Emerging Infections of International Public Health Importance

Objectives:

To be able to describe the changing pattern of food borne disease outbreaks
Describe and understand the epidemiologic and laboratory tools available to investigate these outbreaks.
Understand the need to cross agency and political lines to investigate and control food borne outbreaks.

The discussion will primarily focus on food borne outbreak investigations in the United States. I will present a series of outbreaks which occurred over the past 20 years and give an overview of the process of outbreak investigation.

The old paradigm for foodborne outbreaks was the small "church supper" outbreaks. These were frequently caused by a contaminated food or food handler and were usually related to foods of animal origin such as chicken, milk or eggs.

The new paradigm consists of widely dispersed cases with a central contamination point. Not all foods are of animal origin and more frequently are related to fresh fruit and vegetables.

This new paradigm requires close coordination for outbreak investigation and control:

E. coli is a gram negative bacteria that is commonly found in the gastrointestinal tract of humans and animals. It has several different subtypes many of which can produce disease either by direct invasion of the bacteria or by toxin production. One of these, E. coli 0157:H7 has been associated with several outbreaks in the United States and worldwide.

Of those infected with this organism a small percentage will go on to develop a severe form of the disease called Hemolytic Uremic Syndrome (HUS). The classic clinical triad is microangiopathic hemolytic anemia, thrombocytopenia, and renal failure.

This diagram shows the spectrum of illness and infection with E. coli O157: H7. The majority of persons will get simple diarrhea, a smaller percentage will progress to bloody diarrhea and of these an even smaller percent will go on to HUS.

E. coli O157:H7 was first identified as a problem in 1982 in outbreaks in Oregon and Michigan. These outbreaks were found to be associated with contaminated hamburgers from fast food restaurants after months of investigative epidemiologic work and laboratory testing.

In 1984, we decided to see if Washington State had a problem with this organism. We started routinely testing all stools sent for bacterial culture from a local health maintenance organization (HMO) and tested it for E. coli O157: H7 and found that E. coli was very common here. This initial work led to the identification of our first E. coli outbreak in Walla Walla, WA in 1986. There were 37 culture positive cases of E. coli O157:H7 diarrhea which eventually was linked to a fast food restaurant chain. Some of the severe cases were transferred to Seattle for management and further investigation led to the conclusion that there was a larger outbreak happening around the state. Based on this, E.coli O157:H7 was added to the reportable infectious disease list in Washington State.

Surveillance showed that from 1988-1992 there were 1180 cases of E. coli O157:H7 diarrhea including 3 deaths. The highest incidence rate was seen in children and the most common exposure was ground beef.

In 1993, a pediatrician at the Seattle Children's Hospital notified me of an unusual cluster of bloody diarrhea, O157, and HUS. He contacted us to see if there was a more extensive problem. From the initial epidemiologic investigation there did not appear to be a single exposure source for infection. Emergency rooms and laboratories were alerted to begin active case finding on persons presenting with diarrhea. Within 1 week, 37 cases were identified and 27/37 ate at the same hamburger chain although not the same restaurant.

A case control study was performed. Controls were matched by neighborhood and age and had no diarrhea in 2 weeks before the cases illness began. The case control study implicated the fast food chain restaurant and the organism was isolated in all 9 lots of the regular hamburger patties that had been shipped to this chain throughout Washington. DNA fingerprinting showed that the patient isolate strain was indistinguishable from the strain isolated from the confiscated hamburger. The intervention consisted of a recall of all the hamburger meat, notification of other agencies, and public awareness and education. The time from first notification to the recall of the hamburger meat was 5 days. This timely response was helped by the existing infrastructure in the laboratory, epidemiology and by an astute clinician to make the connection.

This slide shows the epidemic curve of this outbreak by date of hamburger consumption. Once intervention, on January 18th, went into place the curve drops off dramatically.

This outbreak eventually involved 66 restaurants in the Pacific Northwest and an estimated 600 people became ill, the majority with bloody diarrhea. Approximately 250 thousand contaminated hamburgers were recalled and destroyed, thereby saving thousands of cases of illness. The investigation further identified a problem with the proper cooking temperature of the hamburgers. The national standard for hamburgers was to cook to 140 degrees. This was found to not completely cook the hamburgers. Washington State made a new standard which raised the required temperature to 155 degrees. The restaurant chain in question however had not heard of this change and was cooking their hamburgers at the lower temperature. Nationally the investigation found that the contaminated hamburgers came from a California plant which produced 400,000 hamburgers daily. The contaminated hamburgers were produced on a single day and most of the hamburgers were sent to Washington State.

