Lower Prevalence of Antibiotic-Resistant Enterococci on U.S. Conventional Poultry Farms that Transitioned to Organic Practices

Background: In U.S. conventional poultry production, antimicrobials are used for therapeutic, prophylactic, and nontherapeutic purposes. Researchers have shown that this can select for antibiotic-resistant commensal and pathogenic bacteria on poultry farms and in poultry-derived products. However, no U.S. studies have investigated on-farm changes in resistance as conventional poultry farms transition to organic practices and cease using antibiotics. Objective: We investigated the prevalence of antibiotic-resistant Enterococcus on U.S. conventional poultry farms that transitioned to organic practices. Methods: Poultry litter, feed, and water samples were collected from 10 conventional and 10 newly organic poultry houses in 2008 and tested for Enterococcus. Enterococcus (n = 259) was identified using the Vitek® 2 Compact System and tested for susceptibility to 17 antimicrobials using the Sensititre™ microbroth dilution system. Data were analyzed using SAS software (version 9.2), and statistical associations were derived based on generalized linear mixed models. Results: Litter, feed, and water samples were Enterococcus positive. The percentages of resistant Enterococcus faecalis and resistant Enterococcus faecium were significantly lower (p < 0.05) among isolates from newly organic versus conventional poultry houses for two (erythromycin and tylosin) and five (ciprofloxacin, gentamicin, nitrofurantoin, penicillin, and tetracycline) antimicrobials, respectively. Forty-two percent of E. faecalis isolates from conventional poultry houses were multidrug resistant (MDR; resistant to three or more antimicrobial classes), compared with 10% of isolates from newly organic poultry houses (p = 0.02); 84% of E. faecium isolates from conventional poultry houses were MDR, compared with 17% of isolates from newly organic poultry houses (p < 0.001). Conclusions: Our findings suggest that the voluntary removal of antibiotics from large-scale U.S. poultry farms that transition to organic practices is associated with a lower prevalence of antibiotic-resistant and MDR Enterococcus.

volume 119 | number 11 | November 2011 • Environmental Health Perspectives Research Antibiotic use in U.S. conventional poultry production poses potential public health concerns with regard to the selection of antibioticresistant food borne bacteria (Institute of Medicine 1998; Levy and Marshall 2004;Silbergeld et al. 2008). In U.S. conventional poultry production, anti biotics are administered for therapeutic, prophylactic, and nontherapeutic purposes (Sapkota et al. 2007;Wegener 2003). Some researchers have estimated that use of anti microbials in conventional U.S. poultry production (on a per bird basis) increased by 307% from 1985 to the late 1990s, with the use of non therapeutic anti microbial growth promoters (AGPs) accounting for a significant portion of this use (Mellon et al. 2001).
The use of AGPs in conventional poultry production selects for resistant bacterial popula tions in the production environment and retail poultry products (Aarestrup et al. 2000a;Hayes et al. 2001Hayes et al. , 2005Price et al. 2005;Witte 2000). Consequently, the amplification of resistant bacteria in poultry can result in possible increases in the risk of antibioticresistant bacterial infections in human populations (Aarestrup et al. 2000a;Hammerum et al. 2010;Levy 1998). Animal-derived antibiotic-resistant bacteria have been shown to spread from animals to humans through direct contact with animals and through the consumption of meat products (Donabedian et al. 2003;van Loo et al. 2007).
These findings are increasingly reported in mainstream news and have become one of the main drivers influencing consumer demand for organic poultry (Oberholtzer et al. 2006), which is perceived to be safer than conventional poultry (Crandall et al. 2009). This consumer demand has spurred increased production of organic poultry, making poultry one of the fastest growing segments of the U.S. organic products sector Oberholtzer et al. 2006). Retail sales of organic poultry quadrupled between 2003 and 2006 and reached nearly $200 million in 2008 (Oberholtzer et al. 2006).
