Saturday, August 29, 2026

This is my eighth Saturday column. Last Saturday it ran under a different name — Four Outbreaks, One Weekend — and it closed on the observation that in every one of those four, the piece of paper a person would need in order to protect themselves arrived late, arrived incomplete, or has not arrived at all. This week the pattern held and the calendar ran out. The Cyclospora season CDC defines closes on Monday, August 31. On Thursday, four days short of it, the federal outbreak added three states it had not had on Wednesday.

The federal outbreak grew on Thursday, and it started adding states again. CDC’s outbreak page, stamped August 27, now reads 11,458 illnesses, 495 hospitalizations, two deaths and twenty states, three of them new. That is 528 more people and 41 more hospitalizations than the August 20 update, which was the first of the entire outbreak to add no state at all. I read that update at the time as an outbreak deepening inside its own footprint rather than spreading. One week later it took in three more states, and I would rather correct that here than leave it standing. The page does not name the three. FDA’s advisory carries the state list, and it is worth reading against the seventeen it had a week ago.

Two things changed on that page besides the numbers, and they cut in different directions. The alert box has dropped the line saying Taylor Farms initiated a recall on July 17; what remains says only that the best-by dates have passed. That is the second week running something has come off the consumer-facing page — last Thursday it was the paragraph naming the thirty-one states the lettuce went to. In the other direction, CDC has added a direct link to the Taylor Fresh Foods recall notice under See also, where a week ago there was only a link to FDA’s investigation. That is an improvement and it belongs in the same paragraph as the complaint.

CDC’s separate national count moved first this week, and by more. Tuesday’s surveillance updateputs laboratory-confirmed, domestically acquired cases at 17,180 since May 1, with 922 hospitalizations and at least 11,844 more still waiting on confirmation, across forty-eight states plus the District of Columbia. A week earlier those figures were 15,716 and 828. The comparison CDC prints in its own key points is the one to keep: 1,180 cases for the same stretch of 2025. I wrote it up Tuesday night.

Adding up what the fifty state health departments publish about themselves, the floor is 32,506. That was Wednesday night. Six states moved earlier in the week — Illinois, Kansas, Iowa, Virginia, Arkansas and Oklahoma — and the largest single change on the board was a correction to my own chart, which had been carrying Illinois 177 people too low. The last six people in that total are Hawaii’s, entering the chart for the first time. Michigan reported again on Thursday, so the floor is already behind. It remains a floor and not an estimate. It counts only what fifty health departments have published about themselves.

Hawaii is the one worth stopping on. The Hawaii Department of Health announced Wednesday that it is investigating a sixth case of cyclosporiasis — an Oʻahu resident who became ill at the end of July and did not travel outside Hawaiʻi during the period when they were likely exposed. No food source has been identified, and the department says plainly that no recalled Taylor Farms de Mexico lettuce was distributed to or sold in the islands. The five earlier Hawaii cases this season were all travel-associated. This one is not, which is why Hawaii is counted on my chart for the first time. A locally acquired case in a state with no recalled product and no outbreak link is either a coincidence or a question, and so far nobody has said which.

The attributed share of the country’s hospitalizations crossed half this week. CDC counted 922 people hospitalized with this parasite on Tuesday, and on Thursday attributed 495 of them to the Taylor Farms outbreak. A week ago it was 454 against the same 922. Read it loosely rather than to the decimal, because the two figures come off pages stamped two days apart. The roughly four hundred and thirty people on the other side of it are not people anyone has cleared. They are people nobody has said either way about, because attribution here still runs through somebody remembering a meal from June, weeks before anyone knew to ask. Eleven states have been added to this outbreak in August alone.

The honest counterweight is that the curves are falling. North Carolina reported about 250 cases a week in early July and roughly ten in the first two weeks of August, the steepest decline any state has published. Missouri’s weekly increase has come down from roughly four hundred percent to ten. Michigan had its smallest week of the outbreak on August 20. All three are true at the same time as three new states on Thursday, which is what the tail of an outbreak looks like: the places that have been counting longest are finding fewer people, and the places that started late are still finding their first. With a six-week reporting lag, illnesses that began in early August will still be arriving in October. The season closing and the outbreak closing are two different events, and only one of them has a date.

The last congressional deadline on this outbreak came due Thursday. Thirteen letters went out of the House and Senate between July 14 and August 13 — to the Secretary, to the acting heads of FDA and CDC, and twice to the chief executive of Taylor Farms. Seven set a dated deadline that has now passed. The last one still live was Senator Warren’s nine questions, due August 27. I cannot find a published answer to any of the thirteen, and the answers that do exist have gone to a newspaper, to a television network, and onto a company FAQ page revised at least five times and overwritten in place. As of this morning there is no outstanding deadline in front of any agency or any company on this outbreak. That is what a hearing is for.

The sprout recall got fixed, and not by the company. Everything Sprouts, LLC put out its own voluntary recall notice on the night of Friday, August 21, and it was empty: it said the recalled products include the following lot numbers and then listed none, with no best-by dates, no UPC, no product photograph, no store list, and no mention of the Calco brand both federal agencies name. Below a row of hash marks at the foot of the document sat FDA’s blank sample press release — XYZ Inc. of Anywhere, Mississippi, recalling five-ounce packages of Snackies with lot number 666666 and an expiration date of 12/12/99. They published the form with the instructions still in it. The version FDA posted the next day carries what the company left out: lots 222, 223, 225, 226 and 230, the bar codes, the Calco brand, and the two sprout mixes alongside the five-ounce containers of alfalfa sprouts, distributed May 27 through August 21. A shopper can now check a container against a lot code, which is the whole point of the document. It took the agency to write it. The outbreak itself holds at 55 sick in fifteen states, four hospitalized, no deaths, with onsets May 31 through August 8, per CDC and FDA, and FDA is still working on distribution.

A correction of mine belongs here, and it is a happy one. Last week I wrote that FDA’s investigations table is built one row per pathogen, so a genuinely multi-pathogen outbreak could never appear on it as a single entry. FDA has since merged references 1394 and 1404 into one incident, stating that CDC identified four people infected simultaneously with three E. coli strains and Salmonella Agona and combined the associated clusters. The architecture bent. I was wrong, and it is a better outcome than the one I predicted.

The jalapeño outbreak closed its own gap. FDA’s advisory now carries 431 illnesses in thirty-two states, 57 hospitalizations and no deaths, with onsets June 19 through August 2, and it names Coast Citrus Distributors in the opening sentence. The Coast Citrus recall that had no public notice anywhere for two weeks does exist in the enforcement record: my friend Phyllis Entis at eFoodAlert found it — 53,760 cases of jalapeño peppers recalled on July 22, covering all product packed July 1 through July 31, classified Class I. That is the document more than 160 downstream grocery products have been recalled against, and no shopper has been able to read it. The grower in Sinaloa is still unnamed, six weeks in.

San Diego County’s Ramona outbreak sits at twenty-three and still has no recall. Eighteen illnesses tied to house-made mayonnaise sold at a market, five in people who got sick from the eggs themselves, seven hospitalized, most recent case detected August 16. Four weeks in there is no recall and no published customer list. The farm’s position has hardened into an argument rather than a denial — that nine of the ill share the same genotyping but the health department never obtained a test sample from the farm. That concedes the clustering and attacks the missing product isolate, which is the wrong burden. Civil causation has never required a match between a patient isolate and a food, and a negative swab pulled in August tells you very little about eggs laid in June. The split is here.

Two brands of infant formula, made on two continents, turn out to have shared one organism.CDC closed the Nara Organics infant botulism outbreak on Thursday at four babies in California, Pennsylvania and Washington — all four hospitalized, none dead, onsets April 20 through May 31, every lot recalled June 13. The finding that closed the file is the one that matters. Whole genome sequencing shows the strain from one of those four infants matches a clinical sample from an infant in the 2025 ByHeart outbreak, a powdered formula base sample collected in that investigation, and four samples of unopened cans of ByHeart formula. Unopened is the word doing the work: a positive from an opened can invites the argument that a household introduced the organism after purchase, and sealed cans foreclose it. Both brands’ formula was made from milk supplied by Organic West Milk and spray dried by Dairy Farmers of America, and FDA still cannot say whether the milk was contaminated before drying or during it. During the 2025 investigation, Organic West gave FDA a customer list that did not include the maker of the Nara formula. The first Nara baby got sick 160 days after the ByHeart recall. I represent families in both outbreaks, so read me accordingly, and I set the two supply chains side by side on Thursday. On Friday I wrote to both committees again, because a laboratory result answered in August what I could only argue in June, and asked them to put a written question to FDA with a date on it.

The recall ledger, and one item at the top of it. Phyllis Entis carries these every week at eFoodAlert, and one of hers this week belongs first: Piazza Produce, doing business as Cibus Fresh, is recalling Cibus Fresh, Sprig and Sprout, and Jack and Olive brand Garden Salad in 7.75-ounce containers for possible Cyclosporacontamination, with best-by dates running July 13 through July 20. Those dates expired five weeks ago, which means this notice cannot pull a single container off a shelf. It can still tell somebody who has been miserable since July what to ask a doctor to test for, and that is the other thing a recall is for. Elsewhere on her list this week: a Québec spice recall, a feline milk replacer pulled for low vitamin levels, and in the United Kingdom a Tesco salad recalled for Shiga toxin-producing E. coli.

The consumer damage got measured, and the customers named their own terms. Numerator found that more than 6.5 million American households stopped buying salad mixes and kits in the month ending July 26, and that eighty-seven percent of produce shoppers know about this outbreak. The finding I would put in front of every board in the industry is that 82 percent of the people who quit bagged salad say they will buy it again once the outbreak is contained. Not once they are reassured. Not once a campaign runs. Contained is a factual condition they can check for themselves. Meanwhile forty percent of Americans told CBS News and YouGov they are buying less fresh produce because of this, and they are not substituting — canned and frozen barely moved. That is Americans eating fewer vegetables in the same season the department that runs both agencies is telling them to eat more whole food. I wrote to the Secretary about that on Wednesday.

One argument I would rather not have had. FDA’s official account posted a chart on August 21 under the words do not believe the fake news, reporting recalls at their lowest level in over a decade. Susan Mayne, who ran the agency’s food center for eight years, pulled the same dashboard, filtered it to food, and found food recall events running toward the highest total since 2018. The chart was not a food chart at all. And US PIRG, whose data was cited for it, says in writing that the annual number of recalls is not an indicator of whether food is safe — which has been true in every administration I have watched. The reason it matters is not the argument. It is that the claim went out attached to a warning about frozen blueberries, to people who needed to know whether something in their freezer would hurt them.

