The Town That Was Poisoned/Centers for Disease Control
February 28, 1985 CONGRESSIONAL RECORD - HOUSE
Mr. Speaker, I have a letter which is the preliminary report of the health authorities investigating the Salmonella outbreak in The Dalles, to Dr. John A. Googins, M.D., State epidemiologist, Department of Human Resources, Oregon State Health Division, Portland, OR.
It begins: “Dear Dr. Googins, thank you for inviting the Centers for Disease Control to participate in investigating the September 1984, outbreak of Salmonella typhimurium in The Dalles.”
Mr. Speaker, I would like the balance of this letter included in the RECORD at this point, and I ask unanimous consent to do that.
The SPEAKER pro tempore. Without objection.
CENTERS FOR DISEASE CONTROL,
Atlanta, GA, January 18, 1985.
JOHN A. GOOGINS, M.D.,
State Epidemiologist,
Department of Human Resources,
Oregon State Health Division Portland, OR.
DEAR DR. GOOGINS: Thank you for inviting the Centers for Disease Control (CDC) to participate in investigating the September 1984 outbreak of Salmonella typhimurium in The Dalles. It was a pleasure to work with the highly motivated members of your division and of the Wasco-Sherman Public Health Department. This letter is to summarize the preliminary findings of our investigation. A complete report will follow.
INTRODUCTION
On September 17, 1984, the Wasco-Sherman Public Health Department received a call from a person who complained of gastroenteritis after eating at a popular restaurant in The Dalles. In the Next 4 days they received at least 20 additional complaints involving two restaurants. On September 21, Salmonella typhimurium was isolated at the Oregon Public Health Laboratory from a stool sample of one of these patrons. There were anecdotal reports of widespread illness among employees in both establishments, as well. CDC was contacted on September 25 and was asked to participate in the investigation of this outbreak. Following subsequent publicity, hundreds of affected patrons telephoned to report illness. Illness was not limited to patrons of the two initially reported restaurants, but involved employees and patrons from 10 independent establishments. A total of 715 persons met a clinical case definition (below) or had a positive stool culture for Salmonella typhimurium. The dates of onset of clinical illness were between August 28, 1984, and October 26, 1984. Almost all gave a history of having eaten at a restaurant in The Dalles within 72 hours prior to illness. Of those cultured, 342 were found to have Salmonella; all but 3 of 275 were identified as Salmonella typhimurium on further testing. The other 67 isolates were not serotyped. The remaining 373 persons met the clinical case definition of diarrhea and at least three of the following symptoms: fever, chills, headache, nausea, vomiting, hematochezia, and abdominal cramps. An additional 117 people reported an acute gastrointestinal illness but did not meet the clinical case definition. At least 45 persons were hospitalized. No fatalities were reported.
The epidemic curve was biphasic. The first wave, accounting for approximately 15 percent of the total cases, began on September 10 and peaked on September 15 with 17 culture-confirmed cases. The second wave began abruptly on September 22 and peaked on September 24 with 71 culture-confirmed cases. This curve rapidly falls off with fewer than 2 percent of culture-confirmed cases reported after September 28. Eight of the 10 implicated restaurants were associated with culture-confirmed cases in the second wave only; the other two restaurants were associated with confirmed cases in both waves.
An initial investigation by local health officials suggested that having eaten at a salad bar was associated with illness. This observation led to the recommendation on September 25 that the restaurants in The Dalles voluntarily discontinue salad bar service. This action was associated with the subsequent abrupt decline in new cases.
METHODS AND RESULTS
- 1. Investigation of Patron Illness
Information on illness in restaurant patrons was obtained primarily through a passive surveillance system. Initial reports came from unsolicited phone calls form ill patrons. Press releases encouraged reporting. Additional cases were reported from local and regional laboratories. Ill patrons were interviewed by telephone or in person regarding demographic data, nature of illness, risk factors for gastrointestinal infection, and foods eaten at any restaurants in the 3 days prior to onset of symptoms. We attempted to obtain similar information for comparative basis for estimating relative risks. Standardized questionnaires with comprehensive lists of foods available at each restaurant were administered when possible.
No other common exposure was identified among these ill patrons of restaurants. Local residents who became ill were geographically scattered by residence and appeared to be equally likely to live in either region served by the two water sources for The Dalles. Thus, the subsequent investigations focused on food consumption in restaurants.
