Quantification and evaluation of antimicrobial drug use in group treatments for fattening pigs in Belgium

Preventive Veterinary Medicine 74 (2006) 251 263 www.elsevier.com/locate/prevetmed Quantification and evaluation of antimicrobial drug use in group treatments for fattening pigs in Belgium Tom Timmerman *, Jeroen Dewulf, Boudewijn Catry, Bianca Feyen, Geert Opsomer, Aart de Kruif, Dominiek Maes Department of Reproduction, Obstetrics and Herd Health, Veterinary Epidemiology Unit, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, B-9820 Merelbeke, Belgium Received 3 December 2004; received in revised form 25 September 2005; accepted 4 October 2005 Abstract To control the emergence of antimicrobial resistance, knowledge of antimicrobial drug consumption is essential. Because consumption data are not available in Belgium, a study was conducted between March and October 2003 to investigate the antimicrobial drug consumption in pigs, using the treatment incidence based on the animal daily dose pig (ADDpig), the treatment incidence based on the used daily dose pig (UDDpig) (number of ADDpig or UDDpig/1000 pigs at risk/day), and the ratio UDDpig/ADDpig. The sampling frame consisted of 821 pig herds that (a) used a closed or semi-closed production system, (b) were located in the most dense pig areas of Belgium, and (c) had at least 150 sows and 600 fattening pigs each. Of 50 randomly selected herds, all group treatments with antimicrobial drugs, applied to fattening pigs that were within 2 weeks of slaughter (median age 187 days), were collected retrospectively. The treatment incidence based on ADDpig for all oral and injectable antimicrobial drugs was 178.1 per 1000 pigs at risk per day. The treatment incidence based on UDDpig shows that in reality fewer pigs were treated, namely 170.3 per 1000 pigs at risk per day. Proportionally, the most often applied oral antimicrobial drugs were: doxycycline, amoxicillin, combination trimethoprim-sulphonamides and polymyxin E. The most often applied injectable antimicrobial drugs were long-acting amoxicillin and ceftiofur. The distribution of the UDDpig/ADDpig ratio per antimicrobial drug * Corresponding author. Tel.: +32 9 264 75 40; fax: +32 9 264 77 97. E-mail address: Tom.Timmerman@UGent.be (T. Timmerman). 0167-5877/$ see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.prevetmed.2005.10.003

252 T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 shows that 50 75% of the oral formulations were underdosed. Injectable formulations were almost always overdosed (>90%). # 2005 Elsevier B.V. All rights reserved. Keywords: Antimicrobial drug consumption; Swine; UDD; ADD; Group treatment 1. Introduction The intensification of pig production has caused the need for strict sanitary health measures, but the use of antimicrobial drugs remains inevitable (Mateu and Martin, 2001; Mevius et al., 1999; Schwarz et al., 2001). The use of antimicrobial drugs in foodproducing animals is considered to contribute to the emergence of antimicrobial resistance both in veterinary and human medicine (Aarestrup, 1999; Gyssens, 2001). Contamination of carcasses during slaughter can be a source of resistant bacteria to humans (Catry et al., 2003; McKellar, 1998; Schwarz et al., 2001). Because pork is the most consumed meat in Europe (45.9% of the total meat consumption) (EUROSTAT, 2003), possible transfer of resistant bacteria via pig meat is almost inevitable. Therefore, control of antimicrobial resistance in pigs should be a priority. A prerequisite for controlling antimicrobial resistance is a thorough knowledge of the antimicrobial drug consumption (Gyssens, 2001; Mevius et al., 1999; Nicholls et al., 2001). Detailed antimicrobial consumption data can be used to detect inappropriate usage, to identify underlying risk factors for the emergence of antimicrobial resistance, and to quantify the exerted selection pressure (Caprioli et al., 2000; Catry et al., 2003; Gyssens, 2001; Nicholls et al., 2001) from birth till slaughter age. We modified existing antimicrobial drug use parameters from human medicine to quantify the antimicrobial drug consumption in group treatments for fattening pigs (pigs raised for human consumption) that were within 2 weeks of slaughter. Additionally, we evaluated the appropriateness of dosing. 2. Materials and methods 2.1. Study sample and data collection Herds were obtained from the Belgian farm-animal identification and registration database (SANITEL, 2003). The sampling frame consisted of 821 herds that used a closed or semi-closed production system (purchase of breeding animals only). All herds were located in the most dense pig areas of Belgium (West- and East-Flanders, with 1030.8 and 399.0 pigs per km 2, respectively (SANITEL, 2003)), and had at least 150 sows and 600 fattening pigs each. We assigned a computer-generated random number (Excel 1, Microsoft Inc.) to all 821 herds and sorted the herds from low to high random number. Eighty-four herds were contacted by telephone, starting with the herd assigned to random number 1, to obtain 50 cooperative herds (response of 60%). Taking into account an expected treatment incidence

