The Wisconsin Investigation: Rapid communication curtailed 2006 E. coli outbreak

This is EIStories, a series of articles about outbreak investigations conducted by CDC EIS officers.

The EIS is a training ground for many infectious disease professionals, and most look back on that time with fondness for what they learned, the people they met and the service they gave. Epidemiology is an important part of infectious diseases, and we think a look at some of the more important cases may be interesting and informative.

Prompt action by Wisconsin health officials, communication with officials in other affected states and coordination with federal agencies may have stemmed what could have been a far more serious outbreak of Escherichia coli in Wisconsin.

On Sept. 5, 2006, local health departments from three non-contiguous counties in southeastern Wisconsin — Dane, Manitowoc and Ozaukee — reported clusters of laboratory-confirmed cases of Escherichia coli O157:H7 to the Bureau of Communicable Diseases and Preparedness of the Wisconsin Division of Public Health.

Jeffrey P. Davis, MD, chief medical officer and state epidemiologist for communicable diseases at the Wisconsin Division of Public Health, recalled the events of 2006 in an interview with Infectious Diseases in Children.

“There was no clear focus at the outset,” he said. “Local epidemiologists found that four of the five sick individuals from Manitowoc County had visited an animal exhibition at the county fair, but that was the only real signal.”

That said, Davis noted that cases began emerging in Ozaukee County. “There was no apparent connection between the outbreaks in the two counties at first, but we were aggressively obtaining specimens from the beginning,” he said. “This is standard procedure for us with regard to E. coli.”

There was concern over the parallel outbreaks, and results of pulse-field gel electrophoresis (PFGE) analysis were anxiously awaited.

Key phone call #1

The pivotal moment came in the form of a phone call on Sept. 7 from the Southeast Wisconsin Blood Center, according to Davis.

“They got requests for plasma exchanges that occurred in the prior three days,” he said. “They had a hunch that it was to treat hemolytic uremic syndrome, and they turned out to be right. They thought it was unusual and so did I. I consider this to be a sentinel event. That is an awful lot of hemolytic uremic syndrome. In fact, I consider even one case to be a sentinel event.”

By this time, aggressive investigation of the affected counties was underway. When such outbreaks occur, the Wisconsin health department routinely used a long-form questionnaire that covered a broad spectrum of food habits for the previous 7 days.

“Soon after that call on the 7th, the first thing I did was call a staff meeting,” Davis said. “I knew we were conducting investigations in the counties, and I wanted to gather as much other information as I could, rapidly. We had someone calling hospitals, someone sent an email to the health alert network. A lot was going on in that period of time. We called CDC and informed them of a striking occurrence of hemolytic uremic syndrome.”

Initial PFGE results began to trickle in, but no definitive pattern immediately emerged. However, the results were put on Pulse Net, and when other isolates from other parts of Wisconsin came in, some matches were found. Furthermore, there were matched isolates from three other states: New Mexico, Oregon and Utah.

“We were giving an extended form questionnaire because by then it was clear that there were associations with spinach,” Davis said. “This was really important information. As it turns out, when we called CDC on Sept. 13, we found out that health departments in other states were arriving at similar conclusions.”

Key phone call #2

Davis said that an official from the Oregon Health Department called CDC within moments of when he placed a call to CDC on Sept. 14.

“Then, later in the day, we had a conference call involving multiple states, the FDA, CDC and the USDA. Everyone was compelled to act. CDC issued an alert, the USDA issued an advisory.” Davis said. “We wanted a nationwide hold on spinach until more data came in. It took time to hone in on the brand of spinach, but once CDC developed the spinach questionnaire, detailed analysis followed.”

The spinach questionnaire targeted four brands, dubbed brands A-D. Though final data would not come until later, it became clear that the strongest associations with illness were observed in individuals who had eaten brand A spinach.

With more information on the source of infection, the Wisconsin health department conducted a case-control study. The 49 case-patients in Wisconsin were matched by age, sex and residential location with 86 controls. A case-patient was defined as having onset of illness between Aug. 20, 2006 and September 2006.

Final results indicated that the risk for illness associated with spinach consumption was 82.1 (95% CI, 14.7 to <1000). Twenty-six of the 49 patients recalled eating brand A spinach. Multibrand analysis results indicated that only brand A was associated with illness.

Wisconsin health officials obtained 11 previously opened bags of brand A spinach and one non-packaged spinach specimen from the homes of case patients, according to results. Though some failed to recall or recalled incorrectly the brand of spinach they had purchased, Wisconsin investigators asked to see grocery store shopper cards. Shopper card information confirmed that 17 patients had purchased brand A spinach. Two of those patients had not identified that brand in previous interviews.

Lab analysis results indicated that isolates matching the outbreak pattern were found in two bags of spinach corresponding to three case patients. Further analysis revealed that both bags were produced on Aug. 15, 2006 at the same facility during the same work shift.

“The results were pretty conclusive,” Davis said. “This helped in working with the manufacturer. Once the data were there, which it was fairly directly, the manufacturer was compliant.”

Davis said that knowledge of the date, facility and work-shift was useful for officials conducting on-site investigations in California. “Investigators were able to work back to areas where the spinach was likely to have been grown,” he said. “It was narrowed down to fields in three counties. The specificity and efficiency were remarkable.”

E. coli organisms matching the infecting strain were found in water and animal fecal samples in that particular three-county area in California, according to Davis. He said that samples were obtained from livestock manure and from the stool of feral animals that had access to those fields.

Impact and legacy

Though this was not first outbreak of gastrointestinal illness from leafy green vegetable products, it did make national news, in part because the FDA had launched the Lettuce Safety Initiative in August of 2006, which is when the first cases of E. coli were occurring, according to Davis.

“The outbreak focused attention on the initiative, and vice versa,” Davis said.

The size of the outbreak also contributed to the national attention it received. “Whenever you have more than 200 illnesses and three fatalities on record, you are dealing with something substantial,” he said. “Not to mention the hemolytic uremic syndrome, which should always cause concern.”

Davis said that the built-in time factors involved in investigating and curbing an outbreak of gastrointestinal illness from food consumption should highlight the need for speed and efficiency in epidemiology departments.

“There are a lot of steps involved from consumption, to onset of illness, to illness severe enough to see a health care provider, to appropriate diagnosis of Shiga-toxin producing disease, to collection of specimens, to reporting in-house, to lab analysis, to reporting to the public,” he said. “Anything that can be done to cut down the time between these steps needs to be considered.”

For example, the Wisconsin department has a courier service to deliver samples and specimens from one laboratory to another. There is also a person on staff who can read PFGE patterns, which can eliminate the process of waiting for often overworked lab analysts. Davis said that the department is “not shy about redundancy. We have motivated individuals who like to be informed with emails, press releases, postings, whatever it takes.”

Beyond the health department, Davis has some other opinions about why an outbreak like this is significant.

“People eat spinach because it is healthy,” he said. “They think they are doing the right thing. For the most part, they are. But when people get sick from something as healthy as spinach, it can undermine many aspects of the food service and public health industries.”

Davis said that the onsite investigations proved that there will always be risks involved in eating raw vegetables. He said that once the organism is inside the lettuce, it cannot be washed off, and that individuals on the production end are likely more aware of this fact since the investigation.

“Despite the size the outbreak and the unfortunate fatalities, this was a very instructive outbreak,” Davis said. “A lot of systems were helped by this.”— by Rob Volansky