This paper reports the findings from a preliminary study seeking to identify risk factors for sporadic human infection with shiga toxin-producing Escherichia coli (STEC) in South Australia. This phase of the study, conducted between February and September 2002, aimed to make recommendations regarding study methodology, and provided an opportunity to identify any potential risk factors for STEC infections in South Australia. The study design was a prospective age-matched case control study. A case was defined as a person with macroscopic or microscopic evidence of blood in a faecal specimen, and in which a gene associated with the production of shiga toxin (stx 1 or 2) was identified. Two community controls per case were randomly selected from the Social Environmental Risk Context Information System database. Eleven cases and 22 controls were enrolled in the pilot phase of the case control study. Cases were more likely than controls to have eaten berries, including strawberries, blueberries, and blackberries, in the 10 days preceding illness (Mantel Haenszel matched OR 11; 95 per cent CI 1.26-96.12). No other exposures were significantly associated with illness. Due to the small number of study participants, the power of the study was insufficient to expect any significant results. National participation will be vital to obtain sufficient cases in a realistic time, however this would necessitate more consistent ascertainment and reporting of STEC disease between the states and territories. Commun Dis Intell 2004;28:74-79.
Shiga toxin-producing Escherichia coli (STEC) have emerged over recent years as an important cause of gastroenteritis in humans worldwide. Initially acknowledged as a potential foodborne pathogen, the importance of non-food related transmission of STEC is now also being recognised.1,2 Absolute numbers of infections are low when compared to other enteric pathogens, however the severe and potentially life threatening illness caused by STEC has prompted the need for rapid expansion of knowledge about the organism and the disease it causes.3
Outbreaks of STEC in Australia are quite rare, with only three outbreaks being documented prior to 2002.4,5,6 Due to inconsistencies in ascertainment and reporting practices, STEC disease is likely to be largely under-reported in national surveillance data. Data for Australia suggests a yearly average of 47 cases for the past five years.7 Internationally, annual incidence rates vary by region and country and appear to be influenced by numerous local factors including seasonality, dietary habits, food management practice, and animal husbandry methods.3 Globally, incidence rates are almost impossible to interpret due to inconsistencies in surveillance and diagnostic techniques.
Internationally, risk factor information for STEC infection is mostly derived from outbreak investigations and descriptive epidemiological investigations. Studies of common-source outbreaks have identified numerous risk factors for STEC infection, including consumption of ground beef, salami, sliced meats, raw milk, yoghurt, cheese, sprouts, lettuce, and jerky.4,-11 However, it is not clear what role these or other risk factors play in sporadic disease. A small number of international case control studies have investigated risk factors specifically for sporadic STEC infection.12,-16 Risk factors identified in these studies included; consumption of beef burgers from commercial premises, consumption of sliced meats from caterers, consumption of fish, and recreational or work visits to farms.
To date, no epidemiological studies have been conducted in Australia to identify local risk factors for sporadic STEC disease. We conducted a case control study to investigate risk factors for sporadic human infection with shiga toxin-producing Escherichia coli in South Australia. This paper documents the pilot phase of the study that was conducted between February and September 2002. Current practices in South Australia facilitate the detection of STEC. Since 1997, all faecal specimens submitted to laboratories with microscopic or macroscopic evidence of blood have been polymerase chain reaction (PCR) tested for the presence of stx genes.
A prospective age-matched case control study was conducted in South Australia to investigate potential risk factors for STEC infection between February and September 2002. Cases and controls were recruited from the study population defined as the population of South Australia (approximately 1.5 million persons). A case was defined as a person with macroscopic or microscopic evidence of blood in a faecal specimen, and from which the genes associated with the production of shiga toxin (stx 1 or 2) was identified. The identification of cases in this study was facilitated by the presence of a reporting system, the South Australian Notifiable Diseases Surveillance System (NDSS).
Cases were excluded from the study if they were secondary cases within a household; identified as part of an outbreak; had travelled overseas in the month before onset of illness, or had no residential telephone number. Cases that had a mixed infection, such as cases that were positive for both Salmonella and STEC, were also excluded due to difficulties in establishing an accurate onset date for the STEC infection.
