Use of Anticoccidial Drugs in Broiler Chickens in the USA: Analysis for the Years 1995 to 1999

Use of Anticoccidial Drugs in Broiler Chickens in the USA: Analysis for the Years 1995 to 1999 H. D. Chapman 1 Department of Poultry Science, University of Arkansas, Fayetteville, Arkansas 72701 ABSTRACT Data collected by Agri Stats Inc. in the US tent pattern in the sequence with which different programs for the years 1995 to 1999 was evaluated to determine the types of anticoccidial drug programs used by broiler plants, their frequency and duration, and whether there was any correlation with performance of the birds. Information was available for five feed types (prestarter, starter, grower, first withdrawal, and final withdrawal). The most popular program was an ionophore (ION; principally salinomycin) in starter and grower feeds. A chemical (principally nicarbazin) was also used in the starter feed followed by an ION in the grower feed, or two different ION in the starter and grower feeds. Seasonal differences were apparent in the type of program and in the frequency of program changes. There was no consisences were employed. There were no significant differ- in calorie conversion or the number of days to produce a 2.27 kg bird, whether a single ION or a chemical followed by an ION was used, but mortality was significantly higher for the latter. For 1999, calorie conversion and mortality were higher in plants where chemical-ion programs had been used for more than 40% of the time during the previous 4 yr. The duration and frequency with which different programs were employed were similar whether birds were raised to final weights of 1.5 to 2.0 kg, 2.0 to 2.5 kg, or more than 2.5 kg. There were significant differences in the use of salinomycin and nicarbazin in different regions of the US. (Key words: anticoccidial drug, chemical, ionophore, broiler, coccidiosis) 2001 Poultry Science 80:572 580 INTRODUCTION Coccidiosis is an important disease of chickens caused by protozoan parasites of the genus Eimeria. Anticoccidial drugs have been used for the control of coccidiosis for more than 50 yr, and today almost all commercial broilers are reared with an anticoccidial agent in their feed. Many compounds have been introduced, but in the US about a dozen are presently used. They can be categorized as polyether ionophorous antibiotics [ionophores (ION)] that are produced by fermentation or synthetic compounds (often referred to as chemicals) that are produced by chemical (CHE) synthesis. The terms ION and CHE will be used to refer to these two types of drugs. Examples of ION used in the US are lasalocid, monensin, narasin, salinomycin (SAL), and semduramicin. Chemicals include decoquinate, halofuginone, nicarbazin (NIC), robenidine, and zoalene. In some cases, birds are given one drug, but two or more drugs (so-called shuttle programs) may be given 2001 Poultry Science Association, Inc. Received for publication June 15, 2000. Accepted for publication December 19, 2000. 1 To whom correspondence should be addressed: dchapman@mail. uark.edu. 2 Agri Stats, Inc., Fort Wayne, IN 46825. during the life of a flock. It is convenient to provide a particular compound for a period during which one type of feed is given. For example, a CHE may be included in the starter, and an ION may be included in the grower ration. Program will be used to refer to the compounds used for the life of a single flock. By using different permutations of drugs, about 40 different programs have been employed in the US in recent years. A database has been developed that provides accurate, frequent information on a monthly basis from individual poultry plants. This database has been used to assess many aspects of poultry production. 2 A poultry plant may be defined as a reporting unit that, in most cases, represents a broiler complex comprising a group of farms in a common geographical area, which are served by a single feed mill. In addition to the drugs used by each plant, included are the final BW and measurements of performance, such as calorie conversion (a measure of feed efficiency), the number of days taken to grow a flock to a specified average weight, and percentage mortality. Results for the years 1995 to 1999 were examined to provide information on the types of programs and their frequency of use and whether there was any correlation with performance characteristics of the birds. Abbreviation Key: CHE = chemical; ION = ionophore; NIC = nicarbazin; SAL = salinomycin. 572

