The authors have declared that no competing interests exist.
Conceived and designed the experiments: LvR CvM RM LK. Performed the experiments: LvR. Analyzed the data: LvR HO MH LK. Contributed reagents/materials/analysis tools: HO WH. Wrote the paper: LvR WH AK LK.
To be able to analyze the relationship between the level of resistance and the use of antimicrobials, it is necessary to collect detailed data on antimicrobial usage. For this reason, data on antimicrobial use on 495 pig farms from entire Germany were collected and analyzed. In Germany, each application and dispensing of medicines to food-producing animals is documented in detail obligatorily by the veterinarian. This information was collected retrospectively for the year 2011. The analyses undertook separate examinations on the age groups sow, piglet, weaner and fattening pig; both the route of administration and indication per active ingredient, and active ingredient class, were evaluated. In total, 20,374 kg of antimicrobial substances were used in the study population. Tetracyclines were used in highest amounts, followed by beta-lactams, trimethoprim-sulfonamides and macrolides. Concerning the frequency of using an active substance per animal, polypeptides were most commonly administered. In all age groups, respiratory infections were the main indication for using antimicrobials, followed by intestinal diseases in piglets, weaners and fattening pigs and diseases of reproductive organs in sows. Over a period of 100 days, the median number of treatment days with one antimicrobial substance for piglets was 15 days, for weaners about 6 days, for fattening pigs about 4 days and for sows about 1 day. A multifactorial ANOVA was conducted to investigate which factors are associated with the treatment frequency. The factors “veterinarian” and “age group” were related to the treatment frequency, just as the interaction between “veterinarian” and “farm size” as well as the interaction between “veterinarian” and “age group”.
The availability of antibiotics for treating bacterial diseases in animals must be preserved to ensure animal health and welfare. As any use of antimicrobial substances may be associated with the development of resistance [
To be able to assess the relationship between the observed resistance and the use of antimicrobials, it is necessary to provide detailed data on antimicrobial usage. In the Zoonoses Directive 2003/99/EC of the European Commission [
Years ago, active monitoring programs on antimicrobial consumption were established in some European countries, which have been continuously carried out to date (see e.g. [
For this reason a feasibility study was conducted successfully in the years 2007 and 2008 and a method for collecting data on consumption of antimicrobials in farm animals was developed [
The study was set up as a cross-sectional study, based on voluntary participation of veterinarians and farmers. To ensure the representativeness for the whole of Germany, an appropriate sampling plan was created. In a first step, Germany was classified in four regions: Northwest, Middle, East and South, already defined by Merle et al. [
All participants gave consent to use the data for this study, given that all data of pig holding farmer and veterinarian were pseudonymized before the data were analyzed. This was stated in a privacy statement with each participant. The study protocol was approved by the sponsor (the Federal Institute of Risk Assessment) and announced to the data protection officers of both universities.
The recruitment of participants and input of their data was carried out in the period from May 2012 until February 2013. Data were collected retrospectively for the calendar year 2011.
The data collection was implemented following the concepts from the feasibility study in 2007/2008 [
In addition, an animal-specific questionnaire was sent to participating farmers, who should give information about farm management, keeping of animals, and also the number of animals kept as well as livestock places.
The data were entered manually into a database developed during the feasibility study—here the Open-Source relational database system MySQL was used in its current version (currently 5.0. 51a). There were mandatory fields defined in the input mask, which complied with the mandatory information on the treatment and delivery form. To enable fast manual entering of data, drop-down lists were applied. Diagnosis was linked to eight predetermined diagnostic groups (udder disease, skin disease, respiratory disease, intestinal disease, central nervous system disease, joint disease, reproductive organs and other diseases). One entered data record corresponded to one treatment. All data records were checked for completeness and pharmacological plausibility, both during data entry as well as during a final extensive plausibility checkup, taking into account information stated in the additional farmer's questionnaire received.
In the context of the analysis, each age group, defined by weight and production system was considered separately. Thus, the four age groups piglet (4 kg on average), weaner (15 kg on average), fattening pig (50 kg on average) and sow (200 kg on average) were distinguished. Allocation to these groups was done to the best extent possible on the basis of the information documented on the identity of the animals treated or the type of pigs kept on the farm. It was assumed that a sow is housed for more than one year, so the time period under risk of a sow was stated as one year. A piglet is suckled 28 days on average before it is transferred to the next production phase. In Germany the average production phase of a weaner is about 46 days. With a service period of about seven days between groups, this yields approximately 6.89 passages for weaning units per farm per year. In Germany, a fattening pig is fattened 115 days on average. With a subsequent time-window for cleaning and disinfection of 14 days, 2.83 fattening passages arise per year and per farm on average [
In the context of the analysis, the participating pig farms were divided into various farm size groups based on the numbers of livestock places per farm. To achieve this, the farm size groups “small”, “medium” and “large” were defined for each age group on the basis of tertiles derived from our data. Medium-sized farms were defined as those with 80 to 199 sows, 820 to 2,049 piglets, 330 to 799 weaners and 370 to 999 fattening pigs.
