Antibiotic use (ABU) is associated with antibiotic resistance (ABR) (1–3). It is a form of natural selection, whereby the susceptible bacterial population is killed off, allowing the resistant population to survive (4). Antibiotic resistance is considered as one of the top ten threats to human health and it is expected that more people will die due to ABR-related infections than cancer (5). The emergence of ABR is accelerated by multiple factors, which include lapses in infection prevention and control, as well as the overuse of antibiotics in human and veterinary medicine, in particular the use of antibiotic growth promoters in the animal industry (4, 6).

Veterinary antibiotics are governed by two acts in South Africa, namely: (i) the Fertilizers, Farm Feeds, Agricultural and Stock Remedies Act No. 36 of 1947 regulated by the Department of Agriculture, Land Reform and Rural Development (DALRRD) and (ii) the Medicines and Related Substances Control Act No. 101 of 1965 regulated by the Department of Health (DoH) (6–9). Products registered under the Medicines Act must be prescribed by a veterinarian, whereas a prescription is not required to purchase antibiotics registered under the Stock Remedies Act (6). The dual registration of veterinary medicinal products (VMPs) is suboptimal, but it must be considered in a historical context (9). In the 1940s, South Africa had a shortage of veterinarians and the country had a high burden of tick-borne diseases, as well as other animal husbandry issues (9). The Stock Remedies Act was therefore promulgated to ensure that farmers had access to medicines to treat sick animals in the absence of a veterinarian and to ensure food security (9). These regulations had since not changed, but guidelines have been drafted in line with safety and efficacy data per indication of use, according to the required dosage and considering the possibility of ABR emergence per antibiotic (9). In addition, South Africa has committed to follow the International Cooperation on Harmonization of Technical Requirements for Registration of Veterinary Medicinal Products (VICH) guidelines and the Stock Remedies Act is currently under review (6, 9–11). However, selected antibiotics can still be bought over the counter by anyone and be used at their discretion (6, 9–11). Antibiotic growth promoters are also not banned in South Africa, which creates an opportunity for the overuse of antibiotics (11).

A study by Van Boeckel and colleagues (2015) predicted that an average of 63 151 tons of antibiotics, measured in weight of active ingredient, were used in animals among 228 countries in 2010 (12). The authors further estimated that antibiotic consumption will intensify globally between 2010 and 2030 (63,151 ± 1,560 tons to 105,596 ± 3,605 tons); whereas the magnitude of consumption will be even greater (99% increase) in major emerging economies [i.e., Brazil, Russia, India, China and South Africa (BRICS)] (12). Global animal ABU estimates were further revised, using alternative data sources from China, and was calculated at 93,309 tons in 2017 among 41 countries, with an expected increase of 11.5% to 104,079 tons by 2,030 (13). It is important to monitor ABU with these expected increases, as it will guide ABR risk assessments and may give an indication of where resistance may arise in the future (14). However, there is limited ABU data in low- and middle-income countries, especially in sub-Saharan Africa (13, 15).

Some studies on ABU have been conducted in South Africa. Chipangura and colleagues (2017) investigated how antibiotics were used by veterinarians for small companion animals, specifically focusing on dogs through an online questionnaire, but the study did not quantify usage in terms of antibiotic active ingredients (16). A survey was conducted by Eagar and colleagues (2012) on antibiotic sales records, from eight pharmaceutical companies, from 2002 to 2004, in food-producing animals (6). The authors found that 1,538,443 kilograms (kg) active antibiotic ingredients were sold annually, of which the majority were classified as macrolides (42.4%), followed by tetracyclines (16.7%) (6). A more recent report published by the DoH showed that antibiotics sales for animal use increased by 58% from 1,005,763 kg in 2014 to 1,592,842 kg in 2015 (17). This increase is substantial and would require active monitoring to guide policy development and targeted intervention strategies to reduce ABU. However, there is no official platform in South Africa that continuously monitor ABU in the animal sector at species-level (11). Standardized methods to quantify ABU are still under development (18). Nevertheless, there are considerable efforts underway by multiple international agencies and research groups for standardization and harmonization to facilitate comparison and monitor trends over time (19, 20). Numerous units of measurements and calculation methods have been used in the literature, but the applied methodology will be dependent on data availability and should be as close as possible to the point of use (i.e. farm) (18, 21).

The South African agriculture sector is divided into backyard subsistence and commercial farming. On-farm record keeping is generally poor in subsistence farming. However, the purpose of a commercial farm is to be profitable and detailed financial records will easily be available. Financial data may therefore be used to quantify antibiotic purchases at farm-level. This will serve as a proxy for ABU, as medicinal products will not be purchased unnecessarily. Additional information on other VMPs, such as vaccines, would also be available as the use of vaccines is an effective intervention to decrease ABU through disease control (22).

