Verocytotoxin-producing E.coli food poisoning and its prevention

The Institute of Food Science & Technology, through its Public Affairs and Technical & Legislative Committees has authorised the following Information Statement, dated November 2004, prepared by its Professional Food Microbiology Group, and replacing the version dated 14 September 1996.

Summary

Foodborne illness caused by verocytotoxin-producing E.coli (VTEC) – sometimes referred to as enterohaemorrhagic E.coli (EHEC) – was first recognised in the early 1980s. Although the illness does not appear to be very common, VTEC is now regarded as an important pathogen because serious complications may follow infection. The O157:H7 serotype is the predominant cause in the UK and USA of human infections but other serotypes have also been implicated. In comparison with, for example, Salmonella, numbers of cases appear to be low but they seem to be increasing. Infection may produce a mild diarrhoea, or a severe or fatal illness. The infective dose may be very low.

The main reservoir for VTEC is the bovine intestine. More than half of all the cases in the UK follow environmental transmission, which is particularly important in respect of sporadic cases. Food vectors linked to outbreaks include milk and milk products and ground beef. In recent years apple juice and sprouting seeds have become more frequently implicated particularly in the USA. Water has been responsible for many of the largest outbreaks.

Control of VTEC illness in humans requires good slaughterhouse hygiene and heat treatment of raw meat and milk. VTEC is destroyed by heat; adequate cooking of meat (70ħC for 2 minutes internal temperature) and pasteurisation of milk will protect consumers from infection from these sources. It is essential to provide hygienic food handling and good chilled storage conditions to ensure that other foods do not become contaminated.

For most foods, testing for VTEC is not an effective control strategy and the setting of end product specifications may not be appropriate. However, it is noted that in USA, the Department of Agriculture (USDA) has adopted precisely that regulatory strategy for ground beef. Contamination rates in suspect foods are thought to be low, so the chance of isolating the bacteria from a few samples of a batch of food is small. Also, there is little agreement on the best method to use for routine testing. The widely used standard methods for detection and confirmation of E.coli are not appropriate as many VTEC strains grow poorly or not at all at 44ħC. However, extensive research on isolation techniques is being carried out and various commercial kits have become available recently.

The Microbiological Safety of Food Funders Group has published an updated Report on UK publicly funded research relating to VTEC, covering the period from 1999 to 2003.

In the Strategic Plan (2005 to 2010) of the UK Food Standards Agency (FSA) one of the key aims is to continue to reduce foodborne illness in the UK. On 18 January 2005 in London the Agency will conduct a review of the VTEC research programme.

Background

Escherichia coli is a common organism found in the lower intestinal tract of man and animals. Like some other potential pathogens, as a result of contamination via faeces or sewage, it is also readily found in damp, ambient temperature environments e.g. soil, vegetation, moist or wet areas in factories, untreated water. There are many types and strains of E.coli, a few of which are potentially pathogenic. Various strains may cause illness by a variety of infective and toxin-producing mechanisms. Symptoms vary according to the strain of E.coli encountered, and the resistance of the individual to such illnesses. Infants, young children, elderly and sick people are generally more susceptible to E.coli (and other) infections than healthy older children and adults.

VTECs produce one or more verocytotoxins (VT) which are similar to a toxin produced by Shigella dysenteriae type 1. The pathogenicity of these strains is known to involve adhesion to and colonisation of the intestinal tract and the production of powerful toxins which act on the colon. Clinical symptoms range from mild diarrhoea to severe bloody diarrhoea (haemorrhagic colitis), and in some sufferers include haemolytic uraemic syndrome (HUS) and kidney failure, (which can be fatal); although a few have no diarrhoea. In some outbreaks a high proportion of cases develop HUS which is the most common cause of renal failure in children in the UK. In adults, rarely, the infection may progress to cause thrombotic thrombocytopaenic purpura (TTP), a condition like HUS but with neurological complications. Most VTEC infections in the UK and USA arise from strains of serotype Ol57:H7, but other serotypes can produce VT and some have been shown to cause illness. Routine testing only looks for serotype O157:H7, which may account for its apparent predominance.

