Prevalence and antimicrobial susceptibility of thermophilic Campylobacter in organic and conventional broiler flocks
Dr Heuer, Danish Veterinary Laboratory, Department of Poultry, Fish and Fur Animals, Hangøvej 2, DK-8200 Århus N, Denmark (e-mail: firstname.lastname@example.org).
Aims: To determine the flock prevalence and to estimate the within flock prevalence of Campylobacter in broiler flocks from different rearing systems, and to determine the antimicrobial susceptibility of Campylobacter isolates to selected antimicrobial substances.
Methods and Results: One hundred and sixty broiler flocks originating from organic, conventional and extensive indoor production farms were investigated for the presence of Campylobacter at the time of slaughter. Campylobacter isolates from a subsample of positive flocks were subjected to susceptibility testing. Campylobacter spp. were isolated from 100% of organic broiler flocks, from 36·7% of conventional broiler flocks and from 49·2% of extensive indoor broiler flocks. Six of 62 Campylobacter isolates were resistant to one or more of the antimicrobials tested.
Conclusions: These results indicate that the special characteristics of organic broiler production provide a high prevalence of Campylobacter-positive flocks. Antimicrobial resistance was scarce among Campylobacter isolates from all rearing systems.
Significance and Impact of the Study: Organic broiler flocks constitute a strong potential for introduction of Campylobacter to the processing line upon arrival at slaughter.
Campylobacter is a well recognized cause of human enteritis, and food-borne campylobacteriosis is considered a main problem of public health in many developed countries. In 1999, campylobacteriosis was the most frequent food-borne zoonosis in Denmark (78 cases per 100 000 inhabitants) (Anon. 2000a). Poultry products are suspected to be an important source of infection in Denmark as well as in other countries (Kapperud et al. 1992; Hanninen et al. 2000; Studahl and Andersson 2000; Neimann 2001).
Over the last decade, the occurrence and spread of Campylobacter in conventional broiler flocks has been intensively studied, whereas the occurrence of Campylobacter in broiler flocks of organic or other non-conventional origin has received less attention. Organic and other non-conventional broiler products are now readily available for retail in many countries, yet very little is known about the status of these broiler flocks with regard to the prevalence of Campylobacter.
A few reports on the prevalence of Campylobacter in non-conventional broiler production are available. In a French study, 85·7% of faecal samples from one flock of chickens raised in a free-range system were Campylobacter positive (Rivoal et al. 1999). In a study from Tanzania, 76·5% of samples of droppings from indigenous free-range poultry were Campylobacter positive (Kazwala et al. 1993), and in a study of Campylobacter in domestic free range chickens and confined commercial chickens in Peru, the isolation frequencies were 54% and 35%, respectively (Tresierra-Ayala et al. 1995). Although neither of these studies deal specifically with organic broilers, they do indicate that free-range rearing of poultry could be associated with a high prevalence of Campylobacter.
Reports from several countries on prevalence of Campylobacter in conventional broilers are available. Flock prevalences ranging from 18 to 82% have been reported (Humphrey et al. 1993; Kapperud et al. 1993; Jacobs-Reitsma et al. 1994a; Berndtson et al. 1996a; Hald et al. 2000). Although a direct comparison of reports on Campylobacter prevalence from different countries will be strongly biased by differences in sample size, isolation procedures and sample material used, the results reported suggests that in some countries, a considerable proportion of conventional broiler flocks are Campylobacter-negative.
In a few investigations the authors reported the proportion of Campylobacter-positive samples within the samples taken from one or more of the flocks included in the investigation (Berndtson et al. 1996b; Gregory et al. 1997; Evans and Sayers 2000; Hald et al. 2001). The results from these studies suggest that if a broiler flock is Campylobacter infected, a large proportion of the birds within in the flock is infected.
A seasonal variation in the prevalence of Campylobacter-positive broiler flocks has been reported from Denmark (Wedderkopp et al. 2000), as well as from Norway (Kapperud et al. 1993), Sweden (Berndtson 1996) and The Netherlands (Jacobs-Reitsma et al. 1994a).
Antimicrobial susceptibility of Campylobacter isolates from broilers in Denmark has been reported since 1995 through the Danish Integrated Antimicrobial Resistance Monitoring and Research Programme (DANMAP) (Anon. 2000b). However, these data do not provide comparison between different broiler rearing systems. From other countries, reports on antimicrobial susceptibility of Campylobacter isolates from broilers of conventional origin are available (Jacobs-Reitsma et al. 1994b; Berndtson 1996; Saenz et al. 2000), whereas similar knowledge regarding broiler flocks of organic and other non-conventional production origin is absent or scarce.
