Article Text


Antimicrobial use
Sixty years of antimicrobial use in animals: what is next?
  1. Luca Guardabassi, DVM, PhD, DipECVPH
  1. Department of Veterinary Disease Biology, Faculty of Health and Medical Sciences, University of Copenhagen, Stigbøjlen 4, 1870 Frederiksberg C, Denmark
  1. e-mail: lg{at}


This, the last in our series of feature articles celebrating 125 years of Veterinary Record, aims to provide an overview of antimicrobial use in animals. Starting with a journey through the history of antimicrobial use in animals, Luca Guardabassi gives his opinion on the current zoonotic risks associated with antimicrobial resistance and on how these risks might be tackled in the years to come.

Statistics from

ANTIBIOTICS were not discovered before the first issue of Veterinary Record was published in 1888. It was 1909 when the pioneer of antimicrobial chemotherapy, German physician Paul Ehrlic, proved the efficacy of a synthetic compound named Salvarsan in the treatment of syphilis. The first evidence that some microorganisms produce antibiotics came approximately two decades later, when the English microbiologist Alexander Fleming observed, by chance, that moulds growing as contaminants on laboratory agar plates were able to inhibit bacterial growth. Initially, Fleming was not aware of the potential of his discovery in the medical field as his first thought was to employ this mould's property to develop media for selective isolation of the pathogen Haemophilus influenzae from sputum (Fleming 1929). The advent of penicillin and the other antimicrobial agents discovered in the following years contributed significantly to reducing the mortality associated with bacterial infections and increasing life expectancy. It didn't take long for these ‘magic bullets’ to be recognised as important tools for improving livestock production and welfare. In the 1950s, the practice of feeding animals with antimicrobials coincided with the birth of modern livestock farming and signed the beginning of an endless debate about the appropriateness of using these important medicines in animals.

Embedded Image

A short history of antimicrobial use in animals

The history of antimicrobial use in animals goes back to the post-war time, when there was increasing demand for cheap food by industrialised countries. Farmers were facing a period of crisis because of the shortage of animal proteins and consumers were unsatisfied because of the high cost of meat. Similarly to the discovery of penicillin, the growth-promoting effects of antimicrobial agents were discovered by chance, as a consequence of a contamination problem in an experiment designed to search for an inexpensive source of vitamin B12 as a dietary supplement for poultry. The apparent positive effect of vitamin B12 on animal growth observed in the experiment was subsequently shown to be largely attributable to residues of chlortetracycline in the fermentation tank where the vitamin was produced (Jones and Ricke 2003). This discovery was of paramount importance for the growing food industry as it allowed farmers to overcome the high cost of animal-derived proteins in feed by feeding animals low concentrations of antimicrobial agents. Thereafter, antimicrobial agents were also proven to be useful therapeutic agents to treat infections and reduce mortality in livestock (Jukes 1977). The costs of veterinary antimicrobials decreased over time, leading to a rapid growth in the sale of these drugs. Farmers’ production costs dropped, with positive consequences for consumers, who could benefit from lower costs of meat, eggs and milk.

Some of the antimicrobials used for growth promotion and chemotherapy in animals were, and still are, the same as those used in human medicine. Retrospectively, this was a regrettable mistake. It would have been wiser to reserve the most important antimicrobials for human use. This practice inevitably led to discussions on whether use of the same drugs in people and animals was appropriate. In the late 1960s, a committee appointed by the UK government and led by Michael Swann warned that use of antimicrobials for growth promotion could have impaired their efficacy as therapeutic agents through the development of resistance (Swann and others 1969). This warning was widely ignored and antimicrobial consumption continued to increase exponentially in the veterinary sector for both growth promoters and therapeutic agents. The amount of antimicrobial agents produced for animal use surpassed that for people during the 1970s and was destined to increase further in the following decades. The public health concerns associated with antimicrobial use in animals became manifest in the 1990s, when usage of the vancomycin analogue avoparcin as a growth promoter was associated with selection of vancomycin-resistant enterococci (VRE) in chickens and pigs, and possible zoonotic transmission of these multidrug-resistant bacteria through the food chain. In 1997, the European Commission banned the use of avoparcin in accordance with the precautionary principle and in the following years the ban was extended to all antimicrobial growth promoters used in animal feed. Antimicrobial growth promoters are still allowed in the USA, where the total amount of antimicrobials sold for animal use exceeds, several times, that for human medicine.

