Gene-Fingerprint Scan Scopes Out Food Poisoning

Three or four years ago, in the Dark Ages of food poisoning investigation, the little cluster of illness in Virginia would hardly have registered even a queasy blip.

Tens of millions of Americans get sick every year from bad food. Usually there is no reason to suspect they share the misfortune of eating the same bad food unless many fall ill in the same place at the same time.

So the call to the federal Centers for Disease Control and Prevention one day last January was remarkable for the numbers involved. The Virginia state health department wanted to talk about just five people with food poisoning.

All five had gotten salmonella, which spreads through food and causes bad diarrhea. On the surface, they had little else in common: All got sick in November and December. All came down with the same strain of the germ, called Salmonella newport. Most of them were Asian or Latino.

Not long ago, official interest would have ended there, assuming it even got that far. But this time, the medical detectives had one more clue to go on.

The genes in bacteria vary from bug to bug, pretty much the way people’s do. So the state lab used a new test to scan the genes in the victims’ salmonella bacteria--a process similar to the genetic fingerprinting used in criminal cases.

Bingo! The germs were genetically identical.

To pick up the very same bacteria, all five must have eaten the very same thing. But what? No one had a guess.

“They wondered if this was part of a nationwide outbreak, so they called us,” says the CDC’s Dr. Sumathi Sivapalasingam.

About 40,000 cases of salmonella poisoning are reported annually in the United States. Checking its records for November and December, the CDC found there were more cases of Salmonella newport than usual.

The CDC turned next to PulseNet, its nationwide food poisoning network. State and city labs test bacteria, just like Virginia did, and send the results to a database at CDC headquarters in Atlanta. There the genetic profiles are compared to see if identical germs turn up in different places.

The CDC e-mailed all the labs in the network: Test your recent salmonella samples. Send the results to the CDC to see if they match the Virginia bacteria.

“By the end of January, we’d received notice of 78 people who were infected with the same pattern of Salmonella newport living in 13 different states from Maine to California,” says Sivapalasingam. “There were two deaths, and all the illnesses occurred between Nov. 13 and Dec. 27.”

The 73 were probably just the tip of the iceberg. The CDC assumes many more actually got sick. However, most cases of food poisoning never come to their attention, since folks usually just ride out the misery at home.

Now they were certain something sold nationwide had caused a salmonella outbreak. It was time to find it. The CDC questioned the victims. What did they eat? How did their food differ from healthy folks’? One item jumped out.

Mangoes.

Now, here was a surprise. Lots of unlikely things can make people sick. But a tropical fruit? “We never had an outbreak from mangoes that we recognized before,” says Dr. Robert V. Tauxe, the CDC’s chief of foodborne diseases.

“Recognized” is the key word. Mangoes may well have made people sick before. But nobody knew. Not enough mango eaters ever got sick in any single place to raise suspicion.

The next step was to learn how the mangoes got tainted. Investigators from the Food and Drug Administration checked where the victims bought their mangoes and followed their trail from grocery stores back through the food distribution system. Eventually they tracked all the bad fruit to a single packing shed on one farm in Brazil.

The problem there turned out to be a treatment intended to kill fruit flies. Mangoes are dipped first in hot water, then in cooler water. This water is supposed to be chlorinated daily and changed weekly. But somehow the process broke down, and salmonella flourished in the under-chlorinated warm water.

The mangoes actually absorbed the germs through their cut stems, so washing at home did not protect against food poisoning. Investigators say the problem has been solved by improving the way the water is chlorinated and checked.

To CDC officials, this story shows the remarkable power of genetic fingerprinting to make the food supply safer.

“With this system, we are detecting outbreaks that would have been very difficult to see before,” says Tauxe. “It’s not that they are tiny. It’s that they are dispersed.”

The idea for PulseNet grew out of the 1993 Jack in the Box outbreak, in which more than 700 people got sick from fast-food hamburgers in the Northwest. For the first time, CDC investigators used genetic fingerprinting to show that the bacteria that made people ill precisely matched the germs found in hamburger patties.

The CDC worked out standardized ways to do the test, called pulsed-field gel electrophoresis. The goal was to make sure labs everywhere get exactly the same results when they test the same bacteria.

The test takes about a day to run. It uses enzymes to cut bacterial genes into various-size chunks, then sorts them by size. The result is a pattern of bands that looks like a grocery store bar code. Computers can read this code to look for bugs with identical patterns.

PulseNet went online in 1995, linking up state public health labs in Massachusetts, Minnesota, Texas and Washington. At the start, the labs checked for just one especially nasty food poisoning bug, E. coli O157:H7, the culprit in the Jack in the Box affair.

Its first big tryout was in 1997. Two weeks after it joined PulseNet, routine fingerprinting at the Colorado state lab turned up eight identical samples of E. coli. Frozen hamburger patties appeared to be the common link, since the meat was also contaminated with E. coli. But were the bugs in the people identical to the bugs in the hamburger?

Genetic fingerprints from the patients and the meat were sent to the CDC. Bala Swaminathan, chief of the CDC lab that oversees PulseNet, remembers the excitement of that moment.

“We were extremely apprehensive about whether it would work,” he says. A microbiologist spent 10 minutes scrutinizing the scans. A perfect match. “That was a turning point in PulseNet. Everyone said, ‘This is something absolutely fantastic we can do.’ ”

The results led to the biggest beef recall ever, 25 million pounds of Hudson Foods meat.

Eventually genetic fingerprinting was worked out for other common food poisoning bugs, including salmonella, shigella and listeria. This has helped spot combinations of foods and bacteria that have never been suspected before.

Two years ago, Minnesota noticed outbreaks of Shigella sonnei infection among people eating in two different restaurants. Humans are the only carriers of this bug. Typically restaurant outbreaks are blamed on cooks who fail to wash their hands after going to the bathroom. But that explanation seemed unlikely this time, since both outbreaks involved genetically identical germs.

The CDC put out an alert on PulseNet. Soon Los Angeles turned up the same bug. Then other labs joined in. Eventually it was found in 10 states and one Canadian province. This time parsley was the cause. The FDA tracked it back to one farm in Mexico.

Without PulseNet, says Swaminathan, this would have been just another unexplained series of outbreaks.

Does this technology truly make the food supply safer?

Tauxe notes that E. coli outbreaks seem to be getting smaller. Maybe they are cut short because PulseNet picks them up early, but the theory is hard to prove.

After all, Tauxe notes, “once an outbreak is identified and controlled, it stops. You’ve interrupted the flow of history.”

Last year, about 10,000 bacterial fingerprints were sent to the CDC. All but four states--Mississippi, Montana, Nevada and North Dakota--are part of PulseNet, and they will link up by the end of next year. Canada joined last year, and other countries will probably be added. Methods are being worked out to test other bacteria as well as viruses that cause food poisoning.

In fact, food poisoning investigators are awash in germ genes, more than they can handle. If the system has a drawback, users say, it is data overload. Suddenly there are outbreaks everywhere, it seems.

“It has a Pandora’s box aspect to it,” says Deborah A. Shea of the Massachusetts state lab.

Because of PulseNet, the CDC has begun giving outbreaks hurricane names, like Hugo and Irene, just to keep them straight, and it cannot investigate them all. A major chore is deciding which are worth chasing.

“It’s like turning on the kitchen light and seeing all the cockroaches scurrying,” said CDC epidemiologist Paul Mead. “We can’t get to them all before they slip into the woodwork.”

*

On the Net:

PulseNet: https://www.cdc.gov/ncidod/dbmd/pulsenet/pulsenet.htm