Veterinary antimicrobial-usage statistics based on standardized measures of dosage

Preventive Veterinary Medicine 64 (2004) 201–215

Veterinary antimicrobial-usage statistics based on standardized measures of dosage V.F. Jensen∗ , E. Jacobsen, F. Bager Department of Epidemiology and Risk Assessment, Danish Institute for Food and Veterinary Research, Mörkhöj Bygade 19, DK-2860 Söborg, Denmark Received 2 December 2002; received in revised form 19 February 2004; accepted 6 April 2004

Abstract In human medicine, the defined daily dose is used as a technical measure of drug usage, which is independent of the variations in the potency of the active compound and the formulation of the pharmaceutical product—therefore providing a measure of the relative importance of different drugs. A national system of animal defined daily doses (ADD) for each age-group and species has been defined in VetStat (the Danish national system monitoring veterinary therapeutic drug use). The usage is further standardized according to the number of animals in the target population, acquired from production data on the national level or on herd size by species and age in the Danish central husbandry register (CHR). Statistics based on standardized measures of VetStat data can be used for comparison of drug usage between different herds, veterinary practices, or geographic regions (allowing subdivision by animal species and animal production class, route of administration, disease categories, season and geographic location). Individual statistics are available as interactive reports to the control authorities, farmers and veterinary practitioners by a secure access to the database. The ADD also is used in pharmaco-epidemiogical research and to assist in the interpretation of resistance-surveillance data. © 2004 Elsevier B.V. All rights reserved. Keywords: Drug usage; Veterinary; Surveillance; VetStat; DDD; ATCvet; ADD; CHR

1. Introduction Almost all antimicrobials used in veterinary medicine are structurally related to human therapeutics and may select for co- or cross-resistance. There is mounting concern of human health implications from the antimicrobial resistance caused by veterinary use of ∗

Corresponding author. Tel.: +45-72-347349; fax:+45-72-347028. E-mail address: [email protected] (V.F. Jensen). 0167-5877/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.prevetmed.2004.04.001

202

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

antimicrobials (Bager et al., 1997; Low et al., 1997; Aarestrup et al., 1998; Dunlop et al., 1998; Cruchaga et al., 1998; Wegener et al., 1999; Aarestrup et al., 2000; Tollefson and Miller, 2000; van den Boegaard and Stobberingh, 2000; WHO, 1997, 1998). Recommendations on preventive measures to be taken include surveillance of antimicrobial resistance and monitoring antimicrobial usage in food animals—guidelines for which have been defined by EU invitational conferences (Rosdahl and Pedersen, 1998; Anonymous, 2001), by an OIE ad hoc Group of Experts (Nicholls et al., 2001), and by the WHO (WHO, 2000, 2001). Monitoring antimicrobial usage in animals should aim at: • facilitating control and intervention to ensure compliance with developed strategies and regulations on the use of antimicrobials; • ensuring prudent use of antimicrobials; • assisting in the interpretation of resistance-surveillance data; • providing information for research in specific use conditions that govern selection and dissemination of resistant bacteria; and • providing vital information for risk assessment of resistance development at the population level. Acting on the recommendation of 1998 EU invitational conference, the Danish Government decided to implement herd-level monitoring of all drug usage in production animals. The monitoring program, VetStat, yield detailed information about source and usage (Stege et al., 2003). For each prescription item, the data comprise: • date of sale (pharmacies and feed-mills) or use (veterinary practitioner); • source (identity of dispensing pharmacy, feed mill, or the veterinarian using the drug); • drug (quantity and commodity number1 of drug dispensed, providing basic information on the drug such as administration route, formulation, ATC/ATCvet code,2 active components and strength); and • recipient (farm-identity code (CHR-ID) within the Danish Central Husbandry Register (CHR), target animal species, age-group, and disease category). The “age-group” also to some extent represents animal production classes (Table 1). Studies of drug usage face the crucial problem of determining a suitable unit of measurement. Measuring veterinary usage of antimicrobial drugs in terms of weight of active compound has clear limitations as a measure of antimicrobial-drug usage in analyzing the impact on development on resistance (Chauvin et al., 2001). In human medicine, the ATC Index is used widely for classification of drugs. The related “defined daily dose” (DDD) is an internationally accepted technical measure of usage; DDD is independent of the differences in potency between active compounds and different formulations of the pharmaceutical products. The ATCvet/ATC Index has been recommended also for veterinary-drug statistics (WHO, 2001; Nicholls et al., 2001). In the Nordic countries, the ATC or ATCvet code 1 The Nordic commodity number is used to identify the individual packaging of a medicinal product (name of medicinal product, form, strength and size of packaging) with approval of marketing in Scandinavia. 2 The Anatomical Therapeutic Chemical Code identifying the therapeutic ingredient(s) of all human drugs in a hierarchical system with five levels to the active substance level. The ATCvet is the veterinary counterpart (http://www.whocc.no; Anonymous, 2002).

