From January 1995 to December 2002, all stool samples submitted to two hospital laboratories serving the Brugmann, Queen Fabiola, and Saint-Pierre University Hospitals in Brussels, Belgium, were examined macroscopically for consistency, gross blood, and mucus and microscopically for parasites, leukocytes, and erythrocytes. These samples were also cultured for common bacterial pathogens. Stool samples of patients <2 years of age were also evaluated for rotavirus and enteric adenovirus since viral diarrhea is mainly seen in young children.

A specific culture protocol for the recovery of Campylobacter spp. and related organisms consisting of one solid selective medium and a filtration method was systematically applied. In the selective agar method, the fecal suspension (approximately 1 g/mL of saline) was plated onto Butzler's medium comprising Mueller-Hinton agar (Oxoid Ltd, Basingstoke, United Kingdom) containing 5% sheep blood and the antimicrobial supplement cefoperazone 30 mg/L, rifampicin 10 mg/L, and amphotericin B 2 mg/L (Institute Virion, Rüschlikon, Switzerland) (13). The plates were incubated for 48 h at 42°C in a microaerobic atmosphere (5% O₂, 6% H₂, 10% CO₂, and 79% N₂). The plates were examined daily for growth of Campylobacter species.

The membrane filtration method was performed according to Lopez (14). Stool samples were diluted 1:5 in Brucella broth. Cellulose acetate filters (50 mm in diameter) with a pore size of 0.45 µm were placed on the surface of Mueller-Hinton agar plates containing 5% sheep blood. Eight drops of the fecal suspension were placed on the top of the membrane and allowed to filter passively for 30 min at 37°C in air. After filtration, the filters were removed, and the plates were incubated at 37°C in a microaerobic atmosphere for up to 10 days.

Two specific procedures for isolating Arcobacter were used successively. Until April 1995, we used a membrane filtration technique (15). Subsequently, we switched to direct plating on Arcobacter selective medium (16), for which 0.5 g of stool samples was injected into an enrichment broth (Brucella broth + antimicrobial supplement consisting of piperacillin 75 mg/L, cefoperazone 25 mg/L, amphotericin B 10 mg/L, trimethoprim 20 mg/L, and cycloheximide 100 mg/L + 5% laked horse blood) and incubated for 24 h at 25°C in a microaerobic atmosphere. After incubation, 40 µL of enriched broth was plated onto Arcobacter selective medium (16). The plates were then incubated for 3 days at 25°C in a microaerobic atmosphere and examined daily.

Gram-negative, motile, spiral, or curved rods were identified as Campylobacter by morphologic and biochemical characteristics consistent with the genus Campylobacter, as recommended by Vandamme et al. (5). A complementary distinction between A. butzleri and A. cryaerophilus was achieved by using sodium dodecyl sulfate–polyacrylamide gel electrophoresis of whole-cell proteins and by a multiplex polymerase chain reaction assay (17).

To assess the pattern of clinical disease, patients with A. butzleri infection were matched against three randomly selected control patients with C. jejuni infection. To achieve this, all charts of patients with A. butzleri infection and controls were reviewed retrospectively by combining the records of the medical and microbiology departments. A structured, close-ended questionnaire was used to collect the patient's history, age, sex, status (outpatient or hospitalized for >48 h) and history of international travel. The clinical history included diarrhea within the preceding 3 months, duration of symptoms, nature of the diarrhea (watery or bloody and presence of cells), intensity of fever, nausea or vomiting, abdominal pains, asymptomatic carriage (routine screening of enteric pathogens in stools of HIV-infected persons, other immunocompromised states, and patients from a foreign country referred for surgery), and underlying disease. Diarrhea was defined as at least three unformed or liquid stools per day for at least 3 days. Because of the study's retrospective nature, we could only classify diarrhea as acute (duration <15 days) or chronic (duration >15 days). Treatment history included antimicrobial history and use of intravenous fluid therapy, and, finally, clinical status after 1 month of treatment (cure or persistent diarrhea).

All parameters were compared between index patients and matched controls. Comparisons were made by Pearson χ² test for 1:3 control data. Odds ratios (OR) and 95% confidence intervals (CI) were calculated.

From January 1995 to December 2002, a total of 67,599 stool specimens from 40,995 patients were submitted for bacteriologic stool culture. These cultures yielded Campylobacter and related organisms in 1,906 patients, Salmonella species in 1,720 patients, and Shigella species in 244 patients. Other identified pathogens include enteropathogenic Escherichia coli (137 patients), Yersinia enterocolitica (87 patients), Plesiomonas spp. (22 patients), and Aeromonas spp. (21 patients).

Among the 1,906 Campylobacter and related organisms isolated during the study period, 77.2% were C. jejuni, 11.4% were C. coli, and 4.5% were C. upsaliensis (Table 1). Ninety-seven Arcobacter isolates were obtained from 77 patients. Among them, A. butzleri was the most frequently isolated species (84 isolates from 67 patients), accounting for 3.5 % of the Campylobacter and related organisms bacterial group. Thirteen A. cryaerophilus isolates from 10 patients were obtained, but no A. skirrowii isolates were found. Other species, such as C. concisus, C. fetus, C. curvus, C. lari, and C. hyointestinalis, were also found in small numbers (Table 1).

