A retrospective analysis, including the period from January 1, 2010 through December 31, 2011, identified 43 laboratory-confirmed cases of C. coli infections reported to the Montreal Public Health Department; among them, 40 cases with antimicrobial drug susceptibility results were further analyzed. Telephone interviews with the case-patients were conducted by using a standardized questionnaire pertaining to symptomatology of the illness, treatment, exposures, sexual orientation (including practices), and HIV status. The questionnaire was mailed to persons who could not be contacted by phone. Hospital charts for 9 or the 10 outbreak case-patients were reviewed retrospectively.

Statistical analyses, using Fisher exact test to calculate the possibilities, were conducted to test for differences in characteristics between case-patients infected with the outbreak etiologic agent, C. coli pulsovar 1, and those infected with nonoutbreak C. coli. In estimating the odds ratio from a 2 × 2 table that included a zero cell, 0.5 was added to the count in each cell. CIs were calculated by using Miettinen’s test-based method. Statistical analyses were conducted using SPSS software (http://www-01.ibm.com/software/analytics/spss/products/statistics/).

Phenotypic identification of Campylobacter isolates at the genus and species levels was confirmed by cpn60 gene sequencing at Laboratoire de Santé Publique du Québec. C. coli strains were identified by direct sequencing of PCR-amplified partial cpn60 sequences as described by Hill et al. (6). DNA sequences were determined with an ABI 3100 sequencer using a BigDye sequencing kit (Applied Biosystems, Foster City, CA, USA). The sequences were subjected to a BLAST analysis and aligned with the ClustalW program. Phylogenetic analysis was performed using the Lasergene software V6.1 (DNAstar, Madison, WI, USA).

Genetic relatedness was investigated by using pulsed-field gel electrophoresis with SmaI according to PulseNet Canada procedures. Salmonella enterica serotype Braenderup strain H9812 was used as the marker size in each gel (7). For analysis, band position tolerance and optimization values of 1% were used. Similarity coefficient was obtained with the unweighted pair-group method with arithmetic averages. For strains exhibiting similar patterns with SmaI, a second enzyme (KpnI) was used to confirm their pulsed-field gel electrophoresis pattern similarity. The PulseNet Canada SmaI and KpnI pattern designations for the C. coli pulsovar 1 isolate are CASAI.0160 and CAKNI.0078, respectively.

Antimicrobial drug susceptibility testing was determined by using the disk diffusion method for erythromycin, tetracycline, and ciprofloxacin (8) and the Etest (AB Biodisk, Solna, Sweden) method for all 12 agents tested (3). β-lactamase susceptibility was determined as reported (9).

Figure 1

Figure 1. . . Number of cases of Campylobacter coli infection reported to Montreal Public Health, Quebec, Canada, 2010–2011. PFGE, pulsed-field gel electrophoresis.

Figure 2

Figure 2. . . Pulsed-field gel electrophoresis (PFGE) patterns of erythromycin-susceptible, tetracycline- and ciprofloxacin-resistant, Campylobacter coli for SmaI (19 isolates) and KpnI (10 isolates) enzymes, Montreal, Quebec, Canada, 2010–2011. Scale bar indicates percent...

From September 2010 through November 2011, in Montreal, 10 men, 26–57 years of age, were found to be infected with an erythromycin-susceptible, tetracycline- and ciprofloxacin-resistant C. coli pulsovar 1; these men were defined as the outbreak-associated case-patients (Figure 1). An additional 5 women and 4 men were infected with an erythromycin-susceptible, tetracycline- and ciprofloxacin-resistant C. coli strain; however, 9 different pulsovars were involved (Figure 2). Microbiology laboratories at Centre Hospitalier de l’Université de Montréal documented 9 of the 10 outbreak cases, but did not isolate erythromycin-susceptible, tetracycline- and ciprofloxacin-resistant C. coli from December 2011 through November 2012.

Compared with the 30 nonoutbreak case-patients for whom suceptibility results were available, the 10 outbreak case-patients were more likely to be male (p = 0.010), to be 20–59 years of age (p = 0.010), to be men who have sex with men (MSM) (p = 0.0001), to be HIV positive (p = 0.001), and to have had sexual relations within 2 weeks of the beginning of gastrointestinal symptoms (p = 0.017) (Table 1). Of the 8 HIV-positive patients, the CD4 cell count was 210 × 10⁶/L for 1 man and 440–1,150 × 10⁶/L for the 6 other patients, and the HIV viral load was 68 copies/mL for 1 patient and <40 copies/mL for the 5 other patients for whom these data were known.

