Free access
Research Article
December 2013

Risk Factors for Campylobacteriosis in Two Washington State Counties with High Numbers of Dairy Farms


Campylobacteriosis is a frequently reported, food-borne, human bacterial disease that can be associated with ruminant reservoirs, although public health messages primarily focus on poultry. In Washington State, the two counties with the highest concentrations of dairy cattle also report the highest incidences of campylobacteriosis. Conditional logistic regression analysis of case-control data from both counties found living or working on a dairy farm (odds ratio [OR], 6.7 [95% confidence interval [CI], 1.7 to 26.4]) and Hispanic ethnicity (OR, 6.4 [95% CI, 3.1 to 13.1]) to have the strongest significant positive associations with campylobacteriosis. When the analysis was restricted to residents of one county, Hispanic ethnicity (OR, 9.3 [95% CI, 3.9 to 22.2]), contact with cattle (OR, 5.0 [95% CI, 1.3 to 19.5]), and pet ownership (OR, 2.6 [95% CI, 1.1 to 6.3]) were found to be independent risk factors for disease. Campylobacter jejuni isolates from human (n = 65), bovine (n = 28), and retail poultry (n = 27) sources from the same counties were compared using multilocus sequence typing. These results indicated that sequence types commonly found in human isolates were also commonly found in bovine isolates. These findings suggest that, in areas with high concentrations of dairy cattle, exposure to dairy cattle may be more important than food-borne exposure to poultry products as a risk for campylobacteriosis.


Campylobacteriosis is the most frequently reported enteric bacterial disease in Washington State ( Campylobacter jejuni infection in humans causes a self-limiting diarrheal illness, which may be accompanied by abdominal cramps and bloody diarrhea (1). Approximately 17% of cases require hospitalization (2). Guillain-Barré syndrome is a rare but severe sequela to infection that is characterized by flaccid paralysis (3). Epidemiological studies conducted since C. jejuni was first identified as a human pathogen have consistently found that consumption of undercooked poultry is an important source of infection (411). However, among six published studies that reported etiological fractions from Washington State (5), Australia (10), Switzerland (12), the United States (6), and King County, Washington (8), none determined that the food-borne route of exposure to poultry accounted for more than approximately 50% of disease risk, suggesting that other sources and routes of transmission contribute significantly to the total disease burden.
In Washington State, Whatcom County and Yakima County have the highest concentrations of dairy cattle ( Both counties also had significantly higher average annual reported campylobacteriosis rates (expressed as cases per 100,000 population) between 2006 and 2010 than did other Washington State counties ( (see Fig. S1 in the supplemental material). Similar correlations between disease incidence and cattle density have been reported for Manitoba and Ontario, Canada, and Sweden (1315). In a cross-sectional survey of 96 dairy operations in the United States, Englen et al. found that 97.9% of dairy operations and 51.2% of individual samples were positive for Campylobacter spp. (16). In a longitudinal feedlot study, the per-animal prevalence increased with time in the feedlot, from 1.6% to 61.3% just before slaughter (17). Domestic cattle, including both feedlot and dairy cattle, therefore may represent a significant reservoir for this human pathogen.
To test the hypothesis that direct contact with cattle and their environment may contribute significantly to human campylobacteriosis in Washington State and to quantify that contribution, we conducted a case-control study targeting high-cattle-density counties in Washington State. In addition, we compared C. jejuni isolates from human, cattle, and retail chicken sources using multilocus sequence typing (MLST) (18).


Case identification and interviews.

Campylobacteriosis is a notifiable disease in Washington State, so epidemiologists at county local health jurisdictions (LHJs) receive case reports from health care providers and clinical laboratories. Disease investigators at the LHJs routinely conduct case interviews. Yakima County and Whatcom County utilize the Washington State Department of Health (DOH) standardized campylobacteriosis questionnaire, which includes questions about consumption of undercooked meat and poultry, consumption of unpasteurized milk, and other exposures commonly associated with Campylobacter infections ( Because the cattle, farm, and water exposure-related questions on the state questionnaire lacked specificity, a supplemental set of questions was used by the LHJs during the study period (January 2009 through December 2010) (the questionnaire is available on request). Data from the Washington DOH standard questionnaire were transferred electronically from county health personnel to the Washington DOH communicable disease epidemiology division. These data were deidentified and transferred from the Washington DOH to the study project leader at Washington State University (WSU). Supplemental questionnaires were mailed from the LHJs to WSU and subsequently were linked with the Washington DOH data by case identification number.
An eligible case was defined as any laboratory-confirmed campylobacteriosis case reported to a Whatcom County or Yakima County local health jurisdiction. To be eligible, a case patient's primary residence must have been within one of those counties during the study period. Case patients who traveled during their entire exposure period were excluded. The exposure period for a case was defined as the 10 days preceding the onset of symptoms. If more than one person in a household was reported to have a laboratory-confirmed Campylobacter infection, only the first person in the household to acquire the illness (the index case) was included.

Control identification and interviews.

