Background

Campylobacter spp. are common causes of human bacterial gastrointestinal disease worldwide in particular C. jejuni and C. coli are recognized as the leading cause of acute bacterial diarrhea worldwide accounting for nearly 400 million cases per year [1, 2]. Campylobacter-associated gastroenteritis is highly prevalent in South and Southeast Asia and is commonly associated with traveler’s diarrhea (TD) [3,4,5,6]. Over the last decade, recent evidence of other Campylobacter spp. to include C. ureolyticus and C. concisus has highlighted the potential role of these microorganisms to cause gastroenteritis in both industrialized and developing countries [7, 8].

Several Campylobacter spp. (non-C. jejuni/coli) were isolated in the early 1990s from diarrhea stool samples though the ability to culture these organisms was highly variable among laboratories [9, 10]. Many Campylobacter spp. require different conditions for growing due to their fastidious laboratory growth requirements in comparison to other Gram-negative organisms. These factors suggest that other Campylobacter spp. besides C. jejuni and C. coli are likely underestimated as diarrhea etiologic agents as a result of inappropriate detection and isolation procedures which generally support only the growth of C. jejuni and C. coli [11].

Campylobacter concisus historically has been associated as being a human oral pathogen causing gingivitis and peridontitis [12]. More recent studies have shown that C. concisus has been isolated from chronic, more milder diarrheal cases in both adults and children as compared to C. jejuni [13,14,15,16]. Additionally, C. concisus has been implicated in the development of Crohn’s disease and inflammatory bowel disease (IBD) as it has been isolated from fecal samples and colonic biopsies of children with Crohn’s disease and from ulcerative colitis biopsy samples from adults with IBD [17,18,19]. C. ureolyticus, like C. concisus, has been detected more frequently in diarrheal cases. A recent study from Hatanaka et al. showed that C. ureolyticus was detected in 51.9% of diarrheal stool samples from children with gastroenteritis [20]. Similar studies conducted in Ireland and Chile have also implicated C. ureolyticus as an emerging gastrointestinal pathogen [21,22,23]. Like C. concisus, C. ureolyticus has been isolated at high rates from patients with ulcertative colitis and Crohn’s Disease [19, 24].

Currently, there are no to limited data on the prevalence of non-C. jejuni/C. coli Campylobacter spp. specifically, C. ureolyticus and C. concisus, in diarrhea cases in Thailand and Nepal particularly among travelers. In the present/current study, specific pathogen-free stool samples from travelers seeking medical care for traveler’s diarrhea and asymptomatic controls in Nepal and Thailand were assayed for the detection of C. ureolyticus and C. concisus using Campylobacter genus and species specific primers for C. ureolyticus and C. concisus. To our knowledge, this is the first report of C. ureolyticus and C. concisus in stool samples in travelers with traveler’s diarrhea and asymptomatic controls from Nepal and Thailand.

Methods

Sources of samples

Stool samples used in this study were from two separate epidemiological surveillance studies of TD in Thailand and Nepal conducted in 2011–2014 respectively. The period of enrollment date for the Thailand study was 20 January 2012 to 23 December 2014 and the period of enrollment for the Nepal study was 7 November 2012 to 6 November 2014. For both studies, subjects were residents of North America, Western Europe, Australia, New Zealand, Japan, Taiwan, or South Korea who were 18 years of old or older who visited the Out-patient Department (OPD) or In-patient Department (IPD) of Bumrungrad International Hospital located in Bangkok, Thailand or the CIWEC Clinic Travel Medicine Center in Kathmandu, Nepal. Cases were defined as those individuals with three or more unformed stools per 24 h with at least one additional symptom (nausea, vomiting, abdominal cramps, fecal urgency, tenesmus and fever).

Age-matched controls were residents of North America, Western Europe, Australia, New Zealand, Japan, Taiwan, or South Korea who were 18 years of old or older who visited the Out-patient Department (OPD) or In-patient Department (IPD) of Bumrungrad International Hospital located in Bangkok, Thailand or the CIWEC Clinic Travel Medicine Center in Kathmandu, Nepal. Who had no history of diarrhea in the 2 weeks prior to visiting Bamrungrad International Hospital and the CIWEC Clinic Travel Medicine Center. Subjects who had traveled outside of developed countries or had resided outside of their home countries for more than 30 or 90 days, respectively, within the past year were not enrolled. Controls were recruited by the study team after a diarrheal case was enrolled at Bamrungrad International Hospital and the CIWEC Clinic Travel Medicine Center. After informed consent was obtained, a stool sample was provided by the study volunteer. The Thai study was approved by the Bumrungrad International Institutional Review Board (IRB), Thailand and the Walter Reed Army Institute of Research (WRAIR) IRB, Silver Spring, MD, USA. The Nepal study was approved by the Nepal Health Research Council, Kathmandu, Nepal and the WRAIR IRB. A total of 214 pathogen-negative stool samples (Thailand: 26 cases and 30 controls; Nepal: 83 cases and 75 controls) were used in the study that had previously undergone laboratory testing for enteric bacteria, viruses and parasites as previously described [25, 26]. After testing, pathogen- negative stool samples were archived at −70 °C and stored in Bangkok, Thailand.

