Risk Factors Associated with the Presence of Campylobacter Species in Norwegian Broiler Flocks T. M. Lyngstad,*1 M. E. Jonsson,* M. Hofshagen,* and B. T. Heier† *National Veterinary Institute, N-0106 Oslo, Norway; and †Norwegian Institute of Public Health, N-0403 Oslo, Norway ABSTRACT A case-control study was conducted in 2005 to identify risk factors for the presence of Campylobacter spp. in Norwegian broiler flocks. A total of 131 broiler farms (44 cases and 87 controls) were included in the study, and 1 flock from each farm was included in the statistical analyses. Data on farm and flock level were collected by means of a questionnaire designed for the present study. The relationship and strength of association between the presence of Campylobacter spp. in the broiler flocks and several factors were in-
vestigated by unconditional logistic regression. The following factors were found to be associated with an increased risk of testing positive for Campylobacter spp.: water from a private water source, swine holdings closer than 2 km, a specific slaughterhouse, a hired animal caretaker, transport personnel passing through the hygiene barrier when delivering day-old chickens, less than 9 d between depopulation and restocking, and multiple broiler houses on the farm.
Key words: broiler farm, broiler flock, Campylobacter species, epidemiology, risk factor 2008 Poultry Science 87:1987–1994 doi:10.3382/ps.2008-00132
INTRODUCTION Worldwide, campylobacteriosis is the most commonly reported enteric bacterial infection in the human population in developed countries (Blaser et al., 1983; Altekruse et al., 1999; Wagenaar et al., 2006). The incidence of campylobacteriosis in Norway increased substantially during the 1990s and peaked in 2001, with a total of 64 reported cases per 100,000 inhabitants. For approximately half the cases reported, the infection was acquired in Norway (Hofshagen et al., 2006; Norwegian Institute for Public Health, 2006). Food of poultry origin has been identified as a significant source of this disease (World Health Organization, 2001). The control of Campylobacter spp. along the food chain is most effective when the colonization in living animals can be prevented (World Health Organization, 2001; Wagenaar et al., 2006). Poultry easily become colonized with Campylobacter spp. in the gastrointestinal tract. Several epidemiological studies have identified important risk factors for the presence of Campylobacter spp. in broiler flocks, such as a a low level of biosecurity, undisinfected drinking water, season, the number of broiler houses on a farm, and the presence of other animals on the farm or in
the direct vicinity (Kapperud et al., 1993; Gibbens et al., 2001; Refregier-Petton et al., 2001; Bouwknegt et al., 2004; Barrios et al., 2006). A Norwegian study concluded that the outdoor environment is a major source of Campylobacter spp. for colonizing broilers (Johnsen et al., 2006a). An action plan against Campylobacter spp. in Norwegian broilers was implemented in 2001 to reduce consumers’ exposure to Campylobacter spp. through domestically produced broiler meat (Hofshagen and Kruse, 2005). The surveillance program in the action plan revealed that several broiler farms repeatedly produced Campylobacter spp.-positive broiler flocks. In 2005, special attention was paid to broiler farms that repeatedly produced positive flocks over several years to give these farms close follow-up advisory service. This study is part of the follow-up service of these farms. The aim of this study was to identify risk factors for the presence of Campylobacter spp. in Norwegian broiler flocks and to study the strength of association, thereby providing more knowledge to optimize preventive measures in broiler farms in Norway.
MATERIALS AND METHODS Study Population and Observational Unit
©2008 Poultry Science Association Inc. Received March 28, 2008. Accepted June 1, 2008. 1 Corresponding author:
[email protected]
The study population was Norwegian commercial broiler farms that produced broilers slaughtered before 50 d of age. This included virtually all commercial 1987
1988
Lyngstad et al.
broiler farms in Norway. In total, there were 501 broiler farms in Norway in 2004 (Register of Slaughtered Animals, Norwegian Agricultural Authority, Oslo, Norway). The observational unit was 1 broiler flock on each farm. A broiler flock was defined as broilers that were housed in the same broiler house on 1 farm during the same period.