T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 253 Table 1 Characteristics of all pig herds in the sampling frame and of the 84 randomly selected herds asked about their interest in participating in an antimicrobial drug consumption study Study population Number of sows Number of fattening pigs (number of herds) Min Mean Median Max Min Mean Median Max Sampling frame (821) 150 217 200 1600 600 1252 1220 4000 Respondents (50) 150 216 200 380 600 1250 1200 2390 Non-respondents (34) 150 197 185 330 750 1187 1200 1966 of 150 ADDpig per 1000 pigs at risk per day, an accepted error of 10%, and a confidence level of 95%, at least 47 herds had to be sampled (Win Episcope 2.0). Of the non-respondents, 44% (15/34) were not interested in the study, 35% (12/34) were interested but had no time for the interview, 9% (3/34) had stopped their activities recently, 6% (2/34) had stress-sensitive animals and did not allow visitors, 3% (1/34) owned a highhealth pig farm and also did not allow visitors, 3% (1/34) had problems with Aujeszky s disease and did not want to cooperate. Herd sizes of the respondents did not differ from the herd sizes of those that declined to cooperate (Student s t-test; P = 0.46). Moreover, the herd sizes of the volunteers were comparable to those of the herds in the sampling frame (Student s t-test; P = 0.97) (Table 1). All herds were visited by the same investigator on a single occasion between March and October 2003. At the start of the visit, fattening pigs that were within a maximum of 2 weeks to slaughter (age: mean 186 days, median 187 days, min 168 days, max 210 days) and were housed in the same compartment (all-in/all-out) were selected as the on-farm study population (mean 200 pigs, median 150 pigs, min 52 pigs, max 800 pigs). All data pertaining to group treatments with antimicrobial drugs applied to these animals between birth and the time of the herd visit (hereafter called slaughter age ), were collected retrospectively (lifetime-to-date treatment histories). We defined group treatment as the prophylactic or metaphylactic administration of antimicrobial drugs to all animals of the same production stage. Prophylaxis implies the administration of antimicrobial drugs for preventive measures. Metaphylaxis is the application of antimicrobial drugs to groups of animals at times when only some animals of the group exhibit disease symptoms, but when it was expected that most of the group would become affected (Schwarz et al., 2001). The collected group treatment data pertained to: product name, number of treatments, primary intention of treatment (prophylaxis or metaphylaxis), time of treatment, dose, administration route, treatment duration, number of treated animals, indication and body weight. When the body weight was unknown, an estimate was made using a growth table (Straw et al., 1999). On 12 herds, we were able to validate the collected data by checking treatment records. 2.2. Estimate of drug consumption Antimicrobial drug consumption has been quantified in the literature in different ways: using financial units, commercial units or weight indicators (Chauvin et al., 2001). We used the weight indicators: animal daily dose (ADD) (Jensen et al., 2004) and used daily dose

254 T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 (UDD). The terms ADDpig and UDDpig were introduced because this study was based on data collected in pigs. The ADDpig and UDDpig were used to calculate three different parameters which express the antimicrobial drug use within the study population: (1) treatment incidence based on the ADDpig (TI ADDpig )(Jensen et al., 2004), (2) treatment incidence based on the UDDpig (TI UDDpig ) and (3) the UDDpig/ADDpig ratio. All antimicrobials were classified according to the anatomical therapeutic chemical veterinary (ATCvet) classification system (Harr et al., 2004; WHO, 2002). 2.2.1. Animal daily dose pig and treatment incidence based on ADDpig Internationally, the ADDpig is defined as the national defined average maintenance dose per day per kg pig of a drug used for its main indication (Jensen et al., 2004). ADDpig were estimated for the various antimicrobial preparations administered on the different herds. Estimates were based on the dose recommendations in the Belgian Compendium on Veterinary Medicines (Anonymus, 2003a) and in the publications of the Belgian Centre for Pharmacotherapeutic Information (Anonymus, 2003b). These two sources of information are regularly consulted by Belgian veterinarians. In agreement with defined daily dose (DDD) guidelines, the ADDpig for combination preparations were estimated for the main substance. Only for the combination trimethoprim-sulphonamides was the ADDpig calculated for the minor substance trimethoprim (Grave et al., 1999; WHO, 1996, 2002). Treatment incidences based on ADDpig were calculated using the formula in Table 2. Median values were used because our data were not fully normally distributed. The observation period (in days) was equal to the median age (days) of the fattening pigs at time of the herd visit. The number of pigs at risk in kilogram body weight was equal to the number of animals in the observed group multiplied by the median body weight at treatment. In the present study, an observation period of 187 days (mean 186 days, min 168 days, max 210 days) and a median body weight at treatment of 20 kg were calculated (mean 24 kg, min 1 kg, max 100 kg). 