Verotoxigenic Escherichia coli (VTEC) are members of the bacterial species E. coli with the potential to express one or more verotoxins as determined by detection of the verotoxin protein or possession of the verotoxin gene (stx). Verotoxins are bacterial protein toxins of the AB5 family, homologues of the Shiga toxin of Shigella dysenteriae, which terminate protein synthesis in cells by cleavage of a specific adenosine residue from the 28S RNA of the 60S ribosomal subunit (Melton-Celsa, 2014). Verotoxin is also known by the synonymous terms verocytotoxin, Shiga-like toxin and Shiga toxin. These synonyms arose because verotoxin was first identified and reported (Konowalchuk et al., 1977) prior to the establishment of its relationship to Shiga toxin (O'Brien and La Veck, 1983). Thus, the terms verotoxin-producing E. coli, Shigatoxigenic E. coli and Shiga toxin-producing E. coli, should be recognised as synonyms for this class of pathogens.
Subgroups of VTEC with additional biological markers or epidemiological association with severe patient outcomes have been identified in the literature with the terms Enterohemorrhagic E. coli (EHEC) or Hemolytic Uremic Syndrome associated E. coli (HUSEC). The term VTEC is used throughout this document as it is inclusive of these E. coli pathogens.
In addition to VTEC, there are several other enteric pathotypes of E. coli which are distinguished on the basis of specific virulence factors, disease symptoms and pathology (Croxen et al., 2013). In addition to stx, VTEC strains may possess virulence factors or other biomarkers associated with these pathotypes.
VTEC as a Health Hazard
Following ingestion, VTEC can cause enteric illness with a range of symptoms. In the mildest form of the disease, patients develop self-resolving, uncomplicated diarrhoea. Alternatively, patients may develop bloody diarrhoea (BD) or haemorrhagic colitis (HC), which is not always, distinguished clinically, but in both cases the patient experiences blood in the stool (Karpman and Ståhl, 2014). Consequently, this document will refer to BD when discussing both BD and HC. Though the majority of BD cases will self-resolve, a minority of cases will progress to Haemolytic Uremic Syndrome (HUS). The rate of HUS development varies between outbreaks of both VTEC O157 and other VTEC, with reported rates of HUS development of 2 to 22% (Vallis et al., 2018). HUS is a life-threatening illness resulting in death or end stage renal disease in 12% of cases, and with 25 to 30% of survivors experiencing long term renal sequela (Garg et al., 2003; Garg et al., 2009; Spinale et al., 2013).
Determining the relative hazard posed by VTEC isolates from food remains a major challenge as individual strains appear to vary significantly in the likelihood of causing severe illness, with outcomes such as BD, HUS and death. Some clonally related groups of VTEC, such as VTEC serotype O157:H7, have a higher association with severe patient outcomes and outbreaks than others. Other strains have not been reported as clinical isolates, or have only been associated with cases of uncomplicated diarrhoea. Understanding the pathogenic potential of VTEC strains is further complicated by the apparent role of individual patient factors. Even in outbreaks of high risk VTEC, such O157:H7 or sorbitol-fermenting O157, the symptoms experienced by individuals may range from asymptomatic infection to HUS (Jaakkonen et al., 2017; Bayliss et al., 2016). Similarly, a Japanese study of 399 VTEC isolates from asymptomatic healthy adults found that though many of the isolates possessed serotypes and genotypes that are rarely isolated from symptomatic individuals, strains of VTEC with features associated with BD and HUS were also present (Morita-Ishihara et al., 2016).
In the absence of animal models which mimic human responses to VTEC infection, it is challenging to experimentally determine the factors governing the pathogenic potential of VTEC (NACMCF, 2019). Consequently, many putative virulence factors have been identified on the basis of epidemiological association with symptoms of severe illness (BD and HUS) and association with reported outbreaks. Unfortunately, in this approach there is an inherent risk of confounding the three factors which can be expected to determine the apparent pathogenic potential of individual strains:
- The probability that the pathogen strain will come into contact with humans. This will be determined by the ecology of the strain.
- Infectivity: the probability of infection on exposure to a single cell of the strain.
- The probability of infection resulting in severe illness. Patients suffering severe symptoms, such as BD, are more likely to seek medical attention and be reported.
These three factors can potentially occur in any combination. From a food safety perspective, the problem is that an isolate with previously unreported characteristics may have high infectivity and potential to cause BD and HUS, but has previously been unreported due to a low probability of exposure. This appears to have been the case with the VTEC O104:H4 strain with enteroaggregative virulence factors, responsible for the 2011 European outbreak (Beutin and Martin, 2012).
The infectivity of infectious agents is often described in terms of dose response, the probability of a specified response (i.e., illness, infection, or certain sequelae) following exposure to a specified pathogen in a specified population, as a function of the dose (WHO/FAO, 2003).
Dose response is a concept originally developed for characterising toxicological hazards and presumes a proportional relationship between the concentration of an agent and the severity of the resulting illness. However, for infectious agents, the severity of the illness is not determined by the number of cells or virus particles to which an individual is exposed. The severity of illness is instead determined by the virulence factors possessed by the infecting organism and the immunological vulnerability of the infected individual. Thus, the probability of infection can be characterised as either: infectivity, the probability of infection on exposure to single infectious cell or virus particles; or infectious dose, the number of single infectious cell or virus particles which has a high probability of establishing an infection.
