University of Washington Department of Health Services Food-borne E.coli

HOME List of Lectures Return to Title Page

INDEX Introduction to E.coli O157:H7 Ways of Looking at E.coli The 1993 Outbreak Outbreak Response 1984-Washington State Surveillance 1986 Outbreak in Walla Walla Screening and Reporting of E.coli O157:H7 1993- Outbreak in Washington State Some Main Points Regarding that Outbreak Major Decision-Making without all the Facts Cooperation with the Restaurant Chain Legal Power of of States to Recall Products National Recognition Major Technological Improvements in the Laboratory 1996-Apple Juice Outbreak Regulation Changes of Pasteurized Products? 1997 Hamburger Outbreak and Recall Patterns of Outbreaks: Old versus New Creative culture-based intervention - Salmonella Forthcoming Change: What needs to be done? Readings Introduction to E.coli O157:H7 Click on image to enlarge Two antigens are primarily used to classify E.coli. One is the somatic, or O antigen that is related to the body of the bacteria and the other is the flagellar, or H antigen. Generally, 3-4 days after ingestion of the bacteria, persons develop diarrhea, abdominal cramps and may then proceed to bloody diarrhea. The disease can be mild without any serious sequelae or progress to hemolytic uremic syndrome (HUS) or thrombotic thrombocytopenic purpura (TTP). The severe forms of the disease are more common in children. Ways of Looking at E.coli This Venn diagram again shows the disease spectrum for E.coli O157:H7 This slide shows the time sequence for the infection. There is almost always a time delay in the diagnosis of this disease. The 1993 Outbreak In 1993, there was an outbreak of E.coli O157:H7 associated with fast food hamburgers (1). There were about 600 confirmed and probable cases, of which 151 were hospitalized. There were 45 cases of HUS, 28 required hemodialysis and 3 children died. It was the largest outbreak of E.coli O157:H7 reported at that time and remains the largest in the United States. We were able to identify the source of the outbreak within 5 days of our first notification. We were able to recall about 60% of the contaminated meat, which were about ¼ million hamburgers. Our estimate was that if that had not been done, the numbers would have been about twice as bad as what they were. Ability to Respond to the Outbreak Our ability to respond to this outbreak was not accidental or an incredible stroke of good luck. It was a capability that was developed over many years, and we were ready when the 1993 outbreak occurred. 1982-First Identification of E.coli O157:H7 outbreaks E.coli O157:H7 was first identified as a human pathogen in 1982 in two outbreaks (2). One was in southern Oregon, and the other was in Michigan. These were outbreaks of painful bloody diarrhea of unknown cause. Routine laboratory testing ruled out Salmonella and other enteric bacteria. Epidemiologic investigation did implicate hamburgers from a fast food chain. Months later, the CDC identified E.coli O157:H7 as the causative agent. 1984-Washington State Surveillance In 1984, we began to work on E.coli O157:H7 in Washington State. A case-control study was done looking at diarrhea at a large health maintenance organization (3). Everyone with diarrhea who submitted a stool sample was tested for E.coli O157:H7 in addition to the usual enteric pathogens. The laboratory diagnosis for E.coli O157:H7 uses the inability to metabolize sorbitol as a marker. Confirmation can be done using tests for the O and H antigens. These tests are relatively simple and inexpensive. The study showed that there was lots of E.coli O157 in Washington State. It was more common than Shigellosis. The case control study did find a clear relationship with eating undercooked hamburger. 1986 Outbreak in Walla Walla While we had identified several outbreaks, the most notable happened in 1986 in Walla Walla, Washington. There were 37 cases of E.coli O157:H7, where three people developed TTP, two of whom later died (4). This outbreak was related to a fast food taco restaurant. Several features of the outbreak were notable. First, the outbreak was identified not from Walla Walla but from University of Washington where patients with TTP was transferred for care. Second, university physicians made the connection to E.coli O157:H7. Third, a second outbreak of E.coli O157:H7 was taking place in a nursing home close to Walla Walla. We were unable to identify any links between the two sites. We were able to implicate a food source, but were unable to confirm this microbiologically because we were notified of this problem late in the course of the illness. During this time we revised our reportable disease regulations and added E.coli O157:H7 to the list. Screening and Reporting of E.coli O157:H7 