2.2.2. Used daily dose pig and treatment incidence based on UDDpig The UDDpig is defined as the administered dose per day per kilogram pig of a drug. UDDpig can differ between and within herds. When an antimicrobial was administered several times at different dosages within a herd, the UDDpig for that herd was calculated as the median value of the applied dosages. Treatment incidences based on UDDpig were calculated using a similar formula as that for TI ADDpig (Table 2). Table 2 Formulae for the calculation of the treatment incidence based on ADDpig (TI ADDpig ) and UDDpig (TI UDDpig ) Treatment incidence Formula Result TI Total amount of an antimicrobial drug used ðmgþ ADDpig ADDpig ðmg=kgþobservation period ðdaysþkg pigs at risk 1000 pigs at risk Number of ADDpig/1000 pigs at risk/day TI UDDpig Total amount of an antimicrobial drug used ðmgþ UDDpig ðmg=kgþobservation period ðdaysþkg pigs at risk 1000 pigs at risk Number of UDDpig/1000 pigs at risk/day

T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 255 The relative importance of each administered antimicrobial drug was expressed by the proportional TI ADDpig and the proportional TI UDDpig, which were calculated by dividing the TI ADDpig or the TI UDDpig of each antimicrobial by the total TI ADDpig or the total TI UDDpig, respectively (Grave et al., 1999). 2.2.3. The UDDpig/ADDpig ratio The ratio UDDpig/ADDpig was developed to give an indication of the appropriateness of dosing. It compares the administered dose (UDDpig) with the average maintenance dose (ADDpig). Assuming that correct dosing corresponds with the directives of the drug compendia or leaflets (in casu ADDpig), the ratio UDDpig/ADDpig is calculated to give an indication of under- and overdosing. An UDDpig/ADDpig ratio of 1 means the product was correctly dosed, while values below or above 1 are considered under or overdosed, respectively. 3. Results We registered antimicrobial drug use in 10,467 fattening pigs housed in 50 commercial pig herds, which had on average 216 sows and 1250 fattening pigs each (Table 1). Approximately 70 kg of antimicrobial drugs were administered during the entire life span of these fattening pigs (individual treatments not included). The majority of these antimicrobial drugs were administered orally (69 kg), while approximately 1 kg was injected. In 12% of the pig herds (6/50), animals were not treated in groups. Group treatments with only injectable formulations were used in five herds (10%), whereas only oral formulations were used in 13 herds (26%). Group treatments with both injectable and oral antimicrobial drugs were used in 26 herds (52%). When the farmers were asked about the primary intention of the group treatments, it appeared that 56% of all group treatments (90/161) were applied for prophylactic reasons, whereas 44% were applied for metaphylactic reasons. Group treatments with injectable antimicrobial drugs were mainly administered immediately after birth, at the time of castration (not only males were treated, but also females), and when diarrhoea occurred. Group treatments immediately after birth and those at the time of castration were mainly prophylactic treatments with broadspectrum penicillins and cephalosporines. Group treatments for diarrhoea were mainly metaphylactic, using fluoroquinolones and aminoglycosides. The main indications for oral group treatments were (post-weaning) diarrhoea, symptoms of arthritis and/or meningitis, and respiratory disease. Polymyxin E was administered with the intention to prevent postweaning diarrhoea. Symptoms of arthritis and/or meningitis were predominantly prevented with amoxicillin and mostly treated with combinations of trimethoprim-sulphonamides. Methaphylactic treatments for respiratory disease were mainly with tetracyclines and macrolides/lincosamides. Although structurally comparable herds (based on the selection criteria) were monitored, large differences in antimicrobial drug consumption were observed. Some herds were able to raise pigs without or with few group treatments, whereas more than 25%

256 T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 Table 3 Distribution of the overall treatment incidence based on ADDpig (TI ADDpig ) and UDDpig (TI UDDpig ), for all group treatments (oral and injectable) administered to fattening pigs between birth and slaughter age, housed in 50 Belgian closed or semi-closed pig herds between March and October 2003 Distribution of herds TI ADDpig TI UDDpig Minimum 0.0 0.0 5th percentile 0.0 0.0 10th percentile 0.0 0.0 25th percentile 10.3 21.8 50th percentile 82.6 87.5 75th percentile 196.1 147.3 90th percentile 361.4 211.6 Maximum 1369.3 1243.0 of the herds applied at least 80 standard doses (ADDpig or UDDpig) per 1000 pigs at risk per day (Table 3). One herd applied even more than 1000 standard doses (ADDpig or UDDpig) per 1000 pigs per day. TI ADDpig and TI UDDpig for the different oral and injectable antimicrobial drugs applied between birth and slaughter age are shown in Table 4. The TI ADDpig for all oral and