| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Environmental and Enteric Diseases Department, Health Protection Agency (HPA) Centre for Infections, 61 Colindale Avenue, London NW9 5EQ, UK
2 Division of Medicine and Neuroscience, Manchester University, Clinical Sciences Building, Hope Hospital, Stott Lane, Salford M6 8HD, UK
3 Welsh Assembly, Cardiff CF99 1NA, UK
4 HPA Yorkshire and the Humber Regional Microbiology, Bridle Path, York Road, Leeds LS15 7TR, UK
5 Division of Epidemiology and Public Health, School of Community Health Sciences, University of Nottingham Medical School, Nottingham NG7 2UH, UK
6 HPA North West, Rooms 103–112, First Floor, DBH House, 105 Boundary Street, Liverpool L5 9YJ, UK
7 St George's Hospital Medical School, University of London, Cranmer Terrace, London SW17 0RE, UK
Correspondence
Iain A. Gillespie
Iain.Gillespie{at}hpa.org.uk
Received 18 November 2005
Accepted 30 January 2006
Abbreviations: CI, confidence interval; HPA, Health Protection Agency; OR, odds ratio.
The Campylobacter Sentinel Surveillance Scheme Collaborators are public health, environmental health and laboratory staff who serve the populations of the following health authorities: Birmingham, Bradford, Bro Taf, Bury and Rochdale, Dyfed Powys, East Kent, Barnet, Enfield and Haringey, Herefordshire, Leeds, Leicestershire, Manchester, North Cumbria, North Essex, North West Lancashire, Nottingham, Salford and Trafford, South and West Devon, South Lancashire, Southampton and South West Hampshire, Stockport, West Pennine, Wigan and Bolton. In association with: HPA Laboratory of Enteric Pathogens, Campylobacter Reference Unit; HPA Centre for Infections, Environmental and Enteric Diseases Department; HPA Local and Regional Services; HPA Statistics Unit. 
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Infection with Campylobacter jejuni most frequently presents as acute enteritis. Classic medical textbook descriptions of Campylobacter enteritis list the commonly reported symptoms as diarrhoea, malaise, fever and abdominal pain (Blaser, 2000; Skirrow, 1996). The spectrum of diarrhoea ranges from loose stools through profuse watery diarrhoea to frankly bloody stools (Blaser, 2000). Tissue injury can occur along the bowel from jejunum to colon, and gross pathological examination of the gut in severe cases reveals a diffuse, bloody, oedematous and exudative enteritis (Blaser, 2000; King, 1962). Severe disease may be clinically, sigmoidoscopically and histologically difficult to differentiate from ulcerative colitis, and Campylobacter infection forms part of the differential diagnosis for inflammatory bowel disease (Lambert et al., 1979). Abdominal pain may be so severe that Campylobacter enteritis can be confused with acute appendicitis (Blaser et al., 1979; Lambert et al., 1979).
In the first year of a population-based sentinel surveillance study of Campylobacter infection in England and Wales, diarrhoea (95 %), abdominal pain (85 %) and fever (78 %) were the most commonly reported symptoms, whilst vomiting (35 %) and bloody diarrhoea (27 %) were reported less frequently (Communicable Disease Surveillance Centre, 2001). To investigate potential host, micro-organism or environmental factors that might explain the different clinical presentations, we compared the features of C. jejuni cases presenting with vomiting and/or bloody diarrhoea with cases who did not report either clinical manifestation.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Statistical analysis, performed using Stata statistical software release 8.0 (StataCorp, College Station, TX), was restricted to cases of C. jejuni infection who had not travelled abroad in the 2 weeks before illness to control for the confounding effect of foreign travel. Cases who were infected with more than one Campylobacter subtype [as defined by serotyping (Frost et al., 1998), phage-typing (Frost et al., 1999) and/or antimicrobial resistance pattern (Thwaites & Frost, 1999)] were excluded, as were cases whose specimen dates were either prior to, or greater than 31 days from, their onset date.
The date of illness onset was used to define the season in which illness commenced. Season was coded to compare cases with an onset in spring (March to May) with those who were ill in summer (June to August), autumn (September to November) and winter (December to February). Patients' description of their ethnic origin was classified according to the UK census 2001 (Office for National Statistics, 2001). Age values were recoded to compare infants (<1 year of age) with toddlers (1–4 years), young children (5–9 years), older children and teenagers (10–17 years), adults (18–64 years) and the elderly (
65 years). Food exposures were coded to compare cases who had eaten a food once or more than once in the exposure period (2 weeks prior to the onset of symptoms) with those who had not. Contact with raw meat was coded to compare no contact with once, two to five times, six to ten times and more than eleven times. Daily consumption of unboiled tap water was recoded to compare zero consumption with one to four, five to nine and 10 glasses drunk. Binary variables were created to compare the 10 most commonly identified serotypes of C. jejuni with other known serotypes. Individuals with missing information for any of the variables of interest were omitted from the analyses using those variables.