And on Friday came a proposal that would undo the best thing that has happened in this field in my lifetime. The President said he would authorize legal documents giving farmers and ranchers the right to process their own food, calling the four largest processors a monopoly. From 1993 to 2002, at least 95 percent of this firm’s revenue came from E. coli cases linked to hamburger. It is close to zero now, and the reason is a federal inspector standing in the plant to take the sample, hold the lot and stop the shipment. The consolidation complaint underneath the post is real and it has real remedies, none of which is removing that person. Meat with no establishment number cannot be traced or recalled, which means the illnesses would arrive one at a time in counties that never compare notes and most of them would never be counted at all. The Meat Institute and the National Cattlemen’s Beef Association both came out against the idea the same afternoon, which does not happen often. I went through the whole of it on Friday.

We are hiring, and that is not good news. The paralegal posting went up Tuesday — ten years or more in plaintiffs’ personal injury or complex litigation, here on the island, no billable hours and no timesheets. There is an associate opening alongside it. Mia de Graaf opened her Business Insider piece this week on the fact that my business is booming, which is the worst review a food system can get. She is right, and every one of those files is a person who ate something.

That’s the week — a federal outbreak that took in 528 more people, 41 more hospitalizations and three more states on the last Thursday of the season after adding none the Thursday before, a national count 1,464 higher than it was seven days ago, a state with no recalled lettuce finding a case that never left the island, the last congressional deadline passing with nothing behind it, one strain of C. botulinum in two brands of infant formula, a Cyclospora salad recall issued five weeks after its own best-by dates ran out, and a proposal to let beef be slaughtered and sold with nobody inspecting it. The season CDC defines closes on Monday. Nothing else does. Check back next Saturday.

About Bill Marler

William “Bill” Marler has spent more than thirty years as a food safety lawyer and advocate—work that began with the 1993 Jack in the Box E. coli outbreak and has never really stopped since. In the years that followed, he has represented victims in nearly every major foodborne illness outbreak in the United States. That case, and the movement it launched, is the subject of the book “Poisoned” and the Emmy Award–winning Netflix documentary of the same name. Bill’s 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’s fighting food poisoning again”), and many others.

Dozens of times a year, Bill speaks to industry, regulators, and universities across the United States, Canada, Europe, Africa, China, and Australia about a simple idea: outbreaks are preventable. He has testified before Congress on the Food Safety Modernization Act and teaches food safety at institutions including the Harvard T.H. Chan School of Public Health. He writes regularly about food litigation and food safety at Marler Blog, and in 2009 he founded Food Safety News, which he continues to publish.

Federal regulation requires powdered infant formula to be tested for two organisms before it reaches a store shelf: Salmonella and Cronobacter. It does not require testing for Clostridium botulinum. Fifty-two American infants have now been hospitalized with botulism across two formula outbreaks — 48 fed ByHeart formula beginning in 2025, four fed Nara Organics formula in 2026 — many of them on ventilators. On August 27, 2026, the federal government confirmed by whole genome sequencing that the two outbreaks share a common source of contamination. The one organism nobody is required to look for is the organism that put those babies in intensive care.

The government has now confirmed this with a laboratory result rather than an inference.

On August 27 the Centers for Disease Control and Prevention closed the Nara Organics outbreak at four confirmed cases, all four hospitalized, none fatal, with illness onsets between April 20 and May 31, 2026. Sequencing tied a patient sample from that outbreak to material collected during the 2025 ByHeart investigation, including four samples of unopened cans of ByHeart formula. FDA’s traceback determined that the formula sold by both brands was made with milk supplied by Organic West Milk and spray dried by Dairy Farmers of America — one milk supply, one spray dryer, one organism. FDA states that the matching sequences together with the shared supply chain indicate a common source.

What neither agency has determined is a root cause. FDA cannot yet say whether the milk was contaminated before it was dried or during the drying. That question decides how responsibility gets divided among the companies. It does not touch whether the formula was contaminated, and it does not touch whether the formula paralyzed babies. CDC decided both of those on August 27. ByHeart has since announced that FDA told the company it has no further questions about its root cause submission, and has released neither the submission nor the agency’s letter.

A bill that would close the gap has been sitting since April.

H.R. 7867, the Infant Formula Safety Modernization Act of 2026, would add C. botulinum to the required testing list, require environmental monitoring inside the plants, reach foreign manufacturers, and require notice to Congress when an outbreak is identified. The House Energy and Commerce Subcommittee on Health held a legislative hearing on it on April 29, 2026. Nothing has moved since. The Senate passed a related measure the same day, S. 272, by unanimous consent — a genuine accomplishment that nonetheless reaches only Cronobacter and Salmonella, the two organisms already on the books. I wrote to both committees in June and again in August, and those letters are published in full.

Five hundred thirty-five baby bottles arrive on Capitol Hill the second week of September.

Earlier this year I had 535 t-shirts printed, one for every member of Congress, each rolled inside a plastic baby bottle. The front reads, “For Babies’ Sake, Test for Botulism in Infant Formula.” The second week of September my daughter will hand-deliver them. A shirt on its own is a stunt. What would make it something else is mail from constituents arriving the same week.

Four offices decide whether this bill moves, and all four can be written today.

The two committees that control this legislation are the House Committee on Energy and Commerce and the Senate Committee on Health, Education, Labor, and Pensions. Four people run them: Chairman Brett Guthrie of Kentucky and Ranking Member Frank Pallone, Jr. of New Jersey in the House, and Chairman Bill Cassidy of Louisiana and Ranking Member Bernie Sanders of Vermont in the Senate. A chairman decides what gets marked up. That is the whole reason to write these four rather than write generally.

Here is a message you can send. Change whatever you like, and put your own city and state at the bottom, because where you live is the part a congressional office reads first.

Subject: Please mark up H.R. 7867 — botulism testing in infant formula

Dear Chairman Guthrie and Ranking Member Pallone,

Federal regulation requires finished infant formula to be tested for Salmonella and Cronobacter. It does not require testing for Clostridium botulinum. Fifty-two infants have been hospitalized with botulism across two formula outbreaks, and on August 27, 2026 CDC and FDA reported that whole genome sequencing links those two outbreaks to a common source of contamination in a shared milk supply.

H.R. 7867, the Infant Formula Safety Modernization Act of 2026, would add that organism to the required testing list and require environmental monitoring inside the plants. The Subcommittee on Health held a legislative hearing on it on April 29, 2026, and it has not moved since.

I am asking the Committee to do three things: mark up H.R. 7867 and report it favorably; hold an oversight hearing with FDA and the companies in this supply chain at the table; and hear directly from the parents of the injured infants. Please tell me where the Committee stands.

[your name, your city and state]

For the Senate, address it to Chairman Cassidy and Ranking Member Sanders and change the ask. The Senate has already acted once this year: S. 272 cleared the HELP Committee 22 to 0 and passed the full Senate by unanimous consent on April 29, 2026, which deserves saying out loud. It reaches only Cronobacter and Salmonella. Ask for a companion measure that reaches C. botulinum, for the oversight hearing, for the parents at the witness table, and for a written question to FDA with a deadline on whether a rule is coming and when.

How to actually reach those four offices, because the obvious route often fails.

Congressional web forms generally require an address inside the member’s own district or state, so unless you live in Kentucky’s second district, New Jersey’s sixth, Louisiana or Vermont, the form will turn you away. That is not a brush-off; it is how the offices sort constituent mail. Three routes are open to everyone. Call: the Capitol switchboard connects you to any House office at (202) 225-3121 and any Senate office at (202) 224-3121, and the Energy and Commerce majority staff line is (202) 225-3641, the minority staff line (202) 225-2927. Write on paper: Committee on Energy and Commerce, 2125 Rayburn House Office Building, Washington, DC 20515, and Committee on Health, Education, Labor, and Pensions, 428 Dirksen Senate Office Building, Washington, DC 20510 — the Senate publishes the form to use, and warns that outside mail runs a week to ten days late, so post it now if you want it there with the shirts. And use the committee websites’ own contact pages, which are not district-gated.

If your own representative or senator sits on one of these committees, you outrank everyone else writing.

A contact logged from a member’s own district or state carries weight that a message from anywhere else does not, so if one of your members is on the roster, write that person first and the chairmen second. Fifty-four members sit on Energy and Commerce and twenty-three sit on HELP. Thirty-six states have at least one. Here they are.

StateHouse Energy and CommerceSenate HELP
AlabamaGary J. Palmer (R)Tommy Tuberville (R)
Alaska—Lisa Murkowski (R)
CaliforniaJay Obernolte (R); Doris O. Matsui (D); Raul Ruiz (D); Scott H. Peters (D); Nanette Diaz Barragán (D); Kevin Mullin (D)—
ColoradoGabe Evans (R); Diana DeGette (D)John W. Hickenlooper (D)
Connecticut—Christopher Murphy (D)
Delaware—Lisa Blunt Rochester (D)
FloridaGus M. Bilirakis (R); Neal P. Dunn (R, Vice Chair); Kat Cammack (R); Laurel M. Lee (R); Kathy Castor (D); Darren Soto (D)Ashley Moody (R)
GeorgiaEarl L. “Buddy” Carter (R); Rick W. Allen (R)—
IdahoRuss Fulcher (R)—
IllinoisJanice D. Schakowsky (D); Robin L. Kelly (D)—
IndianaErin Houchin (R)Jim Banks (R)
IowaMariannette Miller-Meeks (R)—
Kansas—Roger Marshall (R)
KentuckyBrett Guthrie (R, Chair)Rand Paul (R)
LouisianaTroy A. Carter (D)Bill Cassidy (R, Chair)
Maine—Susan M. Collins (R)
Maryland—Angela D. Alsobrooks (D)
MassachusettsLori Trahan (D); Jake Auchincloss (D)Edward J. Markey (D)
MichiganJohn James (R); Debbie Dingell (D)—
Missouri—Josh Hawley (R)
New Hampshire—Margaret Wood Hassan (D)
New JerseyThomas H. Kean, Jr. (R); Frank Pallone, Jr. (D, Ranking Member); Robert Menendez (D)Andy Kim (D)
New YorkNicholas A. Langworthy (R); Paul Tonko (D); Yvette D. Clarke (D); Alexandria Ocasio-Cortez (D)—
North CarolinaRichard Hudson (R)—
North DakotaJulie Fedorchak (R)—
OhioRobert E. Latta (R); Troy Balderson (R); Michael A. Rulli (R); Greg Landsman (D)Jon Husted (R)
Oklahoma—Alan Armstrong (R)
OregonCliff Bentz (R)—
PennsylvaniaJohn Joyce (R)—
South CarolinaRussell Fry (R)Tim Scott (R)
TennesseeDiana Harshbarger (R)—
TexasDan Crenshaw (R); Randy K. Weber, Sr. (R); August Pfluger (R); Craig A. Goldman (R); Marc A. Veasey (D); Lizzie Fletcher (D)—
Vermont—Bernard Sanders (I, Ranking Member)
VirginiaH. Morgan Griffith (R); Jennifer L. McClellan (D)Tim Kaine (D)
WashingtonKim Schrier (D)Patty Murray (D)
Wisconsin—Tammy Baldwin (D)

Members of the two committees with jurisdiction over H.R. 7867, by state. Rosters retrieved August 29, 2026 from the Office of the Clerk of the House and senate.gov. Check: 54 House members across 25 states plus 23 senators across 23 states, 77 members in all, and 36 of the 50 states appear above.

