“As of Aug. 3, CDC has received reports that 78 cases from 26 states including one death associated with this outbreak ofSalmonellaHeidelberg since March 1, 2011,” he said. “The reported dates of illness onsets are March 9 through July 2. It’s possible there are other cases that have not yet been reported.”
Among the patients from whom CDC has information, 22 were hospitalized, which is higher than hospitalization rates typically seen withSalmonellainfections. “We’re working to explore why this may be,” he said. “One possible reason is the outbreak strain is resistant to several antibiotics. This antibiotic resistance may increase the risk of hospitalization and sometimes lead to treatment failure, which is when a patient doesn’t respond to treatment.”
“I want to note the samples do respond to several other common antibiotics used in clinical practice,” he added.
Braden explained how this complex investigation evolved. “It was a slowly building outbreak in the beginning,” he said. “We had a few cases coming into our PulseNet surveillance system, but along the lines of the baseline of cases we might expect. Over time it became clear there was an unusual clustering of cases. We began investigating that clustering on May 23. These types of investigations can be tricky and initially the interviews with patients didn’t support a link with ground turkey. They did not see a significant proportion of cases reporting they had eaten ground turkey.”
At the same time as some of these initial cases were being reported, four ground-turkey packages purchased at four retail stores as part of routine surveillance in the National Antimicrobial Resistance Monitoring Systems (NARMS) found the same strain ofSalmonellaHeidelberg.
“These positive samples came in one per month, beginning in April through July,” Braden said. “This spurred a more-intensive questioning of cases about ground turkey and other possible exposures. This information eventually did lead to an association with ground turkey.”
Meanwhile, investigators found information on ground-turkey purchases – both the brand and where people bought the ground turkey – from three patients who had become ill from this strain. “Investigators used shopper-card information to trace back these samples to the Cargill plant in Arkansas,” Braden said. “Around July 20, state investigators found a sample of open turkey in one patient’s home and on the July 29 got a positive test report for the outbreak strain.”
David Goldman, M.D., assistant administrator, Office of Public Health Science, FSIS, said FSIS was able to develop a link between some of the reported illnesses in the investigation and the Cargill Springdale, Ark. plant on the evening of Aug. 3. “After discussing the findings with the company, the plant agreed to implement a recall and FSIS issued a recall notice, which we hope will have the affect of preventing additional illnesses,” Goldman said.
He then focused on the traceback investigation, which was principally done by FSIS. “Although, I should note that often state officials assist us with the beginnings of a traceback investigation,” he added. “FSIS has also benefitted from having embedded a senior epidemiologist at CDC in the outbreak branch so we had a very close working relationship and liaison with CDC for investigations of clusters as they occur.”
Traceback is often a key in investigating and in making final determinations about the link between cases and product, Goldman said. It is a very painstaking process. “Our folks at FSIS are tireless at their efforts to obtain info that will link illness to a particular manufacturing plant,” he added. “There was a particular focus on three cases. FSIS began its traceback activities on two of the cases from the same state on July 18...within two or three days, respectively, we were able to confirm the purchases were related to this Cargill plant that was the subject of the recall.”
The traceback investigation into the third case began on July 26 and a similar verification of the purchase of the product was made two days later. “We depend on the state health officials and CDC, as well, to ... confirm the food history reveals this particular exposure as the cause of illness so we felt we had that info in all three of these cases. I also want to emphasize in each of the three cases, shopper-card information was very helpful to us in helping make these determinations.
“Traceback is often helpful at deterring the scope of a proposed recall,” he continued. “If we can find the purchase links back to a particular production date, which is always our effort, then we can determine which days of production in a plant ought to be subject to recall. That can often be a difficult process. In this particular case, there were, we believed at the time based on three cases, multiple days of production that may have contributed to contaminated product.
“On Aug. 3, FSIS took the facts we had available to us at the time based on the information from epidemiology and the traceback to the company and when we provided those facts to them they came back to us in a matter of hours and agreed to the recall you saw posted [the evening of Aug. 3],” he said.
When asked if FSIS inspectors found anything unusual at the plant prior to the recall, Goldman replied, “We are still in the midst of investigating and analyzing the data from the plant. Our processing plants are subject to periodic testing by FSIS. This plant would have been tested in that same way.”
In the weeks ahead, Goldman said FSIS will continue to work with CDC and other state officials to complete a thorough and aggressive investigation into this outbreak to determine if there is any improvement that can be made in the food-safety standards or in recall processes.