Study Design According to the surveillance part of the plan, all flocks were sampled twice and tested for the presence of Campylobacter spp. on the farm approximately 1 wk before slaughter (farm sample) and at the slaughterhouse (slaughterhouse sample). A farm sample was positive when Campylobacter spp. was isolated from a pooled sample consisting of 10 swabs from fresh cecal droppings. Before May 1, 2004, a slaughterhouse sample was positive when Campylobacter spp. was isolated from a pooled sample of 10 cloacal swabs obtained at the slaughter line. After that date, a slaughterhouse sample was positive when Campylobacter spp. was isolated from a pooled sample of contents from 10 ceca. The bacteriological analyses of the farm samples were performed at the National Veterinary Institute in Trondheim. The bacteriological analyses of slaughterhouse samples were performed at laboratories close to the slaughterhouses. The method used was a slightly modified method described by Nordisk Metodikkkomite for Levnetsmidler (NMKL no. 119), and consists of direct-plating on modified charcoal cefoperazone deoxycholate agar (Hofshagen and Kruse, 2005; Nordisk Metodikkkomite for Levnetsmidler, 2008). All Campylobacter spp. isolates were typed to the species level. A broiler flock was defined as positive when Campylobacter spp. was isolated by testing either on the farm, at the slaughterhouse, or both. Otherwise, a broiler flock was defined as negative. This study was designed as a case-control study. A case was defined as a farm that had at least 1 positive flock in 2004 and a total of at least 3 positive flocks from 2001 to 2004. A total of 44 broiler farms fulfilled the criteria to be included as cases. A control was defined as a farm on which no positive flocks had been identified since the action plan against Campylobacter spp. began in 2001, and that had slaughtered at least 1 flock each year in the period from 2001 to 2004. To obtain a control, 2 farms per case were randomly selected from the group of farms fulfilling the control criteria. This was done by adding a random number to each farm and then selecting the 88 farms with the lowest numbers. For farms with 2 or more broiler houses, one of the houses was selected randomly. Figure 1 shows the geographical distribution of the cases and the controls.
Data Collection Information about the potential risk factors was collected by means of a questionnaire designed for the present study. The questionnaire included questions about a wide range of factors, such as identity of the broiler farm and stock, hygiene routines, ventilation, water supply, litter, surroundings, caretaker, visitors, and cleaning and disinfection (Table 1). The questionnaire was filled in during an interview performed either by telephone or by visiting the farm between June 2005 and November 2005. For the cases, the interviewers were asked to fill in the questionnaire referring to the last positive flock in 2004. For the controls, the interviewers were asked to fill in the questionnaire referring to a flock present in August 2004. Interviews of farmers on the case farms were performed by 7 advisors employed by the poultry industry in cooperation with staff from the Norwegian Food Safety Authority (Oslo, Norway). Interviews with farmers on the control farms were performed by the same 7 advisors. The questionnaires were returned to the National Veterinary Institute, where the data were entered into an Excel (Microsoft Corporation, Redmond, WA) database.
Statistical Analyses The statistical analyses were performed in SAS-PC System Version 9.1.3 for Windows (2002-2004, SAS Institute Inc., Cary, NC). The procedures PROC FREQ and PROC UNIVARIATE were used for the descriptive analyses. The majority of variables were categorical. The continuous variables were categorized with both the quartiles and the median. Unconditional logistic regression with Campylobacter spp. status as the outcome variable and broiler flock as the statistical unit was performed by using PROC LOGISTIC. All the independent variables were initially run in univariate analyses, and those having P < 0.25 were selected for a multivariate logistic regression model. The procedure PROC CORR was used to calculate the Spearman correlation coefficients between selected variables. For variables with a correlation coefficient above 0.4, the most biologically plausible was selected for the multivariate analysis. The modeling was performed by using a manually conducted backward selection process by subtracting 1 independent variable at a time in which the likelihood ratio test was used to test significance (P < 0.1). The same approach was used to test the significance of the 2-way interaction terms between the independent variables in the final model. The odds ratios were calculated from the estimated coefficients in the final model and used as a measure of the strength of association.
RISK FACTORS ASSOCIATED WITH CAMPYLOBACTER SPECIES
1989
Figure 1. Geographical distribution of the case and control farms, and the number of commercial broiler farms per municipality in Norway in 2005 (Source: National Veterinary Institute, Oslo, Norway; used with permission).
1990
Lyngstad et al.
