Overview

Identification—Escherichia coli are members of the Enterobacteriaceae family, which is composed of aerobic and facultatively anaerobic, motile or nonmotile, gram-negative rods. Escherichia coli are resident microbes in the large intestine of warm-blooded animals, and hundreds of serotypes have been identified by means of surface structures on the bacterial cell membrane. Each serotype is designated by the somatic O antigen, which is determined by the sugar side chains on the lipopolysaccharide molecule in the gramnegative cell wall and by the flagellar H antigen, which is proteinaceous.¹ Nonmotile E coli serotypes are designated H– or NM. Although most strains are nonpathogenic, many have been associated with disease conditions in humans and animals, including pneumonia; meningitis; and intraabdominal, urinary tract, and enteric infections.

Enteric disease—Numerous serotypes of E coli cause diarrhea in humans. These are grouped into 5 categories according to virulence mechanisms: enterohemorrhagic, enteropathogenic, enterotoxigenic, enteroinvasive, and enteroaggregative. Of these, enterohemorrhagic strains are the most important human pathogens in industrialized countries. Enterohemorrhagic E coli are a subgroup of STEC that produce potent cytotoxins known as Shiga (also Shiga-like or Vero) toxins and cause a spectrum of disease in humans that ranges from mild diarrhea to HC and HUS or thrombotic thrombocytopenic purpura. Shiga toxin–producing E coli may produce 1 or both of the 2 types of Shiga toxins, Shiga toxin 1 and Shiga toxin 2. Shiga toxin 1 is identical to the toxin elaborated by Shigella dysenteriae, which is also associated with HUS²; however, STEC serotypes that produce only Shiga toxin 2 pose the greatest risk for development of HUS.³ After absorption through the intestinal wall, Shiga toxins cause HC and HUS by their action on the vascular endothelium in the intestinal submucosa and renal glomeruli.⁴ The most common STEC serotype in North America and the United Kingdom is E coli O157:H7, which is responsible for an estimated 73,000 cases of illness in the United States annually⁵ and is thought to cause over 90% of cases of diarrhea-associated HUS.² Hemolytic uremic syndrome is the primary cause of acute kidney failure in children in the United States. Approximately 8% of diagnosed STEC infections progress to HUS, mostly in children younger than 5 years and the elderly.⁵ Hemolytic uremic syndrome is fatal in as many as 5% of cases (ie, 61 cases annually on the basis of a 1999 estimate), and approximately 8% of patients who survive have lifelong sequelae, including neurologic impairment, blindness, paralysis, renal compromise, and the need for bowel resection.⁵ Most cases involve young children, but the risk of death is greatest in those over age 65.⁶ Thrombotic thrombocytopenic purpura is a similar condition that usually affects adults. Serogroups other than O157 are increasingly recognized as causes of illness, including HC, HUS, and thrombotic thrombocytopenic purpura. Of those, the O111 and O26 strains, in particular, are important pathogens. An additional 37,000 cases of STEC infection caused by non-O157 E coli strains are estimated to develop annually in the United States.⁷ In clinical case data collected in Canada through 1996, E coli O157:H7 were associated with 93% of human E coli infections, followed in frequency by the O55, O125, and O111 serogroups.⁸ In Austria and Germany, however, non-O157 STEC are responsible for 43% of E coli– associated cases of HUS.⁹ A study¹⁰ of German cattle revealed that O26 strains were more common than any other serotype, and in Australia, O111 strains are common, whereas O157 strains are rare.¹¹ A recent analysis of clinical case data¹² collected from 2000 to 2005 in Connecticut revealed that only 40% of STEC isolates were of the O157 serogroup. The remaining 60% were primarily of serogroups O103, O111, O26, and O45, but 15 other serogroups were isolated; the incidence of identified non-O157 STEC infection increased by 50% during this time. These pathogens are under continuous selection pressure from host immune systems and the environment to undergo frequent genetic rearrangements; therefore, it is highly likely that new genetic variants of virulent STEC will emerge in the future.¹³

Reservoir—Domestic ruminants, especially young cattle, are the most important reservoir for STEC. Other domestic and wild animals can harbor the organism, including pigs, deer, horses, dogs, rodents, and birds, but such animals with positive bacterial culture results have generally lived in close proximity to cattle. In an exception to this rule, rabbits were implicated during an outbreak in visitors to a wildlife center in the United Kingdom.¹⁴ Shiga toxin–producing E coli O157 is a transient member of the resident intestinal flora of cattle and is only rarely associated with clinical disease in neonatal calves (although O5, O26, and O118 serotypes do cause enteric disease in livestock).¹⁵ Cattle lack vascular receptors for Shiga toxin,¹⁶ but Shiga toxin 1 receptors have been detected in bovine intestinal epithelium and kidney.^17,18 The absence of vascular receptors could explain the lack of pathogenicity of most STEC strains in cattle. The bacteria do not invade the intestinal tract of ruminants, and adherence to the mucosa is not required for fecal shedding.¹⁹ Mean duration of detectable shedding in an animal is approximately 30 days but can range from a few days to a year.^20–22 Calves of weaning age are the most frequent shedders of STEC. Because domestic animals are subclinically infected and shed the organism intermittently, identification of infected animals is problematic.

