Late Friday evening, the US Food and Drug Administration (FDA) updated the report regarding the agency’s ongoing investigation of an ongoing 20-state outbreak of E. coli O121 linked to Gold Medal, Wondra, and Signature Kitchen-brand flour processed at a single General Mills facility during November 2015.
The updated release states that the outbreak, which at the time it was first announced was linked to General Mills flour by strong circumstantial evidence collected during the course of the investigation, has now also been confirmed by finding the outbreak strain of E. coli in recalled flour collected from the home of an individual who contracted the outbreak strain.
Challenges of Source Identification in the E. coli O121 ”Flour Outbreak”
It’s surprisingly hard to remember everything that you ate yesterday and the day before, much less a week ago. If you’re skeptical, try writing down everything you ate during the past work week, starting with breakfast on Monday.
Anatomy of an Epidemiological Investigation (Or, Source Identification: Achieving the Impossible)
In an outbreak investigation, one of the investigators’ jobs is to do just that: discern everything that an outbreak case ate in the week or so before they became sick. If a certain food item or restaurant appears in a greater than expected number of responses, investigators have found a solid lead in their search for the source of the outbreak.
The job is deceptively difficult. Ask someone what they ate yesterday, you will in all likelihood get a response something along the lines of “cereal for breakfast; a ham sandwich and a cupcake for lunch; pasta with marinara for dinner.” The chance of receiving a detailed response – “Honey nut cheerios with Borden skim milk, a sandwich on Sunbeam white bread with Oscar Meyer Deli Fresh Smoked Ham, Kraft pre-sliced Swiss cheese, Heinz mustard, and Dole Romaine lettuce with a homemade cupcake for dessert…” – is exceedingly low.
In fact, even if one or two people mentioned a “homemade” item, it is highly unlikely that enough people would do so to spur further investigation into the possibility that the source of the outbreak is some component of various dissimilar homemade goods (to the surprise of this writer, people do use flour to make things other than cupcakes, including lasagna, breads, sauces, etc.).
Investigators Struggle to Find Answers in Initial Investigation
The FDA, US Centers for Disease Control and Prevention (CDC), and state and local officials began investigating the outbreak in February, approximately two months after the date of onset for the first outbreak illness (December 21, 2015). Initial epidemiological investigations by state officials and the CDC seemingly eliminated the most common sources involved in E. coli outbreaks (e.g. ground beef, leafy greens), but provided no real affirmative insight into likely causes.
Stumped, investigators circled back and questioned the individuals sickened in the outbreak a second time, this time using more general questions about what and where the person ate in the days leading up to the onset of their symptoms. Investigators first recognized that a number of the outbreak cases had some exposure to raw dough once they began using the less pointed questions.n Some of the outbreak victims had tasted or played with dough while baking at home, and others reported handling, playing with or even eating raw dough at restaurants.
Traceback From Restaurant Cases Key to Isolating Source
Once investigators isolated flour as a likely cause of the outbreak, they began efforts to pin down the identity of the flour with as much specificity as possible (i.e. brands, production dates, lot codes, etc.). Many of the case patients who were exposed at home were unable to identify the brand of flour they had used – much less more specific identifying information – because they no longer had the packaging.
At that point, information from outbreak victims who contracted the illness while dining out (whether by simply eating the food served at the restaurant or by playing with raw dough) became essential. The restaurants were able to supply information investigators needed to successfully identify likely source(s), such as brand name information.
Despite the fact the FDA and General Mills, which manufactures all three brands of the recalled flour, had been consistently testing samples of flour and had yet to return a positive result for the outbreak strain of E. coli, officials determined traceback efforts revealing implicated restaurants purchased General Mills flour, when coupled with the fact that over half of the interviewees reported using one of the three recalled brands, constituted enough evidence to publicly link the product to the outbreak.
Source (Re)confirmation Using Genome Sequencing
At the time the CDC and FDA first announced the outbreak, the agencies had incredibly strong circumstantial evidence that certain General Mills flour: the dough used in restaurant locations or that had been eaten or handled by people who developed cases of the outbreak strain of E. coli was made using flour produced in a single week during November 2015 at a single General Mills facility. At that time however, none of the flour samples tested by the FDA or General Mills had tested positive for the outbreak strain.
On June 10th, “FDA whole genome sequencing on E. coli O121 isolates recovered from an open sample of General Mills flour belonging to one of the consumers who was sickened was found to be closely genetically related [to] the clinical isolates from human illnesses. The flour came from a lot that General Mills had recalled,” the updated FDA release stated.
Flour Outbreak: Status Update
According to the CDC, as of May 31st, 38 outbreak cases of E. coli O121 had been confirmed in 20 different US states as follows: Alabama (1 case), Arkansas (1 case), Arizona (2 cases), California (1), Colorado (4), Iowa (1), Illinois (4), Massachusetts (2), Maryland (1), Michigan (4), Minnesota (3), Missouri (1), Montana (1), New York (1), Oklahoma (2), Pennsylvania (2), Texas (2), Virginia (2), Washington (2) and Wisconsin (1).
Ten individuals required hospitalization; however, no cases of Hemolytic Uremic Syndrome (HUS), a type of kidney failure, or deaths have been reported.
Ill persons range from 1 to 95 years of age, and first fell ill on dates ranging from December 21, 2015 through May 3, 2016. Outbreak cases that occurred subsequent to May 5th may not be reflected on the June 1st CDC report due to the amount of time that elapses between the onset of illness and when the confirmed case is reported, which according to the CDC takes an average of 2-3 weeks. The time delay, along with a number of other factors, makes it likely that additional cases will be confirmed in the future.