The aim of the present study was to determine the flock prevalence and to estimate the within flock prevalence of Campylobacter in broiler flocks from organic, conventional and extensive indoor production farms in Denmark. Furthermore, the study aimed to determine the antimicrobial susceptibility of Campylobacter isolates from broilers from the three rearing systems to selected antimicrobial substances.
MATERIALS AND METHODS
Broiler flocks of organic, conventional and extensive indoor rearing systems were selected for the study. Organic broiler production was established in Denmark in 1998. The organic producers have adopted strict rules of production, including restricted use of antimicrobial substances, free-range rearing and use of organic feed. A slow-growing breed of broilers is used in the organic production and the minimum age at slaughter is 81 days. Conventional broiler flocks are given high protein and high energy feed. They are reared in a confined environment at high stocking density and the age at slaughter is 36–42 days. Extensive indoor broiler flocks are produced in a less intensive rearing system in a confined environment with a low stocking density. These broilers are given low protein and low energy feed and the minimum age at slaughter is 56 days. Antimicrobial growth promoters are not in use with any of the three rearing systems.
One hundred and sixty broiler flocks from 39 farms were sampled in 1998–2000. Seventy-nine conventional broiler flocks originating from 18 houses at 18 farms, 59 extensive indoor broiler flocks originating from 16 broiler houses at nine farms and 22 organic broiler flocks originating from 12 free-range farms were sampled for the investigation.
The broiler flocks were slaughtered at four different abattoirs. In Denmark, the allocation of broiler flocks to different abattoirs is dependent on geographical location and ownership relations. One abattoir received flocks from all three rearing systems, one abattoir received conventional and extensive indoor broiler flocks, one abattoir received organic flocks only, and one abattoir received conventional flocks only.
Samples consisted of cloacal swabs taken from the broilers at slaughter. In order to enable determination of the flock prevalence as well as estimation of the within-flock prevalence of Campylobacter, 10 broilers from each flock were sampled individually. The sample size considerations were based on the assumption that the within-flock prevalence in Campylobacter-positive flocks would be 50% or higher. During transportation, the swabs were kept in sealed 10 ml tubes containing Brain Heart Infusion broth, 37 g l–1 (Difco), supplemented with 5% calf blood and 0·5% agar (Oxoid). If not sent immediately to the laboratory, the swabs were refrigerated at 5°C for a maximum of 3 days after sampling, before dispatch to the laboratory.
Isolation and identification of Campylobacter
Upon arrival at the laboratory, each swab was immersed in a 2 ml tube containing veal infusion broth supplemented with 4% new-born calf serum. The swabs were left for 10–15 min in the broth. Subsequently, 10 μl of the broth were streaked onto modified charcoal cefoperazone deoxycholate agar (CCDA; Oxoid) and incubated for 48 h at 42°C in incubation jars filled with a gas mixture of 65% N2, 25% H2 and 10% CO2. Campylobacter-like colonies were purified on blood agar and identified to species level on the basis of standard procedures comprising tests for hippurate and indoxyl acetate hydrolysis, catalase production, and susceptibility to cephalotin and nalidixic acid (On and Holmes 1991, 1992).
Susceptibility testing of Campylobacter isolates
The MIC of one Campylobacter isolate from each of 62 flocks to tetracycline, ampicillin, erythromycin, streptomycin and enrofloxacin was determined by the agar dilution method. The test was performed on Mueller-Hinton II agar (Becton Dickinson) supplemented with 5% bovine blood. The test strains were inoculated on the agar plates with a multipoint inoculator (Denley Instruments A400, West Sussex, UK). The dilution ranges used were as follows: tetracycline 0·5–32 μg ml–1, ampicillin 1–64 μg ml–1, erythromycin 0·25–32 μg ml–1, enrofloxacin 0·03125–4 μg ml–1 and streptomycin 1–64 μg ml–1. The following break-points were used: tetracycline, > 4 μg ml–1; ampicillin, > 16 μg ml–1; erythromycin, > 16 μg ml–1; enrofloxacin, > 2 μg ml–1 and streptomycin, > 32 μg ml−1. Four control strains, Escherichia coli ATCC 25922, Enterococcus faecalis ATCC 29212, Pseudomonas aeruginosa ATCC 27853 and Staphylococcus aureus ATCC 29213 were used with every test plate.
Two levels of analysis were performed on the prevalence data: (i) flock level analysis (the proportion of Campylobacter-positive flocks) and (ii) animal level analysis (the proportion of Campylobacter-positive swabs). A broiler flock was considered positive if Campylobacter was isolated from at least one swab sample from the flock. The data structure was hierarchical (animal/flock/house/farm/rearing system) and longitudinal (repeated measurements over time within each broiler house). In order to model the within–group correlation imposed by the hierarchical structure, the data were analysed in a Generalized Linear Mixed Model (GLMM)
with Xβ equal to the fixed part of the model (season and rearing system), ufarm and vhouse as random farm and house effects, respectively, and ɛ as the within-house error term. The error term ɛ was assumed binomial. The model parameters were estimated with Restricted Maximum Likelihood (REML) applying the SAS-macro glmm800.sas (SASv8·00), using the method of Wolfinger and O’Connell (1993). Furthermore, the results were checked in S-plus (Anon. 2000c) applying the reglm procedure, using the method of Schall (1991). Data on differences in numbers of resistant Campylobacter isolates between rearing systems were analysed by the Fisher exact test.