For each case, two controls were randomly selected from the Social Environmental Risk Context Information System (SERCIS). The main role of SERCIS is to conduct health surveys via a computer assisted telephone interview method. Through these surveys SERCIS also recruits participants for other more specific surveys or studies such as case control studies. The reliability of the SERCIS telephone health survey method has been demonstrated.17,18
Two controls were matched to each case by five-year age groups, except for those less than 12 months of age who were matched by 6-month age groups (for example, 0-5 months, 7-11 months, 1-4 years, 5-9 years, >65 years etc). The controls were not matched to cases on any other variables. Controls were excluded if they had a history of gastrointestinal illness in the 14-day period prior to the day of interview; reported overseas travel in the month prior to the day of interview; reported another household member with diagnosed STEC disease in the 10-day period before the day of interview; or had stx detected in a faecal specimen in the 10-day period before the day of interview.
Information on food and environmental exposures for the 10 days prior to the onset of illness for cases and prior to interview for controls was collected using a questionnaire administered by telephone. One hundred and twelve exposures were selected either because they were previously reported or suspected to be related to STEC infections, because they were food items or sources from which STEC had ever been detected, or they were hypotheses generated in a local case series investigation conducted in South Australia between 1994 and 1997 (Table 1). Details were also collected regarding illness, demography, occupation and household contacts. Data were analysed using Epi Info Version 6.04 software. Mantel Haenszel matched odds ratios were calculated for all dichotomous variables.
Table 1. Exposures examined as risk factors for shiga toxin-producing Escherichia coli infection, South Australia, 2002
|Food exposures||Environmental exposures||Travel exposures|
|Beef/beefburger/minced beef/roast beef/beef steak/beef shish kebab/shaslicks/beef sausages/silverside/corned beef||Regular attendance at any of the following institutions:
|Turkey/turkey roll||Occupational exposure to animals|
|Duck||Occupational exposure to raw meat|
|Fish or seafood||Household member with occupational exposure to animals|
|Salami/mettwurst/cabanossi||Household member with occupational exposure to raw meat|
|Jerky||Having a household pet|
|Devon||Having a sick household pet|
|Ham||Handling of raw red meat fed (either cooked or raw) to household pets|
Apples/pears/plums/peaches/ nectarines/apricots/strawberries/other berries/watermelon/rockmelon/honeydew melon
|Exposure to pet foods:
|Fruit salad||Handling of dried meat strips subsequently fed to pets|
Lettuce/cabbage/pre-packaged/bagged salad/ carrots/spring onion/shallots/radishes/celery/cucumber/alfalfa sprouts
|Living in a rural farm or remote property|
|Eggs (raw or cooked)||Visiting a farm|
|Unpastuerised dairy products||Touching animals|
|Consumption of the following types of water:
|Unpasteurised fruit or vegetable juices||Visiting zoos, petting zoos, wildlife parks|
|Handling of any of the following while preparing a meal
|Attending regional shows/fairs where there were animal displays|
|Involved in the slaughter, cutting up, packaging or wrapping of any raw red meat (not part of work)||Contact with fertiliser or compost made from animal faeces|
|Being given a sample of meat to eat in a butcher, supermarket or food store|
In total, there were 11 cases that were enrolled between February and September 2002. The cases were identified from laboratory notifications received by the South Australian Communicable Disease Control Branch that fitted the case definition. Twenty-two age-matched controls were selected from the SERCIS database and enrolled in the study. The median age of the cases was 75 years (age range: 4–90 years) compared with 66.5 years (age range: 2–77 years) for the controls. There was an over representation of females in the control group (59% female) compared with the cases (45% female).
Prevalence of symptoms among the cases varied (Table 2). All cases reported experiencing diarrhoea, with only nine self-reporting that it was visibly bloody. Most cases experienced abdominal pain (n=9), however fewer reported nausea (n=4), vomiting (n=3), headache (n=4), and/or fever (n=4). One case reported constipation prior to diarrhoea onset but subsequent to the onset of other symptoms. One case reported rigors. Five of the 11 cases were hospitalised, and almost all cases sought medical advice from a general practitioner or medical centre (n=10). Three cases reported missing between one and six days of work, recreation time, or childcare.
Table 2. Case illness profile (N=11)
|Symptoms||Number of cases|
* Self reported by cases as blood being visible in faeces.
The consumption of 'berries' in the 10 days before the onset of illness was associated with STEC infection (Mantel Haenszel matched OR 11.00, 95% CI 1.26–96.12). Six of the 11 cases reported eating strawberries, blackberries, or blueberries. The association between STEC and eating strawberries was not significant although the OR was elevated (Mantel Haenszel matched OR 5.00, 95% CI 0.97–25.77). Of the five who reported eating strawberries, two reported obtaining these from supermarkets, while the other three were obtained directly from the farm, of which only one was a commercial berry farm. The case that reported eating blackberries obtained them from his own farm, and the case that reported eating blueberries obtained them from a commercial berry farm.