Identification of age- and county-matched control subjects and subsequent interviews were conducted at the WSU Social and Economic Sciences Research Center (SESRC). To identify potential controls, a random-digit-dialing sample frame was used. To be eligible for the study, potential controls needed to match the age of the related reported case according to the following age categories: <1 year, 1 to 2 years, 3 to 4 years, 5 to 12 years, 13 to 19 years, 20 to 39 years, 40 to 59 years, or ≥60 years. Potential controls also had to reside in the same county as the matched case. If potential control subjects or any of their household members had experienced either fever and abdominal pain or a diarrheal illness during the previous 2 weeks, then they were excluded from the study. To minimize recall bias, questions for control subjects referred to the 2 weeks preceding the interview. The control questions were designed to match the questions asked of case subjects except for questions concerning symptoms and the severity of illness, other than to confirm the absence of a recent history of diarrhea. Our goal was to interview at least 133 case subjects and 266 control subjects (two controls per case), which was estimated to give a statistical power of at least 0.80 to detect an odds ratio (OR) of 3.0 or higher (19).

Data analysis.

Data were analyzed using SAS/PC software (SAS Institute, Inc., Cary, NC). The associations between individual risk factors and disease were assessed using the Cochran Mantel-Haenszel chi-square option in PROC FREQ. A conditional multivariable logistic regression analysis of the matched case-control data was carried out using PROC LOGISTIC in SAS. The conditional logistic model specified the two matching variables (county and age category) in the STRATA statement. The model was built using a backward stepwise method with the requirement for a significance level of ≤0.1 to remain in the model. The population attributable risk (PAR) percentage, or etiological fraction, was calculated using the odds ratio (OR) as an estimate of the relative risk and the following formula: PAR% = [(PEcases) × (1 − 1/OR)], where PEcases is the proportion of cases exposed (20). Ninety-five percent confidence intervals (CIs) around the PAR percentage values were calculated according to the method of Natarajan et al. (21).

Bacterial isolates and genotyping. (i) Human isolates.

Clinical C. jejuni isolates with no identifiers were obtained from two major clinical laboratories in Yakima County and one in Whatcom County (both counties in Washington State). These were obtained from the same region and time period (2009 to 2010) as the epidemiological study cases, but we were not able to obtain clinical isolates linked to the interviewed case subjects.

(ii) Bovine isolates.

Field isolates obtained in 2007 to 2010 from Washington dairy cattle were utilized in this study (M. A. Davis and J. H. Harrison, unpublished data). Culture and isolation were carried out as described previously (22). To represent the diversity among cattle isolates, Campylobacter jejuni isolates from 13 different dairy farms that either were obtained from separate farms or in different years or had different sequence types (STs) were selected (see Table S1 in the supplemental material). Campylobacter isolates were stored at −80°C at the Field Disease Investigation Unit (FDIU) laboratory of the WSU College of Veterinary Medicine.

(iii) Culture and isolation from retail chicken.

Retail chicken isolates were obtained from chicken purchased from local grocery stores in the cities of Yakima in Yakima County and Bellingham in Whatcom County, Washington, in 2009 and 2010. Purchased chicken packages were transported on ice to the FDIU laboratory, where chicken pieces were homogenized in a laboratory blender (Seward Stomacher 80; Seward Laboratory Systems, Port Saint Lucie, FL) and blended in sterile buffered peptone water (Hardy Diagnostics, Santa Maria, CA). Samples (20 ml) from this mixture were transferred into 100 ml Bolton broth (Oxoid Inc., Ogdensburg, NY) and incubated aerobically at 42°C for 48 h. Samples were then filtered (0.45 μm) onto 5% sheep blood agar (SBA) (Hardy Diagnostics), followed by microaerophilic incubation at 37°C for 48 to 72 h. Suspect colonies were transferred to SBA plates and reincubated at 37°C for 24 h under microaerophilic conditions. Suspect colonies were stained with Victoria Blue stain (MP Biomedicals, Solon, OH) and examined under a microscope. Each isolated colony was stored at −80°C in 0.5 ml sterile 10% glycerol (Fisher Chemical, Waltham, MA) in 1.0% proteose peptone (Hardy Diagnostics) until further characterization. Table S1 in the supplemental material describes the provenance of the isolates used in this study.

Multilocus sequence typing.

Genomic DNA was extracted from Campylobacter isolates using a QIAamp DNA Micro kit (Qiagen, Germantown, MD), according to the manufacturer's instructions. Campylobacter jejuni species were differentiated from Campylobacter coli by using PCR primers targeting the Campylobacter lpxA gene and cycling conditions described previously (23). MLST was carried out with extracted genomic DNA as described previously (18, 24). Products were submitted to Functional Biosciences (Madison, WI) for sequencing. Sequences were analyzed using Vector NTI software (Life Technologies, Grand Island, NY), and the resulting sequence data for seven housekeeping genes were submitted to the PubMLST database website ( (25) for allele, sequence type (ST), and clonal complex assignments.