injectable antimicrobial drugs was 178.1 per 1000 pigs at risk per day, which means that for a total of 1000 pigs, 178.1 pigs were treated daily with a standard dose of an antimicrobial drug. A comparison with the TI UDDpig shows that in reality less pigs were treated, namely 170.3 per 1000 pigs at risk per day. In Table 5, the distribution of the UDDpig/ADDpig ratio per antimicrobial drug is shown. Overall, 50 75% of the oral formulations were underdosed. Of the four most frequently used antimicrobials, doxycycline was overdosed in 50 75% of the cases. On the other hand, trimethoprim-sulphonamides were underdosed in 50 75% of the cases. Amoxicillin and polymyxin E were underdosed in 50 and 90% of the cases, respectively. Injectable formulations were almost always overdosed (>90%). Only the narrow spectrum penicillins were underdosed. 4. Discussion 4.1. Methodology Several sources (wholesalers, pharmacists, veterinarians, feed companies, and animal producers) can be used to obtain information on veterinary drug use, but they are not always equally accessible or accurate (Chauvin et al., 2002). Sales statistics from the national drug wholesalers could not be used in the present study because these are not recorded according to animal species, indication, dose and duration of treatment. At the time of the study, Belgian veterinarians, in contrast with those in other countries (Bager, 2000; Grave et al., 1991), were not obliged to make and record prescriptions for the dispension of antimicrobials, except for medicated feed produced and delivered by feed mills. Therefore, herd-specific data could only be obtained by on-farm registrations. These on-farm registrations were performed retrospectively on 50 randomly selected pig herds,

Table 4 Amounts of oral and injectable antimicrobial drugs administered as group treatments to fattening pigs between birth and slaughter age, housed in 50 Belgian closed or semiclosed pig herds between March and October 2003, expressed as animal daily doses or used daily doses per 1000 pigs at risk per day (ADDpig or UDDpig/1000pigs at risk/ day) ATCvet Active substance ADDpig UDDpig TI ADDpig a TI UDDpig a Min Median Max Min Median Max Min Median Max Prop. TI ADDpig % b Oral antimicrobial drugs QJ01A Doxycycline 9.6 1.0 17.6 41.3 0.0 50.2 340.4 0.0 27.4 285.2 28.8 16.2 QJ01C Amoxicillin 17.0 1.0 16.7 66.0 0.0 47.9 267.3 0.0 48.9 401.5 27.5 28.9 QJ01E Trimethoprim-sulphonamides 5.0 2.2 6.3 13.3 0.0 42.8 746.0 0.0 34.1 449.0 24.6 20.1 QJ01F Tylosin 7.5 0.6 2.8 5.5 0.0 1.3 97.7 0.0 3.5 133.0 0.7 2.0 QJ01F Tilmicosin c 12.0 9.8 0.0 4.8 92.7 0.0 5.9 113.3 2.8 3.5 QJ01F Lincomycin c 7.0 3.8 0.0 1.5 85.2 0.0 2.8 158.0 0.9 1.6 QJ01F Lincomycin spectinomycin 10.0 1.3 4.3 28.2 0.0 5.8 227.0 0.0 13.6 78.9 3.3 8.1 QJ01G Neomycin c 10.0 6.2 0.0 0.1 12.6 0.0 0.2 20.3 0.1 0.1 QJ01X Polymyxin E 6.3 0.3 3.8 32.4 0.0 19.8 314.7 0.0 32.8 418.5 11.4 19.4 Injectable antimicrobial drugs QJ01A Oxytetracycline c 10.0 35.7 0.0 <0.1 1.3 0.0 0.0 0.4 0.4 0.4 QJ01C Amoxicillin c 8.5 39.5 0.0 <0.1 2.2 0.0 0.0 0.5 1.0 0.8 QJ01E Amoxicillin long-acting 6.9 9.4 26.8 53.6 0.0 2.7 2.8 0.0 0.7 3.1 69.4 60.7 QJ01E Ampicillin c 7.5 21.5 0.0 <0.1 1.7 0.0 <0.1 0.4 0.7 0.5 QJ01E Ampicillin long-acting 7.5 17.9 35.7 25.2 0.0 <0.1 7.1 0.0 0.1 2.1 1.2 12.1 QJ01E Benzylpenicillin 5.0 1.8 2.7 3.6 0.0 <0.1 0.4 0.0 <0.1 0.4 0.1 0.9 QJ01E Procaine benzylpenicillin c 3.0 1.1 0.0 <0.1 0.1 0.0 <0.1 0.4 <0.1 0.2 QJ01D Ceftiofur 3.0 7.5 10.7 17.9 0.0 0.6 5.4 0.0 0.2 1.9 16.3 15.5 QJ01G Gentamicin c 3.5 14.3 0.0 0.1 1.6 0.0 <0.1 0.4 1.7 1.4 QJ01M Enrofloxacin 2.0 4.0 7.7 35.7 0.0 0.2 4.5 0.0 <0.1 1.6 4.5 4.0 QJ01M Marbofloxacin 2.5 5.4 8.6 10.0 0.0 0.1 5.5 0.0 <0.1 2.0 2.5 2.5 QJ01X Polymyxin E 3.0 7.1 19.9 10.0 0.0 0.1 2.0 0.0 <0.1 0.5 2.1 1.1 a Treatment incidences based on ADDpig or UDDpig. b Proportional TI ADDpig or TI UDDpig. c No minimum or maximum UDDpig available. These antimicrobials were only administered once or several times at the same dose. Prop. TI UDDpig % b T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 257

258 T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 Table 5 Distribution of the ratio UDDpig/ADDpig for all oral and injectable group treatments administered to fattening pigs housed in 50 Belgian closed or semi-closed pig herds between March and October 2003 ATCvet Active substance UDDpig/ADDpig a Min 5th 10th 25th 50th 75th 90th Max Oral antimicrobial drugs QJ01A Doxycycline 0.1 0.8 1.0 1.1 1.8 2.4 3.4 4.3 QJ01C Amoxicillin <0.1 0.2 0.3 0.6 1.0 1.3 2.3 3.9 QJ01E Trimethoprimsulphonamides 0.4 0.5 0.5 0.6 1.3 1.8 2.1 2.7 QJ01F Tylosin <0.1 0.1 0.1 0.2 0.4 0.6 0.7 0.7 QJ01F Tilmicosin b 0.8 QJ01F Lincomycin b 0.5 QJ01F Lincomycin 0.1 0.2 0.2 0.4 0.4 0.6 2.0 2.9 spectinomycin QJ01G Neomycin b 0.6 QJ01X Polymyxin E <0.1 0.2 0.3 0.4 0.6 0.8 1.0 5.2 Total 0.1 0.1 0.3 0.5 0.8 1.3 2.4 5.2 Injectable antimicrobial drugs QJ01A Oxytetracycline b 3.6 QJ01C Amoxicillin b 4.6 QJ01E Amoxicillin LA 1.4 1.9 2.6 3.0 3.9 5.5 6.9 7.8 QJ01E Ampicillin b 4.8 QJ01E Ampicillin LA 2.4 2.4 2.5 2.6 2.9 3.1 3.3 3.4 QJ01E Benzylpenicillin 0.4 0.4 0.4 0.5 0.5 0.6 0.7 0.7 QJ01E Procaine 0.4 benzylpenicillin b QJ01D Ceftiofur 2.5 2.6 2.6 2.9 3.6 5.2 6.0 6.0 