Transmission and control of Escherichia coli O157:H7 A review

Transcription

1 Transmission and control of Escherichia coli O157:H7 A review S. J. Bach 1, T. A. McAllister 1,5, D. M. Veira 2, V. P. J. Gannon 3, and R. A. Holley 4 1 Agriculture and Agri-Food Canada Research Centre, P.O. Box 3000, Lethbridge, Alberta, Canada T1J 4B1; 2 Range Research Unit, Kamloops, British Columbia, Canada V2B 8A9; 3 Health Canada, Animal Diseases Research Institute, Box 640, TWP Road 9-1, Lethbridge, Alberta, Canada T1J 3Z4; and 4 Department of Food Science, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2. LRC contribution no , received 11 March 2002, accepted 9 April Can. J. Anim. Sci. Downloaded from by on 12/08/17 Bach, S. J., McAllister, T. A., Veira, D. M., Gannon, V. P. J. and Holley, R. A Transmission and control of Escherichia coli O157:H7 A review. Can. J. Anim. Sci. 82: Escherichia coli 157:H7 has evolved as an important foodborne pathogen since its initial description in Outbreaks of illness associated with E. coli O157:H7 have been reported throughout the northern hemisphere, most frequently in Canada, the United States, Japan, and the United Kingdom. In Canada, infections due to E. coli O157:H7 appear to be more common in the western provinces than in the east, in rural vs. urban environments, and during summer as opposed to winter months. Undercooked ground beef has been implicated as the primary vehicle in E. coli O157:H7 infection, but contaminated fruits, vegetables and water have also been linked to E. coli O157:H7 outbreaks. Epidemiological investigations demonstrate that dairy and beef cattle are primary reservoirs of this organism, carrying it asymptomatically and shedding it intermittently and seasonally in their feces. Surveys in Canada and the United States indicate widespread distribution of E. coli O157:H7 in cattle operations. The prevalence of E. coli O157:H7 in cattle has been increasing in recent reports, likely due to the development of more sensitive methods for the detection of the organism. Escherichia coli O157:H7 has been isolated from feed, water for livestock, manure, soil and flies, all of which represent potential sources of contamination for cattle and their environment. To date, effective methods for controlling E. coli O157:H7 in cattle have not been identified, although dietary manipulation, vaccination and bacteriophage therapy have been reported to have potential as intervention strategies. Effective control of E. coli O157:H7 requires reducing the frequency and intensity of fecal shedding of this pathogen by cattle, in addition to targeting environmental sources of the organism. Key words: Escherichia coli O157:H7, cattle, sources, diet, transmission Bach, S. J., McAllister, T. A., Veira, D. M., Gannon, V. P. J. et Holley, R. A Transmission de la souche O157:H7 d Escherichia coli et moyens de lutte Récapitulation. Can. J. Anim. Sci. 82: Partout dans l hémisphère nord, on rapporte des foyers de maladie attribuables à la souche O157:H7 de E. coli, mais davantage au Canada, aux États-Unis, au Japon et au Royaume-Uni. Au Canada, les infections causées par cet agent pathogène semblent survenir plus fréquemment dans l Ouest que dans l Est, à la campagne qu en ville, et l été que l hiver. La viande hachée de bœuf mal cuite est souvent la principale blâmée, mais les fruits, les légumes et l eau contaminés sont à l origine de certains foyers. Les études épidémiologiques montrent que les bovins laitiers et les bovins de boucherie constituent un important réservoir du microorganisme. Ces animaux véhiculent la maladie sans en présenter les symptômes et libèrent la bactérie de manière intermittente, saisonnière, dans leurs fèces. Les enquêtes menées au Canada et aux États-Unis révèlent que la souche O157:H7 est largement répandue dans les élevages. Des rapports récents indiquent aussi une hausse de la prévalence de cette souche, sans doute à cause de l avènement de techniques de dépistage plus sensibles. La souche O157:H7 a été isolée dans les aliments du bétail, l eau destinée aux bêtes, le fumier, le sol et les mouches, autant de sources de contamination en puissance pour les animaux et l environnement. Les méthodes de lutte efficaces contre cet agent pathogène n ont pas encore été identifiées, bien qu à en croire certains, la manipulation des aliments, la vaccination et l administration de bactériophages figurent au nombre des stratégies envisageables. Une bonne lutte contre la souche O157:H7 suppose une réduction du nombre de microorganismes libérés dans les fèces et de la fréquence de tels événements chez les bovins, parallèlement à une intervention au niveau des sources de contamination environnementales. Mots clés: Escherichia coli O157:H7, bovins, sources, alimentation, transmission 5 To whom correspondence should be addressed ( mcallister@agr.gc.ca). 475 Abbreviations: EHEC, enterohemorrhagic E. coli; HUS, hemolytic uremic syndrome; SLTEC, Shiga-like toxin producing E. coli; STEC, Shiga toxin-producing E. coli; Stx, Shiga toxin; TTP, thrombotic thrombocytopenic purpura; VFA, volatile fatty acid; VT, Verotoxin; VTEC, verocytoxin-producing E. coli.

