Abstract
Campylobacter species are one of the leading causes of foodborne disease in the United States. Campylobacter jejuni and Campylobacter coli are the two main species that are of concern to human health, and they cause approximately 95% of human infections. The number of studies investigating Campylobacter in chicken livers and gizzards is very limited in the literature. The objective of this study was to determine the prevalence of Campylobacter jejuni and Campylobacter coli in retail chicken livers and gizzards purchased from grocery stores in the Tulsa, Oklahoma, area and to further characterize the isolates obtained through antimicrobial susceptibility testing. A total of 202 retail chilled chicken livers and gizzards (159 livers and 43 gizzards) were purchased on a weekly basis from several grocery stores. The overall prevalence of Campylobacter in chicken livers and gizzards was 136/202 (67%), where 69/202 (34%) of the samples were contaminated with Campylobacter jejuni and 66/202 (33%) with Campylobacter coli. While the prevalence of Campylobacter in chicken livers was 77%, its prevalence in chicken gizzards was lower at 33%. The prevalence of Campylobacter jejuni was slightly higher in chicken livers (36%) than gizzards (26%), while the prevalence of Campylobacter coli was significantly higher in the chicken livers (40%) than chicken gizzards (7%). The prevalence of resistance among C. jejuni and C. coli isolates recovered against 16 antimicrobials were as follows: amoxicillin (98%, 99%), ampicillin (32%, 55%), azithromycin (10%, 25%), cephalothin (92%, 99%), chloramphenicol (4%, 12%), ciprofloxacin (58%, 48%), clindamycin (5%, 19%), doxycycline (39%, 66%), erythromycin (6%, 32%), gentamicin (9%, 43%), kanamycin (11%, 43%), nalidixic acid (50%, 43%), oxytetracycline (99%, 100%), streptomycin (3%, 18%), tetracycline (37%, 60%), and tilmicosin (9%, 16%). Multidrug resistance was higher among Campylobacter coli than Campylobacter jejuni isolates.
Introduction
C
Human campylobacteriosis can be treated with antibiotics such as erythromycin (a macrolide), which can be considered as the optimal choice in treating Campylobacter enteritis for its low cost, safety particularly on children and pregnant women, and low reported resistance (Allos, 2001). Ciprofloxacin (a fluoroquinolone) can also be recommended to treat enteritis, and intravenous aminoglycosides are usually used for the treatment of bacteremia and other systemic infections (Aarestrup and Engberg, 2001). In recent years, there have been reports emerging of Campylobacter that are resistant to these and many other drugs, which may complicate the treatment of human infections (Chatre et al., 2010).
Antimicrobial resistance in foodborne pathogens is of significant concern as it may complicate the treatment of human infections (Summers, 2006; Chatre et al., 2010). This is due to the fact that many of the drugs that are used to treat human infections are used in animal husbandry as prophylactics and feed supplements, which have been shown to promote the selection of resistant isolates that may affect human health if they enter the food chain (Zhu et al., 2006). There has been a correlation shown between the increased use of certain antibiotics and increased resistance to these antibiotics (Zhu et al., 2006; White et al., 2002).
Poultry giblets (particularly chicken livers and gizzards) are not only consumed as human or pet food but are also often used as fishing bait. If a fisherman uses his bare hands to handle the chicken livers bait and then eats a sandwich afterwards without a proper hand wash, he might be at risk from cross contaminating his own food. The number of studies discussing Campylobacter in chicken livers and gizzards is very limited in the literature. The objective of this study was to determine the prevalence of C. jejuni and C. coli in retail chicken livers and gizzards purchased from the Tulsa, Oklahoma area grocery stores and to further characterize the isolates obtained through antimicrobial susceptibility testing.
