An outbreak of Escherichia coli 0157:H7: involving long term shedding and person-to-person transmission in a child care center.

Introduction

During the past several years, increasing research and discussion has focused on the rise in cases involving enterohemoragic pathogens, specifically Escherichia coli O157:H7. First described during two outbreaks in 1982, E. coli O157:H7 has emerged as the single greatest cause of hemolytic-uremic syndrome (HUS) and hemorrhagic colitis in the United States (1,2). Traditionally, the O157:H7 sero-type has been linked to foodborne illness outbreaks in which undercooked meats, especially ground beef, have been consumed. The largest outbreak occurred in 1993 involving over 600 cases in three states (3). These cases were linked to undercooked hamburger patties served at a national fast-food chain. Cases have also been associated with other meats, unpasteurized milk, dry salami, and pressed apple cider (4,5,6,7). Other outbreaks have been associated with recreational bathing waters and child care centers (8,9).

The State of Colorado has required reporting of cases involving E. coli O157:H7 or HUS since 1994. Sporadic cases of E. coli O157:H7 have been reported at the crude rate of 2.25 per 100,000 over the last two years. During July 1995, the Tri-County Health Department (TCHD), which serves three large counties adjacent to Denver, received a telephone call from a local child care center stating that several children at the center had experienced diarrhea. The diarrhea had been severe enough that the children had been taken by their parents to a nearby hospital for evaluation. The hospital initially reported that the tests for enteric pathogens were negative and that E. coli specifically had been ruled out. This incident marked the start of Colorado's first E. coli O157:H7 outbreak.

In this article, a detailed summary of the outbreak investigation is presented, including intervention methods and an analysis of the most salient features of the outbreak - the extensive person-to-person transmission and noted long term shedding of the pathogen.

Methods

Activities at the Child Care Center

The child care center had an enrollment of 141 children and 23 care providers. The center was divided into classrooms assigned by age and [TABULAR DATA FOR TABLE 1 OMITTED] stage of toilet training: infant room (1-11 months), young toddlers room (12-23 months), older toddlers room (24-35 months, not yet toilet trained), 2.5 to 3 years room (1835 months and toilet trained), four-year-olds room (36-47 months) and school-age room. The last room was further divided into two groups: school age one (5-6 year-olds) and school age two (7-12 year-olds). Children arriving early and staying late at the center were grouped into a single classroom until a sufficient number of children had arrived and they could go to their individual classrooms. Meals were prepared at the center by the cook who was assisted by classroom caregivers. The food was then served family-style in the classroom. All children, except the infants, ate the same food.

When TCHD was first notified that the center had several children with a diarrheal disease, and that three of these children had been treated at a nearby hospital for their symptoms, a team of environmental health specialists (EHS) went to the center to investigate and evaluate the site. The specialists conducted educational sessions which focused on reinforcing and stressing enteric precautions the entire staff could utilize to control the further spread of fecal/oral diseases. These precautions included an emphasis on handwashing by providers and by children, cleaning and sanitizing toys and other fomites in the classrooms, and increased diligence in food preparation. An evaluation of food preparation was also conducted at this time.

To limit transmission, symptomatic children were excluded by the center and could return once they were asymptomatic. When the organism was confirmed as E. coli O157:H7, TCHD issued an order prohibiting the center from enrolling new children and requiring the center to arrange a classroom for cohorting all symptomatic children. Once in the cohort, children could not return to the general center population until they had submitted two consecutive negative stool samples 48 hours apart. Parents were also given the option to care for their children at home if they did not accept the cohorting requirement. These parents were given an order from the health officer that excluded their children from any other centers until two negative stool samples had been submitted.

One month after the last new case, a partial rescission order was issued by the health officer, which allowed the center to enroll new children. The final rescission order was issued almost two months later, after the last of the negative sample results were received for the remaining children in the cohort.