The analysis involved the determining of the used active ingredients included in the drug applied and active ingredient quantities per age group. Focus of the analysis was laid on the calculation of amounts of antimicrobial substances and as well on the calculation of number of treatments per active ingredient and animal (treatment unit or number of used daily doses (nUDD)), and the calculation of treatment days per active ingredient per animal (treatment frequency). A treatment unit (UDD) is equivalent to the administration of one active ingredient to one animal on one day [
The treatment frequency specifies how many days an animal in a herd is treated with one active ingredient on average [
To calculate the treatment frequency for each farm per year, the number of animals kept per farm per year had to be determined. For this purpose, the livestock places per farm and age group were multiplied by the average number of production periods per farm per year (sow = 1; weaner = 6.89; fattening pig = 2.83). The number of piglets per farm and year was always estimated for any one farm keeping sows, by multiplying the number of sows on this farm by the average number of litters per sow and year (2.35) and the average number of weaned piglets per sow and litter (10.25).
In the course of the analysis, also factors were analyzed which were assumed to have an influence on the treatment frequency. These factors were analyzed for each age group separately via a multifactorial analysis of variance (ANOVA, target variable: logarithmized treatment frequency, standardized per 100 days of animals kept). Region and farm size were examined as fixed factors; the responsible veterinarian was examined as random factor. For model selection backward selection procedures were used.
Statistical analyses were performed using SAS, version 9.3 TS level 1M2 (SAS Institute Inc., Cary, NC, United States).
In total, 18,151 data records of 495 pig farms from entire Germany were collected and analyzed. On the 495 participating pig farms, one or several age groups were kept, giving a total of 945 groups (sow, piglet, weaner, fattening pig). These groups were allocated as follows: 215 farms keeping sows, 199 farms keeping piglets, 187 farms keeping weaners, as well as 344 farms keeping fattening pigs. In relation to the entire German pork production the study covered 2.19% of all livestock places for sows, 6.14% for piglets, 3.17% for weaners and 2.85% for fattening pigs in Germany (data source from Federal Statistical Office [
The analysis of the data was carried out in two different ways. On the one hand, the amount of antimicrobial agents consumed was determined separately for the different age groups. This was reported with consideration of the season, the administration route and the indication. On the other hand, the frequency of treatments was analyzed separately according to age group.
In total, 20,373.6 kg of antimicrobial agents were used on the study farms in 2011. Tetracyclines were most commonly used with a consumption quantity of 7,275.5 kg (35.7%), followed by beta-lactams with 6,720.9 kg (33.0%), potentiated sulfonamides (trimethoprim-sulfonamides) with 2,126.6 kg (10.4%) and macrolides with 2,827.7 kg (13.9%). Polypeptides (858.4 kg, 4.2%) and aminoglycosides (248.2 kg, 1.2%) were in the fifth and sixth position regarding consumption quantities. The consumption quantities per active ingredient were also evaluated separately for each age group. In addition, seasonal variations in the consumption of certain antimicrobial substances were examined according to the age group (
active ingredient | Consumed quantities in kg | ||||
---|---|---|---|---|---|
spring | summer | autumn | winter | total | |
66.9 | 75.4 | 23.6 | 54.7 | 220.5 | |
23.2 | 27.5 | 26.1 | 21.0 | 97.9 | |
2.4 | 1.4 | 1.5 | 1.0 | 6.3 | |
237.6 | 97.2 | 295.2 | 216.0 | 846.0 | |
592.3 | 134.9 | 100.0 | 118.3 | 945.5 | |
2.6 | 11.3 | 3.6 | 1.0 | 18.5 | |
304.8 | 446.8 | 452.4 | 200.5 | 1,404.6 | |
38.4 | 36.2 | 62.9 | 16.7 | 154.2 | |
40.9 | 68.7 | 60.7 | 36.9 | 207.2 | |
113.4 | 75.8 | 132.1 | 56.9 | 378.2 | |
116.7 | 180.2 | 220.6 | 133.5 | 651.0 | |
40.3 | 21.9 | 19.3 | 21.7 | 103.1 | |
907.2 | 480.9 | 547.8 | 753.6 | 2,689.5 | |
228.0 | 182.0 | 108.1 | 129.6 | 647.7 | |
164.4 | 136.0 | 77.5 | 151.1 | 528.9 | |
153.6 | 134.4 | 111.4 | 187.6 | 587.0 | |
385.2 | 268.4 | 304.2 | 334.0 | 1,291.9 | |
36.9 | 34.8 | 26.4 | 35.3 | 133.4 | |
574.2 | 505.8 | 535.9 | 652.7 | 2,268.6 | |
361.1 | 327.0 | 326.3 | 413.5 | 1,427.8 | |
42.0 | 21.3 | 26.7 | 25.8 | 115.9 | |
91.4 | 107.4 | 173.5 | 104.8 | 477.2 | |
302.2 | 364.5 | 199.9 | 240.1 | 1,106.6 | |
483.8 | 485.4 | 474.6 | 383.8 | 1,827.7 |
Among the sows, the active ingredients of trimethoprim-sulfonamides, tetracycline and amoxicillin were administered most. Tetracycline was used the most in spring. The consumption amount of tetracycline in the spring was about five-fold higher than the amount consumed in the other seasons. The highest amounts of trimethoprim-sulfonamides were consumed in autumn, and almost as much in summer. Amoxicillin was used in summer in highest amounts, but the least in autumn. The highest drug use for all active ingredients was detected in spring.