For the case–case comparison, cases of C. jejuni infection reporting vomiting and/or bloody diarrhoea were considered ‘cases’, whilst those who did not report either symptom were considered ‘controls’. It is important to note that symptoms were self-reported. The demographic and clinical profiles of cases and controls were compared using Pearson's chi square test and Student's t test. Differences in exposure were compared initially using simple logistic regression models. Point estimates and 95 % confidence intervals (CIs) for the Mantel–Haenszel odds ratio (OR) were calculated for each explanatory variable whilst controlling for the effect of age.
Proximal (Victora et al., 1997) exposures with a P value of less than 0.1 were then included in a larger model to obtain maximum-likelihood estimates of the effect of exposures on the outcome of interest, whilst controlling for confounding. The distal (Victora et al., 1997) exposures age, gender, season and ethnicity were included and retained in the model throughout. The significance of exposures was tested using the likelihood ratio (LR) test, and the model was simplified accordingly.
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Amongst those cases who reported their foreign travel status (11 648; 98 %), 2261 cases (19 %) had travelled abroad in the 2 weeks before illness. Cases who reported foreign travel or who did not report their foreign travel status were excluded, as were cases (51; 0.44 %) infected with more than one subtype, and 397 cases (3 %) for whom the recorded specimen date preceded the reported onset date (n=69) or exceeded it by >31 days (n=328). This left 9350 UK-acquired cases of C. jejuni infection for analysis.
Diarrhoea, abdominal pain and fever were reported by 9056 (96.9 %), 8114 (86.8 %) and 7440 (79.6 %) cases, respectively. Vomiting was reported by 3346 cases (35.8 %) and bloody diarrhoea by 2661 cases (28.5 %), and cases who reported one of these symptoms were more likely to report the other (
2 138.19; P<0.001). In total, 4043 cases (43.2 %) reported one or both clinical presentations. These cases were compared with the 3335 cases (35.7 %) who reported no vomiting or bloody diarrhoea. A further 1972 cases (21.1 %) who did not respond to the questions of vomiting and/or bloody diarrhoea were excluded from further analysis. Cases who reported vomiting and/or bloody diarrhoea experienced a longer illness than cases who did not report either symptom (mean 11.8 versus 10.9 days; t test P=0.007) and were more likely to be admitted to hospital (11.8 % versus 5.1 %; 2 99.5; P<0.001).
Single variable analysis
Self-reported vomiting and/or bloody diarrhoea amongst cases of C. jejuni infection decreased with age (Table 1
). Cases who reported vomiting and/or bloody diarrhoea were more likely to be female and were more likely to report the consumption of barbecued foods, pre-packed sandwiches, sausages, and foods eaten in restaurants. They were more likely to have reported a number of water exposures and to have been exposed to pet cats. They were less likely to report the consumption of fish and shellfish, cheese and salad. Reported vomiting and/or bloody diarrhoea appeared to differ depending on infecting serotype.
|
). The likelihood of reported vomiting and/or bloody diarrhoea increased with reported engineering work or problems with drinking water supply and with increasing daily consumption of unboiled tap water. Cases infected with serotype HS37 were less likely to report bloody diarrhoea than those cases infected with other serotypes.
|
Firstly, cases included in this study were those whose infections were confirmed using microbiological methods. These cases are likely to represent the more severe end of the clinical spectrum (Tam et al., 2003). Secondly, exposure data were self-reported over a 2-week period prior to onset, increasing the likelihood of reporting bias. There is no reason to believe, however, that these biases operate differently in those who were included as ‘cases' in the analysis compared with ‘controls’. Thirdly, it has been suggested that co-infections might provide an explanation for the observed vomiting and/or bloody diarrhoea. Data on co-infections were unavailable and a degree of uncontrolled confounding might therefore have occurred. However, in general, low levels of co-infection for campylobacters are reported in the literature (Blaser et al., 1983; The Food Standards Agency, 2000), so this is unlikely to have had a major effect on the findings. Finally, it is possible that additional confounding factors not included in our analysis affected our results. However, we have captured standardized information on a wide range of exposures for a large number of cases of C. jejuni infection, so the likelihood of this taking place has been minimized.
We elected to consider symptoms of vomiting and bloody diarrhoea together. Our a priori hypothesis was that vomiting and bloody diarrhoea did not share a common aetiology. However, the fact that cases who reported one of the symptoms were more likely to report the other suggested that this was not the case. Furthermore, initial analyses (not shown) examining each symptom separately produced very similar findings, reinforcing the need for a combined analysis.