Stool Sampling

When TCHD was informed of the child with hemolytic-uremic syndrome, stool sampling began for all symptomatic children. Bacterial stool collection kits (sterile swab and Cary-Blair transport media) were left at the center for parents and providers. During daily visits to the center, TCHD staff picked up any stool samples collected that morning and took them to the laboratory of the Colorado Department of Public Health and Environment (CDPHE) for analysis. Analysis of the samples followed standard methods which included initial plating on sorbitol-MacConkey agar. Determination was made through biochemical tests and reaction on motility media, with antiserum specific for the H7 antigen, and tube agglutination with antisera for the O157 antigen.

Communication

Once cases had been confirmed for E. coli O157:H7, letters were sent to the parents of all children enrolled at the center explaining TCHD's actions and announcing an informational meeting to address parent concerns. Letters also were sent to other area child care centers, as well as area doctors, alerting them to the presence of the disease in the community and its potential for spread. The parent meeting was arranged by the center and attended by staff from TCHD and CDPHE who presented all relevant information and answered questions.

A team composed of managers from TCHD and CDPHE separate from the day-to-day investigation was formed to address policy issues and provide cohesiveness to all aspects of the outbreak. A TCHD public information officer prepared press releases as new information became available, and fielded all media inquiries.

Results of the Epidemiological Investigation

Over the course of the outbreak, a total of 24 cases were identified out of the 141 children attending the center. A case was defined as any individual at the center who tested positive for E. coli O157:H7, or had diarrhea for two days or more during the described time period. This analysis also includes one sibling who tested positive but did not attend the center, bringing the total to 25. Thirteen of the 25 cases had laboratory confirmed stool cultures for E. coli O157:H7. Five cases required hospitalization, one with HUS.

Symptoms included diarrhea in 24 cases (96 percent), with 18 (72 percent) reporting bloody diarrhea. Other common symptoms were abdominal cramps in 19 cases (76 percent), fever in 13 (52 percent), and 12 cases (48 percent) who reported vomiting.

The room location with the onset date of symptoms for each case is shown in Figure 1. Fifty-six percent of all cases had onsets from June 28 to June 30, suggesting a common source for the outbreak. However, during the investigation, no food item or event was found to be associated with the observed illnesses. No cases were reported in either the infant room (1-11 months old) or within the older school-age students' room (6-12 years old).

The overall incidence rate in the center was 17 percent. Incidence was highest in the young toddlers room (44 percent), followed by the 2.5-3 year-old room (21 percent), the school-age room (20 percent), the four year-old room (nine percent) and the older toddler room (nine percent) (Table 1). Of the 25 cases, there were eight sibling pairs and it could not be determined whether transmission occurred at the center or at home. Person-to-person spread was primarily between siblings of the original cases. There were nine symptomatic children after the initial cluster with six (66.7 percent) of those possibly contracting the infection from siblings. Seven of the eight sets of siblings had apparent causal relationships; that is, the second sibling's onset was within one incubation period after the first sibling's. There was one case of transmission that occurred at home to a sibling who did not attend the center. All cases in the four year-old, young toddler and older toddler rooms (except for one individual) had ill siblings in either the 2.5-3 year-old room or the school-age room. No parents or staff members reported being ill or symptomatic during the outbreak.

One of the more remarkable aspects of this outbreak was the long duration of fecal shedding observed with this organism. Of the 13 cases who tested positive for E. coli O157:H7, the median duration of shedding was 29 days with a range of 11-57 days [ILLUSTRATION FOR FIGURE 2 OMITTED]. The shedding period was measured from the onset of diarrhea, to the first of two consecutive negative stool cultures. Of these cases, 85 percent were found to shed the organism for longer than three weeks. Five of the 13 cases exhibited intermittent shedding as defined by alternating positive and negative cultures. There appeared to be no relationship between the age of the case and the duration of shedding (r = 0.10, not statistically significant). The median duration of shedding may be inflated in this study due to the fact that short term or intermittent shedders could have tested negative on their first cultures, resulting in their not being included in the analysis.

