The Idaho Department of Health and Welfare (DHW) is working with local and state partners to investigate a recent increase in illnesses after consumption of unpasteurized (raw) milk. Since May 19, 2026, nearly 60 people have been identified who became ill after consuming raw milk. At least 45 of those people tested positive for campylobacteriosis, a bacterial infection, although not everyone who is sick was tested. Investigation and interviews of people reported as ill are ongoing and additional illness may be identified. 

The Idaho Division of Public Health is collaborating with Panhandle District Health, Southwest District Health, Central District Health, Southeastern Idaho Public Health, South Central Public Health, and Eastern Idaho Public Health on the investigation.

Most people who were sick reported consuming raw milk from two different milking operations – one in northern Idaho and one in southern Idaho. 

Investigation is ongoing to identify potential batches of concern and test milk samples.

Raw, unpasteurized dairy products can contain bacteria that make people sick, particularly young children, pregnant women, the elderly, and those who are immunocompromised. Pasteurization kills nearly all the germs that can exist in raw milk while maintaining nutritional benefits. 

Common symptoms of campylobacteriosis infection include diarrhea (sometimes bloody), fever, stomach cramps, nausea and/or vomiting. Symptoms usually start two to five days after exposure and last about one week; some people may develop complications that last longer. Anyone who is experiencing symptoms after consuming raw milk or raw milk products is encouraged to promptly seek medical care. 

For over thirty years, I have been beating the same drum, and the last few days were no exception (some argue a bit too loudly and self-serving). 

I have been posting about public health officials — and the FDA — sending out outbreak documents with the names of companies, growers, processors, and retailers blacked out behind gray boxes I do not think belong there. My position has not changed in three decades: most of those redactions should never have been made in the first place. Not surprisingly, my readers wrote back.

I welcome the pushback. Some of it is thoughtful. A little of it I even agree with. So rather than let the arguments for secrecy sit unanswered in my inbox, let me put them all on the table — every justification for non-disclosure I have heard over the years — and tell you exactly what I think of each.

“There is no written policy, but it is the way we have done things for years.”

I keep hearing my mother’s voice: “just because so-and-so does it does not mean you should too.” A practice is not a policy, and “we have always done it this way” is not a reason — it is the absence of one. Like all government habits (and, frankly, like neckwear), change is good.

“The outbreak is over, so there is no immediate public health threat.”

Frankly, that is true in most foodborne illness outbreaks. In nearly every outbreak I have ever seen, the source is figured out long after the peak of illnesses has passed. But that misses the point of disclosure. Telling the public who grew, processed, and sold the contaminated food is not only about stopping today’s outbreak — it is about giving consumers a track record, so they can see which companies have a strong food safety record and which keep turning up in the same sad story.

“Naming the company jeopardizes its cooperation in this and future outbreaks.”

If a company will cooperate only in exchange for a spot in the witness protection program and a promise of permanent anonymity, that tells you everything you need to know about its commitment — and our government’s commitment — to safe food. Cooperation purchased with secrecy is not cooperation. It is a hostage negotiation in which the public never gets a seat at the table.

“Bad publicity may cause economic hardship for the company.”

True. But here is a better business practice than avoiding bad press: not poisoning your customers. The economic hardship of an outbreak belongs to the company that caused it — not to the consumers left guessing.

“The source was an unknown supplier, so naming the point of service unfairly blames that business.”

This one actually makes some sense, and I will give it its due. But ask the obvious follow-up questions. Is this the first time this restaurant or retailer was burned by a faulty supplier, or is it a pattern? And even if it is the first time, is some of that unnamed product still sitting on shelves or in walk-in coolers somewhere? Fairness to the point of service cannot come at the expense of the people it serves.

“The product is perishable, so by the time we announce, it has already been eaten.”

I have heard this one a “bunch” of times, especially in leafy green outbreaks, and it is the one I find most cynical. Yes, the romaine is long gone. But why should the public be left in the dark about the kind of product that sickened people, and about the grower and shipper behind it, when that is precisely the information they need to decide who to buy from next time? “It’s already been eaten” is a reason to disclose, not a reason to hide.

“Going public with the point of service before the investigation is complete compromises the epidemiology.”

I completely agree with this one. This is the hard call — the one I suspect causes public health officials the most genuine angst. They are balancing the need to have enough data to go public and protect people against the risk of pointing the finger too soon. We have all lived through “it’s the tomatoes — no, wait, it’s the peppers.” The answer is not to stay silent forever. The answer is simple: do not go forward until the investigation is complete. Then go forward and go forward fully.

“We are afraid of making an investigation mistake — the tomatoes-then-peppers problem.”

This is exactly why the law gives public health officials immunity from liability for the good-faith decisions they make to protect the public. The system already accounts for honest error. Fear of an honest mistake is not a license for permanent silence.

“Surveillance is underfunded, and there simply are not the resources to finish investigations.”

There is no question this is true, and it is the argument that worries me most — because it is not really an argument for secrecy at all. It is a confession. I have watched investigations get dropped over the last few years. Labs that are not doing the genetic fingerprinting that links sick people to a common source. Tracebacks that stall out for lack of the people needed to find the root cause. That is a scandal in its own right, and the answer is to fund the work — not to redact our way out of admitting it is not getting done.

And now let me say the part the FDA never will. We do not have to guess whether transparency would cause the sky to fall, because we have already run the experiment. For the better part of twenty years, the USDA’s Food Safety and Inspection Service has named the manufacturers of contaminated meat — and, since 2008, the retailers who sold it. The trade secrets of the beef and poultry industries did not collapse. Consumer confidence went up, not down. Chicken Little, the sky did not fall.

The FDA, which oversees roughly eighty percent of the food supply, could have learned that lesson from its sister agency two decades ago. Instead, it still hides behind “confidential commercial information.” And the line is not hard to draw. Formulations, ingredients, and how a product is made — those are trade secrets. Who supplied the tainted raw material, who made the tainted product, and where the tainted product was sold are not, especially during an outbreak. The most egregious example I know of remains the 2017 outbreak tied to I.M. Healthy soy nut butter — a great name for a product carrying a pathogen — which sickened dozens, some of them children, seriously. A recall was announced. No retailers were named. The company went bankrupt and was in no position to help. The public was left to fend for itself.

I have just watched the same instinct play out again. In the fall of 2024, a multistate E. coli O157:H7 outbreak tied to romaine lettuce hospitalized dozens, sent children into kidney failure, and killed someone. The FDA logged a few lines in a table, closed the file, and never issued the named, complete public notice it had issued for every comparable romaine outbreak before it. The processor and the grower sat behind gray boxes while they were free to tell anyone who asked that the evidence did not point to them. It did. The records said so, by name — once someone bothered to make the agency take the boxes off.

So where does that leave me, after over thirty years and more redacted documents than I care to count? Right where I started. Strip away the budget excuses, the cooperation worries, and the trade-secret theater, and you are left with a single question: who gets to decide what the public is allowed to know about the food that can kill them?

For me, the answer is easy. The public has a right to know, and to use that information however it sees fit. And people — especially government employees — have no business deciding for the rest of us what we should and should not be allowed to find out.

A new outbreak of Listeria monocytogenes (ref #1380) linked to a not yet identified product has been added to the table. FDA has initiated traceback and onsite inspections.

A new outbreak of Cyclospora (ref #1375) linked to a not yet identified product has been added to the table. FDA has initiated traceback.

For the outbreak of Salmonella Typhimurium (ref #1377) linked to moringa leaf powder, FDA has initiated sampling. 

A little over a year ago, in January 2025, I wrote in this space asking a simple question about the November 2024 E. coliO157:H7 romaine outbreak: why would the FDA not tell the public who grew, processed, and sold the lettuce that sickened 89 people across 15 states, hospitalized 36, gave 7 of them hemolytic uremic syndrome, and killed one? The agency had quietly announced the outbreak was over, called the vehicle a “common supplier” of romaine lettuce, and left it there. No grower. No processor. No names.

I said then that I would weigh in where the FDA would not. I have now spent the months since doing exactly that — staying on the agency until it unredacted the documents it should never have hidden in the first place. The documents are now in hand, and they tell the story the gray boxes were built to keep from you.

Here is the short version of how this played out.

When the FDA first released its traceback investigation summary for this outbreak (CARA #1280), nearly every name that mattered was blacked out under the (b)(4) exemption — the part of the Freedom of Information Act meant to shield genuine confidential commercial information and trade secrets. The processor was a gray box. The grower was a gray box. The ranch was a gray box. The distribution centers, the brokers, the lot codes — all gray boxes. What the public was left with was a document that confirmed romaine lettuce as the vehicle but told you absolutely nothing about whose romaine it was or where it came from.

So, I did what I always do. I stayed on them. And the FDA, in stages, unredacted the document.

The fully unredacted version now identifies, in plain text, what the (b)(4) boxes were hiding:

  • Taylor Farms of California (Salinas, CA; FEI 3012342127) was the sole processor. The summary states that Taylor Farms “supplied all the romaine lettuce that would have been available at all points of sale during the timeframe of interest.”
  • Anthony Costa & Sons LLC (Soledad, CA) was the single grower.

The records also fill in the other end of the chain, and that too was originally redacted. The unredacted summaries name Andre’s Banquets and Catering of St. Louis, Missouri (1 of the 15 states) as the caterer at the center of the largest cluster of illnesses — the events where many of these people actually ate the lettuce. The traceback ties Andre’s to three separate catered events with meal dates of November 6 through 8, 2024, including a marching band banquet and a Veteran’s Day luncheon at a St. Louis-area high school, with twenty-two cases linked to that point of service alone. The lettuce blend Andre’s served was supplied through its distributor and traced straight back up the chain to Taylor Farms and Anthony Costa & Sons. But the public had no way to know any of that while the caterer’s name, the school events, and the supplier above them all sat behind the same gray boxes — which meant the families who got sick at those events could not see, in the agency’s own records, how the lettuce on their plate connected to the field it came from.

None of that is a trade secret. The identity of the company that processed lettuce that killed someone is not confidential commercial information. The name of the farm that grew it is not a trade secret. A harvest date is not a proprietary formula. These are the basic facts of a public health disaster, and the only thing the redactions accomplished was to delay the moment when the public — and the families of the people who were hospitalized — could learn who was responsible.

I have been making this same point for years, and I will make it again here, because the FDA keeps forcing me to. Formulations, ingredients, and how products are made are trade secrets. Who supplied the tainted raw material, who made the tainted product, and where the tainted product was sold are not — especially during an outbreak. That line is not hard to draw. The agency simply refuses to draw it.

This matters for a reason that goes well beyond my own irritation. While those names sat behind gray boxes, the companies were free to take the position, in the investigation and elsewhere, that the evidence did not point to them. I have watched that move many times: a processor reviews its own internal traceability, announces it “determined there were no commonalities identified,” and lets the redactions do the rest of the work. But the FDA’s own records told a different story. The agency’s traceback identified a single processor and a single grower, found that the implicated product was available at every point of service, and tied it to specific lots. The unredacted document does not leave room for “it wasn’t us.” It was them. The records say so, by name.

That is exactly why over-redaction is not a harmless bureaucratic habit. When the FDA blacks out the name of the processor and the grower in a fatal outbreak, it is not protecting a trade secret — it is handing the responsible companies a period of deniability they did not earn, and the public did not consent to. The (b)(4) exemption was written to protect genuine competitive information, not to spare a lettuce processor the embarrassment of being named in connection with the produce it sold.

