ABSTRACT

Veterinary pharmaceuticals are commonly used in poultry farming to prevent and treat microbial infections as well as to increase feed efficiency, but their use has created public and environmental health concerns. Poultry litter contains antimicrobial residues and resistant bacteria; when applied as fertilizer, the level and effects of these pharmaceuticals and antimicrobial-resistant bacteria in the environment are of concern. The purpose of this study was to investigate poultry litter for veterinary pharmaceuticals and resistance patterns of Escherichia coli. Litter samples were collected from controlled feeding trials and from commercial farms. Feed additives bacitracin, chlortetracycline,oncommercialfarms and added to the feed in the controlled trials, were extracted in methanol and analyzed by liquid chromatography-mass spectrometry techniques. Sixty-nine E. coli were isolated and identified by API 20E. The susceptibility of the isolates to antibiotics was determined using Avian plates and the Sensititer automated system. This study confirmed the presence of antimicrobial residues in broiler litter from controlled environments as well as commercial farms, ranging from 0.07 to 66 mg/L depending on the compound. Concentrations of individual residues were higher in litter from controlled feeding trials than those from commercial farms. All E. coli isolates from commercial farms were multiresistant to at least 7 antibiotics. Resistance to β-lactam antibiotics (amoxicillin, ceftiofur), tetracyclines, and sulfonamides was the most prevalent. This study concluded that broiler litter is a source of antimicrobial residues and represents a reservoir of multiple antibiotic-resistant E. coli.

Key words: antimicrobial residue , veterinary pharmaceutical , antibiotic resistance , poultry litter , Escherichia coli

INTRODUCTION

The poultry industry is an important part of the economy in British Columbia, Canada. In 2006, 85% of the 199 million kilograms of chicken produced in British Columbia were raised in the Fraser Valley (British Columbia Ministry of Agriculture and Lands, 2007). High-density farming techniques have increased this output in recent years, and with increased output came increased waste accumulation. The amount of chicken manure produced by farms in British Columbia is expected to grow to 320,000 tonnes by 2010 as production continues to increase (Timmega and Associates Inc., 2003). As a waste-saving measure, farms are encouraged to reuse poultry manure as fertilizer. Poultry litter is primarily composted for use in specific markets, such as organic farming or mushroom farming, and small amounts of the manure are spread directly onto crops from the barns, either in the Fraser Valley or elsewhere

in the Lower Mainland (Timmega and Associates Inc., 2003). The use of antimicrobial agents as growth promoters

in poultry feed contributed to the increase in poultry production. Most of these antimicrobials are not fully absorbed in the chicken gut and up to 90% of the administered dose can be excreted in the feces (Kumar et al., 2005b). Because litter from farms is commonly used as fertilizer, the presence of antibiotics and their potential effect on humans and the environment is an area of growing concern (Boxall et al., 2003; Rooklidge, 2004). The presence of veterinary pharmaceuticals in the environment has been documented in the literature (Khan et al., 2007; Song et al., 2007). There is also evidence that plants bioaccumulate some antimicrobial agents and can be sources of exposure to such compounds (Kumar et al., 2005a). Escherichia coli is a normal inhabitant of the gastrointestinal tract of humans and animals; however, some strains are known to be pathogenic. These strains induce colibacillosis in chicken, which is an important cause of economic losses for the poultry industry (Amara et al., 1995). In humans, pathogenic E. coli can cause several diseases including urinary tract infections, septicemia, and neonatal meningitis (Ewers et al., 2004, Johnson et al., 2008). The avian intestines have been considered as a reservoir of potential E. coli with zoonotic potential that could be transferred directly from birds to humans (Ewers et al., 2009). The practice of using antimicrobi

als in feed may modify the intestinal flora by creating a selective pressure in favor of resistant bacterial populations (such as resistant E. coli) that could find their way into the environment and food chain (Diarra et al., 2007).

The goal of this study was to investigate the presence of veterinary pharmaceuticals and antimicrobialresistant bacteria in broiler litter collected from controlled feeding trials and commercial farms. Antibiotics commonly used on poultry farms in British Columbia (bacitracin, chlortetracycline, monensin, narasin, nicarbazin, penicillin, salinomycin, and virginiamycin) were quantified and resistance levels of E. coli isolates were investigated.