Results of the Site Evaluation

The initial site inspection revealed a number of operational problems. Improper disinfection of the diaper changing areas was noted in three of the classrooms. In addition, handwashing was not being practiced by the children before eating, and proper handwashing was not being reinforced by the center staff. Environmental health staff dedicated most of the time they spent in the center reviewing and educating staff on enteric control measures. These control measures were checked daily when TCHD staff went to the center to collect stool samples.

A thorough inspection of the kitchen was performed, including a review of all foods served in the center for the week preceding the onset of the initial cases. Staff involved in food service were also interviewed regarding illnesses and personal hygiene. No ill staff or suspect foods were identified during this process. In addition, parents were interviewed about activities, and food histories were recorded.

Discussion

Our results suggest that E. coli O157:H7 in a child care center is a serious health risk. This risk can be attributed to the severe medical complications that can develop from infection, and evidence that the organism is shed for long periods of time, creating an increased period of communicability. In a Minnesota study involving children under five years of age enrolled in child care centers, median duration of shedding was found to be 17 days with a maximum of 62 days (10). Karch et al. have reported both intermittent shedding and long term carriage in 0157:H 7 infections (11). The majority of cases in this study shed the organism for more than three weeks. The increased time of potential infectivity, coupled with lapses in hygienic practices, most likely contributed to the observed person-to-person spread.

Person-to-person transmission of enteric pathogens through direct contact and via fomites has been noted in a number of instances (12). The risk of transmission increases as the factors associated with transmission increase. These include: lack of toilet training, frequent hand-to-mouth contact, and lack of proper hygiene, both at the individual level (handwashing) and at the institutional level (sanitization of diapering areas). These factors favor transmission among young children not yet toilet trained. This is consistent with our findings, as the highest attack rate was seen in the young toddler room.

Eight sibling pairs were identified as part of this outbreak. It was not possible to determine, except in one instance, if sibling-to-sibling spread occurred at home or at the center. Parents' admissions of bringing symptomatic children into the center, both before and during this outbreak, indicate a casual approach to childhood diarrheal illnesses, which may explain how the initial case could have been brought into the center. Although most of the evidence pointed to transmission between siblings, there were no parents known to be symptomatic during the outbreak. Household transmission should be addressed with parents early in an outbreak.

Issues Related to Cohorting

To control this outbreak, cohorting of ill children was used to break the chain of infection. No new cases were found in the center after cohorting was initiated. Because of the severe disease resulting from O157:H7 infections, including HUS, some authors have suggested exclusion of all cases from the center (10). We feel that in situations with one or two cases, exclusion should be implemented. In this child care center with multiple cases, however, cohorting (specifically close control and monitoring of the situation) reduced the risk of transmission to other child care settings. Hoffman and Shillam have reported on cohorting as an effective tool in an outbreak of Shigellosis (13). There are several issues related t,o cohorting within a child care center that merit discussion.

In this case, all the classrooms of the center were occupied, so it became necessary to vacate a room for the cohorting. The children occupying the room chosen for cohorting were moved to other classrooms, and the case children were then moved into the cohort room. This necessitated the mixing of age groups in both the cohort and the general center population, which was a violation of child care license provisions. It was therefore imperative that the licensing authority be contacted to advise them of these unusual conditions. Although the licensing authority for child care centers in this state lies with the Colorado Department of Human Services, the state and local health departments are responsible for investigating public health related issues at child care centers.

The previous national publicity of E. coli had heightened the parents' awareness of this disease, and they were apprehensive about the serious complications that could arise. The parents with seemingly healthy children were concerned with the spread within the center, while the parents of previously symptomatic children were concerned with the possibility of reinfection. Although the possibility of reinfection was probably low, little is known about short-term immunity. The fact that once the cohort was formed there were no cases of reinfection may suggest that reinfection is uncommon.

Some of the parents of symptomatic children decided to take their children out of the center. In an effort to prevent spread to other centers, these parents were issued exclusion orders, which in effect prohibited them from enrolling their children in other centers. In most instances, parents had to take time off from work to stay with their children. As a result, these parents were often understandably difficult to work with. Many parents did not like being told what to do by the agency, and constantly pressured TCHD staff to speed up the sampling process.

