FDA Has Counted Fifty-Two American Sprout Outbreaks Since 1996 and More Than 2,700 Illnesses. Here Is a Thirty-Year Chart of the Big Ones — Twenty-Seven Outbreaks on Four Continents, 14,325 Confirmed Sick and at Least 57 Dead. Here Is What the Experts Have Been Saying, in Their Own Words, Since 1998.

FDA’s own guidance to the sprout seed industry says the agency has observed 52 reported outbreaks of foodborne illness associated with contaminated sprouts in the United States between 1996 and 2020, together causing more than 2,700 illnesses. That is the floor, and it counts only one country. Pulling thirty years of the published record together — CDC final updates, MMWR, peer-reviewed outbreak investigations, WHO, EFSA and ECDC — produces 27 significant sproutbreaks between 1995 and August 21, 2026, on four continents, with 14,325 laboratory-confirmed illnesses and at least 57 deaths. The oldest federal advisory in that stack is dated July 9, 1999. It says essentially the same thing the CDC page posted this week says.

The federal counts do not agree with each other, and all three of them are floors

Three agencies have published three different totals, because they are counting three different things.

FDA’s May 2022 guidance, Reducing Microbial Food Safety Hazards in the Production of Seed for Sprouting, is the broadest: 52 reported outbreaks, 1996 through 2020, more than 2,700 illnesses. FDA’s CORE Network published a narrower slice in Food Safety Magazine in 2021 — 14 multistate sprout outbreak investigations between 2012 and early 2020, 496 laboratory-confirmed illnesses, 82 hospitalizations and three deaths across 45 states. CDC’s National Outbreak Reporting System captures the single-state and restaurant-level events the multistate machinery never sees; a Colorado State University compilation of that NORS data logged at least 53 sprout-associated outbreaks between 2000 and 2020 — 1,498 illnesses, 179 hospitalizations, five deaths, of which 40 were Salmonella, ten were E. coli, three were Listeria, one was norovirus and three were never solved.

None of those three numbers is wrong. They are floors stacked on floors, and every one of them stops before the outbreak announced this week.

Twenty-seven sproutbreaks, 1995 to 2026, on four continents

The table below is the significant published record, not the complete one. Every row is a documented outbreak with a case count tied to a primary source. The single-restaurant events, the recalls with no confirmed illnesses, and the investigations where sprouts were suspected but never confirmed are all left out.

Year	Where	Pathogen	Sprout	Sick	Hosp / died	Source
1995	US, Finland	SalmonellaStanley	Alfalfa	242	—	JID 1997
1996	Sakai, Japan	E. coli O157:H7	White radish	6,561	534 peak / 2	WHO
1996–98	Calif. x6	Salmonella (5), E. coli O157 (1)	Alfalfa, clover	600	— / 2	Ann Int Med
1997	Mich., Va.	E. coli O157:H7	Alfalfa	85*	—	EID 2001
1999	7 states	SalmonellaMuenchen	Alfalfa	157	—	JCM 2001
2005	Ontario	SalmonellaEnteritidis	Mung bean	648	—	Canada
2009	US, 14 states	SalmonellaSaintpaul	Alfalfa	235	3% / 0	CDC
2010	US, 11 states	SalmonellaNewport	Alfalfa	44	7 / 0	CDC
2010–11	US, 26 states	Salmonella I 4,[5],12:i:-	Alfalfa	140	24% / 0	CDC
2011	Germany +15	E. coli O104:H4	Fenugreek	4,075	908 HUS / 50	MMWR
2011	US, 5 states	SalmonellaEnteritidis	Alfalfa	25	3 / 0	CDC
2011–12	US, 11 states	E. coli O26	Clover	29	7 / 0	CDC
2013–14	US, Canada	Salmonella (4 serotypes)	Chia powder	94	21% / 0	Epi Infect
2014	Ill., Mich.	Listeria	Mung bean	5	5 / 2	Frontiers
2014	US, 12 states	SalmonellaEnteritidis	Bean sprouts	115	19 / 0	CDC
2014	Idaho, Wash.	E. coli O121	Clover	19	44% / 0	CDC
2016	US, 12 states	SalmonellaMuenchen and Kentucky	Alfalfa	26	8 / 0	CDC
2016	Minn., Wis.	E. coli O157	Alfalfa	11	2 / 0	CDC
2016	US, 9 states	SalmonellaReading and Abony	Alfalfa	36	7 / 0	CDC
2016	S. Aus., NT	SalmonellaSaintpaul	Bean sprouts	~350*	—	FSN
2018	Ill., Minn., Wis.	SalmonellaMontevideo	Raw sprouts	10	0 / 0	CDC
2018	S. Aus.	SalmonellaHavana	Alfalfa	31	13 / 0	CDI 2019
2020	US, 10 states	E. coli O103	Clover	51	3 / 0	CDC
2022–23	US, 8 states	SalmonellaTyphimurium	Alfalfa	63	10 / 0	CDC
2023–25	10 in Europe	Salmonella (8 serotypes)	Alfalfa	509	—	EFSA
2026	11 nations	SalmonellaBovismorb.	Alfalfa	109	18 / 1	EFSA
2026	US, 15 states	SalmonellaAgona and three STEC	Alfalfa	55	4 / 0	CDC