Incidence

VTEC illness was first recognised in 1982 and since then infections have been reported from more than 30 countries on 6 continents. In 1996 there was an outbreak of gastrointestinal infections with E. coli O157 in Scotland during which 17 people died and hundreds of people became very ill. Nearly 11% of the patients with the outbreak had a diagnosis of HUS which causes kidney failure and TTP which can lead to kidney failure but also affects the brain and the heart (Dundas, et al, 1999). This outbreak was investigated by the Pennington Group (1997) which proposed far-reaching recommendations. These, inter alia. gave rise to the licensing of butchers¡¦ shops and a Hazard Analysis Critical Control Points (HACCP) training programmes for butchers (mandatory for those handling raw meat and ready-to-eat foods). The outbreak was also the subject of a Fatal Accident Inquiry. The FSA Scotland and the Scottish Executive set up a Task Force, which reported in June 2001 with 105 recommendations in areas as diverse as waste recycling, access to the countryside, diagnosis and patient care by health professionals, and outbreak control. In response to the Task Force Report they have established an Action Plan.

In England and Wales, the provisional number of laboratory reports of E.coli O157 fell from a peak of 1084 in 1999 to 595 for the year 2002 but rose to 675 (provisional figure) in 2003. Six outbreaks were reported in 2002, none involving more than 20 people. Only one was thought to be foodborne. In Scotland, from a peak of 506 cases in 1996, there were 229 cases in 2002 and (provisional figure) 148 cases in 2003. In September 2003 an outbreak occurred affecting nine children (and subsequently three relatives) in a nursery in Antrim, Northern Ireland.

The importance of non-food sources of infection is now fully recognised. For example, in addition to waterborne instances referred to later, outbreaks have been associated with visits to open farms and ¡§petting zoos¡¨ (and in the USA, county fairs) where people, and especially children, come into contact with farm animals. In USA there was evidence of airborne transmission. At least 19 people who had gone to a county fair in Ohio in 2001 fell ill with E.coli which apparently spread through sawdust in the air at an exhibition hall, the first time researchers have connected an outbreak to a contaminated building. Testing at the building found E.coli O157 in the rafters, the walls, and the sawdust in some cases 10 months after the fair (Varma et al, 2003; ProMED, 2003).

E.coli O157 is a world wide threat to public health. It is estimated that about 75,000 cases of E.coli O157 occur annually in USA (Perna et al., 2001) with an estimated 2100 cases (2.8%) requiring hospitalisation. However it is less commonly reported in patients in less industrialised countries.

Food Safety

The foods considered most likely to be contaminated are raw meats, particularly beef, and raw milk. There is a need for more information on the incidence of E.coli O157 in cattle and meats.

Cattle are the main source of infection. E.coli O157 has been found in faeces of healthy livestock including sheep, pigs and goats but the organism is not associated with disease in these animals (Zoonoses Report, UK 2002). In 1999 a survey of cattle, sheep and pigs in Great Britain (GB) sent for slaughter for human consumption showed that the faecal carriage rate of E.coli O157 was 4.7% in cattle (n=3939), 1.7% in sheep (n=4171) and 0.3% in pigs (n=2509). In a study funded by the Scottish Executive Environment and Rural Affairs Department (SERAD) for the determination of the prevalence of E.coli O157 in finishing cattle in Scotland, it was reported that 7.9% of cattle and 22.8% herds were shedding the organism. Shedding rates were higher in housed animals and highest in spring (Zoonoses Report, UK 2000). A further similar study in cattle, sheep and pigs arriving for slaughter in GB was commissioned in 2002 to start in 2003 (Zoonoses Report, UK 2002).

Studies have shown considerably different isolation rates, partly because the material studied and methods used have been different. It is thought that the source of the contamination is usually faecal material transferred onto carcasses at slaughter, or into milk at milking. In an 18-month study of the microbiological quality of cow¡¦s milk in the UK, E. coli 0157 was detected in only one of 602 raw milk samples (0.2%) and was not found in any of the 1393 samples of pasteurized milk. In spite of these findings, the consumption of raw milk, cream and cheeses made from raw milk is strongly discouraged because they are a potential source of VTEC infections as well as other pathogens. Raw fresh vegetables and fruit have been contaminated by manure. Other foods have been found to be contaminated in some outbreaks, e.g. lettuce, cheese, sandwiches, unpasteurised apple juice. Waterborne outbreaks have occurred in USA, Scotland and Africa either because clean water was not available (or had become contaminated with run-off from cattle pastures) or perhaps because of chlorination failure. Recreational water has also caused outbreaks. Two recent UK examples (September 2004) are an outbreak associated with a local stream in North Cornwall probably caused by cattle manure carried in by flash floods, and an outbreak associated with a swimming pool in Stretford Leisure Centre, Manchester.