ANTICOCCIDIAL DRUGS IN BROILERS 573 METHODS Data were available for five separate feeds (referred to as prestarter, starter, grower, first withdrawal, and final withdrawal feeds, respectively). To determine the number of plants using an anticoccidial in these feeds, data for the months of January and July, from 1995 to 1999, were expressed as a percentage of the total number of plants reported for that month, and the overall mean was calculated. To simplify the analysis, only drugs used in the starter and grower feeds were taken into account when referring to programs. For January 1995, information from 76 plants was available, but by December 1999, information for 161 plants was recorded in the database. Programs were classified as follows: the same ION in the starter and grower feed (single-drug program; ION), different ionophores in the starter and grower feeds (two-ionophore shuttle program; ION-ION), a CHE in the starter followed by an ION in the grower feed (CHE-ION shuttle program; CHE-ION). Programs with the same CHE, two different CHE, or an ION followed by a CHE in the starter and grower feeds were only occasionally used. Maxiban, a mixture of NIC and narasin, was classified as a CHE for the purpose of the analysis. Performance criteria reported in the monthly feed medication usage vs. performance report include calorie conversion, the number of days to 2.27 kg adjusted for bird weight, and percentage mortality (number of birds settled, divided by number placed). Methods for calculating these variables are given in the live production monthly definition and calculation manual. 2 Programs The number of plants using each program type (ION, ION-ION, or CHE-ION), for each month (from 1995 to 1999), was expressed as a percentage of the total number of plants reporting for that month. Data for each month for 5 yr (n = 5), and for all 12 mo (n = 12), were pooled, and the means were calculated. Plants were categorized according to whether the final BW of birds was within 1.5 to 2.0 kg, 2.0 to 2.5 kg, or greater than 2.5 kg. The percentage using a program of ION or CHE-ION was then determined to establish whether there was any correlation with the type of program used. Results for April for the 5-yr period were combined (n = 5), and means were compared. April was chosen for analysis, because for this month overall use of ION and CHE-ION programs was similar (mean = 46.1 and 46.5% of plants, respectively). Twenty-five plants were randomly selected that had used an ION program, and another 25 plants that had used a CHE-ION program were randomly selected. The average calorie conversion, number of days to 2.27 kg, and percentage mortality were then determined. Results for April for the 5-yr period were combined (n = 5), and means were compared. Fifty plants were randomly selected for which continuous records were available from January 1995 to Decem- FIGURE 1. For each month, the number of broiler plants using a single ionophore (ION), two ionophores (ION/ION), or a chemical followed by an ionophore (CHE/ION) was expressed as a percentage of the total number of plants for that month. Results for each month for the 5-yr period were combined (n = 5), and the means ± SEM were calculated. ber 1999. The number of consecutive months and the sequence for which each program type was used by each plant were determined. For each of the 50 plants, the programs used were compared in successive months for the 5-yr period to establish if program changes (for example from ION to ION-ION, ION to CHE-ION, ION-ION to CHE-ION, and vice versa) were made on a seasonal basis. The number of plants that reported changes from January to February, February to March, and so on throughout the year was expressed as a percentage of the total number of plants. Results for all 5 yr were pooled (n = 5), and the means were calculated and compared. Thirty-six plants were identified in which a CHE-ION program was changed to an ION or ION-ION program, and vice versa. For each plant, data were pooled for calorie conversion, number of days to 2.27 kg, and percentage mortality for the 3 mo preceding the change and the 3 FIGURE 2. For each month, the number of broiler plants using a single ionophore (ION), two ionophores (ION/ION), or a chemical followed by an ionophore (CHE/ION) was expressed as a percentage of the total number of plants for that month. Results for all 12 mo were combined (n = 12), and the means ± SEM were calculated.

