DISCUSSION
This study used data of farm-level technical and economic factors to obtain insight into factors that might be associated with the use of antibiotics on pig farms in the Netherlands. The data set included pig farms that take part in LEI FADN data collection as well as additional data collection on animal medicines, with LEI FADN farms considered representative for all farms in the Netherlands. Based on the 2 selected sets of farm factors, the study (sub)sample was representative for the entire LEI FADN sample of pig farms. The main advantage of this study is that it used real-life data collected on private farms, mostly based on farm accounts, and thus not subject to recording bias.
Antibiotics vary in their potency and pharmokinetic properties, and this is manifested in varying dosages per kilogram of BW. The measurement unit used in this study, NDDi, is suitable for calculating total exposure of pigs to different antibiotics, and making comparisons, for example, between groups of pigs. This unit conforms to international developments in this field and developments in the human sector (Mevius et al., 2007). Other measurement units to quantify antibiotic usage, such as number of treatment days or associated costs, are used. Different measurement units have shown to give different results, highlighting the importance of choosing the right measurement unit (Chauvin et al., 2008). For the aim of the current study (i.e., the selection of farm factors that are associated with the use of antibiotics, NDDi) was considered most appropriate. When the antibiotic treatments themselves are studied in more depth, including the characterization of different classes of antibiotics and their trends in time, other measurement units may be chosen. The NDDi of antibiotics is expressed per average pig present on the farm. This theoretical pig, representing an average pig (having an average BW) that is present on the farm the entire year, receives antibiotics for the specific number of days. For the sow farms, average daily dosages were expressed per average piglet present on the farms since birth, considering the orally given antibiotics only. As oral administration to sows is negligible and the piglets are sometimes also treated with injectable antibiotics, this number may have given an underestimation to the true antibiotic treatment of piglets. Obviously, the average piglet present is theoretical as piglets will not stay on the farm the entire year. The average number of dosages for the true period a piglet is present on the farm can be calculated from this average NDD2. For example, an average of 130 daily dosages per average piglet (overall average NDD2 found in this study) implies that a piglet that is present on the farm for 60 d receives, on average, antibiotics on 21.4 d (35.6% of the period).
This study showed that, both on the fattening pig and sow farms, NDDi increased from 2004 to 2006. Corrected for other effects, this increase was nearly 4 daily dosages per animal at the fattening pig farms and about 23 daily dosages per average piglet present on the sow farms (an entire year). As compared with 2006, NDDi decreased in 2007, reaching the level of 2005 at the fattening pig farms, and the level of 2004 at the sow farms (Tables 3 and 4). For the pig fattening farms, this finding is in contrary to Mevius et al. (2007) reporting an increase of antibiotic use of 4.5 daily dosages per pig year from 2006 toward 2007. The authors attributed this finding to the prohibition of the use of antimicrobial growth promoters (AMGP) in The European Union in 2006. Reduced use of AMGP might increase pig diseases leading to an increased use of therapeutic antibiotics (Casewell et al., 2003). This is supported by findings from Scandinavia, where a temporary increase in antibiotic treatments for clinical disease, mainly diarrhea after weaning, was found after introduction of the ban for AMGP (Bengtsson and Wierup, 2006; Møller Jensen, 2006; Vigre et al., 2008). This study showed that also in the Netherlands, the increase in the use of antibiotics was temporary, lasting during the year 2006.
