Infections with Campylobacter spp. pose a significant health burden worldwide. The significance of Campylobacter jejuni/Campylobacter coli infection is well appreciated but the contribution of non-C. jejuni/C. coli spp. to human gastroenteritis is largely unknown. In this study, we employed a two-tiered molecular study on 7194 patient faecal samples received by the Microbiology Department in Cork University Hospital during 2009. The first step, using EntericBio® (Serosep), a multiplex PCR system, detected Campylobacter to the genus level. The second step, utilizing Campylobacter species-specific PCR identified to the species level. A total of 340 samples were confirmed as Campylobacter genus positive, 329 of which were identified to species level with 33 samples containing mixed Campylobacter infections. Campylobacter jejuni, present in 72.4% of samples, was the most common species detected, however, 27.4% of patient samples contained non-C. jejuni/C. coli spp.; Campylobacter fetus (2.4%), Campylobacter upsaliensis (1.2%), Campylobacter hyointestinalis (1.5%), Campylobacter lari (0.6%) and an emerging species, Campylobacter ureolyticus (24.4%). We report a prominent seasonal distribution for campylobacteriosis (Spring), with C. ureolyticus (March) preceeding slightly C. jejuni/C. coli (April/May).
Campylobacter-related gastroenteritis is one of the most common causes of acute bacterial diarrhoeal illness in both developed and developing countries (Coker et al., 2002). There is a persistently high incidence of campylobacteriosis in European countries, and the 2007; crude incidence rate in Ireland was a direct reflection of the European average of 45.2/100 000 (HSPC, 2007; EFSA, 2009). National surveillance programmes which have been employed by many industrialized countries have contributed to improved reporting of Campylobacter infections (HSPC, 2007). Yet despite the high reported incidence it is estimated that the true rate of infection may be between seven and thirty times higher (Crushell et al., 2004; Moore et al., 2005; Janssen et al., 2008).
Campylobacteriosis is primarily regarded as a zoonosis and the reservoirs for infection vary depending on the species, but encompass a wide range of birds, livestock and domestic animals (McClurg et al., 2002; Moore et al., 2005). To date, 26 species are assigned to the Campylobacter genus (Euzeby, 1997), however, only some of these have been linked with human gastroenteritis (Mandrell et al., 2005; Moore et al., 2005). Campylobacter jejuni and Campylobacter coli combined have been reported as accounting for at least 95% of Campylobacter isolations in humans with gastroenteritis (Kulkarni et al., 2002; Moore et al., 2005), when using traditional culture methods for diagnosis. Advances in molecular diagnostics facilitate detection of non-C. jejuni/C. coli spp. from clinical samples (Duffy et al., 2007; Bullman et al., 2011). However, to date, the true contribution of non-C. jejuni/C. coli species to human gastroenteritis remains unknown, and the aims of our study were to determine the proportion of non-C. jejuni/C. coli Campylobacter in faeces from patients with acute gastroenteritis, and to analyse the complete collection of results with regard to seasonality and age distribution.
In this study we employed a two-tiered molecular-based analysis to examine the distribution of Campylobacter spp. in patients presenting with diarrhoeal illness for the calendar year 2009. The first stage of the investigation was to detect a genus-specific marker for Campylobacter spp., and the second stage sought to determine identity to species level.
Materials and methods
A total of 7194 faecal samples collected over a 1-year period, from patients presenting with diarrhoea were screened for Campylobacter spp. using EntericBio® (Serosep Ltd, Limerick, Ireland). Of these, a total of 349 DNA samples were included in the study. These samples, stored at −20 °C, were extracted from anonymized faecal samples that tested positive for Campylobacter genus using the EntericBio® system. The samples were received by the Department of Medical Microbiology at Cork University Hospital, Ireland, between January 2009 and December 2009.
The following control isolates were used in the study: Campylobacter jejuni ssp. jejuni DSM 4688, Campylobacter coli DSM 4689, Campylobacter lari ssp. lari DSM 11375, Campylobacter fetus ssp. fetus DSM 5361, Campylobacter upsaliensis DSM 5365, Campylobacter hyointestinalis ssp. hyointestinalis DSM 19053 and Campylobacter ureolyticus DSM 20703. All control strains were obtained from DSMZ, Braunschweig, Germany.