Transmission—Because the organism is resistant to acidity, the infectious dose of STEC O157 is extremely low; < 10 viable cells are believed to be sufficient to induce disease in humans.²³ Transmission occurs most commonly through food contaminated with animal feces. Most outbreaks have been associated with undercooked ground beef, but water; vegetables; and unpasteurized milk, juice, and cider have also been implicated. According to a 1999 estimate by the CDC, 85% of STEC-caused illnesses were foodborne.⁷ The estimated overall incidence of infection has decreased by 50% in the recent past (from 2.1 cases/100,000 in 2000²⁴ to 1.06 cases/100,000 in 2005²⁵). Most of this reduction is probably attributable to increased control measures implemented by the USDA in abattoirs to limit the potential for foodborne infection.²⁶ Person-to-person transmission readily occurs, especially in childcare centers, and direct zoonotic and environmental transmissions are increasingly important. Environmental contamination is an important source of transmission because STEC is nearly ubiquitous on livestock farms and because STEC O157 can persist in soil and manure for as long as 5 months and in water trough sediments for at least 6 months.²⁷ Weather also contributes to environmental transmission, with heavy rainfall and warm temperatures associated with increased water-borne outbreaks²⁸ that develop from drinking or swimming in contaminated water. Vegetables can be contaminated by farm effluents or by irrigation with contaminated water. Contaminated radish sprouts caused the world's largest STEC O157 outbreak in Japan in 1996, which resulted in > 6,000 cases of illness.²⁹ In addition, STEC have been found in shellfish in coastal waters,³⁰ and insects, rodents, and wild birds may also be involved in transmission of the agent among animals and from animals to humans.^31–33

Testing and surveillance—Shiga toxin–producing E coli were first recognized as a cause of diarrheal disease in 1982 but did not gain widespread attention until 1993. Infections became nationally notifiable in the United States in 1994,⁵ and reporting has been mandatory in 48 states since 2000.³⁴ Some surveillance programs target only STEC O157, but other STEC serotypes commonly cause human illness and are not detected by testing for O157. To identify all STEC, testing must detect Shiga toxin or the underlying Shiga toxin genes in a sample. A Canadian study³⁵ revealed that testing clinical samples for Shiga toxin rather than for STEC O157:H7 increased the frequency of diagnosis of human STEC infection 3-fold (Shiga toxin was detected in 2.1% of stool samples, and STEC O157:H7 was detected in 0.6% of samples). Many reference laboratories worldwide do not have the capacity for complete serotyping of O and H antigens, which results in delayed reporting or absence of detection of emerging pathogenic serotypes.

Direct Zoonotic Transmission of Enteric Pathogens

History—A 1994 outbreak of STEC O157 infection in England was traced to cattle and goats at an open farm (a farm that is open to the public), and in 1995, a report³⁶ was published in England's Communicable Disease Report warning of the danger of zoonotic disease (including STEC) transmission in open farm settings. Contact with animals was first reported to be a risk factor in the United States in 1997 after a case-control study conducted by the CDC revealed that visiting a farm with cows in the 5 days prior to onset of illness was associated with increased risk of sporadic infection³⁷ (Table 1). Additionally, surveys conducted by FoodNet from 1996 to 1999 revealed that 2% of respondents reported visiting a petting zoo in the previous 7 days.³⁸ Generalization of these results to the larger population means that approximately 6 million humans are at potential risk of infection from animal contact at petting zoos each week.

Direct zoonotic transmission of enteric bacterial and protozoal pathogens (including STEC O157 and non-O157 strains and Salmonella, Cryptosporidium, Campylobacter, and Giardia spp) at venues where the public can come into contact with farm animals is widely recognized as a growing public health threat. The National Association of State Public Health Veterinarians conducted a study³⁹ of all published outbreaks of zoonotic disease traced to animal exhibits from 1966 to 2000. No reports dating from before 1994 were found, but 10 reports of enteric disease outbreaks from 1994 to 2000 were discovered; these outbreaks occurred in the United Kingdom, United States, and Canada and involved STEC O157 and Cryptosporidium and Salmonella spp (Table 2). Several reports of sporadic cases of disease were also discovered. The group then surveyed all state public health veterinarians in the United States to identify other outbreaks from 1990 to 2000. Twelve more outbreaks in that interval involved STEC O157; Campylobacter, Salmonella, Giardia, and Cryptosporidium spp; and multiple enteric pathogens, although 4 of those were attributed to consumption of raw milk rather than to animal contact. In a similar study, LeJeune and Davis⁴⁰ reported an additional 20 outbreaks of enteric disease that occurred from 1990 to 2002 and involved STEC O157 strains and Cryptosporidium and Salmonella spp. Overall, these 2 groups^39,40 of investigators confirmed 38 outbreaks of enteric disease from 1990 to 2002 that involved > 1,000 humans and that were associated with animals but not with food, milk, or water consumption. A literature search revealed another farm-associated cryptosporidiosis outbreak in England in 1989³⁶ and an article linking contact with calves to a case of human E coli infection in Canada in 1993.⁴¹ A review of enteric disease outbreaks associated with public animal venues in the United States from 1990 to 2005 was recently published.⁴²