Table 1. Risk factors for the introduction of Campylobacter spp. in Norwegian broiler flocks—Information collected by questionnaire Type of variable
Example of information collected
Broiler farm
Identity of the broiler farm and broiler flock; Campylobacter spp. status; number of broiler houses on the farm; date when the broiler house was built; date when broiler production started; if, or date when, the broiler house was modernized or renovated; location of the broiler house (e.g., separate or integrated into another building); floor heating in the broiler house; number of chickens in the flock; identity of the hatchery delivering day-old chickens; identity of the slaughterhouse
Hygiene routines
Presence and design of the hygiene barrier in the broiler house; date when the hygiene barrier was introduced; routines for handling the waste bucket with dead chickens; routines for hygiene measures such as hand washing, use of overalls, and change of footwear when passing the hygiene barrier; date when different measures were introduced; whether transport personnel passed by the hygiene barrier when delivering day-old chickens; whether the transport module cart passed from the transport vehicle and into the clean zone of the broiler house when delivering day-old chickens; feeding of the chickens on paper underlay (placed on the floor) during the first few days.
Cleaning and disinfection routines for the floor, walls, roof, drinking equipment, tools, feeding dispensers, ventilation system, overalls, footwear, and service room; whether the broiler house was closed after disinfection before new input; how many days the broiler house was closed before new input; date when routines for washing and disinfection were introduced
Ventilation
Use of net covering of the air inlet to prevent introduction of birds, date when this measure was established, control of the air inlet before arrival of day-old chickens, possibility of birds entering through the air outlet when the ventilation was turned off
Water supply
Use of official approved water or private water sources (unofficial water supply such as a private waterworks and private water sources as bore holes, wells, and lakes), filtering of water before entry into the broiler house, use of UV installation to disinfect water, date when UV equipment was installed, dimensions of UV installation, alarm on the UV installation, service routines of the UV installation
Litter
Kind of litter used in the broiler house (e.g., in packages or not), whether litter was placed in the broiler house before or after disinfecting, whether a tractor was used to distribute litter in the broiler house, temperature of the floor in the broiler house when litter was put into the broiler house
Surroundings
Type of ground surface in front of the doors and entrances to the broiler house; date if asphalt paving was performed; distance between the broiler house and a river or water stream, loch, or pool; chicken manure from one’s own farm; manure from cattle, swine, sheep, or goats; whether any farm deposit of manure was closer than 1 km from the broiler house; production animals closer than 2 km; pet animals or sport animals on the broiler farm
Caretaker, visitors Who is taking care of the broilers (e.g., the owner, hired caretaker, relatives or neighbor, or others), number of persons involved, gender of the animal caretaker, number of times with visitors in the broiler house, observation of pets or other animals, an abnormal number of flies in the broiler house
RESULTS Descriptive Statistics Completed questionnaires were received from all 44 cases and from 87 of the 88 controls. Eleven percent of the farms had multiple broiler houses. The average flock size was 12,700 broilers. Three different hatcheries delivered the day-old chickens and the broilers were slaughtered at 5 different slaughterhouses located in different regions in the middle and southern parts of Norway. Descriptive statistics for the levels of the categorical and categorized variables are presented in Table 2. The majority of farms had implemented hygiene routines, such as a hygiene barrier (94%), change of footwear (97%), overalls (78%), and hand washing (80%), when entering the broiler house. Measures to prevent wild birds from entering the building were also implemented (74%). None of the farms had observed cats or dogs in the broiler house. The majority of the broiler houses had asphalt or another type of solid ground covering in front of the entrance doors (83%). Wood chips were used as litter in most of the broiler houses (98%).
Approximately half of the broiler farms (52%) received water from the official waterworks; the rest received water from private water sources. Ultraviolet filters for disinfecting the water were reported on 35% of the farms. An ultraviolet (UV) filter was installed in 40% of the 44 case farms and in 32% of the 87 control farms. All UV filters were installed after the year 2000, with the majority of these installed in 2002 and 2003. On the majority of these farms (72%), the owners performed a yearly control of the UV installation themselves. Only 13% of the farms collected water samples for bacteriological analyses in 2004. Cattle were the most common (80%) production animals in a 2-km radius around the broiler house, followed by sheep or goats (56%), swine (55%), and poultry (51%). Cats (65%) and dogs (60%) were the most common companion animals on farms. Typing of Campylobacter spp. isolates to the species level revealed that Campylobacter coli was isolated at least once in 16% of all the case farms. Approximately half of the case farms (52%) had swine farms in the neighborhood (closer than 2 km). Of these farms, C. coli was isolated at least once in 30%, Campylobacter lari in 9%, and Campylobacter jejuni in 61%.
