The Cyclospora outbreak is over. The season it was hiding in is not.

On September 11, 2026, the CDC declared the Cyclospora outbreak linked to Taylor Farms de México iceberg lettuce over. The final count is 12,883 illnesses across 21 states, at least 570 hospitalizations and two deaths, both in Michigan. It is one of the largest foodborne outbreaks ever recorded in the United States. And it accounts for barely two-thirds of this year's cases. Over the same months, 19,883 people were confirmed with cyclosporiasis acquired inside the country — which leaves 7,000 illnesses with no named source at all.

What “over” actually means here

An outbreak is declared over when new cases tied to it stop appearing and the implicated food is gone from the supply. Both conditions were met, and the CDC is unusually specific about the shape of it: before the July 17 recall, the outbreak was producing more than 1,000 infections a day at its peak. By August it was averaging fewer than two a day.

That collapse is the strongest evidence anyone has that the traceback was right. A recall that removes the true vehicle stops an outbreak almost vertically; a recall that removes the wrong one does not. This one stopped.

“Over” does not mean the last case has been counted. Cyclosporiasis takes about a week to make someone ill, days more before they see a doctor, and weeks before a confirmed result reaches a national database. Cases from earlier in the summer are still being reported. The CDC moved to periodic rather than weekly updates on September 15 for exactly that reason: the flow has slowed to a trickle of backdated reports.

The 7,000 cases the outbreak does not explain

Subtract the outbreak from the season and 7,000 confirmed infections remain, along with up to 494 hospitalizations. They are not a mystery in the epidemiological sense — cyclosporiasis is seasonal in the United States every summer, and most cases never get linked to a specific food. But the scale is new. A normal season is on the order of a thousand cases nationally. This one produced nearly twenty thousand, and the single largest outbreak in it explains only 65 percent.

Geography makes the point sharply. The outbreak touched 21 states. The season touched 49 states and Washington, D.C. — effectively the whole country. Whatever else was circulating was not confined to one supplier's distribution map.

No public record identifies a source for those cases. That is an honest gap, not an evasion: linking a diffuse parasite to a specific food requires interviews, traceback and a cluster tight enough to analyse, and most of these cases have none of the three.

Why Cyclospora is hard to find in the first place

There is a detail in the CDC's July 14 health advisory that reframes every number above. Cyclospora is not reliably detected by a standard ova-and-parasite stool exam. Finding it requires a specific request: a modified acid-fast stain, or PCR. A clinician who orders the routine panel can get a clean result from a patient who is infected.

That is what the photograph at the top of this article shows — an oocyst that only becomes visible once someone has decided to look for it with the right stain.

The consequence runs in one direction only. Every figure here is a floor, not a ceiling. The 19,883 are the people who were tested, tested correctly, and confirmed. The advisory also asked clinicians to consider cyclosporiasis in anyone with prolonged watery diarrhea between May and August, and to take a food and travel history — a request that would not be necessary if the diagnosis were routinely being made.

It is worth remembering how the season looked from inside it. On July 14, the CDC counted 1,645 confirmed domestic cases with more than 5,100 awaiting confirmation, in 34 states, with no deaths. Nine weeks later the confirmed total was twelve times higher, the map covered nearly every state, and two people had died. The backlog was not noise. It was the outbreak, arriving late in the data.

What this changes for a shopper

Very little, and that is the uncomfortable part. Cyclospora is not destroyed by washing produce, and there is no consumer-level test. Cooking kills it, but the foods it travels on — lettuce, herbs, berries — are eaten raw by definition.

What does help is knowing the symptom that distinguishes it: watery diarrhea that does not resolve, often lasting weeks, sometimes relapsing, frequently with fatigue and loss of appetite. It is treatable — trimethoprim-sulfamethoxazole for seven to ten days — but only once diagnosed, and diagnosis requires the specific test. A patient who mentions the word, and who mentions salad, changes what gets ordered.