Chapter 1. Design and Methods

Antimicrobial Use

What's New

• The 2012 human antimicrobial use data are not included in this report but a summary of the 2012 and 2013 data will be presented in the CIPARS 2013 Annual Report – Chapter 3. Antimicrobial Use.
• The design or the methods of the CIPARS swine Antimicrobial Use Farm Surveillance component hasn't changed in 2012.
• The design or the methods of the Surveillance of the antimicrobials distributed for sale for animals (data provided by the Canadian Animal Health Institute – CAHI) hasn't changed in 2012.

Human Surveillance

To be more timely with the analysis and distribution of the data, the 2012 human antimicrobial use data are not presented in this report but will be presented in the 2013 PHACFootnote 51 report – Human Antimicrobial Use Annual Report and summarized in the CIPARS 2013 Annual Report – Chapter 3. Antimicrobial Use. The most recent information pertaining to the design and methods will be presented in the CIPARS report Human Antimicrobial Use Report, 2011.

Farm Surveillance

Pigs

Farm Swine Questionnaire

In the Farm Surveillance component of CIPARS, sentinel farm data were collected through questionnaires administered by the herd veterinarian (or designated staff) to the producer (or designated farm staff). The questionnaires collected data on antimicrobial use (AMU), herd demographics and animal health.

Questions pertaining to the number of pigs in the population of interest differed by management system: continuous-flow or all-in-all-out. All-in-all-out management is a production system whereby animals are moved into and out of facilities in distinct groups. By preventing the commingling of groups, the hope is to reduce the spread of diseases. Facilities are normally cleaned and disinfected thoroughly between groups of animals. This type of management is generally by room or by barn. In continuous-flow operations, animals are continually being removed and added.

The AMU questionnaire was designed to collect data for herds of pigs in the grower-finisher production phase. No data on individual pigs were collected. Six pens representative of this population were selected for the collection of fecal specimens for bacterial culture and antimicrobial susceptibility testing. Thus, in herds with all-in-all-out management, the population of interest included all pigs that entered and exited the barn in the same group as the sampled pigs. The population of interest in herds with continuous-flow management was pigs that entered the grower-finisher unit with the sampled pigs.

Herd owners/managers were asked about AMU via feed, water, and injections. Data were collected on each diet fed to the specified group of pigs, including medicated and non-medicated feeds (non-medicated feeds did not contain antimicrobials). Information collected on each type of feed fed during the grow-finish period included the average number of weeks each ration was fed and the associated start and end pig weights. Additional information was collected for diets containing antimicrobials: active antimicrobial ingredient(s), their concentration(s) in the feed, and the primary reason(s) for that AMU (growth promotion, disease prevention, or treatment). Secondary AMU reasons were captured if the primary use was for disease prevention or treatment; secondary reasons included: respiratory disease, enteric disease, lameness or other diseases.

Data collected on exposure to antimicrobials through water or injection included active ingredient(s) in the drug(s) used, the reason(s) for use and the proportion of pigs exposed. The primary reasons for AMU in water included: disease prevention and disease treatment with associated secondary reasons for use being respiratory disease, enteric disease, lameness or other diseases. Only disease treatment reasons were collected for AMU administered by injection. The number of pigs exposed to AMU by water or injection were captured as categorical data with ranges of 1-25%, 26-50%, 51-75% or 76-100% of the pigs. No AMU data were collected for any production phase prior to the grower-finisher phase. Any data regarding AMU in pigs weighing less than 15 kg (33 lb) were excluded because this weight is considered below the industry standard for grower-finisher pigs.

Data Analysis

Data were entered into a PostGreSQL Database, and descriptive statistics were obtained with commercially available software^{Footnote 52}.

Antimicrobial exposures were summarized for each herd. An exposure was defined as any reported use of an active ingredient by a given route of administration in 2012. Data were reported as exposure to an active ingredient by a given route of administration, as well as by exposure to an active ingredient by any administration route. These exposures were summarized by antimicrobial class. It is important to note that antimicrobial exposures through feed tend to involve larger groups of pigs and longer durations of use than antimicrobial exposures via water. Alternatively, injectable antimicrobials are generally administered on an individual basis to a limited number of pigs^{Footnote 53}.

Quantitative AMU data (dose and duration) were collected for antimicrobials administered through feed but not for antimicrobials administered through water or by injection. Table 9 summarizes the reported antimicrobial active ingredients and classes used by their Categories of Importance to Human Medicine and ATCvet codes. The amount of an antimicrobial consumed through feed was estimated from the concentration of the antimicrobial in a given ration multiplied by the cumulative tonnes consumed over the duration of exposure. The cumulative feed consumption was calculated using National Research Council feed intake estimates for average performing pigs for the weights indicated by the producer in the questionnaire for each specified ration^{Footnote 54}. Quantitative results for AMU through feed are reported as kilograms of active ingredient per 1,000 pig-days at risk, which standardizes the number of pigs and the duration of exposure for a given antimicrobial use.