Table 3. Selected food exposure frequencies and matched odds ratios in shiga toxin-producing Escherichia coli cases and controls
|Exposure||Exposed (cases)||Exposed (controls)||Matched OR||95% CI|
|Berries (including strawberries, blackberries and blueberries)||6/11||2/22||11.00||1.26–96.12|
|Delicatessen sliced meats||7/11||15/22||0.80||0.16–3.91|
|Being given a sample of meat to eat in a supermarket or food store||2/11||1/22||4.00||0.36–44.11|
|Eating at a national hamburger chain||4/11||3/22||6.00||0.56–64.71|
|Involved in the slaughter, wrapping, or packaging of large quantities of raw red meat*||4/11||2/22||3.50||0.62–19.89|
* If a child, immediately present at the time of slaughter, wrapping, or packaging of large quantities of raw red meat.
No other food items were significantly associated with STEC infection. Despite this, there were several that had elevated ORs in the matched analysis (Table 3). No environmental exposures were associated with illness. There was no association with travel interstate, intrastate or camping. Many of the ORs obtained in the matched analysis were undefined because the Mantel Haenszel numerator or denominator was zero. This was not unexpected due to the small numbers of cases and controls available for the study.
The consumption of berries, including strawberries, blackberries and blueberries, in the exposure period was the only exposure significantly associated with sporadic STEC infection in South Australia. No other food or environmental exposures were significantly associated with STEC infection. Some risk factor exposures did have elevated odds ratios, but their association with illness was not significant. Additionally, due to the small study numbers and zero cell counts, many of the odds ratios were undefined, which restricted the interpretation of the results.
The association between sporadic STEC disease and the consumption of berries though unexpected is a possible vehicle of infection. Outbreaks of STEC and other enteric pathogens due to the consumption of fresh produce have been documented illustrating that it is a biologically plausible route of infection.19,20,21 Implicated raw produce have included radish sprouts, lettuce, alfalfa sprouts, unpasteurised apple juice and apple cider.5,22,23
This is the first reported study to examine risk factors for sporadic infection with STEC conducted in Australia. Conducted as a pilot over a 6-month period, it was not anticipated that any significant results would be obtained from this study. This was due to the limited number of cases of STEC disease that are notified in South Australia, and the requirement for a large study sample to facilitate the detection of any effect. Additionally, it is likely that any association between a risk factor exposure and illness would be small. Given the resource and time limitations, the study was successful in meeting stated objectives. National participation in this study will be vital to identify more reliably the risk factors for STEC in Australia. It would reduce the time frame for the study due to an increased number of cases available for potential inclusion in the study. National participation would require more consistent ascertainment and reporting of STEC disease between the states and territories. The benefits of a national study would include an increased understanding of risk factors for STEC disease, which would enhance the national capability to facilitate the development of targeted prevention and control strategies.
The authors acknowledge OzFoodNet for funding the conduct of the pilot study in South Australia.
1. Trevena WB, Willshaw GA, Cheasty T, Domingue G, Wray C. Transmission of vero cytotoxin producing Escherichia coli O157 infection from farm animals to humans in Cornwall and west Devon. Commun Dis Public Health 1999;2:263-268.
2. Locking ME, O'Brien SJ, Reilly WJ, Wright EM, Campbell DM, Coia JE, et al. Risk factors for sporadic cases of Escherichia coli O157 infection: the importance of contact with animal excreta. Epidemiol Infect 2001;127:215-220.
3. Kerr KG. Infections associated with shiga toxin-producing Escherichia coli: epidemiology, pathogenesis, diagnosis and management. Infect Dis Rev 1999;1:9-14.
4. Clark A, Morton S, Wright P, et al. A community outbreak of vero cytotoxin producing Escherichia coli O157 infection linked to a small dairy farm. Commun Dis Rep 1997;7:R206-R211.
5. Cody SH, Glynn MK, Farrar JA, Cairns KL, Griffin PM, Kobayask J, et al. An outbreak of Escherichia coli 0157:H7 from unpasteurised commercial apple juice. Ann Intern Med 1999;130:202-209.