There were 10 restaurants where at least one case of patron illness was confirmed by stool culture. Eight of these restaurants had salad bars. Food-specific attack rates were analyzed for four restaurants where we had enough complete food histories from both ill persons and controls. These four restaurants accounted for 274 patron cases (38 percent of the total number of cases). Attack rates were highest for salad items in all four restaurants; lettuce was the only item whose consumption was consistently associated with illness. In one restaurant, potato salad, bean salad, and macaroni salad were associated with illness even in patrons who had not reported having consumed lettuce.
- 2. Investigation of Restaurant Employees
Employees of implicated restaurants were interviewed during the outbreak and then again afterward. The initial interviews resembled the patron interviews; the second set of interviews was part of a systematic survey that included controls from six unaffected restaurants. More specific information collected included work duties, restaurant patronage, and recent diarrheal illness.
All employees from implicated restaurants were required to submit a stool sample for culture. Those with a positive culture or a history of recent gastrointestinal illness were excluded form work until two consecutive stools were negative.
There were 325 employees working in the 10 implicated restaurants during September. Of these, 207 (94 percent) were interviewed in depth. Seventy-four (23 percent) reported a diarrheal illness. Sixty-six (20 percent) were culture-positive for Salmonella typhimurium; this group includes 26 employees who denied any illness. Of the eight restaurants where culture-positive patrons were documented from the second wave only, seven had at least one employee with early onset of disease (on or before September 20). In one restaurant, the work history of a salad maker with culture-proven, early onset illness coincided with patron exposure. In another establishment, multiple salad makers became ill early on; no single employee’s schedule coincided with the epidemic curve, but the combination of the ill employees’ schedules did.
Detailed information was available on 47 ill employees. Only one of 16 early cases reported eating at one of the two restaurants associated with the first wave. Comparing the 16 early to 31 late cases of diarrhea among employees, there were no differences with respect to age, sex, consumption of raw eggs or milk, number of restaurants patronized, or in number of meals eaten at other restaurants. Thus, although the early onset of illness in employees in most of these restaurants suggests that they introduced Salmonella typhimurium, the sources of their infections remain obscure.
- 3. Restaurant Inspections
All implicated restaurants were inspected by sanitarians. There were no significant time-temperature violations for holding food. Sanitary inspections revealed minor violations of some restaurants, but none of these violations could plausibly have caused this large outbreak in multiple restaurants.
- 4. Food Sources
Local health department sanitarians and a Food and Drug Administration representative sought out the distributors and original suppliers of foods used in all implicated restaurants. There was no single food item or supplier common to all 10 restaurants or even to the first two establishments implicated. Many of these distributors served large geographic regions in the West, yet only one additional outbreak was recognized in this time interval, and did not appear to be related.
To explain the appearance of Salmonella typhimurium in the first restaurant, several leads were pursued:
a. Five employees of this restaurant, three who were ill with culture confirmed Salmonella typhimurium gastroenteritis, reported drinking raw milk or consuming raw milk products from an uncertified dairy in Washington. An on-site investigation of the dairy by Paul Williams,j D.V.M., Oregon State Health Division, revealed unsanitary practices, including washing milk pails in untreated pond water. However, cultures of raw milk, cow feces, and pond water were negative for any Salmonella species.
b. This restaurant used local produce occasionally. One farm providing cucumbers and tomatoes on September 9 and 20 was inspected by an epidemiologist and a sanitarian. They discovered that a nearby trailer court had a history of septic tank malfunctions in early September, but found no gross signs of contamination of the adjacent vegetable patch at the time of their inspection. Samples of soil and tomatoes were submitted to the Oregon State Laboratory; cultures were negative for Salmonella species. A second farm provided cantaloupes to the restaurant on September 20, but complete harvesting precluded meaningful findings from this inspection.
Samples of two lots of lettuce from each of two restaurants involved in the second wave of illness were submitted to the Oregon Department of Agriculture. They showed no evidence of contamination by any Salmonella species. Cultures of several food items from another restaurant were obtained. Salmonella typhimurium was cultured from a sample of blue cheese dressing only, but not from the dry mix used to prepare it. Given that the sample was obtained well into the epidemic, it is likely that the dressing was contaminated during or after its preparation.