Identification of C. concisus and C. ureolyticus

DNA was extracted from “pathogen negative” stool samples using the QIAamp Fast DNA Stool Mini Kit (Qiagen, Germantown, MD, USA) according to the manufacturer’s instructions. To identify Campylobacter genus positive stool samples, PCR was used to amplify an 816 bp product of Campylobacter 16S rRNA as previously described using the primers C412F (5′-GGA TGA CAC TTT TCG GAG C-3′) and C1288R (5′-CAT TGT AGC ACG TGT GTC-3′) [27, 28]. C. ureolyticus and C. concisus were differentiated via PCR using primers for species-specific heat shock protein genes, Hsp60 and Cpn60 as previously described [22, 29].

To identify C. ureolyticus, the following primers were used to amplify a 429 bp fragment of the heat-shock protein, Hsp60: CU-Hsp60_F (5′-GAA GTA AAA AGA GGA ATG GAT AAA GAA GC-3′) and CU-Hsp60_R (5′-CTT CAC CTT CAA TAT CCT CAG CAA TAA TTA AAA GA-3′) [22]. To identify C. concisus, the following primers were used to amplify a 158 bp fragment of the heat shock protein, cpn60: JH0023 (5′-GGC TCA AAA AGA GAT CGC TCA-3′) and JH0024 (5′-CCC TCA ACA ACG CTT AGC TC-3′) [29]. The following cycling conditions were used: initial denaturing at 94 °C for 5 min, followed by 40 cycles of 1 min at 94 °C, 1 min at 61 °C for C. ureolyticus and 64.6 °C for C. concisus, and 1 min at 72 °C and a final extension at 72 °C for 10 min. Positive samples included C. ureolyticus American Type Culture Collection (ATCC) 43605 and C. concisus ATCC 33237. Negative controls included Salmonella typhimurium ATCC 14028, Shigella flexneri ATCC 12022, Shigella sonnei ATCC 25931, Vibrio cholera N16961, enterotoxigenic Escherichia coli ATCC 35401, enteropathogenic E. coli ATCC 43887, enteroinvasive E. coli ATCC 43893 and Arcobacter butzleri (laboratory strain). Gels were stained with ethidium bromide and amplicons were visualized using gel electrophoresis. The amplified C. concisus (158 bp) and C. ureolyticus (429 bp) fragments were sequenced and sequences were confirmed using the Blast Alignment Tool in the National Center for Biotechnology Information database (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Statistical analyses

Odds ratios and 95% confidence intervals were calculated using multivariate logistic regression for statistical analysis on case–control data. Data were analyzed using GraphPad Prism Version.

7 (GraphPad Software Inc., La Jolla, CA, USA). A P value of <0.05 was considered statistically significant.

Results

A total of 61 previously determined to be pathogen-negative TD stool samples (CIWEC Travel Medicine Clinic, Nepal: cases = 36, control = 14; Bamrungrad International Hospital, Thailand: cases = 9, controls = 2) were positive for Campylobacter 16S rRNA (Table 1). Eighteen isolates (29.5%; 18/61) were positive only for Campylobacter genus and were not further speciated beyond C. concisus and C. ureolyticus.

Table 1 C. concisus and C. ureolyticus detected as monoinfections and in mixed infections by PCR using species specific housekeeping genes in pathogen-negative traveler’s diarrhea cases and asymptomatic controls from CIWEC Travel Medicine Clinic, Kathmandu, Nepal and Bamrungrad International Hospital, Bangkok, Thailand
Full size table

At the CIWEC Travel Medicine Clinic, Nepal, C. consisus was identified significantly more often in TD cases (28.9%; 24/83) when compared to asymptomatic controls (4%; 3/75) (OR = 9.8; 95% CI 2.8–34.1; P = 0.0003) when isolated as the sole pathogen (Table 1). C. ureolyticus was detected in four cases (4.8%; 4/83) and four controls (5.3%; 4/75) respectively (OR = 0.90; 95% CI 0.2–3.7; P = 0.88) when isolated as the sole pathogen.

Campylobacter concisus and C. ureolyticus mixed infections were detected in two cases (2.4%; 2/83) and in one control (1.3%; 1/75) sample (OR = 1.8; 95% CI 0.2–20.5; P = 0.62).

