ARCHIVED - Petting zoo-associated Escherichia coli O157:H7 - Secondary transmission, asymptomatic infection, and prolonged shedding in the classroom

Introduction

On 10 November 2003, two public health units notified the British Columbia Centre for Disease Control (BCCDC) of five laboratory-confirmed and two clinical cases of Escherichia coli O157:H7 associated with school and daycare field trips to a local farm. The main attraction to the farm was its pumpkin patch. A petting zoo also operated on the premises.

Petting zoos have been associated with E. coli O157:H7 outbreaks^(1-3). Given the epidemiologic link between all cases and the farm, the hypothesis under joint investigation by the BCCDC and local health units was that the infections resulted from contact with the petting zoo animals. The purposes of this investigation were to verify the cause of the outbreak, determine the outbreak's scope, and identify activities or behaviours that influenced the risk of infection.

Methods

Farm inspection:The farm was inspected on 10 November 2003. A fresh stool specimen from a goat and an older stool specimen of unidentified origin were collected from a pen that housed goats and sheep. Mud, pumpkin and water specimens were also collected.

Laboratory investigation: Human stool specimens were submitted to local laboratories. Subsequent isolations of E. coli O157:H7 were serotyped, tested for presence of, and production of, Verotoxin by Polymerase Chain Reaction and/or Verotoxin Assay, and subtyped by Pulsed Field Gel Electrophoresis (PFGE) by the Enterics Section of the BCCDC Laboratory Services. PFGE patterns indistinguishable from each other were considered to be from the same source.

The animal and environmental specimens were directly inoculated into a selective enrichment broth for E. coli O157:H7. The pumpkin was rinsed in 200 cc of enrichment broth in a large sterile stomacher bag to remove debris. After incubation, E. coli O157:H7 was purified from the broth using an immunomagnetic separation technique (Dynabeads anti-E. coli O157, Dynal ASA, Oslo, Norway) before plating onto sorbitol Mackonkey agar. Suspect colonies were isolated, confirmed by traditional biochemical means, and subjected to PFGE analysis.

Case finding: The farm operator provided a list of the schools and daycares that had visited the farm. Letters were sent to the guardians of children in these facilities, asking them to contact the local public health authorities if their children had experienced diarrhea following the field trip and to contact their family physicians if they had any concerns. Public health inspectors used a case report form to collect demographic, exposure and symptom information from all people reporting to public health authorities.

Epidemiologic investigation: Cases were categorized as either laboratory-confirmed or clinical, and either primary or secondary. Laboratory-confirmed cases were individuals who had epidemiologic links to the farm and E. coli O157:H7 isolated from their stool. Clinical cases were individuals with epidemiologic links to the farm who experienced diarrhea or bloody diarrhea after 1 October 2003. Due to the concurrent circulation of a virus causing a combination of respiratory and gastrointestinal symptoms in the general population, individuals reporting vomiting and/or respiratory symptoms and no bloody diarrhea were excluded from the clinical case definition.

Primary cases were defined as laboratory-confirmed or clinical cases who developed symptoms within 8 days of a visit to the farm. Secondary cases were defined as laboratory-confirmed or clinical cases who had either not visited the farm or had an onset more than 8 days after the farm visit, and who either reported close contact with or were in the same class as a primary case.

A nested case-control study was conducted in order to identify behaviours or activities related to E. coli O157:H7 infection. Because many of the children visiting the farm were young, teachers and daycare operators were asked to provide information about field trip activities. The 5-minute telephone survey included questions about the class, field trip activities and hand washing behaviours.

Data from the case report forms and teacher surveys were entered into EpiData 2.1b databases⁽⁴⁾. Analyses were conducted using EpiInfo 6.04d⁽⁵⁾. Descriptive frequencies were run to determine demographics of the cases and the symptoms that they experienced.

Each case with a relationship to a class that had visited the farm was linked to that class. The numbers of cases associated with each class were calculated, and classes were categorized as either having experienced primary illness or not. Activities and exposures of the "ill classes" were compared to those of the "not ill classes" using the chi-squared test, odds ratios, Fisher's exact test, the Student's t-test and the Kruskal-Wallis H test. Stratified analyses were conducted to examine the presence of confounding or interaction.