Among piglets, amoxicillin, tetracycline, trimethoprim-sulfonamides and colistin were used in highest amounts, also with seasonal differences. In contrast to the fattening pigs, amoxicillin was least consumed in winter, with a steady increase in consumption amounts from spring to autumn. Amoxicillin was used more than twice as much in autumn than it was in winter. This observation was also made for tetracycline, but with the lowest consumption level being in spring. For all active ingredients, the highest drug use in piglets was detected in autumn.
Among weaners, amoxicillin, tetracycline, chlortetracycline, and about equal proportions of trimethoprim-sulfonamides and colistin were used in highest amounts. In this age group, seasonal differences in regard to the consumed quantities of several active ingredients were also observed. Amoxicillin was consumed the most in spring, twice as much as in summer, with a steady rise in the consumption quantities from summer to spring. Distributed over the year, tetracycline as well as trimethoprim-sulfonamides were consumed constantly. However, the amounts of chlortetracycline were much higher in spring than in the second half of the year. In weaners, colistin was used the least in autumn, but rather constantly over the rest of the year. For all active ingredients, the highest drug use in weaners was detected in spring.
Amoxicillin, tetracycline, tylosin and chlortetracycline were used in highest amounts in fattening pigs and there were different seasonal patterns. Amoxicillin was more often used in winter than in the summer months, whereas the active ingredient tylosin was used about 100 kg less in winter than in the summer. For all active ingredients, the highest drug use for fattening pigs was detected from June to August, the lowest use in autumn.
From all 20,373.6 kg antimicrobial substances used, 19,903 kg (98%) were administered orally, and only 472.3 kg (2%) were administered per injection. However, looking at the data records, there were more than twice as many records on parenteral administration than on oral administration. Among the four age groups, especially for sows there were mainly records on treatment per injection. Regarding the fattening pigs, records on parenteral and oral treatments were approximately the same, with slightly more records on parenteral treatments. In contrast, for weaners records on oral administration were about twice as frequently as records on parenteral treatments. Among the piglets, there were more records on parenteral treatments than for oral administration.
In the course of the analysis, also the consumption amounts of antibiotic active substances according to several indications were determined for the different age groups (
active ingredient | Consumed quantities in kg | |||||||
---|---|---|---|---|---|---|---|---|
respiratory disease | intestinal disease | joint disease | skin disease | other diseases | central nervous system | reproduc-tive organs | total | |
125.8 | 1.0 | 8.5 | 0.8 | 73.8 | - | 10.6 | 220.5 | |
41.1 | - | - | <0.1 | 21.5 | - | 35.3 | 97.9 | |
0.5 | 5.7 | - | - | 0.1 | - | <0.1 | 6.3 | |
39.6 | - | - | - | 768.0 | - | 38.4 | 846.0 | |
776.2 | 4.6 | - | - | 89.6 | - | 75.1 | 945.5 | |
1.1 | 12.5 | 0.9 | - | 2.7 | <0.1 | 1.3 | 18.5 | |
849.1 | 12.2 | 13.9 | 14.5 | 514.9 | - | - | 1404.6 | |
149.3 | 4.0 | - | - | 0.9 | - | - | 154.2 | |
2.1 | 200.4 | - | - | 4.6 | - | - | 207.2 | |
281.3 | 74.3 | - | 4.8 | 17.80 | - | - | 378.2 | |
571.5 | - | - | - | 79.5 | - | - | 651.0 | |
6.8 | 88.1 | - | - | 8.3 | - | - | 103.1 | |
2195.3 | 52.5 | 1.3 | 37.1 | 403.3 | - | - | 2689.5 | |
635.9 | 1.4 | - | - | 10.4 | - | - | 647.7 | |
5.3 | 520.4 | - | - | 3.2 | - | - | 528.9 | |
518.2 | 50.4 | - | 6.4 | 12.0 | - | - | 587.0 | |
1068.1 | 127.3 | - | - | 96.6 | - | - | 1291.9 | |
5.1 | 112.2 | - | - | 16.1 | - | - | 133.4 | |
1924.0 | 92.7 | 14.1 | 6.3 | 231.5 | - | - | 2268.6 | |
1175.7 | 181.6 | - | 2.8 | 67.7 | - | - | 1427.8 | |
6.4 | 95.2 | - | - | 14.5 | - | - | 116.1 | |
413.9 | 26.0 | - | - | 37.2 | - | - | 477.2 | |
830.9 | 103.3 | - | 2.8 | 169.6 | - | - | 1106.6 | |
39.9 | 1747.3 | 0.1 | - | 46.1 | - | - | 1833.4 |
In piglets, the highest amounts of antimicrobial substances were used against respiratory diseases. In second and third place, "other diseases" and "intestinal diseases" were documented as reasons for treatment. Respiratory diseases in piglets were first and foremost treated with amoxicillin, but also with tetracycline. Intestinal diseases in piglets were mostly treated with colistin and tylosin.
Also in sows, the highest quantities of antimicrobial active ingredients were used against respiratory diseases. “Other diseases” were in second place, too, but in third place were diseases of the reproductive organs. Respiratory diseases in sows were particularly treated with tetracycline but also with amoxicillin. Diseases of the reproductive organs were mainly treated with amounts of tetracycline.
In fattening pigs and weaners, the highest quantities of antimicrobial substances were also used against respiratory diseases.