Methods
Samples collection and Campylobacter isolation
A total of 202 retail chilled chicken livers and chicken gizzards were purchased from several Tulsa area grocery stores on a weekly basis from January to June of 2010. Samples were selected to be as variable as possible (with different expiration dates, production codes, etc.). The retail meat samples were transported to the laboratory in ice boxes and processed immediately upon arrival. Samples were then rinsed with Buffered Peptone Water (BPW; EMD, Gibbstown, NJ) in sterile plastic bags (VWR Scientific, Radnor, PA) and massaged briefly by hand for 5 min. Next, 10 mL of the rinsate was added to 10 mL of 2×Bolton broth supplemented with blood and the appropriate antibiotic supplementation. The enrichment was incubated at 42°C for 48 h and then plated onto Campylobacter Charcoal Desoxycholate Agar (CCDA; Remel, Lenexa, KS) plates with the appropriate antibiotic supplementation (Zhao et al., 2001). The plates were incubated at 42°C for 48 h in gas jars containing microaerophilic gas-generating kits (Mitsubishi Gas Chemical, New York, NY). Four to six suspect Campylobacter colonies of each sample from the CCDA plates were then transferred onto Mueller-Hinton (MH; Difco, Sparks, MD) agar plates supplemented with blood and incubated at 42°C for 48 h under microaerophilic conditions, then purified by sub-culturing and kept at −80°C in a freezer for preservation until subjected to molecular identification.
DNA extraction
Bacterial DNA extracts used in polymerase chain reaction (PCR) were prepared from Campylobacter cultures using the single-cell lysing buffer (SCLB) method (Marmur, 1961). Isolates were removed from −80°C storage, struck to MH agar (Difco), and incubated at 42°C for 48 h under microaerophilic conditions. One colony was picked from the plate and suspended in 40 μL of SCLB solution in a 0.2-mL microtube. The SCLB solution consisted of 10 μL of 5 mg/mL proteinase K (Amresco, Solon, OH) and 1.0 mL of TE buffer (10 mM Tris-HCl [J.T. Baker, Phillipsburg, NJ] and 1 mM EDTA [Fisher, Fair Lawn, NJ]). The cells were lysed by heating at 80°C for 10 min, followed by cooling to 55°C for 10 min, using a Mastercycler Gradient thermocycler (Eppendorf, Eppendorf, Germany). The suspension was diluted 1:2 in sterile double distilled water and centrifuged in a Microfuge (Clover Laboratories, Waterville, OH) at 4500×g for 30 s to remove cellular debris. The supernatant was used as DNA template for PCR. All DNA extract samples were stored at −20°C.
PCR identification
All Campylobacter suspect isolates were tested for the identification of Campylobacter genes by multiplex PCR reaction using primers specific for C. jejuni and C. coli (Cloak and Fratamico, 2002). C. jejuni American Type Culture Collection [ATCC] 33560 and C. coli strain 96121033 (Oklahoma Animal Disease Diagnostic Laboratory, Oklahoma State University) were used as positive controls, and sterile water was used as a negative control. Gels were stained with ethidium bromide and viewed and recorded by ultraviolet (UV) transillumination, using a UV imager (UVP, Upland, CA). The expected amplicon sizes were 160 bp for the C-1 gene (Winters and Slavik, 1995), 400 bp for the cadF gene (Konkel et al., 1999), and 894 bp for the ceuE gene (Gonzalez et al., 1997).
Antimicrobial susceptibility testing
Antimicrobial susceptibility testing was performed using the agar dilution method against 16 antimicrobials belonging to nine different antibiotics classes. Isolates were grown on MH agar (Difco) supplemented with 5% laked horse blood (Hemostat Laboratoties, Dixon, CA) and incubated for 48 h at 42°C at microaerophilic conditions. Cultures were then added to MH broth (Difco), adjusted to turbidity equal to a 0.5 McFarland standard, and inoculated onto 6-inch MH agar plates supplemented with 5% defibrinated sheep blood and antimicrobials at different concentrations, including the breakpoint established for each antimicrobial according to the Clinical and Laboratory Standards Institute (CLSI) when available (CLSI, 2011) using a 1-mm 96-point replicator (McDermott et al., 2004) starting at the lowest concentration of each antimicrobial and working up. C. jejuni ATCC 33560 was used as a quality control strain. The plates were incubated at 42°C for 48 h under microaerophilic conditions. The plates were read for growth or no growth, and denoted as resistant or susceptible, respectively, according to the breakpoints for each of the 16 tested antimicrobials. The following 16 