We do not have to guess whether transparency would cause the sky to fall, because we have already run the experiment. For most of the 2000s the USDA’s Food Safety and Inspection Service would name the manufacturer of E. coli-contaminated meat but refused to reveal where it was sold. I still remember people sickened in the 2002 ConAgra outbreak telling investigators they had heard about the recall but figured it did not apply to them because “we bought our meat at Safeway, not at ConAgra.” In 2008, after years of industry hand-wringing about distribution lists, FSIS finally concluded that naming retailers would not cause substantial competitive harm and began releasing the lists. Retailers of USDA-regulated products survived. Trade secrets did not collapse. Consumer confidence went up, not down. Chicken Little, the sky did not fall.

The FDA, which oversees roughly 80 percent of the food supply, could have learned that lesson from its sister agency two decades ago. Instead, it is still hiding behind “confidential commercial information,” and consumers are still left confused — not by too much information, but by too little. The most egregious example I know of remains the 2017 outbreak tied to I.M. Healthy soy nut butter — a great name for a product carrying a pathogen — which sickened dozens, some of them children, seriously. A recall was announced but no retailers were named, the company went bankrupt and had no interest in helping, and the product stayed on shelves and online for months after the recall. That is what secrecy buys you: contaminated product still within reach of the next family while the agency guards a “secret” that protects no one but the seller.

Years ago, a senior CDC official defended withholding company names by saying the practice protects not only public health but also “the bottom line of businesses that could be hurt by bad publicity.” I have a great deal of respect for the people in the diarrheal trenches, but that explanation has always had it exactly backwards. The government does not exist to protect a company’s bottom line from the consequences of selling food that hurts people. Protecting the public sometimes means the responsible company takes a reputational hit. That is not a bug in transparency. That is the point of it.

I would put the principle simply. When people are hospitalized and someone dies, the public’s interest in knowing who grew and processed the food is at its highest, and the commercial interest in staying anonymous is at its lowest. The FDA’s redaction practice in cases like this one inverts that balance. It treats the names of the firms as the secret most in need of protection, when the names are the single most important thing the public is entitled to know.

I am glad the FDA eventually unredacted this one. I should not have had to ask. The next family should not have to wait for a lawyer to pry the names loose months after the outbreak is over and the lettuce is long gone. If the agency genuinely wants to rebuild public trust in how it handles foodborne illness, it can start by drawing the (b)(4) line where the statute actually puts it — and by remembering that the public reporting on an outbreak is supposed to inform the public, not shield the companies.

The lettuce in this outbreak was grown by Anthony Costa & Sons and processed by Taylor Farms. The FDA’s own records say so. There was never a good reason for anyone to have to guess.

When ten people die in a Listeria outbreak traced to a single plant with a documented, years-long record of filth, I pay attention to what the government promises to do about it. And I keep a calendar.

By that calendar, it has now been roughly twenty months since the Department of Justice was formally asked to decide whether Boar’s Head should face criminal charges. It has been roughly twenty months since the USDA Inspector General opened a review of how its own inspectors let the Jarratt, Virginia plant keep running for years while the noncompliance’s piled up. The plant has since closed, been “rebuilt from the inside out,” and reopened. The news cycle has moved on.

And from the two investigations that actually matter for accountability — the DOJ’s and the IG’s — we have heard essentially nothing.

Let me be precise, because precision is the whole point. I am not claiming the investigation was dropped. I do not know that, and neither does anyone outside a grand jury room. What I know is what the public record shows, and what it shows is a long, deliberate silence.

We know there was a real law-enforcement investigation, not just a request for one. When the Associated Press asked USDA for the inspection and enforcement records on Jarratt and eight other Boar’s Head facilities, the agency refused to hand them over. Its stated reason was that the records had been compiled for a law-enforcement purpose covering “both civil and criminal statutes,” and that releasing them could interfere with the investigation. You do not invoke a law-enforcement FOIA exemption over a matter that does not exist. That refusal, in writing, is the strongest confirmation we have that someone with a badge was looking hard at this company.

We know two members of Congress — Senator Richard Blumenthal and Representative Rosa DeLauro — referred the matter to the Attorney General and the Secretary of Agriculture in September 2024 and asked, in plain language, whether criminal charges were warranted. We know the IG opened a review of FSIS’s conduct. We know that as recently as the plant’s February 2026 reopening, reporters were still writing that DOJ “is investigating whether criminal charges are warranted” — present tense — and that the IG “is reviewing” the agency’s handling. Present tense, twenty months on.

What we do not have is a single public document closing the loop. No indictment. No criminal information. No plea. No deferred-prosecution agreement. No consent decree. No public declination explaining why no charge was brought. And no public report from the Inspector General on whether the inspectors who walked past mold, insects, and dripping condensation for years did their jobs. On the central questions, the file is blank.

I want to head off the easy excuse. Someone will say these things take time. They do — and I have made that argument myself. Food-safety criminal cases routinely run two to four years from outbreak to charge. The Peanut Corporation of America prosecution that put Stewart Parnell away for twenty-eight years took years to build. Chipotle’s $25 million deferred-prosecution deal in 2020 followed outbreaks that began long before. So, I am not banging the table because charges haven’t appeared on a schedule I prefer. I am asking a narrower and fairer question: is anyone still working this, and will the public ever be told how it ends?

Because here is what should make every regulator uncomfortable. While we wait, the same company’s other plants keep generating the same kind of paperwork. Records released to the AP around the reopening documented dozens of noncompliance reports at a second Boar’s Head facility in 2025 — dripping condensation, meat residue, and a failure to follow the company’s own written Listeria controls. One inspector noted that a single violation was the fifth of its kind in a month. That is not ancient history dredged up by a plaintiff’s lawyer. That is last year. If the threat of accountability has already faded, the conduct it was supposed to deter apparently has not.

I am asking, on the record, two simple questions.

To the Department of Justice: What happened to the Boar’s Head investigation? If you charged it, say so. If you closed it, say so and tell the families why ten deaths did not clear your bar. The responsible-corporate-officer doctrine and the Federal Meat Inspection Act exist precisely for executives who preside over uncorrected, dangerous conditions. Either those tools applied here or they did not. The public deserves an answer, not a shrug.

To the USDA Inspector General (assuming there is one): What happened to your review of FSIS? You opened it to learn how inspectors and the cooperative state-inspection arrangement let Jarratt operate for years. That answer belongs to the public, because the public is the next set of people who will eat what these plants ship.

Ten people are dead. Sixty were hospitalized across nineteen states. A plant that should never have been allowed to deteriorate the way it did is back in business. The least we are owed — the families, the survivors, and everyone who buys lunch meat trusting that someone is watching — is to be told whether anyone in Washington ever finished the job.

I will keep my calendar open. I would rather fill in an answer than another month of silence.

Legal Basis for a Criminal Investigation Under the Federal Meat Inspection Act

1. The prohibited conduct

The Federal Meat Inspection Act (“FMIA”), 21 U.S.C. § 601 et seq., makes it unlawful to sell, transport, offer for sale or transportation, or receive for transportation in commerce any meat or meat food product that is adulterated or misbranded. 21 U.S.C. § 610 (prohibited acts), including § 610(c). A federal criminal investigation of a meat processor proceeds from the premise that adulterated product entered commerce in violation of § 610.

2. What makes the product “adulterated”

“Adulterated” is defined at 21 U.S.C. § 601(m). Two clauses are directly relevant to a Listeria outbreak: § 601(m)(1) covers a product bearing or containing any poisonous or deleterious substance that may render it injurious to health, and § 601(m)(4) covers a product prepared, packed, or held under insanitary conditions whereby it may have become contaminated with filth or rendered injurious to health. FSIS treats Listeria monocytogenes in ready-to-eat product as an adulterant on a zero-tolerance basis, so confirmed contamination of RTE deli meat establishes adulteration under § 601(m)(1), and a documented record of insanitary plant conditions independently supports adulteration under § 601(m)(4).

3. The criminal penalty structure

Criminal penalties are set by 21 U.S.C. § 676(a), which creates two tiers:

  • Misdemeanor (baseline). Any violation of the FMIA for which no other penalty is provided is punishable by up to one year of imprisonment and a fine. This tier requires no proof of intent or knowledge — it is a strict-liability “public-welfare” offense.
  • Felony (elevated). The offense becomes a felony, punishable by up to three years of imprisonment and a larger fine, if the violation involves intent to defraud or any distribution or attempted distribution of an article that is adulterated (other than the narrow pesticide/additive carve-out at § 601(m)(8), which does not apply to Listeria). Because Listeria-contaminated RTE meat is adulterated under § 601(m)(1), distributing it can supply the felony predicate even without proof of intent to defraud.

Separately, § 676(b) lets the Secretary decline to refer minor violations for prosecution where a written warning will adequately serve the public interest — a discretion provision, not a shield for a repeated, uncorrected pattern of the kind documented at Jarratt.

4. Reaching corporate officers: the responsible-corporate-officer doctrine

Because the FMIA misdemeanor is a strict-liability public-welfare offense, individual executives can be charged under the responsible-corporate-officer doctrine of United States v. Dotterweich, 320 U.S. 277 (1943), and United States v. Park, 421 U.S. 658 (1975). Under Park, an officer who stood in a position of authority and responsibility over the conditions that produced the violation — and who had the power to prevent or correct them — may be convicted of the misdemeanor without any showing that the officer personally participated in or knew of the violation. A documented, repeated, uncorrected pattern of insanitary conditions is the classic factual setting in which these tools are deployed.

5. Who investigates and who charges

  • FSIS. USDA’s Food Safety and Inspection Service investigates suspected FMIA violations through its Office of Investigation, Enforcement, and Audit and refers matters warranting prosecution to the Department of Justice. FSIS’s invocation of the law-enforcement FOIA exemption (Exemption 7, 5 U.S.C. § 552(b)(7)) over the Boar’s Head records is the public marker that such an investigation existed.
  • DOJ. Criminal FMIA matters are prosecuted by the Department of Justice — typically the Consumer Protection Branch (which handles FMIA and Food, Drug, and Cosmetic Act food prosecutions) working with the U.S. Attorney’s Office for the relevant district, here the Eastern District of Virginia. Any charging instrument would appear as an indictment or criminal information on that court’s docket.
  • Grand-jury secrecy. An ongoing federal criminal investigation is shielded by Fed. R. Crim. P. 6(e), which is why continued silence is consistent with either a pending matter or a closed one; the public record cannot distinguish between them.

6. The distinct basis for the FSIS internal investigation

The USDA Inspector General’s review is grounded in different authority — the Inspector General Act of 1978, now recodified at 5 U.S.C. § 401 et seq. (formerly 5 U.S.C. App.) — which empowers the IG to audit and investigate the programs and operations of USDA, including whether FSIS and the cooperative (Talmadge-Aiken) state-inspection arrangement responded appropriately to the documented conditions at Jarratt. This is oversight of the agency itself, separate from any FMIA prosecution of the company, and its findings are ordinarily released to the public in an IG report.

Whole genome sequencing results show that the beef kofta samples collected by FSIS and produced at Olympia Food Industries (Est. 18743) matched the outbreak strain of E. coli O157:H7. FSIS continues to coordinate with the California Department of Public Health and local health departments in California on the outbreak investigation.