2 Can. J. Anim. Sci. Downloaded from by on 12/08/ CANADIAN JOURNAL OF ANIMAL SCIENCE In the 20 yr since its initial description, E. coli O157:H7 has become an important public health problem worldwide (Riley et al. 1983; Griffin and Tauxe 1991). Given the magnitude and severity of recent outbreaks of E. coli O157:H7 infection, there is an urgent need to reduce the human health hazard presented by this pathogen. Foods of bovine origin have been implicated in numerous outbreaks and epidemiological investigations have shown that cattle are a major reservoir of E. coli O157:H7 (Wells et al. 1991; Hancock et al. 1994). The ability of the organism to survive in feed, water, soil, and manure has important implications for its persistence in cattle herds and contamination of water supplies and crops (Lynn et al. 1998; Fenlon et al. 2000; Buchko et al. 2000a; Hancock et al. 2001; Bach et al. 2002). Effective measures to reduce or eliminate E. coli O157:H7 in cattle will reduce not only foodborne illness but also the risk of transmission of the organism into the environment. In this review, we report sources of E. coli O157:H7 and examine the incidence of infections associated with this organism since its emergence nearly two decades ago. Transmission of E. coli O157:H7 and potential methods for its control will also be discussed. The significance of E. coli O157:H7 as a human pathogen underlines the importance of establishing effective strategies to minimize numbers of E. coli O157:H7 on the farm. HISTORY AND ORIGINS OF E. COLI O157:H7 Escherichia coli O157:H7 was first identified as a human pathogen following two outbreaks of gastrointestinal illness in the United States associated with undercooked hamburger patties (Riley et al. 1983; Wells et al. 1983). Most strains of E. coli are benign inhabitants of the gastrointestinal tract of humans and other warm-blooded animals, but pathogenic E. coli strains, such as E. coli O157:H7, also exist which are associated with distinct syndromes of diarrheal disease (Levine 1987). In 1978, Konowalchuk et al. (1978) reported that certain strains of diarrheagenic E. coli produced a potent toxin that killed African green monkey kidney cells (aka Vero cells) and they named the cytotoxic agent as Verotoxin (VT). Subsequent studies revealed immunological and functional similarities between VT and the Shiga toxin produced by Shigella dysenteriae (O Brien and La Veck 1983; O Brien and Holmes 1987; Neill 1997), thus VT also became known as Shiga-like toxin (SLT). Two major types of SLT/VT were recognized (SLT-I and SLT-II, or VT1 and VT2), and in 1996, Calderwood et al. (1996) proposed the nomenclature Stx1 and Stx2, to reflect the high degree of homology of the two toxins that placed them in the Shiga toxin family (Strockbine et al. 1986). Over 100 serotypes of E. coli produce one or both of these cytotoxins. As a group, these organisms have been designated interchangeably as verocytotoxin-producing E. coli (VTEC), Shiga-like toxin-producing E. coli (SLTEC), and Shiga toxin-producing E. coli (STEC). In this review, we preserve the nomenclature used by the researchers whose work is cited. Some strains of STEC are associated with hemorrhagic colitis (bloody diarrhea) and hemolytic uremic syndrome (HUS) in humans. This subgroup is referred to as enterohemorrhagic E. coli (EHEC) (Kaper 1994). Escherichia coli O157:H7 is the predominant EHEC serotype associated with foodborne illness in North America and the United Kingdom, but non-o157 EHEC serotypes, e.g., serogroups O26, O103, O111 and O128, have also been implicated in sporadic enteric infections and are considered to possess outbreak potential (Johnson et al. 1996; Bettelheim 2000). It has been proposed that E. coli O157:H7 evolved from E. coli O55:H7, an enteropathogenic strain of E. coli long associated with diarrhea in infants worldwide (Whittam et al. 1993; Wachsmuth et al. 1997; Mead and Griffin 1998). Characteristic of other enteropathogenic E. coli strains, E. coli O157:H7 can adhere to epithelial cells and produce attaching and effacing (AE) lesions. It is thought that E. coli O55:H7, already possessing a mechanism for adherence to intestinal cells, acquired slt genes from Shigella, and the O157 antigen through horizontal gene transfer and recombination (Whittam et al. 1993; Mead and Griffin 1998). The genetic relatedness between human and bovine E. coli O157:H7 isolates has been studied using an octamerbased genome-scanning method. This methodology provides detailed genome comparisons based on the distances between over-represented, strand-biased octamers. The existence of two distinct lineages of E. coli O157:H7 distributed in the United States has been demonstrated. Among the lineages, human and bovine isolates have been found to be non- randomly distributed, implying that strains from one lineage may not readily cause disease or be effectively transmitted from bovines to humans. It has been hypothesized that the divergence of the lineages is associated with phage mediated events and this provides an explanation for the diversity among E. coli O157:H7 isolates using other molecular methods such as pulse-field gel electrophoresis (Kim et al. 1999). Recent determination of the complete sequence of the E. coli O157:H7 genome has furthered our knowledge about the pathogenicity and evolution of this organism. Comparison of the genomes of E. coli O157:H7 strain EDL933 (isolated from ground beef in Michigan and linked to the 1982 outbreak) and E. coli O157:H7 strain RIMD (isolated from a patient during the Sakai City outbreak in 1996) with the genome of the non-pathogenic laboratory E. coli strain K-12 MG1655 revealed the sharing of a common 4.1 million-base pair backbone with almost identical gene order. Hundreds of sections of DNA unique to each strain are scattered within each genome, with much of the DNA in the O and K islands having been acquired through horizontal gene transfer. The extent of their genetic differences suggests that E. coli O157:H7 and E. coli K-12 shared a common ancestor about 4.5 million years ago (Hayashi et al. 2001; Perna et al. 2001). SOURCES OF E. COLI O157:H7 Cattle as a Reservoir of E. coli O157:H7 Subsequent to the first two outbreaks of E. coli O157:H7 infection in 1982, a bovine source of the pathogen was implicated (Riley et al. 1983; Wells et al. 1983). Early stud-