Risk factors—The first large STEC outbreak associated with farm visits in the United States was reported in Pennsylvania in September 2000.⁴³ Fifty-one recent visitors to a dairy farm developed E coli O157: H7 infection. The farm had hosted visits by the public for decades and estimated that 1,500 to 2,000 people visited each day. Case-control³⁸ and environmental⁴³ studies were performed to determine the source of infection. After multivariate analysis, viewing calves < 6 weeks old was identified as a significant risk factor for infection, as did viewing calves 6 to 35 weeks old. The protective effect of handwashing was not significant (P = 0.08) but may have been clinically relevant (Table 3). In the process of obtaining control samples, investigators found that 4% of persons contacted had visited the farm during the outbreak period. Of those, 16% had developed diarrhea within 10 days after the visit, which was more than twice the expected rate of diarrhea in the general population per 10 days, indicating that there were likely many unidentified cases. The environmental investigation revealed that handwashing facilities at the farm were inadequate; 15% of all cattle and 23% of calves < 6 weeks old were shedding E coli O157:H7; and a biofilm sample from a water trough and a surface swab specimen from a fence were positive for growth of the organism on culture. This outbreak, along with a smaller episode in the state of Washington during the same year, led the CDC to conduct a survey of all state and territorial health departments. It was found that none had laws addressing the issue of zoonotic disease transmission at public venues with animals, and no federal laws existed to address this public health problem. These findings prompted collaboration with the National Association of State Public Health Veterinarians, the USDA Animal and Plant Health Inspection Service, and other groups to draft measures for reducing the risk for farm animal-to-human transmission of enteric infections.³⁸

Outbreaks involving multiple enteric pathogens took place at a farm day camp in Minnesota during 2 consecutive years.⁴⁴ The pathogens isolated were O157 and non-O157 STEC serotypes, Cryptosporidium parvum, Salmonella enterica, and Campylobacter spp; Giardia spp were isolated from calves but not from humans. Fifty-nine illnesses occurred in the summer of 2000, after which preventive measures were instituted (including removal of ill calves, installation of handwashing facilities, and provision of risk information to parents and counselors). Despite these measures, there was a second outbreak in 2001 that affected 25 children. In the first outbreak investigation, taking care of a sick calf was identified as a risk factor for infection by multivariate analysis (Table 3). Washing hands before going home was protective, as was always washing hands with soap after touching a calf. Analysis of data from the second outbreak revealed only visible manure on hands to be a risk factor. Always using alcohol-based handsanitizing gels was protective in univariate analysis but not in multivariate analysis. The authors suggested that it may be virtually impossible to prevent zoonotic transmission of enteric pathogens when children have such close and prolonged contact with young calves as they did in that environment. Substantial changes were made to the structure and operation of the farm, and no further outbreaks have been reported.

Each year, more than 125 million people in the United States visit state and county fairs,⁴⁰ which provide the only contact with livestock that much of the general public will have. Agricultural fair attendance was found to be a strong risk factor for STEC infection by an outbreak investigation in Ohio in 2000.⁴⁵ The Ohio Department of Health initiated an active case finding after recognizing unusually high numbers of STEC O157 isolates in samples submitted from several counties. Attendance at any 1 of 4 agricultural fairs in the region was identified as the primary risk factor for infection (Table 3). Afterward, the investigators examined STEC O157 surveillance data from 1999 and found that the dates of reported infections were significantly associated with exposure to agricultural fairs, supporting the pattern seen during 2000. They concluded that agricultural fairs likely represent an important and underappreciated source of STEC infection and contribute to the yearly peak of cases reported during the summer months.

