Norwalk-like viral gastroenteritis: a large outbreak on a university campus.

Introduction

Viral gastroenteritis is estimated to be the second most common illness in the United States (1), with 30% to 40% of all cases of infectious diarrhea due to viruses (2). Up to 65% of epidemic outbreaks of acute nonbacterial gastroenteritis may be attributable to the Norwalk family of viruses (3). In one review of 38 confirmed outbreaks of the Norwalk virus, only four implicated food as the vehicle of transmission (4). There are few descriptions in the literature of large outbreaks where foodborne transmission is well documented (5-11), and even fewer where airborne transmission was suggested (12, 13). However, when a virus is identified as the etiologic agent in a foodborne outbreak, 90% of the time it is a member of the Norwalk group (14).

This article describes an investigation of a large outbreak of acute nonbacterial gastroenteritis occurring among a university student population where a suspected Norwalk-like virus was transmitted through a common food distribution system.

Background

In November 1990, a large foodborne outbreak occurred on a university campus in Boulder, Colorado. The university had a total student enrollment of 25,176 with 6,008 students housed in 10 on-campus dormitories. Most food is prepared separately in each dormitory kitchen facility. Some meat and all bakery items, however, are prepared at a central kitchen facility and then distributed to each dormitory kitchen for serving.

On the evening of November 9, an unusually large number of students appeared at the student health center with symptoms of nausea, vomiting, and diarrhea. By morning, 144 students were treated and released, with many more telephoning the center for medical advice. Several students remained hospitalized overnight for rehydration therapy. On the morning of November 10, the medical director of the student health center contacted the Boulder County (Boulder) Health Department concerning the outbreak, and an epidemiological investigation was initiated.

Methods

Case Definition

For purposes of this outbreak investigation, a case was defined as any university student or employee who developed, between November 8 and 12, an acute case of vomiting or diarrhea with at least one of the following symptoms: nausea, abdominal pain, chills, muscle aches, or fever.

Case Finding

Several sources were used to determine the extent of the outbreak. Accessing a preexisting disease-reporting network within the county, all area hospitals, laboratories, and selected physicians' offices were surveyed regarding patients who met the case definition. This survey provided an estimate of the community's burden of illness as well as identified any university students who sought medical care off campus. Phone logs and admission records at the student health center provided a census of the number of students reporting illness during the outbreak period. Additionally, the university housing staff conducted a floor-by-floor survey of each dormitory at the request of the health department.

Laboratory Investigation

Fresh stool samples from 12 cases were screened for Salmonella, Shigella, and Campylobacter at the Colorado Department of Health laboratory. Two of these were tested for the presence of Staphylococcus toxin. Seven of the 12 samples were sent to the Centers for Disease Control and Prevention and examined by direct electron microscopy for the presence of viral particles and by solid phase immune electron microscopy (SPIEM) (15) against sera to Norwalk virus previously obtained in volunteer studies. In addition, three of the specimens were tested by SPIEM with homologous convalescent antisera. Nineteen suspect food samples from several meals and various dormitories were submitted to the Colorado Department of Health laboratory for a Staphylococcus enterotoxin assay.

Sanitary Investigation

The central kitchen and several of the dormitory kitchens were inspected for proper food preparation practices and hygienic conditions shortly after the onset of the outbreak. Previous inspection reports were reviewed, and all central and dormitory kitchen staff were interviewed regarding any history of illness.

Case Control Study

A standardized questionnaire requesting basic demographic information, signs and symptoms of illness, and consumed food history was given to the students who sought medical attention and met the case definition. The questionnaire was also distributed at six randomly selected dormitories with a protocol to randomly select ill and non-ill students from each dormitory, not to include ill students who had previously returned a questionnaire after visiting the student health center. In total, 101 ill and 71 non-ill students completed the questionnaire.

