Antibiotics are a finite and precious resource, yet humanity has used them rather profligately in the last 75 years. The result has been the rise of antibiotic resistance, which is already a serious public health problem. In 2019, 1.27 million people died as a direct result of antibiotic resistance. In the four years since its emergence, 6 million people globally have died of COVID-19 – yet I do not hear a clamour to close schools or for everyone to wear masks because of antibiotic resistance.
The causes and spread of antibiotic resistance are complex but there is no dispute that antibiotic use is a critical driver of resistance. What is more controversial is how critical the use of antibiotics in the veterinary field is, particularly in the UK.
What is more controversial is how critical the use of antibiotics in the veterinary field is, particularly in the UK
In 2017, the World Health Organization (WHO) famously condemned the use of antibiotics in food-producing animals, pointing out that more antibiotics were used globally in healthy animals than in unhealthy people. In the UK the use of antibiotics that are important in human medicine as growth promoters was banned in the early 1970s and the EU banned all remaining antibiotic growth promoters in 2006. Despite this, the use of antibiotics in agriculture continued to climb until 2015, when the O’Neill report introduced targets for the veterinary use of antibiotics.
Farm animal antibiotic sales (rather than use) are published annually in the Veterinary Antimicrobial Resistance and Sales Surveillance report (VARSS). In essence, the overall sales of antibiotics for use in agriculture has declined from 62mg/kg in 2014 to 28mg/kg in 2021 (Figure 1). There are, however, some issues with these reports, including the estimation of the total mass of livestock, termed “the population correction unit” (PCU). Horses are mentioned briefly in the report, and the PCU is 400kg, which is not an accurate estimate of the average weight of a horse. Nevertheless, the reports document an impressive reduction in the use of antibiotics in the livestock sector.
Monitoring antibiotic use in equine practice via practice management systems – a study
Monitoring antibiotic use in the horse is quite difficult. A large proportion of the drugs used are not licensed, such as enrofloxacin, or are licensed for use in multiple species, such as oxytetracycline, making analysis of wholesale sales figures unreliable. We concluded that the most reliable way to monitor antibiotic usage would be retail sales obtained from practice management systems.
There are inevitably errors when people forget to book drugs, but equine practices are commercial businesses that depend on accurate booking up for their survival and most practices have systems in place to minimise forgetful losses. The advantage of equine practice is that bulk prescribing on a herd health basis is rare and the data is available on an individual basis.
Step one – create a system
With the help of Brian Whitt from Eclipse Veterinary Software we devised a computer query to report antibiotic use in equine practice. Each individual product sold by a practice could be identified as an antibiotic, including the class of antibiotic and its grouping. We chose to use the WHO classification system, which classifies antibiotics into four groups: important, highly important, critically important and highest priority critically important.
Despite the slightly insulting psychology of beginning with important and moving on up, this system has been widely adopted and most practitioners are aware of “HP-CIAs”. These include third and fourth generation cephalosporins (ceftiofur and cefquinome) and fluoroquinolones (enrofloxacin and marbofloxacin). The European Medicines Agency produces another classification with a similar but less snappily named category B “Restrict” agents.
Step two – record and report
Once the relatively limited number of antibiotic products have been identified and classified, it is necessary to record how many milligrams of antibiotics are used per “sales unit”. For instance, procaine penicillin is typically sold per millilitre, so when a horse is injected with 20ml, a veterinary surgeon would record 20 units in their booking up. By contrast, oral potentiated sulphonamides are typically sold per sachet, so 10 units might be identified if a five-day course of one sachet twice daily was sold. Procaine penicillin is typically 300mg/ml while potentiated sulphonamide sachets are 15,000mg/sachet.
It is then straightforward to report from the computer system how many milligrams of antibiotics have been sold. However, the easiest way to reduce antibiotic consumption would be to treat fewer horses, therefore some denominator is necessary.
As previously mentioned, the UK-VARSS reports use a population correction unit of animal weight and report antimicrobial use in milligrams of active antimicrobial per kilogram of livestock. In equine hospital practice, it was evident we could improve on a simple population correction unit by using the mean weight of all horses treated. The principal confounder of this data would be how many foals had their weight recorded in the practice management software. However, with this data, it was obviously a simple matter to report milligrams of active antimicrobial per kilogram of horses under care.
Step three – troubleshoot
An issue then rapidly came to light. The most commonly used antibiotic in horses was oral potentiated sulphonamides. The licensed dose for a typical 500kg horse is in the region of 15g of combined sulphonamide and trimethoprim once daily, but this is underdosing and responsible practitioners typically recommend twice daily treatment – 30g per day. However, if a horse was treated with marbofloxacin, the dose might be 1g daily. Therefore, a practitioner responsibly using “only” highly important antibiotics rather than the highest priority, critically important ones would see their figures were worse at 30g per horse rather than 1g per horse. We thus decided that the metric used in human antibiotic measurements, the defined daily dose (DDD), is preferred.
A practitioner responsibly using ‘only’ highly important antibiotics rather than the highest priority, critically important ones would see their figures were worse at 30g per horse rather than 1g per horse.
Defined daily dose identifies how many out of 1,000 people are treated with antibiotics on any given day. Slightly bizarrely the European Medicines Agency and Veterinary Medicines Directorate have developed a separate classification of how many days per year the average horse is treated with antibiotics. These are obviously directly related, with the human calculation per 1,000 being 2.7 times larger than the veterinary per 365 days.
Obviously, there are further errors as the precise dose used in any individual case is not known. However, as the mean weight of horses and the commonly recommended dose (as distinct from the licensed dose which is often incorrect) as defined in the British Equine Veterinary Association drugs app are both known, a very reasonable estimate of the number of daily doses can be calculated.
Step four – expand the study
Ultimately, we persuaded 14 practices to take part in this work, and at its peak in 2019, nearly 108,000 horses were under the care of these practices. The British Equestrian Trade Association survey last year estimated that there were 700,000 horses in the UK and therefore a significant proportion of the equine population was accounted for by this study, albeit less than the 95 percent level of coverage achieved by the UK-VARSS reports.
What did we learn about antibiotic use in equine practice?
What was the conclusion of this work (Tallon et al., 2023)? Equine practice uses 46mg of antimicrobial per kilogram of horse, down from 60mg/kg 10 years ago (Figure 2). Viewed through the preferred metric, the DDD/year was 1.52 – down from 1.70 in 2012. We have not shown the dramatic reduction in antibiotic usage seen in the food-producing animal sector. Debatably, of course, maybe equine practice was not overusing antibiotics in the first place?
This data can be compared to human antibiotic consumption data as well. The Netherlands is widely respected as a country with one of the lowest levels of antibiotic consumption, at 8.5 DDD/1,000, while Romania has recently replaced Greece as one of the highest consumers, at 25.7 DDD/1,000, with the UK at 18.2 DDD/1,000. Horses compare favourably here, with a DDD/1,000 of 4.18 (Figure 3).
A cynic might suggest that the DDD/year metric was recommended for veterinary use to make a comparison with human consumption slightly more opaque. These figures do not really support the suggestion that antimicrobial resistance is driven by excessive veterinary use.
Debatably, of course, maybe equine practice was not overusing antibiotics in the first place?
This study is hopefully only the first step. We would like to get more practices involved, to increase the coverage of equine antibiotic usage in the UK. We would like to see more engagement from other software manufacturers. The query described is not rocket science – just identifying the amount of selected products sold in a year and comparing this with the number of animals treated. It should be quite feasible to produce similar information from any practice management system. There needs to be somewhere to put this information, and update it on a regular basis. RCVS Knowledge is leading this challenge at the moment.