The study included all culture-positive Campylobacter infections from May 1, 2001, through June 10, 2002, from two counties with a catchment area of approximately 1.1 million persons (approximately one fifth of the Danish population). The county of Copenhagen, a metropolitan residential area, has a population of 619,000, and the county of Funen, an island with both urban and rural areas, has a population of 472,000. Because of inconsistencies in the patient enrolment from the county of Funen during May and June 2002, patients from this county who were infected after April 31, 2002, were excluded. Epidemiologic data were captured on self-completed standardized patient questionnaires forwarded by the Danish Zoonosis Centre. Patients were interviewed about clinical symptoms, travel history, and exposures to food, water, and animals in the 7 days before illness onset. Completed questionnaires were returned to the Danish Zoonosis Centre and linked with microbiologic data.

All isolates included in the study were tested for resistance to nalidixic acid and erythromycin. All human isolates from the county of Copenhagen and isolates obtained from retail food products and broiler chickens were screened by a disk-diffusion test using Oxoid disks on 5% blood agar plates. On the basis of zone sizes, this method grouped the isolates in two well-separated populations of susceptible and resistant isolates with both antimicrobial drugs. The few isolates that fell between these populations were retested by using the standardized tablet diffusion and E-test procedures described previously (5), with the modifications that resistance to nalidixic acid was defined as an MIC >64 mg/L for the MIC method and a zone size <27 mm for the tablet method. All human isolates from the county of Funen were tested by the standardized tablet diffusion test with both antimicrobial drugs. Finally, all isolates found to be resistant and sensitive to nalidixic acid from our case-comparison study were retested by both the standardized tablet diffusion and E-test procedure.

In the second half of the study period (from December 1, 2001, to June 10, 2002), characteristics of patients with quinolone-resistant and quinolone-sensitive C. jejuni infections were compared. Each patient with a resistant isolate was matched with two randomly selected patients with sensitive isolates. Patients were matched on date of specimen collection.

Patients answered, either by phone or by mail, a short additional questionnaire, which included questions about use of fluoroquinolones the month before onset of illness, use of fluoroquinolones after onset of illness but before specimen collection, use of antimicrobial drugs after specimen collection, and other clinical information. When patients could not answer questions about exposure to fluoroquinolones before fecal sampling, the information was gathered from their healthcare providers.

As part of a national surveillance program, food samples from retail outlet stores were analyzed for Campylobacter at the regional food safety authorities, according to accredited methods of the Nordic Committee on Food Analysis (6). The samples were taken from whole poultry and different cuts of poultry (frozen and fresh), including chicken and turkey. Samples of pork and beef products were also analyzed. Imported as well as domestic food products were sampled.

As part of a national surveillance program for Campylobacter in broiler chickens, chickens were sampled at slaughter and analyzed for Campylobacter. In this study, isolates from broiler chicken farms located in Funen County were included (one isolate per flock). Copenhagen County does not have any broiler chicken farms.

One isolate from the primary isolation on modified charcoal cefoperazone deoxycholate agar (mCCDA) from each patient, as well as one isolate from each retail food sample and broiler chicken fecal sample were characterized at Statens Serum Institut and the Danish Veterinary Institute. Speciation, serotyping, and RiboPrinting (automated ribotyping) were undertaken as previously described (7,8), with the following modifications for the RiboPrinting method: 1-µL eye needle was filled with bacterial culture and dissolved in 100 µL sample buffer. Ten microliters of 10 g/L lysozyme was added, and the solution was left at 37°C for 10 min. From this solution, 30 µL was transferred to a sample carrier for heat treatment. The RiboPrinter was run according to the SEC protocol at 37°C for 2 h.

Conditional logistic regression was applied to calculate a matched odds ratio for the exposure variables. Variables, which reached a significance level of <0.15 in the univariate analysis of the case comparison study, were selected for the multiple logistic regression analysis. Stepwise conditional logistic regression with a backward elimination procedure was conducted to obtain a reduced model. Variables with a p value <0.05 were kept in the model. All excluded variables were retested in the final model. The statistical software SAS Release v.8.00 (SAS Institute Inc., Cary, NC) and Epi Info version 6.04d (Centers for Disease Control and Prevention, Atlanta, GA) were used to analyze the data.