Table 1. Significant sprout-associated outbreaks, 1995–2026, compiled from the sources linked in the final column, August 21, 2026. Arithmetic check: the Sick column sums to 14,325 across 27 rows, and deaths total 57. Three deliberate exclusions, so the total is not inflated: the 2011 Bordeaux cluster (15 cases, 9 with HUS) sits inside the 16-country O104:H4 total, and the 2024 Norwegian outbreak (230) and concurrent Swedish outbreak (about 100) both sit inside the 509-case European total. The 1997 figure marked with an asterisk is the number of isolates found indistinguishable by PFGE, not the full outbreak count, which MMWR put higher; the 2016 Australian figure is a reported approximation, not an agency case count. Sakai deaths are WHO’s count as of August 26, 1996; later clinical reports say three, and Japan’s national 1996 toll that summer was 12.

Two caveats travel with this table. The counts are laboratory-confirmed people, not sick people, and CDC’s own multiplier work says the real figures are many times larger. And the geography is where patients lived, not where the sprouts went.

The two largest are still Japan in 1996 and Germany in 2011, and both of them started in a seed

In July 1996, white radish sprouts from a single farm, served in school lunches, produced what is still the largest E. coli O157:H7 outbreak ever recorded. WHO reported on August 28, 1996 that the Sakai City outbreak had affected 6,309 schoolchildren and 92 school staff across 62 municipal elementary schools, with roughly 160 further cases mostly among family members. Hospitalizations peaked at 534 on July 18. Hemolytic uremic syndrome peaked at 101 cases on July 24. Two girls, ages 10 and 12, were dead by the end of August. Nationwide that summer Japan logged 9,578 O157:H7 cases and 11 deaths; Japanese accounts have since revised the Sakai figure upward to 9,523 sick and the national toll to 12. I wrote about it last month in a different context. The strain isolated there is still one of the reference genomes microbiologists work from.

Fifteen years later the strain nobody had planned for arrived. Between May and July 2011, a rare enteroaggregative E. coli that had picked up Shiga toxin genes — O104:H4 — moved through northern Germany and fifteen other countries. CDC’s own summary puts the final count at 4,075 cases including 908 complicated by HUS, and 50 deaths, in 16 countries. Six of those cases were American, and one, in Arizona, died. Germany’s federal risk institute puts the German death toll at 53. A separate cluster of 15 people near Bordeaux, nine of whom developed HUS, was traced to the same seed. EFSA tied it to a single lot of fenugreek seed imported from Egypt, and the fact worth sitting with is that nobody ever cultured the organism off that seed. The traceback carried it. The microbiology never did.

Neither of those was a dirty-room story. Both were seeds.

Every large outbreak on this list points back at the seed, and Europe proved it again three months ago

FDA’s CORE analysts published the cleanest version of this finding. Of 14 multistate sprout outbreak investigations between 2012 and 2020, tracebacks converged in nine. Of those nine, six converged at the seed distributor or supplier and only three at a sprout grower — and two of the three growers sourced from a single supplier. FDA’s conclusion was that contaminated seed was the likely cause of most sprout outbreaks over that period.

Then Europe ran the experiment at scale, twice.

The first: between January 2023 and January 2025, EFSA and ECDC pulled together what had been investigated as a string of separate national outbreaks in Germany, Finland, Norway and Sweden and found they were one event — 509 laboratory-confirmed cases in ten countries, eight different Salmonella serotypes, all of it running back to one alfalfa seed supplier in Italy who in turn sourced from three primary seed growers in the same geographic area.

The second is three months old and has not been widely noticed on this side of the Atlantic. On June 25, 2026, EFSA and ECDC published a Rapid Outbreak Assessment on Salmonella Bovismorbificans ST377: 109 confirmed cases between January and May 2026 across ten EU and EEA countries and the United Kingdom, 18 hospitalized, one dead. The outbreak strain turned up in water used for sprouted-seed production in the Netherlands and in Northern Ireland. Investigators identified a common seed supplier and alfalfa seed imported from India, and concluded the seed was likely contaminated before it ever entered Europe.

Eleven countries, one seed supply, and it was still moving five weeks before Minnesota’s first patient got sick.

Four organisms in one bag of sprouts is unusual, but it is not unprecedented

Minnesota and Wisconsin have now found Salmonella Agona plus three separate Shiga toxin-producing E. coliserogroups — O26:H11, O103:H25 and O168:H8 — in people who ate the same alfalfa sprouts, and two people are carrying two organisms at once. That is genuinely rare in the United States.

It is not rare in the sprout literature.

