CIDRAP: Nursing home cases were key in Minnesota Salmonella probe

Robert Roos News Editor

Jan 23, 2009 (CIDRAP News) – A physician's report of diarrhea cases in a nursing home, followed by the discovery of cases in several other institutions, provided the key that enabled Minnesota disease detectives to figure out before anyone else that peanut butter was the culprit in the current nationwide Salmonella outbreak.

When epidemiologists compared food product invoices from institutions that had cases, they found that the one thing the places had in common was King Nut peanut butter, Minnesota officials said. That led to testing of the peanut butter and the discovery of the outbreak strain of Salmonella in it.

"If we didn't have those institutional cases, we still might not know it's peanut butter causing this outbreak," said Kirk Smith, DVM, PhD, director of the foodborne disease unit at the Minnesota Department of Health (MDH) in St. Paul.

The outbreak began back in September but did not become public until early this month. So far there have been 491 cases in 43 states and one more in Canada, the Centers for Disease Control and Prevention (CDC) reported in a teleconference with clinicians today. In addition, there have been seven deaths possibly related to the illnesses.

The outbreak has been traced to a Peanut Corp. of America (PCA) facility in Blakely, Ga., which makes peanut butter distributed by King Nut Cos. and others to institutions and food services. PCA also sells peanut butter and peanut paste from the plant to a number of food companies, which process them into cracker snacks, cookies, ice cream, candy, and other products. The outbreak has triggered a long list of product recalls.

Probe began in November
Minnesota cases in the outbreak began showing up in November, according to Carlota Medus, PhD, MPH, a foodborne disease epidemiologist at the MDH. By late in the month the department had found a couple of cases that featured the outbreak strain, Salmonella enterica serotype Typhimurium, she reported.

Minnesota began interviewing case-patients about what they had eaten, using a hypothesis-generating questionnaire developed by the CDC for this outbreak, said Smith. Besides Smith and Medus, investigators included another MDH foodborne disease epidemiologist, Stephanie Meyer, and Team Diarrhea, a group of seven graduate students in public health.

"After we had our first four cases, there was a lead from another state that the cause might be chicken," he said. "So we followed up on the chicken sources on our cases and the [Minnesota] Department of Agriculture [MDA] traced it back to the plant level, and it just wasn't very compelling."

Meanwhile, the early interviews led to some suspicion that peanut butter could be involved. "We had several cases interviewed, and I think most if not all of them had eaten peanut butter, but we were having trouble with brand names, and certain aspects of it were not clear," said Medus.

Big break in the case
By Dec 22 Minnesota had nine cases, but no good evidence as to the cause. But then came the big break in the hunt.

"On Dec 22 we received a call from a clinician who reported an outbreak of Salmonella in a nursing home and that the lab that did the testing had noticed additional cases in the community," said Medus. "At that point we didn't know if these were part of the national outbreak. By the next day the first Salmonella isolates had arrived at our lab, and we were able to confirm that at least one of them was part of the national outbreak." The nursing home was in Brainerd in north-central Minnesota.

Clinical labs routinely send Salmonella isolates to the MDH for typing, but it usually takes several days before they arrive. In this case the samples were sent immediately. "Because of that report, we didn't have to wait for the Salmonella isolates to come in, which would've been a few days down the road," Smith explained.

The physician who reported the nursing home cases didn't know there was a widespread Salmonella outbreak going on, but reported them because they were in a nursing home, according to Medus.

Homing in on peanut butter
The discovery of outbreak cases at one nursing home was significant but not sufficient to implicate peanut butter, said Medus. Shortly afterward, however, cases started showing up at other institutions, including nursing homes, colleges, and an elementary school, according to Smith.

She said that development prompted the next investigative step: collecting menus from the various institutions and, with the help of the MDA, gathering invoices for the food products used. "We went to the invoices and started comparing them, and the one thing they had in common was the King Nut [peanut butter]," she said. In checking the documents, the investigators looked at all foods, not just peanut butter, she noted.

Smith said the regional pattern of the institutional cases offered another important clue: "All the institutions served King Nut peanut butter, and what was interesting was that the institutions were all in northern and northwestern Minnesota. And we found that the King Nut peanut butter they were getting was from a distributor out of Fargo, Sysco.

"What was intriguing to us was that there is a Sysco distributor in the Twin Cities, and they did not distribute King Nut peanut butter, and we were not seeing [institutional] cases in the catchment area for that distribution. So it was all kind of coming together."

With the peanut butter as a suspected source, MDA and MDH labs began testing peanut butter collected from the institutions. Eventually an opened 5-pound tub of peanut butter from the Brainerd nursing home yielded the outbreak strain—or, more precisely, two strains.

Smith explained that the outbreak involves two very closely related Salmonella strains as identified by pulsed-field gel electrophoresis. "We found both strains in the open jar of peanut butter," he said.

Going public with findings
The Minnesota agencies announced the discovery of Salmonella in the peanut butter on Jan 9, and, after further testing, reported on Jan 12 that it matched the outbreak strain. King Nut issued a recall on Jan 10.

The US Food and Drug Administration (FDA), however, did not publicly tie PCA-made products to the outbreak immediately. Because the Salmonella-tainted container had been opened, there was a possibility it had been contaminated at the nursing home, rather than at the plant where it was made. It was not until Jan 19, when Connecticut officials reported finding the outbreak pathogen in an unopened container of King Nut peanut butter, that the FDA publicly linked the outbreak to the PCA facility.

With an opened container, "there's always the concern that somebody could've gotten sick from some other source and then contaminated that tub," said Smith. However, he said that looked unlikely in this case, because Salmonella was found in samples taken from a number of locations within the tub. Moreover, both outbreak strains were found, whereas if the product had been contaminated by an infected person at the nursing home, probably only one strain would have been present.

Also, Smith said that as Minnesota officials were investigating the institutional cases, they were hearing from other states that they also had cases at institutions that had received King Nut peanut butter. "It wasn't just the isolate, it was all the institutions receiving this one brand," he said.

Other products involved
But it also became clear during the investigation that not all cases were related to King Nut peanut butter. Smith said that no institutional cases were found in the Twin Cities area, but there were other cases that couldn't be explained by the peanut butter. "Then other health departments around the country said their cases had been eating peanut butter crackers," he added. That led to a CDC case-control study that tied some cases to peanut butter–containing products such as crackers.

Minnesota investigators have been able to connect almost all the state's cases to particular food products, said Medus. "We have fewer than 5 cases that we can't explain yet, out of 36 cases," she said. And, said Smith, more than half of those have been linked to products other than King Nut peanut butter.

Untraditional epidemiology
Smith and Medus agree that the institutional cases were critical in finding the food culprit.

"If we didn't have the institutional cases, it would've been a lot harder to implicate a product. The noninstitutional cases ate a lot of different products," said Medus. With tens of different products and hundreds of cases, "You won't have a single product common to all, which makes it extremely difficult to figure out a product [source]," she said.

Smith commented, "This was a little untraditional in that it wasn't cracked by looking at the interview data but by noticing the institutional link and getting the menus and getting our Department of Ag to compare and contrast."