Of 975 culture-confirmed Campylobacter infections in the study, 177 (18.2%) were infected with a quinolone-resistant isolate, whereas 3 (0.3%) isolates were erythromycin-resistant. Linked microbiologic and epidemiologic data were obtained from 678 (69.5%) patients. In total, 152 (22.4%) patients had been outside Denmark within 1 week before illness, whereas 526 (77.6%) were domestically acquired infections. The three erythromycin-resistant isolates were all C. coli, two of them were also quinolone-resistant, and these were both isolated from travelers returning to Denmark from Spain and Portugal, respectively.

Quinolone resistance was significantly associated with the origin of infection: 76 (50.0%) of 152 infections among travelers returning to Denmark were quinolone-resistant whereas 52 (9.9%) of 526 of domestically infected patients were infected with a quinolone-resistant strain (p < 0.001) (Table 1). For both C. coli and C. jejuni, a significantly higher proportion of quinolone-resistant infections was found among patients who had been abroad in the week before onset of illness than among patients with domestically acquired infections (risk ratio [RR] 9.3, 95% confidence interval [CI] 1.4 to 63.8, p = 0.004 and RR 4.9, CI 3.6 to 6.7, p < 0.001). A higher proportion of C. coli than C. jejuni infections were acquired abroad (48.3% compared with 21.3% of C. jejuni).

Foreign travel was associated with different prevalences of quinolone resistance, depending on destination (Table 2). No travelers returning from other Scandinavian countries hosted quinolone-resistant Campylobacter isolates, whereas travel to a number of regions and subregions, including southern Europe and Southeast Asia, was significantly associated with a high proportion of quinolone-resistant infections.

C. jejuni infections and C. coli infections did not differ in severity, as assessed by frequency of diarrhea, blood in stool, abdominal pain, fever, vomiting, mean duration of illness, or admission to hospitals. However, the mean duration of illness was longer for the 86 patients with quinolone-resistant C. jejuni infections and a known duration of illness (median 13.2 days) than for the 381 patients with quinolone-sensitive C. jejuni infections and a known duration of illness (median 10.3 days, p = 0.001). The association with extended length of illness was independent of foreign travel. For domestically acquired infections, the mean duration of illness was 12.4 and 10.4 days for quinolone-resistant and quinolone-sensitive infections, respectively. For comparison, the mean duration of illness for travel-associated infections was 13.9 and 10.3 days for quinolone-resistant and quinolone-sensitive infections, respectively. For C. coli, no difference in mean duration of illness was observed between quinolone-resistant and quinolone-sensitive infections.

From December 1, 2001, to June 10, 2002, 42 patients were infected by quinolone-resistant C. jejuni isolates, and these patients were matched with 84 patients with quinolone-sensitive isolates. No patients were connected on epidemiologic grounds of a recognized outbreak. The patients with quinolone-resistant isolates had a mean age of 33 years (interquartile range 20–45 years), and a male-to-female ratio of 1:1.2. For comparison, patients with quinolone-resistant isolates in the larger study, which were not included in the case-comparison study, had a mean age of 31 years (interquartile range 20–45 years), and a male-to-female ratio of 1:1.3. No strains changed susceptibility category after being retested. However, one case-patient was shown to be co-infected with two C. jejuni strains with identical serotype and RiboGroup, but with different susceptibility patterns, i.e., one strain was clearly sensitive (MIC of quinolone = 2 mg/L), whereas the MIC of quinolone for the other one was 128 to >256 mg/L on multiple repeated testings. Subsequent nucleotide sequence analysis indicated a normal consensus sequence in the former and a Thr-86 to Ala-86 mutation and three silent mutations in gyrA in the latter. The patient had not been exposed to fluoroquinolones before stool specimen collection.

Risk factors for a quinolone-resistant C. jejuni infection identified in the univariate analysis and in the multiple logistic regression analysis are presented in Table 3. According to the multiple logistic regression analysis, the only exposures independently associated with an increased risk for quinolone-resistant C. jejuni infection were foreign travel (OR = 16.81), eating fresh poultry other than chicken and turkey (OR = 19.10), and swimming (OR = 5.01). Eating fresh chicken (of presumably Danish origin) was associated with a decreased risk (OR = 0.04). Age group did not affect the findings (younger or older than 15 years of age) in either the univariate or the multiple logistic regression analysis.