In recent years, the public debate has been extended to the therapeutic use of broad-spectrum veterinary antimicrobials, such as fluoroquinolones and cephalosporins, resulting in restrictions and voluntary withdrawals of these drugs in several countries. Among the various activities by international organisations, the Food and Agriculture Organization (FAO), the World Health Organization (WHO) and the World Organisation for Animal Health (OIE) released a joint classification in 2007 of critically important antimicrobials, with the aim of finding an appropriate balance between animal health needs and public health considerations (FAO/WHO/OIE 2008). Companion animals have not been immune from legal initiatives aimed at limiting veterinary use of antimicrobial agents. This year, Sweden has banned the veterinary use of critically important antimicrobials for human medicine (eg, vancomycin and carbapenems) and restricted the use of veterinary fluoroquinolones and extended-spectrum cephalosporins in companion animals. Similar legislation exists in Finland and other countries are about to follow the example set by these Scandinavian countries.

Current consumption of veterinary antimicrobials in Europe

In 2009, the European Medicines Agency (EMA) launched the European Surveillance of Veterinary Antimicrobial Consumption (ESVAC) initiative. This is the first attempt to develop a harmonised approach for reporting data on the use of veterinary antimicrobial agents in the EU. Sales data are expressed using a population correction unit (PCU) that takes into account the estimated size of the animal populations in each country. According to the third ESVAC report, for 2011 (EMA 2013a), there are significant differences in the sales patterns of veterinary antimicrobial products between countries (Fig 1). The countries with the highest antimicrobial consumption are Cyprus (408 mg/PCU) and Italy (370 mg/PCU), followed by Spain (249 mg/PCU), Germany (211 mg/PCU), Hungary (192 mg/PCU) and Belgium (175 mg/PCU). The countries with the lowest consumption use 66 mg/PCU or less. Overall, more than 90 per cent of the sales are pharmaceutical forms applicable to group or mass treatment.

FIG 1:

Sales of veterinary antimicrobial agents for food-producing species (including horses) in 25 EU countries in 2011. Total and relative sales of each antimicrobial class are expressed in mg per population correction unit (PCU). From the third European Surveillance of Veterinary Antimicrobial Consumption (ESVAC) report (EMA 2013a)

Beside quantitative differences, there are also qualitative differences. Notable variations between countries exist in the proportion of broad-spectrum antimicrobial agents, such as third- and fourth-generation cephalosporins (range 0.05 per cent to 0.78 per cent) and fluoroquinolones (range 0.01 per cent to 13.8 per cent). Significant differences between countries were also observed in relation to the consumption of veterinary antimicrobial products in companion animals. For example, the sales of tablets containing the combination of amoxicillin and clavulanic acid ranged between 3 per cent and 100 per cent of the total sales of penicillin tablets. As remarked in the report, the variations observed between countries may be because of a variety of factors, such as geographical differences in the veterinarians’ prescribing behaviour, in the distribution of animal species and of production systems, in the availability and price of veterinary antibacterial products on the market, and in the occurrence of infectious diseases (EMA 2013a). However, it is hard to justify such big differences on the basis of these factors.

Concerns about zoonotic transmission of antimicrobial resistance

Transmission of antimicrobial resistance may take place through different epidemiological routes and mechanisms. People are exposed to resistant bacteria by direct contact with animals and animal-contaminated environments (Fig 2) or indirectly, through consumption of contaminated food of animal origin and vegetables from crops treated with animal manure. Once acquired, animal strains may transiently colonise the human body and transfer resistance genes to the indigenous microflora. Historically, the risks of zoonotic transmission of antimicrobial resistance have been associated with foodborne pathogens, such as Salmonella and Campylobacter. However, the public health burden attributable to antimicrobial resistance in these species is limited, since infections are generally self-limiting and, in most cases, managed without antimicrobial therapy. Of higher concern is the recent emergence in animals of livestock-associated meticillin-resistant Staphylococcus aureus (LA-MRSA) and extended-spectrum β-lactamase (ESBL)-producing Escherichia coli. These multidrug-resistant bacteria are, by definition, resistant to cephalosporins, which are first-line agents in the therapy of S aureus and E coli infections. Both MRSA and ESBL-producing E coli infections have high incidence and are known to increase morbidity, mortality and healthcare costs. This is why the spread of these resistant bacteria in animals is a reason for major concern.