A total of 214 controls were identified that matched cases according to age group and county of residence in Washington State. Because controls were identified after the cases were reported and certain age groups were difficult to match, the number of controls in each age group did not equal 2 times the number of cases. Among 151 Whatcom County residents, 57/69 (82.6%) cases and 1/82 (1.2%) controls were lacking race/ethnicity data. Among 239 Yakima County residents, 5/107 (4.7%) cases and 2/132 (1.5%) controls were lacking race/ethnicity data. Hispanic ethnicity was strongly associated with case status (Table 1); therefore, to address the potential bias introduced by missing ethnicity data, an analysis restricted to Yakima County data was also conducted. The following exposure variables were not significantly associated with campylobacteriosis in the single-variable analysis and were not introduced into the multivariable analysis (chi-square P values of >0.10): eating at a restaurant, sharing a group meal, consuming unpasteurized dairy products, drinking from water sources other than bottled water, hunting, camping, visiting a farm or agricultural fair, having contact with camelids, goats, pigs, or sheep, caring for an animal at home, and touching an animal (Table 1).
Table 1
Table 1 Comparison of single variables for cases versus controls
VariableNo./total (%)Total no. (n = 390)Chi-square Pa
Cases (n = 176)Controls (n = 214)
County of residence    
    Whatcom69 (39.2)82 (38.3)151 
    Yakima107 (61.7)132 (60.8)239 
Male109/176 (61.9)94/214 (43.9)390<.01
    <1 yr9 (5.1)16 (7.5)25 
    1–2 yr25 (14.2)20 (9.4)45 
    3–4 yr13 (7.4)20 (9.4)33 
    5–12 yr17 (9.7)22 (10.3)39 
    13–19 yr15 (8.5)29 (13.6)44 
    20–39 yr38 (21.6)44 (20.6)82 
    40–59 yr35 (19.9)37 (17.3)72 
    ≥60 yr24 (13.6)26 (12.2)50 
    American Indian/Alaskan native2/114 (1.8)9/211 (4.3)  
    White35/114 (30.7)137/211 (64.9)  
    Asian0/1142/211 (1.0)  
    African American1/114 (0.9)3/211 (1.4)  
    Hispanic or Latino75/114 (65.8)37/211 (17.5)  
    Other or multiple1/114 (0.9)23/211 (10.9)325<0.01
Food consumption    
    Ate poultry110/165 (66.7)174/211 (82.5)376<0.01
    Ate undercooked poultryb3/44 (6.8)4/212 (1.9)2560.07
    Handled raw poultry28/176 (15.9)58/214 (27.1)390<0.01
    Ate restaurant food112/174 (64.4)145/214 (67.8)3880.48
    Ate shared group mealb25/103 (24.3)70/214 (32.7)2170.13
    Ate or drank unpasteurized dairy product16/169 (9.5)13/211 (6.2)3800.23
Source of drinking water    
    Bottled water70/176 (39.8)62/214 (29.0)3900.03
    Public water system96/176 (54.6)105/214 (49.1)3900.28
    Well water48/176 (27.3)45/214 (21.0)3900.15
    Untreated water27/168 (16.1)43/214 (20.1)3820.31
    Recreational water exposure22/171 (12.9)37/214 (17.3)3850.23
Outdoor activities    
    Gardening22/163 (13.5)74/214 (34.6)377<0.01
    Hiking6/160 (3.8)18/214 (8.4)3740.07
    Hunting1/176 (0.6)5/214 (2.3)3900.16
    Mountain biking0/1607/214 (3.3)3740.02
    Lawn mowing22/160 (13.8)50/214 (23.4)3740.02
    Sports1/160 (0.6)44/214 (20.6)374<0.01
    Yard work21/160 (13.1)93/214 (43.5)374<0.01
    Camping4/161 (2.5)13/214 (6.1)3750.10
Animal exhibits    
    Visited agricultural fair3/176 (1.7)8/214 (3.7)3900.23
    Visited farm15/176 (8.5)19/214 (8.9)3900.90
    Visited pet shop4/175 (2.3)17/214 (7.9)3890.01
Animal contact    
    Contact with alpacas or llamas2/170 (1.2)1/214 (0.5)3840.43
    Contact with cattle23/171 (13.5)14/214 (6.5)3850.02
    Contact with goats6/170 (3.5)7/214 (3.3)3840.89
    Contact with horses6/170 (3.5)20/214 (9.4)3840.02
    Contact with live poultry27/163 (16.6)17/214 (7.9)377<0.01
    Contact with pigs2/161 (1.2)1/214 (0.5)3750.41
    Contact with sheep0/1703/214 (1.4)3840.12
    Contact with pet animal104/176 (59.1)154/214 (72.0)390<0.01
    Contact with sick petc6/104 (5.8)3/154 (2.0)2580.10
    Contact with raw pet food or treats21/176 (11.9)45/214 (21.0)3900.02
    Any animal contact48/171 (28.1)159/214 (74.3)385<0.01
    Cared for animal at homed7/48 (14.6)22/159 (13.8)2070.90
    Cared for animal at workd12/48 (25.0)1/159 (0.6)207<0.01
    Touched animald6/48 (12.5)16/159 (10.1)2070.63
    Touched surfaced14/48 (29.2)9/159 (5.7)207<0.01
Farm exposure    
    Lived or worked on dairy farm20/176 (11.4)6/214 (2.8)390<0.01
    Lived or worked on nondairy farm26/176 (14.8)17/214 (7.9)3900.03
    Lived or worked on dairy farm or any cattle contact30/171 (17.5)18/214 (8.4)385<0.01
Cochran Mantel-Haenszel chi-square P value.