PCR template preparation
DNA was extracted from the faecal samples using the EntericBio® system in accordance with the manufacturer's instructions. The extracted DNA samples were either used immediately in the EntericBio® system or stored at −20 °C for subsequent use in Campylobacter species-specific and genus-specific PCR.
The 349 Campylobacter genus positive DNA samples were investigated using uniplex species-specific PCR assays for C. jejuni, C. coli, C. lari, C. fetus, C. hyointestinalis, C. upsaliensis and C. ureolyticus, respectively, according to the methods used by Bullman et al. (2011).
S rRNA gene sequencing
A Campylobacter genus-specific PCR targeting the 16S rRNA gene was conducted on DNA samples that were negative for species-specific PCR. Samples that produced a PCR amplicon were subject to 16S rRNA gene sequence analysis (Eurofins MWG Operon, Ebersberg, Germany). Returned sequences were subjected to blast analysis using the NCBI database and aligned by the clustalw method in Megalign from the Lasergene suite of programs (DNAstar). Clinical strains were considered to be identified when sequence similarities with GenBank entries exceeded 99% identity over the entire length of the sequence.
On receipt of the samples in the Microbiology Laboratory, each patient sample was assigned a unique specimen record number and the samples were anonymized for the purposes of this study. Details of the patients' age and gender along with sample details and collection information were provided on patients who were confirmed Campylobacter-positive, from which age and gender profiles were constructed.
Confirmation of EntericBio® positive Campylobacter spp. results
Of 7194 patient faecal samples submitted to the Microbiology Laboratory from January to December of 2009, a total of 349 samples were determined to be Campylobacter genus-positive by the EntericBio® molecular method. Campylobacter species-specific PCR confirmed the presence of Campylobacter spp. in 329 (94.3%) of the 349 EntericBio®-positive samples. Of the 20 samples which were Campylobacter-positive by the EntericBio® but negative for the species-specific PCR, nine samples were positive for the Campylobacter genus-specific PCR (targeting the 16S rRNA gene); however, these samples failed to return results on sequencing. A further two samples were negative for the 16S rRNA gene PCR but were recorded as Campylobacter culture-positive by the clinical Microbiology Laboratory; thus, these 11 samples were reported as non-speciated Campylobacter positives. A total of nine samples were negative for both Campylobacter species-specific and genus-specific PCR and were reported as Campylobacter culture-negative by the clinical Microbiology Laboratory; thus, these nine samples were recorded as false positives. A total of 340 samples of the 349 patient samples which were positive by the EntericBio® molecular method were confirmed by alternative molecular methods (species-specific or genus-specific PCR) and/or routine culture to be true Campylobacter spp. positives.
Speciation of Campylobacter-positive samples
Campylobacter spp. were identified in 329 patient samples by species-specific PCR. There were 373 Campylobacter detections in all, as 33 samples contained mixed infections involving a total of 77 Campylobacter detections. Figure 1 shows the contribution of each species to the total number of organisms detected. The most common species, C. jejuni, was detected in 246 (72.4%) of patient samples and accounted for 66% of the total detections. Campylobacter ureolyticus was detected in 83 (24.4%) patient samples, accounting for 22.3% of the total detections. Campylobacter ureolyticus was the sole Campylobacter species detected in 55 (16.7%) patient samples (95% confidence interval: 12.8–21.2%). Campylobacter coli accounted for 6.7%; present in 25 (7.1%) of patient samples. Collectively C. fetus, C. hyointestinalis, C. upsaliensis and C. lari amounted to 5% of the total detections.