Fourteen states have nobody on either committee: Arizona, Arkansas, Hawaii, Minnesota, Mississippi, Montana, Nebraska, Nevada, New Mexico, Rhode Island, South Dakota, Utah, West Virginia and Wyoming. If you live in one of them your own members can still tell Guthrie, Pallone, Cassidy and Sanders that they want this bill moved, and a member asking a chairman is worth a great deal more than a member asking nobody. The House and the Senate both publish contact pages and district phone numbers. Whoever you write, ask for the same three things.

On August 28, 2026, ByHeart announced that the Food and Drug Administration had told the company in writing that, apart from two pending items, it has no further questions about the roughly 400-page root cause investigation and action plan ByHeart submitted after the November 2025 infant botulism outbreak — 48 sick babies, 28 confirmed and 20 probable, in seventeen states, every one of them hospitalized. The trade press ran it as the agency clearing the company. 

Two documents are now doing public work: ByHeart’s submission and FDA’s written response. Neither has been released. Both should be.

The principle is one every lawyer knows and most people understand on contact. You cannot put a document in issue and then decline to let anyone read it. ByHeart did not have to say anything at all. It chose to announce the conclusion of a document on a national wire, and within hours the announcement was reprinted as news. Having relied on the submission and the letter to tell the public where things stand, the company owns the obligation to produce them.

The page count makes the point rather than answering it. “Approximately 400 pages” is offered as evidence of thoroughness, and nobody can weigh 400 pages they are not allowed to read. The same goes for the letter. The release quotes four words from it — no further questions — and withholds the rest. A company that quotes four words from a letter and will not publish the letter has told you the letter says more than four words.

The announcement does answer some real questions

Start with what it settles, because it settles some things and a fair reading has to say so. FDA has finished reading the submission and said so in writing, which is a dated fact and more than most companies in this position disclose. ByHeart is not coming back to shelves soon: before restarting production, it must give FDA 28 days’ notice and complete verification activities and pre-production commitments it does not describe. The company is changing its whole milk ingredient supplier, which is a concession that the old chain is not being resumed. It is also changing how it treats facility surfaces, adopting a new sporicidal method. And its detection science is public — the research and the protocol were both published rather than held as trade secret. Parents asking whether the formula is returning have an answer: not yet, and not without more steps.

FDA’s own page still says it has not determined a root cause

That is where the release and the record part company. FDA’s outbreak page, current as of June 3, 2026, reports that whole genome sequencing has identified seventeen different strains of Clostridium botulinumacross patient samples, finished product and ingredients, and states that the agency has not yet determined a root cause. The banner across the top of that page says the root cause investigation, with a focus on ingredients, is ongoing. A headline announcing that the root cause points to a milk ingredient states as settled the very thing the agency says is not.

Two smaller things are worth noticing in the same breath. FDA wrote that the inspections identified the organism in a powdered milk ingredient. The release wrote a powdered whole milk ingredient sourced from a third-party ingredient supplier. And FDA’s sentence is that the inspections did not identify any additional factors within ByHeart’s facilities that could explain the cause of this outbreak. The word doing the work is additional. That sentence says nothing further was found beyond what was already found.

“No further questions” means FDA has finished reading a submission. It is not a determination of root cause, not an approval of the action plan, not a clearance to resume production and not an exoneration of anyone. The release concedes as much when it lists the steps still standing between the company and the shelf.

FDA said the day before that the two outbreaks share a common source

On August 27, 2026 — the day before the announcement — FDA updated its advisory on the Nara Organics outbreak, the second infant botulism outbreak tied to powdered formula in seven months. Four confirmed cases in California, Pennsylvania and Washington, onsets between April and May 2026, all four babies hospitalized. CDC declared it over. What FDA published alongside the closeout is the part that matters.

Whole genome sequencing shows that a clinical sample from an infant in the Nara outbreak is a genetic match to a clinical sample from an infant in the 2025 ByHeart outbreak, to a ByHeart powdered infant formula base sample, and to four samples of unopened cans of ByHeart formula. FDA determined that the Nara lots linked to illness were made with milk supplied by Organic West Milk and spray dried by Dairy Farmers of America — the same suppliers as ByHeart. And the agency states plainly that the matching sequencing results and the supply chain commonalities indicate a common source of contamination between the two outbreaks, while adding that it still cannot say whether the milk was contaminated before drying or during it, and that its root cause investigation is ongoing.

Four unopened cans. That forecloses any suggestion that this was something introduced in a kitchen after purchase.

There is one more thing on that page. During the 2025 investigation FDA ran a trace-forward to learn whether Organic West Milk had supplied milk to other formula manufacturers. Organic West gave the agency a customer list that did not include Nara Organics. Only during the 2026 investigation did the company tell FDA the earlier list had been incomplete. Between that list and the second recall, four more babies were hospitalized with botulism.

That was the state of the public record on August 27. On August 28, a company whose entire announcement is about ingredient supply chain root cause issued a release that mentions none of it. And FDA’s own post-outbreak response page — the page the agency says will carry future root cause updates — is still stamped July 13 and has not caught up to its own sequencing result.

The questions still open are exactly the ones the submission would answer

Where in the milk chain did this happen? If an ingredient is the vehicle, was the milk contaminated before it was dried or during drying? Neither agency has said, and the release does not touch the dryer.

Why is a new sporicidal treatment for facility surfaces needed if the facilities were not implicated? That question comes from the company’s own disclosure. Nobody rebuilds surface sanitation over an ingredient somebody else made.

What about the second lot? ByHeart’s own published research reports that an isolate from one finished lot was genetically identical to a base powder, while an isolate from another lot was distinct. One contaminated ingredient lot does not explain a second, different strain in a second lot.

What happened between 2022 and 2025? CDC widened the case definition back to the product’s March 2022 release, and children who got sick in December 2023 and in early 2024 are inside this outbreak only because it did. A more sensitive test adopted in 2026 does not explain why nothing caught this in the years before it. The inspection records went back to 2022 as well, and that history has never been reconciled with a root cause that begins and ends with somebody else’s milk. The company also holds retained samples from that period, and the release does not say whether the new protocol has been run backward across them.

And how does a root cause that begins and ends with somebody else’s milk account for a second brand? FDA has now said the two outbreaks share a common source of contamination. The submission either addresses that, or it does not.

Every one of those questions is answered somewhere in 400 pages, or it is not answered at all. Either way, the document is the answer.

ByHeart is the only party that can release the submission

This is the part most readers will not know, and it closes the obvious door before anyone walks through it. A company’s submission to FDA lands in the confidential commercial information bucket. Ask the agency for it and you wait months for a version with the interesting parts blacked out. That protection belongs to ByHeart, and it is ByHeart’s to waive. FDA cannot publish this. ByHeart can publish it this afternoon.

The letter is easier still. It is short, it is presumably favorable, and it is the entire foundation of the announcement. Posting it costs the company nothing if it says what the company says it says.

And ByHeart has already accepted the principle. This is not a company that went silent. It published the research with IEH Laboratories, it published the testing protocol, and it kept a public timeline of the investigation, including its own positive results and the January genetic match between its formula and whole milk powder collected from its supplier. The argument is not that the company should start being transparent. It is that having chosen transparency in pieces, it should finish.

Redact what genuinely needs redacting and publish the rest

Employee names come out. Genuine formulation secrets unrelated to the hazard come out. That is a small fraction of 400 pages, and saying so up front is what makes this a request rather than a stunt. If material belonging to the milk supplier or the processor is in there and is not ByHeart’s to publish, publish the rest and say which sections were withheld and why.

I will get these documents eventually. There are lawsuits, and discovery does that work. That is precisely the point. A protective order delivers a 400-page root cause analysis to me and to nobody else — not to the next company buying milk powder, not to the pediatrician trying to understand what happened to a patient, not to the parents of the babies in the next outbreak, who do not have a subpoena.

None of this assumes the documents say something other than what the company claims. They may say exactly that. The trouble is that nobody outside ByHeart and FDA can know and ByHeart is the party that made it a public matter.

An investigation nobody outside the company can read has not taught the industry anything. It has only told the industry that somebody learned something.

Six investigations defined the month. As of August 27, 2026, CDC counts 11,458 people sick with Cyclospora in twenty states from Taylor Farms de Mexico iceberg lettuce, 431 sick with Salmonella Javiana in thirty-two states from jalapeños distributed by Coast Citrus, 55 sick in fifteen states from Minnesota alfalfa sprouts carrying four different pathogens, fifteen sick and one dead from requesón cheese in an outbreak that ran three years and four months, and four infants hospitalized with botulism from Nara Organics powdered formula. Two of those are still open. Three closed this month. Each one leaves a household something to do.

The lettuce outbreak reached 11,458 people in twenty states and is still adding cases.

CDC updated the Cyclospora outbreak on August 27, adding Georgia, Tennessee and Texas and bringing the count to 11,458 illnesses in twenty states, at least 495 hospitalizations and two deaths, both in Michigan. Illnesses began June 14 and onset dates run through August 15. Epidemiologic and traceback data pointed at iceberg lettuce from Taylor Farms de Mexico, a processor in central Mexico, and the product was recalled July 17. CDC has since said its genotyping shows the parasites from sick people are genetically related, which is evidence they came from a common source. FDA has moved its onsite inspections and sampling upstream from the processing plant to the iceberg lettuce growers themselves.

The recalled lettuce is gone. Best-by dates passed weeks ago, and nothing in a refrigerator today is part of that recall. The live problem is diagnostic, and it is the reason this outbreak got as large as it did before anyone named it. A routine stool test does not look for Cyclospora. Neither, reliably, does a standard ova-and-parasite exam. A doctor has to order the specific test, and most do not think to.

