Norovirus Outbreak in an Elementary School—District of Columbia, February 2007 FREE

MMWR. 2008;56:1340-1343

1 figure, 1 table omitted

On February 8, 2007, the District of Columbia Department of Health (DCDOH) was notified of an outbreak of acute gastroenteritis in an elementary school (prekindergarten through sixth grade). The school nurse reported that 27 students and two staff members had become ill during February 4-8 with nausea, vomiting, and diarrhea; because symptoms lasted <48 hours, a viral etiology was suspected. DCDOH recommended two preinvestigation interventions, which were implemented the same evening (February 8): (1) more thorough handwashing and (2) bleach cleaning of all shared environmental surfaces with a diluted (1:50 concentration) household bleach solution. This report summarizes the subsequent investigation of the outbreak, which suggested that noncleaned computer equipment (i.e., keyboards and mice) and person-to-person contact resulted in illness. To decrease disease transmission during gastroenteritis outbreaks, public health officials should emphasize good handwashing practices, exclusion of ill persons, and thorough environmental disinfection, including fomites that are shared but not commonly cleaned.

On February 9, DCDOH conducted a site visit and interviewed school personnel to determine the possible etiology of and risk factors for illness and to recommend additional control measures. The school had two to three classes per grade, and one to three staff members were assigned to each class. Although students attended a few classes outside their classroom each day (e.g., art or math), they spent the majority of time in their own classrooms. No outbreaks of gastrointestinal illness in the community were reported to DCDOH during this period.

A case of gastrointestinal illness was defined as illness in a student or staff member who reported nausea, vomiting, or diarrhea and who was present at the school any time during February 2-18. A questionnaire was developed to use in a cohort study of all staff members and students. Because no food was served at the school other than lunches that students brought from home and prepackaged snacks served in prekindergarten classes, foodborne transmission was not suspected; questions focused on illness onset, symptoms, school grade, classroom, special classes (e.g., art), ill contacts, and use of certain facilities or equipment (e.g., library computers) or participation in certain programs (e.g., after-school programs). Questionnaires were sent home by the school principal with all staff members and students the afternoon of February 9. The school nurse identified additional cases beginning February 9 by visiting each classroom daily; she interviewed persons who became ill during school and interviewed absent ill persons or a family member by telephone regarding grade, classes, illness onset, and symptoms. Information regarding ill contacts, facility and equipment use, and participation in programs was unavailable from the participants enrolled by the nurse.

Of 314 students and 66 staff members at the school, 207 (66%) students and 59 (89%) staff members participated in the DCDOH investigation, for a total of 266 participants (70%). A total of 225 (85%) were identified by using the questionnaire, and 41 (15%) were identified by the school nurse interview. Of 266 participants, 103 (39%) met the case definition. Among the 103 ill persons, 79 (77%) were students and 24 (23%) were staff members. The median age of students was 8 years (range: 3-12 years) and of staff members was 41 years (range: 13-66 years). A total of 42 of 77 (55%) students and 22 of 24 (92%) staff members were female. Illness onset occurred during February 4-17, with peak incidence on February 7. Reported symptoms included vomiting (64%), nausea (56%), and diarrhea (47%). Median illness duration was 36 hours (range: 0.2-96 hours). Median length of stay at home after onset of symptoms was 1 day (range: 0-4 days).

The attack rate (AR) among respondents was 39% overall; ARs did not differ significantly between students and staff members or between females and males. Classroom ARs ranged from 18% (kindergarten classroom G) to 71% (first-grade classroom J). Illness was not significantly associated with grade, location (i.e., floor) of classroom, special classes, or certain facilities or programs. Two factors were significantly associated with illness in bivariate analyses (p<0.05, Fisher's exact test): being in first-grade classroom J (AR = 71%; relative risk [RR] = 1.9; 95% confidence interval [CI] = 1.3-2.8) and contact with an ill person (AR = 38%; RR = 1.8; CI = 1.2-2.7). Using a multivariable model, being in classroom J and having an ill contact also were the only two independent and significant risk factors after backward elimination.

First-grade classroom J was the only classroom in the school in which computers were shared among students and staff members. Students in all other classrooms either had their own computer or shared library computers. Library computer use was not associated with illness, and no students in classroom J reported using library computers.

Stool-specimen collection kits were provided during the DCDOH site visit on February 9, and specimens were received from two ill persons. Twenty-five swabs were used to sample environmental surfaces. Although February 9 was the day after the initial bleach cleaning, several surfaces had not been cleaned and were visibly soiled. Sampled surfaces included toilets, faucets, water fountains, doorknobs, mice and keyboards from three computers (each in a different room), school utensils, and toys. Samples were tested by reverse transcription–polymerase chain reaction (RT-PCR) for norovirus and DNA sequencing; stool specimens also were cultured for bacteria.

Laboratory results were available February 13. One (4%) of the 25 environmental swabs, from a computer mouse and keyboard in first-grade classroom J, was positive for norovirus subtype GII. Norovirus subtype GII also was identified in both stool specimens. Noroviruses from the two stool specimens and a single environmental sample were identical by DNA sequencing of region B, the gene commonly used for genetic classification. Bacterial cultures of stool specimens and environmental samples were negative.

