The Bureau of Epidemiology is available 24 hours a day, 7 days a week for consultation at our main number (SunCom 205-4401 or 850/245-4401) PLEASE NOTE: Consultation after 5 p.m. & on weekends is intended for emergencies.

The Florida DOH investigation of the hepatitis A outbreak in Lake and Sumter County, Florida of November-December 2000 has been completed. A summary of this investigation is presented below.

On December 6, 2000, Lake County Health Department (CHD) notified the Florida Department of Health (DOH) of seven hepatitis A cases resulting in five hospitalizations among Lake and Sumter County residents over a 2-week period. In contrast, Lake County reported 8 and 14 cases in 1998 and 1999 respectively over those years. We assisted the Lake CHD in the investigation of this outbreak.

Investigation team members included John Pellosie Jr., DO, MPH, FAOCOPM, Leslie Elliott-Manning, RN, BSN, and Dee Prosser-Shipley, RN, BSN of the Lake County Health Department; Marc Traeger, MD, Zuber Mulla, MSPH, Sandy Roush, MT, MPH, Steven Wiersma MD, MPH, and Richard Hopkins, MD, MSPH of the Bureau of Epidemiology; and Don Windham and Roberta Hammond, PhD, Bureau of Environmental Epidemiology, Food and Waterborne Disease.

A case was defined as a positive test for anti-hepatitis A IgM antibody testies in a Lake or Sumter county resident, person with a clinically compatible illness onset onset between during November 10and -December 16, 2000, residing in Lake or Sumter counties. Cases were identified finding was achieved by reviewing passive reports, contacting local health facilities,and providers, , and reviewing reviewing regional laboratories test resultsy hepatitis A test results.

We identified 23 hepatitis A cases meeting the case definition. Dates of illness onset ranged from November 21 to December 11, 2000. All case patients were white, 52% were female, ages ranged from 15 to 60 years with a median of 38 years, and 15 (65%), were hospitalized. One case-patient was a food-handler at a fast-food outlet and a grocery store delicatessen. No other food handlers working at these locations reported illness consistent with hepatitis A during the outbreak or exposure period.

Two case-control studies were conducted.

The first case-control study was conducted to determine the source of the outbreak. Thirty-eight controls were matched by street of residence to the first 21 cases. The questionnaire administered included questions regarding food eaten at 21 restaurants and any grocery store or deli. These results were analyzed unmatched by univariate analysis and matched by conditional logistic regression. The case-control study revealed a strong association between illness and a fast-food outlet located within a convenience store-gasoline retailer (p<0.001, odds ratio=50.4, 95% confidence interval 9-329 using unmatched univariate analysis). Although not all case-patients initially reported eating food from this outlet, after this study was completed all recalled eating food from the outlet; two additional cases were detected after this study was completed, and both had eaten food from this outlet.

A second case-control study was conducted to investigate a source among food items. Controls were adult patrons of fast-food outlet A in October or November, who did not have an acute gastrointestinal illness in November or December resulting in lost workdays. The second study revealed 6 meal items and 8 ingredients significantly associated with illness. Only two meal items, both of which had the same ingredients (one had larger amounts of some ingredients and was named differently) were eaten by a majority of the case-patients (13 and 12 respectively). These items were nachos-type items with added ingredients. The strongest association between a meal item and illness was taco salad (OR 37.7, CI 4.4-1667, p-value <0.001), although only 10 of 23 case-patients ate the taco salad.

Eight ingredients, including green onions, chips, tomatoes, beans, sour cream, cheese sauce, a salsa in conjunction with a sauce, and guacamole were significantly associated with illness. The ingredient most strongly associated with illness was green onion (OR 29.9, CI 5.7-281, p-value <0.001). Of 23 case-patients, 22 were exposed to sour cream, and 21 were exposed to green onion, chips, tomato, or beans.

serologic testing for recent hepatitis A infectionAll nineteen current and former employees from the convenience store or fast-food outlet A who worked during the exposure period submitted serum for IgM anti-HAV testing. One employee (a case-patient) tested positive and all others tested negative. The employee who tested positive had onset of illness on November 28, 2000, well within the cluster of this outbreak, and thus was not the index case. No index case in this outbreak was identified.