Campylobacter spp. were isolated from all of the 22 (100%) organic broiler flocks, from 29 of 79 (36·7%) conventional broiler flocks and from 29 of 59 (49·2%) extensive indoor broiler flocks. Thus, the proportion of Campylobacter-positive flocks was significantly higher for organic flocks compared with conventional flocks (P < 0·001) and extensive indoor flocks (P < 0·001). The difference between conventional flocks and extensive indoor broiler flocks was non-significant (P > 0·05).
The overall proportion of Campylobacter-positive flocks was significantly higher in the period from May to October (61 of 87) compared with the period from November to April (19 of 73), resulting in an odds ratio (OR) of 6·7. In both seasons, the proportion of Campylobacter-positive flocks was significantly higher for organic flocks compared with conventional flocks (P < 0·001) and extensive indoor flocks (P < 0·001), while the difference between conventional flocks and extensive indoor flocks was non-significant (P > 0·05).
When the total number of positive and negative swab samples within each flock in each rearing system was included in the calculation, the proportion of Campylobacter-positive swab samples was higher for organic flocks compared with conventional flocks (OR: 4·15) and extensive indoor flocks (OR: 3·31). The overall proportion of positive swab samples was significantly higher in the period from May to October compared with the period from November to April (OR: 4·44). The random farm and house effects were small compared with the within-house error, indicating a large variation in the proportion of positive samples in flocks from the same house. The parameter estimates and 95% C.I. of the odds ratios are given in Table 1.
Parameter estimates, odds ratios and 95% C.I. based on all flocks sampled
Within-flock prevalence of Campylobacter
The estimation of within-flock prevalence was based on the proportion of positive samples within each flock. Only Campylobacter-positive flocks were included in the calculation. The estimated within-flock prevalence was 65% for organic flocks, 68% for conventional flocks and 60% for extensive indoor flocks. These differences between rearing systems were non-significant. The mean prevalence of Campylobacter-positive samples in positive flocks was 64·2%. Among positive flocks only, the difference in the proportion of positive swab samples between the period from May to October and the period from November to April was not significant. The parameter estimates and 95% C.I. of the odds ratios are given in Table 2.
Parameter estimates, odds ratios and 95% C.I. based on Campylobacter-
positive flocks only
The species distribution among the positive flocks was as follows: 70 flocks C. jejuni (87·5%), eight flocks C. coli (10·0%), while two flocks (2·5%) yielded a mixture of C. jejuni and C. coli. The species distribution of Campylobacter in the three rearing systems was as follows: organic flocks C. jejuni 91%, C. coli 4·5% and mixed infections (C. jejuni/C. coli) 4·5%; conventional flocks C. jejuni 86·2%, C. coli 10·3% and mixed infections (C. jejuni/C. coli) 3·5%; extensive indoor flocks C. jejuni 86·2%, C. coli 13·8% and mixed infections (C. jejuni/C. coli) 0%.
Antimicrobial susceptibility of Campylobacter isolates
Three of 53 C. jejuni isolates and three of nine C. coli isolates were resistant to one or more of the antimicrobials tested. Among 19 C. jejuni isolates from organic flocks, one isolate was resistant to ampicillin and tetracycline. Two C. coli isolates from organic flocks were susceptible to five antimicrobials. Among 24 C. jejuni isolates from extensive indoor flocks, one isolate was resistant to enrofloxacin and one isolate was resistant to streptomycin. Five C. coli isolates from extensive indoor flocks were tested; one isolate was resistant to erythromycin and one was resistant to streptomycin. Ten C. jejuni isolates from conventional flocks were susceptible to five antimicrobials. Two C. coli isolates from conventional flocks were tested; one isolate was resistant to enrofloxacin and streptomycin. Isolates resistant to enrofloxacin were also resistant to nalidixic acid.