MATERIALS AND METHODS

Experimental Design

Controlled Feeding Trials. Studies were performed with 1,800 one-day-old male broiler chicks in 2 separate studies. In each study, 900 chicks were placed in 18 pens (50 birds per pen), which were randomly assigned to 6 treatments: a control group fed without antibiotics and 5 groups fed diets containing antibiotics as described   in Table 1. The diets and rearing conditions have been described previously (Diarra et al., 2007). At d 36, two litter samples were collected aseptically from each pen in sterile Whirl-Pak plastic bags (Whirl-Pak, Modesto, CA). In the case of bacteriology, a portion of  each  sample was transferred to preweighed tubes containing Car-Blair transport medium (Quelab, Montréal, Qué- bec, Canada) and kept on ice until laboratory analysis was performed the same day. The rest of the sample to be used for chemistry analysis was stored at −20°C.

Commercial Farms. Nine commercial broiler chick- en farms located in the Fraser Valley of British Co- lumbia, Canada, were selected for the study based on feed additives used as described previously (Diarras- souba et al., 2007). Two litter samples were collected aseptically from each farm on d 8 to 10 and d 25 to 28 during the rearing period in sterile Whirl-Pak plastic bags. For bacteriology testing, a portion of each sample was transferred to preweighed tubes containing Car- Blair transport medium (Quelab) and kept on ice until laboratory analysis was performed the same day. The rest of the sample for chemistry analysis was stored at−20°C.

Chemistry Analyses

Individual litter samples were first freeze-dried and divided into 2 containers. Each container was treated separately to calculate values of percentage of mois- ture. Samples were then combined  and  homogenized to fine powder and an aliquot was weighed out for the analysis.

The residues of virginiamycin, monensin, salinomy- cin, narasin, and nicarbazin were analyzed as described previously (Furtula et al., 2005).  Briefly,  5  g  of  sam- ple was placed into a  50-mL  centrifuge  tube,  25  mL  of methanol was added, and the tube was sonicated, shaken, and centrifuged. Supernatant was  decanted  into a 100-mL round-bottom flask, and the extraction procedure was repeated 2 more times. Methanol ex- tracts were combined, reduced by rotary evaporation,and filtered through a 0.2-µm filter for liquid chroma- tography-mass spectrometry (liquid chromatography- mass spectrometry and liquid chromatography-mass spectrometry-mass spectrometry) analysis. The instru- ments used were an Agilent  LC  MSD  SL  system  and  an Agilent 1100 LC system (Agilent Technologies Inc., Santa Clara, CA) coupled with Micromass Quattro Micro/MS/MS (Waters Corporation, Milford, MA). Compounds were separated on a ZORBAX 300SB C-18 column (Agilent Technologies Inc.) using a gradient of acetonitrile (0.1% formic acid) and water as a mobile phase. The residues of bacitracin, tetracyclines, and penicillin were analyzed according to Canadian Food Inspection Agency methods (CFIA, 2004a, 2004b).

Bacteriological Analyses and E. coli Isolation

The generic E. coli population was estimated using

E. coli and Petrifilm coliforms (3M, St. Paul, MN) as described previously (Diarrassouba et al., 2007). Brief- ly, the tubes of sample (5 to 6 g) and transport medium were vigorously vortexed for 1 min, and 10-fold serial dilutions were prepared in sterile saline. Counts of non- specific E. coli were obtained from individual samples and from each farm by plating 1 mL of the appropriate dilution on E. coli and Petrifilm coliforms. After incu- bation at 37°C for 24 h, blue, gas-producing colonies were counted as generic E. coli. Results were expressed as colony-forming units per gram of material. For each sampling, 2 to 3 presumptive colonies from each sample were purified on blood agar and confirmed as E. coli using API 20E strips (bioMerieux, St-Laurent, Québec, Canada) according to the specifications of the manu- facturer.

Antibiotic Resistance

The antibiotic minimal inhibitory  concentrations  were determined for a total of 72 E. coli  isolates (18,  15, and 36 from commercial farms, controlled feeding trial I, and controlled feeding trial II, respectively), us- ing the Sensititer automated system (Trek Diagnostic Systems, Cleveland, OH). The minimal inhibitory con- centrations were determined and classified  according to the guidelines of the Clinical Laboratory Standard Institute as described previously (Diarrassouba et al., 2007). Isolates showing intermediate or full resistance were classified as resistant in this study because these drugs would no longer be useful in treating a human infection once the bacteria has intermediate resistance. The following 18 antimicrobials on Avian plates (Trek Diagnostic Systems) were tested: amoxicillin, peni- cillin, ceftiofur, erythromycin, tylosin, clindamycin, spectinomycin, streptomycin, gentamicin, neomycin, oxytetracycline, tetracycline, enrofloxacin, sarafloxa-  cin, novobiocin, sulfadimethoxine, sulfathiazole, and trimethoprim-sulfamethoxazole. The Avian plate was used in this study because antibiotics included on this plate are of direct interest to human health. The con- trol strain was E. coli ATCC 25922.