An outbreak of STEC O157 infection in Texas in 2003⁴⁶ affected 22 visitors and 3 exhibitors at a local agricultural fair. All 25 persons involved had contact with livestock exhibits, and 12 of the 25 had contact with swine exhibit areas. Multivariate analysis revealed that visiting livestock areas was significantly associated with illness, as was number of days of fair attendance, with risk of illness increased by 51% per additional day (Table 3). Environmental samples had growth of STEC O157 from 10 of 62 sites sampled 46 days after the fair ended. Five of the 46 livestock areas were contaminated. Swine, goat, and lamb areas had negative culture results, but common-use areas and the show arena yielded positive culture results for isolates that matched the PFGE pattern of the outbreak strain. Seven STEC O157 subtypes were identified from multiple sites at the fairground; however, the subtype matching the human outbreak PFGE pattern was isolated from only 4 livestock areas. This incident was unusual in that it is the only report of an outbreak in which livestock exhibitors were affected. Persons having frequent contact with livestock may develop protective immunity against STEC,^47–50 and the continuing decrease in the proportion of the United States population having regular contact with livestock may be contributing to the observed increase in animal-associated outbreaks. Of the 3 affected exhibitors, 2 showed a pig and 1 showed a lamb. Although pigs and lambs occasionally harbor STEC, human infections are not frequently associated with these species. The 3 exhibitors did have contact with commonuse livestock areas and the show arena and may have had direct contact with cattle or cattle areas at the fair.

In 2004 and 2005, E coli O157:H7 outbreaks associated with petting zoos in North Carolina, Florida, and Arizona were reported.⁵¹ The largest of these outbreaks occurred in North Carolina during October 2004 among visitors to the North Carolina State Fair. There were 108 cases of illness, including 15 cases of HUS. Environmental sampling indicated extensive E coli O157 contamination at 1 of the 2 petting zoos. A casecontrol study revealed that illness in children < 6 years old was associated with contact with manure, falling or sitting on the ground, and hand-to-mouth activity such as thumb-sucking or sucking on a pacifier (Table 3). Alcohol-based hand-sanitizer use was not protective; however, awareness among adults accompanying children of a risk for disease from contact with livestock was protective. The CDC subsequently published the Compendium of Measures to Prevent Disease Associated with Animals in Public Settings, 2005 on the basis of findings from this and other investigations.⁵²

Environmental Transmission of STEC

Exposure to animal environments—Three prospective case-control studies in the United Kingdom examined risk factors for sporadic infection with STEC O157. The first⁵³ revealed contact with a farm environment to be an important risk factor (Table 1). This contact included recreational visits by the public to farms and petting zoos, staying on farms for vacations, and work-related visits to farms. The work-related visits generally did not involve touching animals, and only half of the humans exposed through recreational visits reported touching farm animals. Farmers were found not to be at increased risk. The authors concluded that environmental transmission of infection to humans is more important than is generally recognized. In the second study,⁵⁴ environmental risk factors were investigated and results indicated that only likely contact with and contact with animal feces (excluding that from pets) remained significant risk factors after multivariate analysis. Researchers in that study⁵⁴ were intrigued by the unexpected finding that drinking bottled water was strongly protective, suggesting that this result merits further research. Authors of the third study⁵⁵ reported that significant risk factors for sporadic infection included visiting a private farm and having a household member with an occupation involving contact with farm animals. The results of these 3 studies indicate that contact with animal environments, even when indirect, may be as important as direct contact with animals for risk of illness associated with STEC.

Environmental contamination and persistence—It is clear that the environments of domestic ruminants can be an important reservoir for STEC and pose a continual risk of exposure for humans. The variable prevalence of animals with fecal shedding, seasonal pattern of shedding, and inability to routinely culture the organisms from animals and the environment all serve to complicate efforts to study the ecology of STEC. Davis et al⁵⁶ found bedding to be the type of environmental sample that most commonly had positive results of bacterial culture, which sometimes persisted for weeks after culture of feces from the animal residing in the stall yielded negative results. Although isolates were less commonly obtained from water bucket, feed, and coat samples, these substrates were all occasionally found to have bacterial growth, even when the associated animals were not actively shedding the organism at detectable concentrations in feces. Results of that study⁵⁶ also revealed that STEC O157 strains can use bovine urine as a growth substrate. Williams et al⁵⁷ found that persistence of organisms is greatest on moist wood in cool temperatures, with large numbers remaining after 28 days. Numbers declined rapidly on galvanized steel in warm temperatures, but moisture content had a greater effect on survival than did type of material. The latter study also revealed that a mean of 4,000 CFUs of STEC O157 were transferred to a participant's hand by grasping a galvanized steel gate that had been contaminated with a known quantity of E coli–inoculated bovine feces.

Association with agriculture—In an ecologic study⁵⁸ in France, the relationship between cattle density and incidence of HUS in children was analyzed. Multivariate analysis revealed a positive association between HUS incidence and rural degree (total cultivated land/total surface), dairy cattle density, and ratio of calves to children. Two similar studies were performed in Canada, one of which revealed a higher incidence of human STEC infection in rural areas than in urban areas of Ontario and a spatial association between cattle density and human STEC infection.⁵⁹ The second Canadian study⁶⁰ revealed the factors most strongly associated with human STEC infections to be the ratio of beef cattle to human population and the application of manure to agricultural land.