As in piglets, respiratory diseases in weaners were mainly treated with amoxicillin, but also with tetracycline. Intestinal diseases were especially treated with colistin, but also tetracycline and tylosin were used in approximately equal amounts.
Also in fattening pigs respiratory diseases were mainly treated with amoxicillin. In addition to amoxicillin, especially chlortetracycline played an important role in the treatment of respiratory diseases. The treatment of intestinal diseases was primarily carried out with tylosin.
The amount of drugs applied is strongly influenced by the recommended dosage as well as the body weight of the treated animal group. Therefore, tons of used drugs may only give a rough picture of the situation. Thus, treatment units were calculated. To illustrate the difference between these two ways of analyses, the Figs.
In sows, tetracyclines, trimethoprim-sulfonamides and beta-lactams are up front in terms of treatment units as well as the consumed quantities. It is remarkable that fluoroquinolones, which were only used in small quantities, are in fourth place, concerning the treatment units. Also macrolides, polypeptides and pleuromutilins take a relatively large proportion of the treatments in this age group.
In piglets, beta-lactams, tetracyclines and trimethoprim-sulfonamides were most commonly used in terms of quantity. Regarding the treatment units, polypeptides became more prominently used in contrast to the use of tetracyclines. Macrolides, which only are in fifth place in terms of quantity, were nearly used as often as tetracyclines. Only small quantities of aminoglycosides, fluoroquinolones and cephalosporins were used, but they accounted for a relatively large proportion of treatment units.
Among weaners, there are clear differences between consumed quantities and treatment units of each active ingredient class. In terms of quantity, especially beta-lactams, tetracyclines and trimethoprim-sulfonamides were used. Nonetheless, looking at treatment units, polypeptides were the most used antimicrobial class in this age group, followed by beta-lactams and tetracyclines.
In fattening pigs, tetracyclines were most commonly used with regard to the consumed quantities, followed by beta-lactams and macrolides. On the treatment unit scale, macrolides were ahead of beta-lactams. When directly comparing the consumed quantities and treatment unit scale, the total amount of used polypeptides was comparatively low, but the number of treatment units was higher. It becomes obvious that polypeptides and pleuromutilins are much more important as is shown when looking at the consumed quantities only.
Looking at the treatment frequencies depicted as distribution of the farm specific treatment frequencies with regard to one year and one passage, respectively, it was calculated that a sow on a typical farm was treated with one active ingredient on 3.2 days (median) per year on average. A fattening pig was treated with one active ingredient on 4.2 (median) days in its 115-day fattening period, with a combination of two active ingredients on 2.1 days. In its 46-day production period, a weaner was treated with one active ingredient on 3.1 (median) days, with a combination of two active substances on about 1.5 days. Within its four-week suckling period, a piglet received treatment with one active ingredient on 4.1 (median) days.
In the following, the treatment frequencies with one active ingredient per farm and 100 days, given as median value of all farms, are described, to ensure the comparability between the age groups. Converted to 100 days, a median treatment frequency of 0.86 days arose for sows, 14.74 days for piglets, 6.62 days for weaners and 3.67 days for fattening pigs.
The Figs.
age group | treatment frequency per 100 days | |||||||
---|---|---|---|---|---|---|---|---|
n | min | P5 | P25 | median | P75 | P95 | max | |
215 | 0.005 | 0.023 | 0.26 | 0.86 | 2.39 | 13.32 | 26.43 | |
