In February 2016, investigators found the culprit fast. People in several states were sick with salmonellosis, all of them remembered eating alfalfa sprouts, and tracebacks from the restaurants led to one company: Sweetwater Farms of Inman, Kansas. On February 19 the Kansas health department warned consumers, and on February 26 the firm pulled all of its sprout products off the market.
People kept getting sick anyway.
The new cases led to other sprouters. By the end of the investigation it turned out that all of those companies, along with Sweetwater Farms, had used one common lot of alfalfa seed. FDA tested seed from that lot and found Salmonella Cubana carrying the same DNA fingerprint as the Salmonella in Sweetwater’s irrigation water. Final count: 26 people sick in 12 states, eight of them hospitalized.
That leaves one plain question. The lot had already scattered across several operations with the pathogen in it. Whose job was it to test that lot before it scattered? The American answer and the European one differ, and the difference is the whole subject of this article.
Seed is the most common way in
FDA counts 52 sprout-related outbreaks in the United States between 1996 and 2020, with more than 2,700 illnesses. On the source the agency writes without hedging: contaminated seed has historically been, and continues to be, the most common source of sprout contamination. An analysis of 14 outbreaks from 2012 through 2020 reached the same conclusion.
Sampling numbers back it up. In 2014 and 2015 FDA collected 825 samples across 37 states and found Salmonella in 2.35 percent of seed samples and in 0.21 percent of finished sprout samples. The gap between those two figures is statistically significant. The pathogen mostly arrives with the seed, and warm, humid germination simply lets it multiply: a 1999 review in a CDC journal described how cells from inoculated seed can climb past 10 million per gram of sprouts without changing how the product looks.
Where the seed picks it up is visible in the investigations too: rodents and birds inside the seed conditioning plant, chicken manure on the field where the seed grew, dirty transport, irrigation water fed by runoff from a neighboring field. In 2020 FDA sampled soil and water in growing regions and found Salmonella isolates genetically related to clinical isolates from the 2016 and 2018 sprout outbreaks, which means the same organism persists in that environment for years. How one seed lot turns into a multistate outbreak we covered separately.
What the American rule requires of a sprouter
Subpart M of 21 CFR 112 has been in effect since 2015 and is addressed to the people who grow sprouts. Its requirements are specific:
- grow, harvest, pack, and hold sprouts in a fully enclosed building;
- treat the seed with a scientifically valid method, or rely on the supplier’s treatment while holding documentation from that supplier about the method and about how the seed was handled and packaged afterward;
- visually examine the seed and its packaging for signs of contamination;
- test spent irrigation water from every production batch for E. coli O157:H7 and Salmonella, and keep the batch out of commerce until the result comes back negative;
- keep a written environmental monitoring plan for Listeria and act on it when a sample is positive.
Now read that list again. The mandatory testing starts after germination is already under way. Nowhere in the rule is there a requirement to test the seed itself before it goes in.
What the rule does require of a sprouter toward the supplier is notification. If the sprouter knows or has reason to believe a seed lot may be contaminated, it must stop using that lot, keep those sprouts out of commerce, and report the information back to the grower, distributor, or supplier it bought the seed from. What happens next is the supplier’s call. We covered a 2026 case where a seed supplier declined FDA’s recall request.
Why seed fell outside the rule
FDA explains this in plain language. The agency does not consider seed for sprouting to be covered produce under the Produce Safety Rule, so growing, conditioning, and distributing that seed falls outside the rule’s requirements.
The reason is stated in the same place: the end use of the seed is often unknown to the farmer who grows it and to the conditioners and distributors who handle it. A field of alfalfa looks the same whether the crop ends up as animal feed, as planting stock, or as sprouts on a sandwich. Because of that, FDA writes, it did not prescribe specific provisions for growing, harvesting, conditioning, or holding seed for sprouting.
In place of requirements, the agency issued guidance. It came out in May 2022, addressed to seed growers, conditioners, packers, holders, suppliers, and distributors, and the header on every page reads “Contains Nonbinding Recommendations.” The text spells out what that means: the word “should” in FDA guidance signals something suggested or recommended, not required.
The recommendations inside are sensible. Decide at planting time whether a crop is headed for sprouting. Store seed for sprouting apart from seed for other uses. Run shared equipment on sprouting seed first. Use only treated manure. Look at what the neighbors are doing with their animals and their runoff. Tell your sprout customers when you have reason to believe a lot is contaminated. FDA notes outright that producing seed for sprouting calls for a higher level of food safety precautions than producing seed for other purposes. None of it becomes mandatory. For the wider picture of what FSMA changed and what it did not, see our separate piece.
That sprouting seed is food was settled in 1982
There is a detail here that is easy to miss. Seed for sprouting fell outside one specific rule, and it remains food under the law. FDA’s position was written down in July 1982 and revised in February 1989: all seeds used for sprouting are foods. USDA’s classification of mung beans as seed rather than food, the document adds, has no bearing on how FDA regulates food.
The same document explains why planting stock and sprouting seed are different animals. A mercurial fungicide was once allowed on wheat meant only for seed, because a full maturation cycle stands between the treated kernel and anything anyone eats, and what comes out the other end is a separate food, the harvested wheat. Sprouting seed has no such gap. FDA says so directly: the seed is too closely associated with the finished food to allow that kind of exemption, and in many cases the spent hull is not even separated from the sprout before it goes on sale.