QJ01G Gentamicin b 4.1 QJ01M Enrofloxacin 1.3 1.9 2.4 4.0 6.6 9.3 10.8 11.9 QJ01M Marbofloxacin 2.7 2.8 2.9 3.3 3.9 4.4 4.8 5.0 QJ01X Polymyxin E 2.9 2.9 3.0 3.1 3.4 3.7 3.9 4.0 Total 0.4 0.8 1.4 2.7 3.6 5.0 6.0 11.9 A correctly dosed antimicrobial drug has a ratio of 1. a UDDpig/ADDpig ratio: minimum value, 5th percentile, 10th percentile, 25th percentile, 50th percentile, 75th percentile, 90th percentile and maximum value. b No distribution available. These antimicrobials were only administered once or several times at the same dose. which were representative of a modern Belgian closed or semi-closed pig herd. Extrapolation to herds with fattening pigs originating from different sources is not recommended because the infection pressure and, consequently, the antimicrobial drug use in these production systems might be different (Elbers, 1991; Maes, 1998). Antimicrobial drug consumption data were collected retrospectively during the face-toface interviews, ensuring rapid data collection. A systematic external validation of the collected data could not be performed because this required an independent verification from other records, which were not available on all herds. Therefore, recall bias might have occurred due to incorrect recollection of the group treatments. Also intervention/

T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 259 prevarication bias could have occurred if farmers deliberately misstated their treatment regimens, believing that this would lead to a better personal outcome. Both types of bias could have caused an underestimation of the antimicrobial drug consumption. Bias could also explain some of the large variation between herds. It is difficult to estimate precisely to what degree bias has influenced our data. Considering that these group treatments were administered either on a systematic basis or when extensive clinical disease appeared and considering that the farmer was not aware of the study in advance, we believe that the influence of recall bias and intervention/prevarication bias was limited. A disadvantage of the retrospective data collection is that individual treatment data could not be collected accurately; therefore these treatments were not included in this study. Between 80 and 94% of all treatments in pig production are group treatments (Chauvin et al., 2002; MARAN, 2002). Therefore, we believe that a good, though mildly underestimated (biased), approximation of the total antimicrobial drug consumption in fattening pigs was achieved in this study. Because the total consumption of antimicrobial drugs, expressed in kg, does not give sufficiently detailed information on the antimicrobial drug consumption (Chauvin et al., 2001; Jensen et al., 2004), standardized measures of dose based on the weight of pharmacologically active substances have been introduced. In human medicine, the defined daily dose (DDD) was introduced to study drug use on an international level. It is independent of the formulation, package size and sales price of the drug (Chauvin et al., 2001; WHO, 1996). Because an official, internationally accepted veterinary equivalent to the human DDD does not yet exist (unofficial cfr. Harr et al., 2004; WHO, 2002), Jensen et al. (2004) introduced the animal daily dose (ADD). In addition to the ADD, we introduced the used daily dose (UDD), which is an adaptation of the prescribed daily dose (PDD) (Arnold et al., 2004; Chauvin et al., 2001, 2002). Both UDD and PDD describe the actual administered dose and therefore they might differ from the DDD or ADD (Chauvin et al., 2001; Jensen et al., 2004). Possible reasons for these differences are poor evaluation of the patient s weight or age, adaptation to package size, application of a different therapeutic regimen or use for a different indication, etc. (Chauvin et al., 2001; Hekster et al., 1982). The UDD-system is particularly designed for studies at herd level and is robust to noncompliance, a well-known risk factor for the emergence of antimicrobial resistance (Catry et al., 2003). PDD can also be used at herd level, but is not robust to noncompliance; dispensed antimicrobials are not always administered, and if administered, the prescribed dosages are not always applied (WHO, 2002). The ratio UDD/ADD gives an indication of correct dosing, as internationally called for (Mevius et al., 1999). Because an average ratio per active substance can be falsely correct when it consists of an equilibrated mixture of underdosing and overdosing, a good indication of the appropriateness of dosing can be obtained by evaluating the distribution (minimum, percentiles and maximum) of the UDD/ADD ratio. An UDDpig/ADDpig ratio, which equals 1, indicates that the correct dose was administered. Due to dose ranges mentioned in the different leaflets for the same active substance (depending on e.g. company, severity and place of infection, age of the animal), correct dosing can be performed although the UDD/ADD ratio might not be equal to 1. Therefore, the values which are close to 1 must be interpreted carefully. A range of 20% around 1 might be an acceptable margin for correct dosing. To allow a correct interpretation of consumption