Hepatitis A virus RNA detected infrom clinical specimens was sequenced and compared withto other isolates at CDC. The Florida RNA sequences matched each other and those of from a concurrent Kentucky outbreak associated with the same restaurant chain; a serum sample from another outbreak in Nevada also shared this sequence. No food handler or other individual was identified as an index case in the Kentucky outbreak. Implicated restaurants in Florida and Kentucky shared a common supplier for some ingredients, including green onions.

Although most foodborne outbreaks of hepatitis A are due to food contaminated by an infected food preparer, we believe ingredients were contaminated prior to arrival at the outlet in this outbreak. The evidence for this is summarized as follows:

• All food handlers and employees who had worked at the fast-food outlet and the convenience store during the exposure period were tested for recent hepatitis A infection, and no index case was identified.
• Viral sequencing studies indicate all cases were infected by a single source.
• Two other outbreaks are linked by the same viral genetic sequence, indicating a common source for all 3 outbreaks.
• Ingredients from the Florida and Kentucky outbreaks were distributed by a common supplier to the outlet in each state.

The most likely contaminated ingredient is green onion, for the following reasons:

• Green onions held the strongest association between an ingredient exposure and illness in our study.
• Green onions are biologically plausible vehicles of transmission for hepatitis A virus. Green onions are uncooked, and minimally handled or processed between field and restaurant. Green onions contaminated before distribution have been implicated as a source of hepatitis A in at least 3 other outbreak investigations.
• Green onions had the second-strongest exposure association with illness in the linked Kentucky outbreak, although the association was not statistically significant. Green onion exposure was reported by 28 of 32 (88%) of cases in Florida and Kentucky outbreaks combined.
• Other ingredients were either not distributed by the same supplier to both outlets, were less plausible as a vehicle from a biologic standpoint, and/or did not have a statistically significant association with illness.

The An FDA has initiated a trace-back procedure for green onions and tomatoes, as a result of initiated based on our findings and information from the CDC, is ongoing.

Appropriate control measures were taken. We recommend continued emphasis on the importance of hand washing to the public and in particular to employees at food outlets and childcare centers, as the most important measure in preventing enteric disease. Close attention should be made towards proper employee hygiene and food preparation practices in food outlets. Surveillance for additional cases of hepatitis A should be maintained. Use of viral molecular testing should be considered in future outbreaks. Other recommendations may be appropriate based on findings of the FDA trace-back.

On April 25, 2001 the Seminole County Health Department was contacted by the Acting Administrator at an assisted living facility in Winter Springs, Florida. He stated that one resident became ill with diarrhea and vomiting on April 23. Four more residents became ill on April 24, some with only diarrhea and others with both diarrhea and vomiting. The total ill on April 25 was reportedly 16. He also stated that 2 staff members were ill with the same symptoms.

The Seminole County Health Department epidemiologist visited the assisted living facility on April 25, 2001. Information on Norwalk-like virus was given to the facility, as well as some sterile containers to collect stool or vomitus samples. The epidemiologist was given a tour of the facility, as well as a map to visualize the location of the cases. The Acting Administrator provided a line list of those residents and staff that were ill at that time. A more detailed grid was given to the facility to track the illness as it progressed.

This assisted living facility had 80 residents and 46 staff members at the time of the outbreak. No illness was reported among the food handlers at this facility. Preventive measures were put in place at this facility the same day that the health department was contacted. All meals were served in the resident’s rooms, group activities were cancelled, masks were used to clean up any spills and to do laundry, proper hand washing was stressed, and gloves were used when appropriate.