3 BACH ET AL. TRANSMISSION AND CONTROL OF E. COLI O Can. J. Anim. Sci. Downloaded from by on 12/08/17 ies indicated that the prevalence of E. coli O157:H7 in the feces of cattle in the United States was low (Hancock et al. 1994; Faith et al. 1996). Hancock et al. (1994) reported this organism to be present in dairy and beef cattle at 0.28% and 0.71%, respectively, with herd prevalences of 8.3% and 16%, respectively. More recent studies, however, suggest much higher levels of occurrence. Chapman et al. (1997a) isolated E. coli O157:H7 from the feces of 15.7% of cattle (752/4800) over a 1-yr period. Monthly prevalences ranged from 4.8 to 36.8%, and were highest in the spring and late summer. Van Donkersgoed et al. (1999) isolated E. coli O157:H7 from 7.5% of fecal samples collected from cattle at slaughter. In that study, the prevalence of E. coli O157:H7 was higher in samples from yearling cattle than from cull cows (12.4 vs. 2.0%). Even more recently, Elder et al. (2000) reported E. coli O157:H7 to be present in 91 of 327 fecal samples (28%) taken from slaughter cattle in the midwestern United States during July and August. Another survey in the United States found E. coli O157:H7 was isolatable from 63 of 100 feedlots tested, indicating widespread distribution of this organism in cattle operations (Hancock et al. 1997c). The increasing prevalence of E. coli O157:H7 has likely arisen from the use of more sensitive isolation methods. Immunomagnetic separation has increased the sensitivity of detection of E. coli O157:H7 by 10- to 100-fold (Chapman et al. 1994). The prevalence of E. coli O157:H7 in cattle is influenced by numerous variables, including season, the scope, frequency and timing of sampling, and the conditions of sample transport and storage (Van Donkersgoed et al. 1999; Elder et al. 2000). As well, geographic location also seems to influence prevalence. For example, the prevalence of E. coli O157:H7 in cattle is higher in the northwest United States and in northern England (Sheffield), than in other areas (Chapman 2000). Repeated isolation of E. coli O157:H7 from healthy beef and dairy cattle demonstrates that cattle are asymptomatic carriers of the organism (Wells et al. 1991; Hancock et al. 1994; Zhao et al. 1995). Thus, the presence of the organism in an individual animal or livestock operation cannot be signaled by morbidity. At present, fecal analysis is the only practical means of confirming or enumerating E. coli O157:H7. This task is complicated further by the fact that fecal shedding of E. coli O157:H7 by cattle is intermittent. Short periods of relatively high prevalence of excretion are separated by longer periods of reduced or undetectable shedding (Wells et al. 1991; Besser et al. 1997). This situation has undoubtedly contributed to the variance in prevalences reported in the literature. Among those animals reported to be shedding E. coli O157:H7, levels of shedding have ranged from 10 2 to 10 5 cfu g 1 of feces. The shedding itself tends to be seasonal, peaking in the summer months (May through September; Hancock et al. 1994; Van Donkersgoed et al. 1999). The predominant site of E. coli O157:H7 persistence and proliferation in the adult bovine gastrointestinal tract is the colon (Zhao et al. 1995; Grauke et al. 2002). In humans, the virulence of this pathogen is associated mainly with its VT. Receptors for these cytotoxins are present primarily in the vasculature of the colon and the kidney (Schmidt et al. 1995). In contrast, VT receptors are present in the bovine intestinal tract and kidney, but not in the vasculature, which may enable the asymptomatic carriage of E. coli O157:H7 by cattle (Hoey et al. 2002). Other Animal Sources of E. coli O157:H7 Epidemiological investigations of animal populations as reservoirs of E. coli O157:H7 have concentrated on bovine breeds but other animal hosts have also been identified (Beutin et al. 1996). In the United Kingdom, Chapman et al. (1997a, b) reported isolation of E. coli O157:H7 from 2.2% of sheep and 0.4% of pigs at slaughter, but the organism was not detected in samples from 1000 chickens. In other studies in the United States, The Netherlands and Australia, the prevalence of E. coli O157:H7 in sheep ranged as high as 31%, indicating that sheep may also play a significant role in harboring E. coli O157:H7 (Kudva et al. 1996; Heuvelink et al. 1998; Fegan and Desmarchelier 1999). Trevena et al. (1996) suggested that horses and dogs may be vectors of E. coli O157:H7, due to the close contact between farm animals, companion animals and humans. Hancock et al. (1998) reported detection of E. coli O157:H7 in 1.1% (1/90) of horses and 3.1% (2/65) in dogs in the United States. Escherichia coli O157:H7 has also been isolated in the United States from wild deer, white-tailed deer and the feces of wild birds and from birds in the United Kingdom (Rice et al. 1995; Keene et al. 1997; Wallace et al. 1997; Hancock et al. 1998; Sargeant et al. 1999; Fischer et al. 2001). The avian isolations suggest the possibility of transfer of E. coli O157:H7 between animals sharing the same habitat. The high variability of prevalence of E. coli O157:H7 among cattle suggests the possibility of a reservoir of E. coli O157:H7 external to cattle (Hancock et al. 1998). The intermittent and seasonal nature of their fecal shedding of E. coli O157:H7 could reflect cattle functioning more as a vehicle for transmission of the organism than as a reservoir in which the organism is maintained for long periods of time (Hancock et al. 1998). Reports of E. coli O157:H7 in other animal species and birds confirm the existence of non-bovine animal reservoirs or vehicles of E. coli O157:H7. Other than the confirmation of the presence of E. coli O157:H7 in non-bovine species, however, the ecology of E. coli O157:H7 beyond cattle has not been extensively studied. Environmental Sources of E. coli O157:H7 Johnson et al. (1999) outlined many possible habitats for E. coli O157:H7 in the farm environment, including manure heaps, ponds, dams and wells, barns, calf hutches, straw and other bedding, feed and feed troughs, water and water troughs, farm equipment, ground and pasture, and watercourses, and noted that once present in the environment, this organism can be transferred to other sites by rainwater, wind, and removal and spreading of manure, as well as by animals and humans. Aspects of survivability and transmissability of E. coli O157:H7 in many of these habitats have been investigated.

4 Can. J. Anim. Sci. Downloaded from by on 12/08/ CANADIAN JOURNAL OF ANIMAL SCIENCE Feed Escherichia coli O157:H7 was isolated from 6.3% of feed samples obtained from a dairy farm in Wisconsin (Shere et al. 1998). Van Donkersgoed et al. (2001) isolated E. coli O157:H7 from feed in 1.7% of feed bunks in feedlot pens but found that total mixed rations (TMR) did not support the growth of the organism. Conversely, growth of E. coli O157:H7 has been demonstrated in a variety of wet grain mixtures and some silage-based mixtures in vitro and poorly fermented laboratory silage has also been shown to support the growth of E. coli O157:H7 (Lynn et al. 1998; Fenlon and Wilson 2000). Two strains of E. coli O157:H7 inoculated at harvest onto grass ensiled under intentionally poor conditions increased from 10 3 cfu g 1 to 10 7 cfu g 1 over 13 d at 20 C. The poor ensiling technique likely delayed the silage fermentation, creating an environment conducive to growth of E. coli O157:H7. Fecal contamination of forage (e.g., from manure application) followed by inadequate ensiling may therefore be a significant factor in the transmission and persistence of E. coli O157:H7 in the feedlot environment (Fenlon and Wilson 2000). Adequate ensiling, on the other hand, has been shown to reduce the risk of E. coli O157:H7 contamination of forage. In a laboratory-scale mini-silo trial, barley forage was treated at harvest with 10 5 cfu g 1 of O157:H7 and non-o157:h7 strains of E. coli and ensiled for 42 d with and without a bacterial silage inoculant designed to accelerate the fermentation and decline of ph in the silage. All of the E. coli were eliminated within 15 d. Including the silage inoculant significantly reduced the persistence of E. coli O157:H7. Thus, this strategy would allow producers to feed silage out shortly after ensiling with minimal risk of transmitting E. coli O157:H7 to cattle (Bach et al. 2002). The seasonality of fecal