Conclusion

In light of both the long-term shedding of this organism, and the ease of person-to-person transmission in child care settings, we found that cohorting children at the center, proper handwashing, and good hygienic practices were effective tools for controlling the spread of E. coli 0157:H7. For those who are responsible for controlling such outbreaks, it is recommended that a protocol of control measures be developed which explains, before an outbreak occurs, how and when controls such as cohorting will be implemented. Early interventions in the form of exclusion and/or cohorting have a great impact on limiting disease spread. By planning ahead and instituting an established control protocol, much consternation and cost may be avoided.

Corresponding author: Lloyd Williams, Tri-County Health Department, 4857 S. Broadway, Englewood, CO 80110.

Acknowledgements

The authors wish to thank the following: Dr. Richard Hoffman, Dr. Seema Shah and Ms. Pain Shillam from CDPHE for their considerable consultation during the investigation of this outbreak, the microbiological section of the CDPHE laboratory for processing of samples, and Ms. Stacy Weinberg and Dr. Patrick McCool for assistance in reviewing this manuscript.

Facts about E. coli 0157:H7

What is E. coil O157:H7?

E. coli O157:H7 is a bacteria that can be found in raw meat (poultry, lamb, pork and beef) and unpasteurized milk. It is also found in the feces of both humans and animals. Those at risk for serious illness are children under 10 years of age and older people. Complications that can occur include: hemolytic-uremic syndrome (HUS); post diarrheal thrombotic thrombocytopenic purpura (TTP); and serious kidney complications.

What are the symptoms?

* Sudden onset of watery diarrhea (often progressing to bloody diarrhea)

* Severe stomach pains

* Fever (for 1/3 of the cases)

* Vomiting for some

How soon do symptoms appear?

Symptoms occur within two to eight days; median 72 hours. The illness lasts four to ten days.

How is E. coli O157:H7 spread?

* By eating meat (particularly ground beef) or poultry that was not cooked completely.

* When the bacteria gets on hands (from diaper changing or assisting in toilet training) and contaminated hands* are then put in the mouth.

* From an infected person's hands becoming contaminated* and then touching food others will eat.

* When contaminated hands(*) touch objects (toys) that others will put in their mouth.

* Hands become contaminated when feces/stool from an infected person gets on the hands and thorough handwashing does not follow.

How long can an ill person spread E. coli O157:H7?

As long as the organism is present in the stool. In young children, shedding for more than three weeks is common.

How can E. coli O157:H7 be prevented?

* Cook meat and poultry until all pink is gone and juices run clear (ground beef to 155 [degrees] F).

* Drink only pasteurized milk.

* Return any food from a food service establishment that is under-cooked.

* Wash all fruits and vegetables before eating raw - OR - heat to 165 [degrees] F (a slow simmer with steam starting to rise from the pan).

* Wash hands very well with soap and warm water after going to the bathroom or diaper changing, prior to eating, and after handling raw meat.

What should you do if you have it?

* DO NOT go to work if handling food or caring for young children; do not send children to child care settings.

* Call your doctor or local health department for more information.

E. coli O157:H7 in a Child Care Facility

Investigation/Control Tips

1. Investigate any cases of E. coli O157:H7 as soon as possible. Expedient investigation is important due to the potential for. progression to HUS and other life threatening complications.

2. Contact the owner of the child care facility, as well as the operator and their licensing authority, as soon as possible.

3. Make a site visit to the facility to ascertain its condition and reinforce good hygienic practices by staff and children.

4. Conduct a meeting with parents early in the investigation and update them frequently. Stress good hygiene in the home due to the potential for household spread.

5. Cohort the symptomatic children in the child care center away from the children who are well.

6. Culture all children and child care staff for enteric pathogens including E. coli O157:H7. This may extend to certain close contacts of ill children as well.

7. Consider using Glo-Germ[TM] to provide a visible demonstration for children and staff to educate them on disease transmission.