antimicrobials were used to screen the susceptibility of the 293 tested Campylobacter isolates: • Aminoglycosides: gentamicin (breakpoint 8 μg/mL, range 4–64 μg/mL), kanamycin (breakpoint 64 μg/mL, range 32–512 μg/mL), streptomycin (breakpoint 64 μg/mL, range 32–512 μg/mL) • Beta-lactams: amoxicillin (breakpoint 32 μg/mL, range 16–256 μg/mL), ampicillin (breakpoint 32 μg/mL, range 16–256 μg/mL) • Cephalosporins: cephalothin (breakpoint 32 μg/mL, range 16–256 μg/mL) • Fluoroquinolones: ciprofloxacin (breakpoint 4 μg/mL, range 2–32 μg/mL) • Lincosamides: clindamycin (breakpoint 8 μg/mL, range 4–64 μg/mL) • Macrolides: azithromycin (breakpoint 8 μg/mL, range 4–64 μg/mL), erythromycin (breakpoint 32 μg/mL, range 16–256 μg/mL), tilmicosin (breakpoint 8 μg/mL, range 4–64 μg/mL) • Phenicols: chloramphenicol (breakpoint 32 μg/mL, range 16–256 μg/mL) • Quinolones: nalidixic acid (breakpoint 64 μg/mL, range 32–512 μg/mL) • Tetracyclines: doxycycline (breakpoint 8 μg/mL, range 4–64 μg/mL), oxytetracycline (breakpoint 2 μg/mL, range 1–16 μg/mL), tetracycline (breakpoint 16 μg/mL, range 8–128 μg/mL)
Results
Prevalence of Campylobacter in chicken livers and gizzards
A total of 202 retail chilled chicken liver and chicken gizzard samples were purchased from several Tulsa area grocery stores on a weekly basis from January to June of 2010. Samples were selected to be as variable as possible (with different expiration dates, production codes, etc.). The number of chicken livers samples was 159, and the number of chicken gizzards samples was 43 (Table 1). The chicken livers and gizzards samples belonged to two major brands (Brand A and brand C), except one sample of chicken livers and one sample of chicken gizzards that were purchased from an ethnic grocery store and were displayed as their store brand (Brand B in Table 1).
As shown in Table 1, the overall prevalence of Campylobacter (C. jejuni and C. coli) in chicken livers and gizzards was 136/202 (67%), where 69/202 (34%) were contaminated with C. jejuni, 66/202 (33%) of the samples was contaminated with C. coli, and 1/202 (0.5%) of the samples was contaminated with both C. jejuni and C. coli. The prevalence of Campylobacter (C. jejuni and C. coli) in chicken livers was 122/159 (77%), which was higher than its prevalence in chicken gizzards, where it was 14/43 (33%; Table 1). The prevalence of C. jejuni was also slightly higher in chicken livers (36%) than in chicken gizzards (26%), and the prevalence of C. coli was significantly higher in chicken liver samples (40%) than in the chicken gizzards (7%). One chicken livers sample out of the tested 202 chicken livers and gizzards samples was contaminated with both C. jejuni and C. coli (Table 1). While the prevalence of Campylobacter in brand “A” in the chicken livers and gizzards (71%) was comparable to brand “C” (65%), the prevalence of C. jejuni was higher in brand “A” (49%) than brand “C” (20%). On the other hand, the prevalence of C. coli in chicken livers and gizzards was higher in brand “C” (44%) than in brand “A” (22%; Table 1).
Antimicrobial resistance profiling
A total of 293 Campylobacter isolates (270 chicken livers isolates and 23 chicken gizzards isolates) were subjected to antimicrobial resistance profiling against 16 different antimicrobials that belong to nine different antibiotic classes (Table 2). As shown in Table 2, the percentage of resistance of the 293 Campylobacter isolates from chicken livers and gizzards combined to the 16 tested antimicrobials were as follows: amoxicillin (99%), ampicillin (45%), azithromycin (18%), cephalothin (96%), chloramphenicol (9%), ciprofloxacin (52%), clindamycin (13%), doxycycline (54%), erythromycin (20%), gentamicin (28%), kanamycin (29%), nalidixic acid (46%), oxytetracycline (100%), streptomycin (12%), tetracycline (50%), and tilmicosin (13%). As shown in Table 2, the percentage resistance to the 16 tested antimicrobials varied between C. jejuni and C. coli isolates. The percentage of resistance of the 129 C. jejuni and 164 C. coli isolated from chicken livers and gizzards to the 16 tested antimicrobials was as follows: amoxicillin (98%, 99%), ampicillin (32%, 55%), azithromycin (10%, 25%), cephalothin (92%, 99%), chloramphenicol (4%, 12%), ciprofloxacin (58%, 48%), clindamycin (5%, 19%), doxycycline (39%, 66%), erythromycin (6%, 32%), gentamicin (9%, 43%), kanamycin (11%, 43%), nalidixic acid (50%, 43%), oxytetracycline (99%, 100%), streptomycin (3%, 18%), tetracycline (37%, 60%), and tilmicosin (9%, 16%; Table 2).