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

203

Table 1 Standard body weights defined for the VetStat age-groups Animal species

Age-group

Standard weight (kg)

Pigs

Breeders with their suckling pigs Weaners Slaughter pigs

200 15 50

Cattle

Cows, bulls Calves <12 months Heifers, steers

600 100 300

Small ruminants

Sheep, goats >12 months Lambs/kids <12 months

Poultry

Broilers Layers Rearing flocks

Aquaculture Mink Other production animals Horses

Not recorded Not recorded Not recorded Not recorded

50 20 0.2 1 1 1 1 1 500

is assigned to every approved drug and the index is used in veterinary-drug statistics in VetStat. Because an international veterinary equivalent to the DDD does not yet exist, an animal daily doses (ADD) was defined within the VetStat system (the abbreviation ADD is used to designate the national DDD for animals). In this paper, the advantages and disadvantages of the DDD and ADD are discussed. The methodological principles of the standardization of drug use and the implementation into VetStat are described. Suggestions for application of the ADD system are described, and examples of actual and intended use are given.

2. Measuring drug consumption Numerous units for quantifying drug consumption have been described in the literature, including financial units, commercial units, weight indicators, and descriptive units (Chauvin et al., 2001; Merlo et al., 1996). Health authorities and researchers most often use weight indicators, e.g. gram active compound, the defined daily dose and the prescribed daily dose (PDD, the average daily dose prescribed). The unit of measure should be suitable for describing and comparing drug consumption in a wide variety of treatment settings in veterinary medicines in different populations and over time (Chauvin et al., 2001). Studying the total weight consumption of a therapeutic class presents the assumption that each active compound has the same potency—whereas, the PDD in mg/kg has a wide range depending on the active ingredient used (Arnold et al., 2004). If a substance is gradually substituted by a related, more-potent substance, a false impression of a decrease in total drug consumption might be given by mere weight data, while the treatment intensity might be unchanged or even increasing. Statistics based

204

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

on total weight of the active substance and denominator information on population size allow for international comparisons. However, comparisons of the therapeutic effect and the selective pressure in antibiotic consumption studies can be evaluated only when the unit of measure takes into account the potency and formulation of drugs (Chauvin et al., 2001). The WHO has recommended the use of the anatomical therapeutic chemical classification system (ATC) and defined daily dose system for international human drug usage studies (WHO, 2001; WHO Collaborating Center for Drug Statistics Methology, 2002a,b). None of the alternatives (minimum marketed dose, the equipotential dose, and the PDD) appears to offer any advantage over the DDD (Chauvin et al., 2001; Merlo et al., 1996). The PDD, not being a standard unit, can be used appropriately in a second step to explain differences detected by the DDD (Merlo et al., 1996). The PDD can be determined by prescription analysis of a defined population and the PDD is likely to vary between geographical regions and different settings. 2.1. DDD—principles The DDD is the assumed average maintenance dose per day for a drug used for its main indication in adults (when referred to body weight, an adult person at 70 kg). The DDD is nearly always a compromise based on a review of available information: recommended dosages (drug catalogues, published in scientific journals or major international textbooks); data on PDDs used in various countries (when available); an established main indication; and other DDDs within the same chemical subgroup. The DDD is often identical for various dosage forms of the same drug, though different DDDs may be established when the bioavailability is substantially different for various routes of administration or use for different indications. Three years after marketing, the DDD is revised. Otherwise, changes usually are not made unless the difference is at least on the order of 50%. However, minor changes are allowed for very-important drugs (Anonymous, 2002). 2.2. Adjustments of the defined daily dose It is important to maintain stable DDDs over time to allow trends in drug consumption to be studied, without the complication from frequent changes. The PDD most likely differs between countries because the main indication and antimicrobial sensitivity cab vary between countries. In veterinary medicine, there is a large variation in prescription practices and changing the ADD frequently would create a very confusing picture. Unless major changes occur on the national level (for the national ADD), a stable ADD should be preserved as a means of communication. Known discrepancies between the PDD and the ADD should be taken into account when interpreting consumption figures on a regional basis or in research. 2.3. Advantages and limitations of DDD and considerations for a veterinary equivalent The ATC/DDD allows standardization of drugs by group and a stable age measure of drug use—allowing comparison of drug use between countries, regions and health care units. The number of DDDs per inhabitant per year is an estimate of the number of days each inhabitant