Surveillance of Antimicrobials Distributed for Sale for Use in Animals

Quantities of Antimicrobials Distributed for Sale for Use in Animals

As an estimate of antimicrobials used in animals, data on active ingredients distributed for sale were aggregated and provided to the Public Health Agency of Canada by the Canadian Animal Health Institute (CAHI). CAHI is the trade association representing the companies that manufacture and distribute drugs for administration to food (including fish), sporting, and companion animals in Canada. The association estimates that its members’ sales represent over 90% of all sales of licensed animal pharmaceutical products in Canada^{Footnote 55}. CAHI coordinates electronic collection of data from its members on the total kilograms of antimicrobials distributed for sale. Data collection and analysis are performed by a third party, Impact Vet^{Footnote 56}. The CAHI data include information from 15 companies that manufacture antimicrobials products for use in animals in Canada, and 5 major wholesalers/distributors. The CAHI data on the distribution of antimicrobials for use in animals provide a context to interpret other data on antimicrobial use in animals generated through research and farm data collection. They also provide a means to estimate gross temporal changes in antimicrobials used in animals.

The level in the distribution chain that kilograms of active ingredients are reported to CIPARS is at the feed manufacturer/veterinary clinic/over-the-counter outlet feed mill. Antimicrobial use was assigned to either production animal (inclusive of horses) or companion animal by the manufacturers according to label claim, and in the situation where mixed species was indicated on the label, the manufacturer assigned (estimated) the species as either companion animal or production animal based on the veterinary clinic practice profile.

These data do not represent actual antimicrobial use in a given year; rather, they reflect the volume of antimicrobials distributed by manufacturers and wholesalers. Distribution values should approximate amounts used, particularly when data from more than one year are included. However, when data from only one year are included, distribution values may vary from amounts actually used because of the time lag between distribution and actual use, as well as stockpiling of antimicrobials at various points in the distribution system. The sales data also do not account for drug wastage due to drug expiry.

The data do not include antimicrobials imported for personal use (own use importation - OUI) under the personal-use provision of the federal Food and Drugs Act and its Regulations, nor do they include imported active pharmaceutical ingredients (API), which are drugs imported in non-dosage form and compounded by a licensed pharmacist or veterinarian. These data also do not include prescriptions for antimicrobials used in companion animals could be filled at human pharmacies. Hence, the CAHI data are currently an underestimate of the true volume of antimicrobials used in animals in Canada. Also, as the CAHI data represent manufacture and distribution-level data, these data do not capture what happens to the drugs after purchase; hence this data cannot provide information the actual antimicrobial use practices, such as dose, duration, reason for use, detailed species-specific information, or extra-label use.

The CAHI data also include medicines sold directly to pharmacists that have a focus on dispensing for production medicine. It does not include antimicrobial agents moved from veterinarians to pharmacies and then subsequently dispensed by pharmacies. The latter distribution is captured with the veterinary clinic-level data.

CAHI provides the information in categories, with some antimicrobials not independently reported. This is based on a “3 company accounting rule” established by CAHI to comply with the European Union and the United States’ anti-competition regulations. CAHI added in some cases a “90% rule” to be sure not to infringe the regulations in the United States. These accounting rules can result in changes to the categorization of specific antimicrobials over time. For 2012, the antimicrobials are categorized as per Table 4.

Population Correction Unit

Changes in overall sales/distribution of antimicrobials over time may reflect several things: 1) true change in use practices, 2) a change in the numbers or types of animals in the population (requiring antimicrobials), 3) changes in disease prevalence necessitating antimicrobial use, and 4) changes in the types of antimicrobials administered (with different potencies). As one way to adjust the sales data for the changing animal populations over time, a denominator accounting for the number of animals and their standardized weights (animal biomass) was applied. This denominator was based on the methodology currently in use by the European Surveillance of Veterinary Antimicrobial Consumption (ESVAC)^{Footnote 57}.

ESVAC adjusts the sales data by a population correction unit (PCU)Footnote 57; in which a PCU is a proxy for the animal biomass that is at risk of being treated with antimicrobials. The PCU has been described as “currently the best approximation of use, extrapolated from sales data, for changes within a country over time and comparison between countries”Footnote 58. It is a technical measurement only; where 1 PCU = 1 kg of different categories of livestock and slaughtered animals. ESVAC methodology was applied to the greatest extent possible, however population information collected by Statistics Canada and Agriculture and Agri-Food Canada is different in structure somewhat from the data collected by Eurostat and TRACES, hence direct comparisons of PCU’s or mg/PCU with ESVAC participating country data should only be made with due caution.