For this outbreak, a case was defined as a staff member or resident of the assisted living facility in Winter Springs, Florida, who experienced diarrhea or vomiting between April 11, 2001 and May 1, 2001. Twenty-six residents and 12 staff persons fit this case definition.

The overall attack rate was 30% (38/126). The attack rate among the residents was 33% (26/80) and among the staff was 26% (12/46). The cases were not localized to any particular area of the facility. The dates of onset ranged from April 17 to April 27, 2001. Figure 1 shows the epidemic curve, including both residents and staff members.

Conclusions and Recommendations

The source of this outbreak was not determined. Based on the information that was given to the health department, it appears that one or several staff members may have brought this illness into the facility. The virus was likely spread by person-to-person contact rather than through a common food source.

Norwalk-like virus has an incubation period of 12 to 48 hours and duration of 12 to 72 hours. The symptoms include nausea, vomiting, diarrhea, cramps, low-grade fever, chills, headache, body ache, and fatigue. The attack rate for this virus is generally very high among persons that are exposed, probably due to the low infectious dose required to cause illness. It is common in outbreaks of this virus for diarrhea to be prevalent among the adults and vomiting to be more prevalent in the children. Complaints of severe headaches, along with gastrointestinal symptoms, among the ill are a clue that Norwalk-like virus could be the cause of illness. Transmission of this virus occurs through water, food, aerosolization, contaminated surfaces, and person-to-person contact.

This facility was very effective in stopping the continued spread of this virus among its residents and staff members. The acting administrator was very responsive to the suggestions that were presented to him, which likely prevented the virus from spreading even further than it did.

The conditions of this outbreak show how important it is for people to stay home if they are experiencing any type of gastrointestinal illness. Viruses like Norwalk are very easily spread and assisted living facilities are particularly vulnerable to illnesses such as this. Employees and visitors must understand that they should not enter the facility if they are ill. Education about the need to stay home when ill, as well as education on good hand washing techniques, would be extremely helpful in preventing future outbreaks.

On January 25, 2001, the Palm Beach County Health Department’s Division of Epidemiology and Disease Control received notification from a local hospital of a single confirmed laboratory case of Cyclosporiasis. On January 30, 3 additional laboratory-confirmed cases of Cyclosporiasis were reported to the Palm Beach County Health Department. Three of the four cases had exposure to a country club in Boca Raton. The Palm Beach County Health Department began an investigation along with the Bureau of Environmental Epidemiology and the Department of Business and Professional Regulation staff.

For this outbreak investigation, a case was defined as any individual who had exposure to the implicated retail food facility (country club) during the month of January, 2001 and who had onset of illness from January 10 - Feb 8, 2001 and who a) had laboratory confirmation of with Cyclospora cayetanensis or b) reported symptoms of 3 or more episodes of diarrhea in a 24-hour period and at least one other GI symptom or c) had diarrhea and 3 or more GI symptoms consistent with cyclospora infection. Controls were selected from callers who contacted the PBCHD after a media announcement regarding the outbreak, and had eaten at the retail food facility (club) during the same time period.

Of the 109 interviews conducted with symptomatic and asymptomatic area residents (most of them country club members) 39 met the case definition. The seven (7) laboratory-confirmed community cases were not included in the analysis, since their exposure could not be linked to the retail food facility involved. There were two cases with earlier dates of onset (September 9, 2000 and December 12, 2000) that did not meet the case definition. Thirty-nine reported experiencing diarrhea; median 5 episodes per day (range 1-30 days); 30 (77%) reported fatigue; 24 (63%) gas/bloating; 23 (59%) cramps; 21 (55%) loss of appetite; 23 (59%) nausea; 20 (56%) weight loss; 11 (28%) vomiting; 10 (26%) chills; 9 (23%) headache and bodyache; 8 (21%) reported fever (median 100° F); and 5 (13%) constipation. Median illness duration was 4 days (range 1-30 days). There were six (6) hospitalizations. Among those hospitalized, two (2) were country club patrons and 4 were community cases. No deaths were associated with the outbreak. Sixty-six percent (66%) of cases were female. The median age cases was 71 years.