574 CHAPMAN FIGURE 3. Broiler plants were categorized according to whether the final BW of birds was within 1.5 to 2.0 kg, 2.0 to 2.5 kg, or greater than 2.5 kg, and the percentage was determined for use of a single ionophore (ION) or chemical followed by an ionophore (CHE/ION). Results for April for 1995 to 1999 were combined (n = 5), and the means ± SEM were calculated. mo after the change. The means for all 36 plants were calculated and compared. From January 1995 to December 1998, the number of months for which each type of program was used at individual plants was expressed as a percentage of the total number of months (48) for this period. Seventeen plants were then identified for which the percentage use of a CHE had been less than 10%, and a further 17 plants were identified for which the use of CHE had been greater than 40%. These data were compared to determine whether the extent to which CHE had been used during the 4-yr period might influence subsequent performance. For each month of 1999, the average calorie conversion, number of days to 2.27 kg, and percentage mortality, was calculated. Results for the 12 mo were combined, and the means were calculated and compared for each category. Individual Drugs For January and July 1995 to 1999, the percentage of plants using individual drugs was determined, the data were combined, and the means were calculated. For SAL in starter and grower feeds and NIC in starter feed, the percentage was calculated for plants using these drugs for each month of the 5-yr period. Data for plants that reported weekly percentage mortality for April 1998 were analyzed. The percentage mortality was compared from 0 to 7 d, 8 to 14 d, 15 to 21 d, 22 to 28 d, 29 to 35 d, 36 to 42 d, and after 42 d from plants FIGURE 4. The number of program changes reported by 50 broiler plants (expressed as a %) was determined for successive months [January to February (JF), February to March (FM), and so on]. Results for the 5-yr period were combined (n = 5), and the means ± SEM were calculated. that had used NIC in the starter (n = 19 plants) and SAL in the starter and grower feeds (n = 52). The number of plants using different concentrations of SAL during July 1999 was determined. Concentrations of other drugs were recorded for January and July 1999, but because of the small sample size, the number of plants involved is not presented. In the US, broiler production is concentrated in the northeast, southeast, and south-central regions of the country. Data for three regions were available for 1995 to 1999 (Region 20, Delaware, Maryland, North Carolina, Virginia, Pennsylvania, and West Virginia; Region 30, Georgia, South Carolina, Tennessee, and Florida; and Region 50, Arkansas, Texas, and Missouri). The percentage of plants was calculated for use of SAL in the starter and grower feed or NIC in the starter feed. Statistical Analysis Data were analyzed by one-way analysis of variance by using the general linear models procedure of the SAS Institute (1988). Significant effects of treatment means were separated using Duncan s multiple-range test. Feeds RESULTS Most plants provided data for three feeds (starter, grower, and final withdrawal). Approximately 45% of plants used a fourth feed (first withdrawal) before the final withdrawal feed. A few plants (approximately 5%) utilized a fifth feed prior to the starter feed (prestarter). The actual period for which each feed was given varied at different plants. The mean duration of feeding periods during April 1998 was 17 d (range 9 to 25 d) for starter feed, 17 d (range 11 to 30 d) for grower feed, 10 d (range 4 to 22 d) for the first withdrawal, and 8 d (range 1 to 28 d) for the final withdrawal feed (based upon data reported by 155, 155, 152, and 105 plants, respectively). Programs An anticoccidial was used in the starter and grower feeds by 99% of plants and in the first and final withdrawal feeds by 41 and 6% of plants, respectively. Most plants (73%) used three types of programs (ION, ION- ION, and CHE-ION). Two program types (ION, ION- ION; ION, CHE-ION; and ION-ION, CHE-ION) were used by 20, 4, and 2% of plants, respectively. One plant used a single program (ION) during the entire 5 yr. The overall program use of ION was 60.5%, CHE-ION was 27.5%, and ION-ION was 9.2%. Significant differences were evident at various times of the year (Figure 1). An ION program was used more frequently than an ION-ION program throughout the year and more than a CHE-ION program, except for the months of March, April, and May, when use was similar (42.3 to 46.6% of plants). Use of an ION program increased from June to