The U.S. Department of Agriculture’s Food Safety and Inspection Service (FSIS) is issuing a public health alert due to concerns that beef kofta products served at The Kebab Shop restaurant locations may be contaminated with Shiga toxin-producing E. coli (STEC) O157:H7. A recall was not requested because the products are no longer available for purchase.

The beef kofta was produced as a raw ground beef product by Olympia Food Industries, Inc. dba Olympia Foods (Est. 18743) in Franklin Park, Illinois, on January 6, 2026, and supplied to The Kebab Shop restaurant locations in California, Texas, and Florida.
The problem was discovered as part of an ongoing illness outbreak investigation. FSIS, the California Department of Public Health (CDPH), and local health departments in California are investigating a localized outbreak of E. coli O157:H7 that includes 9 sick people in California. As of May 24, 2026, illness onset dates have been reported ranging from March 27, 2026, to April 30, 2026. Because the identified illnesses are limited to California, CDPH is leading this investigation with FSIS. FSIS continues to keep its federal partners informed as the investigation progresses. FSIS collected raw ground beef kofta product samples that tested positive for E. coli O157:H7. Further testing is ongoing to determine if the product samples are related to the specific outbreak strain.

FSIS is issuing this public health alert to ensure that consumers in California, Texas, and Florida are aware of the outbreak. The Kebab Shop stopped selling beef kofta at all of its restaurant locations on May 18, 2026.

E. coli O157:H7 is a potentially deadly bacterium that can cause dehydration, bloody diarrhea and abdominal cramps 2 to 8 days (3 to 4 days, on average) after exposure to the organism. While most people recover within a week, some develop a type of kidney failure called hemolytic uremic syndrome (HUS). This condition can occur among persons of any age but is most common in children under 5 years old and older adults. It is marked by easy bruising, pallor, and decreased urine output. Consumers who ate beef kofta from any location of The Kebab Shop and develop symptoms of STEC infection within 10 days of exposure should contact their health care provider. Consumers should discard any leftover beef kofta from The Kebab Shop.

William “Bill” Marler has been a food safety lawyer and advocate since the 1993 Jack-in-the-Box E. coli Outbreak which was chronicled in the book, “Poisoned” and in the recent Emmy Award winning Netflix documentary by the same name. Bill work has been profiled in the New Yorker, “A Bug in the System;” the Seattle Times, “30 years after the deadly E. coli outbreak, A Seattle attorney still fights for food safety;” the Washington Post, “He helped make burgers safer, Now he is fighting food poisoning again;” and several others. 

Dozens of times a year Bill speaks to industry and government throughout the United States, Canada, Europe, Africa, China and Australia on why it is important to prevent foodborne illnesses.  He is also a frequent commentator on food litigation and food safety on Marler Blog. Bill is also the publisher of Food Safety News.

E. coli:  Marler Clark, The Food Safety Law Firm, is the nation’s leading law firm representing victims of E. coli outbreaks and hemolytic uremic syndrome (HUS). The E. coli lawyers of Marler Clark have represented thousands of victims of E. coli and other foodborne illness infections and have recovered over $900 million for clients. Marler Clark is the only law firm in the nation with a practice focused exclusively on foodborne illness litigation.  Our E. coli lawyers have litigated E. coli and HUS cases stemming from outbreaks traced to ground beef, raw milk, lettuce, spinach, sprouts, and other food products.  The law firm has brought E. coli lawsuits against such companies as Jack in the Box, Dole, ConAgra, Cargill, and Jimmy John’s.  We have proudly represented such victims as Brianne Kiner, Stephanie Smith and Linda Rivera.

If you or a family member became ill with an E. coli infection or HUS after consuming food and you’re interested in pursuing a legal claim, contact the Marler Clark E. coli attorneys for a free case evaluation.

Additional Resources:

I have spent more than thirty years representing the people who get left holding the bill for someone else’s contaminated food. I have read more outbreak reports than I can count. And in all those years, I have learned to pay close attention not just to what a government agency tells the public, but to what it decides to hide.

So, when the FDA finally released its Executive Incident Summary on the ByHeart infant botulism outbreak, the first thing I looked for was the name of the company that supplied the powdered milk. It wasn’t there. In the single most important sentence of the document, the one tracing the contaminated ingredient back through the supply chain, the agency had blacked out two things: the name of the supplier and the place it operates. The sentence now reads that eight whole milk lot powders were traced to 33 fluid milk lots from a company FDA won’t name, located somewhere FDA won’t say.

Let me say plainly why that was the wrong call. And let me start with the part that should trouble any lawyer who has ever filed a public records request.

The wrong exemption for the wrong reason

When a federal agency withholds information, it has to cite a legal basis. FDA stamped these redactions with exemption (b)(5). That is the deliberative-process privilege. It exists to protect the give-and-take of internal agency decision-making, draft opinions, recommendations, the pre-decisional back-and-forth that helps officials reach a conclusion. It is not a tool for hiding facts.

The name of a milk supplier and the state it sits in are not anyone’s deliberations. They are findings. They are the factual result of a traceback investigation, the kind of bedrock fact a public health record exists to convey. Even the exemption that might at least be argued here, (b)(4), which covers confidential commercial information, would be a stretch given what’s already public. But (b)(5) isn’t even the right neighborhood. Using the deliberative-process privilege to black out a company’s name reads less like a careful legal judgment and more like reaching for whatever stamp was closest to hand. When the government has to misapply a privilege to keep a fact from the public, that is usually a sign the fact should have been public all along.

A redaction that protects no one

This outbreak sickened 48 infants across 17 states. Every single one of them was hospitalized. The youngest victims of any foodborne outbreak are the ones who can do nothing to protect themselves, and botulism in a baby is about as frightening a diagnosis as a parent can hear: a toxin that attacks the nervous system and causes paralysis, treatable only with a specialized antitoxin made from pooled human plasma. 

These families did everything right. They bought a premium formula marketed as the next-best thing to breast milk. What they got was a product whose contamination, by FDA’s own case definition, reaches all the way back to March 23, 2022, the very day ByHeart began manufacturing. The agency drew the outbreak’s starting line at the company’s first day in business and then declined to name the supplier that fed into it.

Here is the part that should bother every one of us. The name isn’t actually a secret. Dairy Farmers of America has already confirmed, publicly and on the record, that its plant in Fallon, Nevada dried the milk, that Organic West was the source of the milk in the sample the FDA collected, and that Organic West sold the resulting powder to ByHeart. Trade publications have walked the entire chain in detail: roughly 55 organic farms shipping to Organic West, the milk dried into organic whole milk powder at the DFA facility, and that powder going into ByHeart’s formula before testing found botulinum toxin in sealed cans and in the ingredient itself.

So, the FDA redacted a name that the implicated processor has already volunteered to the world. That is not protecting an investigation. That is not guarding a trade secret. And here is the detail that turns an odd decision into an indefensible one: FDA named these very companies in its own public outbreak advisory. The agency’s February 26 advisory states in plain text that the milk powder isolates were collected at Dairy Farmers of America, the processor for ByHeart’s supplier, Organic West Milk. The agency put the names in one official document and then blacked them out of another, citing a privilege, while the underlying fact sat published on its own website. You cannot credibly claim a name must be withheld to protect anything when you have already printed it yourself.

It matters all the more because the abstract concedes the investigation came up empty on the central question. In FDA’s own words, the findings could not identify the source or root cause of the contamination. When an agency cannot tell the public why a product poisoned 48 babies, the least it owes them is a complete account of what it did learn, including where the ingredient came from. Redacting the one concrete link in the chain, after admitting you never found the cause, leaves the public with less than the facts already on the record.

Transparency is not a courtesy, It is the job.

People sometimes forget what these reports are for. An outbreak summary is not a press release for the companies involved. It is a public health document. Its purpose is to tell parents, pediatricians, regulators, and yes, other formula makers, what went wrong so that it does not happen again.

When you redact the supplier, you take that knowledge away from exactly the people who need it. Other manufacturers buying organic whole milk powder have a direct interest in knowing whose product was implicated and how. Public health officials in 17 states have an interest in a complete record. And the families whose babies spent weeks in hospital beds have a right to know the full path the contamination traveled to reach their kitchen counters. Withholding the name does not make anyone safer. It only makes accountability harder.

I have said for years that the FDA and the CDC too often decide that the public can handle less information than it can. We saw it with the romaine outbreaks. We are seeing it again here. The instinct toward secrecy is presented as caution. In practice it is the opposite of caution, because nothing invites a repeat outbreak faster than a record the next manufacturer cannot learn from.

Secrecy lets the finger-pointing win

There is a practical consequence to all this beyond principle. When the government keeps the supply chain in the dark, it hands the companies involved the power to write the story themselves.

That is already happening. The Dairy Farmers of America has put out a statement saying its powder met all required tests and reminding everyone that manufacturers of end-use consumer products have a responsibility to properly process ingredients to ensure product safety. Translation: not our problem, it was ByHeart’s job to control for this. ByHeart, for its part, declined to identify the source of the milk powder samples its own testing flagged. Each link in the chain is busy positioning blame at the next link, and the FDA’s redaction simply gives them more room to do it in the shadows.

Meanwhile we know what the FDA found inside ByHeart’s own operation. As far back as December 2023, the agency raised warnings, after inspectors documented a leaking roof, mold in a water tank, and 2,500 dead flies in a food production area at the company’s plant. Inspectors also found the facility violating its own rules for maintaining the temperatures needed to kill bacteria before packaging. That is the record. The public deserves to weigh it alongside a full and unredacted account of where the contaminated ingredient came from, not a version with the inconvenient names removed.

What the agency should do now

The fix is simple. Reissue the abstract without the redactions, or at minimum drop the (b)(5) stamp that never fit a factual finding to begin with. Name the supplier and its location the way the agency already named them in its own advisory. Stop treating a public health record as a document to be negotiated with the companies it describes.

I am not asking the FDA to assign legal blame. That is what courts are for, and the civil justice system will sort out responsibility among the farms, the processor, and the manufacturer with the benefit of full discovery. What I am asking is far more basic: that the agency charged with protecting our food supply tell us the truth, completely, when babies have been hospitalized in 17 states and the agency itself admits it never found the cause.

Forty-eight families learned the hard way that the system failed them. The least the FDA can do is give them, and the rest of us, an honest accounting. A black box where a name should be, propped up by a privilege that doesn’t fit, is not an honest accounting. It is the agency deciding, one more time, that we are better off not knowing.

Most people who survive Shiga toxin–producing Escherichia coli hemolytic uremic syndrome (STEC-HUS) recover, yet a meaningful minority are left with lasting neurological problems, and in adults some of that injury declares itself only well after the acute illness has resolved.

The baseline recovery figures are broadly consistent. Spinale and colleagues’ review concluded that roughly 70% of patients recover fully, with the kidneys bearing the bulk of long-term damage and a smaller subset experiencing extrarenal sequelae, including brain injury.[1] Long-term pediatric data align closely: in the Austrian/German cohort followed by Rosales et al. for ten years, 66% of 138 STEC-HUS patients had fully recovered, while 34% retained some residual abnormality — most often reduced glomerular filtration, proteinuria, or hypertension, with persistent neurological symptoms in about 3%.[2] So while frank neurological sequelae affect only a small percentage at a decade out, they are a real and durable component of the residual-disease burden.