6. McDonnell R, Rampling A, Crook S, Cockcroft PM, Wilshaw GA, Cheasty T, et al. An outbreak of vero cytotoxin producing Escherichia coli O157:H7 infection associated with takeaway sandwiches. Commun Dis Rep 1997;7:R201-R205.
7. Communicable Diseases Network Australia. National Notifiable Diseases Surveillance System. Notifications of STEC/VTEC, Australia, 1998-2002. Canberra; 2003.
8. Cases of Escherichia coli 0157 associated with unpasteurised cream. CDR Weekly 1998;8:377
9. Outbreak of vero cytotoxin producing Escherichia coli O157 infection in North Cumbria. CDR Weekly 1999;9:97-98
10. Ackers ML, Mahon B, Leahy E, Goode B, Damrow T, Hayes PS, et al. An outbreak of Escherichia coli O157:H7 associated with lettuce leaf consumption. J Infect Dis 1997;177:1588-1593
11. Willshaw G, Thirlwell J, Jones AP, Parry S, Salmon RL, Hickey M. Vero cytotoxin-producing Escherichia coli O157 in beefburgers linked to an outbreak of diarrhoea, haemorrhagic colitis and hemolytic uremic syndrome in Britain. Lett Appl Microbiol 1994;19:304-307.
12. Parry SM, Salmon RL, Wilshaw GA, Cheasty T. Risk factors for and preventing sporadic infections with vero cytotoxin (shiga toxin)-producing Escherichia coli 0157. Lancet 1998;351:1019-1022.
13. Pierard D, Crowcroft N, De Bock S, Potters D, Crabbe G, Van Loock F, et al. A case control study of sporadic infection with 0157 and non-0157 vero cytotoxin-producing Escherichia coli.Epidemiol Infect 1999;122:359-365.
14. Slutsker L, Ries AA, Maloney K, Wells, JG, Greene KD, Griffin PM. A nationwide case control study of Escherichia coli 0157:H7 infection in the United States. J Infect Dis 1998;177:962-966.
15. O'Brien S, Adak GK, Gilham C. Contact with farming environment as a major risk factor for shiga toxin (vero cytotoxin)-producing Escherichia coli 0157 infection in humans. Emerg Infect Dis 2001;7:1049-1051.
16. Mead P, Griffin PM. Escherichia coli 0157:H7.Lancet 1998;352:1207-1212.
17. Taylor A, Wilson D, Wakefield M. Differences in health estimates using telephone and door-to-door survey methods-a hypothetical exercise. Aust N Z J Public Health 1998;22:223-226.
18. Starr G, Dal Grande E, Taylor A, Wilson D. Reliability of self-reported behavioural health risk factors in a South Australian telephone survey. Aust N Z J Public Health 1999;23:528-530.
19. Ooi PL, Goh KT, Neo KS, Ngan CC. A shipyard outbreak of salmonellosis traced to contaminated fruits and vegetables. Ann Acad Med Singapore 1997;26:539-543.
20. Centers for Disease Control and Prevention. Multistate outbreak of Salmonellapoona infections -United States and Canada, 1991. MMWR Morb Mortal Wkly Rep 1991;40:549-552.
21. Cooks KA, Dobbs TE, Hlady WG, Wells JG, Barrett TJ, Puhr ND, et al. Outbreak of Salmonella serotype Hartford infections associated with unpasteurised orange juice. JAMA 1998;280:1504-1509.
22. Besser RE, Lett SM, Weber JT, Doyle MP, Barrett TJ, Wells JG, et al. An outbreak of diarrhoea and haemolytic uremic syndrome from Escherichia coli 0157:H7 in fresh pressed apple cider. JAMA 1993;269:2212-2220.
23. Centers for Disease Control and Prevention. Escherichia coli O157:H7 outbreak linked to commercially distributed dry-cured salami-Washington and California, 1994. MMWR Morb Mortal Wkly Rep 1995;44:157-160
1. Master of Applied Epidemiology Program, National Centre for Epidemiology and Population Health, Australian National University, Canberra
2. Communicable Disease Control Branch, Department of Human Services, Adelaide, South Australia
Corresponding author: Ms Rebecca Hundy, Epidemiologist, Australian Government Department of Health and Ageing, GPO Box 9848 (MDP6), Canberra ACT 2601. Telephone: +61 2 6289 4539. Facsimile: +61 2 6125 0740. Email: Rebecca.Hundy@health.gov.au
This article was published in Communicable Diseases Intelligence, Volume 28 No 1, March 2004.