Results

The farm was open to the public from 1 October to 31 October 2003. Activities offered to classes included a presentation about pumpkins, a hay ride to and from a pumpkin patch, pumpkin picking, visiting a display of pumpkins, playing on an old wooden boat and visiting a petting zoo enclosure containing goats, sheep, ducks and rabbits. Temporary outhouses were erected for the month of October. Alcohol hand sanitizer dispensers were available in the outhouses; however, the dispensers were placed high on the inside wall, out of reach of some of the children. No running water or other hand washing/sanitizing resources were available and there were no signs recommending hand washing.

The fresh stool specimen from the goat was positive for E. coli O157:H7, with the same PFGE pattern as the laboratory-confirmed human cases. No other etiologic agents were identified in the animal stool specimens; routine ova and parasites and virology were negative.

Seventy-two students, chaperones and close contacts reported symptoms to public health authorities. Of these, 44 met the case definition with eight laboratory-confirmed and 36 clinical cases. One additional child who was not symptomatic but whose parent and siblings experienced symptoms submitted a stool sample. This sample was positive for E. coli O157:H7 and the child was considered a laboratory-confirmed asymptomatic case.

Of the 45 cases, 26 (58%) were primary cases and 19 (42%) were secondary. Onset dates ranged from 15 October to 8 December, with a peak from 27 to 31 October (Figure 1). All cases had epidemiologic links to the farm: 38 were associated with school field trips, six visited the farm with their families and one visited the farm with a social club.

Figure 1. Epidemic Curve: Primary and Secondary Cases of E. coli O157:H7 Associated with Farm Exposure, October-December 2003

Ages of cases ranged from 1 to 39 years, with a median of 5 years. Twenty-four (53%) were female. Forty-five cases (98%) reported experiencing diarrhea; eight of these (six laboratory-confirmed and two clinical cases) reported bloody diarrhea. Other symptoms reported were cramps (60%), fever (29%), vomiting (2%), and nausea (2%).

Fifty-three daycare classes and 119 grade school classes (3,851 students and 1,747 adults) had visited the farm on field trips between 7 and 31 October. Teachers and daycare operators of every class completed the telephone survey for a 100% response rate. Participants from 15 (8.7%) classes reported symptoms to public health authorities and met the primary case definition. The number of cases associated with each of these classes ranged from one to seven (median=2).

All of the classes reported going on the hayride and picking pumpkins. There was no significant difference in the likelihood of ill and not ill classes attending the pumpkin presentation, visiting the petting zoo, playing on the boat, eating while at the farm, or using the outhouses (Table 1). Classes with illness were not significantly more or less likely to participate in hand washing/sanitizing (either at the farm, after arriving back at school, or ever) than classes without illness.

Six (40%) of the ill classes and 77 (49%) of the classes that did not experience illness ate at the pumpkin patch and visited the petting zoo. The sample size was too small to determine whether illness was significantly associated with the relative order of eating and visiting the petting zoo, both controlling for and not controlling for hand washing behaviour (Table 1).

Table 1. Activities of Ill and Not Ill Classes

On average, ill classes had significantly more students, more adult chaperones, and more overall participants than the not ill classes. Mean student-to-adult ratios in the classes did not differ significantly. The mean number of field trip attendees was not related to the relative order of eating and visiting the petting zoo or hand washing/sanitizing behaviours.

The grade distributions of the ill and not ill classes did not differ significantly; however, there appeared to be a statistically insignificant trend for younger classes to be more likely to have ill students (Table 2). Age group was significantly related to the relative order of eating and visiting the petting zoo (chi-square = 7.85, p = 0.020). Of the classes that both visited the petting zoo and ate while at the farm, younger classes were more likely to have eaten during or after their visit to the petting zoo enclosure - 14 (74%) daycare classes, 17 (47%) kindergarten classes, and 9 (32%) primary grade. However, when age group was controlled for, there was no statistically significant difference in the relative order of eating and visiting the petting zoo between the ill and not ill classes. The age group of the students was not related to hand washing/sanitizing behaviours or the number attending the field trip.