The scale of that risk is clearest in population-level work. Merrick et al.’s matched cohort from Wales (1990–2020) tracked neurological endpoints such as epilepsy and cognitive impairment alongside renal, cardiac, and other outcomes, and found that long-term complications were nearly twice as likely after STEC O157 infection and up to eight times as likely after STEC-HUS compared with unexposed controls — a gradient that led the authors to recommend monitoring for at least five years.[3] This reframes neurological sequelae not as rare curiosities but as part of an elevated, measurable long-term risk profile that warrants structured follow-up.

The adult experience adds a distinctive and clinically important wrinkle. Schuppner and colleagues followed 44 adults from the 2011 O104:H4 outbreak at 2, 7, and 19 months using neurological, neuropsychological, MRI, and EEG assessment. At 19 months only about 39% performed normally, and — most strikingly — roughly a quarter of patients showed a secondary decline in cognitive function that was unrelated to the severity of their acute illness.[4] This delayed deterioration means that an apparently good early recovery does not guarantee a durable one, and that acute-phase severity is an unreliable predictor of cognitive outcome.

The pattern of injury helps explain why recovery is so often partial rather than absent. STEC-HUS brain injury is fundamentally microvascular and Shiga-toxin/Gb3-mediated rather than the result of large-vessel occlusion, producing the characteristic symmetric and frequently reversible oedema in the basal ganglia, thalami, and brainstem rather than territorial infarction. Because the insult is diffuse and often partly reversible, many patients improve substantially, but the same microangiopathic mechanism can leave residual cognitive and seizure-related deficits. Tonkovic et al.’s comprehensive review situates these long-term, age-specific outcomes within that mechanistic framework,[5] and the broader concern is underscored by the observation that HUS mortality of around 5% rises toward 40% once the CNS is involved — marking neurological participation as the principal driver of poor outcomes.[6]

Taken together, the evidence supports a few practical conclusions: full neurological recovery is the most common outcome but is not universal; children and adults differ, with adults notably at risk of a delayed cognitive decline; and the elevated long-term risk justifies sustained, multidisciplinary follow-up rather than discharge once renal function stabilizes.

References

Pathophysiology and mechanism (microangiopathy)

1.  [Author byline to verify]. Deleterious consequences of Shiga toxin in the CNS. Microbiol Mol Biol Rev. 2026;90(1):e00301-25. doi:10.1128/mmbr.00301-25

Current mechanism review (clinical + experimental). STEC encephalopathy is described as a main predictor of death; Shiga toxin in circulation rapidly induces endothelial damage via the Gb3 receptor, with consistently reported brain inflammation.

2.  Goldstein J, Nuñez-Goluboay K, Pinto A. Therapeutic strategies to protect the central nervous system against Shiga toxin from enterohemorrhagic Escherichia coli. Curr Neuropharmacol. 2021;19(1):24-44.doi:10.2174/1570159X18666200220143001

Reviews CNS injury and neuroprotective strategies; notes HUS mortality ~5%, rising to ~40% with CNS involvement, plus the contribution of lipopolysaccharide-driven inflammation.

3.  Tonkovic U, Bogicevic M, Manzar A, Andrejic N, Sic A, Atanaskovic M, et al. Neurological manifestations of hemolytic uremic syndrome: a comprehensive review. Brain Sci. 2025;15(7):717. doi:10.3390/brainsci15070717

Primary recommended overview. Dedicated section on Shiga toxin–mediated endothelial injury and microangiopathy (Gb3/CD77 binding, microvesicle transport); spans pathophysiology, clinical spectrum, imaging, and age-specific outcomes for typical (STEC) and atypical HUS, covering both acute presentation and long-term risk.

4.  Kim M, Lee KS, Park JY, Kim CU, Jeong YJ, Lee MS. Shiga toxin induces apoptosis via ROS–caspase activation in human cerebral endothelial cell line hCMEC/D3 and astrocyte co-culture. J Microbiol Biotechnol. 2026;36:e2512.12006. doi:10.4014/jmb.2512.12006

In vitro model of cerebral microvascular endothelial (blood-brain barrier) injury: Stx1a/Stx2a induce MAPK/ER-stress and caspase-mediated apoptosis with tight-junction loss and increased paracellular permeability. Directly relevant to the cerebral microangiopathy mechanism.

Acute-phase illness: clinical presentation and neuroimaging

5.  Fodor TA, Schmook MT, Brücke C. Pearls & Oy-sters: neurologic involvement in Shiga toxin–associated hemolytic uremic syndrome. Neurology. 2024;103(9):e209881. doi:10.1212/WNL.0000000000209881

Concise clinical teaching article: the thrombotic microangiopathy mainly affects the small vessels of the kidneys and brain; acute signs include decreased consciousness, altered mental status, seizures, hyperreflexia.

6.  Nakamura T, Fujikawa H, Uenishi N. Brain magnetic resonance imaging findings of Shiga toxin-producing Escherichia coli hemolytic uremic syndrome-associated encephalopathy. JMA J. 2025;8(1):298-299.doi:10.31662/jmaj.2024-0201

Confirmed O157:H7 case; bilateral symmetric pontine and thalamic FLAIR changes; incidence of neurological involvement cited at 17–34%, with a near-symmetric distribution (basal ganglia, centrum semiovale, thalami, brainstem).

7.  Wengenroth M, Hoeltje J, Repenthin J, Meyer TN, Bonk F, Becker H, et al. Central nervous system involvement in adults with epidemic hemolytic uremic syndrome. AJNR Am J Neuroradiol. 2013;34(5):1016-1021. doi:10.3174/ajnr.A3336

MRI series of 11 adults from the 2011 Hamburg EHEC (O104:H4) outbreak; symmetric brainstem vasogenic edema, with no territorial ischemia, hemorrhage, or blood-brain barrier disruption (consistent with a microangiopathic process).

8.  [Author byline to verify]. Shiga toxin-producing Escherichia coli infection-related acute encephalopathy. Neurol Sci. 2025. doi:10.1007/s10072-025-08034-9

Adult (34-year-old) case of STEC-related acute encephalopathy without concurrent kidney involvement; seizures, cortical blindness, left hemiparesis.

9.  Ylinen E, Salmenlinna S, Halkilahti J, Jahnukainen T, Korhonen L, Virkkala T, et al. Hemolytic uremic syndrome caused by Shiga toxin–producing Escherichia coli in children: incidence, risk factors, and clinical outcome. Pediatr Nephrol. 2020;35(9):1749-1759. doi:10.1007/s00467-020-04560-0

Finnish nationwide pediatric cohort; O157 most common serogroup (66%); acute-phase outcomes and risk factors (age <3 y, stx2/stx2a), bridging to long-term prognosis.

Long-term complications and neurological sequelae

10.  Schuppner R, Maehlmann J, Dirks M, Worthmann H, Tryc AB, Sandorski K, et al. Neurological sequelae in adults after E coli O104:H4 infection-induced hemolytic-uremic syndrome. Medicine (Baltimore). 2016;95(6):e2337. doi:10.1097/MD.0000000000002337

Longitudinal follow-up of 44 adults from the 2011 outbreak at 2, 7, and 19 months (neurologic, neuropsychological, MRI, EEG). At 19 months only ~39% performed normally; a secondary decline of cognitive function occurred in ~1/4 of patients, unrelated to acute-phase severity.

11.  Rosales A, Kuppelwieser S, Giner T, Hofer J, Riedl Khursigara M, Orth-Höller D, et al. Outcome 10 years after Shiga toxin-producing E. coli (STEC)-associated hemolytic uremic syndrome: importance of long-term follow-up. Pediatr Nephrol. 2024;39(9):2739-2750. doi:10.1007/s00467-024-06355-z

Austrian/German pediatric cohort, 138 STEC-HUS patients at 10 years: 66% fully recovered; 34% had decreased GFR, proteinuria, hypertension, or neurological symptoms (3%).

12.  Merrick R, Song J, Fina L, Sawyer C, Jenkins C, King G, et al. Long-term health outcomes of Shiga toxin-producing Escherichia coli O157 (STEC O157) infection and STEC-associated haemolytic uraemic syndrome (STEC-HUS), Wales, 1990–2020. Pediatr Nephrol. 2025;40(7):2363-2374. doi:10.1007/s00467-024-06640-x

Population-based matched cohort (4:1) vs. unexposed comparators; outcomes (kidney, neurological [epilepsy, cognitive impairment], cardiac, GI, respiratory, endocrine) via Cox regression. Long-term complications nearly twice as likely after STEC O157 and up to eight times after STEC-HUS; recommends ≥5-year monitoring.

13.  Spinale JM, Ruebner RL, Copelovitch L, Kaplan BS. Long-term outcomes of Shiga toxin hemolytic uremic syndrome. Pediatr Nephrol. 2013;28(11):2097-2105. doi:10.1007/s00467-012-2383-6

Review of long-term sequelae: ~70% recover fully; kidneys bear most long-term damage, with a smaller number of extrarenal sequelae including brain injury.

Treatment of CNS involvement (acute management bearing on long-term outcome)

14.  Wildes DM, Harvey S, Costigan CS, Sweeney C, Twomey É, Awan A, et al. Eculizumab in STEC-HUS: a paradigm shift in the management of pediatric patients with neurological involvement. Pediatr Nephrol. 2024;39(3):935-940. doi:10.1007/s00467-023-06102-w

Retrospective series (n = 4 children) with CNS involvement (ataxia, altered mental status, visual symptoms, seizures, days 2–7) treated with eculizumab plus supportive care; rapid symptom resolution and complete kidney/neurological recovery at 12 months.

15.  Spagnol R, Alfisi A, Moi M, Bonvecchio I, Bertazza Partigiani N, Vidal E. Eculizumab in severe pediatric STEC-HUS and its impact on neurological prognosis—a systematic review and meta-analysis. Eur J Pediatr. 2025;184(6):360. doi:10.1007/s00431-025-06160-2

Systematic review/meta-analysis; notes neurological complications in 17–34% of affected children, arising from thrombotic events in the cerebral microvasculature, and evaluates eculizumab’s effect on neurological prognosis.

16.  Rosazza C, Cappellari AM, Gandini C, Scola E, Ardissino G. Steroid pulse therapy for severe central nervous system involvement in Shiga toxin-producing Escherichia coli-related hemolytic uremic syndrome. Case Rep Pediatr. 2021;2021:5587050. doi:10.1155/2021/5587050

Pediatric case (7-year-old) where severe CNS involvement (subacute encephalitis → coma) emerged as the TMA was resolving — attributed to reperfusion injury — and resolved without sequelae after high-dose IV steroids.