Both at the fattening pig and the sow farms, NDDi varied greatly between individual farms. In 2004 and 2005, some farms were able to raise pigs without applying antibiotics (Table 2). On the other hand, maximum NDDi values showed that there are farms that are very intensive antibiotic users [e.g., the maximum value of NDD2 in 2007 exceeded 365 implying that the piglets received even more than 1 dosage per day (1.09 doses) over the entire period they were on the farm]. Consistently, the Lorentz curves also show that some farms are “heavy users” as 20% of the farms use about 45% of the antibiotics, both on the fattening pig and sow farms. The large between-farm variation in the use of antibiotics in the Netherlands corresponds to values reported from other countries, including Belgium (Timmerman et al., 2006) and France (Chauvin et al., 2002). Variations may be due to differences in hygiene status, prophylactic use, and treatment decisions made by the farmer or the veterinarian or both (Timmerman et al., 2006). This study showed that both on the sow and fattening pig farms the use of antibiotics on individual farms is relatively stable over the 4 yr of the study period (variation due to year relative to variation due to farm is very low). In other words, farms that use greater amounts of antibiotics in 1 yr will also do so in the next year, whereas farms that use decreased amounts in a particular year may also use less in the other years.
Our study showed that antibiotic use, both on sow and fattening pig farms, was mainly influenced by farm system and number of pigs present on the farm, and for sow farms only, antibiotic use was also affected by the population density in the region of the farm. The greater number of pigs present on the farm may result in a greater probability of infection, which can explain the increased antibiotic treatments on large farms. Increased infection at larger farms has been reported previously (Österberg et al., 2006; Hautekiet et al., 2008; García-Feliz et al., 2009). It may also be hypothesized that workers on large farms spend less time inspecting individual animals and use antibiotics in a more preventive manner as compared with the workers on small farms. However, it remains unclear whether this applies to the current study because the number of pigs present on the farm was positively correlated with factors related to the amount of human labor available. Fattening pigs received the least number of antibiotic treatments when they were raised on farrow-to-finish farms, followed by specialized fattening pig farms, and the greatest amount on mixed farm systems. These findings probably relate to a greater infection risk with increased animal movement to mixed farms and mixture at the farm. For the sow farms, the results of the current study were contrary to this hypothesis, as piglets on specialized sow farms received fewer antibiotics compared with piglets on farrow-to-finish farms. Probably, this finding is due to the fact that on the farrow-to-finish farm systems the term piglet is used for a longer period of time than on the specialized sow farms. The region of the farm location was found to affect the number of treatments given to the piglets, with a greater use on farms located in densely populated livestock areas as compared with farms in less densely populated livestock areas. As was seen in an outbreak of classical swine fever in the Netherlands, the probability of a neighborhood infection decreases with an increased distance to an infected herd (Elber et al., 1999). Therefore, regions with a large number of pig farms may have a greater spread of infection (de Jong et al., 2009). Antibiotic use on sow farms is thus influenced both by the number of animals within a farm as well as by the number of farms within a specified region.
The underlying idea of investigating farm factors that are related to the application of antibiotics on pig farms was to select factors that could be used to identify farms that use more antibiotics. These factors and their underlying reason could probably be used in policy making, to provide farm-specific advice, or for more intensive control of these farms (e.g., in educating farmers and veterinarians on strategies to minimize antimicrobial resistance; European Commission, 2009). Our study results showed that, indeed, some factors were significantly associated with the number of antibiotics applied and, therefore, could be used to identify farms for their antibiotic use. These factors, including farm system, farm size, and region (in case of sow farms), are easy to collect and to register. The study results imply that farm advice and farm investigation to reduce the inappropriate use of antibiotics could be distinguished by these factors. For example, advice to farmers (and their veterinarians) could be differentiated to farm system. As farms that used greater antibiotics in 1 yr were shown to be very likely to also do so in the next years, it is very worthwhile to invest in specific and effective advisory programs. In Scandinavian countries, campaigns on optimal disease preventive management routines and guidelines on antibiotic therapy in food animals proved to be useful in diminishing antibiotic use (Bengtsson and Wierup, 2006).