ANTICOCCIDIAL DRUGS IN BROILERS 575 TABLE 1. Sequential use of different drug programs reported by broiler plants where birds were reared to different final weights from January 1995 to December 1999 1 Final body weight 1.5 to 2.0 kg 2.0 to 2.5 kg > 2.5 kg CDABCABCABCABCAB CABCABA ABABACA CACBABCBABABA CABCABABABA ABA ABCABCABCABACAAB CABCABCA A ACBACACABCABABC CACBACACBCACA ACBABAC?A ABACABABCABCABACA ABABABA CBC ABABACABABCABABACABACBA ABABACABABCABA ABC?A ABCA?A ABABABABABABACABAC CBABCACACACB CBACABABA CACABABABABABADBABABABA ABCACBABCABACACAC CBAC?ABCA?CABA BACABCABABADABABABDA ABCABABCBCACA ABADABA ABA CBABCACABCACB CABABCABCACACABABC BA BCABABACACA ABCABCABACA ABABABABABA - ACACACD CABCABACABCACABCBABC - DCDBABABABABAB ABABA - CACACACABACABABC CACACACABACABABC - ABABABA CBABCACACACACB - 1 Successive types of drug programs used by 43 broiler plants where birds were reared to different final weights. Data is from January 1995 to December 1999. A = same ionophore used in the starter and grower feeds; B = different ionophores used in the starter and grower feeds; C = chemical used in the starter feed followed by an ionophore in the grower feed; D = ionophore in the starter feed followed by a chemical in the grower feed;? = program unknown. September (mean = 75.7% of plants) but subsequently declined. A CHE-ION program was used more frequently than an ION-ION program from December to July, with a peak during March and April (means = 45.9 and 46.5% of plants). This peak was followed by a decline. No seasonal pattern in the use of an ION-ION program was apparent. A progressive increase in the use of an ION program was apparent during the 5-yr period, but use of CHE-ION and ION-ION programs gradually declined (Figure 2). In a few cases (mean = 1.7% of plants), three different drugs were used in the starter, grower, and first or final withdrawal feeds, respectively. These programs involved a CHE followed by two different ION (data not presented). There was no significant difference in the use of an ION or CHE-ION program, whether the final weight of birds was 1.5 to 2.0 kg, 2.0 to 2.5 kg, or greater than 2.5 kg (Figure 3). Seasonal differences were evident in the frequency of program changes (Figure 4). Significantly more changes were made from May to June, June to July, October to November, and December to January than in March to April, April to May, and September to October (P 0.003). The frequency with which different programs were used was extremely variable. Thus ION, ION-ION, and CHE-ION programs were used on an average of 5.0 (range 1 to 11), 3.3 (range 1 to 9), or 2.7 (range 1 to 7) separate occasions during the 5 yr. The average number of changes made by 50 plants per year was 2.1 (range 0 to 4.4). There was a significant difference (P 0.001) in the duration of ION, CHE-ION, and ION-ION programs (means ± SEM were 10.3 ± 1.5, 6.5 ± 1.1, and 2.2 ± 0.4 mo, respectively). There was no significant difference in the frequency or duration of use of programs whether birds were reared to 1.5 to 2.0 kg, 2.0 to 2.5 kg, or greater than 2.5 kg (data not presented). No consistent pattern was evident in the sequence with which different types of programs were employed in which birds were reared to different weights (Table 1).Thus an ION program was equally likely to be followed by an ION-ION or a CHE-ION program. TABLE 2. Performance characteristics reported by broiler plants from April 1995 to 1999 where a single ionophore, or a chemical followed by an ionophore had been employed 1 Program type Variable n ION CHE-ION P Calorie conversion 5 2,855 ± 13 2,853 ± 13 0.91 Days to 2.27 kg 5 46.9 ± 0.7 46.3 ± 0.7 0.57 % Mortality 5 4.78 ± 0.14 b 5.21 ± 0.14 a 0.05 a,b Values followed by a different superscript differ significantly (P < 0.05). 1 The average calorie conversion, number of days to 2.27 kg, and percentage mortality reported by plants that had utilized a single ION in the starter and grower feeds (25 plants) or a chemical followed by an ionophore (CHE-ION) in the starter and grower feeds (25 plants) were calculated. Results for April 1995 to 1999 were combined (n = 5), and the means ± SEM were calculated and compared.