Statistical Analysis

The case control data provided for the calculation of food-specific attack rates, frequency of symptoms, and duration of illness. The overall attack rate was calculated from the floor-by-floor dormitory survey. Fisher's exact test was used to test statistical significance when comparing proportions (16).

Results

The frequency of symptoms as reported by the affected students is shown in Table 1. Nearly three-quarters (73%) of the ill students were found to have recovered from their acute symptoms within 24 hours of initial onset. The epidemic curve [ILLUSTRATION FOR FIGURE 1 OMITTED] shows date and time of onset of symptoms among ill students, dormitory food staff, and central kitchen food staff. Importantly, the onset of illness among the five central kitchen staff members occurred prior to the peak incidence among ill students.

In total, 787 of 6,008 dormitory residents fit the case definition for illness during the three-day outbreak period (attack rate = 13.1%). Each of the 10 dormitories on campus reported ill students with attack rates ranging from 6.6% to 19.9%. Illness was reported among 13 of 60 dormitory food service workers and 5 of 10 central food service workers for attack rates of 21.7% and 50%, respectively. Surveillance for additional cases continued, but no evidence suggested the outbreak had continued beyond the initial three-day period.

Attack rates for the most implicated foods and meals are shown in Table 2. The large variety of foods offered at each meal and the apparent lack of accurate recall by the students complicated the calculation of food-specific attack rates. The menu item "deli and hamburger bar" actually consisted of a variety of food items including various bakery products and leftovers from previous meals.

Table 1. Frequency of Symptoms Reported Among III University
Students, November 1990.

Symptoms           Percent

Nausea               92%
Vomiting             89%
Diarrhea             81%
Abdominal pain       75%
Muscle ache          73%
Fever                55%

All laboratory results for Salmonella and Shigella were negative. One stool sample tested positive for Campylobacter. All submitted stool samples and food samples tested negative for Staphylococcus toxin. Direct electron microscopy revealed 30 nm small round structured virus particles (SRSV) in three of seven stool specimens. None of these three reacted with the homologous convalescent antisera, and none of the seven samples reacted with the Norwalk antisera.

The sanitary inspection and review of past records indicated that the central and dormitory kitchens had been employing proper and hygienic food preparation practices. However, investigation found only one of the five ill central kitchen staff members had removed themselves from food preparation duties while symptomatic. This occurred in spite of a personnel policy for food service workers that discourages ill employees from attending work and provides sick leave compensation.

Discussion

The epidemic curve clearly indicates a large scale, common source outbreak occurring over a three-day period. Surveys of each campus dormitory and interviews with food service personnel found the outbreak had affected 787 dormitory students and 18 food service workers.

While the attack rates were similar among the 10 campus dormitories, they were markedly higher among the central food service staff. In addition, the central food service staff was found to have symptom onsets approximately 24 to 48 hours prior to the peak onset for students. Only one of the five ill staff members took sick leave during their illness. The suspected foods were among those foods that are received or prepared in the central kitchen before distribution to the individual dormitories. We believe some food, particularly bakery items, may have been contaminated in the central kitchen prior to distribution and eventual consumption.

Table 2. Food-specific Attack Rates Among Dormitory Students,
November 1990.

                                Attack Rate
Date of Meal and Food Item   % ill   % not ill   P-value

November 8, Dinner            74%       54%        .04
Baked Ham

November 9, Lunch             79%       46%        .002
Deli Bar

The community water supplier provides domestic water to the university as well as other residents within the city. The water purveyor routinely tests over 100 water samples each month, and no contamination was found before or during the outbreak. Had the outbreak been associated with the community water supply, it would have been expected to equally affect the non-university population, but repeated surveys using the preexisting disease-reporting network did not reveal a community-wide outbreak.