FIG 2:

In the author's opinion, more attention should be paid to potential for transmission of resistance through direct contact with animals

LA-MRSA is primarily associated with pigs (Fig 3), even though it has been isolated from a large variety of animal reservoirs, particularly veal calves, poultry and horses (Guardabassi and others 2013). ESBL-producing E coli are found in all domestic animals, including companion animals, but the distribution of ESBL types differs between species. For example, in Europe CTX-M-1 is the most common ESBL in pigs and cattle, whereas CMY-2 is prevalent in poultry (Ewers and others 2012). The two bugs are spread by different mechanisms and this difference has important implications in relation to the risk of propagation in the human population. LA-MRSA spreads clonally, meaning that a specific clone, generally referred to as sequence type ST398 in Europe, is transferred to people working in contact with livestock. Carriage in primary carriers is usually transient and present knowledge suggests that this animal-adapted clone has limited ability to spread to the human population. On the contrary, ESBLs primarily spread by horizontal transfer of broad-host plasmids that can readily be exchanged between animals and humans without any host barriers. Moreover, the risk that ESBL-producing E coli are transmitted by food is higher than for LA-MRSA, since this route of transmission is unusual and largely unknown for MRSA. Altogether, zoonotic transmission of ESBL-producing E coli is more insidious and difficult to assess and control compared to LA-MRSA. Long-acting cephalosporins are obvious candidate-selective drivers that could be involved in the rapid spread of these resistant bacteria over the past decade. However, co-selection mechanisms are also possible, especially for LA-MRSA, where meticillin resistance is genetically linked with resistance to antimicrobial agents widely used in pig production, such as tetracyclines and zinc (Guardabassi and others 2013).

FIG 3:

Pigs and poultry are currently regarded as important animal reservoirs of livestock-associated MRSA and ESBL-producing Escherichia coli, respectively

Assessing the burden

One of the present challenges is to assess the public health burden of LA-MRSA, ESBL-producing E coli and other resistant bacteria or resistance genes of animal origin. This assessment is made difficult by the lack of data, as well as by the fact that transmission of antimicrobial resistance is a complex and largely unpredictable phenomenon involving different routes and mechanisms of transmission. This exercise is partly facilitated for LA-MRSA, since these bacteria can be distinguished genotypically from MRSA of human origin and are primarily transmitted by exposure to livestock.

Until now, the relative burden of LA-MRSA is relatively low in most countries, with the exception of countries with a low prevalence of MRSA infections and high density of pig production, such as the Netherlands and Denmark, where LA-MRSA accounts for more than 10 per cent of human MRSA infections (Guardabassi and others 2013). In these countries, LA-MRSA is also a threat to the sustainability of the local ‘search and destroy’ control policies used to maintain a low MRSA prevalence in the human population. From a global perspective, the main public health risk is related to the possibility that a LA-MRSA strain may adapt to humans, thereby increasing the MRSA burden in the general population. This possibility is not remote, as suggested by recent data from Denmark, where 21 per cent of human cases recorded in 2012 were not attributable to contact with livestock (DANMAP 2013).

For the reasons explained above, the burden of human ESBL infections attributable to animal sources is extremely difficult to quantify. A recent study has made a first attempt to quantify the burden of human ESBL infections attributable to poultry, using data generated by different studies (Collignon and others 2013). It was estimated that ESBL infections attributable to poultry in the Netherlands accounted for approximately 21 deaths and 908 hospital bed-days in 2007. The authors speculated that the use of cephalosporins and other antimicrobials in poultry would be responsible for 1518 deaths and 67,236 days of hospital admissions if these data were extrapolated to all of Europe. They also admitted that detailed data from more countries are needed to make more accurate estimates of the ESBL burden attributable to cephalosporin use in poultry. In my opinion, there are two other limitations in this study. First, the fact that the same ESBL gene is found in human infections and poultry products does not provide any information on the direction of transmission. Secondly, the number of infections attributable to poultry would have dropped significantly if the authors had considered the diversity of plasmid types associated with ESBL genes (ie, the same gene carried on a different plasmid vector is not indicative of transmission). That said, the authors’ effort of putting solid numbers behind this growing public health concern is valuable and studies like this are needed in the future, possibly using information on plasmid types and subtypes to attribute the sources of human infections.