Further evidence of the risk of environmental exposure to STEC resulting from application of manure to land was reported in Austria in 2003, when 2 cases of HUS caused by an identical strain of E coli O157:H7 in unrelated children from the same village were investigated.⁶¹ The 2 children had had no contact, and neither had visited the nearby cattle farm; however, bacterial culture of feces from 1 cow on the farm yielded positive results for an STEC isolate with a PFGE pattern identical to that isolated from the children, as did the asymptomatic mother of one of the children. Manure from the farm had been spread on an adjacent meadow, which sloped down to a wooded area. One child had played in the meadow, and the other had played in the wooded area a few days before the onset of illness. The area had been affected by heavy rainfall a few days before onset of illness. Investigators theorized that the rain washed the bacteria from the meadow into the wooded area and contributed to transmission of infection, although contamination of the area from infected wild animals could not be ruled out.

Another outbreak of enteric disease that was linked to environmental contamination took place in Scotland in 2000, when a Scout camp was held on an agricultural showground where sheep had grazed until the day before the camp.⁶² There were 337 campers at the event, which was abandoned a day early because of heavy rainfall. A STEC O157 isolate was the primary pathogen involved, but Cryptosporidium and Campylobacter spp were also isolated from some campers. An identical strain of STEC O157 was isolated from 14 of the 30 sheep, and Cryptosporidium spp was isolated from 5 sheep. It was concluded that the source of STEC O157 in that outbreak was the environment and that transmission occurred by contaminated hands, either directly from hand to mouth or via food. Infection was more likely in persons who did not wash their hands before eating and who ate without using cutlery (Table 3). Results of culture of environmental samples remained positive for growth of STEC O157 15 weeks later. This incident led the Scout Association and the region's Environmental Health Services to develop guidelines regarding selection of camping sites and the importance of personal hygiene.

Two large outbreaks of enteric disease were traced to county fairs in the United States and were found to be the result of environmental contamination. The first took place in New York in 1999 and involved STEC O157 and Campylobacter jejuni infections in > 1,000 fair visitors.⁶³ Investigators concluded that manure from a single cow with positive fecal culture results for STEC was washed by recent heavy rainfall into a nearby shallow well, which supplied water to some of the fair's vendors. The second outbreak took place in Ohio in 2001.⁶⁴ In that outbreak, case-control and environmental studies determined that sawdust in a livestock show barn had been contaminated with STEC O157. The barn was the site of a dance held on the last night of the fair, which resulted in dance attendees being exposed to the disease agent in airborne dust. All 6 sawdust samples collected from the barn 42 weeks later still had positive results of STEC culture, 5 of which matched the molecular fingerprint pattern of the outbreak strain.

Prevalence

Cattle—Reported prevalence of STEC O157 in the United States varies from 0% to 28% in individual cattle and from 0% to 72% in herds, with higher rates of shedding seen in the summer months.⁶⁵ More recent surveys that included use of enzyme immunoassay and immunomagnetic separation techniques (which are up to 100 times as sensitive as culture-based techniques^66,67) reveal a considerably higher prevalence than was previously reported. Assays for Shiga toxin in fecal samples also detect a significantly higher prevalence than those that target STEC O157. Neonatal calves and calves at weaning age have the highest prevalence of shedding STEC. In a naturally infected beef herd in Canada,⁶⁸ 25% of calves were actively shedding E coli O157:H7 within 7 days after birth. One week before weaning, the prevalence had decreased to 0% to 1.5% in those calves, but within 2 weeks after weaning, prevalence had increased to 6% to 14%. Cattle apparently do not acquire permanent immunity against STEC infection, and reinfection with the same strain can occur after an infection is cleared. Calves that are reinfected have reduced shedding patterns that are similar to those of an initial infection in adult cattle, indicating that the reduction in shedding is related to changes in normal intestinal flora with age rather than to acquired immunity.⁶⁹ There is evidence, however, that a small percentage of calves are persistent high shedders.⁶⁹

Authors of a prospective study⁷⁰ in France assessed the prevalence and characterization of STEC isolated from cattle, food, and children and found that 70% (330/471) of fecal samples from healthy cattle at slaughter contained the Shiga toxin gene or genes, with significantly higher rates detected in August than at other times of the year. This seasonal pattern was also seen in isolates cultured from samples of beef and cheese. Shiga toxin genes were found in 3% (19/658) of hospitalized children in the region; of those, only 3 children had signs of disease that could be attributed to STEC infection. Shiga toxin 2 strains were more commonly recovered from cattle fecal samples and samples of beef, and a significant association was observed between the Shiga toxin 2 gene and highly cytotoxic isolates. An STEC isolate was obtained in 34% (162/471) of bovine fecal samples. Of those, 30% (48) were serotypes known to be associated with severe disease in humans. A notable finding was that no association could be established between presence of Shiga toxin genes and diarrheal disease in hospitalized children.