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

205

is treated annually. The measure gives an improved basis (over weight-of-active-compound) for tracing drug consumption over extended periods and enables comparison of drug usage between populations and different settings (Wessling and Boethius, 1990). Neither the total weight nor the DDD reflect the variability in PDD. Knowledge about the PDD:DDD ratio enables adjustment of DDD for a geographical region or setting in individual studies (Harris et al., 1994). If the doses used over time are increasing—without changing the DDD—the total consumption expressed in DDD/1000 inhabitants will increase. For most drugs, the differences between the PDD and the DDD is minimal when measured at a national level—but differences in PDD can be observed between prescriptions and between countries (Harris et al., 1994). The DDD should not be used to estimate the prevalence of drug use (number of medicated patients) because of individual differences in dosage regimens related to sex, age, genetic, or metabolic/renal variation (Wessling and Boëtius, 1990; Mantel-Teeuwisse et al., 2001; Zagorski et al., 2002). Together with knowledge of average duration of treatment, the PDD is a better indicator for the prevalence of drug use (Bro and Macke, 1986). However, obtaining the PDD is usually difficult and duration of treatment rarely can be determined; even dosage regimens from prescription data are not the true story of dosing (Tamblyn et al., 1995). Changes in drugs during a treatment course will affect the estimated number of patients—in particular, for drugs that are used in short-term treatments (such as antimicrobials). The number of treatment courses can be estimated only if the duration of treatment is known. Combined analysis of duration of the treatment course and the DDDs used can offer more-reliable information about exposure to antimicrobial agents in humans (Resi et al., 2001). The concept of treatment courses and ADDs has been used in a study analyzing the relationship between disease occurrence (mastitis) and antimicrobial treatment courses in dairy cattle (Hill et al., 2004). However, the concept is not applicable to large parts of the veterinary field, because group medication in fast-growing animals often is used—and variation in dosing and duration of treatment is high (Chauvin et al., 2002). In humans, selective pressure imposed by antibiotic use can be measured in DDD per 1000 inhabitants or per 100 bed-days (Hekster et al., 1982). In food-animal production (where body weight increases rapidly and the treatment rate is high), ADD is unlikely to be a good measure of the proportion of animals treated. The average ADD per animal should be used as a measure of a selective pressure imposed on the herd (herd-exposure), including both animals and the close environment—rather than as selective pressure on a number of individuals treated. Because the treatment dose and the duration of treatment are not inversely related and can vary considerably between practitioners and herds and over time (Chauvin et al., 2002), herd-exposure might be an alternative indicator to the number of animals treated.

3. Standardization of veterinary drug use—the VetStat approach 3.1. Defined daily doses for animals The application of a defined daily dosage in veterinary medicine presents difficulties from metabolic differences among species, the wide range in animal body weights, and the fact that most of drugs are used in growing animals (Chauvin et al., 2001; Nicholls et al., 2001).

206

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

Table 2 Comparison of the PDD and the ADD Deviation of PDD from ADD (%)

Number of antibacterial drug groups Swine No.

Cattle %

No.

Poultry %

No.

%

≤10 10–25 >30

21 2 3

80 8 12

26 11 3

65 28 7

5 0 0

100

Total

26

100

40

100

5

100

The PDDs were provided by a group of specialized veterinary practitioners (see text).

With the introduction of VetStat, it has been become possible to implement a defined daily dose system for animals, because the target animal species and the indication (target organ) for treatment for each package are known. The VetStat approach has been to develop ADDs for each species, taking into consideration the animal weights. For each pharmaceutical product, an ADD is calculated for each of the species it has been approved for and other species for which it is prescribed according to the VetStat records. In the case of human drugs, the ADD is determined by extrapolation from recommendations for use of related generic compounds used both in human and animals, supported by information from the veterinary-pharmacology literature. ADDs also are defined for commonly used extemporaneously prepared medicinal products. (For most topical preparations, general and local analgesics, anaesthetics, neuroleptics, and contrast media, an ADD has not been established.) The ADD is defined as the average maintenance dose for the main indication in a specified species. As a starting point, the ADD is calculated as the median value of the recommended dosage range multiplied by the frequency per day. The range and treatment frequency are taken from the recommendations approved by the Danish Medicine Agency and the recommended doses published by the Veterinary Pharmaceutical Producer Association (VIF) (Anonymous, 2003). The recommendations published by VIF are widely used as a medication guide book in veterinary practice. We verified that the recommendations are in fact representative of regimens used by veterinary practitioners by obtaining information about medication practices used by major cattle, pig and poultry practitioners. To verify this, information on the PDD was requested from a group of specialized veterinary practitioners representing the associations of veterinarians specialized in porcine and bovine diseases, and from the two poultry practices responsible for 56% of the antimicrobial consumption in Danish poultry in 2002. With respect to cattle, the practitioner provided doses for 70 antibacterial products that were divided in 40 groups of similar products (i.e. 40 combinations of formulation, administration route and active chemical compound). For 65% of these groups, the PDD deviated ≤10% from the defined ADD (Table 2); the PDD for 93% (37) of the drug groups were in accordance with the recommendations. The dose of only three drug groups deviated >30% from the ADD, exceeding the recommended dose range. These two groups comprise some, but not all, ␤-lactamase-sensitive penicillin- and streptomycin-based drugs used in cattle.