Statistical Analyses

Bacterial counts were log-transformed and analyzed using the GLM procedure of SAS (SAS, 2000). The as- sociation test of Cochran-Mantel-Haenszel was used to determine the relationship between treatment and the antibiotic resistant in E. coli and using the FREQ pro- cedures of the SAS Institute (SAS, 2000). The statisti-  cal significance level was set at a P-value of P < 0.05.

RESULTS

Antimicrobial Residues

Percentage of moisture for controlled feeding trial I and controlled trial II ranged from 29 to 42% and from 14 to 21%, respectively, with the SD not greater than 25%, indicating consistency and reproducibility from cage to cage. Percentage of moisture for samples from commercial farms had much higher variation, ranging from 21 to 70% (data not shown).

Chemistry results for the litter samples from con- trolled feeding trial I showed that penicillin, salino- mycin, bacitracin, and the combination of salinomyin and bacitracin added to the feed were detected in the litter (Table 1). The same observation was noted for trial II, in which virginiamycin, monensin, narasin, and chlortetracycline were detected in the litter (Table 1). Chemistry results for the litter samples from commer- cial farms collected at d 8 to 10 and d 25 to 28 also showed that each of the pharmaceuticals added to the feed was subsequently detected in the litter (Table 2).

Presence of antibiotics in the poultry litter has also been calculated based on detection  of  antimicrobials  in the litter and inclusion rates (Table 3). In the con- trolled feeding studies, presence of antimicrobials  in  the litter ranged from 3 to 60% in regards to amounts added to the feed.

Bacteriological Analysis

High numbers of E. coli were found in all litter sam- ples and some variation among the same treatment group and farms using the same feed was observed. However, no significant difference (P > 0.05) was found between treatment groups or feed used regarding  E.  coli number. The mean log₁₀ colony-forming units of E. coli recovered were 8.6, 6.9, and 6.8 per gram of litter samples from the trial I, trial II, and the commercial farms, respectively (Figure 1).

Antimicrobial Resistance

Antibiotic-resistant E. coli isolates  have  been  found in litter from all farms regardless of feed used and in both controlled trials independent of which pharma- ceutical was used in feed (P > 0.05). Overall, resistance to tetracycline, β-lactams (amoxicillin and ceftiofur), spectinomycin, streptomycin, and  sulfonamides  was the most prevalent among the E.  coli  isolates  (Figure 2). Resistance to amoxicillin, spectinomycin, strepto- mycin, and sulfonamides was higher in isolates from commercial farms than those from the controlled feed- ing trials (Figure 2) and this difference was found to    be statistically significant (P < 0.05). All isolates were resistant to penicillin, erythromycin, tylosin, clindamy- cin, and novobiocin and were susceptible to enrofloxa- cin (data not shown). Additional resistance to at least one antibiotic has been found in 85.5% of all tested isolates (Table 4). The most common resistance spec- trum patterns were amoxicillin-tetracycline (11 iso- lates), amoxicillin-ceftiofur-tetracycline (6 isolates),  and amoxicillin-spectinomycin-streptomycin-tetracy- cline-sulfathiazole-trimethoprim/sulfamethoxazole (7 isolates). The broadest resistance spectrum (amoxicil- lin-ceftiofur-spectinomycin-streptomycin-gentamicin- tetracycline-enrofloxacin-sarafloxacin-sulfathiazole) was found in one isolate from the treatment group fed salinomycin and bacitracin in controlled trial I (Table  4).

DISCUSSION

In this study, the presence of antimicrobial residues and multiresistant E. coli bacteria in chicken litter from controlled and uncontrolled environments was investi- gated. Despite the controversy about the use of antimi- crobial agents in livestock and poultry production, evi- dence suggests that antibiotic usage has an important effect on the emergence of antimicrobial resistance in bacteria (Smith et al., 1999; Shuford and Patel, 2005).   In the current study, the combination of antimicrobial residues analysis and the determination of antimicrobi- al-susceptible E. coli could offer additional information on potential environmental and public health risks as- sociated with use of untreated poultry litter.