To accommodate increased consumer demand and to profit from the organic poultry niche, some conventional poultry growers are adopting organic practices and transitioning their conventional farms to certified organic poultry farms (Oberholtzer et al. 2006). These transitions-which include cessation in the use of all anti biotics and agri chemicals )-could result in changes in the prevalence of antibiotic-resistant bacteria on newly organic poultry farms and subsequent organic poultry products. European studies suggest that removing the non therapeutic use of anti biotics from poultry farms can result in statistically signifi cant reductions in anti bioticresistant bacteria in animals and food products (Aarestrup et al. 2000b(Aarestrup et al. , 2001Emborg et al. 2003;Hammerum et al. 2007;Heuer et al. 2002;Klare et al. 1999). Reductions in human carriage of resistant bacteria also have been docu mented in association with anti biotic withdrawals in European poultry production (Klare et al. 1999;van den Bogaard et al. 2000).
However, to date, the studies regarding this issue that have been conducted in the United States have been largely cross-sectional in nature (Han et al. 2009;Price et al. 2007). To the best of our knowledge, no prospective studies have been conducted in the United States to quantify on-farm, temporal changes in anti biotic resistance of food borne bacteria when anti biotics are removed from U.S. poultry production environ ments. Voluntary transitions to organic practices among large-scale U.S. poultry producers provide an excellent opportunity to research this issue within the United States. Thus, the objective of this study was to prospectively evaluate the prevalence of anti biotic-resistant enterococci on large-scale conventional poultry farms that transitioned to organic practices. Here we describe the findings from the first year of this study.

Materials and Methods
Study sites. All of the poultry farms participating in this study were located in the Mid-Atlantic United States. Two types of poultry farms were included: large-scale conventional broiler farms that were maintaining conventional practices and using anti biotics (n = 5), and large-scale (previously conventional) broiler farms that had just received organic certification and were producing their first flock of certified organic broilers (n = 5). All participating farms were operating under the guidance of one feed mill that produced both conventional and certified organic poultry feed. Two individual poultry houses from each farm were included in the study, for a total of 20 poultry houses. Characteristics of the conventional and newly organic poultry houses are summarized in Table 1.
All of the newly organic poultry houses were certified organic by a state agency accredited by the U.S. National Organic Program (NOP), which promulgates federal organic standards. An overview of common interpretations of the NOP standards that must be met before a poultry farm can be certified organic is provided in Appendix 1.
The specific anti microbials that were used in feed in the conventional poultry houses were as follows: bacitracin (50 g/ton), virginia mycin (15 g/ton), roxarsone (45.35 g/ton), salinomycin (60 g/ton), nicarbazin (0.0125%), and decoquinate (27.2 g/ton). In addition, gentamicin (GEN) was used at the hatcheries that supplied chicks to conventional poultry houses, and bacitracin, virginia mycin, roxarsone, and salinomycin were used at the breeder facilities that supplied the initial eggs to the hatcheries.
Sample collection. From March to June 2008, poultry litter, water, and feed samples were aseptically collected [in sterile Whirl-Pak ® collection bags (Nasco, Fort Atkinson, WI)] from the conventional and newly organic poultry houses. Litter samples (500 g) from the top 11-5 cm of litter were collected from three randomly selected areas of each poultry house. Two water samples (500 mL) were retrieved: a) one from raw source water before filtration or ultraviolet treatment, and b) one from finished water present in the waterlines. One poultry feed sample (300 g) was collected from the central feed hopper within each poultry house. All poultry litter, water, and feed samples were shipped overnight at 4°C and processed within 24 hr.
Isolation. Poultry litter and feed samples were enriched in a 1:10 weight-to-volume dilution of 100 mL Enterococcosel Broth (Becton Dickinson & Co., Franklin Lakes, NJ) for 24 hr at 41°C. We included positive and negative control broths for quality control and quality assurance. After 24 hr, 10 μL of the enrichment culture was streaked onto Enterococcosel Agar (EA; Becton Dickinson & Co.) and incubated overnight at 41°C. Presumptive colonies of Enterococcus spp. ranged in appearance from brown to black with a brown-black precipitate on EA. Three presumptive Enterococcus colonies from each litter and feed sample were streaked onto separate brain heart infusion (BHI) agar plates for purification and incubated at 41°C for 24 hr. A colony was collected from each BHI purification plate and archived at -80°C in Brucella broth with 20% glycerol.
