|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||
Background Outbreaks of Escherichia coli O157:H7 infections have involved direct transmission from animals and their environment to humans. We describe an outbreak among visitors to a Pennsylvania dairy and petting farm that provides public access to animals.
Methods We conducted both a case–control study among visitors to a farm to identify risk factors for infection and a household survey to determine the rates of diarrheal illness among these visitors. We performed an extensive environmental study to identify sources of E. coli O157:H7 on the farm.
Results Fifty-one patients with confirmed or suspected E. coli O157:H7 infection were enrolled in the case–control study. The median age of the patients was four years, and the hemolytic–uremic syndrome developed in eight. Contact with calves and their environment was associated with an increased risk of infection, whereas hand washing was protective. The household survey indicated that visitors to the farm during the outbreak had higher than expected rates of diarrhea. Environmental studies showed that 28 of the 216 cattle on the farm (13 percent) were colonized with E. coli O157:H7 that had the same distinct pattern on pulsed-field gel electrophoresis that was found in isolates from the patients. This organism was also recovered from surfaces that were accessible to the public.
Conclusions In a large outbreak of E. coli O157:H7 infections among visitors to a dairy farm, predominantly children, high rates of carriage of E. coli O157:H7 among calves and young cattle most likely resulted in contamination of both the animals' hides and the environment.
Outbreaks of E. coli O157:H7 infection caused by transmission from farm animals or their environment were recognized in the United States during 2000.9 We describe an outbreak that is notable for two reasons: the number of cases permitted extensive characterization of risk factors, and a concurrent environmental study of the farm was conducted to define sources of infection.
Methods
The Outbreak
During September 2000, an unexpectedly large number of cases of E. coli O157:H7 infection occurred in Montgomery County, Pennsylvania, and surrounding counties. Officials at the Montgomery County Health Department and the Pennsylvania Department of Health instituted active case finding on September 27. Most patients had been visitors at a popular petting farm in Montgomery County. The farm was a traditional dairy farm that for several decades had hosted visits by the public, especially groups of children, to see and pet the animals. The farm's management estimated that 1500 to 2000 people visited the farm each day. Food and beverages were available at the farm for lunch and snacks. By November 2, 6 patients with the hemolytic–uremic syndrome and 10 patients with diarrhea from Pennsylvania and New Jersey had been reported among farm visitors. The Pennsylvania Department of Health and the Montgomery County Health Department invited the Centers for Disease Control and Prevention (CDC) to assist in the investigation on November 2. The goals of the investigation were to determine the magnitude of the outbreak, to identify risk factors for infection, to interrupt transmission, and to describe the ecology of E. coli O157:H7 in the farm environment.
Epidemiologic Investigation
Surrounding counties and all states were notified of the outbreak. Physicians and medical microbiology laboratories were asked to look for cases associated with the outbreak, to test stools for E. coli O157:H7, to report cases to the Montgomery County Health Department, and to forward isolates to state health department laboratories for molecular subtyping.
To develop hypotheses regarding possible sources of E. coli O157:H7 infections, investigators selected and interviewed three patients on November 6. They used the CDC standard foodborne-illness hypothesis-generating questionnaire (available at http://www.cdc.gov/ncidod/dbmd/outbreak/stand_qu.htm) with an addendum relating to exposures to animals.
Hypothesis-generating interviews suggested that the outbreak occurred among visitors to the farm. Consequently, a case–control study was conducted among farm visitors to identify specific risk factors for infection.
For case-ascertainment purposes and a case–control study, a probable case was defined as acute diarrhea (three or more loose stools in a 24-hour period) in a person beginning within 10 days after visiting the farm after September 1. A confirmed case was defined as acute diarrhea in a person beginning within 10 days after visiting the farm after September 1, and accompanied by either the hemolytic–uremic syndrome or the isolation of E. coli O157:H7 from stool.
Controls were persons who visited the farm after September 1 and did not have diarrhea within 10 days after the visit. Controls were identified by sequential-digit dialing in which patients' telephone numbers were used as a starting point, and all controls were frequency-matched to the patients according to age group (less than 1 year, 1 to 4 years, 5 to 8 years, 9 to 12 years, 13 to 20 years, and 21 years or older). One person per household was included. The goal was to obtain two controls per patient.
A questionnaire was developed that focused on exposures to animals, food and beverages consumed at the farm, and hand-washing practices. The questionnaire was administered by telephone from November 12 through 19. Patients and controls were asked about the dates of their visit to the farm, contact with animals and their environment, foods and beverages consumed, and hand washing and other behavior at the farm. Patients were also asked about their clinical history. When a patient or control was 15 years of age or younger, the interview was conducted with the parent or guardian who had been present during the visit to the farm.
To estimate the rate of illness among petting-farm visitors, a household survey was conducted during the selection of controls for the case–control study. Members of contacted households were asked whether they had visited the farm and those who had were queried about any history of acute diarrhea in the 10 days after the visit.
Environmental Investigation
The layout of the farm was determined, and a complete list of food and beverages available to visitors was compiled. A census of the animals present at the farm during September and October was conducted. The farm's entire population of domestic animals was subjected to rectal- or cloacal-swab sampling between November 13 and 16. Samples were cultured for E. coli O157:H7 and subtyped with the use of pulsed-field gel electrophoresis (PFGE).