What happened since 1993? The US government provided more money to improve foodborne disease control in the National Food Safety Initiative. For Washington State, this meant developing new types of laboratory tests commonly referred to as “DNA fingerprinting.” These are being used to trace epidemics. While DNA fingerprinting existed in 1993, it was only a research tool. Results were only available months after outbreaks occurred. Since 1996, fingerprinting has been performed as isolates are submitted to the state laboratory.

The tests involve breaking apart the DNA, or genetic code, and separating the pieces in an electrical field. Bacteria generating the same patterns may come from the same source.

By 1996 in Washington – all O157 specimens were tested with two methods of "DNA fingerprinting," microrestriction fingerprinting, referred to as "RFLP" on the next slides, and pulsed field gel electrophoresis (PFGE).

In October 1996 there were 2 cases of HUS associated with E. coli O157:H7. Restriction Fragment Length Polymorphism (RFLP) was performed on these cases as well as on 7 other cases and the RFLP results were the same. A case control investigation was undertaken and identified a specific apple juice product as the source of the contamination.

The following slides show the sequence of events for the investigation of this outbreak.

These 2 slides show the initial timeline and after the case-control study led to a product recall. We can see that the recall did not wait for the organism to be identified from the food product which occurred several days later.

Our investigation showed (see slide) that there was contamination and cases in several western states and British Columbia.

This slide shows the various distribution points along the West Coast. In yellow is the amount of apple juice sold and in red are the reported cases. Except for Seattle, where O157 surveillance was increased, the number of E. coli cases are proportional to the amount of apple juice distributed. This argues that the site of contamination was not at the local level but at a central site.

This slide shows the apple juice manufacturing process. Investigation of the apple juice plant did not find any ill workers, no evidence of equipment failure, and environmental cultures were negative. It was felt the contamination did not come from the factory site. However, the plant did not pasteurize the apple juice after processing.

Pre-plant investigation model shows the difficulty with isolating a single source since some of the harvest goes directly to the plant and others go to a packer. We never found the exact source of contamination but could have been any one of 17 different farms or 4 different packing houses. The public health intervention and future prevention was to start pasteurizing the apple juice and this was initiated.

Since 1996, a network of laboratories, the Pulsenet system, was established which performed PFGE in a standardized matter. These laboratories now cover the majority of states in the US. Digital images of the PFGE patterns are stored in computers locally and in a center database at the Centers for Disease Control. It is possible to seach this database, identifying bacterial strains with the same DNA fingerprints.

The following is an example of how the Pulsenet system is used for early detection of outbreaks.

In 1999, 3 Salmonella muenchen cases where isolated from stool cultures. This particular serotype was unusual and therefore caught the attention of the microbiologist who notified the local public health department.

All three cases drank a blended fruit drink produced in different restaurants, but which were part of the same company or “chain.” Since different fruit drinks were made from different combinations of fruits, sick and well people were interviewed about what types of juice they drank at the restaurants. As shown on the next slide, blended drinks consumed by sick people contained bananas, orange sherbet, and unpasteurized orange juice. Since bananas were peeled at the restaurant, and the orange sherbet was pasteurized, the most suspicious common item was the unpasteurized orange juice.

At the same time, an outbreak occurred in Oregon involving the same orange juice.

DNA fingerprinting confirmed that the Salmonella strains in the Washington State and Oregon outbreaks were indistinguishable. This indicated that they outbreaks shared the same contamination source, which was an Arizona company.

This slide shows the onset of illness and number of cases until the source was identified. We can see that cases dropped off dramatically once the juice was recalled.

This slide shows the states where cases where identified which corresponded to distribution by the company.

Investigators were able to get a sample of orange juice from a truck from Mexico before it was unloaded. This was also positive for Salmonella muenchen matching the DNA fingerprints of the isolates from Washington and Oregon. It was thought that the contamination occurred in Mexico either by contamination of the orange juice or by contaminated ice used to refrigerate the orange juice on its 1000 mile journey.

In conclusion, foodborne outbreaks do occur and prevention and control of these outbreaks needs to rely on quick and accurate information that is disseminated across agency and industry lines. The public health infrastructure should be in place to quickly identify outbreak situations and be able to act on an appropriate response.

Study Questions:

Describe some of the epidemiologic tools to investigate a food borne outbreak.
How are most food borne outbreaks detected?
What type of surveillance system is needed to identify and control food borne outbreaks?