576 CHAPMAN TABLE 3. Performance characteristics reported by broiler plants for a period of 3 mo before and after a change of drug program 1 Program change CHE-ION to ION, ION-ION ION, ION-ION to CHE-ION Variable Before (n = 36) After (n = 36) P Before (n = 36) After (n = 36) P Calorie conversion 2,893 ± 18 2,908 ± 18 0.54 2,877 ± 16 2,874 ± 16 0.86 Days to 2.27 kg 46.3 ± 0.5 47.0 ± 0.5 0.28 45.8 ± 0.5 46.2 ± 0.5 0.51 % Mortality 5.32 ± 0.3 5.15 ± 0.3 0.69 5.36 ± 0.4 5.70 ± 0.4 0.52 1 The mean calorie conversion, number of days to 2.27 kg, and percentage mortality ± SEM for 3 successive months reported by 36 plants before and after a program change. Plants had changed from a chemical followed by an ionophore (CHE-ION) to a single ionophore (ION) or to two ionophores (ION-ION), and vice versa. No significant difference in calorie conversion or the number of days to 2.27 kg was found whether an ION, or CHE-ION program was used (Table 2). Mortality was significantly higher (P = 0.056) when a CHE-ION program had been employed. A comparison of performance characteristics for a 3- mo period before and after a change from an ION to CHE-ION program and vice versa is presented in Table 3. No significant differences in calorie conversion, days to 2.27 kg, or percentage mortality were observed. Performance characteristics were compared for which use of a CHE had been reported as less than 10% or more than 40% during the previous 4 yr. No significant difference was observed in the number of days to 2.27 kg. Calorie conversion and percentage mortality were significantly higher in plants where use of a CHE had been greater than or equal to 40% (Table 4). Individual Drugs Salinomycin was the most frequently used anticoccidial in a single-drug program (mean = 42% of plants) followed by narasin (mean = 7.9%), monensin (mean = 7.3%), and lasalocid (mean = 3.3%) (Table 5). Semduramicin was first reported in November 1996 and occasionally used in the starter, grower, or starter and grower feeds (mean = 3.5%). NIC (followed by various ION) was the most frequently used CHE (mean = 10.2%). Maxiban, and CHE such as zoalene, robenidine, and halofuginone were occasionally used (means = 5.5, 4.2, 2.5, and 1.6% of plants, respectively). Decoquinate was used from January to July 1995 (mean = 1.3%). A few plants (mean = 4.1%) used a mixture of sulfadimethoxine and ormetoprim (Rofenaid ) in the prestarter feed (from May 1995 to January 1997). At other times, this drug was only occasionally used (mean = 1%). Significant differences in the use of SAL and NIC were evident (Figure 5). Salinomycin was used more frequently than NIC from June to January and was more popular during summer than the first half of the year. By contrast, the use of NIC increased from January to April but then declined. A progressive increase in the use of SAL (from 26 to 54% of plants) occurred during the 5-yr period (Figure 6), whereas from 1995 to 1999 the use of NIC declined from 21 to 13%. Percentage mortality reported during the first week of age was significantly higher (P 0.022) in plants where NIC was used, compared with those where SAL had been used (Figure 7). No significant differences in mortality were apparent at other times. Various concentrations of SAL were employed in the starter and grower feeds (Table 6). In most cases (50 plants), these programs involved an increase in concentration (step-up programs). Some (13 plants) used a lower concentration in the grower feed than in the starter ration (step-down programs), whereas others (39 plants) used the same level of drug in both feeds (55 ppm, 61 ppm, or the maximum approved level of 66 ppm). Monensin, narasin, and lasalocid were used at a range of concentrations in the starter and grower feeds (Table 6). Maxiban and NIC were also used at a range of concentrations, but robenidine and zoalene were employed at their maximum approved concentrations. There was a significant difference in the use of SAL and NIC in different regions of the US during March and TABLE 4. Performance characteristics reported by broiler plants during 1999 where use of a chemical in drug programs was 10 or 40% 1 Use of a chemical Variable n 10% 40% P Calorie conversion 12 2,848 ± 15.2 b 2,903 ± 15.2 a 0.02 Days to 2.27 kg 12 48.4 ± 0.2 48.1 ± 0.2 0.31 % Mortality 12 4.4 ± 0.2 b 5.1 ± 0.2 a 0.005 a,b Values followed by a different superscript differ significantly (P < 0.05). 1 The average calorie conversion, days to 2.27 kg, and percentage mortality reported by plants that had utilised a chemical for 10% (17 plants) or 40% (17 plants) of the time during the previous 4 yr was calculated. Results for each month of 1999 were combined (n = 12), and the means ± SEM were calculated and compared.

