Use of DNA Fingerprinting Helps Track Tainted Meat


When 68-year-old Patricia Pfoutz died in July after a week of diarrhea and kidney failure, her doctors at a Whitehall, Ohio, hospital were mystified.

In August, public-health detectives at the Centers for Disease Control and Prevention made a startling discovery: Mrs. Pfoutz's death was caused by a strain of E. coli bacteria with a DNA fingerprint that matched a strain found 1,300 miles away in a Greeley, Colo., meat plant.

That plant, owned at the time by ConAgra Foods Inc., was the source of 19 million pounds of ground beef recalled in June and July in one of the largest meat recalls in U.S. history.

For epidemiologists, the genetic match was a powerful illustration of the role DNA fingerprinting can play in food safety and public health. Plaintiffs' lawyers, meanwhile, say the technique has become a formidable legal weapon in product-liability cases against food companies.

The number of lawsuits stemming from food-borne illnesses is small, but the implications of using genetic fingerprinting are huge. "From a burden-of-proof standpoint, it's the difference between the Wright brothers' first flight and a trip to the moon," says William Marler, a Seattle attorney representing the Pfoutz family and 20 other people in talks with ConAgra. "If I am able to prove that your food made my client ill, you are at fault and the only question is how much."

The CDC and the Ohio State Department of Health say Mrs. Pfoutz's death was part of the E. coli outbreak originating at the Greeley plant. The plant's owner, Swift & Co., which remains 45%-controlled by ConAgra, denies responsibility for her death.

Before genetic testing, it was hard for public-health officials to determine where an outbreak of bacterial contamination had occurred. Epidemiologists searching for an outbreak's source would rely primarily on dietary questionnaires filled out by sick people and others in an area. Now, with DNA analysis, there is a more objective way -- although not a foolproof one -- to link each sick person directly with a particular food or product.

Mr. Marler used DNA fingerprinting of bacteria in a 2001 suit filed in Minnesota district court in Hennepin County against supermarket retailer Supervalu Inc., one of its franchisees and a meatpacker. According to the complaint, two-year-old Sonja Pearson became seriously ill after eating meatballs made with ground beef bought at a St. Paul Cub Foods market. Mr. Marler says the suit was settled for a multimillion-dollar sum. Supervalu declined to comment.

And in 2001, a jury decided the Finley, Wash., school district should pay $4.75 million to 11 children and their families for an E. coli outbreak traced, in part by using DNA, to a school lunch of tacos made with bad meat. The producer, Northern States Beef, a Swift & Co. plant in Omaha, Neb., settled with the plaintiffs for an undisclosed sum, Mr. Marler says.

In the Greeley case, Swift says there isn't enough evidence linking their meat to Mrs. Pfoutz, the only person in Ohio known to have been infected in the outbreak and the only person who died. "From a scientific perspective, we look for the two elements that would link a product to an illness: the DNA and the epidemiological link," says Jim Herlihy, a Swift spokesman. "It's our understanding that in this [Pfoutz] case, the epidemiological link is missing."

Indeed, there is a remote chance that Mrs. Pfoutz got sick some other way -- say, from another person infected with the same E. coli subtype. But most public-health experts say that possibility is slim. ConAgra has entered into settlement talks with the other parties but not with Mrs. Pfoutz's family.

Food-industry officials and lawyers emphasize that they support any technology that improves food safety. But privately, many question the validity of DNA testing in these cases. And they fret about what the government plans to do with the CDC's database of bacterial DNA information.

"Bacterial genetic fingerprints, unlike human fingerprints, do not show absolute identity." says Mansour Samandpour, a microbiologist for the University of Washington, Seattle, and consultant for ConAgra and other companies. "They are best looked at as the equivalent of partial human fingerprints."

Some epidemiologists disagree. At a U.S. Department of Agriculture meeting last year, Paul Mead, chief of the CDC's outbreak-response and surveillance unit, described a 1998 listeria outbreak traced to a hot-dog plant. Two years later, a second outbreak, involving deli turkey meat, was traced to the same plant. Both were caused by the same bacterial strain. "Plants may have signature strains," Dr. Mead said.

The E. coli outbreak last year came at the height of summer -- "the season of grilling badly," according to Jim Beebe, a microbiologist with Colorado's public-health lab. He tested fecal samples taken from 20 patients in Colorado, and found many contained bacteria sharing the same genetic fingerprint, suggesting a possible common source.

Dr. Beebe posted the identifying pattern on a CDC computer network that links public-health labs in 35 states and soon found six more people who were sick from the same strain. By late July, CDC investigators had found samples from 34 sick people in eight states with bacterial genetic fingerprints that were "indistinguishable" from the bacteria at the Greeley plant. The CDC said the bug was "a unique strain of E. coli" never seen before among the 10,000 DNA isolates in its database.

Nine days after Mrs. Pfoutz was admitted to the hospital, a stool culture showed that she was infected with E. coli. By that time, a major seizure had left her unconscious and unable to breathe on her own. Richard Pfoutz decided to remove his wife from the ventilator that was keeping her alive. Nine hours later, she died.

Will the legal battles lead to safer food? In 2001, an agriculture department analysis of lawsuits stemming from food-borne illnesses says the legal system provides limited incentives for firms to provide safer food, because few people actually sue. The average jury award was only $25,560, in 1998 dollars, and insurance usually covers it.