The Food Safety Net:Too Many Holes?

Graphic: Cathleen Heard It's what the government doesn't know about food that can kill you, says a federal science-policy analyst. "We do rely on Centers for Disease Control [and Prevention (CDC)] data, and the new numbers on foodborne illness are certainly something we didn't want to see," the analyst says. She is referring to a disturbing CDC report issued in September1 that ups the ante for foodborne diseases to an estimated 76 million incidents per year--twice the number reported in a widely cited 1993 Council of Agricultural Science and Technology survey.2 The CDC report, which is based on proactive monitoring programs and better detection methods, also cites 325,000 hospitalizations and 5,000 food-related deaths in the United States per year. "A large proportion of what CDC has found is attributable to unknown causes," according to the analyst, who is affiliated with a large government watchdog agency and who asked to remain anonymous. "I don't think there's a large proportion of pathogens out there we're not aware of. But a lot of these illnesses are caused by things we haven't been able to investigate properly." This chilling news has to be tempered by CDC's own cautionary statement: "Overall, foodborne diseases appear to cause more illnesses and fewer deaths than previously estimated." Three pathogens, Salmonella, Listeria, and Toxoplasma, cause 1,500 deaths a year, while unknown agents account for a remaining 62 million illnesses, 265,000 hospitalizations, and 3,200 deaths. The illnesses are caused by a variety of agents, including viruses, bacteria, parasites, toxins, metals, and prions, and may manifest themselves in illnesses from gastroenteritis to Guilliame-Barre syndrome to hemolytic uremia, which can be fatal. Until recently, estimates have swung wildly--depending on study source and data--from 6 million to 81 million illnesses per year, and up to 9,000 foodborne deaths. The common belief among food scientists today is that food and water supplies are cleaner and safer than they were even a generation ago, much less 50 or 100 years ago, thanks to improved sanitation, tougher meat-packing regulations, and better medicines and lab detection techniques. Yet many find the new CDC data shocking. "We have to be vigilant in this area because the organisms are changing," explains Robert Gravani, professor of food science at Cornell University. "We're finding the organisms more quickly now, but food processing systems are also more complex. For example, processed foods involve temperature variations [that can promote the growth of pathogens]. Our population is aging, and the immune response is reduced. ... We have to take special precautions." Paul S. Mead, a CDC medical epidemiologist and principal author of the report, believes there is no immediate cause for alarm: "As the report states, these new estimates are not comparable to previous estimates and can't be used to establish trends." The rationale for this is that CDC claims it has better data than ever--data so complete, it is not comparable with previous studies. The CDC estimates are drawn from multiple sources, including the cooperative Foodborne Diseases Active Surveillance Network (FoodNet), a joint federal program started in 1996 that proactively monitors 20.5 million Americans for foodborne illness through laboratories and hospitals selected among eight regions of the United States. Other data sources include the Public Health Laboratory Information Systems; the National Notifiable Disease Surveillance Systems; the Gulf States Vibrio Surveillance System; National Health Care Survey; and many others, including separate published studies. CDC asserts that the new study is based on "better information and new analyses rather than changes in disease frequency over time." Mead continues, "There's a separate question as to whether pathogens are becoming more common or whether we're better at identifying them. My gut feeling is that some are becoming more common, some less common. Listeria, for example, is becoming less common. But there is still a significant problem with foodborne illness in the [United States]. Our findings were surprising to some people because a significant proportion of foodborne disease is coming from viral agents and unidentified pathogens. In addition, new agents are now being recognized." Recent examples of new-generation pathogens include Campylobacter jejuni, Escherichia coli 0157:H7, Listeria monocytogenes, Cyclospora cayetanensi--none recognized as causes of foodborne diseases just 20 years ago. Mead cautions that the vast majority of the 76 million illnesses CDC cites are not serious, with most amounting to bouts of diarrhea and vomiting that pass within a day or so. "The range of illness is from acute diarrhea to cholera," he maintains. "Many different types of foods cause illness; and the issue of what proportion of illness is based on which types of food is one we hope to address in the future." Contaminated food can include anything from Salmonella-infected eggs that are "pooled" (combined) in huge restaurant bowls to raw beef or poultry and organically grown produce infected with E. coli. "There is evidence," Mead continues, "to suggest that E. coli 0157:H7 is more prevalent than it was 50 years ago." Indeed, a preliminary U.S. Department of Agriculture (USDA) report in November cited as much as a 50 percent prevalence rate of E. coli 0157:H7 in U.S. cattle--a figure based on more sensitive detection methods and one set in stark contrast to the estimated 1 to 3 percent prevalence rate USDA had reported previously. "We've used this new, more sensitive test since this past summer, and we've seen an increase in the number of positive tests and also the size of the sample," explains Katherine Wotecki, USDA's undersecretary for food safety. Wotecki says the apparent increase in incidence is based on better identification, not an increase in pathogens. USDA is developing new testing and cleansing measures in the slaughterhouse, including irradiation of