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

207

The swine practitioner provided doses for 51 antibacterial drugs, including 26 groups of products. For all groups, the PDDs were within the recommended range; in 80% of the drug groups, the PDD deviated ≤10% from the ADD (Table 2). In one case (chlortetracycline for oral use), the defined ADD varied >100% from the dose used by the swine practitioner. Prescriptions from two other swine practices were collected at the pharmacies, revealing that the largest applied dose was three times higher than the lowest dose. A special characteristic of this product was that the average recommended dose for this product varied by 100% depending on the organ treated and the doses used were in accordance with the recommendations. For poultry, the doses provided by the practitioners where identical with the ADD for all drug groups used (six products, five groups). When calculating the doses used in practice, rounding may easily cause 10% deviation of doses. Assuming that up to 10% deviation is due to rounding, the doses used in this pilot study equal the defined ADD for 65–100% the antibacterial drugs used in three major production-animal species in Denmark. This pilot study suggests that the Danish veterinary practitioners use doses within the recommended range and there is not a general trend to use either higher or lower doses within the range. This was in accordance with the prescription pattern found in a French survey of group-level antibiotic prescriptions in pig production—showing that the PDD were in the range of dosages recommended in Europe while there was a high variation in duration of treatment (Chauvin et al., 2002). The PDD for tylosin was the only deviation from the recommended dose (10–20 mg/kg) in the French study, but the PDD (4.8–7.4 mg/kg (95% CI)) used in France was close to the Danish ADD (4 or 7.5 mg/kg, depending on the formulation and chemical salt). The doses used by the swine and cattle practitioners were not always consistent for similar products because the recommendations can vary between similar products and because doses in either end of the range may be used. Where similar products have been marketed with different dosage recommendations, a single dose for the active ingredient and formulation is defined (taking into consideration also the potency and bioavailability of other drugs within the same chemical subgroup, i.e. ATC group level 4). The ADD is usually defined per kg bodyweight (ADDkg ); subsequently, an ADD is calculated by multiplication with a defined standard animal bodyweight for each the age-group (Table 1). For one-dose formulations (usually formulated for a specific age-group but independent of bodyweight), only the ADD is defined. The standard body weights for pigs, poultry, and cattle were defined in consultation with the group of specialized practitioners described above. For goats, sheep and mink, researchers with close contact to the production field were consulted. The “standard weight” was set as the estimated average weight at treatment for each age-group, to make recalculation of ADDs easier. For poultry and aquaculture, the “standard weight” was set at 1 kg (because the growth rate is extremely high and in these food production systems, flock treatment based on the population body mass always is used). The defined standard weight could never be the true average of animal weights at treatment, because this will vary depending on the drug. As for the human DDD, the standard weights rather are arbitrary measures, enabling a crude measure of the average number of animals treated. Measuring drug usage in ADDs for “standard animals”, always leaves an opportunity for recalculating the figures to the more precise kg-animal-treated (ADDkg ) or using other defined body weights when relevant for specific purposes.

208

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

3.2. Standardization according to the size of target population To evaluate the drug usage and to produce comparable statistics, the usage needs to be adjusted for the size of the population at risk. At the national level, the size of the population can be inferred from the production figures. Herd-size information must be collected from existing registers or from the individual farmer or possibly veterinary practitioner. In Denmark, the herd-size by species and age-group on the farm may be collected from the central husbandry register (CHR). Like the VetStat database, the CHR is a relational database on an Oracle platform and is part of the so-called GLR/CHR register (operated on behalf of the Ministry of Food, Agriculture and Fisheries). All farms legally are required to be registered in the database. The validity of basic information used for calculating taxes and subsidies (such as postal-, geographic- and area-information) is high, and the information is up-dated weekly from local-authority registers and other-authority registers. The farmer may report changes on herd-size at his own initiative at any time, and annually, a form for reporting current herd-size is send to the farmers. Submission of updates is mandatory when a major change in number of animals has occurred. However, in many cases, the herd-size is not updated as required. From October 2002, reporting the transfer of goat, sheep, and swine to and from farms is compulsory. This information will be used in logical validation of species, age group, and herd-size information held by the CHR, and is expected to improve the data-quality for these species. The updating procedure for cattle differs, in that all movements of cattle are registered and the number of animals registered on the CHR-ID is updated every fortnight from data in the Danish Cattle Database (CDB). The individual head of cattle is identified by a unique individual animal number (CKR-ID on ear-tags) and associated with a CHR-ID. The validity of the data are high because the CKR-ID and CHR-ID are used in the milk-quality control (90% of the cows are tested 6–11 times annually) and cattle cannot be slaughtered for food production if the CHR-ID and CKR-ID are not in accordance with the CDB information. For all species, historical data on animal number and dates are stored in relational tables in the CHR database. A general validation of the number of animals registered for the non-bovine species has not been conducted. In research studies, the number of animals registered must be validated using other sources such as: the register of swine, sheep, and goat transfer (part of the CHR/GLR); information on broiler and egg production; number of pigs slaughtered (13-week periods) by each CHR-ID (registered by the Danish Bacon and Meat Council); or the productivity databases for poultry and pigs containing extensive information on a limited number of herds. The data on herd-size, species and age-group are entered electronically into the VetStat database to produce individual search-based reports on drug use on the farm level. The number of ADDs used for a specific animal species and age-group, as specified in the VetStat records, is compared to a weighted average of herd-size (including all up-dates of herd-size on the CHR-ID within the chosen time-period of interest). 4. Validation of VetStat data In Denmark, virtually all therapeutic drugs are prescription-only, and 95% of the overall sales of drugs prescribed for animal species is sold through the pharmacies. By linking the