Missing data prevented inclusion of this factor in the multivariable analysis.
Among people reporting contact with a pet animal.
Among people reporting any animal contact.
Variables that were associated with campylobacteriosis in the single-variable analysis (chi-square P values of ≤0.10) were entered into the logistic regression model (Table 1). Stepwise elimination resulted in a multivariable model that included Hispanic ethnicity, living or working on a dairy farm, and having contact with a sick pet as independent risk factors for campylobacteriosis. Factors that were significantly associated with being a control subject (“protective” factors) in the multivariable model included eating poultry, handling raw poultry, and gardening. The estimated PAR percentages, or etiological fractions, for living or working on a dairy farm and having contact with a sick pet were 9.7% (95% CI, 2.5 to 16.1%) and 4.9% (95% CI, 0 to 10.5%), respectively (Table 2).
Table 2
Table 2 Results of multivariable conditional logistic regression analysis of all data from both counties (303 observations)a
Risk factorOdds ratio (95% CIb)PAR%c (95% CI)
Hispanic ethnicity6.4 (3.1–13.1) 
Ate poultry0.2 (0.1–0.5) 
Gardened0.3 (0.1–0.9) 
Handled raw poultry0.3 (0.1–0.9) 
Lived or worked on dairy farm6.7 (1.7–26.4)9.7 (2.5–16.1)
Contact with sick pet6.8 (0.8–55.3)4.9 (0.0–10.5)
Each variable in the multivariable model was independently associated with disease.
Wald CI.
PAR%, population-attributable risk percentage.
When the analysis was restricted to Yakima County, Hispanic ethnicity, any contact with cattle, contact with live poultry, living or working on a nondairy farm, and pet ownership were independently associated with having a case of campylobacteriosis, although the 95% confidence interval for contact with live poultry surrounded 1. Factors significantly associated with being a control subject in the Yakima County data included eating poultry, gardening, and having contact with horses. The estimated etiological fractions for contact with cattle, contact with live poultry, living or working on a nondairy farm, and pet ownership were 11.5% (95% CI, 1.5 to 21.0%), 12.9% (95% CI, 0 to 25.4%), 14.3% (95% CI, 0 to 27.0%), and 37.4% (95% CI, 4.6 to 60.0%), respectively (Table 3).
Table 3
Table 3 Results of multivariable conditional logistic regression analysis for Yakima County data (216 observations)a
Risk factorOdds ratio (95% CIb)PAR%c (95% CI)
Hispanic ethnicity9.3 (3.9–22.2) 
Ate poultry0.2 (0.1–0.6) 
Gardened0.1 (0.04–0.6) 
Contact with horses0.1 (0.01–0.6) 
Contact with cattle5.0 (1.3–19.5)11.5 (1.5–21.0)
Contact with live poultry3.1 (0.8–11.8)12.9 (0–25.4)
Lived or worked on nondairy farm3.0 (1.0–9.0)14.3 (0–27.0)
Pet ownership2.6 (1.1–6.3)37.4 (4.6–60.0)
Each variable was included in the multivariable model and was independently associated with disease.
Wald CI.
PAR%, population-attributable risk percentage.
Twenty-eight bovine, 65 human, and 27 retail chicken C. jejuni isolates were assayed using the 7-locus MLST protocol of the Oxford (United Kingdom)-based C. jejuni multilocus sequence typing website ( (25). The ST-21 clonal complex was the most commonly identified clonal complex in our data overall (31/120 isolates [25.8%]) and represented 50.0% of the bovine isolates (14/28 isolates) and 26.2% of the human isolates (17/65 isolates). The ST-45 clonal complex was isolated from all three sources and was the most frequent clonal complex found in retail chicken isolates (11/27 isolates [40.7%]). Among the eight isolates that were assigned to four sequence types (STs) that did not belong to any known clonal complex, none was isolated from both human and chicken sources, but one isolate, ST-934, was isolated from both bovine and human sources (Table 4).
Table 4
Table 4 MLST analysis of C. jejuni isolates in Washington State
Clonal complex (total)STNo. of isolates
Bovine fecesRetail chickenHuman stool
Unassigned922  3
 9341 2
 1698  1
 2514  1
ST-179 (n = 1)4564  1
ST-21 (n = 31)84 1
 21  1
 50  6
 9826 7
 2135  1
 4562  1
ST-22 (n = 5)221 4
ST-257 (n = 4)9294  
ST-283 (n = 9)267 81
ST-353 (n = 9)353  3
 452  2
 939  1
 2132  1
 4489  1
 4561  1
ST-354 (n = 2)354  1
 4559  1
ST-42 (n = 9)42  2
 4594 2
ST-443 (n = 7)51 51
 4525  1
ST-45 (n = 15)45 112
ST-460 (n = 2)460  1
 4563  1
ST-464 (n = 2)464  1
 4677  1
ST-48 (n = 12)48 25
 2521  1
ST-49 (n = 1)3  1
ST-508 (n = 4)132  2
ST-52 (n = 1)52 1 
ST-607 (n = 2)607  2
ST-61 (n = 3)12441 1
Total 282765