The pork industry was chosen as a pilot of the livestock sector, as the industry is highly organized in South Africa, even though it is small (i.e. contributes 0.2% to the global supply of pork). Furthermore, ABU is the highest (172 mg/kg in 2010 and 193 mg/kg in 2017) in pigs and the greatest increase in ABU is expected in pork production over the next decade (12, 13). The South African pork industry is classified into two groups based on the number of pigs owned by a farmer. A formal or commercial farmer is defined as a producer who owns more than 50 pigs, whereas an informal or subsistence farmer owns 50 or fewer pigs. There are 170 commercial pig farmers that own 1,450,713 pigs in South Africa, whereas it is estimated that there are 208,312 households that own 893,262 pigs in the informal sector (23). The purpose of the study was to develop a blueprint to monitor vaccine and ABU in pork production utilizing financial documentation as data sources. This method was piloted in a commercial pig farm over a three-year period (2016–2018) using invoices, monthly statement of accounts and an inventory check to describe and quantify vaccine and antibiotic usage. In addition, the procurement costs of antibiotics and vaccines were calculated.

Methodological variation to quantify ABU is high (18). There are multiple metrics, which are classified as (i) weight- or (ii) dose-based (19, 38). Weight-based measurements are convenient based on accessibility, but dose-based measurements are more accurate (19, 38). Dose-based measurements are calculated from prescription records. However, it is currently not feasible to only use prescription records as a data source in South Africa, as some antibiotics can be purchased over-the-counter (6). Alternative data sources are therefore required to quantify ABU at farm-level in South Africa. The ESVAC performed a pilot study on pigs to determine the most appropriate variables required for the reporting of national sales data on antibiotic consumption (39). The pilot study included invoices as a data source, but ESVAC found that invoices are often unavailable and not frequently used (39). In this study, vaccine and ABU were described and quantified from on-farm invoices from a single commercial pig farm in South Africa over 3 years, indicating the richness of this data source when combined with additional information, such as an inventory check, monthly statements of accounts, livestock counts and slaughter statistics.

Globally, ABU in pigs were estimated to be 172 mg/kg in 2010 and 193 mg/kg in 2017 (12, 13). In the study, the average ABU were calculated at ~276 mg/kg. At first glance, the ABU on the farm seems high, but one should exercise severe caution when comparing ABU data (18, 19). Waret-Szkuta and colleagues (2020) applied different calculation methodologies on the same ABU data set that originated from 70 pig farms located in southwest France (40). The authors found that animal weight (theoretical vs. actual) had a considerable impact on the quantification of ABU (40). One should therefore be mindful of the limitations associated with theoretical weight when interpreting the findings of this study as actual livestock counts, live weight at slaughter and a heavier sow weight were used as input values for the denominator.

The classes of antibiotics used in the study were similar to other studies conducted in Africa (41). Aminoglycosides and ß-lactams were the most common active ingredient present in VMPs. However, tetracyclines, followed by quinoxalines, were used in the highest amounts, both in terms of volume (kg) and adjusted for animal biomass (mg/kg). Colistin was not used on the farm, as an effect of a directive issued by the South African Veterinary Council (SAVC) (11, 42). The council stated that the usage of colistin must be supported by antibiotic susceptibility testing and that veterinarians may be charged with unprofessional conduct if found guilty of misuse (11, 42). This observation highlights the importance of antibiotic stewardship in the veterinary setting and suggests that the implementation of evidence-based regulations is an important tool to reduce ABU in food-producing animals (11). It is important to also note that any ABU can lead to the emergence of antibiotic resistance. A study done by Singh and Bhunia (2019) showed that exposing both Gram-positive and Gram-negative bacteria to sub-inhibitory concentrations of two antibiotic growth promoters (i.e., tilmicosin and florfenicol) can lead to the emergence of resistance toward other antibiotics (i.e., ampicillin, tetracycline and nalidixic acid) (43). In this study, multiple antibiotics and antibiotic growth promoters were used simultaneously and might therefore have unintended consequences on the emergence, selection and transmission of antibiotic resistance to other bystander bacteria in humans, animals and the environment (44).

This study showed that most antibiotics were administered orally via feed. This is similar to a previous study done in South Africa [i.e., 68.5% (722,111.2/1,054,177 kg) of antibiotics sold are administered in-feed] (6). However, in-feed antibiotic usage was largely driven by the use of amoxicillin in weaners (volume: 103.30 ± 7.68 kg and animal biomass: 30.90 ± 3.75 mg/kg). Some studies have shown that most antibiotics are used in the weaning phase (45–48). This can be due to the immunity gap that is experienced in piglets 30 days after birth. Piglets are highly susceptible to disease at this age, due to a decrease in maternal antibodies and their immunological memory still being primed (49, 50). Interventions should therefore focus on the reduction of in-feed antibiotics by improving the piglet's gut microbiome and limiting stress when transitioning from lactation to weaning to ultimately reduce mortality and decrease ABU (51).