Isolation of Enterococcus spp. from water samples was performed in accordance with U.S. Environmental Protection Agency (EPA) Method 1106.1 (U.S. EPA 2006). Briefly, 10-fold dilutions of each water sample were prepared in phosphate-buffered saline (U.S. EPA 2006), and 10 mL of each dilution was filtered through a 0.45-μm, 47-mm mixed cellulose ester filter (Millipore, Billerica, MA). Each filter was placed on a 60-mm plate containing EA, inverted, and incubated at 41°C for 24 hr. Resulting colonies typical of Enterococcus spp. were considered presumptive Enterococcus spp. Of the recovered presumptive Enterococcus spp., three isolates per water sample were purified on BHI and archived in Brucella broth with 20% glycerol at -80°C. Positive (Enterococcus faecalis ATCC 29212; ATCC, Manassas, VA) and negative (Escherichia coli ATCC 25922) controls were used throughout the isolation process.
Identification. All presumptive Enterococcus spp. were streaked from archived stocks onto tryptic soy agar amended with 5% sheep's blood and incubated at 41°C for 24 hr. Presumptive identification of Enterococcus spp. was done by Gram staining and testing for catalase production and pyrrolidonyl arylamidase (PYR) activity. All gram-positive, catalase-negative, and PYR-positive isolates were confirmed and identified to the species level using the automated biochemical identification Vitek ® 2 Compact System (BioMérieux Inc., Hazelwood, MO) in accordance with the manufacturer's specifications.
Antimicrobial susceptibility testing. We performed anti microbial susceptibility testing (AST) on all confirmed Enterococcus isolates (n = 259) by micro broth dilution using the Sensititre™ system (Trek Diagnostic Systems, Westlake, OH) according to the manufacturer's directions. Briefly, colonies from pure 18-to 24-hr cultures were transferred to tubes of sterile Sensititre demineralized water (Trek Diagnostic Systems) to achieve a turbidity equiva lent to a 0.5 McFarland standard. Then, 50 μL of each suspension was added to sterile Sensititre cation-adjusted Mueller Hinton broth (Trek Diagnostic Systems), and 50 μL of the broth solution was then dispensed into micro titer, gram-positive 96-well plates embedded with 17 test anti microbials [National Antimicrobial Resistance Monitoring System (NARMS) Enterococcus Plate Format; Trek Diagnostic Systems]. Plates were then incubated in the Automated Reading and Incubation System (ARIS; TREK Diagnostic Systems) at 37°C for 18 ± 1 hr. The first 100 plates were read both manually and via the ARIS system for quality assurance comparisons of minimal inhibitory concentration (MIC) determinations. After consistency between the two methods was determined, subsequent samples were read by the ARIS exclusively.
We used Clinical and Laboratory Standards Institute (CLSI) interpretive criteria for microbroth dilution methods (CLSI 2008) to evaluate resulting MICs where breakpoints were available, except for quinupristin/dalfopristin (SYN), for which we used the breakpoint from the European Committee on Antimicrobial Susceptibility Testing (EUCAST 2011). Otherwise, we used the provisional breakpoints used by NARMS (Food and Drug Administration 2009). The following specific anti microbials (resistance breakpoints) were used: chloramphenicol (CHL; ≥ 32), ciprofloxacin (CIP; ≥ 4), dapto mycin (DAP; no The time since complete poultry litter change is greater than the time the farm practiced organic methods because the farms were not considered "organic" until the first birds to be produced under organic proctices arrived, and these first flocks did not always arrive immediately after the poultry litter cleanout. b The organic birds did not go outside although they had the opportunity to do so. Statistical analysis. We used the generalized linear mixed model (GLMM) method to evaluate associations between the prevalence of anti biotic-resistant Enterococcus spp. and poultry production type (conventional or newly organic). The GLMM method was used to account for the clustered nature of the study design, which made it necessary to adjust for intra-poultry house and intra-poultry farm variability. Firth's bias correction method was used when zero counts occurred for one group (Heinze and Puhr 2010). All statistical analyses were performed using SAS software (version 9.2; SAS Institute Inc., Cary, NC).