What to do: if you or someone in your family had watery diarrhea that started this summer, went away, and came back, with fatigue and appetite loss and often weeks of it, ask your doctor directly for Cyclospora testing by name. The infection is treatable with an antibiotic, and untreated cases relapse. Report a confirmed case to your local health department. Per Phyllis Entis’s reading of CDC’s surveillance data on August 26, the agency has recorded 17,180 confirmed domestically acquired cases nationally since May 1 with 922 hospitalizations, against 1,180 for all of 2025. The outbreak count and the national count are different series, and both are floors.

Coast Citrus jalapeños sickened 431 people and set off recalls that reached more than 160 grocery products.

FDA and CDC have tied 431 Salmonella Javiana infections in thirty-two states, with 57 hospitalizations and no deaths, to fresh jalapeños grown in Sinaloa, Mexico and distributed by Coast Citrus Distributors. Onsets run June 19 through August 2. Of 224 people interviewed, 203 had eaten at a Mexican-style restaurant, and CDC identified twenty-eight illness clusters across eight states. Chipotle switched jalapeño suppliers for affected stores on July 20 and QDOBA stopped using them July 28. I wrote the count up when it moved from 345 to 431.

The recall is the part worth understanding. One contaminated ingredient moved into finished food made by other companies, and the notices came out one at a time: twenty Taylor Fresh Foods items, an FSIS public health alert covering eighteen meat and poultry products, forty-three items Whole Foods made in its own departments, sixty-three across two Albertsons notices, eighteen NatureBest products, a Salata dressing, and five-pound bags of whole peppers at two Texas Costco warehouses. Most of those products did not carry the name of the company that had the problem. The grower in Sinaloa still has not been publicly named.

What to do: the best-by dates on these recalls have passed, so the refrigerator sweep is over. If you ate at a Mexican-style restaurant in late June or July and got sick, and you are still symptomatic, ask for a stool culture. An infection that is never cultured is never counted and never linked, and the culture is also what tells you what you actually had. When a recall does hit a product you own, match the UPC or the lot code rather than the brand on the front, because the brand is frequently not the manufacturer.

Minnesota alfalfa sprouts carried four different pathogens at once, which points at the seed rather than the kitchen.

CDC and FDA linked 55 illnesses in fifteen states, with four hospitalizations, to alfalfa sprouts grown by Everything Sprouts, LLC of Minneapolis and sold under the Calco and Everything Sprouts brands. Onsets ran May 31 to August 8. The pathogen line is the striking part: three separate Shiga toxin-producing E. coliserotypes, O26:H11, O103:H25 and O168:H8, plus Salmonella Agona, with two people carrying both an E. coliand a Salmonella. FDA inspected the sprouting facility on August 19 and recommended a recall; the company recalled on August 22.

Four organisms in one product is a seed signal. Sprout seed is grown in open fields, where it can pick up contamination from irrigation water or animals, and it then sits for days in warm water at the temperature bacteria like best. There is no cooking step, because the product is sold to be eaten raw. That is why sprouts have caused outbreaks for thirty years, and it is why the same seed lot can seed more than one sprouting operation.

What to do: the recalled containers are 5-ounce plastic with lot numbers 222, 223, 225, 226 and 230, and the same sprouts went into Everything Sprouts Crunchy Protein Sprout Mix and Zesty Garlic Mix cups. Throw them out, do not eat them, and wash whatever they touched. The distribution was to Minnesota and Wisconsin, but the illnesses are in fifteen states, because people travel and eat away from home. The broader point applies past this recall: raw sprouts are not a safe food for children under five, adults over sixty-five, pregnant women, or anyone with a weakened immune system. Cook them or leave them off.

A Listeria outbreak in requesón cheese hospitalized fourteen of fifteen people and killed one before it was closed.

On August 26, CDC and FDA closed a multiyear Listeria monocytogenes outbreak traced to requesón, a fresh soft cheese, made by Clover Hill Dairy, LLC of Mechanicsville, Maryland. Final count: fifteen people sick in Illinois, Maryland, New York and Virginia, fourteen hospitalized, one death in Maryland. Illness onsets ran from March 6, 2023 to July 8, 2026. FDA found Listeria in six product samples of the requesón and in environmental samples at the plant, matching the outbreak strain, and Maryland suspended the dairy’s license. I wrote about the closeout at Food Poison Journal.

Two facts in that record deserve a household’s attention. Everyone whose hospitalization status was known went to the hospital, which tells you Listeria gets counted only when someone is sick enough for a blood or spinal fluid culture. And the cheese was repacked and relabeled before it reached the shelf, so the brand a family remembers buying was often not Clover Hill. A recall notice that asks you to look for a name that was never on your tub is not much of a notice.

What to do: Listeria grows at refrigerator temperature, which almost nothing else does, and its incubation can run weeks, so a soft cheese eaten in June can produce a fever in July. Freezing does not kill it, and requesón is commonly bought in bulk and frozen, so check the freezer as well as the refrigerator. If you are pregnant, over sixty-five, or immunocompromised, heat fresh soft cheeses until steaming or skip them. Fever, muscle aches, and in pregnancy anything flu-like, warrant a call to a doctor, and tell them what you ate.

CDC closed the Nara Organics infant botulism outbreak by linking it genetically to the ByHeart outbreak before it.

CDC declared the Nara Organics outbreak over on August 27: four confirmed infant botulism cases in California, Pennsylvania and Washington, all four hospitalized, no deaths, onsets between April 20 and May 31, with the formula recalled June 13. The finding that matters is not the number. Whole genome sequencing linked a patient sample from the Nara outbreak to material from the 2025 ByHeart investigation, and investigators say the formula sold by both brands was made with milk supplied by Organic West Milk and spray dried by Dairy Farmers of America. The ByHeart outbreak sickened 48 infants in seventeen states. Whether the milk was contaminated before or during drying is still not answered. I have been tracking both on the blog.

Two brands, two outbreaks, one ingredient supply chain. A parent choosing between formula brands had no way to know the two shared a milk powder, because nothing on either can said so. Until ByHeart, Clostridium botulinum was not treated as a hazard reasonably likely to occur in powdered infant formula, and the standard industry test did not catch it in either supply chain.

What to do: recalled formula should not be used, sold or donated. Before you throw a can out, photograph the lot number and use-by date; CDC suggests keeping the container marked DO NOT USE, stored away from your baby’s food, for at least a month, because the state may want to test it. Symptoms can take weeks to appear. Watch for poor feeding, a weak cry, constipation, drooping eyelids, loss of head control and a flat facial expression, and call the pediatrician the same day. Infant botulism is treatable, and treatment works better early.

San Diego County has twenty-three people sick from a Ramona egg farm, and the eggs are still for sale.

San Diego County first announced this outbreak on July 28 as a problem with house-made mayonnaise at a Ramona deli. On August 17 the County named the egg supplier — Happy Hens, a poultry ranch a few minutes up the road — and said cases of the same Salmonella strain were turning up in people with no connection to the market. On August 21 it split the count in two: eighteen illnesses tied to the deli’s Ramona Red Sauce, five in people who got sick from the eggs themselves. Twenty-three cases, seven hospitalizations, most recent case detected August 16. Paul Sisson of the San Diego Union-Tribune has caught every turn in it, including a County update that was never announced. I wrote it up when the County split the numbers.

What makes this one different from the other five is that nothing has been recalled. There is no recall, no published list of the stores and restaurants that received the eggs, and no consumer warning from the state. The County’s action has been to remind retail food facilities that may have received Happy Hens eggs about storage, grading and resales. Happy Hens has publicly denied the link, on social media and on camera, and co-owner Chloe Nevarez says the farm expected its own reports to come back negative. Ramona Family Naturals now says the results are not conclusive. Both of those positions belong in the record, and so does this one: the sequencing on the public record links the sick people to each other. Nothing published ties a patient strain to an egg or to a swab taken at the farm. A California Department of Food and Agriculture spokesperson has confirmed that FDA inspected the ranch.

What to do: this is the one item on the list where a household can act before anyone gets sick. Happy Hens eggs move into grocery accounts across Southern California, and absent a recall there is no notice telling anyone to throw a carton out. Cook eggs until both the yolk and the white are firm, discard cracked or dirty ones, and do not wash a dirty egg — throw it away. Treat any deli mayonnaise, aioli, dressing or sauce made in-house as an unknown and ask whether it is made with pasteurized egg, because California law requires pasteurized egg in exactly that food unless the facility holds a written variance. A premium label is not a control measure. Pasture-raised, organic and certified are claims about how a hen lived, not about whether an egg carries Salmonella internally, which it can from a healthy-looking bird in a clean barn. Marler Clark has filed suit in San Diego Superior Court on behalf of a Ramona woman hospitalized after a deli sandwich she did not know contained raw egg.

The same two things are true of all six, and both of them are about counting.

Every number here is a floor, and the agencies say so on their own pages. Most people who get a foodborne illness recover at home and are never tested, and the ones who are tested take three to four weeks to work through sequencing before they join a count. When you read 431, or 55, or fifteen, you are reading the people who went to a doctor, gave a sample, and had a laboratory grow something. The rest happened and were not written down.

And every one of these ran through a stretch of supply chain no shopper can see: a processor in central Mexico, a Sinaloa grower still unnamed, a seed lot, a cheese repacked under somebody else’s label, a milk powder shared by two formula brands, a carton of eggs with no customer list. Records that follow food through those hands are exactly what the FSMA 204 traceability rule was written to require, and Congress has directed FDA not to enforce it before July 20, 2028. Until then the tracing is done the slow way, by interviewing people after they are already sick.

The one thing a household controls in all of this is whether an illness gets counted. Get seen, ask for the test by name, and ask the laboratory or the health department to keep the isolate. Sequencing an isolate is what connects one person’s bad week to an outbreak, and isolates get discarded routinely. Keep the package, or a photograph of it, and the receipt.

Three questions people are actually asking this week.

Is anything from these outbreaks still in my kitchen?

The lettuce and jalapeño recalls have passed their best-by dates, so no. The sprouts recall is current: 5-ounce Calco or Everything Sprouts containers with lot numbers 222, 223, 225, 226 or 230, plus the Crunchy Protein and Zesty Garlic cups. Recalled requesón should be off shelves, but check the freezer, because that cheese is often frozen for later. Recalled Nara Organics formula should not be used at all.

I was sick in July and never got tested. Is it too late to find out what it was?

For a bacterial infection, largely yes, because a culture needs a sample while you are still shedding. ForCyclospora, which relapses and can run for weeks, it is often not too late, and a blood test cannot substitute; the stool test has to be ordered specifically. If you are still having symptoms from any of this, go in now rather than waiting it out.

How do I check whether something in my house has been recalled?