Age and gender distribution of Campylobacter- positive patients
Age and gender profiles were constructed (Fig. 2) for the 340 patients confirmed as Campylobacter positive; 52.4% were men. A bimodal age distribution was noted in both genders (Fig. 2); 51.8% of all Campylobacter detections were distributed between two age groups; children under 5 years (31.1%) and adults over 70 years (20.7%). In male patients, Campylobacter infection was commonest in those ≤ 5 years (37%), whereas in female patients, it was most frequent in those ≥ 70 years (29%) (Fig. 2). Infection remained at lower levels between late childhood and the 66–70 age group with the exception of a peak in male patients aged 41–45 (10.8%).
The age and gender profiles for C. ureolyticus exhibited a similar bimodal profile to those observed for the combined species, presenting peaks at extremes of age. This species was more commonly detected in female patients (62.7%) and more than 50% of detections in female patients were in the ≥ 70 age group (Fig. 2).
Seasonal prevalence of Campylobacter 2009
The hypotheses that the incidence of C. jejuni/C. coli and of C. ureolyticus is uniformly distributed across the 12 months were conclusively rejected by the chi-square goodness-of-fit test, showing P-values of 0.000, and 0.003, respectively).
As shown in Fig. 3, C. jejuni/C. coli infection peaked in April and May (12.9% and 12.6% of the total, respectively), declined from September onwards and remained low during the winter months.
An abrupt peak was observed for C. ureolyticus infections in early spring; 18.1% of the total number of detections occurred during March. The fewest detections of C. ureolyticus infections (2.4% of the total) were observed in December.
Campylobacter infections remain a significant public health issue, not only in terms of morbidity but also economic cost. In 2006, a total of 175 561 confirmed cases of campylobacteriosis were reported from 21 European countries (EFSA, 2009), a figure which is thought to be an underestimate (Moore et al., 2005; Janssen et al., 2008). Furthermore, the annual cost of Campylobacter infection in the US is estimated to be at least $8 billion; a large economic burden (Crushell et al., 2004). Based on routine culture detection in diagnostic laboratories, C. jejuni and C. coli have been reported as being the most prevalent Campylobacter spp. causing human disease and, as a result, the majority of Campylobacter research has focused on these two species. With recent advances in diagnostic methods, there has been an increasing realization of the limitations of routine culture for the detection of non-C. jejuni/C. coli spp. (Lastovica & le Roux, 2000; Maher et al., 2003; O'Leary et al., 2009; Bullman et al., 2011). Thus, traditional culture underestimates the true occurrence of both Campylobacter-related gastroenteritis and also the emerging campylobacteria that are detectable by molecular methods in cases of gastroenteritis (Lastovica & le Roux, 2000; Maher et al., 2003; Bullman et al., 2011). Herein, we employed a two-tiered molecular detection method, independent of culture, for the direct detection of Campylobacter spp. from the faeces of 7194 patients presenting with diarrhoeal illness.
Of 340 Campylobacter-positive samples, C. jejuni represented the largest proportion of PCR-positives, present in 72.4% of patients; although it has previously been reported that this species accounts for at least 80–90% of Campylobacter-related infections (Kovats et al., 2005; Moore et al., 2005). Campylobacter coli, traditionally ranked as the second most common Campylobacter species found in Campylobacter enteritis (Coker et al., 2002; Moore et al., 2005; Ivanova et al., 2010); was identified in 7.4% of samples. An emerging species, C. ureolyticus, was detected in 24.4% of cases, and was detected solely (without other bacteria or viruses) in the samples of 55 symptomatic patients. There are a number of reasons why C. ureolyticus fails to be identified by routine culture: it is incapable of growth under the microaerobic atmosphere conditions used (5% O2, 10% CO2 and 85% N2) unless hydrogen is supplied. In addition, certain strains appear to be sensitive to the antibiotic used in the Campylobacter selective agars. Finally, most clinical laboratories isolate Campylobacter from stools at incubation temperatures >40 °C; however, C. ureolyticus fails to grow above 37 °C (Vandamme et al., 2010). To our knowledge, this is the first report of C. ureolyticus replacing C. coli as the second most common Campylobacter species in patients presenting with diarrhoeal illness. Furthermore, as previously reported (Bullman et al., 2011), given that only 1.15% of the 7194 samples processed contained this organism, it is unlikely that C. ureolyticus exists as a common commensal in the gastrointestinal tract.