199 | 0.006 | 0.593 | 4.78 | 14.74 | 35.71 | 93.65 | 261.86 | |
187 | 0.011 | 0.126 | 2.20 | 6.62 | 17.75 | 45.50 | 180.80 | |
344 | 0 | 0.031 | 0.92 | 3.67 | 7.84 | 19.09 | 45.62 |
In the course of the analysis, standardized therapy frequencies of individual active ingredient classes were calculated for the different age groups (
active ingredient class | Treatment frequency per 100 days | |||||||
---|---|---|---|---|---|---|---|---|
n | min | P5 | P25 | median | P75 | P95 | max | |
167 | 0.003 | 0.020 | 0.05 | 0.17 | 0.51 | 1.69 | 12.47 | |
65 | 0.001 | 0.005 | 0.02 | 0.06 | 0.26 | 1.08 | 1.62 | |
71 | 0.003 | 0.005 | 0.02 | 0.04 | 0.11 | 0.51 | 1.38 | |
81 | 0.002 | 0.004 | 0.10 | 0.39 | 1.89 | 8.32 | 9.55 | |
7 | 0.064 | 0.064 | 0.95 | 2.23 | 7.56 | 11.34 | 11.34 | |
111 | 0.002 | 0.011 | 0.04 | 0.15 | 0.42 | 1.28 | 3.17 | |
21 | 0.003 | 0.007 | 0.02 | 0.07 | 0.13 | 0.46 | 1.47 | |
12 | 0.003 | 0.003 | 0.07 | 0.26 | 0.78 | 3.65 | 3.65 | |
55 | 0.002 | 0.002 | 0.02 | 0.05 | 0.19 | 1.44 | 10.58 | |
77 | 0.003 | 0.020 | 0.09 | 0.26 | 1.87 | 17.06 | 21.19 | |
17 | 0.002 | 0.002 | 0.01 | 0.02 | 0.04 | 0.27 | 0.27 | |
. | . | . | . | . | . | . | . | |
170 | 0.006 | 0.153 | 1.11 | 4.33 | 10.06 | 28.86 | 58.97 | |
103 | 0.011 | 0.067 | 0.33 | 1.15 | 3.21 | 8.81 | 28.05 | |
44 | 0.015 | 0.045 | 0.18 | 0.56 | 2.30 | 6.03 | 11.86 | |
107 | 0.010 | 0.114 | 1.35 | 3.46 | 8.82 | 25.94 | 49.92 | |
110 | 0.126 | 0.432 | 3.50 | 9.73 | 19.96 | 48.24 | 98.59 | |
120 | 0.033 | 0.070 | 0.27 | 0.56 | 1.51 | 4.39 | 14.01 | |
12 | 0.035 | 0.035 | 0.05 | 0.15 | 0.92 | 7.90 | 7.90 | |
8 | 0.185 | 0.185 | 1.46 | 2.72 | 4.12 | 31.95 | 31.95 | |
97 | 0.013 | 0.064 | 0.64 | 1.42 | 4.45 | 31.12 | 76.22 | |
34 | 0.889 | 1.556 | 2.37 | 6.35 | 10.56 | 68.06 | 78.48 | |
16 | 0.019 | 0.019 | 0.16 | 0.75 | 2.08 | 12.72 | 12.72 | |
. | . | . | . | . | . | . | . | |
135 | 0.003 | 0.081 | 0.74 | 2.20 | 5.44 | 19.42 | 64.68 | |
46 | 0.002 | 0.005 | 0.03 | 0.12 | 0.41 | 6.31 | 25.15 | |
41 | 0.006 | 0.008 | 0.03 | 0.08 | 0.19 | 0.38 | 2.83 | |
106 | 0.001 | 0.090 | 0.97 | 3.20 | 6.25 | 16.04 | 45.04 | |
103 | 0.011 | 0.463 | 2.07 | 5.14 | 9.26 | 21.39 | 68.66 | |
60 | 0.003 | 0.008 | 0.03 | 0.10 | 0.26 | 1.52 | 2.63 | |
17 | 0.013 | 0.013 | 0.16 | 0.41 | 1.16 | 4.58 | 4.58 | |
10 | 0.009 | 0.009 | 0.44 | 1.35 | 2.06 | 3.13 | 3.13 | |
73 | 0.004 | 0.009 | 0.08 | 0.31 | 1.36 | 14.16 | 23.67 | |
41 | 0.024 | 0.121 | 1.89 | 5.00 | 7.33 | 15.43 | 28.38 | |
17 | 0.000 | 0.000 | 0.03 | 0.08 | 0.25 | 1.57 | 1.57 | |
. | . | . | . | . | . | . | . | |
260 | 0.004 | 0.009 | 0.07 | 0.81 | 3.09 | 8.74 | 16.88 | |
92 | 0.001 | 0.004 | 0.02 | 0.09 | 0.46 | 2.15 | 8.00 | |
45 | 0.001 | 0.003 | 0.01 | 0.03 | 0.05 | 0.43 | 1.16 | |
191 | 0.001 | 0.051 | 1.00 | 2.16 | 4.21 | 7.67 | 17.72 | |
62 | 0.004 | 0.161 | 0.66 | 1.78 | 4.06 | 12.90 | 16.57 | |
154 | 0.003 | 0.008 | 0.03 | 0.07 | 0.18 | 0.68 | 8.75 | |
77 | 0.003 | 0.011 | 0.05 | 0.17 | 1.08 | 2.15 | 2.92 | |
45 | 0.002 | 0.006 | 0.46 | 1.13 | 2.39 | 5.44 | 11.67 | |
159 | 0.002 | 0.005 | 0.04 | 0.25 | 1.52 | 5.30 | 25.48 | |
49 | 0.009 | 0.021 | 0.54 | 1.94 | 4.27 | 7.56 | 11.76 | |
57 | 0.001 | 0.005 | 0.02 | 0.04 | 0.12 | 0.77 | 3.80 | |
1 | . | . | . | 0.71 | . | . | . |
As another example, polypeptides, which were used on 110 piglet keeping farms (55% of all farms keeping piglets) for treating piglets, were applied at about 10 days on these farms over a period of 100 days. In contrast, polypeptides were only used on 7 sow keeping farms (3% of all sow keeping farms). Within these farms, polypeptides were applied at 2.2 days to sows, but all in all they were not important for treating sows. Furthermore, polypeptides were used in 55% of all weaner keeping farms but only on 18% of all farms keeping fattening pigs. Within these farms, weaners were treated at about 5 days with polypeptides and fattening pigs at 1.8 days over a period of 100 days.