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

209

data entry inevitably to billing and electronic transfer to VetStat, the risk of human errors is minimized (ensuring the validity of identity and quantity of the drug). A few substances are approved for use in medicated feed on prescription, and the usage is reported directly from the feed mills to VetStat. Data from the feed mills are subject to logical validation of the codes used at data entry; that is, only valid commodity numbers, units, disease category, animal species, and age-groups may be entered. The total national annual usage (data from feed mills and pharmacies) of antimicrobial drugs has been validated against the wholesalers’ statistics. A deviation of 1.2% in kg active compound was caused by erroneous reporting of a wholesaler’s export—confirming that the VetStat data are reliable (DANMAP, 2002). Most practices report through their financing system in connection with billing. Major technical difficulties in the complex system3 for transferring data from the veterinary practitioners have been overcome within 2002–2003, and a system enabling the practitioner to correct erroneous data records detected by a full logical validation was introduced in January 2003. The system validates whether: • • • •

the commodity number exists on the commodity list and the unit of drug is correct; CHR-ID is registered in the CHR database; the combination of animal species, age-group and disease category4 is valid; and the identity-codes of the veterinarian and the practice are registered by the veterinary authorities.

In relation to research studies and national statistics, the amount of drugs reported by the veterinary practitioners is validated against pharmacy data on the total sales of therapeutic drugs for use in practice. Frequently, product registration includes both companion animals and production animal species. Validation of the amounts of drugs used in mixed practices is restricted to drugs that are entirely used in production animals, because drugs used in pet animals only are reported from the pharmacies (Stege et al., 2003). Due to legal regulation of prescription practices, a large part of drugs used in cows (>90%) are used by the practitioner—while the majority of drugs used in poultry (66%) and veal calves (∼80%), and nearly all drugs used in pigs (>98%) are sold directly by the pharmacies or by the feed mills to the farm (DANMAP, 2002). Less than 10 practitioners are responsible for treating production poultry, and errors in their reporting were corrected in 2002. Due to the correction of technical errors and the new electronic validation and correction system, the validity on data from the practitioners are expected to improve considerably in 2003 (affecting primarily the validity of cattle-usage statistics). 5. Potential application of standardized measures Due to dosage variations between animal species, the evaluation of consumption of drugs by means of an ADD depends on knowledge of target species. Data on consumption should 3 The veterinarian practices use different software packages and companies to report the data in connection with billing. Fewer veterinarians report directly to VetStat—either on the VetStat web page or by up-load of discs sent to VetStat. The use of different software companies has given rise to many of technical errors (in particular incorrect coding of package size) affecting the drug amounts. 4 The age-groups are species specific. The disease groups are partially species specific (Stege et al., 2003).

210

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

be collected for each individual medicinal preparation. When analyzing data using the ADD, it might be important to include route of administration as a co-variate, because the dose in mg/kg often differ. The main indication and PDD can vary between countries, but the Danish ADD system can be used in other countries—while correcting for known differences between the local PDD and the ADD. In most countries, national consumption data on drugs that are formulated and used in only one species, can be used for drug statistics based on a DDD for animals (Grave et al., 1999). Changing the measure from weight of active compound might importantly change the picture of importance of different classes of compounds. As an example, the annual usage of antimicrobials used for oral therapy and for growth promotion in pigs is presented in kg active compound and ADDs (Fig. 1). In this example, the ADDkg is used because all age-groups are included. The statistic comprises only antimicrobial drugs that are almost exclusively used for oral treatment in pig (comprising 87% (44 t/51 t) of the oral antimicrobial drug use in pigs in 2001). Tylosin used as growth promoter is measured in ADD for comparison, even though the dose used for growth promotion was considerably lower (20–40 ppm) than the therapeutic dose (100 ppm). Other growth promoters are not included in Fig. 1, because ADDs are not defined for these substances. Tylosin was the major growth promoter used in Denmark before the ban of growth promoters (DANMAP, 2002). After the growth-promoter ban in 1998–1999, the total antimicrobial consumption decreased drastically, although an increase in therapeutic use was observed in the following years (1998–2001). Using the ADDkg instead of kg-active compound, the decrease in antimicrobial use is far more prominent (due to the low therapeutic dose of macrolides, as compared to other therapeutic drugs, such as tetracyclines). The ADD also may be used in research including data on drug consumption and animal species on the farm level: In a study including 120 farrow-to-finish farms, drug-consumption data were provided monthly by the practitioners. Expressing the drug consumption in ADD, the effect of the AGP ban on the monthly total consumption of therapeutic antimicrobials could be followed in the year before and after the ban (Larsen, 2002). 5.1. Use of standardized statistics in VetStat 5.1.1. Online-statistics Individual statistics for herds or veterinary practitioners (combining information in VetStat and the CHR) are available as interactive reports through secure connections via the Internet. The farmers and veterinary practitioners have access to statistics based on their “own” data, whereas the authorities have unlimited access for control purposes. The authorities may use the system to point out farms with deviating patterns of drug usage (possible use contrary to regulations), and to identify and inform the prescribing practitioner. Individual standardized usage statistics are calculated for any sub-population indicated by the user-selected criteria, i.e., herd identity (one or groups of farms), period, species, age-group, veterinarian (optional), veterinary practice (optional), ATC-group, disease categories (optional) or geographic region (optional). For comparative purposes, the unit of the standardized output is %ADD per animal per day, i.e. the average proportion of animals treated per day. Similar statistics on a national level are available for comparison, while correcting the number of animals on the national level (not necessary for