The findings of this study supported the hypothesis that living or working on a dairy farm or having contact with cattle is associated with increased risk for campylobacteriosis in Washington State counties that have high concentrations of dairy cattle. In the overall data, the estimated etiological fraction associated with dairy farm exposure was 9.7%, suggesting a significant contribution to the human disease burden (Table 2). When data from Yakima County were analyzed separately, any contact with cattle was significantly associated with disease and contributed an estimated 12% of disease (Table 3). Contact with a sick pet in the overall data and any pet ownership in the Yakima data were also associated with disease. Because pet ownership was prevalent among Yakima County cases (60.8%), the estimated proportion of disease contributed by this exposure was large (37.4%) (Table 3). Additional exposures significantly associated with disease in this study were contact with a sick pet (Table 2) and contact with live poultry (Table 3), suggesting that live animal contact is an important source of exposure to Campylobacter. The finding that pet ownership and contact with sick pets are risk factors for campylobacteriosis is consistent with previous epidemiological studies that reported contacts with pets, particularly puppies and kittens with diarrhea, as significant risk factors (10, 2628).
In Yakima County, living or working on a nondairy farm and having contact with live poultry were both associated with disease and with etiological fractions of 13 to 14%, although the confidence limits surrounding the latter did not exclude 0. Whether case subjects owned backyard poultry or worked in poultry production was not directly asked of the subjects, although five of 43 subjects who reported working on a nondairy farm listed the type of farm as poultry and four of those were case subjects. Other case-control studies either reported elevated risk associated with contact with farm animals without specifying whether that included live poultry (6, 26) or found elevated risk associated with farm animals, including live poultry (29). A study conducted in Australia found elevated risk associated with domestic chickens (10), and a study conducted in rural Michigan found that, of all exposures studied, poultry husbandry had the strongest association with campylobacteriosis (30). While the confidence intervals surrounding the PAR percentage estimates for live poultry contact, living or working on a nondairy farm, and pet ownership included 0, the upper bounds ranged from 25 to 60%. Therefore, these factors have the potential to be significant contributors to disease and merit further investigation.
Consumption of undercooked poultry was reported more frequently by case subjects than by control subjects but, because of missing data on this variable, it could not be analyzed. However, consumption of any poultry was significantly inversely associated with disease. In the Michigan study in which campylobacteriosis was found to be associated with poultry husbandry, poultry meat consumption was found to be protective (30). The authors speculated that rural populations may be more experienced at food preparation. We currently lack data to support or contradict that explanation. A large U.S. study of sporadic campylobacteriosis found that eating poultry or nonpoultry meat prepared at home was associated with decreased risk, while eating poultry or nonpoultry meat prepared at a restaurant was associated with an increased risk of campylobacteriosis (6). In the single-variable analysis, factors associated with sports and outdoor activities (sports, mountain biking, contact with horses, yard work, lawn mowing, and gardening) were significantly protective; among those factors, gardening and contact with horses remained in the multivariable models. These activities are likely to be associated with other unmeasured factors (e.g., income, level of education, occupation, or dietary preferences) that may directly protect against exposure to Campylobacter.
We found a consistently strong association between Hispanic ethnicity and campylobacteriosis, independent of other exposures. We did not ascertain education or income levels of case and control subjects and so we were unable to determine if the role of ethnicity in risk might have been explained by those two factors. A recent study of campylobacteriosis in children <3 years of age also found an association between Hispanic ethnicity and campylobacteriosis, independent of income level and education level (31). Our study did not identify specific food preferences or handling practices that may have been associated with Hispanic ethnicity and risk. Ethnicity-associated risk for enteric infections is another area of campylobacteriosis epidemiology that warrants further study.
A limitation of this case-control study was that the small sample size resulted in wide confidence limits and a potential inability to detect significant associations. The strength of the association between contact with live poultry and campylobacteriosis supports the idea that contact with live poultry is a risk factor (Table 3), but the 95% confidence limits surrounded 1. A potential source of bias existed because case interviews were conducted by LHJ personnel but control interviews were conducted by SESRC personnel. Thus, cases and controls were interviewed differently, although the questionnaire contents were the same. In addition, the identification of subjects using a random-digit-dialing sampling frame might have biased the data if cases were less likely to have landline telephones than controls; thus, controls might not represent the population from which cases arose. Cases with missing data were excluded from the analysis, and introduction of bias due to differential exposure ascertainment was likely due to differences in approaches by the two interviewer groups, rather than omission of specific questions. Nevertheless, our findings strongly indicate that contact with live food production animals is a significant risk factor for campylobacteriosis in these two counties, possibly more significant than food-borne transmission. Broader-based epidemiological data also support a contribution to human disease from the bovine reservoir. The strongest association (adjusted odds ratio, 21.0 [95% CI, 2.5 to 178]) reported from a FoodNet case-control analysis of sporadic campylobacteriosis was with farm animals among persons 2 to 12 years of age (6), and a FoodNet study on risk factors for campylobacteriosis in infants (27) indicated that visiting or living on a farm contributed significantly to the burden of disease (population attributable fraction, 20.1%). Poultry and bovine husbandry were found to be more significant than eating undercooked poultry or pork in the Michigan study (30), and a study in Norway found that the relative risk of occupational exposure to animals was second only to eating undercooked pork (29).
The results of MLST analysis of isolates obtained from cattle, retail chicken, and humans in the current study also supported a contribution to human infection from cattle. The limited number of retail chicken isolates prevents further inferences about their contribution to the human burden of disease, but the similarity of the MLST type distributions in cattle and human isolates is consistent with our epidemiological finding that living or working on a dairy farm is a significant risk factor. The epidemiological finding that exposure to live poultry is highly associated with human disease does not contradict our finding that relatively few human isolates matched retail chicken isolates by MLST. The retail chicken samples were purchased in the counties where the study took place, in order to detect the MLST types that would represent local food-borne exposures, although the chicken may have originated in poultry production facilities in geographically distant regions of the United States. Another reason that the distributions of MLST types may differ among human populations exposed via food-borne chicken versus live poultry is that Campylobacter MLST types have different rates of survival through the slaughter and postslaughter processes, resulting in a less diverse distribution in retail poultry than in preslaughter poultry. These results suggest that further studies should involve local live poultry sampling in addition to more-extensive sampling of retail poultry.
The most frequent MLST clonal complex in the human and bovine isolates was ST-21 complex (Table 4). This clonal complex is reported frequently for humans, cattle, and poultry in diverse geographic locations (3234). The most frequent ST within the ST-21 complex was ST-982, which included 7 human isolates and 6 bovine isolates. This is consistent with published sequence types of C. jejuni originating in the western United States, where ST-982 originated primarily from cattle or other ruminants ( ST-42 complex and ST-61 complex isolates from cattle in our study were also consistent with reports of cattle-origin C. jejuni from other regions of the United States (33). Multiple molecular genotyping studies have found that Campylobacter isolates from bovine and human hosts frequently have the same genotype by a variety of methods, including pulsed-field gel electrophoresis (PFGE), randomly amplified polymorphic DNA (RAPD) typing, flagellin gene typing, and MLST (3540). A New Zealand study conducted in a dairy farming area that included comparisons between genotypes of Campylobacter from humans, cattle, and poultry found that cattle-origin isolates had a distribution of subtypes more similar to that of human-origin isolates than poultry-origin isolates (36, 41). A study in Denmark using multiple subtyping methods found that subtype sharing was approximately as frequent between poultry and humans as between cattle and humans (39), and a comparison between human, poultry, and raw milk isolates in Québec, Canada, found that three of 10 clonal complexes included isolates from human, raw milk, or water sources (42). Thus, while the relative contribution of cattle-origin Campylobacter to human infections seems to vary regionally, overall a significant proportion of human infection is attributable to ruminant sources. In this case-control study, however, the burden of campylobacteriosis was not completely explained by the risk factors that we investigated. To accomplish this, a more comprehensive study involving complex and exhaustive data-gathering, including obtaining clinical isolates that are linked to epidemiological data, needs to be conducted. This study contributes to the evidence supporting cattle exposure as a significant risk factor for campylobacteriosis. Exposure to live cattle may deserve a higher profile in the informational material provided by public health agencies.