The practical takeaway is short. A packet of garden seed treated before sale does not become edible because you put it in a sprouting jar.
Why testing the seed is a weak instrument
In that same 2022 guidance, FDA says something about seed testing worth reading twice: laboratory analyses have frequently been unable to detect pathogens in seed lots that were conclusively identified as the vehicle in an outbreak. The levels are low and the distribution is uneven, so a sample can simply miss the contaminated grain.
Three conclusions follow, and the agency spells them out for firms that do test seed before selling it:
- testing is no substitute for good agricultural practices in the field and in the plant;
- a positive result tells you the lot is contaminated, while a negative result guarantees nothing about the absence of pathogens;
- a lot that tested positive must not be retested to clear it, because a later negative does not cancel the earlier positive.
Seed treatment is not a complete answer either: it reduces pathogens on the seed rather than eliminating them. That much was known long before the current rules, since the 1999 review reported that treating seed and sprouts with chlorinated water or other disinfectants fails to eliminate the pathogens.
This is the fork in the road. From the fact that seed testing is unreliable you can draw two different conclusions: skip the requirement and work on how the seed is grown, or require the test anyway, limits and all. The United States took the first road. The European Union took the second.
How Europe answered the same question
In May 2011 Europe lived through an outbreak of Shiga toxin-producing E. coli O104:H4 in which sprouts were identified as the most likely origin. Two years later, on March 11, 2013, the European Commission adopted four regulations at once: on traceability of sprouts and sprouting seed, on microbiological criteria, on approval of establishments that produce sprouts, and on certification for imported seed.
The logic behind mandatory seed testing sits in the preamble: pathogens in seed and sprouted seed are uncommon and unevenly distributed, sampling plans have statistical limits, and there is a lack of information about whether good agricultural practices were followed when the seed was grown. Therefore, the text concludes, all batches of seed need to be tested.
The procedure runs like this. The operator draws a representative sample from every batch of seed, at least 0.5 percent of the batch weight in 50 gram subsamples. That sample is germinated under the same conditions as the rest of the batch and tested for Salmonella and for Shiga toxin-producing E. coli of six serogroups, O104:H4 included. Sampling happens no earlier than 48 hours after germination starts, when the odds of detection are highest, and then at least once a month. An exemption from the preliminary test comes only from the competent authority, and only when two conditions hold together: a working food safety management system, and at least six consecutive months in which every batch met the criteria.
Ukraine put the same procedure into force on January 1, 2026, through a ministerial order registered with the justice ministry on October 31, 2025. The figures are identical: 0.5 percent of batch weight, 50 gram subsamples, germination of the sample under identical conditions, the 48 hour floor, monthly sampling, and an exemption only by decision of the competent authority after six clean months. It binds market operators rather than advising them. The rest of the requirements around leafy produce are gathered in our food safety section.
The rule that would close the gap waits until 2028
The United States does have a document written about exactly this gap. Sprouts sit on the Food Traceability List, and the listing is specific: all varieties of fresh sprouts, irrespective of seed source.
The rule requires whoever initially packs a lot of sprouts to keep the seed’s entire path on record: who grew it and when it was harvested, who conditioned it and under which seed lot code, who packed it, who supplied it, the master lot and sub-lot codes, a description of the seed including any antimicrobial treatment, and the date the sprouter received it.
One line in that list marks the limit of what is achievable. The location of the seed grower and the date of seed harvesting are to be recorded “if either is available.” Even the traceability rule concedes that the first link in the chain may be impossible to establish.
And the rule is not operating yet: Congress barred enforcement until July 20, 2028. Against that background, it gets easier to see why the same companies keep appearing in outbreak records decade after decade.
What this means for you
If you grow sprouts commercially in the European Union or in Ukraine, testing every seed batch is your legal obligation, and the lab has to be lined up before the batch goes in, not after a customer complains. If you grow them commercially in the United States, no one in the chain is required to test the seed, so the supplier you pick is doing that job by proxy.
If you sprout at home, you have no lab, no paperwork, and no standing to demand an analysis. Three levers remain.
- Seed choice. Buy seed sold for sprouting, with information about treatment on the label. Garden planting stock does not qualify: it is often treated before sale, which is the exact situation FDA wrote about back in 1982.
- A question for the seller. Under the American rule a sprouter may rely on the supplier’s treatment when it holds documentation of the method, such as a certificate of conformance. That is a fair question to ask any seller of sprouting seed, and the answer tells you a great deal about the company.
- Cooking. Heat is the one step on a home kitchen counter that reliably removes the risk. For young children, pregnant women, older adults, and anyone immunocompromised, raw sprouts stay in the higher risk column.
One more thing worth keeping in mind when you choose what to grow. The risk in sprouts comes from a warm, wet environment that multiplies whatever arrived with the seed, which is why sprouts and microgreens carry different risk profiles: microgreens are cut above the medium and the seed never reaches the plate. The 2026 alfalfa sprout outbreak, where one seed lot went out to 16 states, Puerto Rico, and several countries abroad, shows how the mechanism plays out in practice.
The gap in the chain stays open as long as responsibility for the seed rests with whoever receives it last. Europe’s answer, and now Ukraine’s, is to test the batch before it goes in, knowing full well that the test is imperfect. The American answer is to work on how the seed is grown and hope the farmer knows where the crop is headed. Once the crop is dried and milled, the question changes hands entirely, as the moringa powder outbreak shows.