Distance to river or stream from the broiler house
Surroundings
High number of flies observed in the broiler house
Hired animal caretaker
Rabbits, guinea pig, or hamster on the farm
Dogs on the farm
Cattle closer than 2 km
Swine closer than 2 km
>1 1 A B C A B C D E A B C D Yes No No Yes No Yes No Yes No Yes 1 km ≤1 km >1 km Yes No Yes No Yes No No Yes Yes No Yes No
Categories 14 (7/7) 117 (37/80) 34 (17/17) 69 (21/48) 28 (6/22) 15(11/4) 38 (15/23) 44 (11/33) 8 (2/6) 26 (5/21) 30 (18/12) 29 (7/22) 34 (11/23) 38 (8/30) 13 (8/5) 118 (36/82) 16 (8/8) 115 (36/79) 36 (15/21) 95 (29/66) 34 (15/19) 97 (29/68) 59 (24/35) 72 (20/52) 65 (28/37) 66 (16/50) 33 (14/19) 97 (29/68) 63 (30/33) 68 (14/54) 101 (38/63) 30 (6/24) 112 (39/73) 19 (5/14) 59 (23/36) 72 (21/51) 105 (38/67) 26 (6/20) 53 (22/31) 78 (22/56) 113 (41/72) 18 (3/15) 15 (8/7) 116 (36/80) 5 (3/2) 126 (41/85)
2.2 1 3.7 1.6 1 11.6 2.7 1.4 1.4 1 5.6 1.8 1.2 1 3.6 1 2.2 1 1.6 1 1.8 1 1.7 1 2.4 1 1.7 1 3.5 1 2.4 1 5.1 1 1.6 1 1.9 1 1.8 1 2.8 1 2.5 1 3.1 1
Odds ratio
0.5, 19.3
0.9, 7.5
0.8, 10.4
0.9, 3.8
0.7, 5.1
0.7, 3.2
1.1, 23.2
0.9, 6.4
1.6, 7.6
0.8, 3.9
1.1, 5.0
0.8, 3.6
0.8, 4.1
0.7, 3.5
0.8, 6.3
1.1, 11.9
1.9, 16.4 0.6, 5.1 0.4, 3.8
2.6, 52.0 0.8, 8.8 0.4, 4.6 0.2, 9.1
1.2, 11.3 0.6, 4.5
0.7, 6.6
95% confidence interval for odds ratio
0.22
0.09
0.11
0.12
0.21
0.24
0.04
0.08
0.001
0.19
0.02
0.12
0.13
0.23
0.14
0.03
0.007
0.01
0.05
0.18
P-value
2
Three different hatcheries labeled A, B, and C. Five different slaughterhouses labeled A, B, C, D, and E. 3 A = counties of Hedmark and Oppland; B = counties of Rogaland, Sogn, and Fjordane, and Hordaland; C = county of Trøndelag; D = counties of Østfold, Akershus, Buskerud, Vestfold, and Telemark.
1
Caretaker, visitors
Private water source or official waterworks
Water supply
Distance to manure from the broiler house
Net covering of air inlet to prevent introduction of birds
Number of days after depopulation and before restocking
Disinfection of the service room
Disinfection of footwear
Disinfection of the drinking installation
Cleaning of roof in the broiler house
Transport personnel passing through the hygiene barrier when delivering day-old chickens
Region for location of the broiler farm3
Slaughterhouse collecting broilers for slaughter2
Hatchery delivering day-old chickens
Ventilation
Hygiene routines
Number of broiler houses on the farm
Broiler farm 1
Risk factor
Variable
Number (case/control)
Table 2. Distribution of the independent variables and results from the univariate analyses with Campylobacter spp. status as an outcome variable in a study of 131 broiler farms (44 cases and 89 controls) in Norway in 2005
RISK FACTORS ASSOCIATED WITH CAMPYLOBACTER SPECIES
1991
1992
Lyngstad et al.
Table 3. Results from the multivariate logistic regression analysis with Campylobacter spp. status as an outcome variable in a study of 131 broiler farms (44 cases and 89 controls) in Norway in 2005
Variable
Risk factor
Categories
Broiler farm
Number of broiler houses on the farm
>1 1 A B C D E Yes
Slaughterhouse collecting broilers for slaughter1
Hygiene routines
Transport personnel passing through the hygiene barrier when delivering day-old chickens Number of days after depopulation and before restocking
Water supply
Private water source or official waterworks
Surroundings
Swine closer than 2 km
Caretaker, visitors Hired animal caretaker 1
No