[1]Spinale JM, Ruebner RL, Copelovitch L, Kaplan BS. Long-term outcomes of Shiga toxin hemolytic uremic syndrome. Pediatr Nephrol. 2013;28(11):2097-2105. doi:10.1007/s00467-012-2383-6

[2]Rosales A, Kuppelwieser S, Giner T, et al. Outcome 10 years after Shiga toxin-producing E. coli (STEC)-associated hemolytic uremic syndrome: importance of long-term follow-up. Pediatr Nephrol. 2024;39(9):2739-2750. doi:10.1007/s00467-024-06355-z

[3]Merrick R, Song J, Fina L, et al. Long-term health outcomes of Shiga toxin-producing Escherichia coli O157 (STEC O157) infection and STEC-associated haemolytic uraemic syndrome (STEC-HUS), Wales, 1990–2020. Pediatr Nephrol. 2025;40(7):2363-2374. doi:10.1007/s00467-024-06640-x

[4]Schuppner R, Maehlmann J, Dirks M, et al. Neurological sequelae in adults after E. coli O104:H4 infection-induced hemolytic-uremic syndrome. Medicine (Baltimore). 2016;95(6):e2337. doi:10.1097/MD.0000000000002337

[5]Tonkovic U, Bogicevic M, Manzar A, et al. Neurological manifestations of hemolytic uremic syndrome: a comprehensive review. Brain Sci. 2025;15(7):717. doi:10.3390/brainsci15070717

[6]Goldstein J, Nuñez-Goluboay K, Pinto A. Therapeutic strategies to protect the central nervous system against Shiga toxin from enterohemorrhagic Escherichia coli. Curr Neuropharmacol. 2021;19(1):24-44. doi:10.2174/1570159X18666200220143001

What is infant botulism?

Infant botulism is a rare but serious illness affecting babies under one year of age. It is caused by the toxin made by the bacterium Clostridium botulinum (and, rarely, related Clostridium species). It develops when a baby swallows bacterial spores that take hold in the immature intestine, grow, and release a powerful nerve toxin directly inside the gut. [2] [8]

This is the crucial difference from classic foodborne botulism. In foodborne botulism a person swallows toxin that has already formed in contaminated food. In infant botulism the baby swallows spores, and the toxin is produced inside the body. A baby’s gut is uniquely vulnerable because it has not yet developed the protective “normal” bacteria that keep C. botulinum in check in older children and adults. [4] [2]

About 95% of cases occur in infants under six months, with a median age around three months. Most cases are caused by toxin type A or type B. [5] [31]

Causes and how infants are exposed

The spores of C. botulinum are everywhere in the environment — in soil and dust and carried in the air. Because they are so common, in many individual cases the exact source is never identified. Two sources are firmly recognized: environmental soil and dust, and honey. [5] [7]

Recognized and suspected sources

  • Soil and household dust. The leading recognized source; spores can be inhaled or swallowed. [7] [9]
  • Honey. The only well-established dietary risk factor. Babies under 12 months should never be given honey or honey-containing foods. [30] [2]
  • Other infant foods (occasional). Surveys have recovered C. botulinum spores from dry cereals, fruits, vegetables and, rarely, powdered formula — though a causal link to illness is usually not established. [33]
  • Corn syrup (unproven). Sometimes suggested, but never clearly linked. [6]

What the toxin does

Once produced in the gut, the toxin is absorbed and binds — essentially irreversibly — to the junctions between nerve and muscle, blocking release of acetylcholine, the signal that tells muscles to contract. The result is a descending, “floppy” paralysis. Recovery requires the body to grow new nerve endings, which is why the illness can be prolonged. [7] [1]

Signs and symptoms

Unlike botulism in adults, infant botulism usually comes on gradually and worsens over hours to days. Constipation is often the first sign — present at the outset in about 90% of cases — followed by weakness that begins in the face and head and spreads downward. [2] [8]

Common signs include:

  • Constipation, often for several days
  • Poor feeding; weak sucking and swallowing
  • A weak, altered or soft cry
  • Drooping eyelids (ptosis); sluggish pupils; reduced eye movement
  • Generalized weakness and low muscle tone — the classic “floppy baby”
  • Lethargy and a diminished gag reflex
  • Breathing difficulty, progressing in severe cases to respiratory failure — the most dangerous complication

Because these early signs mimic more common conditions, a high index of suspicion matters. A baby with poor feeding, a weak cry and increasing floppiness — especially with any breathing difficulty — needs urgent evaluation. [3] [4]

Diagnosis

Diagnosis is primarily clinical, confirmed in the laboratory — most reliably by testing the baby’s stool for the toxin or organism (the most accurate test, though results can take about a week). Treatment should not wait for laboratory confirmation: because the antitoxin works best given early, clinicians treat as soon as infant botulism is reasonably suspected. [1] [4] [2]

Treatment

Infant botulism is highly treatable, and the great majority of babies recover fully. Care rests on two pillars: specific antitoxin and supportive care. [1] [3]

Antitoxin (BabyBIG)

The specific treatment is human botulism immune globulin, given intravenously and known by the brand name BabyBIG(BIG-IV). First-line since 2003, it neutralizes toxin still circulating in the blood and, given early, shortens the duration of ventilation, hospital stay and tube feeding. In the United States it is obtained from the Infant Botulism Treatment and Prevention Program (IBTPP) in California. [1] [2] [3] [34]

Supportive care, and what is avoided

Supportive care may include intensive-care monitoring, respiratory support up to mechanical ventilation, and nutritional (often tube) feeding until safe feeding returns. Antibiotics are generally not given for the botulism itself — killing the bacteria can release more toxin — and aminoglycoside antibiotics in particular are avoided because they can worsen the weakness. [3] [2]

Prognosis and long-term complications

The outlook is excellent: reported survival is roughly 98–100% and the case-fatality rate for hospitalized infants is well under 1%. Most babies recover fully over several weeks to months, usually without lasting effects. [1] [3]

  • Respiratory failure, apnea, aspiration — the main reason a ventilator may be needed (on average ~3 weeks when required). [3] [7]
  • Prolonged feeding difficulty — safe oral feeding can take many weeks to return. [7]
  • Autonomic effects such as urinary retention; and secondary infections (e.g. pneumonia, ear infection) acquired in hospital. [3] [8]
  • Prolonged general weakness during recovery. [9]

Lasting neurological after-effects are seldom seen, and most infants regain normal strength and development. Severe long-term harm is rare. Follow-up with neurology and physiotherapy is advisable after a severe episode. [7] [9] [10]

Why infant botulism outbreaks are so rare

A key point sets infant botulism apart from foodborne botulism. In a foodborne outbreak, several people can be traced to one shared contaminated food. Infant botulism, by contrast, comes from spores picked up from the wider environment, so it almost always occurs as isolated, sporadic single cases — genuine outbreaks are extremely uncommon. [33]

That is why the 2025 infant-formula outbreak is so notable: it is the rare instance in which many infant botulism cases were linked to a single manufactured product. According to the FDA it is the first known botulism outbreak tied to infant formula anywhere in the world since the condition was first described about fifty years ago. [12] [16]

The 2025 infant-formula outbreak (ByHeart)

In October 2025, California’s Infant Botulism Treatment and Prevention Program noticed that three infants with suspected infant botulism had all been fed the same powdered formula. State officials alerted the CDC, and a multistate investigation followed. The case total was revised as the investigation widened, the case definition was expanded to capture earlier illnesses, and some early-suspected cases were later reclassified. [14] [13]

ProductByHeart Whole Nutrition powdered infant formula (cans and single-serve “Anywhere Pack”); on sale nationwide since March 2022
Cases48 infants (28 confirmed + 20 probable); revised down from an interim 51 after 3 cases were reclassified
Geography17 U.S. states
Hospitalized / deathsAll 48 hospitalized and treated with BabyBIG; no deaths reported
Illness onset rangeDecember 24, 2023 – November 29, 2025 (most clustered Aug–Nov 2025)
Toxin typeBoth type A and type B among outbreak cases. The formula/milk-powder strain and its genetically-matched cases were type A (subtype A1); the unusual appearance of type A where type B normally dominates was a key outbreak signal
Laboratory evidenceC. botulinum type A detected in finished formula (multiple lots) and in an opened can from a patient’s home
Likely root causeGenetic (whole-genome) match between the formula organism and C. botulinum in organic whole milk powder from a supplier; full source investigation ongoing
RecallTwo lots recalled 8 Nov 2025; all products recalled 11 Nov 2025
SignificancePer the FDA, the first known botulism outbreak tied to infant formula anywhere in the world since infant botulism was first described ~50 years ago

The milk-powder link

The detail most relevant to milk and milk powder is the suspected root cause. Investigators detected C. botulinum type A in finished ByHeart formula, and the New York State Wadsworth Center reported that whole-genome sequencing of the organism in a sample of organic whole milk powder — collected at a processor supplying ByHeart — matched the strain in the finished formula and in a sick infant. Contaminated milk powder used as an ingredient is the leading explanation for how spores entered the product. The full source investigation remained ongoing into 2026. [11] [15]

Because spores can be unevenly distributed through a batch of powder, not every baby who drank the formula became ill. All infants under one year are nonetheless considered at risk, which is why every ByHeart product was recalled and parents were urged to stop using it and discard it. [13] [11]

Why formula — normally so safe — was vulnerable here

Powdered infant formula is not sterile. The pathogens usually watched for (such as Cronobacter and Salmonella) do not form spores and are killed by pasteurization. Spore-forming bacteria like C. botulinum are different: their heat-resistant spores can survive normal processing. The 2025 event prompted the FDA and international food-standards bodies to begin a formal risk assessment of spore-forming organisms in powdered infant formula. [16] [12]

Honey: the principal dietary risk factor

Among foods given to infants, honey is the one with a well-established link to infant botulism. Honey samples across many countries have tested positive for C. botulinum spores, and numerous studies tie honey consumption to cases. This is why public-health bodies advise that no honey — and no honey-containing foods — be given to babies under 12 months. [30] [32]

Honey is not a bee disease; rather, spores are an occasional environmental contaminant carried into the hive on dust or water. The simple, effective preventive step is to keep all honey away from infants until their first birthday. [32]

Infant botulism by the numbers

A snapshot of recent United States surveillance — the country that diagnoses the great majority of the world’s reported cases. [31] [7]

MeasureFigure (United States, recent years)
Most common form of botulism in the USInfant botulism — roughly 60–70% of all reported cases
Infant cases reported in 2021181 (highest annual count since the disease was characterized in 1976)
Typical ageUnder 12 months; ~95% under 6 months; median age ~3 months
Toxin types (2021)Type B ~59%, Type A ~40% (rare Ba/Bf)
Geography~90% of the world’s reported cases are diagnosed in the US; California and Pennsylvania report the most
Deaths (2021)None reported among US infant cases

What the literature shows

  • Until 2025, C. botulinum was essentially never found in commercial dried milk or formula. Dedicated surveys — including nonfat dry milk, evaporated milk and canned formula — came back negative. [21] [22]
  • The 2025 ByHeart outbreak is the turning point and the strongest milk-powder evidence to date (§8): whole-genome sequencing tied type A C. botulinum in organic whole milk powder to the finished formula and to a sick infant’s own strain. [11] [17]
  • Earlier formula links were single cases with caveats: a 2001 UK case and a 2023 Chinese investigation each recovered the organism from formula milk powder, but causation was never firmly proven. [18] [19] [20]
  • Fluid/raw milk can carry the organism and toxin during bovine botulism, and spores survive pasteurization — though pasteurization strongly degrades pre-formed toxin. [23] [24] [25]

Infant formula and dried (powdered) milk

The category most relevant to the milk/formula question, and the one transformed by the 2025 outbreak (covered in depth in §8). [11]