At the Bamrungrad International Hospital, Thailand, C. consisus was identified in only two TD cases (7.7%; 2/26) and in none of the controls (0%; 0/30) (Table 1). C. ureolyticus was detected in one case (3.8%; 1/26) and in one control (3.1%; 1/30) (OR = 1.16; 95% CI 0.07–19.5; P = 0.92) in the stool samples collected at Bamrungrad. C. concisus and C. ureolyticus as a mixed infection were detected in one case (3.8%; 1/26) and in none of the control samples.

In Table 2, the number of C. jejuni and C. coli isolates detected by culture are listed in both monoinfections and mixed infections by site. C. jejuni was detected significantly more often in the cases (7.9%; 38/480) when compared to the controls (2.9%; 6/209) at the CIWEC Travel Medicine Clinic, Nepal where C. jejuni was the sole pathogen isolated (OR = 6.5; 95% CI 2.2–19.1; P = 0.0007). C. coli was detected slightly more often in the cases (1.5%; 7/480) when compared to the controls (1.4%; 3/209) though not significantly (OR = 1.0; 95% CI 0.3–3.9; P = 0.9) when C. coli was the sole pathogen detected. In mixed infections, C. jejuni was detected significantly more often in the cases (8.1%; 39/480) when compared to the controls (1.9%; 4/209) (OR = 20.2; 95% CI 2.7–152.9; P = 0.0036) (Table 2). C. coli was detected more often in the cases (1.3%; 6/480) when compared to the controls (0.9%; 2/209) though not significantly in mixed infections (OR = 1.3; 95% CI 0.3–6.6; P = 0.74) (Table 2).

Table 2 C. jejuni and C. coli isolated as single pathogens and from mixed infections via culture in traveler’s diarrhea cases and asymptomatic controls from CIWEC Travel Medicine Clinic, Kathmandu, Nepal and Bamrungrad International Hospital, Bangkok, Thailand
Full size table

At Bamrungrad International Hospital, Thailand, C. jejuni was detected significantly more often in the cases (13.9%; 24/173) when compared to the controls (2.4%; 4/165) when C. jejuni was the sole pathogen isolated (OR = 2.9; 95% CI 1.2–6.9; P = 0.01) (Table 2). C. coli was detected only in the cases (2.3%; 4/173) when C. coli was isolated as a single pathogen. C. jejuni was detected significantly more often in the cases (11%; 19/173) when compared to the controls (0.6%; 1/165) (OR = 20.2; 95% CI 2.7–152.9; P = 0.0036) in mixed infections. In mixed infections, C. coli was detected only in the cases (2.9%; 5/173) at Bamrungrad International Hospital, Thailand.

Discussion

The prevalence of Campylobacter spp. as etiologic agents for TD in Nepal and Thailand has been well documented, but the vast majority of these studies have focused on the detection of C. jejuni and C. coli as the primary Campylobacter spp [4, 5, 30,31,32]. More recently, other Campylobacter spp. to include C. concisus and C. ureolyticus have been implicated in human gastroenteritis due to improved methods of laboratory detection. Most clinical laboratories do not regularly test for other Campylobacter spp. such as C. concisus and C. ureolyticus due to both the fastidious nature of the organisms, the requirement for an enriched H2 environment in order for growth and both species being unable to grow on the selective media that is commonly used for C. jejuni and C. coli [16, 20, 33]. It is highly probable that due to these reasons, gastroenteritis caused by C. concisus and C. ureolyticus is highly underestimated. In TD etiologic studies, C. concisus and C. ureolyticus are rarely tested for and therefore the incidence of C. concisus and C. ureolyticus in TD is largely unknown.

In this study, C. concisus was detected significantly more often in TD stool samples from Nepal as compared to asymptomatic controls. These stools samples were previously categorized as pathogen-negative after undergoing a comprehensive laboratory workup for enteric bacteria, viruses and parasites. In the present study, 92.3% of the C. concisus cases were from Nepal. Similarly, 80% of the C. ureolyticus cases were from Nepal though there were only five total cases of C. ureolyticus detected. There are no previous published data on the prevalence of C. concisus and C. ureolyticus from the CIWEC Travel Medicine Clinic, Kathmandu, Nepal and at Bamrungrad International Hospital, Bangkok, Thailand as these microorganisms are not routinely tested for in the clinical laboratories at these facilities. In a previous TD case–control study conducted at the CIWEC Travel Medicine Clinic from 2001 to 2003, Campylobacter was the most prevalent pathogen detected in cases (17%), but the Campylobacter isolates were not speciated in this study [4]. Multiple studies have also shown that Campylobacter spp. is the most common pathogen in Thailand to cause acute gastroenteritis in travelers, deployed US military personnel and the indigenous population [5, 6, 31, 34, 35]. Similar to cases from CIWEC, C. jejuni was the most common detected pathogen from cases at Bamrungrad International Hospital. To our knowledge, this is the first report showing detection of C. concisus and C. ureolyticus in TD samples in a case–control setting from Nepal and Thailand and that these results strongly suggest a potential role of C. concisus in the development of TD in travelers to Nepal.