The negative bacteriological analyses, explosive spread, type and frequency of symptoms, incubation period, and duration of illness are all consistent with reports of Norwalk and Norwalk-like virus-mediated gastroenteritis in the literature (5-10, 17). Kaplan et al. found a high correlation between outbreaks fitting the Norwalk-like pattern of illness and serologic confirmation of a Norwalk-like viral infection (3). Kaplan's proposed criteria for classifying an outbreak as due to a Norwalk-like virus are negative bacterial pathogen cultures, an incubation period between 24 and 48 hours, duration of illness 12 to 60 hours, and vomiting in [greater than or equal to]50% of the cases (3). Hedberg and Osterholm (18) recently expanded this definition to include an increased frequency of vomiting relative to fever rather than a simple vomiting frequency [greater than or equal to]50%. All of the data herein are consistent with both definitions and suggests that the spread of a Norwalk-like virus through contaminated food caused this outbreak.

The case-control study did not implicate a single food served at one particular meal. This is consistent with our belief that ill food workers from the central kitchen facility contaminated several food items, including bakery products, which were served during several meals. Various contaminated food items served at consecutive meals would obscure the findings of a case-control study. The serving of leftover foods in a buffet-style "deli-bar" would further dilute the analysis.

While attempts at visualizing the viral particles by direct EM and SPIEM were inconclusive, the findings herein are consistent with the proposed etiology (9, 19). In 24 serologically-confirmed Norwalk virus outbreaks, SRSV particles were only seen in five. In only two of these five did the particles aggregate upon exposure to the matching convalescent antisera (4). In samples from this outbreak, relatively few SRSV particles were seen. Immunoglobulin-resembling material, which frequently blocks the reaction to appropriate antisera in the laboratory, precoated many of the observed particles. It is also possible that improper handling may have compromised the integrity of the samples. The samples were collected nearly a week before they were prepared and studied at the CDC laboratories. Prompt refrigeration and analysis within 24 hours has yielded increased numbers in past studies (Humphrey, CDC, personal communication, 1992). Generally, [10.sup.6-7] particles/gram are needed in the sample for adequate visualization (10,17,20), and so optimal timing of collection and preservation are important. Detection of a suspected virus in implicated food is rarely attempted due to the difficulties with available methods and the inability to culture the common foodborne viruses (21).

A set of circumstances and food handling practices were in place that created a high potential for a large foodborne outbreak to occur. First, the Norwalk group of viruses is moderately hardy, able to survive both freezing (22) and mild heating (23), although most of the reported Norwalk virus foodborne outbreaks implicate uncooked food (5,6,7,8,11,19,24). The virus is also highly infectious, and outbreaks of Norwalk-like gastroenteritis are characterized by their explosive nature (25,26). The role of the infected food worker in the propagation of this illness has been well described (5,8,12,16,26). In the largest outbreak found in the research conducted, an estimated 3,000 persons became ill over a four-day period from one common source (6). Volunteer studies show that the infective dose can be quite low. From 10 to 100 ingested particles can produce illness (17,19). Industry accepted standards for hand washing may not remove all infectious particles. Viral particles may be shed up to 48 hours after recovery from symptoms (8,19), and infection may be asymptomatic (28). Given the high infectivity, low dosage, and asymptomatic shedding that is possible, one infected food worker can contaminate large quantities of food in a short period of time without an overt disregard for good hygienic practices.

Second, the suspect food items were received and handled in a central kitchen facility for distribution to outlying dormitory kitchens. This food distribution channel creates a situation where an ill food worker could unwittingly infect many hundreds of students. It is clear why many of the described outbreaks occur in institutions, cruise ships, and other settings where a few people prepare food for a much larger group. Fortunately, this food distribution network also contributed to the outbreak's quick subsidence. The ill workers had quickly recovered by the time local health authorities were involved in the outbreak. Once there was no longer a common source for the virus, the opportunity for continued transmission via handled food was eliminated. Secondary transmission (direct-contact) of enteric viruses from primary cases is usually limited (29) and did not contribute to the continuation of this outbreak.