Source (study, year, place)What was testedKey finding
FDA / CDC outbreak investigation, 2025–26 (USA, multistate)Finished ByHeart formula; ingredient samples incl. organic whole milk powder; one clinical isolateOUTBREAK (detailed in §8). Type A C. botulinum in finished formula was whole-genome matched to organic whole milk powder AND to a sick infant’s strain — a closed genetic chain. Final total: 48 cases (28 confirmed, 20 probable) in 17 states, no deaths; cases spanned type A and type B. [11] [13] [14]
Harris et al., Front. Microbiol., 2026 (USA)Unopened formula containers and “base powder” (bulk formula before packaging)C. botulinum found in both finished product and base powder, with genetic identity between one finished lot and a base powder. Notably present even where the usual indicator (sulfite-reducing clostridia) was non-detectable — standard screening would have missed it. [17]
Brett et al., J. Med. Microbiol., 2005 (UK, 2001 case)14 home foods of a 5-month-old with infant botulismType B C. botulinum recovered from an OPENED can of infant formula milk powder; two isolates were DNA-fingerprint matches to the baby’s own strain. A unique type B was also found in an UNOPENED can of the same batch, and type A in opened dried rice pudding. [18]
Johnson et al., J. Clin. Microbiol., 2005 (UK, same case)Re-analysis of the case strains (PFGE), incl. retained sealed cansConfirmed the infant’s strain matched organism in the opened formula but judged the unopened brand an unlikely original source. Net verdict: a possible link, never definitively proven; environmental exposure could not be excluded. [19]
Luo et al., Chin. J. Food Hygiene, 2023 (China)30 batches of infant formula milk powder from a company tied to an infant botulism caseToxin NOT detected directly; 4 batches caused botulism signs in mice, but C. botulinum was isolated from only 1 batch (group I, type B, subtype B2). Authors stressed that isolation alone is not proof. [20]

Dedicated surveys of dried milk and infant foods

Before 2025, researchers periodically screened infant foods — including dried and evaporated milk and formula — specifically for C. botulinum. The organism was not found, even when related (harmless-to-adults) clostridia were. An earlier survey by Kautter and colleagues (1982) reached the same broad conclusion. This decades-long baseline is why the 2025 finding was so surprising. [33]

Source (study, year, place)What was testedKey finding
Guilfoyle & Yager, J. AOAC, 1983 (USA, NYC)236 infant-food samples incl. nonfat dry milk, evaporated milk, canned formula, canned baby food, honey, dry cerealNONE of the 236 samples contained C. botulinum spores. Concluded incidence in these commercial products is not widespread.[21]
Barash et al., J. Pediatr., 2010 (USA, California)30 formula samples used by infant-botulism patients + 9 market-bought unopened formulasClostridial spores in 17% of patient-used and 78% of market formulas — but none were C. botulinum (related Clostridium species). Formula can carry clostridia, but not C. botulinum itself. [22]

Fluid and raw milk

Most evidence here comes from veterinary medicine. When a dairy herd suffers botulism (often from contaminated silage), the organism and even pre-formed toxin can appear in raw milk; because spores survive pasteurization, this is treated as a wider food-safety question, even though milk-borne human botulism is rare. [24]

Source (study, year, place)What was testedKey finding
Böhnel et al., Vet. Record, 2013 (Germany)99 milk + 51 udder samples from dairy cows with suspected bovine botulismBotulinum toxin in milk from 8.1% of farms and C. botulinum in 5.4%; udder-tissue toxin in 19.6%. The organism/toxin can appear in raw cow’s milk during herd botulism. [23]
Lindström et al., Crit. Rev. Food Sci. Nutr., 2010 (review)Literature review across the dairy chainContaminated silage can let C. botulinumreach raw milk; pasteurization does NOT destroy spores, and many dairy products allow growth/toxin formation. Large dairy-linked outbreaks are rare but documented.[24]
Appl. Environ. Microbiol., 2010 (“Botulinum toxin in milk”)Raw milk spiked with purified BoNT/A and BoNT/B, then pasteurizedStandard pasteurization (72°C / 15 s) inactivated ≥99.99% of free BoNT/A and /B — reassuring that pasteurization strongly degrades pre-formed toxin in milk. [25]
Rasooly & Do, J. Agric. Food Chem., 2010 (USA)Heat stability of BoNT type B in milkReported BoNT type B as relatively heat-stable in milk and not fully inactivated by pasteurization — a more cautious counterpoint; outcome differs by toxin type, form and conditions. [26]
Glass et al., J. Food Prot., 1999 (USA)Pasteurized 2% milk inoculated with C. botulinum spores, stored warmSpore-inoculated milk left at 21°C produced botulinal toxin within days — surviving spores can germinate and make toxin if milk is temperature-abused. [27]

Two of these entries appear to disagree on whether pasteurization neutralizes botulinum toxin in milk. The tension is explainable: the outcome depends on toxin serotype, whether it is free toxin or in its protective complex, and the exact temperature/time. The firmer points: pasteurization does not kill C. botulinum spores, and surviving spores can later make toxin if milk is left warm. [25] [26] [27]

Other milk products

Outside infant feeding, soft high-moisture dairy products show what C. botulinum can do in a milk matrix when conditions allow. Italian mascarpone is the clearest example, with both a survey detecting spores and a real outbreak. [28]

Source (study, year, place)What was testedKey finding
Franciosa et al., J. Food Prot., 1999 (Italy)1,017 mascarpone cheese samples + 260 other dairy products32.5% of mascarpone carried botulinal spores; 0.8% of samples from an outbreak-linked plant also held type A toxin. Of other dairy products, 2.7% were spore-positive. (Soft high-moisture milk product — not infant food, but the clearest evidence of C. botulinum in a dairy item.) [28]
Aureli et al., Eur. J. Epidemiol., 2000 (Italy)Investigation of a foodborne botulism outbreakA 1996 Italian outbreak traced to a dessert made with mascarpone cream cheese (type A) — the real-world illness behind the mascarpone surveys above. [29]

How to read this evidence

  • Spores vs. toxin. A spore is the dormant, heat-resistant form; toxin is the harmful product made when spores germinate. Spores in milk powder signal a colonization risk for an infant’s gut; toxin signals immediate danger. [24]
  • Detection vs. causation. Recovering the organism from an opened container (as in the UK and Chinese cases) cannot rule out contamination after opening; the strongest cases use genetic matching across food, product and patient — exactly what set 2025 apart. [18] [11]
  • Detection methods matter. The 2026 laboratory paper showed C. botulinum can be present even when routine indicator organisms are not — so ordinary screening can miss it. [17]
  • Rarity remains the headline. Across decades of surveys, C. botulinum in commercial dried milk or formula was vanishingly rare until 2025; the dairy-cow and cheese findings are real but specialized circumstances. [21] [23] [28]

References

[1] Cleveland Clinic. Infant Botulism: Symptoms, Causes & Treatment. https://my.clevelandclinic.org/health/diseases/infant-botulism

[2] Merck Manual, Professional Edition. Infant Botulism. https://www.merckmanuals.com/professional/infectious-diseases/anaerobic-bacteria/infant-botulism

[3] StatPearls (NCBI Bookshelf). Infantile Botulism. https://www.ncbi.nlm.nih.gov/books/NBK493178/

[4] Children’s Hospital Los Angeles. What to Know About Infant Botulism. https://www.chla.org/blog/advice-experts/what-know-about-infant-botulism

[5] American Academy of Pediatrics – HealthyChildren.org. Botulism. https://www.healthychildren.org/English/health-issues/conditions/infections/Pages/Botulism.aspx

[6] Nemours KidsHealth. Infant Botulism. https://kidshealth.org/en/parents/botulism.html

[7] American Family Physician (AAFP). Infant Botulism. https://www.aafp.org/pubs/afp/issues/2002/0401/p1388.html

[8] MedLink Neurology. Infant botulism. https://www.medlink.com/articles/infant-botulism

[9] Public Health Agency of Canada. Clostridium botulinum – Pathogen Safety Data Sheet. https://www.canada.ca/en/public-health/services/laboratory-biosafety-biosecurity/pathogen-safety-data-sheets-risk-assessment/clostridium-botulinum.html

[10] Botulism Sequelae: A Systematic Review (PMC). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12798721/

[11] U.S. FDA. Outbreak Investigation of Infant Botulism: Infant Formula (November 2025) — ingredient & whole-milk-powder WGS results. https://www.fda.gov/food/outbreaks-foodborne-illness/outbreak-investigation-infant-botulism-infant-formula-november-2025

[12] U.S. FDA. FDA’s Actions to Respond to Clostridium botulinum Illnesses Associated with Consumption of Powdered Infant Formula. https://www.fda.gov/food/outbreaks-foodborne-illness/fdas-actions-respond-clostridium-botulinum-illnesses-associated-consumption-powdered-infant-formula

[13] CDC. Investigation Update: Infant Botulism Outbreak, November 2025 (final case data, 4 Mar 2026). https://www.cdc.gov/botulism/outbreaks-investigations/infant-formula-nov-2025/investigation.html

[14] Khouri JM, et al. Multistate Infant Botulism Outbreak Associated with Powdered Infant Formula. NEJM Evidence / CIDRAP Public Health Alert (2026) — toxin types A and B. https://evidence.nejm.org/doi/full/10.1056/EVIDpha2600020

[15] New York State Dept. of Health, Wadsworth Center. Biodefense Laboratory Identifies Contamination Source in Multistate Infant Botulism Outbreak (2026). https://www.wadsworth.org/news/wadsworth-center-biodefense-laboratory-identifies-contamination-source-multistate-infant

[16] Contemporary Pediatrics. Infant botulism outbreak linked to powdered formula under FDA investigation. https://www.contemporarypediatrics.com/view/infant-botulism-outbreak-linked-to-powdered-formula-under-fda-investigation

[17] Harris RA, et al. Detection and characterization of Clostridium botulinum isolated from powdered infant formula. Frontiers in Microbiology (2026). https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2026.1800624/abstract

[18] Brett MM, et al. A case of infant botulism with a possible link to infant formula milk powder. J Med Microbiol (2005). https://pubmed.ncbi.nlm.nih.gov/16014431/

[19] Johnson EA, et al. Characterization of Clostridium botulinum Strains Associated with an Infant Botulism Case in the United Kingdom. J Clin Microbiol 2005;43(6):2602-7. https://pmc.ncbi.nlm.nih.gov/articles/PMC1151885

[20] Luo H, et al. Isolation and typing of Clostridium botulinum from milk powder of an enterprise associated with a case of infant botulism. Chinese Journal of Food Hygiene 2023;35(10):1475-81. https://doaj.org/article/30d4cb5e20434fad9a47a75d7afcd132

[21] Guilfoyle DE, Yager JF. Survey of infant foods for Clostridium botulinum spores. J Assoc Off Anal Chem 1983;66(5):1302-4. https://pubmed.ncbi.nlm.nih.gov/6355058/

[22] Barash JR, et al. Clostridial spores in powdered infant formula. The Journal of Pediatrics (2010). https://www.jpeds.com/article/S0022-3476(10)00070-3/fulltext

[23] Böhnel H, et al. Presence of Clostridium botulinum and botulinum toxin in milk and udder tissue of dairy cows with suspected botulism. Vet Rec (2013). https://pubmed.ncbi.nlm.nih.gov/23585115/

[24] Lindström M, et al. Clostridium botulinum in cattle and dairy products. Crit Rev Food Sci Nutr 2010;50(4):281-304. https://pubmed.ncbi.nlm.nih.gov/20301016/