shedding of E. coli O157:H7 implies that environmental replication may play a key role in its ecology in the farm environment. Given that this organism can replicate in a variety of feeds, thus amplification of low levels of E. coli O157:H7 may increase the possibility of further dissemination of the organism on the farm. Water for Livestock Water for livestock may become contaminated by bacteria from numerous sources. Cattle may contaminate their troughs with fecal material, feed, bedding or saliva (LeJeune et al. 2001b). In some cases, water may be contaminated even before it enters the trough (Van Donkersgeod et al. 2001). Several studies have documented isolation of E. coli O157:H7 from animal drinking water (Faith et al. 1996; Hancock et al. 1998; Buchko et al. 2000a; LeJeune et al. 2001a, b; Van Donkersgoed et al. 2001). This organism survived for over 6 mo in the sediments of microcosms designed to simulate cattle water troughs, and these contaminated microcosms effected colonization of 10-wk old calves with E. coli O157:H7. Wang and Doyle (1998) found that E. coli O157:H7 may exist in water in a viable but nonculturable (VBNC) state for long periods of time, all the while retaining pathogenic potential. Escherichia coli has also been shown to form biofilms. Active bacterial growth and sloughing in response to biocides suggested that a protective capacity of biofilms allowed these microorganisms to persist in water distribution systems (Daly et al. 1998). Thus, water troughs can serve as environmental reservoirs of E. coli O157:H7 and contribute to subsequent infection of cattle (Lejeune et al. 2001a). There is evidence that drinking water for livestock may become contaminated by oral rather than fecal routes. Escherichia coli O157:H7 was detected in drinking water obtained from covered water tanks equipped with ball-valved watering ports (Shere et al. 1998). Those researchers discounted fecal contamination of that watering system, given that delivery of water required physical depression of the ball in the port by the animal. Recovery of E. coli O157:H7 from the tonsils (Cray and Moon 1995) and saliva (Buchko et al. 2000a; Keen and Elder 2002) of cattle further supports the possibility of oral contamination of their drinking water. Moreover, recovery of E. coli O157:H7 in the feces of previously O157-negative cattle subsequent to their having ingested water harboring that bacterium suggests that drinking water may be a source of infection (Shere et al. 1998). In a study of two feedlots in southern Alberta, E. coli O157:H7 was isolated from 12% of water troughs (27/219), and 1 of the 21 incoming water sources was also confirmed positive for E. coli O157:H7 (Van Donkersgoed et al. 2001). Environmental data collected in one of the feedlots revealed relationships between prevalence of E. coli O157:H7 in the troughs and the season, ambient temperature maxima, total precipitation in the week prior to sampling, and total coliforms and E. coli in the water troughs. Isolation of similar E. coli O157:H7 subtypes from the feed, water and feces in the lot implied transmission of the organism among these sources. Conversely, however, different subtypes were also isolated from the water at source, the water troughs and fecal samples, which suggested contamination of the water by E. coli O157:H7 from a source other than fecal material (Van Donkersgoed et al. 2001). LeJeune et al. (2001b) isolated E. coli O157:H7 from 6 of 473 water troughs located at 99 different cattle operations. The extent of contamination of water troughs with E. coli was positively correlated with proximity of the troughs to the feed bunk, protection of the trough from direct sunlight, warmer environmental temperatures and reduced protozoal concentrations in the water troughs. Reducing the protozoal population in an experimental microcosm also resulted in increased populations of E. coli O157:H7 (LeJeune et al. 2001b). These results suggest that many factors that dictate microbiological quality of animal drinking water are controllable. It is evident that water troughs may serve as long-term reservoirs of E. coli O157:H7. Survival of E. coli O157:H7 in water for extended periods of time, in addition to the potential of contaminated water as a point source of infection, emphasizes the importance of control of E. coli O157:H7 in drinking water for livestock. Manure Reports regarding the relationship between application of manure to grazing land and the prevalence of E. coli

5 BACH ET AL. TRANSMISSION AND CONTROL OF E. COLI O Can. J. Anim. Sci. Downloaded from by on 12/08/17 O157:H7 in cattle have been inconsistent. Hancock et al. (1994) reported a tentative association between applying slurry to grazing land and the E. coli O157:H7 infection status of the herd. Subsequent studies, however, reported no association between manure application to pastures or forage crops for cattle and E. coli O157:H7 fecal shedding by cattle (Garber et al. 1995; Hancock et al. 1997b). More recently, a study by Bolton et al. (1999) indicated that current waste management practices such as spreading manure on pastureland may increase the carriage rate of E. coli O157:H7 in herds. Survival of E. coli O157:H7 in manure and manure slurry has been observed under various experimental and environmental conditions. Kudva et al. (1998) found that E. coli O157:H7 from experimentally inoculated sheep survived for 21 mo in manure exposed to environmental conditions. In aerated piles of bovine and ovine manure, the organism survived for 4 mo and 47 d respectively. In bovine feces inoculated initially with 10 5 cfu g 1, E. coli O157:H7 survived for 70, 56 and 49 d, at 5, 22 and 37 C, respectively (Wang et al. 1996). Fukushima et al. (1999) found that E. coli O157:H7 inoculated at 10 3 cfu g 1 survived in bovine feces for 2 to 14 wk at 5 C, 1 to 18 wk at 15 C, and 3 to 12 wk at 25 C. Contamination with bovine manure has been suggested in several cases of human infection associated with non-bovine food products. Inadequately washed vegetables were implicated as the source of E. coli O157:H7 infection following isolation of the pathogen from a garden fertilized with manure (Cieslak et al. 1993). The use of manure as fertilizer could explain foodborne outbreaks of E. coli O157:H7 associated with unpasteurized apple cider, potatoes, radish sprouts and lettuce (Morgan et al. 1988; Besser et al. 1993; Chapman et al. 1997b; Ackers et al. 1998; Michino et al. 1999). Manure may be a risk factor for transmitting E. coli O157:H7 to produce grown on manure-fertilized soil and may serve as a source for the maintenance of E. coli O157:H7 among cattle herds (Chapman et al. 1997b; Chart 1998). Given that E. coli O157:H7 can survive in manure and manure slurry for extended periods of time, proper manure management is of utmost importance in preventing the spread of this organism to the environment, food and animal crops, and back to cattle. Composting is an effective method for eliminating pathogens such as E. coli O157:H7 from manure. In preliminary studies, pathogenic organisms such as Giardia and Salmonella were eliminated from compost after 14 d, whereas Campylobacter and E. coli O157:H7 were eliminated after 7 d (Olson 2000). Few pathogens are able to withstand the heat generated during the composting process, but it is important that 60 C be attained and maintained for threshold periods