The organism is heat-sensitive and should be destroyed by the same temperature that is recommended to eliminate Salmonella and Listeria. The advice in the UK is that minced beef and minced beef products, including beefburgers, should be cooked to a minimum internal temperature of 70¢XC for 2 minutes or equivalent. Industry should provide cooking instructions with burgers to ensure that they are adequately cooked, that the meat juices run clear and there are no pink areas inside cooked products. In USA the Food & Drug Administration (FDA) now recommend that ground beef products should be cooked so that all parts of the food are heated to at least 68.3¢XC for a minimum of 15 seconds. Pasteurisation of milk also effectively eliminates VTEC: at 72¢XC for 16.2 seconds more than 104 cells/ml will be killed. Several of the recorded outbreaks of VTEC illness are likely to have been caused by undercooking of beefburgers and similar products.

The minimum pH for growth is considered to be pH 4.5, but some strains of the organism can survive in low pH products such as mayonnaise and yoghurt, particularly in refrigerated storage¡¦ where survival can be for several weeks. More work is needed to assess the importance of these observations. Refrigeration below 5¢XC is thought to prevent growth of VTEC and is an important hygiene measure. However, any organisms present are likely to survive at these temperatures perhaps for several weeks.

Outbreaks may also be caused by cross-contamination of ready-to-eat foods from raw foods or dirty utensils and by post-pasteurisation contamination of milk. Person-to-person spread also occurs and has caused outbreaks in hospitals, day care centres, infant schools and nursing homes. It has not been possible to pinpoint the source of infection in many sporadic cases and small outbreaks. Normal good manufacturing/catering practices should ensure that the chance of cross-contamination occurring is minimised. The measures needed to protect consumers from VTEC are the same as those needed to protect against Salmonella, Campylobacter, Listeria and most other non-spore-forming foodborne pathogens.

Food handlers suffering from E.coli O157 infection must be excluded from work until two negative faecal specimens taken at intervals of not less than 48 hours have been obtained. (The Food Safety (General Food Hygiene) Regulations, 1995)

Quality Assurance

Quality assurance programmes in slaughterhouses should stress the need to minimise faecal contamination of carcases and to chill meat rapidly. Measures must be taken in slaughterhouses to minimise faecal contamination of carcases.

In July 2003 researchers at the Institute of Agriculture and Natural Resources, University of Nebraska, announced the results of a two-year study in which a new vaccine and a beneficial bacterial feed additive each significantly reduced E.coli O157:H7 in feedlot cattle.

One of the slaughterhouse measures is an assessment by a veterinary surgeon of the fleece/hide cleanliness of animals arriving at the slaughterhouse. Animals are graded from 1 to 5 with the highest score indicating high faecal contamination. The surgeon has to decide if the animals are to be rejected, cleaned for re-submission to ante mortem inspection or particular attention paid to hygiene procedures during processing.

The USDA Agricultural Research Service has developed and tested a commercial faecal detection system capable of scanning an entire beef carcase to reveal trace levels of contamination which are invisible to the human eye. It can be used prior to and after trimming, and after carcase-cleaning treatments, which include high-temperature steam from steam vacuums or steam cabinets (USDA ARS, 2002).

Elsewhere in the food industry, procedures to ensure that incoming food materials and ingredients are of good quality should be in place and should be adhered to. Screening of raw meats for VTEC is not an effective control mechanism because isolation rates from raw beef are low. Since the organism has been found in the faeces of a small proportion of healthy cattle, it is currently unlikely that it can be eliminated at source. Simple and reliable methods suitable for routine VTEC detection in foods have not been widely available. Similarly, because of the low contamination rate of VTEC in other foods, routine screening specifically for this organism is considered unlikely to be successful. The value of screening raw meat is being debated widely. Undoubtedly screening will detect some contaminated material and this can then be designated to a secure heat treatment process. But as screening can never detect all contaminated lots it is a poor control procedure. Nevertheless in USA, the USDA has adopted exactly that regulatory strategy for ground beef, with consequent huge recalls (e.g. Hudson Foods in 1997, ConAgra Foods in 2002) (USDA FSIS Final Rule, 1999).

In 1999 the USDA approved the use of irradiation to treat ground beef in order to inactivate pathogenic bacteria, particularly VTEC. While the use of this technique would improve the safety of ground beef, adequate cooking of meat remains the most practical and sure way of eliminating the danger of VTEC infection from this source.