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

211

Fig. 1. Consumption of antimicrobial growth promoter and oral therapeutics used for pigs in Denmark. Data from 1996 onwards are from usage statistics from the Danish Medicines Agency (1996). Data before 1996 were collated by the Danish Bacon and Meat Council (Rønn and Jacobsen, 1995). Data on antimicrobial growth-promoter (AGP) were collated by the Danish Plant Directorate. Only therapeutics that are almost exclusively used for oral treatment of pigs are included. AGPs other than Tylosin are not included (see text). (a) Antimicrobials measured in kg active compound. (b) Antimicrobials measured in ADDkg . ADDs used: Tylosin: 4 mg/kg; Lincomycin/spectinomycin and Lincomycin: 10 mg/kg; Neomycin: 15 mg/kg; Tiamulin and Valnemulin: 6.5 mg/kg; Tilmicosin: 16 mg/kg.

212

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

cattle). The VetStat online-system works as a motivating factor for up-dating the CHR information because the statistics on farm-level are sensitive to the data quality of the CHR. 5.1.2. National statistics National data on antimicrobial usage (in kg active compound) have been collected to assist in the analyses of the resistance surveillance data in Denmark and presented in the DANMAP reports since 1996. In the recent DANMAP reports, the antimicrobial-consumption data were based on VetStat data and the standardized doses were introduced (DANMAP, 2001, 2002). Statistics applying the defined ADD reflects the body mass treated, and may be compared with the number of animals or kg meat produced nationally. The detailed standardized utilization data are expected to make an important contribution to the analysis and estimation of the relative importance of different antimicrobials for generating resistance and to the identification of potential risk factors. 5.1.3. Research The VetStat data provide information for research in specific use conditions that govern selection and dissemination of resistant bacteria and for risk assessment of resistance. The complexity of dealing with a large number of active compounds in epidemiologic studies is reduced with the ADD. The detailed data allow subdivision of drug use on animal species, animal-production classes, and disease categories and allows modeling of associations between usage and resistance. The modeling could include spatial characteristics, potential seasonal influences, administration route and data on farm-management practices (such as range condition and vaccine strategies). An ADD has been assigned to most drugs used in the veterinary field. This enables pharmaco-epidemiologic research with a direct comparison between usage of antimicrobials and other medicines on farm level.

5.2. Limitations to the use of ADDs When a drug is licensed, a recommended dose is set for the species and indication based on clinical trials—but the dose is often labile due to decreasing bacterial sensitivity. Dose should be defined based on updated knowledge of pharmaco-kinetic and pharmaco-dynamics of the drug, but such information often is not available. The recommended doses have been used as a starting point for calculation of dose, because the recommendations are presumed to have an important impact on the dose used by the practitioner. The decision to treat a specific animal or group of animals most often is made by the animal caretaker. The dosage recommendations might not always be used—either due to lability of the therapeutic dose, the use of sub-therapeutic doses, or because the clinically effective dose might be lower than the dose for obtaining bacteriological cure (Craig, 2001). The ADD system provides a fixed unit of measurement independent of formulation and potency, but will give only a statistical measure for comparing usage and not an exact picture of the actual use (e.g. dose schedules used in the field). Information on the actual dose rate and the length of treatment is desirable for pharmaco-epidemiological research on antimicrobial resistance, but this only would be obtainable directly from the farmer.

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

213

6. Conclusion Standardized statistics are necessary for international comparison of usage, monitoring of national drug usage and for comparative studies of drug use in subpopulations. A measure taking into account the pharmacological activity provides a better measure of the total selection pressure applied to a particular environment under study. The defined animal daily dose is a better measure than the weight of active compound, but interpretation of trends in drug consumption also should include current knowledge of changes in the prescribed daily doses. To enable direct comparison of usage, the number of animals in the target population should be used as a baseline. Depending on the scope, information on number of animals may be obtained from different sources (such as animal-production data, existing registers or direct contact with the individual farmers). Due to the detailed data in VetStat, standardized statistics of drug use on a national, regional or herd-level are available. Improvement of the CHR data on herd size for all species but cattle is necessary, to fully use the data on drug usage on herd-level. Application of standardized statistics will improve the basis for research and the ability to respond to problems of antimicrobial resistance in a targeted way and to evaluate the effect of informative or regulative measures taken to ensure prudent use. The national ADD is seen by us as a predecessor to a future international DDD for animals. The ADD system will be adjusted to international standards. When implementing an international DDD, compromises must be made regarding differences in dose and main indications between countries.