in all parts of the compost pile. Pell (1997) reported that drying manure was also effective for reducing the presence of pathogens. Soil Inoculated at a rate of 10 6 cfu g 1 of soil, E. coli O157:H7 survived in loam and clay soils for 25 wk and in sandy soil for 8 wk (Fenlon et al. 2000). The organism was detectable for up to 7 d after incorporation into the uppermost 2.5 cm of the soil, and for up to 7 d on grass plots inoculated with a fecal slurry from dairy cattle at an application rate of E. coli O157:H7 of 660 cfu m 2. Approximately 2% of the initial inoculum of E. coli O157:H7 was transported to the deeper layers of the soil. Transport was mainly associated with rainfall, which occurred 3 and 7 d after slurry application, and which led to a cumulative loss of 7% of the applied E. coli O157:H7 (Fenlon et al. 2000). Maule (2000) measured the persistence of E. coli O157:H7 in soil using cores as a model system. Survival of the organism was greatest in soil cores containing rooted grass; viable numbers declined from 10 8 cfu g 1 soil to between 10 6 and 10 7 cfu g 1 soil after 130 d at 18 C. Tilling practices, soil type and method of pathogen delivery were each found to affect vertical transport of E. coli O157:H7 in soil (Gagliardi and Karns 2000). Steady rainfall was applied to intact (no-till) and disturbed (tilled) soil cores treated with manure inoculated with E. coli O157:H7. Manure enhanced the survival of E. coli O157:H7 in intact (no-till) soils but not in disturbed (tilled) soils. High levels of recovery and low leachate levels of E. coli O157:H7 from clay loam soils as compared to silt loam or sandy loam soils indicated that growth of E. coli O157:H7 in soil could occur if leaching losses were minimal. The research confirmed vertical movement of the organism through the soil for more than 2 mo after initial application, provided soil pores did not become clogged. This migration represents another possible route of contamination of food and water (Gagliardi and Karns 2000). Barker et al. (1999) found that E. coli O157:H7 was able to survive and replicate within cells of Acanthamoeba polyphaga, a common environmental protozoan and suggested a dynamic and mutually beneficial interaction between E. coli O157:H7 and A. polyphaga in laboratory microcosms. In some instances, E. coli O157:H7 was digested as a food source by the protozoa but in others, the bacteria multiplied. Thus, intracellular growth in protozoa widely distributed in soils may provide a protective niche for pathogens such as E. coli O157:H7. In addition to providing physical protection from adverse conditions, bacterial physiology may be altered in such a way to optimize their potential for virulence (Barker and Brown 1994). Survival of E. coli O157:H7 in amoeba cysts may also enhance its distribution, as has been shown for Legionella pneumophila, Vibrio cholerae and Mycobacterium avium, which may be blown through the air (Brown and Barker 1999; Steinert et al. 1998). Protozoa containing E. coli O157:H7 distributed to grass and silage may subsequently be ingested by grazing cattle and represent a significant vector for the transmission of E. coli O157:H7 (Brown et al. 2002). Flies The increased presence of flies around cattle during the summer months represents a potential mechanism for the spread of E. coli O157:H7 among animals. A recent study was conducted on the potential of transmission of E. coli O157:H7 to fresh-cut apple tissues by fruit flies. Following contact with E. coli O157:H7, the fruit flies were contami-

6 Can. J. Anim. Sci. Downloaded from by on 12/08/ CANADIAN JOURNAL OF ANIMAL SCIENCE nated with the organism both externally and internally, and subsequently transmitted the bacterium to uncontaminated apple wounds (Janisiewicz et al. 1999). Sasaki et al. (2000) detected E. coli O157:H7 at a level of cfu in the crop of houseflies (Musca domestica) immediately after their having fed on a bacterial preparation. Numbers of E. coli O157:H7 in individual drops of excreta exceeded 10 4 after 1 h, averaged bacteria after 3 h, and continued to decline over the next 24 h. Isolation of E. coli O157:H7 from flies collected on dairy farms and feedlots has been reported (Rahn et al. 1997; Hancock et al. 1998; Shere et al. 1998). Thus, flies are potentially effective vectors for dispersion of E. coli O157:H7 in the environment and to humans. OUTBREAKS OF E. COLI O157:H7 INFECTION Following identification of E. coli O157:H7 as the causative agent of human illness in 1982 (Riley et al. 1983), numerous outbreaks and sporadic cases of illness associated with E. coli O157:H7 have been reported from Argentina, Australia, Belgium, Canada, China, Czechoslovakia, England and Wales, Ireland, Italy, Japan, Korea, Mexico, New Zealand, Romania, Scotland, Spain, Thailand and the United States (Griffin and Tauxe 1991; Besser et al. 1993; Chapman 1995; Ackman et al. 1997; Keene et al. 1997; Swerdlow and Griffin 1997; Cody et al. 1999; Michino et al. 1999). Increased recognition of E. coli O157:H7 as a cause of human illness has resulted in increased surveillance for E. coli O157:H7 worldwide. Seemingly more frequent isolation of the organism in developed countries is likely due to more advanced reporting systems and more sensitive technologies for detection. Clinical Manifestations of E. coli O157:H7 Infections Infection with E. coli O157:H7 causes three principal manifestations of illness in humans: hemorrhagic colitis, HUS, and thrombotic thrombocytopenic purpura (TTP)(Griffin and Tauxe 1991). The initial symptoms of hemorrhagic colitis (abdominal cramps, bloody diarrhea and dehydration) generally occur 3 to 5 d after ingestion of E. coli O157:H7. Whereas most cases are self-limiting and resolve without harmful effects, complications resulting in HUS or TTP occur in about 5% of cases. Characteristic symptoms of HUS include intravascular destruction of red blood cells (hemolytic anemia), depressed platelet counts (thrombocytopenia) and acute renal failure. Children under 5 yr of age, the elderly and immunocompromised individuals are most susceptible to HUS, with a mortality rate of 3 5%. By comparison, TTP causes less renal damage and has more neurological involvement (seizures, strokes and coma) than HUS (Buchanan and Doyle 1997). The production of one or both types of VT, VT1 and VT2, is the major virulence factor of EHEC, such as E. coli O157:H7 (Nataro and Kaper 1998). The receptors for VT1 and VT2, are endothelial cells that express globotriaosylceramide (Gb 3 ) and are present primarily in the cortex of the human kidney (Schmidt et al. 1995). The resultant disruption of protein synthesis leads to the death of any cells possessing the Gb 3 receptor (Nataro and Kaper 1998). Damage to renal endothelial cells by this process is likely the primary phenomenon resulting in HUS where endothelial cells of the renal vasculature are the principal site of damage (Nataro and Kaper 1998). Geographic Distribution Although illnesses due to E. coli O157:H7 have been reported in over 30 countries on six continents, E. coli O157:H7 infections have been reported most frequently in Canada, the United States and the United Kingdom (Doyle 1991; Griffin and Tauxe 1991; Waters et al. 1994; Boyce et al. 1995; Chapman 1995; Mead and Griffin 1998; Nataro and Kaper 1998; Michino et al. 1999). Escherichia coli O157:H7 is also an important pathogen in Japan