Third, while there is a personnel policy for sick leave to exclude ill food service staff from reporting to work, four of five ill central kitchen workers continued to work while symptomatic. While there is no direct evidence that it was these specific workers who eventually contaminated the suspected food, this disregard for existing policy definitely contributed to the potential risk of an outbreak. Given the persistence of the virus in stools during the convalescent phase, it has been recommended that workers not return to work until 48 hours after symptoms have resolved (7,10,30,31).

To reduce the potential for future outbreaks, personnel policies regarding time off for ill kitchen staff, or at least reassignment to duties uninvolved in food preparation, must be strictly followed. Also, proper food handling and hand washing techniques should be rigorously enforced. In states such as Colorado that have not fully adopted the FDA's food handling regulations (32,33), employers should consider voluntary compliance with those regulations that prohibit direct, bare-handed contact of ready-to-eat foods by food employees. The authors further recommend that food distribution systems in large institutions be modified to receive and prepare potentially hazardous foods at the local kitchen where people are served.

References

1. Dingle, J., G. Badger, A. Feller, et al. (1953), "A Study of Illness in a Group of Cleveland Families: 1. Plan of study and certain general observations," Am J Hyg, 58:16-30.

2. Blacklow, N.R., and H.B. Greenberg (1991), "Viral Gastroenteritis," N Engl J Med, 325(4):252-264.

3. Kaplan, J.E., R. Feldman, D.S. Campbell, et al. (1982), "The Frequency of a Norwalk-like Pattern of Illness in Outbreaks of Acute Gastroenteritis," Am J Public Health, 72:1329-1332.

4. Kaplan, J.E., G.W. Gary, R.C. Baron, et al. (1982), "Epidemiology of Norwalk Gastroenteritis and the Role of Norwalk Virus in Outbreaks of Acute Nonbacterial Gastroenteritis," Ann Intern Med, 96:756-761.

5. Griffin, M.L., J.J. Surowlec, D.I. McCloskey, et al. (1982), "Foodborne Norwalk Virus," Am J Epidemiol 115(2):178-184.

6. Kuritsky, J.N., M.T. Osterholm, H.B. Greenberg, et al. (1984), "Norwalk Gastroenteritis: A community outbreak associated with bakery product consumption," Ann Int Med, 100:519-521.

7. White, K.E., M.T. Osterholm, J.A. Mariotti, et al. (1986), "A Foodborne Outbreak of Norwalk Virus Gastroenteritis," Am J Epidemiol 124(1):120-26.

8. Fleissner, M.L., J.E. Herrmann, J.W. Booth, et al. (1989), "Role of Norwalk Virus in Two Foodborne Outbreaks of Gastroenteritis: Definitive virus association," Am J Epidemiol 12991):165-172.

9. Riordan, T., J. Craske, J.L. Roberts, et al. (1984), "Foodborne Infection by a Norwalk-like Virus (small round structured virus)," J Clin Pathol, 37:817-820.

10. Reid, J.A., E.O. Caul, D.G. White, et al. (1988), "Role of Infected Food Handler in Hotel Outbreak of Norwalk-like Viral Gastroenteritis: Implications for control," Lancet, 321-3.

11. Nelson, M., T.L. Wright, M.A. Case, D.R. Martin, et al. (1992), "A Protracted Outbreak of Foodborne Viral Gastronenteritis Caused by Norwalk or Norwalk-like Agent," J Environ Health, 54(5):50-5.

12. Sawyer, L., J.J. Murphy, J.E. Kaplan, et al. (1988), "25- to 30-nm Virus Particle Associated with a Hospital Outbreak of Acute Gastroenteritis with Evidence for Airborne Transmission," Am J Epidemiol, 127(6):1261-1271.

13. Gellert, G.A., S.H. Waterman, D. Ewert, et al. (1990), "An Outbreak of Acute Gastroenteritis Caused by a Small Round Structured Virus in a Geriatric Convalescent Facility," Infect Control Hosp Epidemiol, 11(9):459-464.