Reported vomiting and/or bloody diarrhoea was strongly related to the age of the case. This might relate to host susceptibility or might represent ascertainment bias. It is possible that the relatively immature colonic flora (Stark & Lee, 1982) and intestinal immune system (Davies, 1988) of the infant might lead to a more severe infection and hence clinical presentation (Blaser, 2000). Alternatively, these are alarming symptoms that parents might be more likely to report on behalf of their children. The observed inverse ‘dose-response’ relationship with increasing age group, however, suggests perhaps that the effect is a genuine one.
The direction of the association amongst the food variables independently associated with vomiting and/or bloody diarrhoea warrants comment. Whether or not these findings represent different levels of contamination with C. jejuni is uncertain, but since it is possible that some foods are more highly contaminated than others, this might help to explain these observations (i.e. it is not the food items per se, but rather the contamination levels that are important).
The direction of association with the water-exposure variables and in particular the dose-response relationship between drinking unboiled tap water and presenting with vomiting and/or bloody diarrhoea were intriguing. Case–case comparisons, as used in this study, should not be used to make statements about the magnitude or direction of population risk (McCarthy & Giesecke, 1999; Gillespie et al., 2002). Nevertheless, finding a dose-response relationship adds weight to an association between an exposure and disease (Hill, 1965), suggesting that there might be a real effect on clinical presentation. We cannot be absolutely certain, however, whether increased consumption of unboiled tap water was a cause of infection (and hence symptoms) or a consequence of it.
Outbreaks of Campylobacter infection linked with municipal water supplies have been reported in the past (Communicable Disease Surveillance Centre, 2000a; Melby et al., 1991; Mentzing, 1981; Palmer et al., 1983; Vogt et al., 1982). In all of these studies, either the water supply was unchlorinated (Melby et al., 1991; Mentzing, 1981) or a serious challenge to a chlorinated supply occurred prior to the outbreak (Communicable Disease Surveillance Centre, 2000a; Palmer et al., 1983; Vogt et al., 1982). Case-control studies of sporadic infection have failed to identify municipal mains water as a source of infection (Adak et al., 1995; Eberhart-Phillips et al., 1997; Kapperud et al., 1992; Neal & Slack, 1997; Rodrigues et al., 2001). This might represent a negligible/zero risk, or might relate to the high prevalence of exposure to drinking mains water amongst the general population in many countries, making an effect difficult to detect.
Under current UK regulations, drinking water supplied for domestic purposes or to food production premises should be regarded as wholesome if it does not contain any micro-organisms or parasites at a concentration or value that would constitute a potential danger to human health (Department of the Environment, Transport and the Regions, 2000). Indicator organisms (historically coliforms and Escherichia coli, but current legislation includes enterococci) are used to ensure this quality and to assess the effectiveness of water treatment (especially disinfection). However, indicator organisms such as E. coli might not be an adequate indicator for the presence of campylobacters in water (Environment Agency, 2002), and statutory testing for campylobacters is not undertaken.
There was only one (inverse) association between infecting serotype and presentation with vomiting and/or bloody diarrhoea in this study. This suggests, perhaps, that there is no fundamental relationship between pathogenicity and expression of surface antigens detected by phenotyping (Frost et al., 1998). Further work on the genetic basis of pathogenesis could be undertaken by the molecular subtyping of the isolates referred in this study.
Conclusions
Vomiting and/or bloody diarrhoea characterized the more severe end of the disease spectrum, as evidenced by a longer illness and increased hospital admissions. Host susceptibility might be important, as witnessed by strongly age-related symptom presentation. Infective dose might also be a factor. The role of unboiled tap water as a potential source of C. jejuni infection in England and Wales requires further investigation.
|
ACKNOWLEDGEMENTS |
|---|
Membership of the Campylobacter Sentinel Surveillance Scheme Steering Committee:
Mr A. Charlett (Head, Statistics, Modelling and Bioinformatics, HPA Centre for Infections); Dr J. M. Cowden (Consultant Epidemiologist, Health Protection Scotland); Mrs J. A. Frost (Welsh Assembly, Cardiff); Mr I. A. Gillespie (Clinical Scientist, Environmental and Enteric Diseases Department, HPA Centre for Infections); Ms J. Millward (Principal Environmental Health Officer, Birmingham City Council); Dr K. R. Neal, (Clinical Reader in the Epidemiology of Communicable Diseases, Division of Epidemiology and Public Health, University of Nottingham); Professor S. J. O'Brien (Division of Medicine and Neuroscience, Manchester University); Dr M. J. Painter (Consultant in Communicable Disease Control, Manchester Health Authority); Professor Q. Syed (Regional Director, HPA North West); Dr D. Tompkins (Regional Microbiologist, HPA Yorkshire and the Humber).
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