The contribution of veterinary antimicrobial use to resistance problems in human medicine has always been and probably always will be a controversial topic. This topic is subject to multiple opinions and divergence as it involves ethical issues on animal welfare and human health, as well as economic interests by the pharmaceutical industry, the food industry and various professional categories, including farmers, veterinarians, pharmacists and researchers. As a consequence of all these factors, the debate on antimicrobial use in animals is often vigorous and not always scientifically unbiased. My personal opinion, based on 17 years of research in this area, is that the vast majority of the resistance problems observed in human medicine, especially the most critical ones associated with multidrug-resistant bacteria in nosocomial infections, including MRSA and ESBL-producers, are largely attributable to human antimicrobial use. However, even relatively small numbers may be important in this context since the level of acceptable risk is likely close to zero when human life is put at risk by the use of antimicrobials in animals.

Risk management options

Zoonotic transmission of antimicrobial resistance can be controlled in two ways: by reducing antimicrobial use or by preventing transmission of resistant bacteria from animals to humans. Antimicrobial use can be reduced by a variety of preharvest on-farm interventions, including the development of treatment formularies and guidelines to rationalise antimicrobial prescription; bans or restrictions on the use of certain antimicrobials; limitations of veterinarians’ profit on the sale of antimicrobial agents; taxes and penalties to increase the cost of antimicrobial use; and vaccination programmes to prevent infectious diseases. Preharvest interventions aimed at reducing antimicrobial use have been demonstrated to be effective in several circumstances in Nordic countries (Wegener 2012). However, this type of intervention requires variable time to reduce the prevalence of target resistant bacteria in the animal population. For example, it took 16 years to reduce vancomycin-resistant enterococci (VRE) to undetectable levels in Danish poultry after the ban of avoparcin (DANMAP 2012).

In situations where the occurrence of resistant bacteria is not directly related to the use of specific antimicrobial classes, the most rapid and effective way to prevent transmission to consumers would be to employ postharvest interventions, such as improving hygiene standards during slaughtering and food processing. Besides technological innovations for generic reduction of the bacterial load in food products, postharvest interventions comprise regulatory measures setting thresholds for the acceptable level of resistant bacteria. However, no food safety standards exist today to control trade of food products and living animals carrying resistant bacteria of zoonotic concern, such as LA-MRSA and ESBL producers. Under these circumstances, it is impossible to control dissemination of these bacteria on a global scale and any national efforts to reduce their occurrence in animals and foodstuffs are scarcely effective in the absence of international legislation and coordination. This problem is well illustrated by recent data from Denmark indicating a significantly higher rate of ESBL and LA-MRSA contamination in imported broiler meat products compared to Danish products (DANMAP 2012).

Continuous monitoring of antimicrobial resistance and antimicrobial use in animals is essential to guide risk management through detection of patterns of irrational antimicrobial use and new resistance trends, respectively. At present, national surveillance programmes are not yet established in all EU countries, but recent efforts by the European Food Safety Authority (EFSA) and the EMA have set the basis for establishing a joint European surveillance framework. Ideally, the framework should be extended to companion animals, for which data are lacking in relation to both antimicrobial use and prevalence of resistance, and comprise early warning systems for detection of new resistance phenotypes of high zoonotic potential. Resistance to carbapenems and colistin in Gram-negative bacteria and resistance to the new drugs developed against Gram-positive bacteria are likely to be the next challenges.