[25] The Case of Botulinum Toxin in Milk: Experimental Data. Applied and Environmental Microbiology 2010;76(10):3293. https://aem.asm.org/content/76/10/3293.full

[26] Rasooly R, Do PM. Clostridium botulinum neurotoxin type B is heat-stable in milk and not inactivated by pasteurization. J Agric Food Chem 2010;58:12557-61. https://pubmed.ncbi.nlm.nih.gov/21038876/

[27] Glass KA, et al. Toxin production by Clostridium botulinum in pasteurized milk treated with carbon dioxide. J Food Prot (1999). https://pubmed.ncbi.nlm.nih.gov/10456739/

[28] Franciosa G, et al. Clostridium botulinum spores and toxin in mascarpone cheese and other milk products. J Food Prot 1999;62(8):867-71. https://pubmed.ncbi.nlm.nih.gov/10456738/

[29] Aureli P, et al. An outbreak in Italy of botulism associated with a dessert made with mascarpone cream cheese. Eur J Epidemiol 2000;16(10):913-8. https://pubmed.ncbi.nlm.nih.gov/11338122/

[30] Association between honey consumption and infant botulism (PubMed). https://pubmed.ncbi.nlm.nih.gov/12432974/

[31] CDC. National Botulism Surveillance Summary, 2021. https://www.cdc.gov/botulism/php/national-botulism-surveillance/2021.html

[32] UC Master Food Preserver Program. Infants and Honey (September 2025). https://ucanr.edu/program/uc-master-food-preserver-program/article/whats-buzz-infants-and-honey-september-2025

[33] Harris RA, Dabritz HA. Infant Botulism: In Search of Clostridium botulinum Spores (review). Curr Microbiol 2024;81:306. https://link.springer.com/article/10.1007/s00284-024-03828-0

[34] CDC / California Dept. of Public Health (CDPH). Infant Botulism Treatment and Prevention Program & outbreak notice. https://www.cdph.ca.gov/Programs/OPA/Pages/NR25-017.aspx

Since 2000, the U.S. Centers for Disease Control and Prevention (CDC) and the U.S. Food and Drug Administration (FDA), working with state and local partners, have publicly reported a series of multistate Escherichia coli outbreaks linked to romaine lettuce (and, in several cases, to leafy-green products in which romaine was the implicated or strongly suspected component). The below summarizes each of those publicly reported outbreaks in chronological order, with primary-source citations to the relevant CDC and FDA notices provided in the footnotes.

Context: The Pre-2010 Period and Why the Catalog Begins in 2010. Although this compilation addresses the period since 2000, the earliest outbreaks the CDC and FDA reported as standalone, romaine-specific investigations date to 2010 (O145 / Freshway) and 2011 (O157:H7 / St. Louis). Before roughly 2010, the agencies did not routinely publish individual produce-outbreak notices in the web format adopted later; earlier produce events are generally captured in retrospective CDC surveillance reviews — for example, Rangel et al., “Epidemiology of E. coli O157:H7 Outbreaks, United States, 1982–2002” (Emerg. Infect. Dis., 2005), and Heiman et al., “E. coli O157 Outbreaks in the United States, 2003–2012” (Emerg. Infect. Dis., 2015) — rather than as named public outbreak reports. The FDA frames the modern record as 28 STEC outbreaks with a confirmed or suspected link to leafy greens between 2009 and 2017.

Just as important, the prominent, widely publicized leafy-green outbreaks of the 2000s were attributed to greens other than romaine, so they do not belong on a romaine-specific list. The two most significant are noted here for completeness and explicitly distinguished:

September 2006 — Fresh bagged spinach (not romaine).  An E. coli O157:H7 outbreak tied to fresh spinach caused roughly 205 illnesses and three deaths across about 26 states; the probable origin was a California cattle ranch adjacent to spinach fields. FDA later treated this as the benchmark large produce outbreak that the 2018 Yuma romaine outbreak surpassed (FDA Environmental Assessment).

December 2006 — Shredded iceberg lettuce (not romaine).  Green onions were suspected first, but a CDC case-control investigation pointed to lettuce, and FDA identified the vehicle as shredded iceberg lettuce served at Taco Bell restaurants in the northeastern U.S. (about 71 illnesses). A concurrent Taco John’s outbreak in Iowa and Minnesota — a genetically different strain — was likewise traced to iceberg lettuce (CIDRAP; FDA Q&A).

Romaine and other lettuces did cause E. coli outbreaks before 2010 — including a 1995 Montana leaf-lettuce outbreak and assorted restaurant clusters — but those were generally recorded in aggregate surveillance rather than as romaine-specific public outbreak reports. Accordingly, the catalog below is believed to be complete for publicly reported, romaine-attributed CDC/FDA outbreaks since 2000, and it begins in 2010 for that reason.

Publicly Reported CDC/FDA Romaine Lettuce E. coli Outbreaks Since 2010

  • 2010 — E. coli O145, Shredded Romaine Lettuce (Freshway Foods)

Pathogen:  Shiga toxin-producing E. coli O145

Vehicle:  Shredded romaine lettuce processed by Freshway Foods (Sidney, OH); an unopened bag tested positive for the outbreak strain

Scope:  31 cases (27 confirmed, 4 probable) across 5 states (MI, OH, TN, PA, NY)

Outcomes:  14 hospitalized (45%); 3 HUS; no deaths

Notes:  Documented as the first U.S. foodborne outbreak of STEC O145; CDC supported the investigation, FDA led the traceback and recall

Source(s):[1]

  • 2011 — E. coli O157:H7, Romaine Lettuce (St. Louis / Schnucks salad bars)

Pathogen:  E. coli O157:H7

Vehicle:  Romaine lettuce (illnesses linked to salad bars; traced to a single processor/grower)

Scope:  58 cases across 9–10 states (concentrated in Missouri)

Outcomes:  33 hospitalized (67%); 3 HUS; no deaths

Notes:  CDC posted a final outbreak summary in March 2012 identifying romaine lettuce as the source

Source(s):[2]

  • 2017–2018 — E. coli O157:H7, Leafy Greens (U.S.) / Romaine Lettuce (Canada)

Pathogen:  E. coli O157:H7

Vehicle:  Leafy greens in the U.S.; the Public Health Agency of Canada identified romaine lettuce as the source of the genetically linked Canadian outbreak

Scope:  25 U.S. cases across 15 states

Outcomes:  9 hospitalized; 2 HUS; 1 death (California)

Notes:  CDC did not name a specific food in the U.S. because no common source was identified before the short shelf-life product was gone

Source(s):[3]

  • Spring 2018 — E. coli O157:H7, Romaine Lettuce (Yuma, AZ growing region)

Pathogen:  E. coli O157:H7

Vehicle:  Romaine lettuce from the Yuma, Arizona growing region; the strain was found in Yuma-region canal water

Scope:  210 cases across 36 states

Outcomes:  96 hospitalized; 27 HUS; 5 deaths (AR, CA, MN ×2, NY)

Notes:  The largest U.S. E. coli O157:H7 outbreak since the 2006 spinach outbreak

Source(s):[4]

  • Fall 2018 — E. coli O157:H7, Romaine Lettuce (Central Coastal CA / Santa Maria)

Pathogen:  E. coli O157:H7

Vehicle:  Romaine lettuce from the Central Coastal growing regions of northern/central California; strain found in an agricultural water reservoir at an Adam Bros. Farming ranch

Scope:  62 cases across 16 states and the District of Columbia

Outcomes:  25 hospitalized; 2 HUS; no deaths

Notes:  Prompted a nationwide consumer warning later narrowed to specific California counties; coincided with the rollout of harvest-region labeling

Source(s):[5]

  • October 2019 — E. coli O157:H7, Possibly Linked to Romaine Lettuce (FDA disclosure)

Pathogen:  E. coli O157:H7

Vehicle:  Romaine lettuce (likely source); product past shelf life by the time the source was identified

Scope:  23 illnesses

Outcomes:  No deaths reported

Notes:  Disclosed by FDA after the fact (Oct. 31, 2019); distinct from the larger November 2019 Salinas Valley outbreak

Source(s):[6]

  • November 2019 — E. coli O157:H7, Romaine Lettuce (Salinas Valley, CA)

Pathogen:  E. coli O157:H7 (same strain as 2017 leafy-greens and 2018 romaine outbreaks)

Vehicle:  Romaine lettuce from the Salinas Valley growing region; strain detected in romaine products and the surrounding environment

Scope:  167 cases across 27 states (CDC final); a related analysis reported 172 across 28 states

Outcomes:  85 hospitalized; 15 HUS; no deaths

Notes:  CDC advised consumers not to eat Salinas-grown romaine during the investigation

Source(s):[7]

  • Fall 2020 — E. coli O157:H7, Leafy Greens (romaine-related strain)

Pathogen:  E. coli O157:H7 (genetically related to the fall 2019 romaine strain)

Vehicle:  Leafy greens; a specific type or brand was not confirmed. Romaine was among the leafy greens commonly reported by cases

Scope:  40 cases across 19 states

Outcomes:  20 hospitalized; 4 HUS; no deaths

Notes:  Outbreak strain matched in cattle feces collected uphill from a leafy-greens growing area

Source(s):[8]

  • Fall 2020 — E. coli O157:H7, Romaine Lettuce (Tanimura & Antle; “Unknown Source 3”)

Pathogen:  E. coli O157:H7

Vehicle:  Romaine lettuce; the outbreak strain was identified in a single-head package of Tanimura & Antle romaine, recalled Nov. 6, 2020

Scope:  18 cases across 9 states

Outcomes:  6 hospitalized; no HUS; no deaths

Notes:  Investigators could not confirm that any ill person had eaten the specific recalled product

Source(s):[9]

  • The November 2024 Outbreak: Reported Privately, Not Publicly

A multistate E. coli O157:H7 outbreak coded by PulseNet on November 25, 2024, as 2411MOEXH-2 and reported to the National Outbreak Reporting System (NORS) as NORS ID 511856. The investigation began after Missouri public-health colleagues identified illnesses linked to events served by the same Missouri-based caterer between November 6 and 8, 2024.

As of January 15, 2025, the investigation included 89 cases across 15 states (AR, CO, IL, IN, KS, KY, MO, MT, ND, NE, OH, PA, SD, TN, WI), with the heaviest concentration in Missouri (50). Reported onset dates ran from November 4 to November 30, 2024. Among cases with available outcome data, 36 (49%) were hospitalized, there were 7 reported cases of HUS, and one death was attributed to the outbreak. Seven subclusters — including three Missouri catered events, a secondary school, two restaurants, and an event run by a different caterer — were tied together by a salad containing an iceberg/romaine lettuce blend.