Results
Prevalence of Enterococcus spp. Enterococcus spp. were isolated from 100% of all conventional and newly organic poultry houses. Poultry litter was the principal environmental media for the recovery of Enterococcus spp. from both farm types, with 100% of all litter samples testing positive; however, these microorganisms were also recovered from water and feed samples (Table 2).
Overall, 46% of Enterococcus spp. were identified as E. faecalis and 43% as Enterococcus faecium. Enterococcus durans, Enterococcus gallinarum, and Enterococcus hirae were also isolated from both types of poultry houses in several types of environmental media (Table 2). We found no significant differences in species prevalence between farm types.
MICs. MIC ranges, MIC 50 s (MIC for 50% of the bacteria are less than or equal to this MIC) and MIC 90 s (MIC for 90% of the bacteria are less than or equal to this MIC) for E. faecalis and E. faecium recovered from the poultry houses are shown in Table 3. For E. faecalis, 53% of MIC ranges (8 of 15 antibiotics, excluding LIN and SYN because of inherent resistance) differed depending on farm type; for E. faecium, 56% of MIC ranges (9 of 16 anti biotics, excluding FLA because of inherent resistance) differed depending on farm type (Table 3). Some MIC ranges differed depending on species (Table 3). Similarly, we also observed some differences in MIC 50 s and MIC 90 s between isolates recovered from different farm types, and between species (Table 3).
Acquired anti biotic resistance. Among E. faecalis isolates, acquired anti biotic resistance against nine anti microbials (CHL, ERY, FLA, GEN, KAN, NIT, PEN, STR, and TYL) was lower among E. faecalis from newly organic poultry houses compared with conventional poultry houses ( Figure 1A). The differences in percent resistance were statistically significant for ERY (p = 0.004) and TYL (p = 0.004) ( Figure 1A).
Among E. faecium, we observed acquired resistance to GEN and TYL only among isolates recovered from conventional poultry houses ( Figure 1B). We observed no resistance to CHL, LZE, or VAN among any of the E. faecium recovered from conventional or organic poultry houses ( Figure 1B).

Sources of antibiotic-resistant bacteria.
Most antibiotic-resistant E. faecalis were isolated from poultry litter samples (Table 4). E. faecalis isolated from conventional feed samples also expressed acquired resistance to eight anti microbials (CHL, ERY, FLA, KAN, NIT, STR, TET, and TYL), indicating that conventional poultry feed could be a potential source of exposure to anti biotic-resistant E. faecalis among broilers (Table 4). No resistant E. faecalis were isolated from organic poultry feed or source water or waterline samples retrieved from either conventional or newly organic poultry houses.
Most anti biotic-resistant E. faecium also were isolated from poultry litter samples (Table 4). Antibiotic-resistant E. faecium were also recovered from feed and water line samples from conventional poultry houses and from feed, source water, and water line samples from newly organic poultry houses (Table 4). Conventional feed was contaminated with E. faecium that expressed acquired resistance to 10 anti microbials (CIP, ERY, KAN, LIN, NIT, PEN, STR, SYN, TET, and TYL), whereas organic feed was contaminated with E. faecium that expressed acquired resistance to 6 anti microbials (ERY, KAN, LIN, NIT, SYN, and TET) ( Table 4). No conventional source water samples were contaminated with resistant E. faecium, whereas one organic source water sample was contaminated with one LINresistant E. faecium isolate. Conventional waterline samples were contaminated with E. faecium that expressed acquired resistance to 11 antimicrobials (CIP, ERY, GEN, KAN, LIN, NIT, PEN, STR, SYN, TET, and TYL), whereas organic waterline samples were not contaminated with anti biotic-resistant E. faecium (Table 4). The differences in waterline contamination between poultry house types could be attributed to the fact that conventional    (Table 1), allowing more time for contamination to occur. Acquired MDR. The percentage of MDR E. faecalis was statistically significantly lower among isolates from newly organic poultry houses compared with isolates from conventional poultry houses (10% vs. 42%; p = 0.02; Figure 2). The percentage of MDR E. faecium also was statistically significantly lower among isolates from newly organic poultry houses compared with isolates from conventional poultry houses (17% vs. 84%; p < 0.001; Figure 2). Predominant MDR patterns are shown in Table 5.
