The incidence of non‐thermophilic Campylobacter species was assessed in an unselected population‐based study in a mixed urban and rural community in North Jutland, Denmark. In a 2‐year study period, 11 314 faecal samples from 8302 patients with gastroenteritis were cultured with supplement of the filter method. We recovered a high incidence of Campylobacter concisus (annual incidence 35/100 000 inhabitants), almost as high as the common Campylobacter jejuni/coli. In contrast, there was a very low incidence of other non‐thermophilic Campylobacter species, such as Campylobacter upsaliensis. Campylobacter concisus was, unlike C. jejuni/coli, found more frequently among small children (<1 year) and the elderly (≥65 years). Around 10% of the patients with C. consisus had co‐infections dominated by Clostridium difficile and Salmonella enterica, whereas co‐infections occurred in about 5% of C. jejuni/coli patients. We observed a seasonal variation in C. jejuni/coli with a peak incidence in late summer months and autumn, whereas there was an almost constant monthly prevalence of C. concisus. Among patients participating in a questionnaire sub‐study, there was a higher degree of close contacts with animals, especially dogs, as well as a higher travel exposure among C. jejuni/coli patients compared with C. concisus patients. We did not culture any C. concisus in stool samples from a small cohort of healthy individuals. Future studies have to focus on the clinical follow‐up and the long‐term risk of inflammatory bowel diseases in C. concisus‐positive patients. We conclude that there is a high incidence of C. concisus in Denmark.

Campylobacter concisus infections of the gastrointestinal tract can be accompanied by diarrhea and inflammation, whereas colonization of the human oral cavity might have a commensal nature. We focus on the pathophysiology of C. concisus and the effects of different clinical oral and fecal C. concisus strains on human HT-29/B6 colon cells. Six oral and eight fecal strains of C. concisus were isolated. Mucus-producing HT-29/B6 epithelial monolayers were infected with the C. concisus strains. Transepithelial electrical resistance (Rt) and tracer fluxes of different molecule size were measured in Ussing chambers. Tight junction (TJ) protein expression was determined by Western blotting, and subcellular TJ distribution was analyzed by confocal laser-scanning microscopy. Apoptosis induction was examined by TUNEL-staining and Western blot of caspase-3 activation. All strains invaded confluent HT-29/B6 cells and impaired epithelial barrier function, characterized by a time- and dose-dependent decrease in Rt either after infection from the apical side but even more from the basolateral compartment. TJ protein expression changes were sparse, only in apoptotic areas of infected monolayers TJ proteins were redistributed. Solely the barrier-forming TJ protein claudin-5 showed a reduced expression level to 66±8% (P<0.05), by expression regulation from the gene. Concomitantly, Lactate dehydrogenase release was elevated to 3.1±0.3% versus 0.7±0.1% in control (P<0.001), suggesting cytotoxic effects. Furthermore, oral and fecal C. concisus strains elevated apoptotic events to 5-fold. C. concisus-infected monolayers revealed an increased permeability for 332 Da fluorescein (1.74±0.13 vs. 0.56±0.17 10−6 cm/s in control, P<0.05) but showed no difference in permeability for 4 kDa FITC-dextran (FD-4). The same was true in camptothecin-exposed monolayers, where camptothecin was used for apoptosis induction.In conclusion, epithelial barrier dysfunction by oral and fecal C. concisus strains could mainly be assigned to apoptotic leaks together with moderate TJ changes, demonstrating a leak-flux mechanism that parallels the clinical manifestation of diarrhea.

This study emphasizes the importance of infection control measures to limit transmission of VREfm between patients. However, the diversity of the VREfm isolates points to the fact that other important factors may also affect the VREfm increase in Denmark. Finally, WGS is suitable for typing of VREfm and has replaced PFGE for typing of VREfm in Denmark.