Wildlife—Wild animals occasionally harbor STEC. Researchers in Georgia⁷¹ found that experimentally inoculated white-tailed deer had patterns of fecal STEC O157 shedding similar to those in calves. The organism was recovered from 0.5% of fecal samples from free-ranging deer and from 4.3% of cattle at the same site. The authors concluded that it is unlikely that deer serve as a reservoir for STEC in the southeastern United States but that the same precautions should be taken when handling deer carcasses and venison as are recommended for handling beef and pork. Another group⁷² found the prevalence of STEC O157 to be 2.5% in white-tailed deer in Kansas, which was similar to the rate in cattle in the same area. A prevalence of 0.25% was found in free-ranging deer in Nebraska,⁷³ but the prevalence of shedding in cattle in that region was not reported. Other studies have revealed nonruminant wildlife to be colonized by STEC. Samples of intestinal contents from a pigeon and a raccoon from separate Wisconsin dairy farms yielded STEC O157 isolates⁷⁴; interestingly, the raccoon lived on a farm with no STEC-shedding cattle. An investigation of wild animals living in close proximity to cattle farms in Denmark³² revealed a low prevalence overall, but a Norway rat and a starling were discovered to be shedding an STEC strain that was identical to the cattle isolate. Again, the researchers concluded that wild animals most likely do not serve as an important reservoir for the organism, but can become infected from farm animals and vice versa and possibly have a role in transmission of infection to domestic animals and humans.

Agricultural fairs and dairy farms—In a study³³ of STEC O157 prevalence at 32 state and county agricultural fairs, livestock, animal barns, and insects were examined. Of all livestock fecal samples tested, 6.4% had positive culture results, including 11.4% of cattle samples (13.0% of beef cattle and 4.1% of dairy cattle). Additionally, 1.2% of pigs, 4.4% of sheep, and 1.8% of goats had positive results for culture of STEC O157. Poultry, rabbits, equids, and camelids all had negative culture results. Of fly pools originating from beef barns, a swine barn, and an outdoor manure pile, 5.2% had positive culture results. Overall, STEC O157 was isolated from 6.3% of all samples and from 31 of 32 (96.9%) fairs. Environmental samples were tested again 10 to 11 months after the fairs ended, and 4 samples with positive results were obtained from beef barns at 3 fairgrounds. Husbandry and management of fair animals are much different from those of commercial animals because fair animals and their environments are generally kept clean, yet the prevalence of STEC shedding in beef cattle and swine at fairs is similar to the overall prevalence in US commercial herds. The indication is that animal hygiene is unlikely to substantially affect public exposure to STEC at agricultural fairs.³³

In 2001 and 2002, researchers in Minnesota collected manure samples from conventional and organic dairy farms and county fairs for Shiga toxin detection.⁷⁵ Shiga toxin–producing bacteria were found in 2.3% of samples from conventional farms, where 65% of farms had at least 1 animal with positive results, and in 6.6% of samples from organic farms, where 87.5% of farms had at least 1 animal with positive results. Shiga toxin was also detected in 17.4% of samples from county fairs. Of the 71 samples that had positive results for Shiga toxin, 46 STEC isolates belonging to 26 different serotypes were recovered.

Potential Control Measures in Cattle

Cattle manure, of which more than 1.2 billion tons is produced annually in the United States,⁷⁶ is the most important source of STEC infection for humans and other animals. The rate of STEC shedding ranges from < 100 to > 1 million CFUs/g of feces.⁶⁵ Measures to reduce fecal shedding are being investigated as a means of reducing human exposure because the transient and nonpathogenic nature of the infection in cattle makes testing and removal or treatment of infected animals impractical as a control measure. Multiple potential methods of limiting STEC shedding in cattle are being investigated, including vaccination, administration of probiotics, administration of antimicrobials, bacteriophage treatment, feeding alterationsand dietary supplements, and changes in management.

Vaccination—Some infected calves develop antibodies against the O-polysaccharide antigen and Shiga toxin 1 (but not Shiga toxin 2) of STEC, but these antigens do not correlate with clearance of infection or prevention of reinfection with and subsequent shedding of the same strain,^77,78 so a vaccine containing these antigens would not likely be effective. A vaccine containing intimin of STEC O157:H7 has some promise. Vaccination with a preparation containing this antigen reduced colonization and intestinal lesions in suckling piglets of a vaccinated sow,⁷⁹ and a vaccine containing type III secreted proteins of STEC O157:H7 significantly reduced the amount and duration of fecal shedding in cattle in experimental conditions and the prevalence of shedding in feedlot cattle.⁸⁰ Some investigators, however, consider vaccination unlikely to be successful given that mucosal colonization is not required for fecal shedding.¹⁹

Use of probiotics—Probiotics are suspensions of commensal bacteria introduced into the gastrointestinal tract to reduce the number of pathogens by either competitive or antagonistic mechanisms. Administration of probiotic bacteria can substantially reduce STEC shedding in neonatal⁸¹ and weaned^21,82 calves by means of competitive exclusion. Administration of probiotics has had mixed results in practice, possibly in part because of conflicting management techniques such as administration of antimicrobials at subtherapeutic dosages.⁸³ In some European countries in which subtherapeutic antimicrobial usage in production animals was banned altogether, the incidence of disease initially increased in some food animal species and necessitated an increase in therapeutic antimicrobial use, but changes in husbandry and management practices were effective countermeasures against this effect.⁸⁴ Administration of probiotics may prove to be useful as such countermeasures in future US food animal production as pressure increases for producers to phase out use of subtherapeutic doses of antimicrobial feed additives.

