As of this morning, CDC’s count in the E. coli O26:H11 outbreak linked to Sierra Nevada Cheese Company’s Graziers grass-fed raw milk cheese stands where it stood on September 25: 13 sick in nine states, eight hospitalized, three with hemolytic uremic syndrome, no deaths. Over half of the sick are children 5 and younger. Illnesses began between July 7 and August 26. All nine people interviewed ate Graziers raw milk cheese, against a background rate of 4.9 percent.

The company’s recall notice finally appeared on FDA’s site on September 29. It lists ten products, from 8-ounce squares to 40-pound cheddar blocks sold to wholesalers, distributed nationwide through distributors, retailers and direct customers, and it still names no lot code, date code or production date. If it is Graziers raw milk cheese, it is recalled.

Marler Clark represents the family of a child who developed HUS after eating Graziers cheese, and families in last fall’s Twin Sisters Creamery raw milk cheese outbreak.

The rates are high, and that is the point

Eight of 13 hospitalized is 62 percent. Three HUS of 13 is 23 percent. CDC’s own burden estimates put the hospitalization rate for laboratory-confirmed non-O157 STEC infections at about 13 percent, and for O157 at 46 percent, and CDC’s rule of thumb is that 5 to 10 percent of people diagnosed with a STEC infection develop HUS, with young children at the high end. The three other raw milk cheese E. coli outbreaks of the last 31 months ran 11 sick, five hospitalized and two HUS (Raw Farm, 2024); 10, one and none (Twin Sisters, 2025); and nine, three and one (Raw Farm, 2026). This one is above all of them.

There are two explanations. Either the strain is unusually dangerous, or the count is missing the people who got less sick. Both are probably true, and they point the same direction.

The strain

E. coli O26:H11 is not one bug. The best-known O26 outbreak in this country, the Chipotle outbreak of 2015, put 21 of 55 people in the hospital and caused no HUS at all. The O26:H11 strains that carry the stx2a toxin gene are a different animal. In France, stx2-carrying O26:H11 has become the leading cause of pediatric HUSover the last decade, and the French reference laboratory reports a rising share of those strains in food, including raw milk products.

France lived through this outbreak in 2019. Raw cow’s milk Saint-Félicien and Saint-Marcellin cheeses carried O26:H11 with stx2a. Between March 25 and April 27, French pediatric departments reported 19 suspected STEC-HUS cases, against five to 10 in the same weeks of earlier years; the outbreak was ultimately counted at 16 cases, 14 of them children with HUS, aged six months to four years, and the clinical series reached 20 HUS patients in eight pediatric nephrology centers by September. Italy saw it in 2013: an O26:H11 stx2 outbreak put 20 children with HUS, median age 17 months, into hospitals in one region, two with severe neurological damage. The authors’ conclusion was that O26 infections in children have a high probability of progressing to HUS.

CDC has not said which toxin genes the Graziers strain carries. PulseNet has the sequences, and so do the California and Nevada laboratories. It is a one-sentence answer, and it matters: if this is the stx2a clone, three HUS cases out of 13 is not a surprise. It is the pattern.

The undercount

The 13 are the people whose stool sample grew an isolate that was sequenced and matched in PulseNet. A toddler with three days of diarrhea who never saw a doctor is not in the count. A patient whose laboratory ran a Shiga toxin panel but never cultured an isolate is not in the count. The sicker you are, the more likely you are to be tested, cultured, sequenced and counted, which is exactly why the counted cases skew toward hospitalization and HUS. France 2019 shows the mechanism in its purest form: 14 HUS cases and two cases of plain diarrhea, because the surveillance system was built to find HUS and the mild cases were never looked for.

CDC says so itself: the real number of sick people is likely much higher than the number reported, and the outbreak may reach beyond the nine states. Its burden study estimates that for every laboratory-confirmed non-O157 STEC infection, about 107 go undiagnosed; for O157, about 26.

Expect the count to move. The most recent onset is August 26, CDC’s lag from illness to a counted case runs three to four weeks, the recall notice with product codes did not post until September 29, the cheese keeps for months, and 40-pound blocks went to food service. September’s illnesses are still working their way through the system.

Not the plant’s first positive

As far as the CDC and FDA record shows, this is the first outbreak attributed to Sierra Nevada Cheese Company. It is not the first pathogen found in its cheese. On August 27, 2025, California’s State Veterinarian ordered a statewide recall and quarantine of the company’s Bella Capra raw goat milk cheddar after the California Department of Food and Agriculture found Listeria monocytogenes in a routine sample collected August 14 at the Willows manufacturing and packaging facility. One lot, no illnesses reported. The company sold that cheese as aged a minimum of 120 days. The Graziers labels say “Aged at least 60 days,” and the sharp cheddar “Aged over 120 days.” Thirteen months apart, at the same plant, two pathogens outlasted the only safeguard federal law requires of a raw milk cheese.

The 60-day rule and the petition

FDA’s standards of identity for cheddar and Monterey jack allow unpasteurized milk if the cheese is held at least 60 days at not less than 35°F, a rule adopted in 1949. The joint FDA and Health Canada risk assessment released in 2013 estimated the per-serving listeriosis risk of soft-ripened cheese made from raw milk at 50 to 160 times that of the pasteurized version, and FDA said then that it would reevaluate the 60-day rule. Its own sampling of 1,606 aged raw milk cheeses in 2014 through 2016 found Salmonella, Listeria monocytogenes and Shiga toxin-producing E. coli. Four outbreaks in 31 months: 43 sick, 17 hospitalized, six children with HUS, all from cheese that complied with the rule.

I have been working on a citizen petition to FDA under 21 CFR 10.30. It asks for two things: a prescribed warning statement on cheese made from unpasteurized milk, added to 21 CFR 101.17, the regulation that has required a warning on unpasteurized juice since 1998; and completion of the 60-day reevaluation FDA promised in 2013. The petition will disclose that the firm represents families in the Twin Sisters and Graziers outbreaks.

The draft is posted here. I want it to be right before I file. Every count, date and quotation in it is footnoted, and I would rather hear what I got wrong now than from FDA later. Send comments to bmarler@marlerclark.com by Friday, October 9. Then it gets filed.

Thirteen is where the count stands today. It is not where it will end.