Recent findings indicate that the risk of zoonotic transmission of antimicrobial resistance is high for people who work or live in close contact with animal carriers. This risk has been clearly shown for LA-MRSA and I would not be surprised if a similar risk was also demonstrated for ESBL-producing E coli. In my opinion, the occupational hazards resulting from direct contact with animals deserve more attention in the future. It appears that the role of foodborne transmission might have been overemphasised in the past by both the scientific community and public health authorities, with a consequent underestimation of non-foodborne sources of antimicrobial resistance. Studies assessing the health consequences in farm workers, veterinary staff and their family members are warranted. Similar studies should be extended to people in close contact with companion animals, which are recognised reservoirs of MRSA, VRE and ESBL-producing E coli and other resistant bacteria of zoonotic potential (EMA 2013b).

What next?

Antimicrobials are essential in the cure of bacterial infections, but have an important side effect: they promote spread of resistance, thereby reducing their therapeutic efficacy over time. The literature is full of scientific evidence indicating that antimicrobial use selects for the occurrence of resistant bacteria in patients, hospitals and the community, and similar evidence is growing in the veterinary sector. The remarkable differences regarding consumption of veterinary antimicrobial products in EU countries are indicative of irrational antimicrobial use. Denmark and, more recently, the Netherlands have shown that consumption of veterinary antimicrobials can be reduced drastically without significant losses in productivity. It would, therefore, be desirable if the example provided by some countries was taken as a model in Europe, with stricter control of antimicrobial prescription, if needed.

Interventions should be tailored to the target, taking into consideration geographical, drug and animal-specific needs. For example, interventions to control antibiotic use in companion animals should take into account legal, ethical and economic implications to veterinary antimicrobial chemotherapy that differ between food-producing and companion animals. Moreover, antimicrobial use is not the only driving force promoting spread of resistant bacteria in animal populations, as indicated by the vertical spread of resistant Salmonella in poultry and pigs as well as by the recent epidemics of ESBL-producing E coli in broiler meat, which are driven by vertical transmission of strains introduced into flocks from the top of the production pyramid (Berytsson and others 2011). Furthermore, control of antimicrobial use alone might not lead to conclusive results since the spread of resistance is favoured by complex mechanisms of co-selection. As such, it is necessary to have a holistic approach that goes beyond the control of antimicrobial use and involves research aimed at developing new tools for preharvest and postharvest solutions to control zoonotic transmission (ie, vaccines, alternative treatment strategies, evidence-based diagnostic protocols, clinically effective dosage regimes, and new technologies to reduce the risk of meat contamination).

Last, but not least, it is time to redirect this historical debate about antimicrobial use in animals on to a different set of questions. Between 1999 and 2030, per capita meat consumption of livestock products is expected to rise by nearly 50 per cent in developing countries, where demand will grow faster than production, producing a growing trade deficit (FAO 2003). Should the agricultural trade deficit of developing countries be overcome by an increase in the industrial production of livestock by developed countries, or should we reconsider, politically, the way we produce livestock, with less attention to market demands and more focus on animal welfare and environmental sustainability? In the latter scenario, would consumers accept paying a higher price for livestock products obtained using lower stocking densities and fewer antimicrobials? And, ultimately, would such a radical change in industrial livestock production be economically sustainable on a global scale? There is a concrete risk that the debate will continue without coming to any rational conclusion if we do not wrestle with these questions.

What made the biggest impact?

In celebration of Veterinary Record's 125th anniversary, in April we asked readers to vote for one of 10 developments that might be considered to have had the biggest impact in the veterinary sphere (VR, April 20, 2013, vol 171, pp 415-416). Topics included the eradication of rinderpest, one health issues, BSE, diagnostic imaging, the rise of the internet and specialisation.

In all, over 1000 people voted and fewer than 40 votes separated the top two. We found that 200 people considered that effective antibiotics and anthelmintics had had the biggest impact. However, the top topic, voted for by 237 people, was developments in anaesthesia and analgesia. When considering how the field has developed from ‘doing it quickly on a wing and a prayer’ to maximum control, which has produced conditions supporting most of the major developments in surgery and improved the welfare of animals after trauma or while undergoing surgery or medical treatment, we can see why so many people thought it had had the most impact.

To wrap up the celebrations, in spring 2014, we will be publishing an article by Polly Taylor on the winning topic – the development and use of anaesthesia and analgesia in veterinary medicine.


  1. 9.
View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.