In food manufacture and processing, quality assurance of raw and in-process materials, finished products and the manufacturing environment should be based on needs identified by a HACCP evaluation, and the end product specification. Following a series of outbreaks of E.coli O157 associated with unpasteurised apple juice, culminating in the large Odwalla outbreak, the US Food and Drug Administration (FDA) in 1998 introduced a rule for warning labelling of unpasteurised juice, and in 2001 introducing a rule (finalised in 2004) requiring that juice processors must use FDA specified HACCP principles for juice processing in accordance with 21. CFR Part 120 and Health Canada introduced similar requirements in August 2000. The presumed problem with apple juice arises from use of faecally-contaminated fallen apples in orchards where livestock have been grazing.

Monitoring trends of indicator organisms, e.g. Enterobacteriaceae and Standard Plate Count should indicate deviations from quality standards A general method for E.coli detection i.e. for non-pathogenic and potentially pathogenic strains, is also useful as levels of all types of E.coli should be minimised in food production. Any increase in normal levels or indicators should trigger an active investigation of the reasons for the increased levels. Effective process control of all cooking/pasteurisation stages is essential to ensure that the correct heating temperatures and times are achieved.

Detection

It is not easy to detect VTEC in raw meats and foods where low levels of E.coli may be swamped by high numbers of other bacteria. It is estimated that E.coli O157 has an infective dose of only 10 to 100 organisms; therefore with such a small infective dose cross contamination of high risk foods with raw food is a potential problem in retail outlets and at home. So far, there are no widely recommended methods for routine food examination. Methods used in medical laboratories to detect the organism from stools are more successful, probably because the number of E.coli O157 cells present in the stools of someone made ill by the organism is relatively high in comparison to the background flora.

Currently traditional isolation methods from foods involve enrichment in a selective broth followed by plating on to sorbitol MacConkey agar with additives. This medium is suitable only for O157:H7 strains (most but not all E.coli O157 strains do not ferment sorbitol). The composition of the enrichment broth and plating medium is important if VTEC is to be isolated from contaminated materials, and several groups are working on determining the optimum combination of selective agents.

For E.coli O157 (but not all VTEC serotypes) commercial kits are available for isolation and identification (ELISA methods) and for confirmation of suspect colonies (latex agglutination). An immunoassay kit is available for Shiga-like toxins and it is reported to detect several different VTEC strains. Recipes for several effective broths and agars have been published but there is no consensus yet on which is the best. Another technique used to enhance isolation of VTEC from enrichment broths is the use of commercially-available immunomagnetic beads coated with specific O157 antiserum. An immuno-blotting technique is available for rapid identification of O157 colonies on agar plates. Biochemical and serological confirmation tests may also be needed to exclude false positives. In clinical cases, confirmed strains should be tested for the production of VT or the presence of VT genes. Confirmation can only be done at very specialised laboratories. Most laboratories are only able to test for E.coli O157.

It is important to note that many VTEC strains do not grow well at 44¢XC. Many food and water E.coli methods rely on the ability of most strains to grow and produce gas and indole at 44¢XC. Such methods would probably fail to detect VTEC. Methods which rely on enrichment followed by plating on to selective media and biochemical confirmation, all at 37¢XC would be more likely to isolate VTEC. Incubation at slightly elevated temperatures, e.g. 37¢XC to 42¢XC, may be incorporated in the future; method development using cultural and rapid techniques is progressing fast.

The use of sub-typing by pulsed-field gel electrophoresis and comparison of patterns by the US Center for Disease Control¡¦s National Molecular Subtyping Network for Foodborne Disease Surveillance (PulseNet) has also increased the ability to identify outbreaks. PulseNet is not a surveillance system itself but a laboratory sub-typing method used in surveillance.

Research

The Microbiological Safety of Food Funders Group has published an updated Report on UK publicly funded research relating to VTEC, covering the period from 1999 to 2003.

In the Strategic Plan (2005 to 2010) of the UK FSA one of the key aims is to continue to reduce foodborne illness in the UK. On 18 January 2005 in London the Agency will conduct a review of the VTEC research programme.

Conclusion

VTEC can cause very serious illness in humans. Numbers of reported cases are usually low but increasing. The main source of the organism is cattle and it is not likely that VTEC can be completely eliminated from raw meats or the cattle population. Manufacturers, caterers and consumers need to understand how they can each reduce the chance of causing foodborne VTEC illness. Thorough cooking of raw meats, pasteurisation of milk and the avoidance of cross-contamination from raw meats to other foods are the most effective ways of preventing VTEC infections.