The 2016 Sweetwater Farms outbreak was Salmonella Muenchen and Salmonella Kentucky in patients, with Kentucky and Cubana turning up in irrigation water and sprouts. The 2016 Sprouts Extraordinaire outbreakwas Reading and Abony, and five people carried both. The 2013–14 sprouted chia seed powder outbreak ran four serotypes across the United States and Canada. And in Norway in the autumn of 2024, investigators identified four Salmonella serovars — Newport, Typhimurium, Kisarawe and Kinondoni — in 230 people, a third of whom were hospitalized. It was the largest Salmonella outbreak in Norway since the 1980s. The investigators wrote that the diversity of serovars suggested contamination was not a single event but multiple contamination points along the production and supply chain.

That is the argument. One dirty pipe gives you one organism. Four organisms in one product is a seed-supply signal, and it has been a seed-supply signal every time it has happened.

The experts have been saying the same thing for twenty-eight years, and here it is in their own words

1998 — Larry Beuchat, food microbiologist at the University of Georgia, on why seed disinfection does not work: “All of the sanitizers have some effect on the pathogens, but none eliminates them.” He added that a single surviving pathogen on a seed can multiply to a million or ten million per gram during sprouting.

1999 — FDA, health advisory P99-13, July 9: “People in high risk categories should not eat raw sprouts.” Everyone else was told that anyone wishing to reduce their risk was advised not to eat raw sprouts at all.

1999 — Jane Henney, M.D., FDA Commissioner, announcing the first sprout guidance in October: “FDA is committed to preventing foodborne illness associated with this product.” The guidance was built on a National Advisory Committee on Microbiological Criteria for Foods review FDA had commissioned in 1997.

1999 — Taormina, Beuchat and Slutsker, in CDC’s Emerging Infectious Diseases, concluded that until effective preventive measures were identified, anyone wanting to reduce their risk should not eat raw sprouts, and that the elderly, children and the immunocompromised should not eat them at all.

2001 — Mohle-Boetani and colleagues, in the Annals of Internal Medicine, after investigating six sprout outbreaks in California: “As currently produced, sprouts can be a hazardous food.” They estimated those six outbreaks caused about 22,800 illnesses, against 600 culture-confirmed cases.

2002 — FDA, Talk Paper T02-37, extending the warning to mung beans: “hospitals, day care centers, nursing homes and senior centers should not serve sprouts.”

2011 — Christopher Braden, M.D., then of CDC, on why sprout outbreaks are so hard to solve: he called sprouts a “stealth vehicle,” because people do not remember eating something tucked into a sandwich.

2016 — Michael Doyle, director of the University of Georgia Center for Food Safety, asked what he personally avoids: raw bean sprouts made the list, alongside medium-rare hamburger and bagged salad.

2022 — FDA, in the current seed guidance, describing why the document exists at all: the agency’s “serious concern with the continuing outbreaks of foodborne illness” associated with raw and lightly cooked sprouts. Twenty-three years after the first advisory, the word FDA reached for was continuing.

2025 — the Norwegian outbreak investigation team, writing in Infection, recommended that alfalfa sprouts be formally recognized as a high-risk product and that testing rules be revised, because a supplier reporting negative Salmonella results may be reporting the limits of its testing rather than the absence of contamination.

2026 — Barbara Kowalcyk, director of the Institute for Food Safety and Nutrition Security, on why raw sprouts are on her no-list: “there’s really no good way to clean them.”

2026 — CDC, in the standing advice at the top of every sprout outbreak page it has published for a decade: “Raw and lightly cooked sprouts are a known source of foodborne illness.”

Twelve voices over twenty-eight years, from FDA, CDC, two universities, a peer-reviewed journal and the Norwegian outbreak team. Read them in order and the striking thing is not the disagreement. There isn’t any.

The advice has not changed since 1999 because the product has not changed

I have written some version of this before. In February 2016 I counted 75 sproutbreaks and said sprouts were still not safe to eat. Ten years on, the count is higher, the genomics are better, the tracebacks are faster, and the answer to the underlying question is identical. There is still no validated kill step for a food grown in warm water and eaten raw.

That is not a criticism of the people working this outbreak. The Minnesota epidemiologists found the 2016 alfalfa sprout outbreak, they found this one, and they named the grower before FDA had a page up. The problem sits upstream of every one of them. Seed for sprouting is not covered produce under the Produce Safety Rule — FDA says so in the 2022 guidance and explains why, which is that the person growing the seed often does not know it will be sprouted. The single input responsible for most of the outbreaks in the table above is therefore the one input nobody is required to grow as food.

Fifty-two American outbreaks. More than 2,700 American illnesses. Twenty-seven big ones on four continents and 14,325 confirmed sick. Fifty-seven people dead, most of them in two events fifteen years apart, both of which started in a seed. My firm has handled sprout claims for more than twenty-five years and filed the Caudill cases in 2009, when contaminated seed went to growers all over the country. The seed is still the story.

The 1999 advisory told people at high risk not to eat raw sprouts. Twenty-seven years later it is still the best advice anyone has. Our firm notice on the current outbreak, with what to check for in your refrigerator, is here.