Due to the potential for long term sequela or death, and the absence of effective treatment, experimental determination of VTEC infectivity is not possible. Instead, the infectivity or infectious dose of VTEC has been estimated from data on exposure levels in outbreak foods. Published estimates of the infectivity of VTEC are based on outbreaks of VTEC O157:H7.
A commonly quoted value for the infectious dose of VTEC O157:H7 is less than 100 CFU, based on a review of reported exposure levels from eight outbreak reports (Todd et al., 2008). However, the term infectious dose implies that there is an exposure below which there is negligible probability of infection in otherwise healthy adults. This interpretation is misleading for VTEC as the probability of infection may be significant upon exposure to a single cell.
In a model developed from exposure data from eight outbreaks (six foodborne, one water, one mud) of VTEC O157:H7, the probability of infection was estimated at between 1% and 10% (Teunis et al., 2008). Higher infectivity has been estimated in individual outbreaks, the mean probability of infection per cell, in an outbreak involving salad with seafood sauce was 26% for children and 17% for adults (Teunis et al., 2004). The differences in infectivity estimates from individual outbreaks may be due to multiple variables, including differences in the virulence factors possessed by the infecting strain, the protection of ingested cells from digestion processes by the food matrix, and host vulnerability.
The conclusion that there is significant probability of infection with exposure to a single cell of VTEC O157:H7 is supported by the low levels of the pathogen reported in variety of outbreak food vehicles (Strachan et al., 2001; Gill and Oudit, 2015; Hara-Kudo and Takatori, 2011; Gill and Huszczynski, 2016). It is currently unknown how variable infectivity is between individual VTEC strains or wider phylogenetic groups. Comparison of the levels of non-O157 VTEC in outbreak food vehicles, with reports for VTEC O157, indicates that the infectivity of non-O157 VTEC strains can approach that of VTEC O157:H7 (Paton et al., 1996; Buvens et al., 2011; Gill et al., 2019a).
Features of Vulnerable Populations
It has long been recognised that the young and elderly are at greater risk of VTEC illness and experiencing severe health outcomes. This relationship between age, the risk of illness, and of severe illness is illustrated by US data summarised in Table 1 (CDC FoodNet Fast, 2018). The rate of infection is highest for children under 5 years at 8.08 per 100,000. The rate declines steady with each increased age cohort to a minimum of 0.80 per 100,000 for 40 to 49 years old. The rate of infection then increases with age to a maximum of 1.48 for those 70 years and older. Similar patterns are seen for the rate of hospitalisation and death due to VTEC. Though patients over 70 years have the greatest risk of death, 0.06 per 100,000, of any age group, including under 5 years, 0.04 per 100,000.
Females have a higher rate of infection (2.19 per 100,000) and hospitalisation (0.66 per 100,000) than males (1.90 per 100,000 and 0.57 per 100,000), but the death rates of the two genders are the same. This gender imbalance of infection could be attributed to differences in food consumption patterns, particularly raw fruits and vegetables, as proposed to explain the predominance of female patients in the 2011 outbreak of a VTEC/Enteroaggregative E. coli (EAEC) hybrid associated with sprouts (Frank et al., 2011).
Individual serology has also been proposed as a factor governing the likelihood of VTEC illness (Karmali, 2018). The regular occurrence of infected asymptomatic individuals in outbreaks of VTEC strains that can cause BD and HUS is well documented (Jaakkonen et al., 2017; Bayliss et al., 2016; Kanayama et al., 2015). Mohamed Karmali (2018) noted that the presence of antibodies to the Locus of Enterocyte Effacement (LEE) colonisation factor intimin is associated with protective immunity to VTEC illness. Karmali has hypothesised that mid-twentieth century improvements to sanitation in industrialised countries resulted in a "...reduction in EPEC outbreaks in the 1970's and 80's..." leading "...to a decline in population immunity to intimin which, in turn may have contributed to the emergence of VTEC O157:H7 outbreaks and HUS in industrialised countries in the 1980s." If this hypothesis is correct, it can be expected that individuals who have experienced higher levels of exposure to environmental E. coli, particularly Enteropathogenic E. coli (EPEC), may have lower likelihood of developing illness from LEE positive VTEC strains.
The primary features of VTEC as a pathogen are summarised below:
- VTEC are coli with the potential to express one or more verotoxins (Shiga toxins). Synonyms for VTEC are verotoxin-producing E. coli, Shigatoxigenic E. coli and Shiga toxin producing E. coli. The term VTEC is inclusive of Enterohemorrhagic E. coli and Hemolytic Uremic Syndrome associated E. coli.
- VTEC infection follows ingestion, with potential outcomes including, asymptomatic infection, uncomplicated diarrhoea, bloody diarrhoea, haemolytic uremic syndrome and death.
- Individual health status contributes to the probability of serious illness.
- Strains of both VTEC O157 and non-O157 VTEC can be highly infectious, with a significant risk of infection on exposure to a single cell.
- The rates of VTEC infection, hospitalisation and death from VTEC illness varies with age. Midlife adults have the lowest rates, with higher rates for the young and old. Children under 5 years of age and adults over 70 have the highest death rates.
- Females have a higher rate of infection and hospitalisation than males, but the death rates of the two genders are the same.