We were the first state to require the reporting of this disease. Washington state has a long history of aggressive food-borne disease investigation. For many years we would report more food-borne outbreaks to the CDC than any other state by number except for New York, and by population except for Hawaii. That's not to say that our food is any worse than anywhere else in the country, but that people are aggressive about identifying and reporting outbreaks. That laid a lot of groundwork for what was done with E.coli. Much of the work to identify the cause of HUS was done by 2 physicians, Dr. Phil Tarr and Dr. Peggie Neill (5, 6). Their hypothesis was that HUS is caused by E.coli O157:H7. They were able to study most of the cases in the state with HUS since the majority came to Seattle for kidney dialysis. Since then, we found found that almost all the HUS cases could be ascribed to E.coli O157:H7. 1993-Outbreak in Washington State Consequently, when Phil Tarr called me on January 12, 1993, saying, "I've never seen anything like this before." I knew something was very wrong. In 36 hours, there were 11 cases of HUS or bloody diarrhea at Children's Hospital which were probably caused by E.coli O157:H7. This was the start of our epidemiological investigation into the source. Most of the cases were initially reported in King county however after the initial investigation, 3 counties were also reporting cases. By the week's end, we has 37 cases, and we became progressively more suspicious about a single fast food chain. Seventy five percent of the cases ate there. There were about 13 restaurants initially named by the cases. We found that this restaurant chain had not increased the cooking temperature for their hamburgers to the required 155 degrees. We instructed them to change their cooking temperatures immediately. At the same time, the epidemiological investigation, a case control study, was being done. This confirmed the suspicion of the chain as the culprit since ¾ of the cases ate there and none of the controls. This prompted other states including Nevada and California to look for cases of E.coli O157:H7 and they found outbreaks as well. [Question: When did you isolate the E.coli O157:H7?] It was about five days after we made the public announcement. While we don't usually deal with outbreaks of this size, it is common for us to make decisions and public announcements based on epidemiologic data, before laboratory confirmation is available. Some Main Points Regarding that Outbreak First, the epidemiological tools we used to identify the source were not exotic. It was a case control study, field epidemiology and food questionnaires. The second point is practice helps in outbreak investigations. Needless to say, when you are dealing with a problem like this, you do not want to make a mistake. The chances of being right get better if you have done it before. As I mentioned, there is a history of good food-borne investigations in Washington State. Major Decision-Making without all the Facts In major decision making situations like this, decisions have to be made with incomplete information. You don't have all the data. If we waited for all the data to be available, all the hamburgers would have been eaten. On the other hand, the quality of partial information, or the ability to make good decisions with partial information increases markedly if you get that data from established relationships. In Washington State, we had been working together for a long time. How much weight to put on critical pieces of information is easier if you've been working with folks before. Cooperation with the Restaurant Chain In investigations like this, we meet with the implicated food producer or restaurant and try not to have a hostile adversarial relationship. We work together and try to find out what is going on. Many times, the food producer or the restaurant chain may have ideas on how things were handled that may help solve the problem. Legal Power of Washington State to Recall Products [Question: What kind of legal power do you have as the State Health Department, to go in and shut them down or require that they give you access…] Regarding recalls, there is general wording within the Washington Administrative Code, that says that either the State or the local health officer have the authority to quarantine, potential sources of infectious diseases, etc. However, it works much better is there is a cooperative working agreement which expedites the appropriate public health action from a company. Most businesses in this situation voluntarily follow our recommendations. [Question: If I remember correctly, you mentioned in your talk that they took away 60% of the contaminated meat. How did they do that?] Most fast food chains have