- 5. Water Supply Investigation
There was no epidemiologic evidence to indicate water contamination as the source for this outbreak. Records of routine surveillance by the local water company showed no increase in coliform counts during September. Supplemental samples were collected from five of the implicated restaurants, and no Salmonella was cultured from any of these sources.
- 6. Special Studies
a. The restaurant accounting for the largest number of cases also had a busy home delivery service. All items on the menu were routinely delivered except salad. Telephone interviews of over 120 people consuming delivery orders disclosed no illness except where there were other potential exposures at an implicated restaurant. This study provided further evidence supporting salad as the vehicle of the outbreak.
b. Several restaurants had banquet facilities. Although a number of groups had been served during the outbreak, illness was not reported from any of them. These banquets had salad bars prepared by each restaurant’s usual personnel and from the usual ingredients, but the salad bars had fewer items, and perhaps more importantly, were taken down within a short period of time. In contrast, most of the other salad bars remained open as long as the restaurant was open.
c. The first wave of illness at the first affected restaurant was investigated using the original waitress food order slip and records available for patrons who had charged their meals to their rooms. A total of 49 such patrons who had eaten from September 11 to 15 were contacted and their clinical status ascertained. Seven (14 percent) reported a diarrheal illness in the week after their meal. All seven, versus 17 of the 42 who remained well, had eaten food from the salad bar (p=0.004, FET, 2-tailed). For statistical analyses, we pooled food histories from this group with those from other patrons of the restaurant and found that lettuce and salad dressings were not associated with illness. Consumption of sliced fruit was of borderline significance. Consumption of any mixed salads, e.g. macaroni, was highly significant (11/15 ill vs. 11/43 well, p=0.002, FET, 2-tailed).
SUMMARY
1. A large outbreak of diarrhea due to Salmonella typhimurium occurred in The Dalles in late September 1984.
2. During this time, only one outbreak was recognized in the Northwest outside of The Dalles.
3. Amplification and transmission appears to have occurred at 10 restaurants via contamination of one or more salad items.
4. No single source for food items or single distributor common to all or most of the implicated restaurants was identified.
5. At most restaurants implicated, food handlers were identified with gastroenteritis and onset of illness prior to exposure of patrons, suggesting that food handlers may have contaminated the salad bars.
6. The food items implicated were uncooked and likely to be handled during their preparation without rewashing prior to serving.
7. There was no epidemiologic evidence to suggest deliberate contamination of foods.
8. Sanitary practices in implicated restaurants were not found to be grossly deficient.
9. The public health actions of the Wasco-Sherman Health Department appear to have been very effective in terminating the outbreak.
FURTHER INVESTIGATION
1. Plasmid profile studies of Salmonella typhimurium isolates are being performed to test the hypothesis that a single isolate was responsible for disease acquired at multiple restaurants. These studies will also allow comparison with strains collected elsewhere and may shed light on how Salmonella typhimurium was introduced into the community.
2. The Oregon State Health Division will conduct studies of temperature changes in foods held on salad bars.
3. Complete data analysis in cooperation with Centers for Disease Control is in progress.
RECOMMENDATIONS
1. Continued exclusion from work of culture-positive food handlers until they have had at least two consecutive negative stool cultures.
2. Continued surveillance for new cases of salmonellosis.
3. Further emphasis on hygienic practices in public food-handling establishments.
Sincerely yours,
Thomas J. Torok, M.D., EIS Officer,
New Mexico Division of Field Services,
Epidemiology Program Office.
Robert Tauxe, M.D., M.P.H., EIS Officer,
Enteric Diseases Branch, Division of Bacterial Diseases,
Center for Infectious Diseases.
Paul A. Blake, M.S., M.P.H.,
Chief, Enteric Diseases Branch, Division of Bacterial Diseases,
Center for Infectious Diseases.
Robert P. Wise, M.D., M.P.H., EIS Officer,
Division of Surveillance and Epidemiologic Studies,
Epidemiology Program Office.
John M. Horan, M.D., Assistant Director,
Division of Field Services,
Epidemiology Program Office.
This work is in the public domain in the United States because it is a work of the United States federal government (see 17 U.S.C. 105).
Public domainPublic domainfalsefalse