The distribution of the multidrug resistance (MDR) among the 270 Campylobacter (C. jejuni and C. coli) chicken liver isolates was as follows: 60 isolates resistant to one to four antimicrobials, 152 isolates resistant to five to seven antimicrobials, and 81 isolates resistant to more than seven antimicrobials (data not shown). A dendrogram created using BioNumerics software indicated that MDR was higher in C. coli isolates than in C. jejuni strains (Fig. 1).
A dendrogram created using the BioNumerics software showing a simple comparison of the isolates by species showing antimicrobial resistance profiles. Campylobacter jejuni are labeled with black squares, and Campylobacter coli are labeled with gray squares. The meat source is listed by each isolate code as chicken livers or chicken gizzards. The 16 tested antimicrobials are listed on the top of the dendrogram with their three-letter abbreviations as follows: ampicillin (AMP), erythromycin (ERY), nalidixic acid (NAL), tetracycline (TET), streptomycin (STR), kanamycin (KAN), oxytetracycline (OXY), amoxicillin (AMX), gentamicin (GEN), ciprofloxacin (CIP), clindamycin (CLI), azithromycin (AZI), doxycycline (DOX), chloramphenicol (CHL), tilmicosin (TIL), and cephalothin (CEP).
Discussion
The fact that the overall prevalence of Campylobacter in chicken livers and gizzards in this study was 67% (Table 1) is in agreement with some recent reports (Sallam, 2007; Suzuki and Yamamoto, 2009). In contrast, a higher prevalence (over 90%) was reported in some earlier studies (Fernandez and Pison, 1996; Shih, 2000). While the prevalence of Campylobacter in chicken livers was 77%, its prevalence in chicken gizzards was lower at 33%. Sallam (2007) reported similar results, but an earlier study with very few samples tested indicated that Campylobacter prevalence in liver samples (29%) was lower than those in gizzards (36.4%) (Denis et al., 2001).
The prevalence of C. jejuni was significally higher in chicken liver samples (36%) than in chicken gizzards (26%), while the prevalence of C. coli was significantly higher in the chicken liver samples (40%) than in the chicken gizzards (7%; Table 1). C. jejuni was the most prevalent species found in livers and gizzards in most of the earlier studies (Sallam, 2007; Watanabe et al., 2005; Kawamori et al., 2003), but Fernandez and Pison (1996) detected higher rates of C. coli (78.6%) than of C. jejuni (21.4%). The variation in prevalence data of Campylobacter in the literature can be explained by variable processing facilities, locations, collection time of the year, and variable methodologies. To our knowledge, our study is unique in using over 200 chicken liver and gizzard samples to determine the prevalence and antimicrobial resistance of C. jejuni and C. coli. The majority of previous studies on the prevalence of Campylobacter in chicken livers and gizzards used smaller number of samples and were generally part of larger studies investigating retail chicken meats.
The distribution of the MDR among the 270 Campylobacter chicken livers isolates showed that the majority of the isolates (152) were resistant to five to seven antimicrobials and that 81 isolates were resistant to more than seven antimicrobials. A dendrogram created using the BioNumerics software indicated that MDR was clearly higher in C. coli isolates than in C. jejuni strains (Fig. 1). It was also apparent from Table 2 that the percentage of C. coli isolates resistant to most of the 16 tested antimicrobials was higher than that for C. jejuni. For example, while the percentage of C. jejuni resistance to erythromycin was only 6%, it was 32% in case of C. coli. Ge et al. (2003) found the resistance rate to erythromycin was 61% in the case of C. coli.
The Campylobacter isolates screened in this study were shown to be highly resistant to oxytetracycline (100%), amoxicillin (99%); and cephalothin (96%); the lowest resistance rates were to chloramphenicol (9%), clindamycin (13%), streptomycin (12%) and tilmicosin (13%). High resistance rates have been previously detected for oxytetracycline (Ishihara, et al., 2004). Velazquez et al. (1995) detected high rates of resistance to cephalothin (up to 100%). Cephalothin is a cephalosporin that has been widely used, which might explain the increased resistance (Allos and Blaser, 1995). Antibiotic resistance in Campylobacter could be due to several factors such as increased antimicrobial use in animal husbandry as growth promoters and prophylactics, and this could have led to the proliferation of antimicrobial resistance due to selection (Aarestrup and Engberg, 2001). Also, Campylobacter contains the cmeABC and other efflux pumps, which could help in the active removal of the antimicrobials from the inner membrane space, hence making the drugs ineffective (Lin et al., 2002; Pumbwe et al., 2004). Resistance genes and efflux pumps make very effective resistance mechanisms against most antibiotics, which means that it is harder to not only treat the animals for infections but also eventually humans too (Summers, 2006). Campylobacters have been known to carry the tetO gene in plasmids, which confers resistance to tetracycline (Saenz et al., 2000; Sougakoff et al., 1987; Avrain et al., 2004). Production of beta-lactamases by bacteria have increased the resistance against beta-lactam antibiotics (Aarestrup and Engberg, 2001).