Antimicrobial Residues

Veterinary pharmaceuticals were detected in all sam- ples from farms and controlled feeding trials. The concentration of salinomycin in the litter from controlled feeding trial I was the same order of magnitude as for salinomycin added as a single compound or in combina- tion with bacitracin. Consistency of detected concen- trations of antibiotics in replicates in controlled studies is satisfactory taking into consideration the number of variables in the system.

In the controlled study II, residues of all compounds added to the feed were detected in the litter as well.   The concentration of narasin in litter from this trial.

was  the  same  order  of  magnitude  as  in  litter  samples from the farms, but with less variation.

The  chlortetracycline  concentrations  detected  in  the litter samples were surprisingly high compared with lit-

erature  values  from  previous  studies  (Webb  and  Fontenot,  1975).  Chlortetracycline  binds  to  the  soil  andhas  a  relatively  long  half-life  in  soil;  therefore,  its  presence  could  influence  soil  microbes  and  contribute  to development of tetracycline resistance in environmentalbacteria  (Chander  et  al.,  2005).  Furthermore,  the  long persistence  of  chlortetracycline  could  explain  the  high level of resistance of our isolates to this antibiotic.

The  presence  of  virginiamycin  in  the  litter  was  low,which  may  indicate  a  short  half-life  in  the  litter;  this is   contrary   to   reported   half-life   of   87   to   173   d   for this   compound   in   soils   under   laboratory   conditions

(Weerasinghe and Towner, 1997). It could also indicate a  higher  absorption  efficiency  in  the  chicken  gut,  lead-

ing  to  a  smaller  portion  of  the  dosage  being  excreted compared  with  the  other  pharmaceuticals  used  in  this study.  Increase  of  the  virginiamycin  doses  in  feed  from11 to 22 mg/kg resulted in increased amounts detected.Virginiamycin  is  not  used  in  humans  but  it  is  relatedto   dalfopristin,   a   streptogramin   antibiotic   of   humanimportance  (Khan  et  al.,  2007).  Because  of  this  relationship  and  possible  development  of  cross-resistance,virginiamycin has been banned for use in the European

Union but is still in use in many other countries includ-ing Canada and the United States (Soltani et al., 2000;

Smith et al., 2003). Samples from commercial farms showed that the presence of the antibiotics in chicken litter followed the pattern of their addition to the feed. Introduction of bacitracin and salinomycin to the feed at farms d, e,

and j at the grower phase resulted in detection of these compounds in the litter.

The detection of nicarbazin in samples from farms a and i from the second collection of samples confirmed the nicarbazin persistence in the litter.

Variations in amounts of antibiotics found in the litter from the farms with the same feed regimen may be due to the great nonhomogeneity of the samples. Reanalysis of the samples with extreme results confirmed these findings (for example, sample j-2 result for narasin). The amounts of salinomycin, narasin, and nicarbazin detected in this study are similar to those reported in the literature (Furtula et al., 2005). Therefore, it could be expected that the concentrations of antimicrobials found in the litter could persist in the environment. Samples from commercial farms as well as controlled studies were homogenized mixtures of litter and bedding. If fresh litter is directly applied to fields as fertilizer, there would be localized areas with concentrations of pharmaceuticals even higher than those

reported in this study, exposing soil microbes to a range of concentrations. Presence of veterinary pharmaceuticals in the litter increases demand for development of best management practices and stricter regulations for treatment of litter before it is used as fertilizer. Ionophores, narasin, and salinomycin have similar presence in the litter in controlled studies. The high concentration of penicillin and chlortetracycline in the litter is concerning because these are antibiotics commonly used in human treatment. All isolates in this study were 100% resistant to penicillin including isolates from control groups. Comparison of percentage of antibiotics in the litter from commercial farms and litter from controlled studies indicates that the presence of antimicrobials under the controlled condition is gen

-erally higher and more consistent and for salinomycin and bacitracin significantly different. This difference in

antibiotic concentration in litter may be explained by the fact that for the controlled studies, the individual antibiotics were administered continuously through feed (starter, grower, and finisher phases) from d 0 to 35, except for chlortetracycline, which was withdrawn from the finisher diet as recommended by the manufacturer. In commercial diets, given antibiotics are included during specific growth periods that help to decrease their residues in poultry meat for human consumption and that was reflected in the litter as shown by data in this study.