FDA posts recalls of produce, dairy, packaged food and formula; USDA’s Food Safety and Inspection Service posts meat and poultry. Match the UPC or the lot code printed on the package rather than the brand name on the front, because in three of this month’s six, the name on the label was not the name of the company that made the food. And remember that a food can be the identified source of an outbreak and still never be recalled at all.

A parasite from a lettuce field, a pepper from Sinaloa, a seed, a cheese vat that stayed contaminated for three years, a lot of milk powder that went to two different babies’ formulas, and a carton of eggs still on a shelf in San Diego County. None of it was bad luck. Every one of these was somebody’s supply chain, and every one of them was findable earlier than it was found.

On August 28, 2026, President Trump said he would give farmers and ranchers the right to process their own food, in a Truth Social post calling the four largest meat processors a monopoly. Here it is in full:

“Ranchers and Farmers have always been a number one priority for me. They work very hard, are smart, efficient, and immaculately CLEAN, but for years I have heard that they have had a tremendous problem with the Big Processors, who many say are a nasty Monopoly. There are, essentially, 4 of them, a very non competitive number, and they make life miserable for our wonderful Farmers and Ranchers, and I can’t let that happen, can I? So, in order to break this powerful monopoly, with much of its ownership based outside of the U.S., I am authorizing legal documents to be drawn in order to allow Farmers and Ranchers to be given the right to PROCESS THEIR OWN FOOD. This should move quickly. Thank you for your attention to this matter! President DONALD J. TRUMP”

Reuters reporters Leah Douglas, Katharine Jackson, Daphne Psaledakis and Susan Heavey reported the same day that the White House did not respond to requests for detail, and Bloomberg noted that it is unclear whether the President has the authority to do this at all. The Washington Post put the substance of it plainly, describing a plan to let ranchers bypass USDA-inspected meat plants. CNN reported where the idea came from: on Wednesday’s Glenn Beck program, Beck told the President that “the inspections are insane,” and Trump asked whether ranchers and farmers would do a good job of processing beef with less regulation, then said he would look at it that day. Whatever those legal documents turn out to say, they run straight at the reason I no longer make a living suing beef companies.

From 1993 to 2002, at least 95 percent of my firm’s revenue came from hamburger cases.

I started this work in January 1993, representing children sickened in the Jack in the Box outbreak. Hundreds were sick. Four children died. From that outbreak until the ConAgra recall in the summer of 2002, at least 95 percent of Marler Clark’s revenue came from E. coli cases linked to hamburger. I could count on a significant ground beef outbreak and recall arriving like clockwork every spring or summer. In 2002 I wrote an op-ed for the Denver Post asking the beef industry and USDA to put me out of business.

They largely did it. Ground beef E. coli is now close to zero as a share of what this firm does, and I have said so in print more than once. The millions the industry spent on interventions and the work of a generation of food safety professionals made the difference. It is the single best thing that has happened in food safety in my lifetime, and I do not say that grudgingly.

One sentence in a September 1994 speech did most of the work.

On September 28, 1994, FSIS Administrator Michael Taylor told the American Meat Institute that the agency considered raw ground beef contaminated with E. coli O157:H7 to be adulterated under the Federal Meat Inspection Act. The exact language is preserved in the Federal Register. The industry sued and lost. On October 17, 1994, FSIS began sampling raw ground beef for the organism, and in 1998 it raised the sample size from 25 grams to 325 grams because the smaller sample was missing low-level contamination. After the 2002 ConAgra recall in Greeley, the agency ordered every ground beef plant to reassess its controls.

On June 4, 2012, six more Shiga toxin-producing strains — O26, O45, O103, O111, O121 and O145 — became adulterants in raw non-intact beef, a change Marler Clark had petitioned for in 2009. In June 2020 FSIS extended that testing from beef manufacturing trimmings to ground beef, bench trim and other components.

None of that happened on a farm. Every piece of it happened inside a federally inspected establishment, with federal personnel standing there to take the sample, hold the lot and stop the shipment. Take the inspector out and the entire apparatus has nothing to attach to.

A President cannot write a new exemption into the Federal Meat Inspection Act.

The inspection requirement is statutory, and so are the exemptions. For livestock there is one meaningful exemption, custom slaughter under 21 U.S.C. § 623, and the product goes back to the animal’s owner marked not for sale. FSIS says plainly in its own guideline that inspection is required, unless exempted, for meat from cattle, sheep, swine and goats, and it lists every exemption there is: personal use, custom slaughter, retail stores, restaurants, and slaughter in an unrecognized Territory. Selling uninspected beef to the public is not among them. Changing that takes Congress.

FSIS says absorbing 2,000 more plants would compromise its food safety mission.

Set aside the law for a moment and look at the arithmetic, starting with what thinning this system already did. Juan Vassallo reported for Investigate Midwest in May that complaints about the safety of meat, poultry and egg products rose nearly 40 percent in a single year, from 1,443 to 2,016, in a year when USDA cut its workforce by 18 percent and FSIS lost about 9 percent of its staff. That is this system with an inspector still in every plant, only fewer of them. FSIS spends about $1.24 billion a year to cover roughly 7,100 federally inspected establishments. The agency’s own fiscal 2027 budget justification warns that state reimbursement rates are falling toward 38 percent, that states may pull out of their inspection programs, and that taking on an estimated 2,000 additional establishments could severely strain its resources and compromise its food safety mission.

Two thousand plants is described by the agency as an existential problem. There are roughly 1.9 million farms in this country. Nobody is sending an inspector to the ranch, and nobody is proposing to. The only way to give a farmer the right to process his own beef for sale is to delete the requirement that anyone inspect it.

An immaculately clean steer carries E. coli O157:H7 in its intestines.

The President’s post says ranchers are immaculately clean, and I have no doubt most of them are. That has never been the point, and the confusion between the two is the oldest one in this field. E. coli O157:H7, Salmonella and Campylobacter live in the intestinal tracts of healthy cattle on clean farms. The organism reaches the carcass off the hide and the gut during dressing, and grinding then takes what was on the surface and distributes it through the interior where cooking may not reach it. As few as ten organisms can make a child sick. None of it is visible.

Everything the industry built — hide-on interventions, carcass washes, trim testing, test and hold — exists because what you can see on an animal has almost nothing to do with the pathogen load you cannot. A rancher’s good character has never been a kill step.

Meat with no establishment number cannot be traced or recalled.

This is the part that worries me most, and it gets almost no attention. A grant of inspection carries an establishment number, production records and lot identity. That is what makes a traceback possible, and a traceback is what makes a recall possible. Product slaughtered on a farm and sold down the road has none of it.

Illnesses from that product would be sporadic and scattered, arriving one at a time in counties that never compare notes, the kind of cluster that never crosses the threshold to be recognized as an outbreak in the first place. Most of them would never be counted. When somebody’s child ends up on dialysis, the first thing the family loses is any ability to find out why — and the second is any way to keep it from happening to the next family.

The consolidation complaint is real, and custom slaughter is not the fix.

I want to be fair about what sits underneath the President’s post, because it is not made up. The administration’s own figures, which an independent fact-check confirmed, put the four largest packers at about 85 percent of beef processing, up from 36 percent in 1980, and independent producers really do wait months for a kill date. That is a genuine problem with genuine remedies. The Cooperative Interstate Shipment program lets state-inspected plants under twenty-five employees ship across state lines with an inspector still on the floor. USDA’s Small Processors Action Plan put $60 million into processing expansion grants and cut overtime inspection fees. Those tools exist and are underfunded. Agriculture Secretary Brooke Rollins said Friday that a slate of beef processing announcements begins Monday, and Noah Wicks and Olivia M. Bridges of Agri-Pulse published the full list she posted: waiving red tape in processing, rescinding outdated guidance, growing support for small processors, and expanding ranchers’ ability to sell across state lines. That last one is the Cooperative Interstate Shipment program, and it is the right direction to move in.

What binds a small plant is capital, labor and scale. It is not the inspector. Removing the inspector does not build a plant, does not add a kill line and does not move the price of ground beef. It only removes the one person in the building whose job is the public’s.

I am not alone in this. The Meat Institute, the trade association for the very packers the President’s post calls a monopoly, came out against the idea the same afternoon, saying that allowing “uninspected meat to be sold to unwitting consumers” is the wrong approach and risks this country’s reputation for producing the safest meat in the world. It answered the price rationale in the same breath: if the goal is lower beef prices for American families, lowering food safety standards is the wrong response. Its fuller answer, reported by Noah Wicks and Olivia M. Bridges at Agri-Pulse, goes to the premise rather than the risk: farmers and ranchers already have several routes to process and market the livestock they raise, through the custom exemption or by opening a state or federally inspected plant of their own. The post assumes the door is closed. It is not.

The National Cattlemen’s Beef Association was blunter in the same story. Weakening federal meat inspection and food safety standards, it said, “is not the answer,” and beef sold to American consumers should go on meeting rigorous, science-based standards no matter the size of the processor. Asked what would actually help, the cattlemen named lower fuel and fertilizer costs, protection of the herd against foreign animal disease, and expanded opportunities for mid-size and regional processors. That is the same place I arrived from the other direction. It is not often that the packers, the cattlemen and the plaintiffs’ lawyer land together on a question, and it is worth noticing when we do.

I spent nine years making a living on hamburger, and I have spent the twenty-four years since telling anyone who would listen that the beef industry met the challenge and that losing the work was the best outcome I could have asked for. I meant it then and I mean it now. I would rather not get it back.

On August 11, the FDA published its proposed rule on Substances Generally Recognized as Safe, 91 FR 51834, Docket No. FDA-2025-N-3262, with comments due December 9, 2026. HHS announced it the day before as a landmark food policy reform. The fact sheet describes it simply: manufacturers would have to tell FDA when they decide on their own that an ingredient is safe. 

Page 51841 provides the agency’s example. In 2022, a popular at-home food-delivery company used tara flour in a human food product and recalled it after roughly 400 adverse event reports describing gastrointestinal distress, hepatotoxicity, and hospitalization. The company conducted its own root cause analysis and identified tara flour. FDA asked for records showing that the company had reached a GRAS conclusion on the ingredient and received none. Tara flour had never been the subject of a GRAS notice, or even a pre-submission meeting.

The company was Daily Harvest. Those roughly 400 adverse event reports are our clients. Marler Clark represented 339 of the approximately 470 people sickened by Daily Harvest’s French Lentil + Leek Crumbles[1], as well as 33 people sickened by Revive Superfoods’ Mango & Pineapple Smoothies—another product sold in the U.S. and Canada that contained tara flour. Nearly forty of them lost a gallbladder. 