central warehouses that distribute the meat to their outlets. So they usually recall the products from a few prime sources rather than from multiple places. They can also track where products were delivered. In this particular situation it was easier. You are dealing with one chain. The chain could send out information through their communication system. For a few days they stopped cooking hamburgers entirely. [Question: How did E.coli O157H7 originally get to the main grinder?] Where does E.coli O157:H7 come from? We do not have complete information on all of the reservoirs. You can have transmission from one person to another in day care centers, nursing homes, etc. A major reservoir is cattle. Why do some cattle carry it and others don't? No one really knows. The basic hypothesis is that while the cattle are being butchered, there is fecal contamination of the meat, the meat is served to people and people get sick. Hamburger is a much higher risk product than other cuts of beef because it is ground from multiple sources. Contamination from the surface of one cut of beef, can be spread through the entire product. The unanswered question is, where did it really come from? Were they able to trace it back? The CDC, the USDA and the California Health Department did studies and tracing, and were unable to find the source. [Question: Since E.coli O157:H7 is becoming reportable in other states and we have become aware of it nationally, are you still seeing Northwest predominance in cases?] There was a geographic predominance in who was looking. There may well be a certain component of regional differences. There are other types of E.coli that carry a shiga-like toxin, and there appears to be global differences. It wouldn't surprise me that the long term answer is that there is a little bit of both, but I think a substantial amount of the north-south difference is related to where people were looking. Back in 1993, there were 11 states that were requiring reporting, and now 42 or 44 states require reporting. [Question: I know that the state public health department has a very good relationship with the media here, and I think that is very helpful in terms of outbreaks and notification of the public. Can you talk briefly about how you establish those relationships to get the word out?] One of the things that you always do with outbreak investigations or major events is decide who will will be the main person in your agency to work with the media. In the 1993 outbreak, that person was me. Consequently, I was spending a great deal of my time talking to the media, which allowed others to investigate and control the outbreak. In general, it is important to have a coordinated response. Public affairs helped organizing press conferences, and press calls, but most calls came directly to me. It is also a good idea to be as forthcoming as possible with the media. We have worked out a good relationship in that regard. In fact, the media was very helpful in warning the public about secondary transmission. We were concerned about it, because there were about 60 kids who had been in daycare centers who were cases. It was a major concern that there might be more spread. So I talked about secondary transmission all weekend long, after our first public announcement on the previous Monday. I had four television interviews on Sunday. National Recognition Since 1993, there has been national recognition of E.coli O157:H7 as an emerging pathogen, and a lot more work on this organism has occurred on the national scale. Major Technological Improvements in the Laboratory There were major technological improvements in the laboratory, which involved DNA fingerprinting of E.coli O157:H7 isolates. The method is to take DNA from bacteria of a known type, apply certain enzymes to "chop up" the DNA, and separate the components of varying molecular weights on a gel. This allows epidemiologists to use these different subgroups as markers when an outbreak occurs. In this way we can match cases and more readily pinpoint the source of the contamination. While DNA fingerprinting of E.coli isolates existed in 1993, it was a retrospective tool only, available months after the fact. Since then, fingerprinting has become a routine part of O157 surveillance, being performed as isolates are submitted to the state laboratory. For Washington State, we have the luxury of two methods: microrestriction fingerprinting, performed by Mansour Samadpour's laboratory at the University of Washington, and pulsed field gel electrophoresis (PFGE) at the Washington State Public Health Laboratory. We have chosen to use PFGE because it is a standardized and published method performed at the CDC and several state health departments. Currently, the protocol used at all these sites is a one day procedure developed by Romesh Gautom of the Washington State Public Health Laboratory (7). 