8. Have culture kits readily available and develop a plan that includes the arrangements for transportation of samples and reporting of results with a local laboratory that has the capability to detect E. coli O157:H7.

REFERENCES

1. Riley, L.W., P.A. Blake, M.L. Cohen, N.T. Hargrett, R.J. Hebert, L.M. Johnson, et al. (1983), "Hemorrhagic Colitis Associated With A Rare Escherichia coli Serotype," N. Engl. J. Med., 308(12):681-685.

2. Karmali, M.A., G.S. Aubus, L. Corazon, D.C. Fleming, H. Lior, and M. Petric (1985), "The Association Between Idiopathic Hemolytic-Uremic Syndrome and Infection by Verotoxin-Producing Escherichia coli.," J. Infect. Dis., 151(5):775-782.

3. Bell, B.P., R. Baron, T.J. Barrett, C.A. Bartleson, M.A. Davis, P.M. Goldoft, et al. (1994), "A Multistate Outbreak of Escherichia coli O157:H7-Associated Bloody Diarrhea and Hemolytic-Uremic Syndrome from Hamburgers," The Washington Experience," JAMA, 272(17): 1349-1353.

4. Doyle, M.P., and J.L. Schoeni (1987), "Isolation of Escherichia coli O157:H7 from Retail Fresh Meat and Poultry," Appl. Envtl. Microbiol., 53(10):2394-2396.

5. Martin, M.L., J. Arnoldi, J.P. Davis, K. Hedberg, ME. Potter, and L.D. Shipman (1986), "Isolation of Escherichia coli O157:H7 from Dairy Cattle Associated with Two Cases of Haemolytic-Uraemic Syndrome," Lancet, 2(8514):1043.

6. Centers for Disease Control and Prevention (1995), "Escherichia coli O157:H7 Linked to Commercially Distributed Dry-Cured Salami - Washington to California, 1994," Morbidity and Mortality Weekly Report, 44(9):157-160.

7. Besser R.E., T.J. Barrett, M.P. Doyle, P.M. Griffin, S.M. Lett, J.T. Weber, and J.G. Wells (1993), "An Outbreak of Diarrhea and Hemolytic-Uremic Syndrome from Escherichia coli O157:H7 in Fresh-Pressed Apple Cider," JAMA, 269(17):2217-2220.

8. Brewster, D.H., J. Bimson, M.I. Brown, G.L. Houghton, D. Robertson, and J.C.M. Sharp (1994), "An Outbreak of Escherichia coli O157:H7 Associated with a Childrens' Paddling Pool," Epidemiol. Infect., 112(3):441-447.

9. Spika, J.S., P.A. Blake, R.A. Gunn, D. Nordenberg, J.E. Parsons, and J.G. Wells (1986), "Hemolytic-Uremic Syndrome and Diarrhea Associated with Escherichia coli O157:H7 in a Day Care Center,"J. Pediatr., 109(2):287-291.

10. Belongia, E.A., D.A. Ammend, J.E. Braun, K.L. MacDonald, M.T. Osterholm, and J.T. Soler (1993), "Transmission of Escherichia coli O157:H7 Infection in Minnesota Child Day-care Facilities," JAMA, 269(7):883-888.

11. Karch, H., Heesemann, H. Russmann, H. Schmidt, and A. Schwarzkopf (1995), "Long-Term Shedding and Clonal Turnover of Enterohemorrhagic Escherichia coli O157:H7 in Diarrheal Diseases," J. Clin. Microbiol., 33(6):1602-1605.

12. Pickering L.K., A.V. Bartlett, and W.E. Woodward (1986), "Acute Infectious Diarrhea among Children in Day Care: Epidemiology and Control," Rev. Infect. Dis., 8(4):539-547.

13. Hoffman R.E., and P.J. Shillam (1990), "The Use of Hygiene Cohorting, and Antimicrobial Therapy to Control an Outbreak of Shigellosis - Facts on E. coli O157:H7," Am. J. Diseases of Children, 144(2):219-221.