Out of the initial selected sets of 16 and 19 factors for fattening pig and sow farms, respectively, only a few factors showed to be significantly associated with the use of antibiotics. Excluding factors that were correlated with each other and selecting the model with the best set of explanatory factors were done with great care, considering underlying biology/physiology as well as statistical arguments. Therefore, the applied selection methodology is considered to have resulted in the best set of explanatory factors, from both biological
DISCUSSION
This study used data of farm-level technical and economic factors to obtain insight into factors that might be associated with the use of antibiotics on pig farms in the Netherlands. The data set included pig farms that take part in LEI FADN data collection as well as additional data collection on animal medicines, with LEI FADN farms considered representative for all farms in the Netherlands. Based on the 2 selected sets of farm factors, the study (sub)sample was representative for the entire LEI FADN sample of pig farms. The main advantage of this study is that it used real-life data collected on private farms, mostly based on farm accounts, and thus not subject to recording bias.
Antibiotics vary in their potency and pharmokinetic properties, and this is manifested in varying dosages per kilogram of BW. The measurement unit used in this study, NDDi, is suitable for calculating total exposure of pigs to different antibiotics, and making comparisons, for example, between groups of pigs. This unit conforms to international developments in this field and developments in the human sector (Mevius et al., 2007). Other measurement units to quantify antibiotic usage, such as number of treatment days or associated costs, are used. Different measurement units have shown to give different results, highlighting the importance of choosing the right measurement unit (Chauvin et al., 2008). For the aim of the current study (i.e., the selection of farm factors that are associated with the use of antibiotics, NDDi) was considered most appropriate. When the antibiotic treatments themselves are studied in more depth, including the characterization of different classes of antibiotics and their trends in time, other measurement units may be chosen. The NDDi of antibiotics is expressed per average pig present on the farm. This theoretical pig, representing an average pig (having an average BW) that is present on the farm the entire year, receives antibiotics for the specific number of days. For the sow farms, average daily dosages were expressed per average piglet present on the farms since birth, considering the orally given antibiotics only. As oral administration to sows is negligible and the piglets are sometimes also treated with injectable antibiotics, this number may have given an underestimation to the true antibiotic treatment of piglets. Obviously, the average piglet present is theoretical as piglets will not stay on the farm the entire year. The average number of dosages for the true period a piglet is present on the farm can be calculated from this average NDD2. For example, an average of 130 daily dosages per average piglet (overall average NDD2 found in this study) implies that a piglet that is present on the farm for 60 d receives, on average, antibiotics on 21.4 d (35.6% of the period).
This study showed that, both on the fattening pig and sow farms, NDDi increased from 2004 to 2006. Corrected for other effects, this increase was nearly 4 daily dosages per animal at the fattening pig farms and about 23 daily dosages per average piglet present on the sow farms (an entire year). As compared with 2006, NDDi decreased in 2007, reaching the level of 2005 at the fattening pig farms, and the level of 2004 at the sow farms (Tables 3 and 4). For the pig fattening farms, this finding is in contrary to Mevius et al. (2007) reporting an increase of antibiotic use of 4.5 daily dosages per pig year from 2006 toward 2007. The authors attributed this finding to the prohibition of the use of antimicrobial growth promoters (AMGP) in The European Union in 2006. Reduced use of AMGP might increase pig diseases leading to an increased use of therapeutic antibiotics (Casewell et al., 2003). This is supported by findings from Scandinavia, where a temporary increase in antibiotic treatments for clinical disease, mainly diarrhea after weaning, was found after introduction of the ban for AMGP (Bengtsson and Wierup, 2006; Møller Jensen, 2006; Vigre et al., 2008). This study showed that also in the Netherlands, the increase in the use of antibiotics was temporary, lasting during the year 2006.