The mode number of anti biotics that E. faecalis expressed acquired resistance against was one and three, among isolates from newly organic houses and conventional houses, respectively ( Figure 3A). The mode number of anti biotics that E. faecium expressed acquired resistance against was one and four, among isolates from newly organic houses and conventional houses, respectively ( Figure 3B). These findings show that newly organic poultry houses are characterized by individual E. faecalis and E. faecium isolates that express resistance to fewer numbers of anti biotics compared with their conventional counterparts.

Discussion
In this study, we observed a significantly lower prevalence of anti biotic-resistant and MDR E. faecalis and E. faecium on large-scale poultry farms that had just transitioned to organic practices compared with large-scale poultry farms that were maintaining conventional practices. To our knowledge, these are the first U.S. data to show immediate, on-farm changes in anti biotic resistance when antimicrobials are voluntarily withdrawn from large-scale U.S. poultry production.
These findings are in agreement with earlier European and Asian studies that have documented reductions in anti biotic-resistant Enterococcus spp. after governmental bans and/or voluntary withdrawals of the use of anti biotics in animal production (Aarestrup et al. 2001;Lauderdale et al. 2007). Using data from the Danish program for surveillance of anti microbial resistance in bacteria recovered from animals, foods, and humans (Danish Integrated Antimicrobial Resistance Monitoring and Research Programme; DANMAP), Aarestrup et al. (2001) reported signifi cant decreases in the percentages of E. faecalis and E. faecium resistant to avilamycin, ERY, avoparcin, and virginia mycin, four anti biotics banned by the Danish government for use as AGPs in the late 1990s. For example, from 1997 to 2000 the percentage of ERY-resistant E. faecium isolated from broilers decreased from 76.3% to 12.7%, and the percentage of virginia mycin-resistant E. faecium isolated from broilers decreased from 66.2% to 33.9% (Aarestrup et al. 2001). In the present study, we observed that the prevalence of E. faecium resistant to ERY was 13% and 10% among isolates from conventional and newly organic farms, respectively, whereas the prevalence of E. faecium resistant to SYN (a virginia mycin analogue) was 20% and 7% among isolates from conventional and newly organic farms, respectively ( Figure 1B).
Reductions in percent resistance to ERY and other anti biotics observed among Enterococcus spp. from newly organic poultry farms in the present study may not be as dramatic as those observed by Aarestrup et al. (2001) and other European researchers because poultry houses in the present study were sampled during the production of the first flock of certified organic broilers. Although these poultry houses under went extensive and comprehensive cleaning events before they could be certified as organic, reservoirs of resistant bacteria may have remained in the packed dirt floor and on fomites within the poultry houses, contributing to persistent low levels of anti biotic-resistant entero cocci in newly organic poultry houses. Similarly, Heuer et al. (2002) demonstrated that VAN-resistant enterococci can persist in broiler flocks for > 5 years after anti biotic-selective pressures are removed from the production environment.
Two additional factors likely play significant roles in the persistence of low rates of anti biotic-resistant enterococci observed in newly organic poultry houses in this study. First, U.S. organic certification standards, a E. faecalis is intrinsically resistant to LIN and streptogramin A (dalfopristin) (Dina et al. 2003), and E. faecium is intrinsically resistant to FLA; therefore, these species/drug combinations were excluded from the MDR analysis.