ANTICOCCIDIAL DRUGS IN BROILERS 577 TABLE 5. Overall use of individual drugs in the starter and grower feed reported by broiler plants 1 Grower feed Salinomycin Narasin Monensin Lasalocid Total Starter feed % plants Salinomycin 42.0 1.7 1.8 0.2 45.7 Narasin 0.9 7.9 1.7 0 10.5 Monensin 0.6 0.9 7.3 0.2 9.0 Lasalocid 0.4 0.3 0.6 3.3 4.6 Nicarbazin 4.4 0.6 4.4 0.8 10.2 Maxiban 1.4 1.6 1.9 0.6 5.5 Zoalene 1.0 0.6 1.6 1.0 4.2 Robenidine 1.8 0 0 0.7 2.5 Halofuginone 1.2 0.2 0.2 0 1.6 Total 53.7 13.8 19.5 6.8 93.8 1 The percentage of plants using individual drugs in the starter and grower feed was determined. Data for January and July from 1995 to 1999 were pooled, and the overall means were calculated. April (Table 7). NIC was used more frequently in the north and southeast (Regions 20 and 30), whereas SAL was used more frequently in the south-central region (50). DISCUSSION Quantitative aspects of the use of anticoccidial drugs and the drug programs employed by the poultry industry have not been published in the scientific literature. The availability of an extensive database (representing more than 90% of the US broiler industry) enables individual companies to evaluate the performance of their flocks and to make cost comparisons with those of other producers. In this study, the database has been used to provide information concerning how anticoccidial agents have been employed by the broiler industry during the last 5 yr and to try to establish whether differences in performance may result from the use of various types of drug program. Anticoccidial drugs are almost universally included in the starter and grower feeds provided to commercial broilers (99% of plants). These feeds are given for approximately the first 33 d after hatch when birds are considered to be most at risk from coccidiosis. Drugs are also used to some extent by plants that use a first withdrawal feed but only occasionally in the final withdrawal. Those deciding when to withdraw anticoccidial drugs from the feed may take various factors into consideration. These factors include potential savings in the costs of medication and the perceived risk of clinical coccidiosis occurring if birds are exposed to infection after drug withdrawal. A few plants (5%) use a prestarter feed that usually contains a potentiated sulfonamide (Rofenaid ). This drug has antibacterial as well as anticoccidial efficacy and is presumably employed to control bacterial infections in very young birds. The most frequent program type involved use of the same drug (an ION) in the starter and grower feeds. The use of a single ION increased over the 5-yr study period, whereas use of a CHE followed by an ION declined. There may be several reasons for the decreased use of CHE. These include their high cost compared with ION, lack of new drugs with novel modes of action, problems in some cases of rapid resistance development, toxicity and side effects, and perceived interference with immunity development. Despite reports of the acquisition of resistance by field strains of Eimeria (Watkins, 1997), there is little evidence that the efficacy of ION has declined in the field. These compounds remain the most popular drugs for the control of coccidiosis. Ionophores are thought not to interfere with the development of natural FIGURE 5. For each month, the number of broiler plants using salinomycin in the starter and grower feed (SAL/SAL), or nicarbazin in the starter feed followed an ionophore (NIC/ION) was expressed as a percentage of the total number of plants for that month. Results for each month for the 5-yr period were combined (n = 5), and the means ± SEM were calculated. FIGURE 6. For each month, the number of broiler plants using salinomycin in the starter and grower feed (SAL/SAL) or nicarbazin in the starter feed followed an ionophore (NIC/ION) was expressed as a percentage of the total number of plants for that month. Results for all 12 mo were combined (n = 12), and the means ± SEM were calculated.

578 CHAPMAN FIGURE 7. Broiler plants that reported weekly percentage mortality for April 1998 were categorized according to whether salinomycin (SAL/SAL) was used in the starter and grower feed (n = 52), or nicarbazin (NIC/ION) in the starter feed followed by an ionophore (n = 19). The percentage mortality from 0 to 7 d, 8 to 14 d, 15 to 21 d, 22 to 28 d, 29 to 35 d, 36 to 42 d, and after 42 d (Weeks 1 to 7+) was compared. immunity, which in recent years has been considered desirable in broilers as it may permit extended withdrawal periods for drugs (Chapman, 1999). Approximately 9% of plants used different ION in the starter and grower feeds. The rationale for this change in ION is unclear because there have been no published studies to indicate that the use of two ION in shuttle programs can result in improved control of coccidiosis. Seasonal differences were apparent in the use of