On December 27, 2023, I wrote to Jim Jones, then FDA’s Deputy Commissioner for Human Foods[2], and laid out the chain: tara flour milled by Molinos Asociados in Peru, imported by Smirk’s in Colorado, put into the Crumbles by Stone Gate Foods in Minnesota, and sold to consumers who had no way to know that the ingredient had never been evaluated for safety in the United States. 

We asked FDA to determine whether anyone had ever sought GRAS status for tara flour. Our search of FDA records showed that nobody had. We also asked what Smirk’s had done to satisfy its Foreign Supplier Verification Program obligations, given that FDA’s own CORE incident summary said the importer had an FSVP plan for tara protein flour but that more information was needed. And we pointed FDA to what Canada had already done. On September 28, 2023, the Canadian Food Inspection Agency told Canadian businesses to stop buying and selling tara flour as an unassessed novel food[3].

Credit where it is owed. Ilana Korchia in our office built the regulatory spine of that work[4]. She is the one who took a mass personal injury case about a bag of frozen lentil crumbles and turned it into a question about 21 U.S.C. §§ 321(s), 342, 348 and 384a: whether the ingredient itself was lawful and who in the supply chain was responsible for verifying that it was. 

On April 10, 2024, FDA’s Division of Food Ingredients issued a memorandum on tara flour derived from the seed germ of Caesalpinia spinosa[5], concluding that the use of tara flour in food did not meet the criteria for general recognition of safety; that there was no common use in food before 1958; that the available toxicity evidence was insufficient to establish safe use; and that, as a result, the ingredient was an unapproved food additive and food containing it was adulterated under section 402(a)(2)(C)(i) of the Federal Food, Drug, and Cosmetic Act. FDA posted that determination on May 15, 2024, added tara flour to its post-market not-GRAS inventory, and began screening it at ports of entry. 

That brings us back to the proposed rule. FDA is now using tara flour as its example of why the voluntary GRAS system is not enough.

The proposed rule is worth supporting, and I will file a comment saying so. It would make notification mandatory rather than optional, requiring companies to put their GRAS determinations on the public record. Notices would appear in a public, searchable GRAS notice inventory. That public record matters. A filed notice, or the conspicuous absence of one, is evidence. When a company later tells a jury it had no reason to know its ingredient was unsafe, the record will tell a different story.

But this is not premarket approval, and FDA says so plainly. A company can put a substance into interstate commerce and notify afterward. It can keep selling while its notice sits in the queue. It can keep selling even if it never files a notice at all. The stated consequence for skipping the requirement is simply that the violation becomes a factor in how FDA prioritizes substances for post-market review. 

Run the facts of 2022 through the new rule and ask what actually changes. Daily Harvest still designs the product. Smirk’s still imports the flour. Stone Gate still puts it into the Crumbles. The product still ships in April. People still start getting sick in June. Somewhere in there a form gets filed, or not, and the only difference is that afterward we would have a better paper trail for the lawsuits. That is progress, not prevention.

The timeline is the second problem. Comments close in December. A final rule follows, whenever it follows. It takes effect 60 days after publication. Compliance with the notification requirement is proposed for 18 months after that. On the most optimistic schedule, the first year in which anyone is actually obligated to notify FDA about an ingredient is 2029. 

The “legacy” pathway is the third problem. For substances already on the market under a self-affirmed conclusion, a streamlined submission asks for the identity of the substance, its conditions of use, and evidence it was already in commerce. It does not require the underlying safety documentation. A company that quietly self-affirmed an ingredient in 2015 can register it in 2029 without ever showing anyone why it believed the ingredient was safe.

Then there is the import problem, which is where our clients actually got hurt. Of the more than 1,200 GRAS notices FDA has filed, 617 came from foreign firms and 503 of those from countries where English is not the primary language. Tara flour traveled from a mill in Peru through an importer in Colorado to a manufacturer in Minnesota, and the law already required somebody in that chain to verify it. FSVP has been on the books since FSMA. It did not stop this. Layering a notification duty on top of a verification duty that no one audits does not produce verification. 

What I will ask for in the comment period is straightforward. Notice should come before introduction into commerce, not after. Noncompliance should carry a stated legal consequence rather than a place in a prioritization queue. The streamlined pathway needs a minimum evidentiary requirement and a hard deadline to convert to a full notice. FDA also needs reviewers, and Congress needs to fund them. Finally, the public inventory should be designed for the public—not merely searchable, but understandable and usable by a parent standing in a grocery aisle.

Comments are due December 9.

[1]Daily Harvest, Inc., Daily Harvest Issues Voluntary Recall of French Lentil + Leek Crumbles Due to Potential Health Risk(company announcement posted by FDA, June 23, 2022), and Daily Harvest, Updates on Our Voluntary Recall of French Lentil + Leek Crumbles (advisory page, updated May 31, 2023).

[2]Letter from William D. Marler to James “Jim” Jones, Deputy Commissioner for Human Foods, U.S. Food & Drug Administration, Re: Status of Tara Flour, a Novel Ingredient not Generally Recognized As Safe, in the United States (Dec. 27, 2023). The letter collects the underlying record as attachments, including the CFIA notice, the FDA recall announcement, FDA’s CORE incident summary, and the Chittiboyina, Chan & Smith, and Choi et al. publications.

[3]Canadian Food Inspection Agency, Notice to Industry: Tara Protein Powder (Tara Flour) Not Assessed for Safety by Health Canada (Sept. 28, 2023), Attachment No. 1 to the December 27, 2023 letter. 

[4]Marler Clark memorandum, Foreign Supplier Verification Program and Food Additive Authorities (May 14, 2024) (collecting 21 U.S.C. §§ 321(s), 342, 348, and 384a).

[5]Memorandum, U.S. Food & Drug Administration, Center for Food Safety and Applied Nutrition, Office of Food Additive Safety, Division of Food Ingredients, Regulatory status and review of available information pertaining to tara protein/flour derived from the seed germ of the plant, Caesalpinia spinosa: lack of general recognition of safety for its use in foods (Apr. 10, 2024).

On June 23 and June 29, 2026, I wrote to the House Committee on Energy and Commerce and to the Senate Committee on Health, Education, Labor, and Pensions about Clostridium botulinum in powdered infant formula — 48 hospitalized infants in the ByHeart outbreak, four more in the Nara Organics outbreak, and a federal rule that still requires no one to test infant formula for the organism that paralyzed them. Neither letter drew a reply. On August 27, 2026, the federal government confirmed the central factual claim in both of them, and confirmed it with a laboratory result rather than an inference. On August 28 I wrote to both committees again. All four letters are below.

I asked both committees for the same three things

The June 23 letter went to Chairman Brett Guthrie and Ranking Member Frank Pallone, Jr., in support of H.R. 7867, the Infant Formula Safety Modernization Act of 2026, which Representative Rosa DeLauro introduced in March and which received a Subcommittee on Health legislative hearing on April 29. I asked the Committee to mark the bill up and report it favorably, to hold an oversight hearing, and to put the parents of the injured infants at the witness table. That letter is published here, and the heart of it is not my argument. It is Appendix A — fourteen families describing, in their own words and with their children’s photographs, what infant botulism did to a two-month-old.

The June 29 letter went to Chairman Bill Cassidy and Ranking Member Bernie Sanders, and it began with credit, because the Senate had earned it. On April 29 the HELP Committee reported the Protect Infant Formula from Contamination Act, S. 272, by a vote of 22 to 0, and the full Senate passed it the same day by unanimous consent. In a Washington that agrees on almost nothing, that is not a small thing. My point to that Committee was narrow: S. 272 requires one-business-day notice only for the organisms already on the books — Cronobacter and Salmonella — and does not reach the one that put these babies in intensive care. That letter is published here.

Neither letter was answered, and that is not what this post is about

Congressional committees receive an enormous volume of mail, most of it from people with an interest in the outcome, and I am one of them. A letter from a plaintiffs’ lawyer has no claim on anyone’s reply, and I am not writing to complain about the silence. I am writing again because the record changed. In June I could tell these committees what I believed the evidence would show. In August the government showed it.

On August 27 the government said the two outbreaks were one contamination

The Centers for Disease Control and Prevention declared the Nara Organics outbreak over. The final count is four infants — two in California, one in Pennsylvania, one in Washington — with illness onsets on April 20, May 17, May 20 and May 31, 2026. All four were hospitalized. None died. CDC found that epidemiologic, traceback and laboratory data showed that Nara Organics Whole Milk Organic Infant Formula was contaminated with Clostridium botulinum and made infants sick. Its earlier update had said only that the formula might be contaminated. The agency dropped the qualifier.

Then came the finding that ought to move this from a lawyer’s argument to a committee calendar. Whole genome sequencing linked a patient sample from the Nara outbreak to material collected during the investigation of the 2025 ByHeart outbreak — a patient sample, product samples and an ingredient sample. FDA describes the same result more specifically, reporting a match to a ByHeart clinical sample, a powdered infant formula base sample, and four samples of unopened cans of ByHeart formula. CDC states that FDA’s traceback determined that the formulas sold by both brands were made with milk supplied by Organic West Milk and spray dried by Dairy Farmers of America, and that the matching sequences indicate a common source of contamination for the two outbreaks.

In June I wrote that the contamination did not strike two unrelated companies by chance, and that it traveled through one shared supply stream the system missed twice. The support for that sentence was a traceback and a trade-press report. The support for it now is the government’s own sequencing. The same organism has been found in an infant fed one brand and in the product and the ingredient of the other. A strain does not move between two babies who never consumed the same product by accident. It moves through the thing they shared. I have been making that argument about this ingredient since June, and I would rather have been wrong.

One question the sequencing does not answer, and I would rather say so myself

Neither agency has determined a root cause. FDA says it cannot yet tell whether the milk was contaminated before it was dried or during the drying process. That question decides how responsibility is divided among the brands, the milk supplier and the processor, and it is the question I expect to spend the next two years on. It does not touch whether the formula was contaminated, and it does not touch whether the formula paralyzed babies. CDC decided both of those on August 27.

My June letters undercounted, and the correction runs the wrong way

Both June letters said three Nara infants and fifty-one babies in total. The agencies added a fourth Nara case as the investigation continued, so the correct totals are four and fifty-two. Both follow-up letters state the correction openly rather than quietly restating the number, and I wrote about the fourth case when it was added. Fifty-two infants hospitalized across two outbreaks, many of them on ventilators, each treated with BabyBIG antitoxin that the State of California prices at $69,300 a dose.