1996 Apple Juice Outbreak The next large outbreak occurred in 1996 and involved apple juice juice (8). This started as 2 HUS cases were identified at Children's hospital. DNA fingerprinting performed by Dr. Samadpour's laboratory at the University of Washington confirmed that these 2 cases had the same DNA fingerprint and were therefore linked. As the interviews were continued, 8 more cases were identified. DNA fingerprinting and showed these were all identical. The outbreak investigation showed that on brand of unpasteurized apple juice had been consumed by all the cases and by none of the controls. This led to the recall of the implicated juice. This outbreak leads to some strong conclusions: First, the use of DNA fingerprinting as a tool to identify cases was essential to the identification of the source of the outbreak. It allowed us to exclude other cases with different DNA fingerprints from investigation. Second, good field epidemiology led to rapid identification of the source of the outbreak so a recall and public awareness campaign could go into effect. Third, the DNA fingerprinting test that was used was rapid, allowing us to get matching cases within 24-48 hours. Regulation Changes of Pasteurized Products? [Question: Did they change the regulations and the sale of pasteurized products or the labeling of pasteurized products?] There are several issues in the food processing "dogma": One is that acidic products like apple or orange juice not have been a problem because E.coli O157:H7 grows poorly in an acidic environment. Unfortunately, this is incorrect, E.coli O157:H7 can grow well in an acidic environment. There have been previous outbreaks of E.coli related to apple juice where apples that fell into a cow pasture were used without being properly washed. On the other hand, the implicated producer in the 1996 outbreak was using state-of-the-art processing, except for pasteurization. Now they are pasteurizing their juices. However, defining pasteurization for all apple juices or all juices is difficult. There are FDA policy statements that juice should be pasteurized, but its application is more difficult. 1997 Hamburger Outbreak and Recall What has happened in Washington State since the 1996 apple juice outbreak? In 1997, there was an E.coli O157:H7 outbreak in Colorado, associated with contaminated hamburger. Although the number of human illnesses was relatively small, the resulting recall was nationwide, eventually involving 25 million pounds of beef. While we were on the periphery of this outbreak, there was pubic concern on whether contaminated beef was distributed in Washington State, and if there were human illnesses. Because PFGE is a standardized method performed at the CDC and several state health departments, it is possible to perform direct comparisons of PFGE images tranmitted through the internet rather than having to mail and retest isolates. When we receive an image, our computer software searches our database for matches. In 1998, a communication sytem called Pulsenet connecting several states and the CDC was established. In 1997, Washington State was the first to establish such a connection. Soon after the Colorado hamburger outbreak was identified and the recall initiated, PFGE profiles of involved strains were available at the CDC. The images were sent to Washington State through the internet, and our database was searched for similar patterns. No matches were found, which agreed with the initial distribution data for Washington State. Having two independent surveillance methods which agreed was reassuring. Each has its limitations. Surveillance data requires that ill people are tested for E.coli, and are reported to the health department. On the other hand, distribution data can be complex and changing as the recall is expanded. Patterns of Outbreaks: Old versus New A general comment about the patterns of outbreaks is that you can look at it as an old scenario outbreak and a new scenario outbreak. Changing Scenarios for Food-borne Outbreaks Old New "Church supper" or other event Error in food preparation Increase in "sporadic" cases Geographically dispersed Low attack rate Subtyping of strains often essential Traceback intergal part of investigation Dialogue with regulators, industry and academia essential to promote change The old scenario outbreak is frequently involved single food item at a meal. The new scenario outbreak involves an increase in scattered cases around the country due to centralized food production and mass marketing. Creative culture-based intervention This is the Yakima Mexican-style Cheese Story. This represents another form of intervention that really needs