14. PHLS Working Party on Viral Gastroenteritis (1988), "Foodborne Viral Gastroenteritis (with a brief comment on Hepatitis A)," PHLS Microbiol Digest.

15. Humphrey, C.D., E.H. Cook, Jr., D.W. Bradley (1990), "Identification of Enterically Transmitted Hepatitis Virus Particles by Solid Phase Immune Electron Microscopy," J Virol Meth, 29L177-188.

16. EPI-INFO Version 5 (1990), Centers for Disease Control and Prevention, Atlanta, Ga.

17. Sekla, L., W. Stackiw, S. Dzogen, et al. (1989), "Foodborne Gastroenteritis Due to Norwalk Virus in a Winnapeg Hotel," CMAJ, 140:1461-4.

18. Hedberg, C.W., and M.T. Osterholm (1993), "Outbreaks of Foodborne and Waterborne Viral Gastroenteritis," Clin Microbiol Reviews, 6:199-210.

19. "Community Outbreak of Norwalk Gastroenteritis - Georgia," (1982), MMWR, 31:405-7.

20. Iversen, A.M., M. Gill, C.L.R. Bartlett, et al. (1987), "Two Outbreaks of Foodborne Viral Gastroenteritis Caused by a Small Round Structured Virus: Evidence for prolonged infectivity in a food handler," Lancet, 556-8.

21. Cliver, D.O. (1994), "Viral Foodborne Disease Agents of Concern," J Food Prot, 57(2):176-8.

22. "Outbreak of Viral Gastroenteritis-Pennsylvania and Delaware" (1987), MMWR, 36:709-11.

23. Morse, D.L., J.J. Guzewich, J.P. Hanrahan, et al. (1986), "Widespread Outbreaks of Clam- and Oyster-associated Gastroenteritis: Role of Norwalk virus," N Eng J Med, 314:678-81.

24. Kuritsky, J.N., M.Y. Osterholm, J.A. Korlah, et al. (1985), "Norwalk Virus: Statewide assessment of foodborne transmission," (letter), J Infect Dis, 151:568.

25. Alexander, J.W., R.J. Holmes, J.F.E. Shaw, et al. (1986), "Norwalk Virus Outbreak at a College Campus," South Med J, 79(1):33-40.

26. Lieb, S., R.A. Cunn, R. Medina, et al. (1985), "Norwalk Virus Gastroenteritis: An outbreak associated with a cafeteria at a college," Am J Epidemiol, 1(2):259-68.

27. Thornhill, T.S., A.R. Kalica, R.G. Wyatt, et al. (1975), "Pattern of Shedding of the Norwalk Particle in Stools during Experimentally Induced Gastroenteritis in Volunteers as Determined by Immune Electron Microscopy," J Infect Dis, 132:28-34.

28. Blacklow, N., G. Cukor, M. Bedigan, et al. (1979), "Immune Response and Prevalence of Antibody to Norwalk Enteritis Virus as Determined by Radioimmunoassay," J Clin Microbiol, 10:903-9.

29. Cliver, D.O. (1994), "Epidemiology of Viral Foodborne Disease," J Food Prot, 57(3):263-6.

30. Curry, A., T. Riordan, J. Craske, et al. (1987), "Small Round Structured Viruses and Persistence of Infectivity in Food Handlers," (letter), Lancet, 864-5.

31. Appleton, H. (1990), "Foodborne Viruses," Lancet, 1362-4.

32. U.S. Public Health Service (1993), Food Code 1993, Food and Drug Administration, Washington, D.C.

33. Colorado State Board of Health (1989), Rules and Regulations Governing the Sanitation of Food Service Establishments in the State of Colorado, Colorado Revised Statutes, Section 12-44-204.

Robert A. Brockmann, M.S., Laboratory Services, Boulder County Health Department, 3450 N. Broadway, Boulder, CO 80304.