New tests that promise to speed up diagnosis of food poisoning pose an unexpected problem: They could make it more difficult to identify dangerous outbreaks like the one that sickened people who ate a variety of Trader Joe's peanut butter this fall.
The new tests could reach medical laboratories as early as next year, an exciting development for patients. They could shave a few days off the time needed to tell whether E. coli, salmonella or other foodborne bacteria caused a patient's illness, allowing faster treatment of sometimes deadly diseases.
The problem: These new tests can't detect crucial differences between different subtypes of bacteria, as today's tests can. And that fingerprint is what states and the federal government use to match sick people to a contaminated food.
"It's like a forensics lab. If somebody says a shot was fired, without the bullet you don't know where it came from," explained E. coli expert Dr. Phillip Tarr of Washington University School of Medicine in St. Louis.
The federal Centers for Disease Control and Prevention expects private labs to rapidly adopt these next-generation tests — and warns that what is progress for individual patients could hamper the nation's efforts to keep food safe. Already, 1 in 6 Americans gets sick from foodborne illness each year, and 3,000 die.
So even before these tests hit the market, the agency is searching for solutions. Unless one is found, the CDC's Dr. John Besser said the tests' unintended consequence could be that ultimately, more people become sick.
"In the past 20 years, there's been a fantastic ability to fingerprint bugs: Is this an organism that's causing multiple infections and we can interdict it? Or is this a once-only event?" added Tarr, the E. coli specialist. "Without that organism in hand, the state can't do it. The government can't do it. You lose the ability to get the evidence."
It all comes down to what's called a bacterial culture — whether labs grow a sample of a patient's bacteria in an old-fashioned petri dish, or skip that step because the new tests don't require it.
Here's the way it works now: Someone with serious diarrhea visits the doctor, who gets a stool sample and sends it to a private testing laboratory. The lab cultures the sample, growing larger batches of any lurking bacteria to identify what's there. If disease-causing germs such as E. coli O157 or salmonella are found, they may be sent on to a public health laboratory for more sophisticated analysis to uncover their unique DNA patterns — their fingerprints.
Those fingerprints are posted to a national database, called PulseNet, that the CDC and state health officials use to look for food poisoning trends.
There are lots of garden-variety cases of salmonella every year, from runny eggs to a picnic lunch that sat out too long. But if a few people in, say, Baltimore have salmonella with the same molecular signature as some sick people in Cleveland, it's time to investigate, because scientists might be able narrow the outbreak to a particular food or company.
But culture-based testing takes time — as long as two to four days after the sample reaches the lab, which makes for a long wait if you're a sick patient.
What's in the pipeline? Tests that could detect many kinds of germs simultaneously instead of hunting one at a time — and within hours of reaching the lab. Those tests essentially work by searching for an identifying piece of a germ's DNA without first having to grow a culture.
This isn't just a science debate, said Shari Shea, food safety director at the Association of Public Health Laboratories.
If you were the patient, "you'd want to know how you got sick," she said.
PulseNet has greatly improved the ability of regulators and the food industry to solve those mysteries since it was launched in the mid-1990s, helping to spot major nationwide outbreaks in ground beef, spinach, eggs and cantaloupe in recent years. Just this fall, PulseNet matched 42 different salmonella illnesses in 20 different states that were eventually traced to a variety of Trader Joe's peanut butter.
Food and Drug Administration officials who later visited the plant where the peanut butter was made found salmonella contamination all over the facility, with several of the plant samples matching the fingerprint of the salmonella that made people sick. That New Mexico-based company, Sunland Inc., recalled hundreds of products that were shipped to large retailers all over the country, including Target, Safeway and other large grocery chains.
The source of those illnesses probably would have remained a mystery without the national database, since there weren't very many illnesses in any individual state. And these types of nationwide outbreaks are only increasing as global food companies ship farther and faster all the time.
To ensure that kind of crucial detective work isn't lost, the CDC is asking the medical community to send samples to labs to be cultured even when they perform a new, non-culture test.
But it's not clear who would pay for that extra step. Private labs only can perform the tests that a doctor orders, noted Dr. Jay M. Lieberman of Quest Diagnostics, one of the country's largest testing labs.
A few first-generation non-culture tests are already available. When private labs in Wisconsin use them, they frequently ship leftover samples to the state lab, which grows the bacteria itself. But as more private labs switch over after the next-generation rapid tests arrive, the Wisconsin State Laboratory of Hygiene will be hard-pressed to keep up with that extra work before it can do its main job — fingerprinting the bugs, said deputy director Dr. Dave Warshauer.
Stay tuned: Research is beginning to look for solutions that one day might allow rapid and in-depth looks at food poisoning causes in the same test.
"As molecular techniques evolve, you may be able to get the information you want from non-culture techniques," Lieberman said.
Follow Mary Clare Jalonick on Twitter at http://twitter.com/mcjalonick