meats. Meanwhile, E. coli 0157:H7 has also been found in produce, other meats, milk, in fruit juice, and even in daycare settings, transmitted to children through water. "Any number of these organisms can be transmitted via any number of routes," including water and person to person, Mead remarks. "We need to identify gaps in our knowledge and define the actual frequency of these diseases." Tracking Complexity CDC'S report is "the best scientifically presented data ever in terms of foodborne illness" in the United States, says Michael Doyle, pictured here with research coordinator Tong Zhao at the University of Georgia's Center for Food Safety and Quality Enhancement. Because the complexity of food systems is growing dramatically, epidemiologists today must often track separate outbreaks worldwide and find the causes of illness in seemingly isolated cases. Many scientists believe that globalization of trade, the ease of international travel, and changes in American lifestyle--specifically eating out more--are also affecting the food safety picture. "Certainly things are no worse and may be better than they were before," says Kathryn Boor, an assistant professor of food science at Cornell University. "But many factors in food processing systems, such as the scale of mass production in food plants, can affect large numbers of people." The virulence of certain types of pathogens is also grabbing media and consumer attention, adds Gravani: "A relatively small number of [pathogenic] organisms ... can get people mighty sick." Gravani and Boor believe that food science this past decade has developed dramatically improved surveillance techniques; one is proactive regional monitoring. "We now have genetic and microbial tools to identify whether clusters of illness in one part of the country are related to outbreaks in another part of the country," Boor said. FoodNet, a cooperative project of the Food and Drug Administration, CDC, and the USDA, has an active surveillance system in which health departments routinely contact laboratories around the country to monitor reports of pathogen incidence (e.g., the Port of New York, Oregon, parts of California, Minnesota, Georgia, and Massachusetts, among other regions). Before FoodNet, surveys of food-related illnesses were essentially "passive," more random, and in some instances, anecdotal. Today scientists use molecular fingerprinting and a system called PulseNet to link a variety of states' data, according to Boor's colleague, Martin Weidmann, an assistant professor of food science at Cornell. "PulseNet can detect outbreaks in one state and relate the pathogens to those identified in another state," he explains. "CDC has led in these systems because [it] can recognize illness caused by the same DNA fingerprint." More sophisticated lab techniques have turned up new pathogens--a prime example is Campylobacter bacteria. "Five or six years ago, it was barely detectable," Weidmann says. "Today we've got better cultural media, and Campylobacter has become the most increased bacteria in terms of identification. It is now the most highly reported foodborne disease." An Increase in Chutzpah? Despite surveillance progress (or indeed, because of it), CDC has drawn fire for some of its boldest assumptions. One critic of the new report is Dean Cliver, professor of food safety at University of California, Davis, an expert in hepatitis A and mad cow disease, among several others. "As far as the new set of CDC estimates goes, they can't prove their numbers or assumptions are correct," he asserts. "What we have here is a great increase in chutzpah." Cliver believes that CDC's statistical methodology is suspect because "for every agent listed in [the report] table, a different set of assumptions was necessary. We [meaning CDC] are now willing to make assumptions and draw conclusions that are, in many cases, not supported by actual data. The 76 million disease figure makes your hair stand on end. When CDC tabulates illnesses of known cause, they run a grand total of 38.6 million illnesses of the kinds that are known to be foodborne sometimes, of which 13.8 million are probably foodborne. That means that one in three of these illnesses is [likely to be] foodborne. But CDC took another multiplier for how many illness are out there for which the causes are not known, and came up with 76 million. I'm not terribly comfortable with that. Whoever guesses the highest gets the most ink." For certain pathogens such as Salmonella, Cliver continues, "the proportion of foodborne illness out of the total incidence rate is going to be relatively high, about 98 percent. But for other illnesses, no one knows what proportion is foodborne." For example, CDC lists Norwalk-like viruses (NLVs), which cause nausea, vomiting, and diarrhea, at an estimated 23 million cases every year. "Of these, they estimate that nine million cases of [NLVs] are foodborne. But they have no data to support that number. That's the weakness of the study--for every agent a different set of assumptions. Whereas some [suspected pathogens] will go to a laboratory for confirmation, something that's highly prevalent [such as Norwalk] doesn't go to the lab because the clinical outcome isn't terribly serious. In this age of medicine governed by HMOs, the only way a physician can send samples to the lab is that it has to affect the [clinical] outcome of the patient. In the case of most [NLVs and other suspected pathogens], the patients will almost certainly get better whether physicians do anything or not." Cliver's greater concern is the apparent serendipity of government-funded pathogen tracking and reporting. "For example, the FoodNet program includes [tracking] no viruses whatever; they've targeted certain disease agents--actually seven bacterial and two parasitic foodborne diseases [within the monitored population]. But if FoodNet doesn't ask [about other pathogens and viruses], no one will tell them, and they've chosen not to ask." Mead explains that FoodNet was established in 1996 "to deal with some very specific issues related to