Antimicrobial use in animals: When more might be better

Antimicrobial resistance (AMR) poses an important threat to human and veterinary medicine. In humans, it has been estimated that 4.7 million deaths were associated with bacterial AMR in 2021 (1). The World Organization for Animal Health (WOAH), however, has estimated that AMR in livestock will account for a global GDP loss of US$575 billion between 2025 and 2050, and that by 2050, AMR will account for production losses equalling the food consumption needs of 746 million to 2 billion humans (2). Limited information is available about the impact of AMR on companion animals, but it is undoubtedly substantial and increasing. As a result, antimicrobial stewardship efforts are needed to optimize antimicrobial use in both humans and animals, maximizing positive clinical impacts and minimizing AMR.

Antimicrobial resistance is a complex problem, and antimicrobial stewardship requires multimodal and multidisciplinary approaches. A key focus of this is often reduction of antimicrobial use (AMU). An emphasis on reduction in AMU in animals is logical but there are many nuances that must be considered for there to be effective action. An important component is the unit of measure. A variety of approaches have been used, consisting of count-based (e.g., number of animals or farms treated, number of days treated), weight-based (mass of active ingredient overall or based on animal biomass, numbers, patient visits, or prescribers) or dose-based (e.g., defined daily doses, overall or per number of individuals, at risk period of prescribers) metrics. Various metrics have individual advantages and disadvantages, and there is no standard approach that would be optimal for all species, sectors, and surveillance goals. Indeed, specific metrics are more or less appropriate for various situations. However, the metrics used can affect assessments and conclusions (3).

At global levels and as part of advocacy efforts, total mass of antimicrobials tends to be the focus (4). Recent international commitments have reinforced this concept, including the 2022 Muscat Manifesto, from the Third Global High-level Ministerial Conference on Antimicrobial Resistance (AMR) in Muscat, Oman that included a target to “Reduce the total amount of antimicrobials used in the agri-food system at least by 30–50% by 2030 from the current level” (5). The 2024 United Nations General Assembly High Level Meeting on AMR political declaration avoided specific targets but included a commitment to “Strive to meaningfully reduce, by 2030, the quantity of antimicrobials used globally in the agri-food system from the current level” (6).

Whereas a reduction in AMU may seem, at face value, to always be a desirable outcome, that may not always be the case. Total mass of antimicrobial is most reported and targeted because it is often the easiest information to obtain; however, it does not necessarily provide an accurate understanding of AMU, particularly across diverse animal species, animal health systems, and nations. Broad recommendations for reductions also do not adequately consider the realities of antimicrobial access, animal health, and animal production across the spectrum of high-, middle- and low-income countries. Understanding these nuances and complexities is required for meaningful discussions about antimicrobial stewardship and for effective target setting.

ALL ANTIMICROBIALS ARE NOT THE SAME

Antimicrobial surveillance has a weight problem. Total mass (e.g., kg, tonnes) is a commonly used metric to report AMU in animals, but it is crude and potentially misleading. A major flaw is that it does not account for differences in importance of the drug (in either humans or animals) or how the drug is used.

There are various categorization systems for antimicrobials that are used in animals, including international efforts such those by the World Health Organization (WHO) (7) and the European Medicines Agency (8), along with various national guidelines. These classify antimicrobials used in animals into categories that reflect their importance in human medicine and the concerns about use in animals leading to resistance in bacteria affecting humans, based on recognition that issues related to use of antimicrobials differ among drugs. The resistance risks posed by 1 kg of oxytetracycline use in animals are presumably much less than those posed by use of the same mass of fluoroquinolones, 3rd generation cephalosporins, or carbapenems. Yet, they are equivalent when assessing overall mass.

For example, based on data generated by the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS), tetracyclines are the dominant drug class administered to animals in Canada (9). Reduction of the use of tetracyclines, particularly prophylactic use, is an important stewardship target and could result in a significant decrease in overall AMU in animals. On many levels, that would be useful. However, to highlight the potential issues with focusing on overall mass, we can consider a scenario in which there is a moderate decrease in tetracycline [classified as a ‘highly important antimicrobial’ (HIA) by the WHO] but (related or unrelated) marked increase in use of a higher priority drug such as enrofloxacin [classified as a ‘highest priority critically important antimicrobial’ (HPCIA)]. Based on 2022 CIPARS data (9) data, if tetracycline use was reduced by 10% (from 510 758 to 464 325 kg) but enrofloxacin use increased to 500% of 2022 levels (897 to 4485 kg), there would have been a net decrease of total antimicrobial by 47 488 kg (9.3% of total AMU). At a crude, overall mass level, if one evaluated either tetracycline use or total AMU, this would look like a useful decrease in AMU and could be considered a stewardship success. However, in actuality, it would presumably be a very concerning change in AMU, because of the relative importance of those 2 drugs and the marked increase in HPCIA use.