To investigate which factors are associated with the treatment frequency in general, a multifactorial ANOVA was conducted for each age group separately. In
sow | piglet | weaner | fattening pig | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
n | GM | Std | n | GM | Std | n | GM | Std | n | GM | Std | |
75 | 0.62 | 6.7 | 76 | 9.87 | 5.48 | 72 | 3.78 | 6.88 | 168 | 2.29 | 8 | |
86 | 0.76 | 5.65 | 71 | 9.25 | 5.23 | 68 | 4.28 | 5.87 | 149 | 1.6 | 9.3 | |
7 | 3.31 | 4.1 | 6 | 14.1 | 4.98 | 6 | 15.67 | 9 | 9 | 3.4 | 11.03 | |
47 | 0.8 | 4.03 | 46 | 18.07 | 3.92 | 41 | 10.45 | 2.69 | 18 | 2.91 | 2.21 | |
68 | 0.65 | 5.01 | 56 | 7.73 | 6.97 | 62 | 2.96 | 4.07 | 115 | 0.93 | 12.71 | |
63 | 0.63 | 5.85 | 60 | 13.65 | 3.86 | 60 | 4.55 | 6.8 | 113 | 2.24 | 5.53 | |
84 | 0.96 | 6.01 | 83 | 12.5 | 4.65 | 65 | 9.93 | 5.59 | 116 | 3.85 | 5.7 |
Following the ANOVA-multi-factorial approach:
neither “region” nor “farm size” had a significant impact on the treatment frequency; except for weaners, where these two fixed factors seemed to have a slight influence (
178 | 81.6707230 | 0.4588243 | |||
214 | 121.2449435 | ||||
2 | 1.34617447 | 0.67308724 | 2.64 | 0.1858 | |
2 | 0.07275842 | 0.03637921 | 0.06 | 0.9459 | |
13 | 19.42976993 | 1.49459769 | 3.26 | 0.0002 | |
14 | 9.12067360 | 0.65147669 | 1.42 | 0.1478 | |
4 | 1.01988934 | 0.25497233 | 0.56 | 0.6951 |
163 | 49.26920456 | 0.30226506 | |||
198 | 98.08760267 | ||||
2 | 0.97108955 | 0.48554477 | 1.36 | 0.3793 | |
2 | 2.72531103 | 1.36265551 | 1.46 | 0.2682 | |
13 | 33.55440307 | 2.58110793 | 8.54 | <.0001 | |
13 | 12.14350519 | 0.93411578 | 3.09 | 0.0004 | |
3 | 1.06889828 | 0.35629943 | 1.18 | 0.3196 |
149 | 59.2655024 | 0.3977550 | |||
186 | 108.3508443 | ||||
2 | 1.16987186 | 0.58493593 | 7.69 | 0.0426 | |
2 | 5.76430524 | 2.88215262 | 3,96 | 0.0415 | |
13 | 18.64776524 | 1.43444348 | 3.61 | <.0001 | |
15 | 10.90852450 | 0.72723497 | 1.83 | 0.0355 | |
4 | 0.30418793 | 0.07604698 | 0.19 | 0.9427 |
298 | 212.2190436 | 0.7121444 | |||
343 | 288.7672195 | ||||
2 | 0.68633442 | 0.34316721 | 0.92 | 0.4557 | |
2 | 2.83907983 | 1.41953991 | 1.11 | 0.3513 | |
17 | 23.66596891 | 1.39211582 | 1.95 | 0.0139 | |
18 | 23.03261344 | 1.27958964 | 1.80 | 0.0251 | |
5 | 1.85830095 | 0.37166019 | 0.52 | 0.7597 |
Data presented here was created in the setting of a cross-sectional study taken into account the variability in German farming practice. Therefore a stratified approach was used with four regions of different agricultural areas. The formal study base used was the official forms of recording the use of pharmaceutical drugs which are obligatory by law in Germany. Therefore all calculations were based on individual application data on farm level, including the drug, the amount, the number of days, the number of animals treated, and the animal age group respectively on an individual level of treatment. This concept is in contrast to other (monitoring) studies, where mostly the entire amount per farm or even more general groups was recorded and numbers of animals treated and other variables of interest have to be estimated from this aggregated information (see e.g. studies in Denmark [
As sales data do not contain any information on the usage of the drugs, it is not possible to combine sales data with real antibiotic application for animals. Bondt et al. [
This study is based upon voluntary participation of farmers and veterinarians. It has to be taken into account that especially those veterinarians and farmers may be open to participation, who usually care about a diligent and thoughtful handling of drugs and who are seeking to improve their management. Thus, it is possible that the real use of antimicrobial substances is higher and a selection bias cannot be completely ruled out. To ensure compliance with the study, all information was pseudo-anonymized and no information was transferred to veterinary authorities. This resulted in a study base of farms, which is closely connected to the figures known from the demographic structures in pig-holdings farms in Germany concerning the regional distribution as well as the structures of farming sizes and general facilities. Therefore, the coverage of livestock places in the study population compared to the total livestock population in Germany is sufficiently high. These figures were evaluated on the basis of the annual German livestock microcensus (Federal Statistical Office, microcensus 2011). In addition these results show a huge variety of treatment frequencies per farm with some extended results (e.g. a farm with a treatment frequency within the 42 day flat-deck period of approximately 75, i.e. a treatment with two active ingredients per day for all animals). This should be seen as an argument that a severe selection bias is not present in the data.