and Europe, whereas in countries such as Australia, Argentina, Chile and South Africa, non-o157 serotypes of EHEC are more frequently associated with hemorrhagic colitis and HUS (Bettelheim 1998; Nataro and Kaper 1998). Infections due to E. coli O157:H7 are apparently more common in Canada than in the United States (Griffin and Tauxe 1991; Boyce et al. 1995). Within the United States, outbreaks are more common in the northern states than in the southern states (Griffin and Tauxe 1991; Slutsker et al. 1997; Nataro and Kaper 1998; Chapman 2000). In Canada, reports of E. coli O157:H7 infection have been more frequent in the western provinces than in the east (Griffin and Tauxe 1991; Nataro and Kaper 1998). Rural vs. Urban Environments Michel et al. (1999) investigated the relative risks of VTEC infection associated with living in rural vs. urban environments by examining the geographic distribution of VTEC infections in Ontario, Canada, between 1990 and Reported annual rates of VTEC infection were consistently higher in rural areas (5.4 per ) as compared to urban areas (4.4 per ). Spatial analyses revealed that cattle density was a significant indicator of VTEC infection in many regions of Ontario. Rural residents were at an increased risk for VTEC infection. That finding may also explain the relatively high incidence of VTEC infections in Prince Edward Island, where the cattle density (0.17 per hectare) approaches that of the regions of Ontario with the highest incidence of VTEC infection (0.2 cattle per hectare; Michel et al. 1999). Factors which may be associated with the increased risk include recurrent exposure to cattle, surface water and shallow well contamination with cattle manure used as fertilizer and consumption of locally processed foods (Michel et al. 1999). While living in a rural area with a high cattle density may result in an elevated risk for VTEC infection, continual exposure to VTEC (O157 and non-o157 serotypes) in the environment may offer protection (Reymond et al. 1996; Wilson et al. 1996; Michel et al. 1999). Reymond et al. (1996) reported that the frequency of O157 lipopolysaccharide antibodies was threefold higher in Ontario dairy farm families as compared to urban families (12.5 and 4.7%, respectively) and that the incidence of

7 BACH ET AL. TRANSMISSION AND CONTROL OF E. COLI O Can. J. Anim. Sci. Downloaded from by on 12/08/17 Table 1. Foods other than ground beef associated with outbreaks of E. coli O157:H7 infection Year Location Food Reference 1985 United Kingdom Potatoes Morgan et al. (1988) 1986 Canada Unpasteurized milk Borczyk et al. (1987) 1987 United Kingdom Turkey roll Salmon et al. (1989) 1991 United States Apple cider Besser et al. (1993) 1993 United Kingdom Unpasteurized milk Chapman et al. (1993b) 1993 United Kingdom Yogurt Morgan et al. (1993) 1994 United States Mayonnaise Weagent et al. (1993) 1995 United States Dry-cured salami CDC (1995) 1995 United States Lettuce Ackers et al. (1998) 1995 United States Venison jerky Keene et al. (1997) 1995 United Kingdom Potatoes Chapman et al. (1997b) 1996 Japan Radish sprouts Michino et al. (1999) 1996 United Kingdom Meat products Liddell (1997) 1996 United States/Canada Apple juice Cody et al. (1999) 1997 United States Alfalfa sprouts CDC (1997) 1998 Canada Genoa salami Williams et al. (2000) VT1-neutralizing antibodies was sixfold higher in dairy farm families (42 and 7.7%, respectively). Continued or recurrent exposure to VTEC may have resulted in a subclinical immunizing effect in the dairy farm residents. Exposure to the farm environment may result in more serious disease in urban visitors to the farm, children with declining maternal immunity, the elderly and the immunocompromised, due to less prior exposure to the organism (Wilson et al. 1996). Seasonal Influence Human infections associated with E. coli O157:H7 appear to be more common during the summer months, with the majority of cases occurring during the months of May through September (Pai et al. 1988; Ostroff et al. 1989; Waters et al. 1994; Boyce et al. 1995; Slutsker et al. 1997; Coia 1998; Mead and Griffin 1998; Chapman 2000). Ostroff et al. (1989) suggested that the increased shedding of E. coli O157:H7 by cattle during summer months may contribute to a similar seasonal pattern of E. coli O157:H7-associated foodborne illness in humans. Parallel seasonality of E. coli O157:H7 prevalence in cattle and illness in humans has subsequently been convincingly documented (Hancock et al. 1994, 1997a; Boyce et al. 1995; Chapman et al. 1997a; Hancock 1997; Mechie et al. 1997; Lynn et al. 1998, Van Donkersgoed et al. 1999). Incidence It has been estimated that E. coli O157:H7 infection accounts for more than cases of illness and result in as many as 250 deaths each year in the United States (Boyce et al. 1995; Coia 1998). Results of the recently initiated Foodborne Disease Active Surveillance Network (Foodnet) by the Center for Disease Control (CDC) reported the incidence of E. coli O157:H7 in the United States in 1996 to be 3 per population (CDC 1996). In 1999, a total of 38 outbreaks of E. coli O157:H7 infection in the United States were reported to the CDC. The outbreaks were reported from 30 states and affected 1897 people (CDC 2001b). The Laboratory Center for Disease Control (LCDC) in Canada reported 1334 cases associated with E. coli O157:H7 in 1995, which represents approximately 5 per population (LCDC 1995). In the United Kingdom, the highest rates of infection were reported in Scotland, with 9.8 cases per reported in 1996 (Coia 1998). Lower rates of infection (1.29 per population) were reported in England and Wales in 1996 (Coia 1998). MODES OF TRANSMISSION OF E. COLI O157:H7 TO HUMANS Food Vehicles Contaminated and undercooked cooked ground beef has been the principle vehicle implicated in E. coli O157:H7 infections (Borczyk et al. 1987; Bell et al. 1994). Fecal contamination of carcasses during slaughter and processing is the likely route by which E. coli O157:H7 is transferred to beef (Chapman et al. 1993a; Elder et al. 2000). Grinding the beef may further compound the problem by introducing the pathogen into the interior of ground meat patties where it is more likely to survive inadequate cooking (Boyce et al. 1995; Mead and Griffin 1998). Ground beef often comprises meat from many carcasses, thus a few infected animals could potentially contaminate a large quantity of ground beef (Boyce et al. 1995). Numerous foods other than ground beef have been linked to E. coli O157:H7 outbreaks (Table 1). Cross-contamination as a result of contact with bovine products or contamination with feces of wild or domestic animals has been suspected in the majority of these outbreaks (Armstrong et al. 1996; Mead and Griffin 1998). The levels of contamination in the food samples incriminated in E. coli O157:H7 outbreaks support the hypothesis of a low infectious dose (Armstrong et al. 1996; Neill 1997). A community outbreak associated with beef burgers in Wales in 1993 was used to estimate the presence of E. coli O157:H7 in the burgers. Fewer than two organisms per 25 g sample of meat caused illness, implying that the infectious dose of E. coli O157:H7 could be as low as 10 organisms (Willshaw et al. 1994). The infectious dose of E. coli O157:H7 is also affected by variables such as stomach ph, food composition and host susceptibility (Cassin et al. 1998; Gannon 1999).

