In the fall of 2006, health officials in Minnesota and Vermont reported multiple cases of Salmonella Typhimurium poisoning with matching genetic patterns.  Working with federal officials, they began looking for a source.

A month later, their investigations had focused on fresh tomatoes, sliced and consumed in a variety of restaurants. And the tomatoes were, in turn, traced to a packing house in Ohio. Case closed.

As outbreaks go, that outbreak was not terribly unusual.  About 190 people were known to have been sickened across 21 states, mostly in the upper Midwest and New England.  Twenty-four were hospitalized.  (The actual number of sick people could have been as many as 6,000, the CDC says.)

But that outbreak has continued to draw attention from epidemiologists, most recently with a research paper published by officials at the Centers for Disease Control and Prevention.  And it has drawn attention precisely because it was not unusual.

In October, 2006, health officials had little to go on – clusters of sick people across several states with matching genetic analysis, but no obvious cause.  The response was what researchers call a “dynamic, iterative process of hypothesis generation” – intensive interviews with sickened people, asking what they had eaten in the week before their illness.  The interviews touched on more than 300 possible food items. 

The objective is to come up with a “hypothetical” cause, a single food item that had been consumed by all or most of the cases.

Initially, the results merely suggested that the sick people had all eaten in restaurants – but not at the same restaurant. In fact, the eateries ranged from a fast-food burger joint to a gourmet delicatessen. So interviewers called the victims back, asking specifically what they ate.

Tomatoes. Big, round tomatoes, sliced for sandwiches or salads.

Armed with that hypothesis, officials questioned managers at the restaurants, and learned their tomatoes all came from a single packing house in Ohio. Alas, the packer had closed down for the winter, and the suspect fields had been planted with other crops. There was no recall. The epidemiology ended in the fields of Ohio.

This, researchers point out, underscores one of the obstacles to tracking bacteria in tomatoes and other produce. Tomatoes have a shelf life of about a month; so, by the time health investigators focus on a potential culprit, the offending food is likely to have been consumed or discarded. That means there is no opportunity for genetic testing of the suspect food.

But the problem remains, and people continue to get sick.

There was a time when Salmonella was popularly identified mostly with uncooked chicken or other poultry.   Over the past two decades, however, outbreaks have been linked to virtually every uncooked food imaginable, and tomatoes have become – and probably always were – a popular vehicle for the bacterium.

Americans eat about 5 billion pounds of fresh tomatoes per year – about 16 pounds per person. (A similar amount is cooked.) The CDC reports three or four outbreaks per year linked to tomatoes. So the odds of getting sickened are quite remote.

But not as remote as they should be, researchers say. Fresh tomatoes  can be contaminated by domestic or wild animals in the fields or in the packing houses, or by unclean irrigation water. And, once contaminated, they serve as “an efficient medium for bacterial amplification,” the CDC report warns.

Food researchers are trying to understand the problem better. A Florida study suggests for instance that oblong-shaped Roma tomatoes appear to be more susceptible to bacteria than round Romas. Another study says green tomatoes are less friendly to bugs. But nobody knows why.

Still another study seemed to suggest that tomatoes coated with a harmless bacterium were less likely to be contaminated with Salmonella – possibly because the Salmonella didn’t take well to competition.

The 2006 outbreak underscored the fact that tomatoes, like other uncooked food, can be hazardous, researchers said.  And  the hazards are not limited to industrial farms in Florida or California; Salmonella contamination is just as likely to emerge from a small farm in Virginia or Ohio.  “These outbreaks suggest that the problem is not localized to one particular region,” they report.

Contamination, they add, is “probably occurring early in the distribution chain, such as at the farm or packing house, rather than at multiple restaurants.”