260 T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 figures on a regional basis or in research, this ratio should be taken into account (Jensen et al., 2004). A limitation of the use of the UDD/ADD ratio, with regard to antimicrobial drug resistance, is that treatment interval and treatment duration, two factors which might influence the development of antimicrobial resistance (Zhao and Drlica, 2001), are not incorporated. 4.2. Antimicrobial drug consumption Although in most herds, the fattening pigs were treated in group at some stage in their lives, 12% of the herds were able to raise pigs without group treatments. Similar levels of antimicrobial drug use in finishing herds were observed in the Netherlands, the United States and Canada (Dewey et al., 1997; Dunlop et al., 1998a,b; Elbers et al., 1990). There was a large variation in antimicrobial drug use between the separate herds (Table 3). Differences in management and herd-specific factors have an important impact on the antimicrobial drug use, i.e. hygiene status, prophylactic or metaphylactic use and treatment decision made by the farmer, technician or veterinarian (Chauvin and Madec, 2004; Nielsen et al., 2002). Control and optimization of these factors might reduce the antimicrobial drug consumption. Limiting the prophylactic use, which is often not based on bacteriological diagnosis (Schwarz et al., 2001; Ungemach, 2000) might be especially effective. Discrepancies between the TI ADDpig and the TI UDDpig of the different antimicrobial drugs could be observed. These variances were mainly a consequence of inappropriate dosing, possibly due to the lack of veterinary advice, misinterpretations of the instruction leaflet, or wrong evaluation of the body weight. Treatment incidences were difficult to compare with other studies because, to our knowledge, no studies have been published using this calculation method in pigs. Therefore, only proportional-use data could be used to compare our data with other studies. The proportional TI ADDpig (Table 4) were compared with the figures in the Dutch MARAN report (2002). These figures show a much higher use of tetracyclines (57%), while the use of amoxicillin was much lower (1%). These differences are caused by the high use of amoxicillin in the present study. Amoxicillin was frequently applied in the combination therapy with polymyxin E around weaning age, against symptoms of arthritis and/or meningitis and post-weaning diarrhoea. The combination of amoxicillin, which has a broadspectrum activity and a good absorption from the intestine (Prescott, 2000a) and polymyxin E which has a Gram-negative spectrum and no absorption from the intestine (Prescott, 2000b), ensures a broadspectrum treatment of the piglets. Although a wide range of antimicrobial drugs was used, one registered veterinary antimicrobial class was not applied, namely the pleuromutilins (tiamulin and valnemulin). In Belgium, these antimicrobials are registered for the treatment of Mycoplasma and Brachyspira infections. In the literature, they are also used for the control of ileitis (Friendship, 2000). The use of other antimicrobials (i.e. tetracyclines, macrolides, lincosamides, etc.) to control these diseases in growing-finishing pigs can explain the absence of these antimicrobial drugs. Broadspectrum antimicrobial drugs, belonging to the third-generation cephalosporines (15.9%) and the fluoroquinolones (6.5%), were regularly used for injectable group treatments, though their contribution to the total amount of antimicrobial drugs used in

T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 261 group medication was less than 1%. This is in agreement with the Dutch results of the MARAN (2002) report in which the use of fluoroquinolones and cephalosporines in group medication was almost negligible (0 1%). Because of the possible public health implications due to direct and indirect resistance transfer of these compounds, the routine use of these antimicrobial drugs should be discouraged and retained as second- or even third-line antimicrobial drugs (Mevius, 2002). The differences between the ADDpig and the UDDpig were used to evaluate the appropriateness of dose. Almost all injectable antimicrobials were overdosed. This is probably due to the use of a standard therapy for young piglets, which is not based on a correct estimation of the body weight. Another possible reason might be the difficulty of administering small amounts (<0.5 ml) to piglets. Most oral antimicrobial drugs were also not correctly administered. These differences are probably caused by administering antimicrobials per 1000 kg feed or per 1000 L water, instead of per kilogram body weight. Because a lot of other factors might contribute to inappropriate dosing, a high level of accuracy and prudence in preparing and administering oral antimicrobials is needed (Ungemach, 2000). Based on this information, there is a need for clear recommendations about correct dosing, especially because underdosing is a main risk factor for the development and spread of antimicrobial resistance (Catry et al., 2003) and overdosing can influence the actual withdrawal time. Treating per kilogram body weight should be encouraged. Group treatments, especially those for prophylactic reasons, should be reduced and replaced by other disease-preventive measures (Wierup, 2000). 