Recent studies have shown that the prevalence of C. concisus and C. ureolyticus typically are equal or higher when compared to C. jejuni and C. coli in studies in which they are both included in the laboratory workup [36,37,38]. In Australia, Underwood et al. showed that C. concisus had a prevalence of 49.7% in patients with gastroenteritis as compared to C. jejuni (5%) and that it was detected as a single agent in 78% of the C. concisus positive cases indicating a potential role for C. concisus in the development of gastroenteritis in these patients [16]. In a 2 year study conducted in Denmark, Nielsen et al. showed that C. concisus was prevalent in 26.1% of patients with gastroenteritis as compared to C. jejuni/coli detected in 31.9% of infected patients [15]. In the first report of C. ureolyticus in children with gastroenteritis in Japan, Hatanaka et al. showed that C. ureolyticus was prevalent in 51.9% of the infected children as compared to C. jejuni/coli infecting 15.5% of the children [20]. In a study from Southern Ireland, C. ureolyticus was prevalent in 41% of diarrheal patients as compared to C. jejuni at 50.7% and C. coli at 5.7% [28]. In the present study, only the incidence of C. concisus in the Nepal pathogen-negative TD cases (28.9%) was higher than the incidence of both C. jejuni (16.0%) and C. coli (2.7%) in the Nepal TD cases that yielded at least one enteric pathogen under routine laboratory testing. It is likely that the incidence of C. concisus and C. ureolyticus are highly underestimated due to the fact that detection of these organisms was not included in the initial routine laboratory testing of the TD cases from both CIWEC and Bamrungrad hospital.

There were several limitations to the study. As mentioned previously, only pathogen-negative stools were selected which limited the number of samples used. Between the two sites, there were 653 cases and 374 controls (total = 1027). This resulted in a smaller sample size particularly for the TD samples from Thailand (26 cases and 30 controls) as well as an underestimation of the overall prevalence of both C. concisus and C. ureolyticus in all of the cases from both sites. The selection of pathogen-negative stools also inhibited our ability to address co-infections with C. concisus and C. ureolyticus. However, since pathogen-negative samples were solely used, we can infer from the data that those samples containing C. concisus or C. ureolyticus were monoinfections with no other enteric pathogens being detected previously. As both organisms have been implicated in the development of gastroenteritis, there have been recent studies to include whole genome sequencing aimed at delineating virulence factors to elucidate the pathogenic mechanisms of each organism. Both C. concisus and C. ureolyticus contain genes similar to C. jejuni for adhesins, invasins and toxins that are known to play an important role in the development of acute gastroenteritis [37, 39,40,41,42]. These virulence factors were not tested for in the present study. Additionally, PCR assays should have been included for the detection of C. jejuni and C. coli as PCR has shown to be far superior for the detection of both C. jejuni and C. coli when compared to culture [43,44,45,46]. It is likely that the overall estimate of C. jejuni and C. coli by PCR would increase to those that were positive by culture. Future plans include testing all 826 samples using PCR for C. jejuni, C. coli, C. concisus and C. ureolyticus to better compare the prevalence of each organism using a similar molecular method of detection and in the same batch of samples.

As this is the first report of C. concisus and C. ureolyticus in acute gastroenteritis in travelers to Nepal and Thailand, further studies with long term clinical follow up of infected travelers should be initiated to assess the development of any long term post infectious sequelae to include IBD and ulcerative colitis that have been observed in studies examining the long term effects of Campylobacter infection [47, 48]. In conclusion, as it is evident that at least C. concisus from this study is involved in the development of acute gastroenteritis in travelers to Nepal, it is imperative that more studies include both C. concisus and C. ureolyticus are included in the laboratory diagnostic methods to better ascertain their true prevalence as etiological agents of acute gastroenteritis. Additionally, there are limited data on the geographic distribution of C. concisus and C. ureolyticus in developing and developed countries and further epidemiological studies to ascertain this are warranted.

Conclusions

The prevalence of C. ureolyticus and C. concisus in traveler’s diarrhea cases in travelers to Nepal and Thailand largely unknown. In this study, pathogen-negative stool samples from travelers seeking medical care for traveler’s diarrhea and asymptomatic controls in Nepal and Thailand were assayed by PCR for the detection of C. ureolyticus and C. concisus using Campylobacter genus and species specific primers for C. ureolyticus and C. concisus. The results of the study show that C. concisus is potentially involved in the development of acute gastroenteritis in travelers to Nepal as it was detected significantly more often in cases when compared to controls. Future traveler’s diarrhea studies should include C. concisus and C. ureolyticus in their laboratory diagnostics menu to better estimate their potential as etiological agents of traveler’s diarrhea.