8 Can. J. Anim. Sci. Downloaded from by on 12/08/ CANADIAN JOURNAL OF ANIMAL SCIENCE Water Outbreaks of E. coli O157:H7 infection have been associated with drinking water and with swimming (Swerdlow et al. 1992; Keene et al. 1994; Ackman et al. 1997; Paunio et al. 1999). In 1989, a large outbreak of E. coli O157:H7 infection occurred in Cabool, Missouri. This outbreak was associated with drinking water from an unchlorinated water supply (Swerdlow et al. 1992). The outbreak resulted in 243 cases of illness and four deaths. Epidemiological studies of three swimming-associated outbreaks implicated fresh-water lakes as the source of E. coli O157:H7 infection, with illness presumably due to bathers swallowing small amounts of contaminated lake water (Swerdlow et al. 1992; Keene et al. 1994; Paunio et al. 1999). An outbreak of E. coli O157:H7 infection in June 1998 was associated with visiting a water park near Atlanta, Georgia, and affected residents of eight states. Evidence suggested that contamination of park water by human feces was the source of infection (Georgia Department of Public Health 2002). In May 2000 the system supplying the drinking water to the town of Walkerton, Ontario, became contaminated, primarily with E. coli O157:H7 and Campylobacter jejuni, resulting in more than 2300 illnesses and seven deaths (The Walkerton Inquiry 2002). Person-to-person Transmission It has been suggested that the spread of E. coli within families and in institutionalized settings may occur through means other than foodborne transmission (Doyle 1991; MacDonald et al. 1996). Approximately 11% of the illnesses recorded during a large E. coli O157:H7 outbreak in the United States in 1993 due to contaminated beef burgers were determined to be secondary infections (Bell et al. 1994). In Wales, between 1990 and 1998, 17% of the identified cases of E. coli O157:H7 infection were the result of contact with an index case (Parry and Palmer 2000). The household transmission rate of E. coli O157:H7 infection in Wales between 1994 and 1996 was estimated to be 7%, with children aged 1 4 yr and adults aged yr most likely to contract infection from an index case through household contacts. This pattern of transmission suggests that personto-person spread is likely via fecal-oral routes, as opposed to foodborne transmission. Person-to-person spread of E. coli O157:H7 has also been identified as the primary means of transmission of infection in day care centers (Swerdlow and Griffin 1997). Person-toperson transmission has been identified in secondary cases of illness in a nursing home outbreak and has been implicated in an outbreak involving venison jerky (Ryan et al. 1986; Keene et al. 1997). Bovine-to-human Transmission Three incidents of apparent direct transmission of E. coli O157:H7 from bovines to humans have been reported (Renwick et al. 1993; Synge et al. 1993; Rice et al. 1996). In Ontario, Canada, in 1992, a 13-mo-old boy was hospitalized with hemorrhagic colitis following prolonged contact with calves on a dairy farm. In Scotland in 1993, a 15-mo-old child who lived adjacent to a farm became ill with E. coli O157:H7 (Synge et al. 1993). It was suggested that the family dogs may have carried the organism into the household, resulting in the infection of the child. An additional case of E. coli O157:H7 infection acquired from livestock occurred in a 10-yr-old boy who was actively involved in raising livestock (Rice et al. 1996). In 1999, an outbreak of E. coli O157:H7 infection among visitors to a petting zoo in Ontario resulted in 159 illnesses and four confirmed cases (Warshawsky et al. 1999). Direct contact with farm animals as a result of school visits to farms resulted in outbreaks of E. coli O157:H7 infection in Pennsylvania and Washington in the spring and fall of 2000 (CDC 2001a). Children were allowed to touch cattle, calves, sheep, goats, llamas, chickens and pigs. Lack of appropriate handwashing prior to eating and eating while petting the animals may have been associated with infection (Warshawsky et al. 1999; CDC 2001a). CONTROL OF E. COLI O157:H7 IN CATTLE Using a simulation model of E. coli O157:H7 infection in cattle to assess pre-slaughter intervention strategies, Jordan et al. (1999) reported that vaccinating cattle against E. coli O157:H7 and use of substances to reduce fecal shedding of the organism by cattle held the greatest potential for reducing contamination of beef carcasses. Elder et al. (2000) reported a positive correlation between fecal and hide prevalence of E. coli O157:H7 and subsequent contamination of carcasses with the organism. This suggests that an opportunity exists for controlling E. coli O157:H7 at the bovine source. There is ample opportunity for cross contamination once the animal enters the packing plant, thus intervention strategies to reduce the levels of E. coli O157:H7 in the feces of cattle may subsequently reduce the cases of foodborne illness associated with this pathogen. Effective control of E. coli O157:H7 requires a multi-faced approach. In addition to reducing the frequency and intensity of E. coli O157:H7 fecal shedding by cattle exposed to the organism, control of E. coli O157:H7 in environmental sources such as water troughs, feed and manure is critical in breaking the cycle of infection and re-infection of livestock. Diet The effect of diet on fecal shedding of E. coli O157:H7 has been investigated using experimentally inoculated sheep and cattle, but results are inconclusive. Sheep fed a highfiber, low-energy diet (grass hay) shed E. coli O157:H7 in larger numbers and for longer periods than did those fed a low-fiber, high-energy diet of corn and pelleted alfalfa (Kudva et al. 1995, 1997). The organism was shed for 8 to 15 d with the high-fiber diet, and for 2 to 6 d with the latter. Hovde et al. (1999) found that cattle fed hay shed E. coli O157:H7 in their feces for a significantly longer period (39 to 42 d) than did those fed grain (4 d). Other researchers have found no association between diet type and fecal shedding intensity or duration of shedding by inoculated calves (Tkalcic et al. 2000) or heifers (Magnuson et al. 2000). Interrelationships between feed quality, ruminal volatile fatty acid (VFA) concentrations and ruminal ph have been documented. High concentrate diets increase VFA concentrations and decrease ruminal ph as compared to ruminants