The sequencing speaks to four specific provisions of the House bill

H.R. 7867 is not a general expression of concern. It does six concrete things, and the August 27 result speaks directly to four of them. The testing list: federal regulation still requires finished-formula testing only for Salmonella and Cronobacter and says nothing about Clostridium botulinum, so the organism now sequenced across two brands, two outbreaks and fifty-two hospitalized infants is the one organism nobody is required to look for. Environmental monitoring: a single strain persisting through this ingredient stream across two manufacturers, a national recall, a federal investigation and a formal closeout is not an incidental event — it is a resident organism, and Zone 2 and Zone 3 monitoring is how a resident organism gets found. Foreign manufacturers: the Nara formula was made abroad, its milk came from an American supplier, it was dried by an American processor, and the sequencing has now tied it to an American outbreak — the route the bill anticipated is the route the organism actually took. Notice to Congress: consider what this committee would have known in 2025 had that provision been law, and how much of what I am reporting today you would already have had.

The mechanism built to prevent a second outbreak failed on a customer list

During the 2025 ByHeart investigation the FDA did what the playbook calls for. It ran a trace-forward and asked Organic West Milk who else it supplied, so the next brand could be found before the next baby was hurt. The customer list did not include Nara Organics, and Organic West has since told the FDA that it should have. ByHeart recalled all of its product on November 11, 2025. CDC declared that outbreak over on February 26, 2026. The first Nara infant became ill on April 20, 2026 — 160 days after the recall, and 53 days after the government told the country the first outbreak was finished. The one mechanism designed to prevent a second outbreak existed, the agency used it, and an incomplete answer defeated it. We now know at the level of the genome exactly what got through. I have been asking what Nara knew and when it knew it since June, and this is the part of the record that answers a different question — what the government asked, and what it was told.

The three asks are unchanged, and there is now a fourth

First, pass a bill that reaches botulism. The House Committee should mark up H.R. 7867 and report it favorably; the hearing is behind it and the next step is procedural. The Senate should take up a companion measure or write its own. Second, hold the oversight hearing. Bring the FDA, Dairy Farmers of America, Organic West Milk, Nara Organics and ByHeart to the table. The questions are sharper than they were in June because there is now a laboratory result to ask about: how long has this organism been in this ingredient stream, who tested for it, by what method and how often, and who prepared the 2025 customer list, who reviewed it, and when did each company learn that the same milk had reached a second brand. Third, put the parents at the witness table. They are willing to do more than write, and no expert and no lawyer, myself included, can substitute for the parent who lived it.

The fourth is new. Put a written question to the FDA, with a deadline. The agency’s root cause investigation has been open since November 2025, and its own public page on this subject has not been updated since July 13, 2026 — it still says there is not enough evidence to determine whether the shared suppliers are the source of the contamination, language its own sequencing has now overtaken. Ask the Commissioner what the root cause investigation has found, whether the agency intends to require testing or environmental monitoring for Clostridium botulinum by regulation, and if so on what schedule. A date certain would tell Congress whether legislation is necessary or merely useful. The absence of one would tell you something too. I made a version of this point directly to the agency in June, and the answer so far has been letters.

We have done this before

The Jack in the Box tragedy moved USDA to declare E. coli O157:H7 an adulterant in ground beef in 1994 — one decision that has kept countless children out of hospital beds in the decades since. That was a rule. What the powdered infant formula industry has received on botulism is a series of letters, the first of them sent on March 8, 2023, naming the organism by genus and species, three and a half years ago. Letters are not a rule. Fifty-two families already know what the sequencing means, and they would like to know that Congress does too.

My thanks to both committees for their work on infant formula this year, and my standing offer: I will answer any question from either committee or its staff, and I will make the families available to them.

The four letters

https://www.marlerblog.com/files/2026/08/HR_7867_Support_Letter_6_23_26_WDM_FINAL.pdf

Letter to the House Committee on Energy and Commerce, June 23, 2026, with Appendix A — statements and photographs of fourteen families.

https://www.marlerblog.com/files/2026/08/Marler_Senate_HELP_Letter_Letterhead.pdf

Letter to the Senate Committee on Health, Education, Labor, and Pensions, June 29, 2026.

https://www.marlerblog.com/files/2026/08/HR7867_Follow_Up_Letter_8-28-26.pdf

Follow-up letter to the House Committee on Energy and Commerce, August 28, 2026.

https://www.marlerblog.com/files/2026/08/Senate_HELP_Follow-Up_Letter_8_28_26_WDM.pdf

Follow-up letter to the Senate Committee on Health, Education, Labor, and Pensions, August 28, 2026.

Two different infant formula brands, made by different companies on different continents, sickened infants with Clostridium botulinum seven months apart. What the two products had in common was not a factory. It was an ingredient. FDA’s traceback determined that the ByHeart formula implicated in the November 2025 outbreak and the Nara Organics formula implicated in the 2026 outbreak were both made with organic whole milk supplied by Organic West Milk and spray dried into powder by Dairy Farmers of America. On August 27, 2026, the laboratory work caught up to the paperwork: whole genome sequencing showed that C. botulinumisolated from an infant sickened by Nara formula is a genetic match to C. botulinum isolated from an infant sickened by ByHeart formula, to a powdered infant formula base sample collected in the ByHeart investigation, and to four samples of unopened cans of ByHeart formula. The same organism reached babies through two brands. The only place the two supply chains overlap is upstream of both formula plants — at the milk and at the dryer.

The production point of view.

Follow the powder forward. Organic West Milk supplies raw organic whole milk. Dairy Farmers of America takes that liquid milk, and spray dries it into organic whole milk powder — the bulk dairy base that a formula manufacturer buys by the tote and blends with oils, carbohydrates, vitamins and minerals to make finished infant formula. That powder then goes to whichever formula maker bought it. In 2025 it went to ByHeart, whose formula is made in the United States. In 2025 and 2026 it also went into the Nara Organics product, which FDA says was manufactured in Europe at two firms the agency had inspected before these illnesses, issuing observations citing deficiencies to both. Two brands, two sets of formula plants, two countries of manufacture — and one shared dairy base flowing into both.

That is why the shared ingredient matters more than anything either brand did on its own canning line. A contamination event at a formula plant contaminates that plant’s product. A contamination event in the milk, or in the drying step that converts milk to powder, contaminates every downstream customer that bought from the affected production — regardless of how clean the formula plant is, whose name is on the label, or which continent the can was filled on.

Production stepByHeart (Nov. 2025 outbreak)Nara Organics (2026 outbreak)
Raw organic whole milkOrganic West MilkOrganic West Milk
Spray drying into whole milk powderDairy Farmers of AmericaDairy Farmers of America
Formula manufacture and canningByHeart, United StatesTwo firms in Europe
Brand and labelByHeart Whole Nutrition Infant FormulaNara Organics Whole Milk Organic Powdered Infant Formula
RetailNational, incl. Target, Walmart, KrogerNational, Target stores, Target.com, Nara.com
Reported illnesses48 infants, 17 states4 infants, 3 states

Sources: FDA, Post-Outbreak Response Activities (July 13, 2026); FDA outbreak advisory updates of June 26, July 6 and August 27, 2026; CDC outbreak pages for both investigations. The two chains are identical through the dryer and diverge only after it.

Two features of this product make an upstream contamination event carry all the way to a bottle. First, C. botulinum forms spores, and FDA states plainly that spore-formers survive typical pasteurization and harsh environmental conditions — so heat treatment of the milk upstream does not remove them. Second, powdered infant formula is not a sterile product and has no terminal kill step after blending and canning; the finished powder is not retorted or otherwise treated in a way that would destroy surviving spores. Contamination that is present in the milk or introduced during drying is therefore not corrected by anything that happens later. It is packaged.

FDA has not resolved which of those two points the organism entered at. Its position on August 27 remains that there is not enough evidence to determine whether the milk was contaminated before drying or during the drying process, and its root cause investigation stays open with a focus on ingredients. That is an unresolved question between two defendants’ operations, not a gap in the link between the two outbreaks.

The laboratory point of view — what whole genome sequencing actually showed.

Whole genome sequencing reads the full DNA of a bacterial isolate. Two isolates that are indistinguishable, or separated by only a few base differences, are the same strain — which in practice means they came from the same contamination event rather than from two unrelated encounters with an organism that is common in soil and dust. It is the difference between saying two products shared a supplier and saying two products shared a contamination.

What was published in the ByHeart investigation:

  • FDA’s inspections of the ByHeart facilities identified C. botulinum in a powdered milk ingredient.
  • FDA and ByHeart sampled dairy ingredients at Dairy Farmers of America. Two samples from one lot of organic whole milk powder were shown by WGS to match a clinical sample and a positive infant formula sample. That is a closed chain: milk powder at the dryer, to finished formula, to a sick baby.

What FDA added on August 27, 2026:

  • One clinical sample from an infant in the Nara outbreak is a genetic match to a clinical sample from an infant in the 2025 ByHeart outbreak; to a powdered infant formula base sample collected during the ByHeart investigation; and to four samples of unopened cans of ByHeart formula.

The word doing the most work there is unopened. A positive from an opened can invites the argument that a household, a scoop or a countertop introduced the organism after purchase. Sealed cans foreclose it. The strain was in the product as it left the manufacturer.

Put the two publications together and the chain runs the length of the supply. The Nara infant’s isolate matches the ByHeart infant’s isolate. The ByHeart infant’s isolate was previously reported to match organic whole milk powder sampled at Dairy Farmers of America. If the clinical sample described in the August 27 update is the same patient isolate described in the earlier one — which FDA has not stated, and which is the first thing to confirm — then the organism recovered from a baby fed Nara formula is indistinguishable from the organism recovered from a lot of milk powder at the dryer that served both brands. That is the inference the published record supports. It should be pled as an inference and confirmed with the isolate identifiers, the NCBI accessions and the SNP distances before it is stated as a fact.

Sources.

CDC declared the Nara Organics infant botulism outbreak over on August 27, 2026. Four babies in three states — California, Pennsylvania and Washington — were sickened by Nara Organics Whole Milk Organic Powdered Infant Formula, with illness onsets running from April 20 through May 31, 2026. All four were hospitalized. None died. Nara recalled every lot on June 13. And the laboratory work that closed the file ties one of those four infants to the ByHeart formula outbreak that CDC had already declared over three months earlier.

Four babies, four hospitalizations, and no case added since early July.

The case count has not moved since the July 6 update. FDA says no new case has been added since July 3; CDC says none has been added since its own July 6 update. Either way the investigation sat still for roughly eight weeks before CDC closed it. The four illness onsets are not evenly spaced, and the gap in the middle is the part worth looking at.

Illness onset dateSick infantsDays since previous onset
April 20, 20261—
May 17, 2026127
May 20, 202613
May 31, 2026111
Total441-day onset span

Source: CDC, “When People Got Sick: Infant Botulism Outbreak, June 2026,” downloadable data file, retrieved August 27, 2026. Interval column computed from the onset dates in that file. Check: 1 + 1 + 1 + 1 = 4, which matches the total of 4 cases CDC and FDA both report.