drugs. A single ION (SAL) was used most frequently during summer and winter (July to February), whereas a CHE (usually NIC) followed by an ION was most frequently used in spring (March and April). No seasonal pattern in the use of two ION was apparent. Changes in program type were made most frequently in early summer (May to July) and late fall (October to November). Regional differences in the use of drugs were evident at certain times of the year. Nicarbazin was used more frequently during March and April in the northeast and southeast, whereas SAL was more popular in the south-central region. There was no consistent pattern in the sequence with which different types of programs were used. A single ION was equally likely to be followed by a two ION, or a CHE-ION program. The frequency with which programs were employed was also variable. For example, a single ION was used from 1 to 11 occasions (mean = 5) during the 5-yr period. Significant differences in the duration of use of single ION, CHE-ION, and two ION programs were apparent (means = 10.3, 6.5, and 2.2 mo, respectively). Most plants used three program types (single ION, CHE followed by ION, and two ION) during the 5-yr period, but a few used just two (single ION, and two ION). Salinomycin was the most frequently used drug in single-drug programs, but ION such as monensin, narasin, and lasalocid were also employed. Nine different programs were identified in which various concentrations of SAL were used in the starter and grower feeds. The most popular (49% of plants) involved an increase in drug level from 44 to 55 ppm (step-up programs). The initial use of a low drug level will permit some parasite development and hence stimulation of an immune response in the bird (Chapman, 1999). A higher level of drug may then be used to prevent the possibility of clinical coccidiosis during the growing phase. This procedure is the opposite of the traditional step-down procedure in which the concentration of drug is progressively reduced to permit the gradual acquisition of immunity. Approximately 13% of plants used a lower concentration of SAL in the grower feed, and 38% of plants used a constant level of drug. Nicarbazin was the most frequently used CHE, but Maxiban (a mixture of NIC and narasin) and zoalene were also employed. Robenidine and halofuginone were occasionally used. Maxiban and NIC were sometimes used at less than their maximum approved use concentrations. No differences in calorie conversion, or the number of days to produce a 2.27 kg bird, were observed whether plants utilized a single ION or a CHE-ION shuttle program. Plants using a CHE-ION program reported significantly higher mortality. There were no differences in the number of days to produce a 2.27 kg bird, whether a CHE had been used for less than 10% of the time during the TABLE 6. Concentrations of various drugs in the starter and grower feeds reported by broiler plants during July 1999 1 Concentrations used (ppm) Concentrations used (ppm) Drug n Starter feed Grower feed Drug Starter feed Grower feed Salinomycin 40 44 55 Monensin 99 to 121 99 to 112 Salinomycin 1 50 61 Narasin 59 to 66 66 to 79 Salinomycin 9 55 61 Lasalocid 76 to 100 88 to 100 Salinomycin 2 61 55 Maxiban 69 to 99 Salinomycin 5 66 55 Nicarbazin 100 to 125 Salinomycin 6 66 61 Robenidine 33 Salinomycin 9 55 55 Zoalene 125 Salinomycin 5 61 61 Salinomycin 25 66 66 1 n = number of plants (data for drugs other than salinomycin are not presented). Approved use levels in the US are salinomycin, 44 to 66 ppm; monensin, 99 to 121 ppm; narasin, 59 to 79 ppm; lasalocid, 75 to 125 ppm; Maxiban, 59 to 99 ppm; nicarbazin, 125 ppm; robenidine, 33 ppm; zoalene, 125 ppm (Miller Publishing Company, 1999).

ANTICOCCIDIAL DRUGS IN BROILERS 579 TABLE 7. Use of salinomycin and nicarbazin reported by broiler plants from three regions of the US during March and April 1995 to 1999 1 % plants Region n Salinomycin Nicarbazin P 20 10 35.2 ± 2.9 b 65.8 ± 2.9 a 0.0001 30 10 39.1 ± 2.7 b 60.9 ± 2.7 a 0.0001 50 10 62.2 ± 4.9 a 37.8 ± 4.9 b 0.0023 a,b Values followed by a different superscript differ significantly (P < 0.05). 1 Each observation is the mean percentage of plants using salinomycin in the starter and grower feed or nicarbazin in the starter feed. Results for March and April 1995 to 1999 were combined (n = 10), and the means ± SEM were calculated. Region 20 = Delaware, Maryland, North Carolina, Virginia, Pennsylvania, and West Virginia; Region 30 = Georgia, South Carolina, Tennessee, and Florida; Region 50 = Arkansas, Texas, and Missouri. previous 4 yr or more than 40% of the time. Calorie conversion and mortality were significantly higher in flocks in which CHE had been employed more frequently. Analysis of weekly mortality during April (1998) indicated that mortality was significantly higher during the first week posthatch in plants that used NIC in