References Aarestrup, F.M., Rasmussen, S.R., Arthursson, K., Jensen, N.E., 1998. Trends in the resistance to antimicrobial agents of Streptococcus suis isolates from Denmark and Sweden. Vet. Microbiol. 63, 71–80. Aarestrup, F.M., Agersoe, Y., Gerner-Schmidt, P., Madsen, M., Jensen, L.B., 2000. Comparison of antimicrobial resistance phenotypes and resistance genes in Enterococcus faecalis and Enterococcus faecium from humans in the community, broilers, and pigs in Denmark. Diagn. Microbiol. Infect. Dis. 37, 127–137. Anonymous, 2001. The Microbial Threat. Progress Report on Antimicrobial Resistance. Visby, Sweden, June 13–15, 2001. Art. No. 2001-123-68. ISBN 91-7201-586-1. Anonymous, 2002. Guidelines for the STC classification and DDD assignment. WHO collaborating Center for Drug Statistics Methology. ISSN 0802-8753, ISBN 82-90312-38-5. Anonymous, 2003. Veterinærmedicinsk Productkatalog 2003. Veterinærmedicinsk Industriforening. ISSN 1395-7430. http://www.vif.dk. Arnold, S., Gassner, B., Giger, T., Zwahlen, R., 2004. Banning antimicrobial growth promoters in feedstuffs does not result in increased therapeutic use of antibiotics in medicated feed in pig farming. Pharmacol. Drug Safety 13, 323–331. Bager, F., Madsen, M., Christensen, J., Aarestrup, F.M., 1997. Avoparcin used as a growth promoter is associated with the occurrence of vancomycin-resistant Enterococcus faecium on Danish poultry and pig farms. Prev. Vet. Med. 31 (1/2), 97–112. Bro, F., Macke, C.E., 1986. Use of antibiotics in general practice in Denmark. Scand. J. Prim. Health Care 4, 101–104. Chauvin, C., Madec, F., Guillemot, D., Sanders, P., 2001. The crucial question of standardization when measuring drug consumption. Vet. Res. 32, 533–543. Chauvin, D., Boloeil, P.A., Orand, J.P., Sanders, P., Madec, F., 2002. A survey of group-level antibiotic prescriptions in pig production in France. Prev. Vet. Med. 55, 109–120.