9 BACH ET AL. TRANSMISSION AND CONTROL OF E. COLI O Can. J. Anim. Sci. Downloaded from by on 12/08/17 consuming roughage-based diets (Owens and Goetsch 1988; Magnuson et al. 2000). Tkalcic et al. (2000) observed more extensive proliferation of E. coli O157:H7 in vitro in ruminal fluid from cattle fed a high roughage diet than a high concentrate diet, but did not observe this effect in vivo. Harmon et al. (1999) also found no correlation between daily fluctuations in ruminal VFA concentrations and ruminal or fecal E. coli O157:H7 numbers. In feces, the percentage of E. coli that were acid tolerant was higher with concentrate-fed cattle than with those fed hay (Diez- Gonzalez et al. 1998). Similarly, Hovde et al. (1999) determined that coliforms from hay-fed cattle were more sensitive to acid shock than were those from grain-fed cattle, but E. coli O157:H7 shed by the two groups of animals were equally acid resistant. Development of E. coli O157:H7-related illness in humans requires that the pathogen survive the acidic environment of the human stomach (Benjamin and Datta 1995). It was considered that feeding high-grain diets may increase the risk of human illness by increasing the acid resistance of E. coli, but findings to date are inconclusive. Thus, the significance of acid-resistant E. coli O157:H7 in bovine feces remains controversial. The effects of three different grain-based diets (85% barley, 85% corn or 15% whole cottonseed/70% barley) on fecal shedding of E. coli O157:H7 by experimentally inoculated steers was recently evaluated (Buchko et al. 2000a). Although diet did not affect the numbers of E. coli O157:H7 shed by the steers, fecal shedding was detected in a significantly higher number of steers when the barley diet was fed, as compared with the other two diets. In that study, E. coli O157:H7 was recoverable from mouth swabs of the steers, as well as from feed, water and manure. This suggests that this organism can survive in the environment and the alkaline saliva of cattle long enough to enable transfer among cattle (Buchko et al. 2000a). The bovine gastrointestinal tract is a complex system, thus survival of E. coli O157:H7 in this system is undoubtedly influenced by multiple factors. Further study is necessary to determine the effect of diet on E. coli O157:H7 in the gastrointestinal tract of cattle before any recommendations can be made in terms of diet as a means of controlling this pathogen. Fasting Although some studies have shown that fecal shedding of E. coli increased in response to dietary stress, feed withdrawal was found to exert little effect on E. coli O157:H7 shedding (Kudva et al. 1997; Cray et al. 1998; Jordan and McEwen 1998; Harmon et al. 1999; Buchko et al. 2000b). Fasting reduced ruminal VFA concentrations, with a resultant increase in the ruminal ph, but a correlation between VFA concentration and ruminal and fecal levels of E. coli O157:H7 was not observed (Harmon et al. 1999). However, diet-stressed calves were more susceptible to E. coli O157:H7 infection than well-fed calves (Cray et al. 1998). In cattle actively shedding E. coli O157:H7, feed withdrawal had little effect on growth of the organism in the rumen, or on the numbers shed in the feces (Harmon et al. 1999). This suggests that fasting prior to slaughter does not increase fecal shedding of E. coli O157:H7 or the number of E. coli O157:H7 positive animals entering the food chain. Probiotics The protective intestinal microflora that establishes in animals soon after birth is inherently very stable, but it is subject to influence by certain dietary and environmental conditions. These include antibiotic therapy, overcrowding, changes in feed, handling, shipping and new environments. When animals are stressed, the balance between beneficial and undesirable organisms may become upset resulting in diarrhea, gastroenteritis and reduced feed intake and growth (Fox 1988). Fuller (1989) defined probiotics as live microbial feed supplements which beneficially affect the host animal by improving its intestinal microbial balance. The concept of administering probiotics to animals involves the hypothesis that these beneficial microorganisms will combat the effects of stress and prevent undesirable microorganisms from establishing themselves in the intestine. The specific mode of action of probiotics may comprise a direct antagonistic effect against a specific group of organisms. Alternatively, suppression of bacterial numbers may be achieved by production of antibacterial compounds, or by competition for nutrients or adhesion sites. Probiotics may also alter microbial metabolism by affecting enzyme activity and/or by stimulating immunity, resulting in increased antibody levels and macrophage activity (Fuller 1989). Probiotic bacteria have proven effective for reducing the duration of ruminal carriage of E. coli O157:H7 in cattle. Eighteen bacterial isolates (Proteus mirabilis and 17 strains of E. coli) that produced metabolites inhibitory to E. coli O157:H7 were isolated from cattle not currently shedding E. coli O157:H7. The 18 bacterial strains were suspended in a 2% dilution of sterilized skim milk and administered orally to calves 2 d prior to their inoculation with E. coli O157:H7. Calves that received the probiotic bacteria carried E. coli O157:H7 ruminally for only 9 17 d, compared with d in the control calves. Fecal shedding was reduced from d in the control group to d in those receiving the probiotic (Zhao et al. 1998). Vaccination A potential intervention strategy for reducing carriage of E. coli O157:H7 by cattle may be vaccination. The goal of vaccination is either to reduce the susceptibility of cattle to colonization by E. coli O157:H7, or to decrease the duration of such colonization. In order for E. coli O157:H7 to be shed, it must first be able to survive in the gastrointestinal tract of cattle. Researchers believe that E. coli O157:H7 adheres to the wall of the large intestine by secreting virulence factors directly into host cells (Nataro and Kaper 1998). A vaccine that induces production of antibodies against these virulence factors could prevent adherence of the organism and result in its elimination from the gastrointestinal tract. In preliminary studies, Potter and Finlay (2000) reported that vaccinating cattle with two recombinant antigens, Tir and EspA, reduced fecal shedding of E.