Seven men (70%), all MSM, lived in surrounding neighborhoods of Montreal’s so-called Gay Village. The antimicrobial drug treatment regimen was known for 1 outbreak case-patient who received oral azithromycin. In the previous 15 years, 1 HIV-positive man (who also had sex with men) and had C. coli pulsovar 1 and S. flexneri in fecal specimens, exhibited 7 other sexually transmitted diseases. Among the 10 outbreak case-patients, 2 HIV-positive MSM were simultaneously infected with an S. flexneri isolate, and 1 of these 2 patients experienced C. coli septicemia.

The following data were reported for 10 outbreak C. coli and 30 nonoutbreak C. coli case-patients, respectively: diarrhea, 100% (6/6) and 88% (23/26); abdominal cramps, 60% (3/5) and 83% (19/23); blood in stool specimen, 20% (1/5) and 30% (6/20); fever, 60% (3/5) and 61% (14/23); and hospitalization, 17% (1/6) and 19% (5/26) (p>0.05 for all data). Exposures to potential sources of infection did not differ between outbreak and nonoutbreak case-patients. All patients reported having consumed meat, dairy products, tap water or commercially bottled only, and no nonchlorinated water. Exposures to animals, farms, and other persons with known cases of diarrhea were rarely reported. Travel history outside of the island of Montreal in the 2 weeks before symptom onset was documented in none (0/6) of C. coli pulsovar 1 case-patients and in 48% (12/25) of other C. coli case-patients (p = 0.059); all 12 had traveled outside Canada.

The 10 outbreak isolates were susceptible to erythromycin, azithromycin, ampicillin, gentamicin, imipenem, clindamycin, chloramphenicol, and tigecycline and were β-lactamase negative. All 10 were resistant to ciprofloxacin, nalidixic acid, tetracycline, and cefotaxime (Table 2).

Epidemiologic and molecular data confirmed a cluster of erythromycin-susceptible, tetracycline- and ciprofloxacin-resistant, C. coli pulsovar 1 infections in MSM in Montreal, Quebec, Canada, during September 2010–November 2011. The epidemiologic data reported in Table 1, the 14-month outbreak duration, the simultaneous S. flexneri infection in 2 HIV-positive MSM, and the absence of any reported common food exposure suggest a sexually transmitted enteric infection. A cluster of erythromycin- and ciprofloxacin-resistant, tetracycline-susceptible C. jejuni subsp. jejuni infections from 1999 through 2001 (13) and 7 clusters of Shigella spp. infections from 1999 through 2011 (14,15; unpub. data), which were sexually transmitted, have been documented in MSM in Montreal and surrounding neighborhoods. Among MSM, Shigella spp. infection is, in most cases, sexually transmitted (15).

C. coli infection clusters are infrequently reported (4,5). Campylobacter should be identified to species level by phenotypic and, if needed, by molecular characterization. Association of cluster cases with the correct Campylobacter species is the first step of suspecting an outbreak and can lead to improved outbreak detection. Antimicrobial drug susceptibility testing, at least to erythromycin and ciprofloxacin, is recommended for every isolate (10). The erythromycin, ciprofloxacin, and tetracycline susceptibilities were epidemiologic markers in the Campylobacter spp. clusters documented in Montreal (present study; 13). Nine different pulsovars were documented in 9 nonoutbreak case-patients, indicating a high heterogeneity of C. coli.

If necessary, the first-choice antimicrobial drug treatment for patients infected with C. coli pulsovar 1 would be a macrolide as it is for C. jejuni and C. coli enteric infections because of increasing fluoroquinolone resistance in these bacteria (1–3). HIV-positive or AIDS patients may have a higher incidence of Campylobacter infections with more septicemia and more complicated outcome than healthy patients have (1,2). MSM should be counseled on methods to avoid or reduce the risk of sexual transmission of enteric infections such as those caused by Campylobacter or Shigella (13).