Another major challenge with overall mass is drug potency and associated differences in dosing regimens. Some drugs are used at higher doses and/or administered more frequently. These are often older, lower-tier antimicrobials that should be prioritized when appropriate over newer, higher-tier and broad-spectrum options. Use of these older drugs will result in greater mass of use (Table 1). Indeed, changing from use of penicillin and tetracycline (HIAs) to drugs such as ceftiofur and enrofloxacin (HPCIAs) for the same number of treated animals and treatment duration would result in a marked decrease in mass of use but would be counterproductive to stewardship efforts. Therefore, in situations where this type of issue might be present, metrics that evaluate dosing (e.g., defined daily doses, treatment courses, percentage of animals treated) or that provide drug class-specific data would be preferable to a reliance on overall mass. The ability to collect data of adequate quality and detail for alternative metrics varies. Farm-, clinic- or veterinarian-level data could provide very useful information but may be more complicated to collect, especially in sectors or countries in which electronic medical record systems are not used or accessible.

TABLE 1.

Examples of dosing regimens for drugs belonging to various World Health Organization Medically Important Antimicrobial (WHO MIA) list drugs (7). Drug doses are from label recommendations from authorized products in Canada or international treatment guidelines (15,17).

Antimicrobial	WHO MIA List category	Dosing
Ceftiofur sodium	HPCIA	1 mg/kg q24h (cattle)
		2 mg/kg q24h (horses)
		2 mg/kg q12 to 24h (dogs)
		3 mg/kg q24h (swine)
Enrofloxacin	HPCIA	7.5 to 12.5 mg/kg once (cattle)
		2.5 to 5.0 mg/kg q24h (cattle)
		7.5 mg/kg once (swine)
		5 to 20 mg/kg q24h (dogs)
		5 mg/kg q24h (cats)
Marbofloxacin	HPCIA	2.75 to 5.5 mg/kg q24h (dogs, cats)
Cefovecin	HPCIA	8.0 mg/kg once (dogs, cats)
Tulathromycin	CIA	2.5 mg/kg once (cattle)
Gentamicin	CIA	5 to 8 mg/kg q24h (cats)
		9 to 14 mg/kg q24h (dogs)
		5 mg/1- to 3-day-old piglets, once (swine)
		6.6 mg/kg q24h (horses)
Procaine penicillin	HIA	6000 to 18 000 IU/kg q24ha (cattle, horses, dogs, swine)
Oxytetracycline	HIA	6.6 mg/kg q24 (cattle, pigs)
Trimethoprim/sulfonamide	HIA	16 mg/kg q24h (cattle, swine)
		15 to 30 mg/kg q12h (dogs, cats)
Cephalexin	HIA	15 to 25 mg/kg q12h (dogs, cats)
Florfenicol	HIA	40 mg/kg once (cattle)
		15 mg/kg 2 doses (swine)
Amoxicillin	HIA	10 to 20 mg/kg q24h (chickens)
		20 mg/kg q24h (swine)
		11 to 22 mg/kg q8 to 12h (dogs, cats)
Amoxicillin/clavulanate	HIA	12.5 mg/kg q12h (dogs, cats)
Doxycycline	HIA	5 to 10 mg/kg q12 to 24h (dogs, cats)

^aHigher doses are typically used as the current label doses are old and widely considered to be inadequate.

HPCIA — Highest priority critically important antimicrobial; CIA — Critically important antimicrobial; HIA — Highly important antimicrobial.

Overall mass, or adjusted metrics [e.g., mg/population correction unit (PCU) (3)] have a role in surveillance and are likely most effective when being used for refined comparisons that are restricted to single antimicrobials in single species, where there are fewer confounders and a greater ability to understand and describe potential limitations. For example, mass-based metrics may be useful to monitor change in use of a specific drug or drug class in response to an intervention if the main expected change would be to use of that drug (not a switch from one drug to another).

INCREASED COMPLIANCE WITH GUIDELINES

Clinical antimicrobial use guidelines aim to improve antimicrobial use practices, including optimizing drug selection and dosing. Use of higher-tier antimicrobials is common in companion animals and is often deemed to be excessive (10–13). Guidelines can improve AMU practices by steering clinicians towards effective, older, lower-tier options, although doing so can increase the mass of AMU.

As an example, a random sampling of 1000 dogs treated by veterinarians for bacterial cystitis from a published study (14) was taken and the distribution and dosing of antimicrobials was recorded. Using the median dose, median duration of treatment, and using a 15 kg dog as the exemplar, antimicrobial mass data are presented in Table 2. International Society for Companion Animal Infectious Disease (ISCAID) guidelines recommend amoxicillin, amoxicillin-clavulanic acid, or trimethoprim-sulfonamides as first-line options for bacterial cystitis (15). If treatment was changed to be fully consistent with guidelines through proportionately assigning dogs treated with other drugs to 1 of the 3 first-line options, the impact on AMU is presented in Table 2. Complete compliance with guideline drug selection would result in a 31% increase in antimicrobial mass, while eliminating the use of all HPCIAs. Few could argue that this was not a desirable outcome, despite the increase in antimicrobial mass.

TABLE 2.

Evaluation of impacts of improved compliance with antimicrobial treatment guidelines for treatment of bacterial cystitis in dogs, based on a published dataset (14).