The high prevalence of Campylobacter in retail chicken livers and gizzards is alarming and might be due to cross contamination since the giblets are recovered from several chickens and pooled together. In a study by Fernandez and Pison (1996), the authors reported a high prevalence (92.9%) of Campylobacter in commercial chicken livers and suggested that this is due to cross contamination and not due to the existence of infected chicken livers. Boukraa et al. (1991) isolated C. jejuni and C. coli from only 18.8% of 223 livers of broiler chickens with necrotic hepatitis lesions. Also, Cox et al. (2006) detected Campylobacter in only 9% of 43 liver/gallbladders of commercial broiler breeder hens, and in another study, Cox et al. (2009) reported the presence of Campylobacter in 13–17% of liver-gallbladder of laying hens. To determine whether chicken livers infected with C. jejuni are seeded in vivo or contaminated after slaughtering, Barot et al. (1983) found that, of 56 livers positive for C. jejuni, two yielded tissue growths only, while 54 showed growth from the surface. The authors of this study concluded that the commercially available chicken livers in New York at that time were contaminated after slaughtering and this did not reflect the presence of an in vivo acquired hepatitis C. jejuni infection, particularly given that neither scarring nor macroscopically overt abscesses were present in any of the livers examined. As discussed above, the high prevalence of Campylobacter in retail chicken livers in several studies despite a relatively lower prevalence in the liver as an internal organ of the chicken supports the theory of surface cross contamination. Ghafir et al. (2007) suggested that the high level of recovery of Campylobacter from livers is probably due to the fact that the liver surface stays moist, which might protect this foodborne pathogen.
In a cohort study of a campylobacteriosis outbreak in the United Kingdom in June of 2010, analyses strongly supported the hypothesis that this outbreak was caused by the consumption of chicken liver parfait (Inns et al., 2010). Few recent reports in England indicated that Campylobacter outbreaks that were linked with chicken liver dishes increased significantly, indicating the escalating importance of this type of food as a public health concern (Little et al., 2010; HPA, 2009). Adding to the importance of chicken livers as a public health risk is the recent discovery by Strachan et al. (2012) that molecular source attribution by multi-locus sequence typing (MLST) demonstrated that Campylobacter strains from chicken livers were most similar to those found commonly in humans, which provides further evidence that chicken liver is a probable source of human infection. Thorough cooking of chicken liver would be a good food safety practice to reduce such a risk, but cooking for too long of a time is not always desired since the livers become gray and unappealing; if cooked inadequately, however, they may still harbor Campylobacter. In an effort to solve this issue, Whyte et al. (2006) reported that pan frying of chicken livers at or above an internal temperature of 70°C for 2–3 min would destroy Campylobacter while preserving the appealing pink color. The fact that people often use chicken livers as fishing bait is troubling to some. If a fisherman uses his bare hands to handle the chicken liver bait and then eats a sandwich afterwards without a proper hand wash, he might be at risk from cross contaminating his own food. The increased incidence of antimicrobial resistance seen in this study raises concern about the safety of retail chicken products particularly livers and gizzards. There have been very limited studies on foodborne pathogens in poultry livers and gizzards in the United States, and the results of this study warrant additional research, particularly molecular typing to investigate similarities to strains causing human infections.
Footnotes
Acknowledgments
We are grateful to Dr. Brenda Love (Oklahoma Animal Disease Diagnostic Laboratory, Oklahoma State University) for providing Campylobacter coli strain 96121033. We would like to acknowledge financial support from the University of Tulsa Office of Research and Sponsored Programs (Student Research Grant Program) in subsidizing the purchase of the retail meat samples used in this study.
Disclosure Statement
No competing financial interests exist.