Administration of antimicrobials—Treatment with antimicrobials can be used therapeutically to reduce STEC in the bovine intestinal tract. One study⁸⁵ revealed that oral treatment with neomycin sulfate reduced shedding to undetectable levels. In another study,⁸⁶ it was reported that adding the treatment to feed for 3 days reduced shedding of STEC O157 in commercial feedlot cattle from 46% to 0%. Neomycin has a 24-hour meat withdrawal period and, hence, could be administered shortly before slaughter for the purpose of reducing abattoir contamination, but it is not currently approved for this use by the FDA. Neomycin is not an important antimicrobial with regard to use in human medicine, but the issues of resistance and crossresistance must be considered given that other aminoglycoside antimicrobials are used in humans.

Antimicrobial resistance in E coli is an important concern at present and has been correlated in some studies with antimicrobial administration in production animals. In 1 US study,⁸⁷ 40% of all E coli isolates from dairy cattle were multidrug resistant; in another study,⁸⁸ two thirds of all STEC O157 isolates from southwestern US dairy farms were multidrug resistant. Among Wisconsin dairy farms, the greatest antimicrobial resistance was found in STEC isolates from cows on farms where antimicrobials were used the most,⁷⁴ and significantly higher levels of resistance for 10 of 17 antimicrobials were found in E coli isolates from conventional versus organic farms.⁸⁹ Escherichia coli is thought to be an important carrier of resistance genes in fecal flora.^90,91 Norwegian researchers⁹¹ found that antimicrobial treatment targeted at a pathogen in 1 organ system can alter resistance traits in the endogenous flora of other organ systems in a given animal and in the environment. The benefits of using antimicrobials to selectively eliminate STEC from food animals would therefore have to be weighed carefully against the risks.

Treatment with bacteriophages—Lytic bacteriophages are highly specific bactericidal viruses; colicins are proteins produced by bacteria of certain strains of E coli and related Enterobacteriaceae that inhibit susceptible strains of E coli. Phages have been used successfully as antibacterial agents in both human and veterinary medicine. Experimental studies in which phages^92,93 and colicins^94,95 were used to target STEC O157 have had more success in vitro than in vivo; however, Raya et al⁹⁶ reported a 100-fold reduction of STEC O157 in sheep within 2 days after receiving a single oral dose of bacteriophage, compared with control sheep. Although such research findings are promising, more work is needed to determine whether treatment with phages is a viable option for reducing STEC shedding in cattle.

Chlorate—Facultatively anaerobic bacteria such as the Enterobacteriaceae can respire by reducing nitrate to nitrite via the enzyme nitrate reductase. Nitrate reductase does not differentiate between nitrate and chlorate as a substrate. Chlorate is reduced to chlorite, and accumulation of chlorite is bactericidal.⁹⁷ Sodium chlorate supplementation in feed experimentally reduced STEC O157 populations in swine,⁹⁸ and chlorate added to drinking water for 24 hours significantly reduced STEC O157 populations in cattle and sheep,^99,100 indicating that oral chlorate supplementation could potentially be a feasible control strategy.

Supplementation with Ascophyllum nodosum— Administration of a commercial feed supplement derived from a type of brown seaweed may benefit animal health and carcass quality.^101,102 Supplementation improves immune function^103–105 and reduces the establishment and shedding of STEC from the digestive tracts of cattle. Feeding the product to experimentally infected cattle in a feedlot environment reduced the numbers of STEC O157 shed by 10 to 100 times, compared with control cattle,¹⁰⁶ and supplementation in a large commercial feedlot for 14 days prior to slaughter significantly reduced the frequency of hide and fecal samples containing STEC O157.¹⁰⁷

Animal management—Given that rates of fecal shedding are similar among dairy, range, feedlot, and show cattle,^74,108,109 animal management and hygiene probably have little effect on STEC bacterial loads. However, water troughs (especially those containing sediments) can harbor the organism for extended periods of time,¹¹⁰ and regular cleaning of troughs and treatment of water by chlorination or ozonation may help reduce horizontal transmission and reinfection within a herd. Feeding changes can temporarily alter rumen and intestinal microflora, which may have a practical use for reducing shedding at particular times.¹¹¹ The type of grain fed can substantially affect STEC shedding,¹¹² and when cattle are abruptly switched from grain to a 100% hay diet, fecal E coli populations decline substantially. Diez-Gonzalez et al¹¹³ reported a 1,000-fold decline in fecal E coli populations within 5 days of switching to an all-hay diet, and acid-resistant E coli populations (which pose a greater risk for human enteric infection) declined 100,000-fold. Although the effect is shortlived, it may be useful at certain times.