Drug, median dose	Baseline		Guideline compliant
	Number of dogs	Antimicrobial mass	Number of dogs	Antimicrobial mass
Amoxicillin (15 mg/kg q12h)	211	949 500 mg	290	1 305 000 mg
Amoxicillin-clavulanic acid (20 mg/kg q12h)	500	3 000 000 mg	690	4 140 000 mg
Cefpodoxime (7.5 mg/kg q24h)	49	55 125 mg	0	0
Enrofloxacin (10 mg/kg q24h)	79	118 500 mg	0	0
Marbofloxacin (5 mg/kg q24h)	95	71 250 mg	0	0
Cefovecin (8 mg/kg once)	65	7800 mg	0	0
TMS (30 mg/kg q12h)	11	99 000 mg	20	180 000 mg
Total		4 301 175 mg		5 625 000 mg

^*Based on a 15 kg dog and 10-day treatment course.

Tables 1 and 2 also highlight issues that are present with long-acting antimicrobial formulations. These are often higher-potency (lower mg/kg dosing) and higher-tier (e.g., HPCIA) drugs, and beyond the typically low mass of use, single-dose treatments with long durations of antimicrobial effect can complicate measures such as days on treatment and defined daily doses. Accordingly, metrics that evaluate days on treatment, treatment courses, or appropriateness of use (e.g., percentage of treatments consistent with a guideline) may be more useful when long-acting formulations are commonly used for a given species, sector, or condition.

INCREASED ANIMAL PRODUCTION

In a burgeoning world, there is a growing demand for food, including animal protein. Population growth and increased wealth in low- and middle-income countries have greatly increased the demand for animal protein. For example, it has been estimated that, in Ethiopia, meat and milk production will increase 203 to 234% between 2015 and 2050 (16). It is simply unrealistic and inequitable to expect a developing economy to reduce overall antimicrobial use when there are marked increases in animal production. That should not be taken as license to use antimicrobials abundantly and unnecessarily; however, measurement of antimicrobial use must account for the expanding livestock population. An increase of 50% in animal production over a 10-year period that is accompanied by a 10% increase in antimicrobial use would likely reflect an important improvement in AMU, despite the overall increase in mass of use. Measures to optimize AMU are particularly important for developing economies and increases in production are ideally accompanied by parallel increases in animal health systems, preventive medicine, veterinary care, and other factors that will improve and, at the animal-level, reduce AMU. However, absolute reduction in AMU at the mass level is unrealistic. Targeting such reductions is both a recipe for failure and a way to drive nations away from antimicrobial stewardship efforts that they see as unattainable or that are interpreted (not unreasonably) as the established economies of the Global North dictating practices to the Global South and telling them that they cannot do what the Global North did during its phase of accelerated agricultural production. Although we do not want the same mistakes that accompanied rapid livestock production expansion in high-income countries, there cannot be an assumption that veterinarians and farmers in LMICs can immediately replicate the animal health systems and infrastructure that the Global North developed over decades. In rapidly expanding production systems, measures that control for animal numbers (e.g., percentage of animals treated) or biomass (mg/PCU) (3) are particularly important to help differentiate changes in use that are simply from having more animals from those associated with changes in AMU patterns.

INCREASED ACCESS

The focus on appropriate AMU should be the use of antimicrobials to improve animal health and welfare. Use as little as possible but use enough is an important consideration. Zero AMU in animals is not a realistic goal. The lower limit is unknown and is certainly well below current global practices, but there may be situations in which there is inadequate AMU. Access to veterinary care and appropriate antimicrobials remains a challenge in some regions, particularly low-income countries. Efforts are needed to increase access to antimicrobials in some regions, preferably alongside concurrent improvements in veterinary access, education, vaccine access, animal husbandry, and other components of antimicrobial stewardship. Yet, an increase in AMU in animals is a valid goal in some situations to improve animal health and welfare.

CONCLUSION

These examples are not meant to dismiss efforts to reduce AMU in animals. Rather, they are meant to improve them, to ensure that time, energy, and money are spent addressing important, relevant and, most importantly, actionable issues. There is clearly a need to optimize AMU in animals to reduce the incidence and impacts of AMR in both animals and humans. Concerted efforts from individual users to international bodies are required. However, these must focus on evidence-based, acceptable, and actionable approaches. Moving beyond simplistic approaches to AMU is critical. There are no perfect AMU metrics. Specific metrics have varying strengths and weaknesses and are better or worse in specific situations. Indeed, mass may be a relevant indicator in some situations in which issues related to the above scenarios are not involved (e.g., measuring use of a specific drug in a specific animal species where there are no corresponding changes in animal numbers or relative use of other antimicrobials). Yet, it is clearly a metric that must be used with caution. Although simple, in relative terms, to collect, focusing on total antimicrobial mass may drive ineffective or even harmful decisions. A range of metrics, particularly those that consider impacts of drug (or drug class or drug tier), animal numbers, treatment regimens, and appropriateness of use may be more informative and actionable, but more consensus building about measurement approaches is required. Antimicrobial stewardship efforts and discussions need to embrace and address the complexity in AMU monitoring and antimicrobial stewardship efforts to generate equitable, sustained, and effective measures.