The results indicate that the special characteristics of organic broiler production provide a high prevalence of Campylobacter. Thus, organic broiler flocks constitute a strong potential for introduction of Campylobacter to the processing line upon arrival at slaughter. In other studies, a high prevalence of Campylobacter in poultry reared under free-range conditions has been reported (Kazwala et al. 1993; Tresierra-Ayala et al. 1995; Rivoal et al. 1999). Unlike conventional and extensive indoor broiler flocks, which are reared in a confined environment, organic broiler flocks have unimpeded access to soil and water in the open. Horizontal transmission from the environment has been put forward as a likely route of Campylobacter infection in broilers (Kazwala et al. 1990; Jacobs-Reitsma et al. 1995), and the use of hygiene measures has been shown to reduce the risk of infection in conventional rearing systems (van de Giessen et al. 1998; Evans and Sayers 2000; Hald et al. 2000; Gibbens et al. 2001). Hence, the presence of Campylobacter in water in the open (Hanninen et al. 1998; Obiri-Danso and Jones 1999) is a possible explanation for the high Campylobacter prevalence observed in organic broilers.
However, the high age at slaughter (> 81 days) of organic broilers may influence the flock prevalence of Campylobacter. Other investigators have reported that the risk of flock infection increased with the age of the broilers (Berndtson et al. 1996b; Evans and Sayers 2000). The use of a slow-growing breed of broilers is a special characteristic of organic broiler production. Different breeds of chicken may not be equally susceptible to Campylobacter colonization. Stern et al. (1990) found that resistance to caecal colonization by C. jejuni was significantly influenced through chicken host lineage.
Thus, due to co-linearity among factors such as age, breed, housing, feed, abattoir and rearing system, no single factor related to organic broiler production can be pointed out as the sole determinant of high Campylobacter prevalence. Rather, the prevalence results reported in this study reflect the combined effect exerted by factors that are inextricably related to each broiler rearing system.
A relatively high prevalence of Campylobacter-positive flocks was observed among extensive indoor flocks (49·2%) compared with conventional flocks (36·7%). Although not significant in the present investigation (P > 0·05), the observed difference between the two groups can probably be attributed to the effect of age at the time of testing, 56 days for extensive indoor flocks vs. 36–42 days for conventional flocks. Higher age would make extensive indoor flocks more likely to be Campylobacter infected (Lindblom et al. 1986; Berndtson et al. 1996b; Evans and Sayers 2000). The flock prevalence of 36·7% observed in conventional broiler flocks differs from the prevalence of 52% reported in a previous Danish study performed in 1995 (Hald et al. 2000). Meanwhile, the widespread implementation of hygiene barriers in recent years, and pricing of Campylobacter-free flocks in Denmark, are potential explanations for this difference.
Although large and significant differences in flock prevalence of Campylobacter were observed between organic and conventional flocks, and between organic and extensive flocks, the differences in within-flock prevalence between the three rearing systems were non-significant. Thus, the origin of a Campylobacter-positive flock did not seem to affect the within-flock prevalence. Results of other studies indicated that a large proportion of the individual broilers in Campylobacter-positive flocks were colonized (Berndtson et al. 1996b; Gregory et al. 1997; Evans and Sayers 2000; Hald et al. 2001) while in this study, most positive flocks yielded some negative swab samples. This may reflect actual differences in within-flock prevalence of Campylobacter at the study locations at the time of sampling, or differences in sample handling procedures and sample material used (caecal samples, swab sampling of faecal droppings or cloacal swabs).
For conventional and extensive indoor flocks the observed species distribution (C. jejuni/C. coli) was in agreement with results of previous studies of Campylobacter in Danish broiler flocks (Hald et al. 2000; Wedderkopp et al. 2000). For organic flocks, the proportion of C. jejuni-positive flocks and flocks that yielded a mixture of C. jejuni and C. coli, was slightly elevated compared with the other production categories.
A low level of antimicrobial resistance was found among Campylobacter isolates from the three broiler rearing systems. This observation correlated with results reported through the DANMAP (2000b). A low level of antimicrobial resistance in Campylobacter isolates from broilers was also reported from Sweden (Berndtson 1996). Other investigators found high levels of resistance to ampicillin and tetracycline in Campylobacter isolates from broilers (Jacobs-Reitsma et al. 1994b; Saenz et al. 2000). These variations might reflect differences in consumption of antimicrobial agents in the broiler production in different countries. In the present investigation, low numbers of resistant isolates (six of 62 isolates) hampered comparison of resistance patterns of C. jejuni as well as of C. coli isolates between the three rearing systems. Thus, no relation between resistance pattern and origin of the Campylobacter isolates could be established.
The authors thank Dr Niels Tornøe, Danpo A/S, for his invaluable advisory support and collaboration, Dr Flemming Bager, DVL, for review of the manuscript and Dr Lars Bogø Jensen, DVL, for comments on the susceptibility testing. They thank the technical staff at Danpo A/S for their co-operation, and Eva Haarup Sørensen, Kirsten Cristensen, Camilla Hansen, Susanne Obsen and Anita Fogh Hansen at DVL, Århus, for excellent laboratory assistance. This study was funded by The Danish Directorate for Food, Fisheries and Agro Business.