to be considered in responses to food-borne outbreaks. There was a outbreak of about 60 or so cases of Salmonella in Yakima County related to Mexican-style cheese. Unpasteurized milk was being used to make the cheese. While it is illegal to buy and sell unpasteurized milk from almost all dairies, just saying that is not going to solve the problem. Many people are convinced that the raw milk cheese tastes better and certainly they were used to eating cheese from raw milk. In response, "Project Abuela,"(9) was started by Val Hillers and the agricultural extension program at WSU. First they studied who makes this cheese in Yakima County and found that there were about 225 cheese makers; all were small home operations and mostly run by elderly women. Hence the word abuela, which means "grandma" in Spanish. There was no single commercial operation. They then worked with some folks to develop a recipe for making Mexican-style cheese using pasteurized milk using buttermilk to put some bacteria into the milk, etc. They were able to produce a recipe for pretty good tasting Mexican-style cheese using pasteurized milk. They translated this recipe into Spanish and, with money from the dairy association, distributed aprons, measuring cups, thermometers etc., to 15 abuelas in Yakima County. They trained these women on the proper use of pasteurized milk to make Mexican-style cheese with the idea that these abuelas would in turn train other people in their community. They are doing a before and after study to see what effect has occurred. The important message is to work with the community and to find culturally acceptable alternatives which produce safer food. Forthcoming Change: What needs to be done? Improve rapid diagnosis of potential food sources of an outbreak. Improve the ability to identify the exact place in the food processing chain where contamination occurred. Develop a system to be able to trace all food products from its point of origin to the final destination. This was useful in the 1996 apple juice outbreak where distribution patterns of the implicated lots matched with the distribution of human illnesses.. Increase industry awareness of the necessity and importance of use of bar codes or other tracer mechanism for public health surveillance purposes. Develop and enhance community awareness of food-borne outbreaks. Develop creative programs to retrain local food producers in mechanisms to reduce the possibility of a food-borne outbreak. References: 1. Bell BP, Goldoft M, Griffin PM, Davis MA, Gordon DC, Tarr PI, Bartleson CA, Lewis JH, Barrett TJ, Wells JG, Baron R, Kobayashi J "A Multistate Outbreak of Escherichia coli O157:H7 - Associated Bloody Diarrhea and Hemolytic Syndrome From Hamburgers: The Washington State Experience. JAMA 272:1349 November 2, 1994. BACK 2. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, Hebert RJ, Olcott ES, Johnson LM, Hargrett NT, Blake PA, Cohen ML. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N-Engl-J-Med. 1983 Mar 24. 308(12). P 681-5. BACK 3. MacDonald KL, O'Leary MJ, Cohen ML, Norris P, Wells JG, Noll E, Kobayashi JM, Blake PA. "Escherichia coli O157:H7, an Emerging Gastrointestinal Pathogen: Results of a One Year Prospective Population Based Study." JAMA 259:3567 3570 June 24, 1988. BACK 4. Ostroff SM, Griffin PM, Tauxe RV, Shipman LD, Greene KD, Wells JG, Lewis JH, Blake PA, Kobayashi JM. "A Statewide Outbreak of Escherichia coli O157:H7 in Washington State." American Journal of Epidemiology 132:239 1990. BACK 5. Neill MA, Tarr PI, Clausen CR, Christie DL, Hickman RO. Escherichia coli O157:H7 as the predominant pathogen associated with the hemolytic uremic syndrome: a prospective study in the Pacific Northwest. Pediatrics. 1987 Jul. 80(1). P 37-40. BACK 6. Tarr PI, Hickman RO. Hemolytic uremic syndrome epidemiology: a population-based study in King County, Washington, 1971 to 1980.Pediatrics. 1987 Jul. 80(1). P 41-5. BACK 7. Gautom, RK Rapid pulsed-field gel electrophoresis protocol for typing of Escherichia coli O157:H7 and other gram-negative organisms in 1 day. J-Clin-Microbiol. 1997 Nov. 35(11). P 2977-80. BACK 8. Outbreak of Escherichia coli O157:H7 infections associated with drinking unpasteurized commercial apple juice--British Columbia, California, Colorado, and Washington, October 1996. MMWR-Morb-Mortal-Wkly-Rep. 1996 Nov 8. 45(44). P 975. BACK 9. Click here for a link to a description of the Abuela project. Note: there is no return link...bookmark this page if you want to return. BACK Readings: Armstrong, Gregory L. "Emerging Foodborne Pathogens: Escherichia coli 0157:H7 as a Model of Entry of a New Pathogen into the Food Supply of the Developed World." Epidemiologic Reviews, Vol. 18, No. 1, 1996, pp.29-51. Return to Title Page HOME  |  List of Lectures