The outbreak appears to have occurred among residents of a country club and among residents of the surrounding area in the southern section of Palm Beach, Florida. Also involved were several of the employees of the country club facility. Review of the onset dates suggests that the employees were exposed at the same time as the patrons. Figure 1 shows an epidemic curve of the laboratory-confirmed cases in Palm Beach County from January 10-February 8. The epidemic curve in Figure 2 suggests the possibility of three clusters or subgroups in the outbreak. The shape of the epidemic curve in cluster A suggests a common point source exposure, and includes most of the laboratory-confirmed country club cases. This cluster first brought the outbreak to the attention of the Palm Beach County Health Department on January 25, 2001.

Illness in these cases began January 10, 2001, peaked on the January 16, and tapered on or about January 20. Based on the incubation period of cyclospora, the potential exposures of this cluster may have occurred as early as January 1, or as late as January 19, 2001. Cluster B occurred in the interim week between clusters A & C, and involved laboratory-confirmed community cases with onset of illness January 22 through February 1, and probable exposure to the etiologic agent occurring anytime from January 8 through January 31, 2001. The third cluster (cluster C) once again involved mostly country club patrons with onset of illness January 28 through February 8, 2001 with potential exposure dates January 14 through February 7, 2001.

The bell-shaped distribution of the epidemic curve for cluster A includes only country club cases and employee cases. This suggests a common point source exposure to the agent that may have occurred at the club facility. The employees are not residents of the local area and are not likely to shop at local food outlets or specialty markets nor do they have other common exposures with resident cases. The employees are encouraged to eat at the facility and many do so, suggesting that employees and patrons alike were exposed to the infectious agent at the country club.

The community cases not associated with the club (cluster B) may have been exposed by another source, sharing a common supplier of the contaminated food product with the club. Cluster C, consisting mostly of resident cases, may have been exposed to a similarly contaminated source as in cluster A, through cross-contaminated products, or through newly delivered contaminated products. Alternatively, the community cases (cluster B) and country club cases (cluster C) may have acquired the agent from a common community source not associated with the club. There are several local food outlets and specialty markets frequented by the local area residents and club members. Review of the common suppliers to the club and the local specialty markets revealed that common produce suppliers are frequently shared.

The environmental investigation of the food service facility revealed several improper food handling procedures including: cross contamination of soiled wiping cloths onto food contact surfaces; wiping cloths not stored in proper sanitation dilution; the dishwasher was not properly sanitizing; and hand washing facilities were not being used properly. Six of the ten ill employees tested positive for Cyclospora cayatenensis.

Produce purveyors for the country club were interviewed about the suppliers of their produce. A list of the produce purveyors for the country club was also obtained. Purveyor A provides 95% of the produce to the country club. Purveyor B provides 5% of the produce to the country club. The interview process also revealed that there were several local produce markets frequented by both country club cases and community-acquired cases. These local markets were also asked about the origin of their produce.

Considering the history of cyclospora causing illness through contamination imported produce (most notably raspberries and leafy lettuces) one plausible explanation for the outbreak is that contaminated produce eaten by case-patients from all three clusters may have originated from a common food supplier made available through multiple food retail outlets.

Statistical analysis of the data collected did not reveal any significant association with any food group investigated. Univariate analysis of foods most frequently associated with cyclospora contamination did not establish a significant association with illness within the outbreak.

Fueled by a continuing drought [1], wildfires are currently burning in several portions of the state. According to the Florida Division of Forestry, over 1400 wildfires have burned approximately 100,000 acres in Florida between January and March of 2001 [2].

The wildfires may cause an increase in the air pollution in the areas surrounding the fires. High levels of smoke and pollutants may cause health problems for some individuals. This degraded air quality can adversely affect the respiratory system and may aggravate heart and lung problems such as asthma and emphysema.