For evaluating and describing the antimicrobial usage, various definitions and variables have been introduced over time [
However, active ingredients, which were applied at a high dosage over several days, carry higher weight than those which were applied at a lower dosage. In addition, the frequency of application may be different, depending on the type of drug used and whether they maintain the effect of the active ingredient for a longer time. This may lead to an overvaluation or undervaluation of the situation [
Using an appropriate variable is essential for interpreting the antibiotic usage [
In the context of the analysis presented here, the antibiotic use was examined regarding the used quantities as well as the frequency of treatments. To do this, the variables "treatment unit" and especially "treatment frequency" were used. This term specifies the average number of days an animal of a defined population is treated with one active ingredient within a certain period of time. The treatment frequency does not distinguish between dosages, but it merely indicates how often an animal is treated with one active ingredient. Therefore, low and high dosed active ingredients, as well as treatment of young or elder animals, are equally weighted, unlike pure consumption quantities or terms which are based on dosages and standardized body weights (Jensen et al. [
Whereas the concept in general is comparable to the nDDDA (number of defined daily doses animal; see ESVAC [
The treatment frequency is comparable between years and regions, and may also refer to an arbitrary population [
In general, among all age groups in pigs the highest amounts of antimicrobials used were tetracyclines, followed by beta-lactams and trimethoprim-sulfonamides. Also in regard to the sales data for all animals species in Germany, tetracyclines and beta-lactams are in first and second place, followed by trimethoprim-sulfonamides, macrolides and polypeptides [
In the feasibility study of the years 2007/2008, Merle et al. [
The finding of our recent study in general is in line with sales data for 2011 in Germany, where penicillins were the antimicrobial class in second place as regards the sold amount. The trend towards increasing use of beta-lactams and decreasing use of tetracyclines was already discussed by Merle et al. [
In 2011 sales data on antimicrobial substances used in veterinary medicine in Germany were collected obligatory for the first time. These data based on the announcements of pharmaceutical companies and wholesales, who sold drugs to veterinarians in Germany and were published in context of the third ESVAC report [
But, sales data is a total on suggested use in all animal species, which prohibited a direct comparison for the data presented here for pigs only. In a recent publication we compared the distribution of active ingredient classes for these sales data with the consumption data of our study for all animal species under study, which are closely related. Taken into account that data on turkey and other species was not available, an extrapolation from study data to sales data was sufficient [
Jordan et al. [
It is important to note that these are active ingredient classes, which are highly important (tetracyclines and sulfonamides) or even critically important (beta-lactams) for human medicine [
Some authors also report, that these active ingredient classes were often used for metaphylactic matters [
In sows, tetracyclin, trimethoprim-sulfonamides and amoxicillin were mostly used in this study. In piglets, amoxicillin, tetracyclin, trimethoprim-sulfonamides and colistin were mostly used.
Dunlop et al. [
According to Dunlop et al. [
Also, in the study of Jensen et al. [
Jensen et al. [
Similar to this study, Dunlop et al. [
In weaners, amoxicillin, tetracyclin, chlorotertacyclin as well as trimethoprim-sulfonamides and colistin in almost equal quantities, were mostly applied in this study. In fattening pigs the active ingredients amoxicillin, tylosin, chlorotetracyclin and tetracyclin were mostly used.
According to Dunlop et al. [
According to Callens et al. [
98% of the used quantity was administered orally, whereas in this connection especially treatments of weaners and fattening pigs led to this distribution. This observation corresponds with the in ESVAC reported high amounts of oral powder of antibiotic substances [
Among all age groups, respiratory diseases represent the main indication for the use of antimicrobials. It should be stressed here again that in Germany prophylactic usage of antibiotics is not allowed. The evaluation of reasons for using antimicrobial substances based on the documentation on official forms, on which any treatment administered to the animal is documented obligatory by the veterinarian. Therefore, the documentation and report of misuse, off-label use and prophylactic use is not possible with this study design.
Respiratory diseases are primarily treated with beta-lactam antibiotics in piglets, weaners and fattening pigs, but not in sows. Except for sows, intestinal diseases are the next most frequent indication for antibiotic treatment. In our study, mainly macrolides (tylosin) and polypeptides (colistin) are used for treating intestinal diseases. Chauvin et al. [
The second most common indication for sows is diseases of the reproductive organs, which are primarily treated with tetracyclines. According to Chauvin et al. [
Considering seasonal effects in view of antimicrobial consumption might give an allusion to seasonal dependencies on disease patterns. However, it must be emphasized that different consumption levels by season may be caused by differences in the number of animals kept in the individual time periods. This could not be answered with our data due to the documentation of average numbers of animals kept over one year only.
Thus, the highest drug use in piglets was detected in autumn in our study. In autumn, amoxicillin was used in highest quantities. Furthermore, amoxicillin was used for treating respiratory diseases, which was the main indication for treating piglets in our study. Based on these results, it could be speculated that there was an increase in respiratory diseases due to falling temperatures in autumn and the associated higher susceptibility to infections in younger animals. Nevertheless, this could not be confirmed for weaners and fattening pigs; there the hightest amount of antimicrobials usually consumed for treating respiratory diseases was in winter and spring. In contrast to theses observations, recently, Sanchez-Vazquez et al. [
As regards colistin and tylosin, the antimicrobials commonly used for treating intestinal disorders, again there was no consistent picture. Low amounts were used in winter for piglets and fattening pigs, but not for weaners. As regards weaners and fattening pigs, quite high amounts were used in spring to treat dysenteriae.
When looking at sows, the highest drug use in our study was detected in spring, with the highest amounts of tetracycline used in spring. In addition, tetracycline is mainly used for treating respiratory diseases and diseases of the reproductive organs, which were at the top of the indications in sows in our study. However, sows are continuously used throughout the year for producing piglets, so that a seasonal accumulation of diseases of the reproductive organs could be ruled out.