Swabs of the railings around animal enclosures were obtained. Water samples were collected at sites around the farm. Biofilm samples were collected from watering troughs. All samples were cultured for E. coli O157:H7 and subtyped with the use of PFGE.
Laboratory Investigation
Human fecal samples and biofilm samples from watering troughs were plated on sorbitol–MacConkey agar after immunomagnetic separation. Rectal and cloacal swabs from animals and water samples were plated on sorbitol–MacConkey agar. Immediately after collection, all swab samples were placed in Carey–Blair transport medium and were cultured within 48 hours. Isolates of E. coli O157:H7 were confirmed serologically and tested for toxin production with the use of an enzyme immunoassay at the Pennsylvania State Public Health Laboratory. Environmental-surface swabs were forwarded to the CDC for culture for E. coli O157:H7. Molecular subtyping of all E. coli O157:H7 isolates was performed at the Virginia State Public Health Laboratory, the Pennsylvania State Public Health Laboratory, and the CDC with the use of standard methods of PFGE.10,11 Shiga toxin genes were detected by multiplex polymerase chain reaction with the use of established primers.12
Statistical Analysis
Statistical analyses were conducted with the use of Epi Info, version 6.04 (CDC), SAS System for Windows, release 8.0 (SAS Institute), and LogXact for Windows, version 4.1 (Cytel Software). A multivariate logistic-regression model was built from the set of measured quantities to identify independent variables that were significant risk factors for infection. Factors considered for inclusion in the final model were demographic variables, direct contact with animals, environmental exposures, hand–mouth activities, foods and beverages, and hand-washing behavior. Where appropriate, combined exposure variables were defined. The analysis was stratified according to age to account for differences in exposure that might have been the result of age.
Results
Epidemiologic and Clinical Information
As of November 12, 2000, 15 confirmed cases and 36 probable cases of E. coli O157:H7 had been identified. Patients were from five counties in eastern Pennsylvania and one county in New Jersey. No residents or employees of the farm reported having diarrhea during the outbreak period. Of the 51 patients, 25 (49 percent) were female. They ranged in age from 1 through 52 years (median, 4), and 47 (92 percent) were 10 years of age or younger. Dates of onset could be accurately determined for 49 patients and ranged from September 4 through November 8, 2000 (Figure 1). Patients reported having bloody diarrhea (37 percent), fever (45 percent), and vomiting (45 percent). Sixteen patients (31 percent) were hospitalized. The hemolytic–uremic syndrome developed in eight patients (16 percent), all of whom were 10 years old or younger. End-stage renal failure developed in one of these eight patients, and renal transplantation was required. No patient died. Within days after the farm and public health officials prohibited public access to animals on November 4, no further cases were reported (Figure 1).
|
|
|
Environmental Investigation
The farm was a small, working dairy farm. Calves younger than 6 weeks of age were kept in hutches, and calves 6 to 35 weeks of age were kept in a barn. Both areas were fully accessible to the public. Heifers and mature cows were kept separate from the calves and were less accessible to the public. In addition, pigs, donkeys, llamas, sheep, goats, peafowl, chickens, cats, and dogs were displayed. A store and concession stands sold food and drink that could be consumed in the animal areas. Raw milk was not served. Hand-washing facilities were limited, not configured for use by children, and unsupervised.
Thirty-three of 216 cattle (15 percent) were colonized with E. coli O157:H7. Of these, 28 of 33 (85 percent) were colonized with E. coli O157:H7 that had the same PFGE pattern as isolates recovered from farm visitors who became ill (Table 3). Overall, calves and heifers were more often colonized than older cattle. Of calves less than six weeks of age, 3 of 13 (23 percent) were colonized with E. coli O157:H7; 2 of 3 isolates (67 percent) had the same PFGE pattern as the isolates from the outbreak. Of calves 6 to 35 weeks of age, 5 of 38 (13 percent) were colonized with E. coli O157:H7, all 5 of which had the same PFGE pattern as the isolates from the outbreak. Of small heifers, 10 of 25 (40 percent) were colonized with E. coli O157:H7; 9 of 10 isolates (90 percent) had the same PFGE pattern as the isolates from the outbreak. Of large heifers and dry cows, 11 of 40 (28 percent) were colonized with E. coli O157:H7; 8 of 11 (73 percent) had the same PFGE pattern as the isolates from the outbreak. Of nine "pre-fresh" cows (cows that were within two weeks before parturition), one (11 percent) was colonized with E. coli O157:H7; the isolate had the same PFGE pattern as the isolates from the outbreak. Of 91 lactating cows, 3 (3 percent) were colonized with E. coli O157:H7, and each isolate had the same PFGE pattern as the isolates from the outbreak.
|
Laboratory Results
The number of isolates of E. coli O157:H7 and the results of molecular subtyping are summarized in Table 3. All 8 E. coli O157:H7 isolates from patients and 29 (83 percent) of 35 isolates from nonhuman sources had the same rare PFGE pattern (EXHX01.0070, pattern 70) and produced both Shiga toxin 1 and Shiga toxin 2.