5. Conclusions We demonstrated large differences in antimicrobial drug consumption between herds. We also demonstrated that some older antimicrobial drugs (i.e. tetracyclines, trimethoprim-sulphonamides, amoxicillin) are still frequently used. The ratio UDDpig/ ADDpig appears to be a potential tool to evaluate the appropriateness of dosing. Using this ratio, we have shown that most antimicrobial drug treatments are not administered at the correct dose. Acknowledgements The authors thank all the farmers for their participation and hospitality. Dr. M. Coryn is acknowledged for his critical reading the manuscript. This study was supported by the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen), grant number SB-023200. References Aarestrup, F.M., 1999. Association between the consumption of antimicrobial agents in animal husbandry and the occurrence of resistant bacteria among food animals. Int. J. Antimicrob. Agents 12, 279 285. Anonymus, 2003a. Belgian Compendium on Veterinary Medicines, 10th ed. Belgian Pharmaceutical Industry Association, Brussels, Belgium.

262 T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 Anonymus, 2003b. Gecommentarieerd geneesmiddelenrepertorium voor diergeneeskundig gebruik 2003. Belgisch Centrum voor Farmacotherapeutische Informatie, Brussels, Belgium. http://www.bcfi-vet.be. 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., 2000. DANMAP: monitoring antimicrobial resistance in Denmark. Int. J. Antimicrob. Agents 14, 271 274. Caprioli, A., Busani, L., Martel, J.L., Helmuth, R., 2000. Monitoring of antibiotic resistance in bacteria of animal origin: epidemiological microbiological methodologies. Int. J. Antimicrob. Agents 14, 295 301. Catry, B., Laevens, H., Devriese, L.A., Opsomer, G., de Kruif, A., 2003. Antimicrobial resistance in livestock. J. Vet. Pharmacol. Ther. 26, 81 93. Chauvin, C., Beloeil, 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. Chauvin, C., Madec, F., Guillemot, D., Sanders, P., 2001. The crucial question of standardisation when measuring drug consumption. Vet. Res. 32, 533 543. Chauvin, C., Madec, F., 2004. Quantitative analysis of antibiotic consumption in turkey broiler production. In: Reid, S.W.J., Menzies, F.D., Russell, A.M. (Eds.), Antimicrobials. Proceedings of the Society for Veterinary Epidemiology and Preventive Medicine Conference, Martigny, Switzerland 24 26 March, pp. 127 133. Dewey, C.E., Cox, B.D., Straw, B.E., Bush, E.J., Hurd, H.S., 1997. Associations between off-label feed additives and farm size, veterinary consultant use, and animal age. Prev. Vet. Med. 31, 133 146. Dunlop, R.H., McEwen, S.A., Meek, A.H., Black, W.D., Clarke, R.C., Friendship, R.M., 1998a. Individual and group antimicrobial usage rates on 34 farrow-to-finish swine farms in Ontario, Canada. Prev. Vet. Med. 34, 247 264. Dunlop, R.H., McEwen, S.A., Meek, A.H., Friendship, R.M., Clarke, R.C., Black, W.D., 1998b. Antimicrobial drug use and related management practices among Ontario swine producers. Can. Vet. J. 39, 87 96. Elbers, A.R., Cromwijk, W.A., Hunneman, W.A., Tielen, M.J., 1990. Logboek registratie op vleesvarkensbedrijven in het Integrale Keting Beheersing (IKB) project. Part I. Gebruik van medicijnen en vaccins. Tijdschr. Diergeneesk. 115, 249 261. Elbers, A.R., 1991. The use of slaughterhouse information in monitoring systems for herd health control in pigs. PhD Thesis, University of Utrecht, The Netherlands, 1 152. EUROSTAT, Animal products, in: Agricultural statistics Quarterly bulletin, no. 4/2003, Office for official publications of the European Communities, Luxembourg. http://europa.eu.int/comm/eurostat/public/ datashop/print-product/en?catalogue=eurostat&product=ks-nt-03-000-3a-n-en&mode=download. Friendship, R.M., 2000. Antimicrobial drug use in swine. In: Prescott, J.F., Baggot, J.D., Walker, R.D. (Eds.), Antimicrobial Therapy in Veterinary Medicine, 3rd ed. Iowa State University Press, Ames, Iowa, USA, pp. 602 617. Grave, K., Bangen, M., Engelstad, M., Soli, N.E., 1991. An evaluation of the information given on veterinary prescriptions in Norway. Compliance with legal regulations. J. Vet. Pharmacol. Ther. 14, 150 155. Grave, K., Greko, C., Nilsson, L., Odensvik, K., Mørk, 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. Gyssens, I.C., 2001. Quality measures of antimicrobial drug use. Int. J. Antimicrob. Agents 17, 9 19. Harr, L.M.F., Litleskare, I., Grave, K., Rønning, M., 2004. DDDanimal a unit of measurement. In: Proceedings of the 2nd International Conference on Antimicrobial Agents in Veterinary Medicine, Ottawa, Canada, 13 17 June, p. 81. Hekster, Y.A., Vree, T.B., Goris, R.J.A., Boerema, J.B.J., 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. J. Clin. Hosp. Pharm. 7, 251 260. Jensen, V.F., Jacobsen, E., Bager, F., 2004. Veterinary antimicrobial-usage statistics based on standardized measures of dosage. Prev. Vet. Med. 64, 201 215. Maes, D.G.D., 1998. Respiratory disease in slaughter pigs: epidemiological aspects and effect of vaccination against Mycoplasma hyopneumoniae. PhD Thesis, Ghent University, Belgium, 1 227.