On February 15, DCDOH recommended the following additional interventions: (1) clean computer equipment (e.g., mice and keyboards) and other shared surfaces that were overlooked during the February 8 cleaning with a 1:50 concentration household bleach solution, and (2) exclude ill persons from school for at least 72 hours after resolution of illness because of continued fecal shedding of infectious virus.¹ The last person reported with a case of illness had symptom onset February 17.

R Diggs, MPH, A Diallo, PhD, H Kan, PhD, C Glymph, MPH, BW Furness, MD, District of Columbia Dept of Health. SJ Chai, MD, EIS Officer, CDC.

Norovirus (family Caliciviridae) causes the majority of acute gastroenteritis outbreaks in the United States.² Person-to-person spread through the fecal-oral route, contaminated food and water, and aerosolized vomitus are known to transmit norovirus; contact with contaminated environmental surfaces also has been implicated in transmission.³ Laboratory studies have demonstrated that fingers contaminated with norovirus can transfer the virus to environmental surfaces, which can subsequently contaminate clean fingers with detectable amounts of norovirus.⁴ Because of shared computer use in health-care facilities, schools, and workplaces, certain researchers have suggested that computer equipment might be a route of bacterial disease transmission.⁵ A surrogate marker for norovirus, feline calicivirus, has been shown to persist on computer mice and keyboards for 8-48 hours.⁶

This outbreak is the first report of norovirus detected on a computer mouse and keyboard, which highlights the possible role of computer equipment in disease transmission and the difficulty in identifying and properly disinfecting all possible environmental sources of norovirus during outbreaks. The contaminated computer was located in first-grade classroom J, the only classroom that was independently associated with illness and the only classroom in which computers were shared by students and staff members. No other high-risk modes of transmission explain the increased attack rate in classroom J; for example, no food was served, water-fountain samples were negative for norovirus, and no episodes of vomiting were reported. These factors, together with previously documented hand-to-fomite and fomite-to-hand norovirus transmission, suggest that computer contamination might have played a role in norovirus transmission in classroom J and possibly elsewhere in the school.

Person-to-person contact also likely played a role in this outbreak. Contact with an ill person was one of two significant risk factors for illness in bivariate and multivariable analyses. School children might be at increased risk for person-to-person norovirus transmission because of close quarters and poor hygiene.⁷ Because an ill person is infectious while symptomatic and possibly for 3-14 days or longer after recovery because of continued fecal shedding, the short exclusion time of ill persons from school (median: 1 day after symptom onset) might have facilitated person-to-person transmission in this outbreak. Student person-to-person contact during the weekend was reported anecdotally, and onset of new cases continued after the weekend.

The findings in this report are subject to at least four limitations. First, because data regarding ill contacts, facilities, and program participation were unavailable for nurse-interviewed participants, bivariate and multivariable analyses of these variables were limited to data from survey respondents. Second, certain uninfected persons might have been misclassified as ill because of the broad case definition and subjective reporting of symptoms, which might have resulted in an overestimate of ARs. Third, data were not collected regarding which students in classroom J used computers; consequently, the risk associated with computer contact could not be directly assessed. Finally, because several fomites were cleaned before sampling and not all fomites were sampled, the extent of environmental contamination and the possible transmission role of fomites unrelated to computers could not be characterized.

Proper washing with soap and water can eliminate norovirus from hands⁴; alcohol-based sanitizers also reduce feline calicivirus on hands.⁸ Potentially (but nonvisibly) soiled surfaces are best disinfected with a solution of 1:50 to 1:10 concentration of household bleach in water (1,000-5,000 ppm chlorine) by vigorous wiping for ≥10 seconds.^4,9 However, because a 1:10 household bleach solution is caustic, only corrosion-resistant surfaces should be cleaned with this concentration. Laptop computer keyboards have been shown to withstand >300 disinfections with 80 ppm bleach solution without visible deterioration.⁵ When cleaning environmental surfaces that are visibly soiled with feces or vomitus, masks and gloves should be worn, a disposable towel soaked in dilute detergent should be used to wipe the surface for ≥10 seconds, and a 1:10 household bleach solution should then be applied for ≥1 minute.^4,9 Disposable towels used to clean visibly soiled surfaces should be discarded appropriately after use because they can transfer norovirus to fingers and other surfaces.⁴ Although quaternary ammonium compound-based cleaners typically are not recommended for eliminating norovirus, certain newer formulations^* are effective; alcohol-only cleaners are less effective.¹⁰

This report is based, in part, on contributions by J Davies-Cole, PhD, District of Columbia Dept of Health; S Lyss, MD, and J Blair, PhD, Office of Workforce and Career Development, CDC.

*A list of cleaning products effective against norovirus approved by the Environmental Protection Agency is available at http://www.epa.gov/oppad001/list_g_norovirus.pdf.