Table 2. Illness in Classes by Grade and Number of Participants

Discussion

The isolation of E. coli O157:H7 from the goat with a PFGE pattern indistinguishable from that isolated from human stool indicates that the human cases most likely acquired their infections from contact with the animals in the petting zoo.

An inverse trend of illness with age was observed, although it was not statistically significant. Younger classes were more likely to have eaten their snacks during or after their visit to the petting zoo enclosure. This is of concern because children < 5 years of age are most at risk for developing hemolytic uremic syndrome (HUS), a complication of E. coli O157:H7 infection⁽⁶⁾.

Although the sample size was too small to perform a valid chi-squared analysis, half of the classes who combined their snack time with visiting the petting zoo had at least one ill student. Interestingly, a smaller proportion of the ill classes had eaten after visiting the petting zoo than classes that did not experience illness. This result may have been an artifact of the small sample size - only six of the ill classes had visited the petting zoo and eaten while at the farm. Another explanation relates to the finding that younger children were more likely to have eaten after visiting the petting zoo in both ill and not ill classes. Young children may be more likely to put their fingers in their mouths whether they were eating or not, equalizing the risk of illness between those who ate after visiting the petting zoo and those who did not. This may explain why stratified analyses showed that illness was not associated with eating after petting the animals, even when controlling for hand washing. With very young children, hand washing immediately following contact with the animals should be emphasized rather than hand washing prior to eating.

The size of the group on the field trip was the only exposure or behaviour associated with the likelihood of illness in the class. The total number of people on the field trip, but not the relative amount of adult supervision, was related to illness. Group size was not associated with hand washing or the relative order of eating and visiting the petting zoo. However, larger groups may have been more chaotic, resulting in risky behaviours not being observed or prevented.

The lack of hand washing/sanitizing facilities at the farm may have also contributed to illness among participants. The farm operator will provide hand washing facilities if the farm opens in October 2004.

Forty-five cases were identified through this investigation - 19 were secondary. One laboratory-confirmed case with the outbreak PFGE pattern developed symptoms on 8 December - more than 7 weeks after his daycare group had visited the farm. Other children in his daycare had reported illness, suggesting secondary spread. Another child tested positive for E. coli O157:H7 and was completely asymptomatic. A third child developed symptoms on 19 October and tested positive for E. coli O157:H7 in a stool sample that was collected 6 December, long after symptoms had resolved. This evidence of secondary transmission, asymptomatic infection, and prolonged shedding provides ideal circumstances for ongoing transmission of E. coli O157:H7 among groups of young children, which has been observed in other outbreaks⁽⁷⁾.

Due to the presence of ongoing secondary transmission, public health inspectors visited the daycare centres experiencing high rates of illness, reminded them of the appropriate sanitizing measures for the classroom and enforced exclusion of symptomatic children. On follow-up visits, classroom hygiene was much improved.

One limitation of this study was its small sample size. Conducting the analysis on the classroom rather than on the individual level significantly decreased the study's power. However, it would have been impossible to obtain reliable exposure information on the individual level, because it would involve interviewing very young children.

As there are many illnesses that can cause diarrhea in young children, the clinical case definition used in this study was not specific to E. coli O157:H7. Since 67% of laboratory-confirmed but only 6% of clinical cases experienced bloody diarrhea, it is possible that some of the clinical cases had etiologies other then E. coli O157:H7 . Counting individuals not affected by E. coli O157:H7 as cases may have diluted differences in exposures between true classes and control classes.

This investigation highlights the previously identified risk of E. coli O157:H7 associated with petting zoos, particularly for young children. Provincial guidelines, similar to other jurisdictions'^(8-10), are currently being developed for petting zoos and public health. Petting zoo operators may also consider sending pre-visit information to school groups.