The case-comparison study identified 12 quinolone-resistant cases that were domestically acquired. However, to determine the sources of infection for the domestically acquired quinolone-resistant infections, an unmatched subanalysis on domestically acquired infections (quinolone-resistant versus quinolone-sensitive) was performed. Infections treated with fluoroquinolones before specimen collection were excluded. In this model, the parameter estimates did not change substantially from the primary model, but because of the lower sample size, the confidence intervals increased, and only eating fresh poultry other than chicken and turkey had p value <0.05. In 10 (11.9%) of 84 domestically acquired infections, patients reported eating fresh poultry other than chicken and turkey compared with 4 (9.5%) of 42 infections in persons with travel-related infections.

Overall, we found information on antimicrobial drugs received by 122 of 126 patients. Forty patients (32.8%) were treated with antimicrobial agents for their campylobacteriosis; of these, 33 patients (27%) received a fluoroquinolone, 6 patients (4.9%) received a macrolide, and 1 patient (1%) received both a fluoroquinolone and a macrolide for the C. jejuni infection.

The human isolates were included in a database and compared with 180 Campylobacter isolates obtained from retail food products (chicken [n = 139], turkey [n = 39], and pork [n = 2]) and 49 isolates from broiler chicken fecal samples obtained from the same geographic area and time period as the human isolates. Most (63%) food isolates were from Danish-bred food animals; the remaining isolates were from imported food from France (n = 48), Italy (n = 7), and the United Kingdom (n = 9). The origin of three chicken isolates was unknown. Of 180 isolates obtained from food products of both domestic and foreign origin, 153 (85%) isolates and 27 (15%) isolates were C. jejuni and C. coli, respectively (Table 1). Thirteen (8.5%) of 153 C. jejuni isolates and 8 (29.6%) of 27 C. coli isolates were resistant to nalidixic acid. Three (2.0%) of 153 C. jejuni isolates and 5 (18.5%) of 27 C. coli isolates were resistant to erythromycin. Two isolates (one C. jejuni and one C. coli) from domestic chicken products were resistant to both antimicrobial agents. A subanalysis of resistance status by origin of 139 retail chicken products (domestic versus imported) showed that 7 (8.0%) of 87 C. jejuni isolates and three (60%) of five C. coli isolates from domestic raised chicken products were resistant to nalidixic acid. Of isolates from imported chicken products, 5 (14.7%) (3 isolates from France and 2 isolates from the United Kingdom) of 34 C. jejuni isolates and 1 (10%) (from France) of 10 C. coli isolates were resistant to nalidixic acid.

Of 49 isolates from broiler chicken fecal samples, 39 (79.6%) were C. jejuni and 10 (20.4%) were C. coli (Table 1). Two isolates (4.1%) (both C. jejuni) were nalidixic acid–resistant; one was also erythromycin-resistant. Five (10.2%) isolates (four C. jejuni, one C. coli) were erythromycin-resistant.

We found 133 combinations of serotypes and RiboGroups (hereafter subtypes) among 496 typed isolates (10 isolates were not tested or nontypeable) from domestically acquired C. jejuni infections (Table 4). Eighteen (13.5%) subtypes were identified exclusively among quinolone-resistant isolates, 102 (76.7%), exclusively among quinolone-sensitive isolates, and 13 (9.8%) among both resistant and sensitive isolates.

Five of 11 subtypes of quinolone-resistant C. jejuni found among isolates from retail food products, broiler chickens, or both were also found among quinolone-resistant domestically acquired C. jejuni isolates from humans, and 34 of 88 subtypes of quinolone-sensitive C. jejuni found among isolates from retail food products, broiler chickens, or both were also found among quinolone-sensitive domestically acquired C. jejuni isolates from humans.

Patients with domestically acquired quinolone-sensitive C. jejuni infections were more likely to have a C. jejuni subtype that was also identified among retail food products and broiler chickens than were patients with domestically acquired quinolone-resistant infections (270 of 444 vs. 15 of 51, RR = 2.07, CI 1.34 to 3.18, p < 0.001).