This work was funded by the National Institute of Allergy and Infectious Diseases, National Institutes of Health, under contract N01-AI-30055.
This publication made use of the Campylobacter multilocus sequence typing website ( developed by Keith Jolley and sited at the University of Oxford (43). The development of this site has been funded by the Wellcome Trust.

Supplemental Material

File (zjm999092943sd2.xlsx)
File (zjm999092943so1.pdf)
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.


Butzler JP. 2004. Campylobacter, from obscurity to celebrity. Clin. Microbiol. Infect. 10:868–876.
Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, Widdowson MA, Roy SL, Jones JL, and Griffin PM. 2011. Foodborne illness acquired in the United States: major pathogens. Emerg. Infect. Dis. 17:7–15.
Humphrey T, O'Brien S, and Madsen M. 2007. Campylobacters as zoonotic pathogens: a food production perspective. Int. J. Food Microbiol. 117:237–257.
Altekruse SF, Stern NJ, Fields PI, and Swerdlow DL. 1999. Campylobacter jejuni: an emerging foodborne pathogen. Emerg. Infect. Dis. 5:28–35.
Denno DM, Keene WE, Hutter CM, Koepsell JK, Patnode M, Flodin-Hursh D, Stewart LK, Duchin JS, Rasmussen L, Jones R, and Tarr PI. 2009. Tri-county comprehensive assessment of risk factors for sporadic reportable bacterial enteric infection in children. J. Infect. Dis. 199:467–476.
Friedman CR, Hoekstra RM, Samuel M, Marcus R, Bender J, Shiferaw B, Reddy S, Ahuja SD, Helfrick DL, Hardnett F, Carter M, Anderson B, and Tauxe RV. 2004. Risk factors for sporadic Campylobacter infection in the United States: a case-control study in FoodNet sites. Clin. Infect. Dis. 38(Suppl 3):S285–S296.
Gormley FJ, Macrae M, Forbes KJ, Ogden ID, Dallas JF, and Strachan NJ. 2008. Has retail chicken played a role in the decline of human campylobacteriosis? Appl. Environ. Microbiol. 74:383–390.
Harris NV, Weiss NS, and Nolan CM. 1986. The role of poultry and meats in the etiology of Campylobacter jejuni/coli enteritis. Am. J. Public Health 76:407–411.
Mughini Gras L, Smid JH, Wagenaar JA, de Boer AG, Havelaar AH, Friesema IH, French NP, Busani L, and van Pelt W. 2012. Risk factors for campylobacteriosis of chicken, ruminant, and environmental origin: a combined case-control and source attribution analysis. PLoS One 7:e42599.
Stafford RJ, Schluter PJ, Wilson AJ, Kirk MD, Hall G, Unicomb L, and OzFoodNet Working Group. 2008. Population-attributable risk estimates for risk factors associated with Campylobacter infection, Australia. Emerg. Infect. Dis. 14:895–901.
Wingstrand A, Neimann J, Engberg J, Nielsen EM, Gerner-Smidt P, Wegener HC, and Molbak K. 2006. Fresh chicken as main risk factor for campylobacteriosis, Denmark. Emerg. Infect. Dis. 12:280–285.
Buettner S, Wieland B, Staerk KD, and Regula G. 2010. Risk attribution of Campylobacter infection by age group using exposure modelling. Epidemiol. Infect. 138:1748–1761.
Green CG, Krause DO, and Wylie JL. 2006. Spatial analysis of Campylobacter infection in the Canadian province of Manitoba. Int. J. Health Geogr. 5:2.
Nygard K, Andersson Y, Rottingen JA, Svensson A, Lindback J, Kistemann T, and Giesecke J. 2004. Association between environmental risk factors and Campylobacter infections in Sweden. Epidemiol. Infect. 132:317–325.
Thompson JS, Cahoon FE, and Hodge DS. 1986. Rate of Campylobacter spp. isolation in three regions of Ontario, Canada, from 1978 to 1985. J. Clin. Microbiol. 24:876–878.
Englen MD, Hill AE, Dargatz DA, Ladely SR, and Fedorka-Cray PJ. 2007. Prevalence and antimicrobial resistance of Campylobacter in US dairy cattle. J. Appl. Microbiol. 102:1570–1577.
Besser TE, Lejeune JT, Rice DH, Berg J, Stilborn RP, Kaya K, Bae W, and Hancock DD. 2005. Increasing prevalence of Campylobacter jejuni in feedlot cattle through the feeding period. Appl. Environ. Microbiol. 71:5752–5758.
Dingle KE, Colles FM, Wareing DR, Ure R, Fox AJ, Bolton FE, Bootsma HJ, Willems RJ, Urwin R, and Maiden MC. 2001. Multilocus sequence typing system for Campylobacter jejuni. J. Clin. Microbiol. 39:14–23.
Dean AG, Dean JA, Coulombier D, Brendel KA, Smith DC, Burton AH, Dicker RC, Sullivan K, Fagan RF, and Arner TC. 1994. Epi Info, version 6: a word processing, database, and statistics programs for epidemiology on microcomputers. Centers for Disease Control and Prevention, Atlanta, GA.
Coughlin SS, Benichou J, and Weed DL. 1994. Attributable risk estimation in case-control studies. Epidemiol. Rev. 16:51–64.
Natarajan S, Lipsitz SR, and Rimm E. 2007. A simple method of determining confidence intervals for population attributable risk from complex surveys. Stat. Med. 26:3229–3239.
Bae W, Kaya KN, Hancock DD, Call DR, Park YH, and Besser TE. 2005. Prevalence and antimicrobial resistance of thermophilic Campylobacter spp. from cattle farms in Washington State. Appl. Environ. Microbiol. 71:169–174.