In the feasibility study, no seasonal effects were observed with regard to used quantities in all age groups covered in the study [
In our study, the highest median treatment frequency—for a standardized period of 100 days—was calculated for piglets with 14.7 days, followed by weaners with 6.6 days, and fattening pigs with 3.7 days. The lowest rate was calculated for sows with median 0.9 days per 100 days. In the previous feasibility study, arithmetic means had been reported. There, the treatment frequency for piglets was lower with 6.1 days per 100 days, but at the same magnitude for sows with a treatment frequency of 0.9 days per 100 days. For fattening pigs, the average treatment frequency, which was observed in the feasibility study, was higher with 4.6 days per 100 days on average [
In summary, while the treatment frequencies of sows and fattening pigs calculated in these three studies are similar to each other, the treatment frequency of piglets is more than twice as high in our study than in the feasibility study. This could be explained by a misclassification of weaners and piglets, which may have had impact on the results, because data is from veterinary practitioners, who were note standardized in beforehand. In this case, treatment frequency for piglets may have been overestimated and the treatment frequency of piglets may have been less than the calculated value, while the actual treatment frequency of weaners was possibly much higher than the calculated median value.
Merle et al. [
Bos et al. [
Looking at the treatment frequencies concerning individual active ingredient classes, beta-lactams represent the active ingredient class that was used on most farms among all age groups, with a treatment frequency of 4.3 days for piglets, 0.2 days for sows, 2.2 days for weaners and 0.8 days for fattening pigs, respectively. In contrast, Vieira et al. [
However, in the feasibility study Merle et al. [
As another example, on all farms which used polypeptides for treating infections, piglets were treated 9.7 days (median) per 100 days with polypeptides. Here, especially colistin plays the major role. Polypeptides are used for the treatment of respiratory infections as well as intestinal infections [
As polypeptides are quite lowly dosed compared to sulfonamides and tetracyclines, which are mainly used against intestinal and respiratory diseases, according to Jensen et al. [
Vieira et al. [
Considering the treatment frequency within the age groups it becomes obvious that piglets are far the most commonly treated age group, followed by weaners and fattening pigs. In contrast, sows are treated rarely. In addition, the model has shown that in this study the farm size only had a significant influence on the treatment frequency in interaction with the veterinarian. This could be justified due to the fact that several veterinarians advise farms of different sizes and that they follow different treatment strategies. Furthermore, the veterinarian alone has a significant impact on the treatment frequency. This may also be justified with different specializations of the different veterinarians. Thus, some veterinarians are specialized in the supervision of fattening pigs, whereas others are specialized in the supervision of sows and piglets, whereby here also different treatment strategies are followed.
Callens et al. [
The invariably right-skewed distribution of the treatment frequencies shows for all age groups the majority of farms having rather low and only a few farms having higher treatment frequencies. This observation is also documented by Bos et al. [
In Germany a representative scientific collection of data on the antibiotic use in food-producing animals is feasible. The legally binding requirement to document all treatments and deliveries provides an excellent data source, allowing for a penetrative analysis at an individual animal level. Therefore, conclusions about the actual usage of antibiotics are allowed, both on the level on amounts of drugs as well as on the level of number of used daily doses (per antibiotic class).
In total amounts, tetracyclines, beta-lactams and trimethoprim-sulfonamides are the most commonly used active ingredient classes in pigs in Germany. This general observation does not differ from the results of other antibiotic consumption studies in other countries. However, if data is stratified in the different age-classes of pig's life, the using pattern changes and beta-lactams, macrolides and also polypeptides seem to have become more important in treating pigs in Germany.
In the VetCAb study, the median treatment frequency—for a standardized period of 100 days—was calculated for piglets with 14.7 days. Weaners are treated at 6.6 days, and fattening pigs at 3.7 days. The lowest rate was calculated for sows with median 0.9 days per 100 days. These treatment frequencies only differ little from the treatment frequencies calculated in earlier antibiotic consumption studies, which were carried out in Germany.
Most consumption data based on amounts and uses theoretical DDDA for estimating the antibiotics usage. The data used in the VetCAb study based on UDDA, a fact that does not allow a direct comparison. However, most important is that nUDDA uses the number of treated animals as well as the applied dosage directly without any estimation. This may reduce the variance of the given information. Thus, actual treatment frequencies possibly differ from estimated treatment frequencies.
The data shows, that same antimicrobial substances both in human and veterinary medicine are used, which may cause a development and a transfer of resistance. But, linking the data collected in our study with human health is not allowed due to a lack of similar information for human treatment.
Our study is the first cross-sectional study on antimicrobial usage in pig farming in entire Germany. As a first reference point our data may be used to develop rules for regular or elevated use of antibiotic substances of German farming practice to describe the situation. However, without accompanying studies on the resistance level in livestock in Germany, it remains difficult to realize comparative observations on the two research sectors.
The authors wish to thank the veterinarians and the farmers who participated in the study and provided data voluntarily. Further thanks go to the employees of the veterinary departments of the study districts and farmers' associations, which were a great help in contacting appropriate participants. We would also like to acknowledge the veterinary chambers for assistance in the recruitment of veterinarians all over Germany. Finally, we would like to thank the Federal Institute for Risk Assessment for the research assignment and financial support.