E. coli Number

Although some benefits have been documented, little is known of the effect of antimicrobial feed supplementation on the poultry litter microflora. Poultry litter is a mixture containing fecal material and bedding and is known as a source of infectious agents (McCrea et al., 2008). Escherichia coli is used as an indicator bacterium to detect the possible presence of fecal pathogens. In our studies, wood shavings were used as bedding, which are known to harbor 108 viable E. coli per gram of litter after 5 wk of growth (Macklin et al., 2005). We also found a comparable high number of E. coli in our litter samples. The microbial ecosystem of broiler litter is undoubtedly influenced by management practices including feeding. For example, it has been reported that the addition of purified lignin and mannan oligosaccharides to broiler feed increased the population of beneficial bacteria in the ceca and decreased those of E. coli in the litter (Baurhoo et al., 2007). We found that the numbers of generic E. coli recovered in litter materials were not significantly influenced by antimicrobial feed supplementation.

E. coli Resistance

Previously, we showed that multidrug-resistant E. coli was found in chicken guts and that feed supplementation with some antimicrobial agents like salinomycin or bacitracin could influence the distribution of the resistance phenotype in E. coli (Diarra et al., 2007). In the present study, we examined the effect of such supplementation on the resistance patterns of isolates from litter. As expected, multiple antibiotic-resistant E. coli were found in all studied litter samples, suggesting that broiler litter should be considered as a significant reservoir of such E. coli. There were no significant differences in resistance levels between isolates collected from different treatment groups in controlled feeding trials I and II or between the isolates from commercial farms using each of the different feeds. Additional resistance to amoxicillin, ceftiofur, gentamicin, spectinomycin, streptomycin, and sulfonamides was present in some of the isolates in the control groups (no antibiotic added to the diet), implying that the bacteria are colonizing the chicks from birth. This finding agrees with previous studies that have found resistance to many antibiotics, including trimethoprim-sulfamethoxazole, amoxicillin, streptomycin, gentamycin, and tetracycline, in E. coli isolated from broilers in Canada and the United States

when raised without antimicrobials (Smith et al., 2007; Chander et al., 2008; Thibodeau et al., 2008). The supplier of the chicks for the current studies used eggs from all over British Columbia and some parts of the United States and so it was not possible to pinpoint where the resistant bacteria were coming from. The presence of this resistance in the control groups has been cited in the literature as evidence that the use of antimicrobials in feed has no pronounced effect on resistance levels (Phillips et al., 2004). The higher resistance to amoxicillin, spectinomycin, streptomycin, and sulfonamides found in isolates from commercial farms compared with those from the controlled feeding trials could imply that the use of commercially available feeds containing multiple antimicrobials could be contributing to the increase in resistance. Based on these differences, it is possible that the spread of antibiotic resistance could be reduced by changes

in the current practices (for example, by reducing the number of antibiotics in the feed).

The resistance to ceftiofur found in the isolates was surprising and concerning, considering a similar study in the United States that found no ceftiofur-resistant isolates (Smith et al., 2007). Ceftiofur is a cephalosporin antibiotic resistant to bacterial β-lactamase, which makes it the drug of choice for treating many infections, including salmonellosis in children (Frye and Fedorka-Cray, 2007). The high resistance to tetracycline may be explained by the common use of this antibiotic for a long time in poultry. The level of resistance to sulfonamides was higher than expected because sulfonamides are no longer widely used in the Fraser Valley poultry industry, although high levels of resistance have been previously reported in the literature (Guerra et al., 2003; Smith et al., 2007). This could be partially explained by the fact that sulfonamides degrade very slowly in the environment (Dolliver et al., 2008). The particularly striking increase in resistance to trimethoprim-sulfamethoxazole on commercial farms is of concern because this is the drug of choice for treating urinary tract infections caused by E. coli in women(Talan et al., 2000; Russo and Johnson, 2003). There is some suggestion that sulfonamide resistance genes have been maintained due to their association with other resistance genes (Diarrassouba et al., 2007). This theory is supported by the fact that each of our isolates that was resistant to sulfonamides was also resistant to amoxicillin and tetracycline. Based on data presented in this study, poultry litter poses an environmental risk and should be treated in some way before being applied as fertilizer.

This study confirmed the presence of veterinary pharmaceuticals used as feed additives in the poultry industry and antibiotic-resistant E. coli in broiler litter. These antimicrobials and resistant E. coli are trans ferred to the environment from contaminated litter and can contaminate surface and ground water through water run-off during storage of piles of litter. More research is needed to determine the pathogenicity of the described E. coli isolates as well as the potential effect of the veterinary pharmaceuticals on microbial flora in the environment.

ACKNOWLEDGMENTS

This research was funded by Health Canada and by Agriculture and Agri-Food Canada through the Agriculture Policy Framework GAPs program (Pacific AgriFood Research Centre, Agassiz contribution number 756).

• [cf7ic]