range of countries with a relatively low prevalence of AMR. For example, in 2012 the prevalence of amoxicillin clavulanic acid resistance in E. coli isolates from New Zealand dogs (6%) was similar to those reported in Swedish and Danish dogs (3–7%) but significantly lower than in Italian, Spanish and Portuguese dogs (26–48%). For the same bacterial species, overall prevalence of resistance to trimethoprim-sulphonamides and fluoroquinolones was 8 and 2%, respectively, in New Zealand dogs, whereas in European dogs resistance to these drugs ranged from 8–32% and from 1–32%, respectively. Altogether these data indicate that the prevalence of AMR in bacteria isolated from companion animals is similar to Scandinavian countries, which are considered at the forefront of prudent antimicrobial use in small animal veterinary medicine.

This special issue also helps us to not forget the important role played by education and communication in the control of AMR. The article by McDougall et al. (2017) clearly illustrates that veterinarians are regarded by dairy farmers as the key source of advice on antimicrobial selection. It also highlights the importance of educating herd owners and farm workers about prudent antimicrobial usage. My experience suggests that veterinarians around the world usually do not receive adequate graduate and postgraduate training in antimicrobial stewardship, which is a coordinated programme that promotes appropriate use of antimicrobial drugs, i.e. the best drug, route of administration, dosage and duration of treatment for each disease condition. This lack of knowledge may affect the veterinarian’s role in educating farmers, and let economic interests and convenience factors prevail over the principles of prudent antimicrobial usage. There is also a general lack of experts able to teach antimicrobial stewardship at veterinary schools or as part of continuing education programmes. This is an issue that deserves attention (and resources) because antimicrobial stewardship is a cornerstone in the control of AMR. The availability of national experts is also required to ensure that national guidelines on antimicrobial use are appropriate and up-to-date for each main animal species and production type. International guidelines are available for specific disease conditions but they do not take into consideration important local factors such as drug availability on the market and trends in AMR and antimicrobial usage. Adequate resources should also be allocated for effective promotion and implementation of the guidelines through regional seminars and meetings with veterinarians and farmers.

In conclusion, the overall impression is that in New Zealand antimicrobial use and AMR in animals are sufficiently controlled even in the absence of an integrated programme of epidemiological surveillance. Although realisation of a comprehensive One Health surveillance system is desirable, the national action plan on AMR that is presently under development should also address the other the key elements embraced by the WHO global action plan, namely research, communication, education, infection prevention and control, and rational antimicrobial use (Anonymous 2015c). Surveillance is a useful tool to elucidate the complex interactions between antimicrobial usage and the occurrence of AMR in animals and people, but resources should not be subtracted from risk management interventions based on the existing evidence. Such interventions include efficient communication and education of antimicrobial prescribers and users, effective sanitation, hygiene and infection prevention measures, and optimisation of antimicrobial use, which in turn requires investments for updating and implementing national guidelines as well as for developing new medicines, diagnostic tools, vaccines and other interventions.