The PCU is calculated by multiplying the numbers of livestock and slaughtered animals in each species/production state (n) by the theoretical (standardized) weight at the most likely time of treatment^{Footnote 57Footnote 59}.

PCU (kg) = n * weight(kg)

AMU = (Antimicrobials distributed in milligrams) / (PCU in kilograms)

National denominator data regarding the number of livestock and slaughtered animals for 2006 to 2012 were obtained from Statistics Canada^{Footnote 60}, Agriculture and Agri-Food Canada^{Footnote 61}, and Fisheries and Oceans Canada^{Footnote 62}, and Equine Canada^{Footnote 63} websites. Validation of the data accessed for the animal populations is currently underway, the PCU measures as provided in this report should be considered provisional.

The average weights at treatment used in these calculations, as per ESVAC, can be found in Table 5. Canadian average weights were not used for this surveillance reporting period, as there is current on-going discussion with industry stakeholders to determine appropriate weights in the Canadian context. However, the intention is that future reporting of the CAHI data will additionally include average weights of treatment/average weights of the production stage more specific to the Canadian context. There may also be alterations in the production categories included in a Canadian PCU denominator; hence future reports using this metric will vary depending upon the outcomes of these discussions. Future reports will articulate a clear distinction in the results/methods as to which denominator is applied.

Detailed inclusions and exclusions for the PCU denominator: As per ESVAC, exported animals were added to the PCU, whereas imported animals were subtracted, based on the ESVAC assumption that animals are treated in their country of origin. However, it was noted that in the Canadian context, this would vary depending upon the production stage that is crossing the border. For the purposes of calculating the PCU, production animal species with the largest populations were included, using the same production classes as ESVAC, with the exception that we additionally included beef cows (not included by ESVAC). Species currently excluded from our PCU calculations include game animals (e.g., moose), “pocket” companion animals (e.g., hamsters, guinea pigs, pet birds), reptiles, and amphibians. Import and export data for poultry are included in a different structure before and after 2009, based on the data available from Statistics Canada. For cattle, international export data was not stratified by the type of cattle (i.e., dairy versus beef cattle), as this stratification was not available in the data accessed. The total number of cattle slaughtered per year as provided/accessed was not stratified by type of cattle (beef versus cull dairy); hence it was assumed that the total slaughtered includes all cattle types (including cull dairy).

Provincial Stratification of the Numerator and Denominator

There may be subsequent distribution of antimicrobials across provincial borders after being distributed to the veterinary clinics (in particular the movement of medicated feed - for example, anecdotal information is that New Brunswick has a negligible feed-mill industry, they generally purchase their medicated feed from Québec), hence caution should be applied when interpreting the quantities of antimicrobials distributed for sale within each province. An effort was made to calculate a PCU at the provincial-level, however there is ongoing discussion with industry stakeholders regarding the inter-provincial movement of animals. As inter-provincial export data is not available for all species in all provinces, provincial calculations of PCU will be postponed pending further discussion.

Overall Discussion of Strengths and Limitations

CIPARS currently has farm-level surveillance in the swine and poultry sectors. The CAHI data provides a rough measure of antimicrobials distributed for sale for all animal species, including those not covered by CIPARS farm-level surveillance (with appropriate caveats regarding OUI/API). With respect to the PCU, as stated in the United Kingdom’s surveillance report on antimicrobials sold for use in animals^{Footnote 65}, the population is an important denominator, as the greater the number of animals, the greater the potential need for antimicrobial therapy. The PCU metric currently does not take into account the lifespan of the animal, which may affect the interpretation of the quantities of antimicrobials administered to animals. Also, use of a static standard weight may not reflect an industry shift in production affecting the average weights of animals treated, related to weather, trade, or other reasons. Measures of antimicrobial use as reported by broad categories and by a PCU denominator do not account for the individual potencies of the drugs that make up the category. For example, a decrease in the mg/PCU reported for a given year could potentially reflect a switch to using a more potent drug, as opposed to reflecting a decrease in the actual exposure of animals to antimicrobials. The CAHI data should be interpreted as one measure describing antimicrobials used in animals, strong caution should be applied with making inferences to any use practice for a particular animal species. CIPARS continues to work to improve this measure and other appropriate measures, to best reflect antimicrobial use in the Canadian context.