Heterogeneous shedding—A group of researchers in the United Kingdom¹¹⁴ studied the concentration and prevalence of STEC O157 in cattle feces at the time of slaughter. The individual prevalence was 7.5%, whereas group prevalence was 40.4%. Of the 44 infected cattle detected, 9% shed high numbers of organisms, with concentrations of STEC O157 > 10⁴ CFUs/g in feces. Organisms shed by those 9% represented > 96% of the total STEC detected in all cattle tested. Most (61%) shedding cattle were low shedders (< 10² CFUs/g).

These results were supported by those of a later study,⁶⁹ in which the role of heterogeneous shedding on the transmission dynamics of STEC O157 was assessed. A model was developed to relate infectiousness to bacterial count under a range of assumptions regarding withinand between-host variability in bacterial carriage. Analysis of prevalence data from Scottish cattle farms revealed a highly skewed distribution across cattle groups, indicating that a small number of cattle have much higher transmission rates. By obtaining bacterial counts for samples with positive results from 474 cattle farms and fitting variables into the model, the investigators found that approximately 20% of infected cattle would be responsible for approximately 80% of transmission. These results have interesting implications for control. In that study,⁶⁹ the authors determined that preventing infection in the 5% of cattle with the highest level of shedding would be sufficient to control the spread of infection. Also, given that high levels of shedding tend to be associated with intestinal colonization and persistence, they suggested that testing and removal of high shedders from the herd could be an effective control measure. Moreover, interventions, such as administration of probiotics, undertaken to limit colonization and persistence of STEC in the gastrointestinal tract may be effective in preventing high bacterial loads and would thereby limit transmission of the organism. The model predicted that preventing very high bacterial counts (10⁴ to 10⁵ CFUs/g) would effectively limit transmission.

Conclusions

Direct zoonotic and environmental transmission of STEC is a growing public health concern. Most human infections continue to be foodborne in origin, although the USDA's implementation of a Pathogen Reduction/Hazard Analysis Critical Control Point in abattoirs from 1997 to 2000, along with additional requirements placed in 2002, apparently substantially reduced the burden of human disease.²⁶ Just as consumers must be responsible for properly handling and preparing food to avoid foodborne disease, they must also be responsible for protecting themselves and their children in situations in which there is contact with potentially infectious animals. Escherichia coli will not be eliminated from animals; therefore, educating the public about risk remains the most effective means of preventing human infection, whether foodborne or zoonotic. Elimination of all petting zoos has been proposed as a response to the problem, but this is shortsighted and an overreaction to an uncommon route of infection. If, as was estimated in the FoodNet survey, 6 million people in the United States visit a farm or petting zoo each week,³⁸ the percentage of visitors that actually contract zoonotic infection at these venues is very low. The crude attack rate in the 2003 county fair outbreak in Texas was 0.015%,⁴⁶ which was almost identical to that of the 2004 North Carolina State Fair outbreak (0.014%).⁵¹ For perspective, this rate is similar to the incidence of severe adverse reactions to administration of some common pharmaceuticals. Handwashing is by far the most important preventive measure, but the frequency with which it is undertaken depends on awareness of a risk. The investigation of the North Carolina outbreak revealed that knowledge of a risk for disease associated with animal contact was protective. Among the study cohort that visited the implicated petting zoo, awareness among adults was highly protective for all visitors and for children < 6 years, even in the absence of readily accessible handwashing facilities.¹¹⁵ This finding indicates that effective public education would substantially reduce the risk for both adults and children. Education should be provided by the animal venues along with appropriate venue design, provision of handwashing facilities, and operation as detailed in the Compendium of Measures to Prevent Disease Associated with Animals in Public Settings, 2007.¹¹⁶ Children are still more likely to get sick from contact with other children than from contact with animals, and it would be unfortunate to deny them the recreational and educational value of animals as a result of unjustified fear. Veterinarians have an important role to play in providing the public with accurate information and allaying unnecessary fear.

The problem of environmental contamination will be more difficult to solve. Practical measures to reduce shedding in livestock and more careful disposal of animal waste are needed. For this, more research will be necessary to discover cost-effective and practical means of limiting infection and shedding in livestock. Also, producers should be educated regarding their role and should take responsibility in addressing the problem. Veterinarians are in a unique position to educate both the general public and producers on preventive measures. Foodborne and animal-associated outbreaks garner much negative publicity, result in high costs from morbidity and litigation, and create public anxiety. Veterinarians in all sectors are expected to be experts on zoonotic diseases and are frequently asked for information and advice. Knowledge of this increasing risk represents an opportunity for veterinarians to make an important contribution to public health.