Werno AM, Klena JD, Shaw GM, and Murdoch DR. 2002. Fatal case of Campylobacter lari prosthetic joint infection and bacteremia in an immunocompetent patient. J. Clin. Microbiol. 40:1053–1055.
Dingle KE, Colles FM, Falush D, and Maiden MC. 2005. Sequence typing and comparison of population biology of Campylobacter coli and Campylobacter jejuni. J. Clin. Microbiol. 43:340–347.
Jolley KA, Chan MS, and Maiden MC. 2004. mlstdbNet: distributed multi-locus sequence typing (MLST) databases. BMC Bioinformatics 5:86.
Domingues AR, Pires SM, Halasa T, and Hald T. 2012. Source attribution of human campylobacteriosis using a meta-analysis of case-control studies of sporadic infections. Epidemiol. Infect. 140:970–981.
Fullerton KE, Ingram LA, Jones TF, Anderson BJ, McCarthy PV, Hurd S, Shiferaw B, Vugia D, Haubert N, Hayes T, Wedel S, Scallan E, Henao O, and Angulo FJ. 2007. Sporadic Campylobacter infection in infants: a population-based surveillance case-control study. Pediatr. Infect. Dis. J. 26:19–24.
Marcus R. 2008. New information about pediatric foodborne infections: the view from FoodNet. Curr. Opin. Pediatr. 20:79–84.
Kapperud G, Espeland G, Wahl E, Walde A, Herikstad H, Gustavsen S, Tveit I, Natas O, Bevanger L, and Digranes A. 2003. Factors associated with increased and decreased risk of Campylobacter infection: a prospective case-control study in Norway. Am. J. Epidemiol. 158:234–242.
Potter RC, Kaneene JB, and Hall WN. 2003. Risk factors for sporadic Campylobacter jejuni infections in rural Michigan: a prospective case-control study. Am. J. Public Health 93:2118–2123.
Patrick ME, Mahon BE, Zansky SM, Hurd S, and Scallan E. 2010. Riding in shopping carts and exposure to raw meat and poultry products: prevalence of, and factors associated with, this risk factor for Salmonella and Campylobacter infection in children younger than 3 years. J. Food Prot. 73:1097–1100.
Jorgensen F, Ellis-Iversen J, Rushton S, Bull SA, Harris SA, Bryan SJ, Gonzalez A, and Humphrey TJ. 2011. Influence of season and geography on Campylobacter jejuni and C. coli subtypes in housed broiler flocks reared in Great Britain. Appl. Environ. Microbiol. 77:3741–3748.
Sanad YM, Kassem II, Abley M, Gebreyes W, LeJeune JT, and Rajashekara G. 2011. Genotypic and phenotypic properties of cattle-associated Campylobacter and their implications to public health in the USA. PLoS One 6:e25778.
Zhang M, Gu Y, He L, Ran L, Xia S, Han X, Li H, Zhou H, Cui Z, and Zhang J. 2010. Molecular typing and antimicrobial susceptibility profiles of Campylobacter jejuni isolates from north China. J. Med. Microbiol. 59:1171–1177.
Fitzgerald C, Stanley K, Andrew S, and Jones K. 2001. Use of pulsed-field gel electrophoresis and flagellin gene typing in identifying clonal groups of Campylobacter jejuni and Campylobacter coli in farm and clinical environments. Appl. Environ. Microbiol. 67:1429–1436.
Garrett N, Devane ML, Hudson JA, Nicol C, Ball A, Klena JD, Scholes P, Baker MG, Gilpin BJ, and Savill MG. 2007. Statistical comparison of Campylobacter jejuni subtypes from human cases and environmental sources. J. Appl. Microbiol. 103:2113–2121.
Gilpin BJ, Thorrold B, Scholes P, Longhurst RD, Devane M, Nicol C, Walker S, Robson B, and Savill M. 2008. Comparison of Campylobacter jejuni genotypes from dairy cattle and human sources from the Matamata-Piako District of New Zealand. J. Appl. Microbiol. 105:1354–1360.
Keller J, Wieland B, Wittwer M, Stephan R, and Perreten V. 2007. Distribution and genetic variability among Campylobacter spp. isolates from different animal species and humans in Switzerland. Zoonoses Public Health 54:2–7.
Nielsen EM, Engberg J, Fussing V, Petersen L, Brogren CH, and On SL. 2000. Evaluation of phenotypic and genotypic methods for subtyping Campylobacter jejuni isolates from humans, poultry, and cattle. J. Clin. Microbiol. 38:3800–3810.
Schouls LM, Reulen S, Duim B, Wagenaar JA, Willems RJ, Dingle KE, Colles FM, and Van Embden JD. 2003. Comparative genotyping of Campylobacter jejuni by amplified fragment length polymorphism, multilocus sequence typing, and short repeat sequencing: strain diversity, host range, and recombination. J. Clin. Microbiol. 41:15–26.
Mullner P, Shadbolt T, Collins-Emerson JM, Midwinter AC, Spencer SE, Marshall J, Carter PE, Campbell DM, Wilson DJ, Hathaway S, Pirie R, and French NP. 2010. Molecular and spatial epidemiology of human campylobacteriosis: source association and genotype-related risk factors. Epidemiol. Infect. 138:1372–1383.
Levesque S, Frost E, Arbeit RD, and Michaud S. 2008. Multilocus sequence typing of Campylobacter jejuni isolates from humans, chickens, raw milk, and environmental water in Quebec, Canada. J. Clin. Microbiol. 46:3404–3411.
Jolley KA and Maiden MC. 2010. BIGSdb: scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics 11:595.