The farmer had an extensive vaccination programme that targeted seven different diseases caused by various bacterial and viral pathogens. Vaccination coverage for neonatal pig diarrhea was below one (0.41), which means that not all breeding animals present on the farm received this vaccination based on procurement records. This could be due to cost containment strategies, due to the Listeriosis crisis in 2017/2018 that had a devastating impact on the profitability of pork production in South Africa (52). Furthermore, this study showed that it was more expensive to procure vaccines than antibiotics. The calculated cost of vaccination is potentially an underestimation as the cost of needles, syringes and labor were not calculated. This has important implications, as vaccines are advocated as alternatives to antibiotics (22). Vaccines are considered by both veterinarians and farmers to be more acceptable to reduce mortality, increase herd health and reduce ABU (22, 53). The government can consider tax incentives to increase vaccination uptake (54). This will not only reduce mortality but will also increase herd health, which will lead to greater production output using fewer antibiotics.

The study had several limitations and potential biases. Only a single farm was recruited and findings can therefore not be generalized to the rest of South Africa, as ABU is farm-specific and one can expect farm-to-farm variability (45, 55). Furthermore, the blueprint has only been piloted in a single farm and it is unknown if this will be useful in other farms. Participant recruitment was challenging, as there is no official farm register available in the public domain in South Africa to design a representative sampling plan. In addition, this methodology might be impractical in backyard and subsistence farming, as financial record keeping is usually not done. Furthermore, financial record keeping on this farm was done in a very organized and exact manner, which might not be the case in other farms. Lost invoices or insufficient description of items purchased (i.e., lack of data) in invoices may be potential bias in reporting. However, if only a single metric is used in a surveillance system it can be used for benchmarking as all farms will have the same biases and the percentage of missing invoices can be calculated from a monthly statement of accounts. In this study, the authors refrained from calculating the amount of ABU per pig population separately as the appropriate livestock counts per emptying period were not available to calculate the denominator. Furthermore, the same active ingredient was used in multiple populations simultaneously and it was unknown if the use distribution was equal for the different populations. The authors also did not calculate animal-defined daily dose (DDD_vet) as the exact number of animals treated were unknown and the DDD_vet is also not a suitable metric to quantify antibiotic growth promoters use.

Despite these limitations, this study had shown that invoices and monthly statement of accounts are a useful data sources to quantify vaccine and ABU at farm-level for a particular food-producing animal species. This study may in future serve as a blueprint to quantify usage in more commercial farms across different food-producing species in the absence of prescription records. This blueprint can serve as a template on how ABU can be reported, by showing which variables are required and how calculations can be performed in a future species-specific vaccine and ABU surveillance system in South Africa. A system, as such, would be important to guide risk assessments, residue monitoring and antibiotic resistance surveillance programmes, which in turn can be used the South African government in the development of evidence-based animal health policies. This system can also be used as a tool to measure the effectiveness of interventions, as antibiotic growth promoters are phased-out to ultimately improve animal health and welfare, without negatively impacting productivity and economic viability.

The datasets presented in this study can be found in online repositories. The accession numbers to the repositories can be found in the Supplementary Material.

The studies involving human participants were reviewed and approved by Human Research Ethics Committee (Medical), University of the Witwatersrand, Johannesburg (M190244) and the Faculty of Health Sciences Research Ethics Committee, University of Pretoria, Pretoria (406/2020). The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by Research and Animal Ethics Committee of the Faculty of Veterinary Science (REC0055-20) and the Research Ethics Committee of the Faculty of Humanities, University of Pretoria, Pretoria (HUM027/0620). In addition, the Animal Research Ethics Committee of the National Institute for Communicable Diseases (NICD), Johannesburg (AEC003-19) approved the study. A section 20 clearance certificate (12/11/1/1/13) was obtained from Department of Agriculture, Land Reform and Rural Development, South Africa.

WS conceptualized the idea, collected, entered, analyzed, and interpreted the data and wrote the manuscript. ML assisted in data collection and independently double-checked calculations used in the summary of product characteristics. EE provided support in statistical analysis. OP and EE provided commentary on the manuscript and interpretation of the results. All authors reviewed the manuscript.

Research reported in this publication was supported by the South African Medical Research Council (SAMRC) as a sub-grant received from the Bill and Melinda Gates Foundation (Project code: 96086). The contents and findings reported/ illustrated are the sole deduction, view, and responsibility of the researcher and do not reflect the official position and sentiments of the SAMRC or the Bill and Melinda Gates Foundation. WS was also supported by a Fogarty International Center Global Infectious Disease research training grant, National Institutes of Health, to the University of Pittsburgh and the National Institute for Communicable Diseases (D43TW011255).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.