T. Timmerman et al. / Preventive Veterinary Medicine 74 (2006) 251 263 263 MARAN, 2002. Usage of antibiotics in animal husbandry in the Netherlands. In: Mevius, D.J. (Ed.), Monitoring of Antimicrobial Resistance and Antibiotic Usage in Animals in the Netherlands in 2002. http://www.cidclelystad.nl/docs/maran-2002-web.pdf. Mateu, E., Martin, M., 2001. Why is anti-microbial resistance a veterinary problem as well? J. Vet. Med. B 48, 569 581. McKellar, Q.A., 1998. Antimicrobial resistance: a veterinary perspective. Brit. Med. J. 317, 610 611. Mevius, D.J., 2002. Standpunt werkgroep veterinair antibioticumbeleid. Verantwoord gebruik quinolonen gewenst. Tijdschr. Diergeneesk. 127, 622 624. Mevius, D.J., Sprenger, M.J.W., Wegener, H.C., 1999. EU conference The Microbial Threat. Int. J. Antimicrob. Agents 11, 101 105. 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. Nielsen, J.P., Petersen, H.H., Bak, H., 2002. Antimicrobial consumption in Danish finishing herds. In: Proceedings of the International Pig Veterinary Society 17, vol. 2, Ames, Iowa, USA, 2 5 June, p. 9. Prescott, J.F., 2000a. Chapter 6: beta-lactam antibiotics: penam penicillins. In: Prescott, J.F., Baggot, J.D., Walker, R.D. (Eds.), Antimicrobial Therapy in Veterinary Medicine, 3rd ed. Iowa State University Press, Ames, Iowa, USA, pp. 105 133. Prescott, J.F., 2000b. Chapter 9: peptide antibiotics: polymyxins, glycopeptides, streptogramins and bacitracin. In: Prescott, J.F., Baggot, J.D., Walker, R.D. (Eds.), Antimicrobial Therapy in Veterinary Medicine, 3rd ed. Iowa State University Press, Ames, Iowa, USA, pp. 177 190. SANITEL, 2003. Federal Agency for the Safety of the Food Chain, World Trade Center III, Boulevard Simon 30, 1000 Brussels. Schwarz, S., Kehrenberg, C., Walsh, T.R., 2001. Use of antimicrobial agents in veterinary medicine and food animal production. Int. J. Antimicrob. Agents 17, 431 437. Straw, B.E., Meuten, D.J., Thacker, B.J., 1999. Physical examination. In: Straw, B.E., D Allaire, S., Mengeling, W.L., Taylor, D.J. (Eds.), Diseases of Swine, 8th ed. Iowa State University Press, Ames, Iowa, USA, pp. 3 18. Ungemach, F.R., 2000. Figures on quantities of antibacterials used for different purposes in the EU countries and interpretation. Acta Vet. Scand. Suppl. 93, 89 98. WHO, 1996. WHO Collaborating Centre for Drug Statistics Methodology, Guidelines for ATC Classification and DDD Assignment, 1st ed. Oslo, Norway. WHO, 2002. WHO Collaborating Centre for Drug Statistics Methodology, Guidelines for ATCvet Classification and DDD Assignment, Oslo, Norway. http://www.whocc.no/atcvet. Wierup, M., 2000. The control of microbial diseases in animals: alternatives to the use of antibiotics. Int. J. Antimicrob. Agents 14, 315 319. Zhao, X., Drlica, K., 2001. Restricting the selection of antibiotic-resistant mutants: a general strategy derived from fluoroquinolone studies. Clin. Infect. Dis. 33, 147S 156S.