pathogen-control strategies of direct concern to USDA [which monitors beef and poultry] and FDA [which monitors fish and produce]. The agencies were specifically interested in regulatory issues for certain pathogens for which they're trying to initiate control strategies, such as Salmonella [and] E. coli." Mead says that CDC estimates of viruses are important and are based on published data and the opinions of experts, although laboratory data is, in fact, lacking. "We know there are outbreaks of foodborne illnesses due to [NLVs]," he continues. "The counter question would be how could we not [make an estimate]? Not to estimate is to say that [NLVs] don't create these outbreaks, and they do. Even if our estimates are off considerably, the data suggest that [NLV] is an important foodborne pathogen and should be looked at to try to establish surveillance." The Best Estimates Yet Defenders of the CDC report say that the authors have drawn from a wide variety of survey data sources on viruses, parasitic infections, and bacteria. Further, the data interpretation is rich in detail and does not shirk from asserting that foodborne illnesses, hospitalizations, and deaths have been underreported. Many reputable scientists staunchly defend the CDC report. "This is the best we've got," declares Michael Doyle, director of the Center for Food Safety and Quality Enhancement, University of Georgia. "The report is the best scientifically presented data ever in terms of estimating foodborne illness in the [United States]. In the past, much of the data used were based on past outbreaks, retrospective studies that were done following outbreaks, and some of these estimates were best guesses. But with FoodNet data that CDC has used, at least for those organisms under surveillance, those data are much harder than what they used in the past." Doyle believes that even the estimates of foodborne illnesses due to unknown causes "look reasonable." In addition, some foodborne pathogens and associated human illnesses have actually decreased. For example, the incidence of human infection caused by Salmonella has gone down significantly since the three-year period that CDC has done trends analysis, and the USDA has implemented a safety process known as Hazard Analysis Critical Control Point. This is a process by which scientists look at all the hazards that might occur to a food and then apply a systematic approach to controlling the hazards--determining where best to intervene in the farm-to-table process. In the case of ground meats, for example, cooking to 160 degrees Fahrenheit for at least one minute can destroy E. coli and other pathogens. (USDA recommends a thermometer; meat color is not a good indicator of temperature.) "Our own data on Salmonella [pathogens show] a significant decrease," reports Wotecki. "We now have two good data sets showing that Salmonella levels in meat and poultry products and human illnesses are going down." That's good news. The question now becomes what scientists, government officials, and food safety experts will do with the expansive data cited in the CDC report. At USDA, for example, officials must support a "zero tolerance" policy for contaminated hamburger, which means that ground meat from any infected animal cannot be sold. The apparent increases in E. coli contamination in cattle will almost surely result in stronger "early testing" methodologies to determine contaminated carcasses, as well as new "steam pasteurization," vacuuming, and irradiation techniques, Wotecki says. Similarly, the FDA is supporting the president's Food Safety Initiative and has been allotted millions of dollars to conduct surveys and inspections of food establishments. "We're also working on the presidential directive on imported food safety," reports an FDA policy analyst. "We're trying to prevent 'port shopping,'" a term that applies to food importers that may try to introduce contaminated products into a different U.S. port once customs officials have identified a contaminated shipment and denied entry here. Among food scientists themselves, a large burden goes beyond the laboratory to food safety education--both for consumers and selected patient populations. For example, since populations of the elderly and immunocompromised are increasing (e.g., those with AIDS, cancer, and other illnesses), surveillance and a total "systems approach" to preventing contamination are becoming important, according to Linda Harris, an extension specialist in microbial safety at UC-Davis. "I educate on all aspects of the food chain--from food industry to farmers to consumers. For consumers, if you look on the Internet, www.fightback.org, there's a partnership for food safety education. Basically we've come up with simple slogans that are easy to remember and focus on problem areas we presume are in the kitchen. Though it's hard to measure the impact on reducing foodborne illnesses, it shouldn't stop us from pursuing that angle." Bottom line: The data scientists are discovering must be translated into a strong medical consumer education process. "We do believe that a lot of foodborne illness is due to mistakes made in the home, at the restaurant level, and the food service," Gravani says. "We need [pathogen detection], monitoring of processing and distribution, and education for consumers. People aren't doing the right things in their own homes. Start with cooking food thoroughly and not mixing raw meat products on a plate with cooked products. Mundane handwashing. Improper composting. We need to be more vigilant all the way around." Arielle Emmett (ArielleEm@aol.com) is a freelance writer in Wallingford, Pa. References 1. P.S. Mead et al., "Food-related illness and death in the United States," Emerging Infectious Diseases, 5:607-25, September-October 1999. 2. Foodborne Pathogens: Risks and Consequences, Ames, Iowa, Council of Agricultural Science and Technology, 1994. For More InformationGateway to Government Food Safety Informationwww.foodsafety.com

The Food Safety Net:Too Many Holes?

Interested in reading more?

Become a Member of