In support of the findings of Maher et al. (2003), we determined that non-C. jejuni/C. coli spp. accounted for 27.4% of the total Campylobacter organisms detected; a much greater percentage than the previously reported 1–5% (Galanis, 2007; Ivanova et al., 2010). The large proportion of non-C. jejuni/C. coli species in this study is primarily due to the novel identification of C. ureolyticus, which as mentioned previously, is not normally isolated by routine Campylobacter culturing techniques. This finding highlights the advantages of molecular diagnostics for the detection of such fastidious organisms. Recent observations by O'Leary et al. (2009), when comparing a multiplex PCR assay to routine culture indicated that approximately 30% of samples positive for Campylobacter by the molecular method failed to grow on culture. Similar results were reported by Bessède et al. (2011), who reported considerably lower sensitivity for culture-based techniques compared to molecular methods. The hypothesis that the true proportion of Campylobacter detections due to C. jejuni/C. coli is 95% or more in the current study is conclusively rejected by the appropriate statistical test (P-value = 0.000).
The reported incidence of the less commonly grown campylobacteria varies in different regions of the world (Moore et al., 2005), and interestingly, we have detected C. fetus in 2.4% of the Campylobacter-positive samples (0.11% of the 7194 samples analysed). Campylobacter fetus has been considered to be a very rare cause of gastroenteritis in humans and acts most often as an opportunistic pathogen in immunocompromised individuals (Umehara et al., 2009); our findings may suggest the need for further investigation of this species as a gastrointestinal pathogen.
Human Campylobacter infection in temperate countries has been acknowledged as having a seasonal occurrence. Reports demonstrate a trough during winter months followed by a two- to threefold higher peak in late spring or early summer (Kovats et al., 2005). Our findings reflect this for the combined Campylobacter spp. as well as for C. jejuni/C. coli and C. ureolyticus individually. Campylobacter jejuni/C. coli and C. ureolyticus infections peak in spring. Figure 3 suggests a slightly earlier peak (March) for C. ureolyticus in comparison to C. jejuni (April and May).
A second epidemiological feature of Campylobacter infection is that incidence shows associations with patient age and gender (Crushell et al., 2004). European studies report (HSPC, 2007, EFSA, 2009) that in developed countries Campylobacter infection is most common in children under 5 years; however, a second peak is noted in the young adult, the 15–24 year age group, which is not observed in developing countries. In our study, Campylobacter was most frequently detected in the extremes of age, with more than half of all detections in children ≤5 years and adults ≥70 years. Campylobacter ureolyticus was shown in this study to be more common in female patients. Previously, Campylobacter infections have been reported as more common in male patients (Moore et al., 2005; EFSA, 2009). In addition, the peak which we report in the elderly (particularly for C. ureolyticus) accords with a recent study in the UK (Gillespie et al., 2009), which found that the elderly are the population most at risk of Campylobacter infection.
In conclusion, the results of our investigation for the period of January–December inclusive, 2009, in Southern Ireland, suggest that non-C. jejuni/C. coli spp. contribute considerably to the proportion of Campylobacter-positive samples from patients presenting with diarrhoeal illness. This finding illustrates the power of molecular-based detection over less sensitive culture techniques (O'Leary et al., 2009; Bullman et al., 2011). The resulting positive impact of increased sensitivity of diagnostics on clinical care is likely to be significant. Although Campylobacter-related gastroenteritis is usually self-limiting in the immunocompetent host, it imposes a high economic burden primarily due to hospitalization and man hours lost to industry. As Campylobacter infections are notifiable in Ireland, increased sensitivity of detection might be expected to heighten infection control efforts, in particular for organisms such as C. ureolyticus, whose reservoir is currently unknown.
We acknowledge the assistance of Mr John Murphy, Department of Biological Sciences, Cork Institute of Technology. S.B. is supported by a scholarship from the Irish Research Council for Science, Engineering and Technology (RS/2009/1670). R.D.S. is an ESCMID Fellow. Funding was provided by Serosep Ltd, Ireland. The authors declare that they have no competing interests.