As health authorities raced to find the source of the unprecedented E. coli epidemic sweeping through Germany this past spring, epidemiology was faulted when the first case-control study erroneously pointed to cucumbers, tomatoes and lettuce as possible suspects.

But it was microbiology, not epidemiology, which implicated the wrong vegetable. After lab tests found pathogenic E. coli on four cucumbers, three of them from Spain, it took days to determine that the contaminated Spanish cucumbers did not carry E. coli O104:H4, the unusual outbreak strain.

In the end it was epidemiology, not tell-tale DNA evidence, that showed the probable source of more than 4,300 illnesses and at least 55 deaths was organic sprouts grown from tainted fenugreek seeds  — sprouts that in many cases had been served in salads along with cucumbers, tomatoes and lettuce.

The difficulty in picking out the correct salad ingredient arose because so few patients had mentioned eating sprouts, researchers at the Robert Koch Institute recount in the New England Journal of Medicine. 

“The one dish that frequently exposed guests to sprouts was the side salad, which contained tomatoes, cucumbers, three sorts of leaf salads and sprouts,” the researchers wrote in an outbreak profile published Wednesday. “Sprouts may have been the ingredient that visitors recalled least in such a mixed salad.”

Using menu cards to spur restaurant patrons’ memories and to help investigators identify ingredients of various dishes eventually helped crack the case. After June 10, when health authorities warned the German public to not eat raw sprouts, the number of illnesses “dropped substantially,” according to the NEJM profile.

“Although international guidelines generally recommend otherwise, this experience suggests that food items or ingredients that are deemed to be hard to remember should be included in analytical studies, even if they are mentioned by less than 50% of those surveyed,” the researchers suggest.

Back in May and June, Europe’s deadly E. coli outbreak may have seemed a distant threat to the United States – until the spread of Listeria on contaminated cantaloupes this fall. Listeria, a common contaminant in deli meats, had never been detected in cantaloupes before.

In an accompanying NEJM editorial, Dr. Martin Blaser observes that “infectious disease epidemics constantly arise, usually involving familiar pathogens but with combinations of known and unknown virulence factors or in new vehicles, causing novel outbreaks and clinical consequences.” 

The surprise in Germany’s massive outbreak was the uncommon E. coli O104:H4 strain, which Dr. Blaser describes as “well-armed for mayhem.” And while all outbreaks are different, the NEJM profile details how the epidemic in Europe “exemplifies the threat posed by foodborne pathogens with their propensity to cause large common-source outbreaks.”

Other insights from the profile:

— The outbreak lasted roughly from May 1 to July 4, growing “dramatically” from May 8.

— The estimated incubation period – the time between exposure to the pathogen and the initial symptoms – was 8 days, compared with the typical 3- to 4-day incubation for E. coli O157:H7.

— Both children and adults affected experienced painful abdominal cramps, but adults were more likely to have bloody diarrhea. Vomiting occurred more often in children than adults.

— About 22 percent of those stricken developed hemolytic uremic syndrome (HUS), the potentially fatal kidney disorder associated with E. coli infections. In E. coli O157:H7 infections, the HUS rate is typically 6 percent.

— Among patients who died, death occurred a median of 10 days after the onset of the disease.

— HUS predominantly affects very young children, but 88 percent of the case patients in the German outbreak were adults. This could reflect food choice — the implicated sprouts — or the lack of previous immunity to a novel bug.

— Among women, the incidence of HUS peaked in the 30 to 34 age group. Among men, it was in the 25 to 29 age group.

— Even among the 101 pediatric patients with HUS, the median age — 11 — was considerably higher than is typically seen in an E. coli outbreak. Only 2 percent of the case patients were younger than 5.

— Women were disproportionately affected, likely because they were more likely to have eaten sprouts. Fifty-eight percent of those infected, and 68 percent of those with HUS, were female. Of those with HUS who died, 75 percent were women.

— Toward the end of the outbreak, when secondary household transmission caused more infections than consumption of sprouts, there was a shift to a more even gender distribution.

— The outbreak strain combined the virulence properties of two different kinds of diarrhea-causing E. coli — enteroaggregative, the common cause of diarrhea in travelers, and Shiga-toxin producing.

— The strain’s host remains unclear, but because typical enteroaggregative E. coli normally is transmitted by people, the fecal contamination of the fenugreek seeds could have been by humans or animals.

— The organic farm that grew the tainted sprout seeds was inspected routinely and used satisfactory hygienic measures. But given that sprouts and sprout seeds are so vulnerable to contamination, the researchers suggest they should be tested for Shiga-toxin producing E. coli.