214

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

Craig, W.A., 2001. Does the dose matter? Clin. Infect. Dis. 33 (Suppl. 3), S233. Cruchaga, S., Echeita, A., Aladuena, A., Garcia-Pena, J., Frias, N., Usera, M.A., 1998. Antimicrobial resistance in salmonellae from humans, food, and animals in Spain in 1998. J. Antimicrob. Chemother. 47, 315–321. Danish Medicines Agency, 1996. http://www.laegemiddelstyrelsen.dk/en/products/pp/numbers.asp. DANMAP 2001. Use of antimicrobial agents and occurrence of antimicrobial resistance in bacteria from food animals, foods and humans in Denmark. July 2002. ISSN 1600-2032. http://www.dfvf.dk/Files/Filer/ Zoonosecentret/Publikationer/Danmap/Danmap 2001.pdf. DANMAP, 2002. Use of antimicrobial agents and occurrence of antimicrobial resistance in bacteria from food animals, foods and humans in Denmark. July 2003. ISSN 1600-2032. http://www.dfvf.dk/Files/Filer/ Zoonosecentret/Publikationer/Danmap/Danmap 2002.pdf. Dunlop, R.H., McEwan, S.A., Meek, A.H., Clarke, R.C., Black, W.D., Friendship, R.M., 1998. Associations among antimicrobial drug treatments and antimicrobial resistance of fecal Eschericia coli of swine on 34 farrow-to-finish farms in Ontario, Canada. Prev. Vet. Med. 34, 283–305. Grave, K., Greko, C., Nilsson, L., Odensvik, K., Mork, T., Rønning, M., 1999. The usage of veterinary antibacterial drugs for mastitis in Cattle in Norway and Sweden during 1990–1997. Prev. Vet. Med. 42, 45–55. Harris, C.M., Cullen, J.W., Roberts, D.J., 1994. Consider national standards for daily dosages. BMJ 308, 207. Hekster, Y.A., Vree, T.B., Goris, R.J., Boerema, J.B., 1982. The defined daily dose per 100 bed-days as a unit of comparison and a parameter for studying antimicrobial drug use in a university hospital. A retrospective study of the effects of guidelines and audit on antimicrobial drug use. J. Clin. Hosp. Pharm. 7, 251–260. Hill, A., Chriel, M., Jensen, V.F., Vaarst, M., Stockmarr, A., Bruun, J., Greiner, M., 2004. Use of existing surveillance data to detect welfare problems in Danish cattle. An evaluation of available data sources, with detection of existing herd health problems and associated risk factors, in preparation. Project report, Theme 6, International EpiLab, 22 March 2004. (contact www.dfvf.dk/EpiLab). Larsen, P.B., 2002. Consequences of termination of AMG use for pig health and usage of antimicrobials for therapy and prophylaxis. In: Working Papers for the WHO International Review Panels Evaluation: Impacts of Antimicrobial Growth Promoter Termination in Denmark. http://www.dfvf.dk/Files/Filer/ Zoonosecentret/Publikationer/WHO%20Rapport/Working papers Report - final.pdf. Low, J.C., Angus, M., Hopkins, G., Munro, D., Rankin, S.C., 1997. Antimicrobial resistance of Salmonella enterica typhimurium DT104 isolates and investigation of strains with transferable apramycin resistance. Epidemiol. Infect. 188 (2), 97–103. Mantel-Teeuwisse, A.K., Klungel, O.H., Verschuren, W.M.M., Porsius, A., de Boer, A., 2001. Comparison of different methods to estimate prevalence of drug use by using pharmacy records. J. Clin. Epidimiol. 54, 1181–1186. Merlo, J., Wessling, A., Melander, A., 1996. Comparison of dose standard units for drug utilization studies. Eur. J. Clin. Pharmacol. 50, 27–30. Nicholls, T., Acar, J., Anthony, F., Franklin, A., Gupta, R., Tamura, Y., Thompson, S., Threlfall, E.J., Vose, D., van Vuuren, M., White, D.G., Wegener, H.C., Costarrica, M.L., 2001. Antimicrobial resistance: monitoring the quantities of antimicrobials used in animal husbandry. Rev. Sci. Tech. Off. Int. Epiz. 20, 841–847. Resi, D., Castelvetri, C., Vaccheri, A., Montanaro, N., 2001. The therapeutic course as a measure complementary to defined daily doses when studying exposure to antibacterial agents. Eur. J. Clin. Pharmacol. 57, 177–180. Rønn, N.E., Jacobsen, E., 1995. In: Rønn, N.E. (Ed.), Use of Antibiotics in the Pig Production. The Danish Bacon and Meat Council, November 1995 (Vurdering af antibiotikaforbruget i dansk svineproduktion) (In Danish). Rosdahl, V.T., Pedersen, K.B. (Eds.), 1998. The Copenhagen Recommendations. Report from the Invitational EU Conference on the Microbial Threat, 9–10 September 1998, Copenhagen, Denmark. Stege, H., Bager, F., Jacobsen, E., Thougaard, A., 2003. VETSTAT—the Danish surveillance system of the veterinary use of drugs for production animals. Prev. Vet. Med. 57, 105–115. Tamblyn, R., Lavoie, G., Petrella, L., Monette, J., 1995. The use of prescription claims database in pharmacoepidemiological research: the accuracy and comprehensiveness of the prescription claims database in Quebec. J. Clin. Epidemiol. 48, 999–1009. Tollefson, L., Miller, M.A., 2000. Antibiotic use in food animals: controlling the human health impact. J. ADAC Int. 83, 245–254. van den Boegaard, A.E., Stobberingh, E.E., 2000. Epidemiology of resistance to antibiotics. Links between animals and humans. Int. J. Antimicrob. Agents 14, 327–335.

V.F. Jensen et al. / Preventive Veterinary Medicine 64 (2004) 201–215

215

Wegener, H.C., Aarestrup, F.M., Gerner-Schmidt, P., Bager, F., 1999. Transfer of antibiotic resistant bacteria from animals to man. Acta Vet. Scan. Suppl. 92, 51–57. Wessling, A., Boethius, G., 1990. Measurement of drug use in a defined population. Evaluation of the defined close (DDD) methodology. Eur. J. Clin. Pharmacol. 39, 207–210. WHO, 1997. The medical impact of the use of antimicrobials in food animals. In: Report and Proceedings of a WHO Meeting, Berlin, Germany, 13–17 October 1997, WHO/EMC/ZOO/97.4. WHO, 1998. Use of quinolones in food animals and potential impact on human health. In: Report and Proceedings of a WHO Meeting, Geneva, Switzerland, 2–5 June 1998, WHO/EMC/ZDI/98.12. WHO, 2000. WHO global principles for the containment of antimicrobial resistance in animals intended for food. In: Report of a WHO Consultation, Geneva, Switzerland, 5–9 June 2000, WHO/CDS/CDR/APH/2000, pp. 1–21. WHO, 2001. Monitoring antimicrobial usage in food animals for the protection of human health. In: Report of a WHO Consultation, Oslo, Norway, 10–13 September 2001, WHO/CDS/CSR/EPH/2002.11. WHO Collaborating Center for Drug Statistics Methology, 2002a. Guidelines for ATC Classification and DDD Assignment, 5th ed. Oslo, Norway. WHO Collaborating Center for Drug Statistics Methology, 2002b. Guidelines for ATCvet Classification, 4th ed. Oslo, Norway. Zagorski, B.M., Trick, W.E., Schwartz, D.N., Wisniewski, M.F., Hershow, R.C., Fridkin, S.K., Weinstein, R.A., 2002. The effect of renal dysfunction on antimicrobial use measurements. Clin. Infect. Dis. 35 (12), 1491–1497.