Percentage of E. faecalis
promulgated through the NOP, apply starting on day 1 of a chick's life . No organic certification standards need to be met before the first day of life. Thus, some breeder facilities that supply eggs to hatcheries, and hatcheries that ultimately produce "organic" chicks, do not have to meet any organic standards and can therefore use anti biotics among breeder stocks and inject anti biotics into eggs. These practices can result in exposures to anti biotics among "organic" broilers before the first day of life. Second, organic broilers can be exposed to anti biotic-resistant bacteria through feed and water. Organic poultry feed is required by the NOP to be free of anti biotics, slaughter by-products, and genetically modified organisms ). However, our data show that contamination of organic feed with anti biotic-resistant bacteria can occur (Table 4). The question remains as to whether feed is contaminated at the feed mill, during transport, and/or during storage at poultry houses via bio aerosols, insects, rodents, or other factors. Beyond feed, we observed that one source water sample from newly organic poultry houses was contaminated with one LIN-resistant E. faecium (Table 4) and one water line sample from organic poultry houses was contaminated with one MDR E. gallinarum (data not shown).
We find it encouraging that the percentages of MDR E. faecalis and MDR E. faecium in the present study were significantly lower on newly organic poultry farms compared with farms that were maintaining conventional practices. E. faecalis recovered from newly organic and conventional farms expressed acquired resistance against a mode number of one and three anti biotics, respectively, and E. faecium from newly organic and conventional farms were resistant against a mode number of one and four anti biotics, respectively. These data are in agreement with a recent study by Miranda et al. (2007) that showed that rates of MDR Enterococcus spp. were significantly lower among isolates recovered from organic chicken and turkey products compared with conventional products.
As with all field-based studies, the present study had several limitations. As discussed above, we could not control for the fact that organic broilers may have been exposed to anti biotics before the first day of life. This could have influenced the rates of anti biotic resistance observed among Enterococcus spp. recovered from newly organic poultry houses; however, because we could not include a control farm that produced chicks that were known to have never been exposed to antibiotics, we could not estimate the contributions of these potential exposures to observed resistance rates. The study is also limited in terms of geographical location. All poultry farms included in this study are located in the Mid-Atlantic United States and under the advise ment of one feed mill. Thus, it is unclear whether our results are generalizable across the United States and across the various largescale contract growers that dominate the U.S. poultry industry. Larger-scale studies based in varying geographical areas at farms managed by different companies are necessary. Finally, this study is limited by the fact that separate conventional poultry farms served as control farms for the newly organic poultry farms. Although it would have been preferable to also sample the newly organic poultry farms before their conversion from conventional to organic practices, this was not possible.

Conclusions
This study provides the first on-farm U.S. data describing the impacts of eliminating antibiotics from large-scale U.S. poultry production on rates of anti biotic-resistant enterococci. The findings support the hypothe sis that removing anti biotic use from large-scale U.S. poultry farms transitioning to organic practices can result in immediate and statistically significant reductions in on-farm anti biotic resistance.

correction
In Table 1 of the manuscript originally published online, the value for months since complete poultry litter change was incorrect for organic broilers. It has been corrected here. Producer must create and implement an • organic system plan and manure management plan. Broilers must be produced under con-• tinuous organic management starting "no later than the second day of life." All feed components must be organically • produced and contain no anti biotics, other animal drugs, slaughter by-products, or genetically modified organisms. No antibiotics may be used for animal • treatment. Producer must establish preventative • broiler health care practices, and diseases can be prevented with vaccines, bio security meas ures, pre biotics, and pro biotics. Maximum stocking densities of broilers is • not specified by the NOP, but certifying agencies often require at least 0.14 m 2 per bird. An outdoor access area must be provided • to ensure access to fresh air, exercise, and sunlight.

Appendix 1. Overview of Common
Clean and dry bedding must be provided • in an indoor area. Sanitizers and cleaners used on the prop-• erty must be on approved products lists. Agrichemicals cannot be used on the • property.