Medical records were available for 61 of the 67 patients with A. butzleri infection. Ages of patients with A. butzleri infection were 30 days–90 years; there were slightly more female than male patients. Fourteen patients were hospitalized for >48 h. Four patients had traveled abroad before onset of symptoms. Ten patients had underlying disease: 4 were HIV seropositive, and 3 were immunocompromised (postrenal graft, celiac disease, and chemotherapy for cerebellar astrocytoma). Other chronic illnesses included dementia (n = 1), insulin-dependent diabetes mellitus (n = 1), and hepatitis C (n = 1).

Thirty-one patients complained of acute diarrhea (>24 watery stools) lasting for 3 to 15 days, and 10 had persistent or recurrent diarrhea lasting >2 weeks–2 months. Six patients had abdominal pain without diarrhea. Twelve patients were asymptomatic. Sixteen patients received antimicrobial therapy, but only 7 were treated empirically with an antimicrobial agent for which the strain was susceptible. The symptoms resolved in all patients except one, regardless of the antimicrobial agent used. Among patients treated symptomatically, three patients had persistent or recurrent symptoms.

Sixty-seven patients with A. butzleri infection were matched against 201 patients with C. jejuni infection. The age and sex distributions were similar for patients colonized by each species. No significant differences in international travel were observed. However, proportionally fewer patients with A. butzleri (79.1%) were treated as outpatients than those with C. jejuni (90.5%) (OR 0.40, 95% CI 0.17–0.90) (Table 2).

Rectal bleeding, inflammatory exudates, or both were significantly less common in A. butzleri than in C. jejuni infection (OR 0.15, 95% CI 0.05–0.46). A concomitant infection with another enteric pathogen occurred in a higher proportion of patients with A. butzleri infection than patients with C. jejuni infection, but the difference was not significant. Twelve patients with A. butzleri infection had a coinfection with one of the following enteric pathogens: Salmonella enterica (n = 4) (2 ser. Enteritidis, 1 ser. Typhimurium and 1 ser. Virchow), Rotavirus (n = 3), C. jejuni (n = 2), Giardia lamblia (n = 2), and Clostridium difficile toxin B positive (n = 1).

In 24 patients with Campylobacter jejuni infection, we found a coinfection with one another enteric pathogen: S. enterica (7) (3 ser. Enteritidis, 3 ser. Typhimurium, and 1 ser.Virchow), Rotavirus (5), Adenovirus (4), Giardia lamblia (2), Shigella flexneri (2), Yersinia enterocolitica (1), and Clostridium difficile toxin B positive (1). In two additional cases of campylobacteriosis, we found a coinfection with two other enteric organisms: one patient was infected with S. dysenteriae and Hymenolepis nana and the other patient with Rotavirus and Salmonella enterica serotype Typhimurium.

To assess the prevalence of A. butzleri and Campylobacter jejuni in diarrheic stool specimens, we considered only stools with loose or liquid macroscopic aspect as diarrhea. Among the 67,599 stool specimens received, 12,413 were solid stools, and 55,186 were diarrheic. Among the diarrheic stool specimens, we isolated A. butzleri and C. jejuni in 77 (0.14%) and 3,209 (5.81%) stools, respectively. Among the nondiarrheic stool specimens, we isolated A. butzleri and C. jejuni in 7 (0.06%) and 205 (1.65%) stools, respectively. A. butzleri was more frequently isolated from diarrheic stool specimens than from nondiarrheic stools specimens. This difference was significant (OR 2.48, 95% CI 1.10–5.86) (p = 0.0175). We observed a similar result for the recovery of C. jejuni from diarrheic stool specimens compared with nondiarrheic stools specimens. This difference was even more significant (OR 3.68, 95% CI 3.18–4.25) (p < 0.0001).

Because medical records were unavailable for 6 patients with A. butzleri and 10 patients with C. jejuni, we compared the clinical features of 61 patients with A. butzleri matched against those of 191 patients with C. jejuni. Although diarrhea was a common clinical feature of both groups, it was more frequent in the C. jejuni–infected patients. The characteristics of the diarrhea differed significantly, however. Patients with A. butzleri were more likely to have persistent diarrhea (OR 3.97, 95% CI 1.4–11.3), or watery diarrhea (OR 2.20, 95% CI 1.18–4.13) than those with C. jejuni infection, but they were less likely to have acute diarrhea (OR 0.38 95%, CI 0.20–0.79). Other clinical features did not differ significantly. Asymptomatic carriage was more frequently encountered in patients with A. butzleri infection than in those with C. jejuni infection, but not significantly (OR 2.22, 95% CI 0.94–5.21).

Proportionally more patients with C. jejuni infections were treated with antimicrobial agents (41.4%) than patients with A. butzleri infections (26.2%) (OR 0.50, 95% CI 0.25–1.0). Among them, only 43.8% of patients with A. butzleri infections were treated empirically with an antimicrobial agent for which the strain was susceptible, whereas 79.5% of patients with C. jejuni received an appropriate antimicrobial drug.

In this study, Arcobacter was the fourth most common Campylobacter or Campylobacter-like organism isolated from stool specimens in our laboratories. Our observations suggest that A. butzleri display similar microbiologic and clinical features as C. jejuni. However, compared with C. jejuni, A. butzleri are more frequently associated with a persistent and watery diarrhea and less associated with bloody diarrhea. This first study on Arcobacter in humans could be the beginning of future research to better understand the pathogenesis and epidemiologyof these non–jejuni/coli Campylobacter.