Information & Contributors


Published In

cover image Journal of Clinical Microbiology
Journal of Clinical Microbiology
Volume 51Number 12December 2013
Pages: 3921 - 3927
PubMed: 24025908


Received: 3 June 2013
Returned for modification: 12 July 2013
Accepted: 14 August 2013
Published online: 21 December 2020


Request permissions for this article.



Margaret A. Davis
Department of Veterinary Microbiology and Pathology, Washington State University, Pullman, Washington, USA
Paul G. Allen School for Global Animal Health, Washington State University, Pullman, Washington, USA
Danna L. Moore
Department of Veterinary Microbiology and Pathology, Washington State University, Pullman, Washington, USA
Social and Economic Sciences Research Center, Washington State University, Pullman, Washington, USA
Katherine N. K. Baker
Department of Veterinary Microbiology and Pathology, Washington State University, Pullman, Washington, USA
Nigel P. French
Massey University, Palmerston North, New Zealand
Marianne Patnode
Yakima Health District, Union Gap, Washington, USA
Joni Hensley
Whatcom County Health Department, Bellingham, Washington, USA
Kathryn MacDonald
Washington State Department of Health, Shoreline, Washington, USA
Thomas E. Besser
Department of Veterinary Microbiology and Pathology, Washington State University, Pullman, Washington, USA


Address correspondence to Margaret A. Davis, [email protected].

Metrics & Citations


Note: There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.

Citation counts come from the Crossref Cited by service.


If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. For an editable text file, please select Medlars format which will download as a .txt file. Simply select your manager software from the list below and click Download.

View Options

Figures and Media






Share the article link

Share with email

Email a colleague

Share on social media

American Society for Microbiology ("ASM") is committed to maintaining your confidence and trust with respect to the information we collect from you on websites owned and operated by ASM ("ASM Web Sites") and other sources. This Privacy Policy sets forth the information we collect about you, how we use this information and the choices you have about how we use such information.
FIND OUT MORE about the privacy policy