Persons in the affected areas who suffer from respiratory conditions or heart disease should avoid physical exertion and outdoor activity. This caution also extends to young children and the elderly. Individuals who have respiratory conditions and do not have access to air conditioning should consider spending time in areas such as shopping malls, public libraries or museums. If you feel you are experiencing adverse health affects, contact a physician for an evaluation or advice.

The article below appeared previously in the "Epi Update" and summarizes the epidemiologic response to the wildfires of 1998.

References

During the summer of 1998, concern about the possible adverse health effects of the wildfires led the Volusia County Health Department and the Bureau of Epidemiology to conduct hospital-based surveillance of 13 conditions including asthma, acute bronchitis, and conjunctivitis (see previous article).

The 1999 wildfire season has started in Florida. If our past experience is any indication, county health departments may be asked by politicians, physicians, and the general public to determine if the fires are causing an increase in the frequency of certain conditions. Since asthma, acute bronchitis, and conjunctivitis, are not reportable to the Department of Health and county health departments (CHDs) by law, data on these conditions must be specially requested from the local health care providers.

Last year when gearing up for this project we learned a few lessons that might be helpful for others contemplating active surveillance for fire and smoke related morbidity

First - We found that collaborating with both the hospital infection control practitioners (ICPs) as well as staff from the medical records department was beneficial. The ICPs, at times, acted as liaisons between the medical records staff at their respective hospitals and the Volusia CHD. This strategy seemed to facilitate the rapid transfer of data.

Second - Our overall approach to data collection was to keep it simple. We chose a series of 13 conditions that were easy for the hospital staff to identify and collect information on. Next time, perhaps we would choose fewer. Our request did not ask for the number of cases of a particular condition by ICD-9 code. We also found that coding practices varied somewhat from hospital to hospital. For example, asthma is represented by several codes. A particular hospital may never use all of these codes when coding an asthma case. In order to avoid missing cases, we furnished the hospitals with plain English text and avoided the used of codes.

Third – We chose a particular time frame for the current fire event and the same period for the previous year for comparison and used the same parameters for all hospitals surveyed.

Fourth – We used a simple data collection instrument and took the time to explain it to staff involved in its completion. This time on the front end paid dividends by promoting cooperation, avoiding delays and encouraged completion by a certain date. Whenever possible the form was hand delivered to the hospital so staff participating collection of the information requested had an opportunity to meet us and ask questions they might have. If this was not possible we faxed the form and followed up with a personal phone call. A sample table is shown below.

Frequency of Emergency Room Visits for Selected Conditions, and Percentage Change during May 1998 and 1999, X County

Fifth – In hindsight, we now believe it is only necessary to ask for data on emergency room (ER) visits. Last year we also asked the hospitals to furnish the numbers of admissions for the same 13 conditions. This no doubt delayed our data request, and, furthermore, the number of hospital admissions did not convey much information. The numbers were small and the overall change between 1997 and 1998 was small.

Sixth – We also learned to recognize and appreciate that coding practices may vary within one hospital over time. For instance, a large increase or decrease between 1998 and 1999 for a particular condition for a particular hospital may mean that in 1998, ER staff coded ER visits, but this year medical records staff coded ER visits.

Finally – One must always keep in mind that these are surveillance data – nothing more. Any increase in the frequency of the conditions over time is not necessarily due to the fires. Also, certain persons who suffered from these conditions may have never presented at a hospital because they chose not to seek medical care or were seen by their private physician.

Those with any additional questions regarding the institution of an active surveillance at hospital ER’s in their area may direct them to Mr. at hospital ER’s in their area may direct them to Mr. Zuber Mulla at (407) 623-1212 or SC 334-1212 ext. 178. Mr. Mulla is also on the departmental Intranet via cc-mail.

Save them as favorites or put them on their desktop for quicker access.