Prevalence of Arcobacter Species among Humans, Belgium, 2008–2013

We examined fecal samples from 6,774 patients with enteritis in Belgium, 2008–2013. Members of the genus Arcobacter were the fourth most common pathogen group isolated, and the isolation rate was higher than previously reported. Culturing Arcobacter in a microbiology laboratory is feasible and should thus be tested for in cases of diarrheal disease.

and raw or undercooked food that is eaten or handled are the most frequent sources of human infection (5,6). An association of A. butzleri and A. cryaerophilus with enteritis, colitis, and septicemia has been proposed in epidemiologic studies, outbreak reports, and case reports (7), although A. cryaerophilus can also be present in healthy humans (8). More recently, A. skirrowii has been implicated as the causal agent of bacteremia and colitis in individual case reports (9).
Information about transmission, colonization in the gut, and virulence in humans of members of the Arcobacter genus is insufficient. The presence of virulence genes and their distribution within the genus have been proven, but their role in pathogenicity in selected host groups has not yet been demonstrated (10).
Most laboratories do not use appropriate culture methods or conditions to detect species other than Campylobacter jejuni and C. coli from feces. Furthermore, identification of Campylobacter and related organisms to species level is not always performed or is executed with phenotypic methods, leading to conflicting results. Hence, data on the incidence and clinical importance of Arcobacter remain scarce. The aims of this study were to assess the feasibility of Arcobacter culture in a clinical microbiology laboratory and to determine the recent prevalence of Arcobacter spp. in humans with gastrointestinal disease in Belgium.

The Study
Fecal samples were collected from outpatients and patients with symptoms of enteritis who were admitted for <72 hours to the Sint-Lucas Hospital in Ghent, Belgium, during 2008-2013. All age groups were represented. We examined the fecal samples macroscopically for consistency and presence of blood and mucus and microscopicallyfor leukocytes and cultured them for all common bacterial pathogens. Feces of patients <2 years of age were also tested for rotavirus and enteric adenovirus, and those of adult patients with loose or mucouscontaining stools were also tested for toxigenic Clostridium difficile. Samples were tested for parasites on clinical demand only.
We cultured for Arcobacter using an isolation method for fecal samples carried out in veterinary medicine that was previously validated for use on human feces (11): 1 g of feces was inoculated into a liquid selective broth (24 g/L Arcobacter broth with 50 mL lysed, defibrinated horse blood and an antimicrobial supplement of 16 mg/L cefoperazone, 10 mg/L amphotericin B, 100 mg/L 5-fluorouracil, 64 mg/L trimethoprim, and 32 mg/L novobiocine [Oxoid, Cambridge, United Kingdom]) and incubated for 24 h at 25°C in a microaerobic atmosphere of 6% O 2, 7% CO 2, 7% H 2 , and 80% N 2 (Advanced Instruments, Inc., Norwood, MA, USA). Then, 40 μL of enriched broth was plated onto a solid Arcobacter-selective medium with the same composition as the broth but with the addition of 12 g/L agar technical no. 3 (Oxoid) and without the horse blood. The plates were incubated for 72 h at 25°C in a microaerobic atmosphere and examined daily. Arcobacter colonies were detected by screening the selective transparent agar as described with Henry transillumination microscopy for bluish colonies (11). Definitive identification was performed by a genus-and species-specific multiplex PCR, and subsequently, by amplification fragment length polymorphism (12,13).
Of the 8,994 eligible samples received during the study period, 6,774 (75.31%) samples were cultured for Arcobacter; 2,220 samples (24.68%) were excluded be- Arcobacter spp. were isolated in samples from 89 patients (1.31%), ranking members of this genus as the fourth most commonly isolated pathogen group in the study. This ranking is comparable to those of other reports, although a higher isolation rate was obtained during this study (14,15). The distribution frequency of pathogens and of A. butzleri versus A. cryaerophilus, 49 (0.72%) and 38 (0.56%) isolates, respectively, is shown in Table 1. A. skirrowii was not recovered, but the first 2 known isolates of A. thereius (0.03%) from humans were documented. In the excluded population (n = 2,220), comparable recovery rates of Campylobacter spp. (6.53%), Salmonella spp. (2.88%) and C. difficile (2.07%) were observed. In 6 patients whose samples were positive for Arcobacter, a second pathogen was isolated: 1 patient was positive for Salmonella spp., samples from 2 patients showed Campylobacter spp. and A. butzleri, and toxigenic C. difficile was detected in 2 patients who tested positive for A. butzleri and in 1 patient who tested positive for A. cryaerophilus.
We collected retrospective data about age, hospital stay, presence of diarrhea, clinical diagnosis of enteritis, colonoscopy results, and underlying disease for the positive-culture patient group. We assessed differences between study populations and results using t-test and Χ 2 statistical methods as appropriate.
The mean age of the included population (29 years) differed significantly (p<0.01) from the excluded group (22 years), because of an excess of infants from whom sample material was insufficient in the latter group. The positiveculture patient group had a mean age of 42 years and demonstrated a significantly older age distribution pattern (p<0.01) (Figure). The consistency of fecal samples was not predictive (p = 0.62) for the presence of Arcobacter. Within the Arcobacter-positive group, the mean age of patients whose specimens shed A. butzleri (49 years) was significantly higher (15 years, p<0.01) than that of patients whose specimens shed A. cryaerophilus (34 years). Presence of A. butzleri in patients admitted to the hospital and in outpatients was not significantly linked with underlying disease (p = 0.12). A. cryaerophilus was more frequently observed in outpatients who had uncomplicated gastroenteritis than in hospitalized patients with coexisting conditions (p<0.001) ( Table 2).

Conclusions
Arcobacter species were the fourth most common pathogen group isolated from fecal samples from persons with acute enteric disease. The high isolation rate could possibly be explained by the inclusion of an outpatient population, local indications for sampling, and the use of an enrichment culture method. The Arcobacter-positive patients tended to belong to older age groups. No notable association of A. butzleri enteritis with coexisting conditions was observed. The feasibility of selective culturing Arcobacter from fecal material in a routine microbiology hospital laboratory was confirmed, but the slow turnaround times for culture results show a need for optimization of methods. Arcobacter species should be considered and tested for in cases of diarrheal disease.

Acknowledgments
We thank the technicians of the microbiology laboratory in Sint-Lucas Hospital for their technical skills and support, and give special thanks to Cindy Germis and Annick Van Der Straeten; we also thank Sandra Vangeenberghe, technician in the veterinary public health laboratory, Ghent University, Merelbeke.