Both at the fattening pig and the sow farms, NDDi varied greatly between individual farms. In 2004 and 2005, some farms were able to raise pigs without applying antibiotics (Table 2). On the other hand, maximum NDDi values showed that there are farms that are very intensive antibiotic users [e.g., the maximum value of NDD2 in 2007 exceeded 365 implying that the piglets received even more than 1 dosage per day (1.09 doses) over the entire period they were on the farm]. Consistently, the Lorentz curves also show that some farms are “heavy users” as 20% of the farms use about 45% of the antibiotics, both on the fattening pig and sow farms. The large between-farm variation in the use of antibiotics in the Netherlands corresponds to values reported from other countries, including Belgium (Timmerman et al., 2006) and France (Chauvin et al., 2002). Variations may be due to differences in hygiene status, prophylactic use, and treatment decisions made by the farmer or the veterinarian or both (Timmerman et al., 2006). This study showed that both on the sow and fattening pig farms the use of antibiotics on individual farms is relatively stable over the 4 yr of the study period (variation due to year relative to variation due to farm is very low). In other words, farms that use greater amounts of antibiotics in 1 yr will also do so in the next year, whereas farms that use decreased amounts in a particular year may also use less in the other years.
Our study showed that antibiotic use, both on sow and fattening pig farms, was mainly influenced by farm system and number of pigs present on the farm, and for sow farms only, antibiotic use was also affected by the population density in the region of the farm. The greater number of pigs present on the farm may result in a greater probability of infection, which can explain the increased antibiotic treatments on large farms. Increased infection at larger farms has been reported previously (Österberg et al., 2006; Hautekiet et al., 2008; García-Feliz et al., 2009). It may also be hypothesized that workers on large farms spend less time inspecting individual animals and use antibiotics in a more preventive manner as compared with the workers on small farms. However, it remains unclear whether this applies to the current study because the number of pigs present on the farm was positively correlated with factors related to the amount of human labor available. Fattening pigs received the least number of antibiotic treatments when they were raised on farrow-to-finish farms, followed by specialized fattening pig farms, and the greatest amount on mixed farm systems. These findings probably relate to a greater infection risk with increased animal movement to mixed farms and mixture at the farm. For the sow farms, the results of the current study were contrary to this hypothesis, as piglets on specialized sow farms received fewer antibiotics compared with piglets on farrow-to-finish farms. Probably, this finding is due to the fact that on the farrow-to-finish farm systems the term piglet is used for a longer period of time than on the specialized sow farms. The region of the farm location was found to affect the number of treatments given to the piglets, with a greater use on farms located in densely populated livestock areas as compared with farms in less densely populated livestock areas. As was seen in an outbreak of classical swine fever in the Netherlands, the probability of a neighborhood infection decreases with an increased distance to an infected herd (Elber et al., 1999). Therefore, regions with a large number of pig farms may have a greater spread of infection (de Jong et al., 2009). Antibiotic use on sow farms is thus influenced both by the number of animals within a farm as well as by the number of farms within a specified region.
The underlying idea of investigating farm factors that are related to the application of antibiotics on pig farms was to select factors that could be used to identify farms that use more antibiotics. These factors and their underlying reason could probably be used in policy making, to provide farm-specific advice, or for more intensive control of these farms (e.g., in educating farmers and veterinarians on strategies to minimize antimicrobial resistance; European Commission, 2009). Our study results showed that, indeed, some factors were significantly associated with the number of antibiotics applied and, therefore, could be used to identify farms for their antibiotic use. These factors, including farm system, farm size, and region (in case of sow farms), are easy to collect and to register. The study results imply that farm advice and farm investigation to reduce the inappropriate use of antibiotics could be distinguished by these factors. For example, advice to farmers (and their veterinarians) could be differentiated to farm system. As farms that used greater antibiotics in 1 yr were shown to be very likely to also do so in the next years, it is very worthwhile to invest in specific and effective advisory programs. In Scandinavian countries, campaigns on optimal disease preventive management routines and guidelines on antibiotic therapy in food animals proved to be useful in diminishing antibiotic use (Bengtsson and Wierup, 2006).
Out of the initial selected sets of 16 and 19 factors for fattening pig and sow farms, respectively, only a few factors showed to be significantly associated with the use of antibiotics. Excluding factors that were correlated with each other and selecting the model with the best set of explanatory factors were done with great care, considering underlying biology/physiology as well as statistical arguments. Therefore, the applied selection methodology is considered to have resulted in the best set of explanatory factors, from both biological
การแปล กรุณารอสักครู่..