the starter feed than in those plants that used SAL. No obvious explanation is apparent for this higher mortality. Nicarbazin can cause elevated mortality if birds are subject to heat stress (Chapman, 1994). Higher mortality was reported in floor pen experiments in which NIC-ION shuttle programs were compared with continuous ION programs, although the differences observed were not significant (Watkins and Bafundo, 1993). Mortality resulting from coccidiosis does not usually occur before 3 to 4 wk of age, and, therefore, failure of NIC to control coccidiosis is unlikely to be the cause of the elevated mortality. The data presented show that alternation of drug programs is widely practiced by the poultry industry in the US, although there is no evidence that this is done on a systematic basis. Alternation usually involves changing from a single drug in the starter and grower feeds (principally SAL) to shuttle programs comprising a CHE (usually NIC) followed by an ION or two different ION. According to Watkins and Bafundo (1993), shuttle programs employ a CHE in the starter to provide maximum control of coccidial challenge and then an ION during the grower period to provide maximum growth, feed efficiency, and immunity. Shuttle programs have been used by the broiler industry for many years, but little evidence is available that their use can improve overall performance compared with single-drug programs. Few comparative studies of the effects of shuttle programs have been reported in the scientific literature. In one investigation, body weights, feed efficiencies, and mortalities of birds reared in floor pens were similar by 47 d, whether birds were given a continuous ION program (monensin, narasin, or SAL) or NIC followed by these ION (Watkins and Bafundo, 1993). Repeated exposure to the same anticoccidial drug can result in the selection of drug-resistant strains of Eimeria (Chapman, 1997), which might be expected to result in reduced efficacy of anticoccidial drugs. Alternation of drugs with differing modes of action (for example between ION that share a similar mode of action and CHE agents) has, therefore, been proposed as a means of slowing the rate of development of resistance (Chapman, 1997). Ideally, any resistant parasites that arise during use of the first drug will be eliminated by the second and so on. If the acquisition of resistance is prevented or delayed, improved drug efficacy and, perhaps, better flock performance might be anticipated. In the present study, however, no differences in calorie conversion, number of days to 2.27 kg, or percentage mortality were observed when a single ION was changed to a CHE-ION program, or vice versa. According to McDougald and Reid (1991), the shuttle program is intended to improve control of coccidiosis, and because intensive use of ION has produced strains of coccidia with reduced sensitivity, it is common practice to use a drug such as NIC to bolster the anticoccidial control and take some pressure off the ionophore. In the present study, however, there was no evidence that changing the type of drug program resulted in improved performance. Ideally, CHE agents should be employed in starter and grower feeds if resistance development is to be delayed, because any resistant forms from a previous flock may survive in the litter for the short period that drugs are usually included in the starter feed. Drug programs of this type were not employed during the 5 yr covered by this study. ACKNOWLEDGMENTS I would like to thank Agri Stats. Inc. and Michael Donohue for kindly giving permission to use the Agri Stats database. REFERENCES Chapman, H. D., 1994. A review of the biological activity of the anticoccidial drug nicarbazin and its application for the control of coccidiosis in poultry. Poultry Sci. Rev. 5:231 243. Chapman, H. D., 1997. Biochemical, genetic and applied aspects of drug resistance in Eimeria parasites of the fowl. Avian Pathol. 26:221 244.

580 CHAPMAN Chapman, H. D., 1999. Anticoccidial drugs, and their effects upon the development of immunity to Eimeria infections in poultry. Avian Pathol. 28:521 535. McDougald, L. R., and W. M. Reid, 1991. Coccidiosis. Pages 780 797 in: Diseases of Poultry. B. W. Calnek, ed. Iowa State University Press, Ames, IA. Miller Publishing Company, 1999. Feed Additive Compendium. The Miller Publishing Company. Minnetonka, MN. SAS Institute, 1998. SAS/STAT User s Guide. SAS Institute Inc., Cary, NC. Watkins, K. L., 1997. Are sensitivity test results good predictors of anticoccidial field efficacy? Page 53 in: Control of Coccidiosis into the Next Millenium. M. W. Shirley, F. M. Tomley, and B. M. Freeman, ed. 7th International Coccidiosis Conference, Oxford, UK. Institute for Animal Health, Compton, Newbury, UK. Watkins, K. L. and K. W. Bafundo, 1993. Effect of anticoccidial programs on broiler performance. J. Appl. Poult. Sci. 2:55 60.