Discussion
We describe a large outbreak of E. coli O157:H7 infections that were directly transmitted from animals and their environment to people, largely children. Contact with calves up to 35 weeks old and their environment was associated with illness. Among all farm visitors, the data showed a trend toward hand washing as providing protection.
The proportion of cattle colonized with E. coli O157:H7 at the farm (15 percent) was higher than in other studies of cattle herds (rates are typically in the range of less than 0.5 percent to 2 percent).13,14 High rates of colonization occur among young cattle, and occasionally, periodic, sharp fluctuations in prevalence, to rates above 10 percent, have been identified.15 Determinants of the rates of E. coli O157:H7 colonization in cattle herds are not yet well understood.15,16 Collection and transport of samples and culture methods may affect the rates. Although some investigators have sampled fresh fecal pats,13 our use of rectal swabbing was consistent with the methods of other studies.14 Had we used immunomagnetic separation for the rectal swabs obtained from animals, as we did for the human fecal specimens, the isolation rate would probably have been even higher.17 The high rate of colonization of cattle may indicate that the E. coli O157:H7 strain involved in the outbreak had recently been introduced to the farm, leading to the peak in prevalence that may occur when a new strain sweeps through a previously unexposed herd.15 The high prevalence of E. coli O157:H7 colonization in the herd at the farm most likely contributed to increased contamination of cattle hides and the environment during the outbreak period. This increased contamination in turn created greater risk to farm visitors than might have occurred had the colonization rate of cattle been lower. Despite this, farm workers reported no diarrheal illness during the outbreak period. This finding is consistent with studies showing that previous infection and frequent reexposure to E. coli O157:H7 may confer some protection against symptomatic infection.18,19,20
Evidence from a household survey suggests that higher than expected rates of diarrhea occurred among visitors to the farm, indicating excess illness that was not identified by routine case finding. This highlights the potential scale of zoonotic outbreaks of E. coli O157:H7 infection and suggests that the importance of this mode of transmission may previously have been underestimated. It also underscores the risk associated with bringing children, a group at increased risk for severe illness due to E. coli O157:H7 infection, together with cattle, major reservoirs of E. coli O157:H7, in an uncontrolled environment where eating is encouraged and hand-washing facilities are inadequate. The outbreak at the farm may not be an isolated event. A case–control study conducted by the Foodborne Diseases Active Surveillance Network (FoodNet) of the CDC during 1996 and 1997 found that living on or visiting a farm with cows during the seven days before illness was associated with an increased risk of sporadic E. coli O157:H7 infection.21 Studies in other countries have found similar associations.22,23 Two random-digit-dialing telephone surveys of approximately 9000 persons each, conducted by FoodNet during 1996–1997 and 1998–1999, showed that 2 percent of respondents reported visiting a petting zoo in the five to seven days before the interview,24,25 indicating that the population at risk is large. Furthermore, settings such as agricultural fairs attract large numbers of visitors and create conditions similar to those found at the farm.8
E. coli O157:H7 can survive in the environment for months26,27,28 and thus pose an ongoing source of infection of humans, even in the absence of direct contact with animals. The frequency of E. coli O157:H7 in U.S. cattle herds combined with the environmental persistence of the organism supports the recommendations that all cattle should be handled as if they are colonized and that all cattle environments should be approached as if they were contaminated with E. coli O157:H7.
Farm environments can be made safer for visitors. Prevention strategies were developed to help reduce the risk of transmission of enteric pathogens at petting zoos, farms open to the public, animal exhibits, and other venues where the public has contact with farm animals.9 The strategies include the use of hand washing, controlled and supervised contact with animals, and clear separation of food-related activities from areas housing animals.
Evidence is growing that contact with farm animals and their environment is a substantial contributor to the risk of E. coli O157:H7 infection. This outbreak underscores the need to consider zoonotic transmission during searches for the source of E. coli O157:H7 and other enteric infection and that simple measures such as effective hand washing can make contact with farm animals and their environments safer.
Source Information
From the Epidemic Intelligence Service (J.A.C.) and the Public Health Prevention Service (C.S.), Division of Applied Public Health Training, Epidemiology Program Office, and the Foodborne and Diarrheal Diseases Branch (J.A.C., A.C.S., A.J.L., S.B.H., S.K.W., P.M.G., T.J.V.) and the Biostatistics and Information Management Branch (R.M.H.), Division of Bacterial and Mycotic Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta; the Montgomery County Health Department, Norristown, Pa. (A.S.C., R.G., M.B.); the Pennsylvania Department of Health, Harrisburg (M.M.); the Pennsylvania Department of Health Bureau of Laboratories, Lionville (G.W.); and the Virginia Division of Consolidated Laboratory Services, Richmond (D.M.T.).
Address reprint requests to Dr. Crump at the Foodborne and Diarrheal Diseases Branch, Division of Bacterial and Mycotic Diseases, National Center for Infectious Diseases, MS A-38, Centers for Disease Control and Prevention, 1600 Clifton Rd., Altanta, GA 30333, or at jcrump{at}cdc.gov.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