One baby got sick on April 20 and then nothing for 27 days. Three more got sick inside a two-week window in late May. The recall came on June 13 — 54 days after the first baby got sick, and 13 days after the last one.

CDC’s own data file puts two of the four cases in California. FDA’s map caption says one.

CDC’s locations page maps four infants across three states. The file behind that map has California at two. FDA’s August 27 advisory reports the same four cases from the same three states and then describes its map as one case in Washington, one in Pennsylvania and one in California. That adds to three. The count is four; the caption is wrong, and the caption is what a reporter reads.

StateSick infantsCDC display range
California21 to 5
Pennsylvania11 to 5
Washington11 to 5
Total43 states

Source: CDC, “Where Sick People Lived: Infant Botulism Outbreak, June 2026,” downloadable data file, retrieved August 27, 2026. Check: 2 + 1 + 1 = 4, matching the reported total. FDA’s narrative map caption on the same date describes one case per state, which sums to 3.

The strain from one sick Nara baby matches a ByHeart baby, ByHeart base powder, and four unopened cans of ByHeart formula.

This is the finding that matters. FDA reports that whole genome sequencing of one clinical sample from an infant in the Nara outbreak is a genetic match to a clinical sample from an infant in the 2025 ByHeart outbreak, to a powdered infant formula base sample collected during the ByHeart investigation, and to four samples of unopened cans of ByHeart formula. One organism. Two outbreaks. Two brands. Seven months apart.

Read the match set carefully, because the shape of it tells you what is still missing. Everything the Nara baby’s isolate matched is on the ByHeart side of the ledger — a ByHeart patient, ByHeart base powder, ByHeart cans. FDA does not report a match to any Nara product sample, and it does not say the Nara can it was testing in July has been sequenced at all.

Both brands’ formula was made from milk supplied by Organic West Milk and dried by Dairy Farmers of America.

FDA’s traceback determined that the Nara lots linked to these illnesses were made with milk supplied by Organic West Milk and spray dried by Dairy Farmers of America — the same milk supplier and the same dryer behind the ByHeart formula that sickened 48 infants in 17 states. FDA says the matching sequences and the shared supply chain together indicate a common source of contamination, and then says the thing everyone should sit with: there is still not enough evidence to determine whether the milk was contaminated before drying or during the drying process. Nine months into the first outbreak, the agency cannot yet say which side of the spray dryer the C. botulinum came from.

This is the whole-milk-powder problem I have been writing about since the first outbreak. Both products were built on it, and FDA warned this entire industry in writing in March 2023 that C. botulinum belonged on its risk list.

Nara was left off the customer list Organic West gave FDA during the ByHeart investigation.

During the 2025 outbreak, FDA ran a trace-forward to find out which other formula makers were getting Organic West milk. By FDA’s own account, Organic West produced an incomplete customer list that did not show its milk going to the manufacturers of Nara Organics formula. Organic West told FDA during this investigation that the earlier list was incomplete and should have included Nara.

Put the dates next to each other. ByHeart broadened its recall to all product on November 11, 2025. CDC declared the ByHeart outbreak over on February 26, 2026. The first Nara baby got sick on April 20, 2026 — 160 days after the ByHeart recall and 53 days after federal officials announced that outbreak was finished. A complete customer list in November would have pointed investigators at a second brand drawing from the same milk while there was still time to do something about it. That is not hindsight. That is the ordinary purpose of a trace-forward.

FDA still has not said what the unopened Nara can from the same lot showed.

In July, FDA reported that C. botulinum had been confirmed in an open can of Nara formula fed to one of the sick infants, and that it was testing an unopened sample from the same lot with results expected in the coming weeks. The closeout says samples were collected from patients, from finished product and from ingredients — and then reports a single clinical match. No finished-product result. No ingredient result. If the unopened can came back positive, that is a fact families and pediatricians should have. If it came back negative, that is a fact the industry should have, because it goes straight to how well anyone can sample their way out of this hazard.

One more gap. The FDA advisory tells readers to see its Post-Outbreak Response Activities page for more information. That page is stamped current as of July 13, 2026 — 45 days ago — and still says there is not enough evidence to determine whether the shared suppliers are the source or route of contamination. It has not caught up to the sequencing result FDA announced today.

The recalled formula carries Best By dates into 2027, which means some of it is still in kitchens.

Powdered formula does not spoil in any way a parent can see, and the recalled product carries Best By dates ending in 2027. It was sold nationwide through Target stores, Target.com and Nara.com between July 2025 and June 2026, and all of it was recalled on June 13. An outbreak being over does not make a leftover can safe.

  • Do not feed it to your baby. Throw it away, or return it.
  • Before you discard a container, photograph the label, the lot number and the Best By date.
  • Wash anything that touched the formula — bottles, counters, scoops, dispensers — in hot soapy water or the dishwasher.
  • Know the signs: constipation, poor feeding, a weak or altered cry, loss of head control, difficulty swallowing, decreased facial expression. Get care immediately. Physicians should call the Infant Botulism Treatment and Prevention Program at 510-231-7600, available around the clock.

Questions parents are asking today.

I still have a can of Nara Organics formula. Is it safe now that the outbreak is over?

No. The recall covers every lot, and the closing of an outbreak investigation says nothing about the can in your cupboard. Throw it away.

My baby drank this formula. How long do I have to watch?

Symptoms of infant botulism can take as long as several weeks to appear. CDC’s guidance is to watch for a month after your infant last consumed the product, and to seek care immediately for poor feeding, loss of head control, difficulty swallowing or a decrease in facial expression. The clinical overview is here.

Is any formula on the shelf right now affected?

No other brand has been named. Recalled Nara product should no longer be for sale anywhere. FDA’s root cause investigation is still open and is focused on ingredients, which is the honest answer to a question about what else might be involved: the agency does not yet know.

Two brands, one strain, and a testing rule that still does not name this organism.

I represent families in both the ByHeart and the Nara outbreaks, and we have sued over this formula, so read me accordingly. But you do not need a stake in it to follow the sequence. The same milk went into two brands. The list that would have connected them was incomplete. A strain from a can of one brand turned up in a baby fed the other. And the federal regulation that tells formula makers what to test finished product for still names Salmonella and Cronobacter and does not name C. botulinum.

Four babies is a small number and a total one if you are one of the four. What I want out of this closeout is not a headline. It is a complete customer list every time a regulator asks for one, a finished-product and ingredient testing requirement that names the organism that keeps showing up, and a supplier that cannot hand two brands the same problem while the file on the first one is being closed. I have asked Congress for exactly that. The lab result FDA published today is the strongest argument for it yet.

The fifty-state Cyclospora floor stands at 32,506 tonight, August 26, 2026, and the news is in a state that has never been part of this outbreak. The Hawaii Department of Health announced today that it is investigating a sixth case of cyclosporiasis, an Oʻahu resident who became ill at the end of July and did not travel outside Hawaiʻi during the period when they were likely exposed. No suspected food source has been identified, and the department says plainly that no recalled Taylor Farms de Mexico lettuce was distributed to or sold in the islands. The five earlier Hawaii cases this season were all travel-associated.

That reason is gone, so all six are counted here and the total rises by six. North Carolina published today as well, at 940 cases statewide with hospitalizations unchanged at 32 and the source still unknown, which is the same 940 my chart was already carrying because the row was right and the date on it ran a week early. Federal figures have not moved since Tuesday. CDC counts 17,180 laboratory-confirmed domestically acquired cases with data current through August 24, against 1,180 for the same stretch of 2025, and the outbreak investigation stands at 10,930 illnesses in seventeen states with at least 454 hospitalizations and two deaths. Michigan reports tomorrow, and the season closes on Sunday.

Here is where all fifty states stand tonight. The Lettuce column is whether FDA has confirmed that recalled Taylor Farms de Mexico lettuce reached that state. The last column is the different and more important question: whether anybody has tied that state’s illnesses to Taylor Farms at all.

State2026 casesAs ofAll of 2025LettuceTaylor Farms
Michigan14,277Aug 2050YesLinked
Ohio5,452Aug 1376YesLinked
Missouri2,152Aug 23none reportedYesLinked
Indiana2,069Aug 1927YesLinked
Illinois1,437Aug 18301YesLinked
North Carolina940Aug 26300YesLinked
New York939Aug 17694YesUnknown
Kansas807Aug 1929YesLinked
Kentucky620Jul 2945YesLinked
Oklahoma544Aug 1841YesLinked
Florida407Aug 8210YesUnknown
Iowa327Aug 2064YesLinked
Texas315Aug 11512YesUnknown
West Virginia313Aug 218YesLinked
Arkansas287Aug 2413YesLinked
Virginia234Aug 24115YesUnknown
Nebraska218Aug 651YesLinked
Wisconsin216Aug 565YesUnknown
Colorado150Jul 17205NoUnknown
Massachusetts145Aug 574YesLinked
Alabama85Aug 319YesUnknown
Pennsylvania82Jul 29not notifiableYesLinked
Washington71Aug 21not availableMay haveTravel link
Maryland69Jul 17123YesNo link found
New Hampshire69Aug 47YesLinked
New Jersey46Jul 11180YesUnknown
California41Jul 14not availableMay haveUnknown
Minnesota41Jul 17none reportedNoUnknown
Connecticut35Jul 1741YesUnknown
Maine27Aug 74YesLinked
Oregon23Jul 24not availableNoTravel link
Arizona19Jul 1749NoUnknown
Georgia11Jul 17116YesUnknown
Tennessee11Jul 1738YesUnknown
Mississippi7Jul 22not notifiableYesUnknown
Hawaii6Aug 26not availableNoNo link found
Alaska5Jul 166NoUnknown
Rhode Island4Jul 176May haveNo link found
South Dakota3Jul 163May haveUnknown
Louisiana1Jul 1799YesUnknown
Utah1Jul 1737NoNo link found
Delawareno count published—1May haveUnknown
Idahono count published—not notifiableNoUnknown
Montanano count published—8NoUnknown
Nevadano count published—not notifiableNoUnknown
New Mexicono count published—9NoNo link found
North Dakotano count published—4May haveUnknown
South Carolinano count published—45YesUnknown
Vermontno count published—3May haveUnknown
Wyomingno count published—0NoUnknown
All fifty states32,506  3117

Counts as published by each state health department, compiled from the linked sources on August 26, 2026. Forty-one states publish a 2026 count and those forty-one sum to 32,506. Hawaii is counted for the first time here, because the five travel-associated cases that kept it out of the sum are now six and the sixth reports no travel. Nine states publish no count at all. Every figure is a floor.