Investigations into foodborne illness are being radically transformed by whole genome sequencing, which federal officials say is enabling them to identify the source of an outbreak far more quickly and prevent additional cases.
Previously, samples from sick patients were sent to state and federal labs, where disease detectives ran tests to see if the infections were caused by the same bug. When enough matches emerged, typically a dozen or so, epidemiologists interviewed sick people, looking for a common food that was causing the outbreak.
But the testing wasn’t definitive, and linking one case to another took time. “While all of this was going on, more contaminated product was getting out into the public,” said Dr. Steven Musser, deputy director for scientific operations at the U.S. Food and Drug Administration’s Center for Food Safety and Applied Nutrition.
Now, the FDA is building a network of state and federal labs equipped to map out the exact DNA sequence of strains of Listeria, Salmonella and other foodborne pathogens found in sick patients. These sequences are then uploaded to a public database housed at the National Institutes of Health. The technology can not only differentiate a pathogen from multiple related species, but can also show slight mutations within the same strain.
At the same time, the FDA has begun sequencing pathogens found during routine plant inspections and adding those to the database. One benefit of that, they say, is being able to quickly connect patients within an outbreak. Another is the potential to identify the source of an outbreak after just a few patients fall ill, shortening the time it takes to get tainted food off store shelves.
To increase the odds of a match, the FDA wants manufacturers to contribute samples of pathogens found during their own plant inspections. Some contamination is common in food plants. When it is found in the manufacturing facility, but not in food products, companies generally are required only to clean it up without recalling products.
But eliminating pathogens is tough, and convincing companies to offer up potentially incriminating evidence has been a hard sell, according to interviews with public health officials, food manufacturers and experts on recalls.
“That is not something that we’ve solved yet,” said Ruth Timme, an FDA microbiologist who has talked to 10-15 companies over the past year about the benefits of sequencing.
‘You’ve Got ‘Em’
The FDA became convinced of the superiority of the new approach during a 2014 outbreak of salmonella affecting peanut butter made by nSpired Natural Foods of Oregon.
The FDA had just activated a network of state, federal and academic laboratories to do whole genome sequencing, and the agency had also begun sequencing pathogens it collected from swabbing surfaces during factory inspections. All of these codes were uploaded to the database, known as GenomeTrakr.
When people started getting sick, FDA scientists and partners searched GenomeTrakr, looking for matches with inspection samples. They found the DNA of bugs taken from two sick patients were “almost indistinguishable” from salmonella the FDA had found at nSpired Foods, said Dr. Eric Brown, director of FDA’s Division of Microbiology.
The match allowed officials to quickly recall tainted peanut butter. Only six people got sick.
“You catch things far earlier” with sequencing, said Dr. David Lipman, director of the National Center for Biotechnology Information. “It can be two cases. If you see a match, Bam! You’ve got em.”
More Precise Tool
Since the start of GenomeTrakr in 2012, 25,000 genomes from a variety of pathogens have been added to the database, and several state and federal partners, including the U.S. Department of Agriculture and the U.S. Centers for Disease Control and Prevention, have signed on.
The participants agree that sequencing offers huge advantages over the 20-year-old genetic fingerprinting technique used previously, known as PFGE.
David Acheson, a former official at both the USDA and the FDA, who now advises companies on food safety, likens the difference to a witness in a hit and run accident. While PFGE might identify the vehicle as a brown Toyota Corolla, whole genome sequencing provides the license number and even the vehicle identification number.
“They both help identify the culprit,” he said, but one identifies the specific bug implicated.
The CDC began experimenting with whole genome sequencing in the fall of 2013, joining the FDA and USDA in a pilot project to sequence all reported cases of Listeriosis in patients and upload those sequences on the GenomeTrakr database.
For the pilot, CDC compared whole genome sequencing to PFGE, and found that sequencing reduced the number of red herrings – cases that look similar but aren’t.
In the first year of the Listeria project, whole genome sequencing identified 19 Listeria clusters and solved four outbreaks. That compared with identifying 14 dusters and solving one outbreak the year earlier.
For the food industry, the ability to more quickly match bacteria from a patient sample with their plants is a mixed blessing.
FDA’s Musser thinks the technology will ultimately be embraced by industry, which could use sequencing to see if suppliers are bringing contamination into their plants or if bacteria has taken up residence in a corner of the plant.
“These are things industry could just never get to before,” Musser said.
Bernie Steves of Aon Risk Solutions Crisis Management Practice, who advises companies on product recalls, said sequencing is tying illnesses in people to the source “faster than we’ve seen before,” allowing companies to “nip things in the bud.”
Conversely, he said, “it gives regulators another tool to find out where a potential problem is being sourced from.”
That could mean more recalls. According to the Center for Science in the Public Interest, only about 40 percent of reported foodborne disease outbreaks from 2002-2011 were ever solved, letting many manufacturers off the hook.
To allay some of those concerns and still get more samples, the FDA and its partners are working on ways to allow companies to provide blind samples through a third party.
One such program called VoluntaryNet at the University of Georgia would allow companies to provide samples anonymously. For companies, the program could alert them pathogens residing in their plants. If a submitted pathogen starts making people sick, public health officials could alert companies about the problem, and possibly pull food off market shelves more quickly.
FDA statistician Errol Strain said several companies have agreed in principle to supply samples and they are working out details on how to participate anonymously.