Smaller petting zoos are frequently established in conjunction with other events, often for short times. There is no central listing of petting zoos and therefore no way to contact all petting zoos in order to distribute guidelines. Teachers and daycare operators should also be informed about the appropriate precautions when visiting petting zoos^(8-10).

Conclusion

The E. coli O157:H7 infections in these cases were most likely caused by contact with the animals in the petting zoo. Although specific modifiable risk or protective behaviours could not be identified, petting zoos are a well-established risk factor for E. coli O157:H7 outbreaks(1-3). Young children are the target audience for petting zoos; however, they are more likely to become infected due to poor hygiene and suffer severe complications like HUS. Secondary transmission, asymptomatic infection, and prolonged shedding all contribute to propagation of E. coli O157:H7 infection in the classroom and within families.

Guidelines for petting zoo operators should highlight the need for adequate hand washing/sanitizing facilities. Adults escorting children to petting zoos should also be informed of the importance of hand washing immediately following contact with animals.

Acknowledgements

The authors thank the following for their assistance: the farm operator; the school boards, teachers, and daycare operators involved; the BCCDC Enterics and Food Poisoning laboratories; and the two health authorities involved.

References

1. Gage R, Crielly MS, Baysinger M et al. Outbreaks of Escherichia coli O157:H7 infections among children associated with farm visits - Pennsylvania and Washington, 2000. MMWR 2001;50(15):293-97.

2. Helwig D. E. coli outbreak linked to fall fair. CMAJ 2000;162(2):245.

3. Payne CJ, Petrovic M, Roberts RJ et al. Vero cytotoxin-producing Escherichia coli O157 gastroenteritis in farm visitors, North Wales. Emerging Infect Dis 2003;9(5):526-30.

4. Lauritsen JM, Bruus M, Myatt MA. EpiData (version 2). A comprehensive tool for validated entry and documentation of data. The EpiData Association, Odense Denmark, 2002. URL: <http://www.epidata.dk>.

5. Centers for Disease Control and Prevention, USA & World Health Organization, Geneva, Switzerland. EpiInfo 6 (version 6.04d). January 2001.
   URL: <http://www.cdc.gov/epiinfo/>.

6. Chin J, ed. Control of communicable diseases manual. 17^th ed. Washington, DC: American Public Health Association, 2000.

7. Galanis E, Longmore K, Hasselback P et al. Investigation of an E. coli O157:H7 outbreak in Brooks, Alberta, June-July 2002: The role of occult cases in the spread of infection within a daycare setting. CCDR 2003;29:21-8.

8. Communicable Disease Control Branch & Environmental Control Branch, Department of Human Services, Government of South Australia. Petting zoo infection control guideline: for petting zoo operators, education and childcare services and environmental health officers. February 2002.
   URL: <http://www.dh.sa.gov.au/pehs/branches/environmental-surveillance.htm>.
   Date of access: 2 September 2004.

9. Ohio Department of Agriculture & Ohio Department of Health. Voluntary guidelines for animal exhibitions in Ohio. April 2003.
   URL: <http://www.odh.state.oh.us/ODHPrograms/ZOODIS/ZDis/zoodis1.htm>.
   Date of access: 2 September 2004.

10. The Middlesex-London Health Unit. An E. coli O157:H7 outbreak associated with an animal exhibit: Middlesex-London Health Unit investigation and recommendations. 2002. URL: <http://www.healthunit.com/new_template.asp?id=859>.
    Date of access: 2 September 2004.

Source: ST David, MHSc, Canadian Field Epidemiology Program, Population and Public Health Branch, Health Canada and Epidemiology Services, British Columbia Centre for Disease Control (BCCDC); L MacDougall, MSc, Epidemiology Services, BCCDC; K Louie, CPHI(C), Fraser Health Authority, Fraser South; L McIntyre, BSc, AM Paccagnella, BSc, RT, Laboratory Services, BCCDC; S Schleicher, BSc, CPHI(C), Fraser Health Authority, Fraser East; A Hamade, CPHI(C), Richmond Health Department, British Columbia.