Epidemiologic and traceback evidence pointed to romaine lettuce. Of cases who could recall the type of leafy green consumed, 88% reported romaine — well above the roughly 49% background rate in the FoodNet population survey. A three-leg traceback converged on a single processor sourcing romaine from a single grower, and four implicated lots tied to a common ranch. The CDC closed the investigation on January 15, 2025, as an outbreak with a confirmed vehicle of romaine lettuce. According to the memorandum and the FDA records cited within it, the processor of the implicated lettuce was Taylor Farms of Salinas, California, and the grower was Anthony Costa and Sons, LLC.[10]

Why This Outbreak Is Treated as “Not Reported”

Unlike the outbreaks listed above — each of which received a public CDC investigation notice, periodic case-count updates, and a final report — this outbreak never received a comparable public communication. The FDA logged a brief, source-anonymous entry in its CORE outbreak investigation table (carried under reference #1280, described only as an “iceberg and romaine lettuce blend served at catering events, restaurants, and a school”), with case counts that grew over time from 69 to 75 to 86 across as many as 12 named-only-by-count states. The agency then closed the investigation in February 2025 without issuing a final outbreak report, without naming the implicated supplier or grower, and without identifying the affected venues.[11]

Reporting by NBC News, working with attorney Bill Marler, subsequently established that the closed investigation corresponded to the 89-case, one-death outbreak — information the public had not been given through any ordinary CDC or FDA outbreak announcement.[12]

Bill Marler’s Position

Marler — a food-safety lawyer whose firm represents victims of the outbreak — has argued that this outbreak should have been publicly reported in the same manner as the earlier romaine outbreaks. After obtaining the FDA’s investigation records, his firm’s on-staff epidemiologist concluded that the common link among genetically matched cases across multiple states was consumption of Taylor Farms romaine lettuce during the outbreak period. Marler contends that, had the agencies completed their normal process, “they would have publicized the same conclusion.”[13]

He has tied the silence to reduced federal public-health capacity, asserting that informing the public and identifying the cause of outbreaks had ceased to be a priority, and stating that his firm would “step up to inform and protect the public” in the agencies’ place. To press the point publicly, Marler Clark filed several lawsuits naming Taylor Farms as the source of the romaine lettuce and amended earlier suits accordingly. Taylor Farms denies that its product was the source, stating that its raw and finished-product testing found no evidence of contamination.[14]

The core of Marler’s criticism is one of transparency rather than epidemiology: the investigative work was done, a vehicle was identified internally, and a person died — yet, in his view, the public was deprived of the timely, named, and complete outbreak notice that the CDC and FDA had routinely issued for every comparable romaine outbreak in the preceding fifteen years.

Here is a bit more detail:

On November 25, 2024, PulseNet coded an outbreak of E. coli O157:H7, given the code 2411MOEXH-2. This outbreak was coded following notification from colleagues in Missouri after they identified and investigated multiple illnesses linked to events catered by the same Missouri-based caterer. These events occurred between November 6 and November 8. All events included the same menu items with a few modifications. A case was defined as a person with an E. coli O157:H7 infection, with an isolate related to the outbreak strain within 0-4 alleles by cgMLST, and an isolation date ranging from November 7 to December 1, 2024. Specimen were collected from November 7, 2024, to December 1, 2024, all related within 0-4 alleles by cgMLST.  

This outbreak was related to six historical outbreak investigations: 2302MLEXH-1, 2210MLEXH-3, 2210MLEXH-2, 2209MLEXH-1, 2112MLEXH-1, and 2106CAEXH-1. The only vehicle identified was for 2112MLEXH-1, which was closed with a confirmed vehicle of organic power greens. The NCBI tree for this outbreak strain, 2411MOEXH-2, included numerous nonclinical beef isolates.

Aa of January 15, 2025, this investigation included 89 cases from 15 states: AR (2), CO (1), IL (7), IN (8), KS (1), KY (1), MO (50), MT (1), ND (2), NE (3), OH (8), PA (1), SD (1), TN (1), WI (2). Reported onset dates ranged from November 4 to November 30, 2024 (n=83). Ages ranged from 4 to 90 years, with a median age of 24. Sixty of 88 cases (68%) were female. Outcome information was available for 74 cases, of which 36 (49%) were hospitalized. There were seven reported cases of HUS and one death attributed to this outbreak.

In total, seven subclusters were identified across the multistate outbreak. These included three Missouri catered events, a secondary school, two different restaurants, and an event catered by a different caterer. In the Missouri investigations, Missouri colleagues conducted a retrospective cohort study at two of the events and found that salads were the only statistically significant menu item across both events. Salads contained an iceberg/romaine lettuce blend, carrots, purple cabbage, onions, canned pimento, canned artichokes, parmesan cheese, and a house made salad dressing. Salads were the common link across all seven subclusters, and cases in all subclusters ate an iceberg/romaine lettuce blend. The CDC deployed a focused questionnaire on November 26, 2024, and 27 questionnaires were returned. Epi information was available for 65 cases, of which 60 (95%) reported consuming any type of leafy green prior to illness. Of 57 cases who could remember the exact type of leafy green consumed, 50 (88%) consumed romaine lettuce. This is statistically significantly higher than the background rate of 49% from the FoodNet Population survey. 

A traceback investigation was initiated in response to the E. coli O157 outbreak with leafy greens as the suspected vehicle. Each case included in the traceback investigation reported consumption of leafy greens prior to illness onset. Based on information available at the points of service (POS), the traceback focused on iceberg and romaine lettuce. The investigation consisted of three traceback legs representing 28 cases and five POS. The three traceback legs identified four distribution centers, one broker, two processors, one grower, and one ranch. The traceback investigation determined that a sole processer sourced romaine lettuce from a single grower that would have been available at all points of service during the timeframe of interest. Additionally, romaine lettuce supplied to four of the five POS was traced back to a common ranch and lot. Through analysis of records, four lots of romaine lettuce were implicated, resulting in confirmation of romaine lettuce as the vehicle.

Epidemiologic and traceback data supported the conclusion that romaine lettuce was the source of illnesses in this outbreak. CDC closed this investigation on January 15, 2025, following the elapsing of the surveillance reporting lag period and lack of new uploads. CDC closed this investigation as an outbreak with a confirmed vehicle of romaine lettuce. This outbreak was reported to NORS with NORS ID: 511856.

Based on the documents obtained from the FDA and in discovery, the processor of the contaminated subject lettuce was Taylor Farms, Salinas, CA, and the grower of the contaminated subject lettuce was Anthony Costa and Sons, LLC. 

[1]U.S. Centers for Disease Control and Prevention (CDC), Multistate Outbreak of E. coli O145 Infections Linked to Shredded Romaine Lettuce (Final Update / 2010); J.M. Taylor et al., “Multistate Outbreak of Escherichia coli O145 Infections Associated with Romaine Lettuce Consumption, 2010,” Journal of Food Protection (2013), PubMed: pubmed.ncbi.nlm.nih.gov/23726187; U.S. Food and Drug Administration (FDA), Environmental Assessment of the Freshway Foods O145 outbreak.

[2]CDC, 2011 E. coli Outbreak Linked to Romaine Lettuce (posted Mar. 23, 2012), archived at archive.cdc.gov/…/ecoli/2011/romaine-lettace-3-23-12.html; R.B. Slayton et al., PLoS ONE (2013), DOI:10.1371/journal.pone.0055300.

[3]CDC, Update: Multistate Outbreak of E. coli O157:H7 Infections (Media Statement, Jan. 10, 2018), archived at archive.cdc.gov/…/media/releases/2018/s0110-update-ecoli.html. The Public Health Agency of Canada identified romaine lettuce as the source of the linked Canadian outbreak.

[4]CDC, 2018 E. coli Outbreak Linked to Romaine Lettuce (Yuma growing region) – Final Update (June 28, 2018), archived at archive.cdc.gov/…/ecoli/2018/o157h7-04-18/index.html; FDA, Investigated Multistate Outbreak of E. coli O157:H7 Linked to Romaine Lettuce from the Yuma Growing Region, fda.gov/…/romaine-lettuce-yuma-growing.

[5]CDC, 2018 E. coli Outbreak Linked to Romaine Lettuce (Central Coastal / Santa Maria, CA) – Final Update (Jan. 9, 2019), archived at archive.cdc.gov/…/ecoli/2018/o157h7-11-18/index.html; FDA, Investigation Summary: Factors Potentially Contributing to the Contamination of Romaine Lettuce Implicated in the Fall 2018 Multi-State Outbreak, fda.gov/…/implicated-fall.

[6]FDA, FDA, CDC and Other Health Partners Investigated Outbreak of E. coli O157:H7 Possibly Linked to Romaine Lettuce; Outbreak Appears to Be Over (FDA In Brief, Oct. 31, 2019), fda.gov/…/possibly-linked-romaine. This smaller outbreak was distinct from the November 2019 Salinas Valley outbreak and was disclosed after the product was past shelf life.

[7]CDC, 2019 E. coli Outbreak Linked to Romaine Lettuce (Salinas, CA) – Final Update (Jan. 15, 2020), archived at archive.cdc.gov/…/ecoli/2019/o157h7-11-19/index.html; FDA, Outbreak Investigation of E. coli O157:H7 – Romaine from Salinas (2019–2020).

[8]CDC, 2020 E. coli Outbreak Linked to Leafy Greens – Final Update (Dec. 22, 2020), archived at archive.cdc.gov/…/ecoli/2020/o157h7-10-20b/index.html; FDA, Outbreak Investigation of E. coli – Leafy Greens (December 2020), fda.gov/…/leafy-greens-december-2020. A specific type/brand was not confirmed, but the strain was genetically related to the 2019 romaine outbreak.

[9]CDC, 2020 E. coli Outbreak Linked to Unknown Source 3 – Final Update (Dec. 18, 2020), archived at archive.cdc.gov/…/ecoli/2020/o157h7-11-20/index.html. The outbreak strain was identified in a single-head package of Tanimura & Antle romaine lettuce, recalled Nov. 6, 2020.

[10]Uploaded outbreak memorandum re PulseNet code 2411MOEXH-2 / NORS ID 511856 (citing FDA Executive Incident Summary, Att. No. 18; FDA Additional Records re Romaine Lettuce Outbreak (Taylor Farms, unredacted), Att. No. 19; FDA Diagram TF00000163, Att. No. 20).

[11]FDA, CORE Outbreak Investigation Table, E. coli O157:H7 outbreak reference #1280 (“iceberg and romaine lettuce blend served at catering events, restaurants, and a school”), investigation opened Dec. 2024 and closed Feb. 2025 without a public final report naming a source; Food Safety News, “Publisher’s Platform: FDA closes romaine lettuce outbreak with 89 sick and says nothing more?” (Feb. 14, 2025), foodsafetynews.com.

[12]S. Chuck & A. Lozano, “A deadly E. coli outbreak hit 15 states, but the FDA chose not to make the details public,” NBC News (Apr. 17, 2025), nbcnews.com/news/us-news/ecoli-bacteria-lettuce-outbreak-rcna200236; “Uncovered: FDA Did Not Disclose Fatal E. coli Outbreak Linked to Lettuce in 2024,” Food Safety Magazine (Apr. 17, 2025), food-safety.com.

[13]Marler Clark, Inc., P.S. press release, reprinted in Food Poison Journal, “Marler Clark reports on 89 person E. coli Outbreak not reported by CDC and FDA” (Apr. 17, 2025), foodpoisonjournal.com; see also Marler Blog, billmarler.com.

[14]Marler Clark federal lawsuits and amended complaints naming Taylor Farms (Apr. 17, 2025); see NBC News (Apr. 17, 2025) and Salon, “The name of the grower behind a deadly E. coli lettuce outbreak has been disclosed” (Apr. 24, 2025), salon.com. Taylor Farms denies that its product was the source, stating its raw and finished-product testing found no evidence of contamination.