FDA and CDC updated the multistate Cyclospora outbreak today, August 27, 2026: 11,458 laboratory-confirmed illnesses, at least 495 hospitalizations and two deaths across twenty states, with illness onsets running from June 14 through August 15, all of it tied to processed iceberg lettuce from Taylor Farms de Mexico recalled on July 17. Georgia, Tennessee and Texas were added this week. And here is the figure that answers the question I keep being asked. Of the 9,256 people in this outbreak for whom CDC has an illness onset date, 8,959 of them, 96.8 percent, got sick before the recall.
Cyclosporiasis is the intestinal illness caused by Cyclospora cayetanensis, a microscopic parasite that reaches produce when human waste reaches the field or the irrigation water, and that washing does not reliably remove. Two people are dead in Michigan. I want that sentence in everything I write about this outbreak.
The exposure ended in July, and that is worth saying plainly.
CDC published the epidemic curve behind today’s update, and it is one of the cleanest curves I have seen. Illnesses climb through late June. Then the single week of July 6 through July 12 carries 5,287 people, 57 percent of everyone in the outbreak with an onset date, in seven days. The peak is July 10, at 1,001 people in one day. After that it falls off a cliff. Fifteen people in the entire outbreak have an August onset date, and most of the daily counts in August are zero or one.
Illnesses in the multistate Cyclospora outbreak by the date symptoms began. Source: CDC outbreak epidemic curve data table, downloaded August 27, 2026, built on the 9,256 people of 11,458 for whom an onset date is available. Rust bars are onsets before the July 17 recall; blue bars are on or after it. The bars sum to 9,256.
The shape does the arguing. A curve that rises for three weeks, spikes for one, and then falls to single digits is a curve describing a food that was in commerce for a defined window and then was not.
When FDA says it remains confident that all the recalled lettuce is off the market, and when CDC writes in the past tense that this outbreak was a major driver of the rise in cases this year, the curve backs them up. The lettuce stopped making people sick around the middle of July. I am not going to pretend otherwise, and anyone who tells you the exposure is ongoing has not looked at the chart.
What is still growing is not the outbreak. It is the counting.
Today’s update added 528 illnesses, 41 hospitalizations and three states. Almost none of those people got sick this week. CDC says on its own pages that it takes about six weeks from the first symptom to the day a case reaches Atlanta, and for this parasite the lag is worse than usual: Cyclospora is not on a routine stool panel, so a clinician has to think to order the test; patients wait days or weeks before seeking care; laboratories in the hardest-hit states ran short of kits; and the interview that assigns a case to an outbreak happens after the laboratory result, not before it.
Every number that moved today is a number about July. That is the same lag that produced the two-peaked curve Phyllis Entis flagged last month, and it is why a falling curve and a rising count are not a contradiction.
CDC has now published exact state numbers for this outbreak for the first time.
All summer the federal outbreak map has shown states in colored bands rather than figures. The data table behind today’s map is exact and it is per state. This is the first state-level federal attribution ever published in this outbreak, and it is the answer to a question I have been asking since the middle of August. Here is the whole of it.
Where the 11,458 illnesses are. Bands match those CDC uses on its own outbreak map. Grey states have no illness attributed to this outbreak, which is not the same as having no cyclosporiasis. Source: CDC outbreak map data table, August 27, 2026.
| State | Illnesses attributed to the outbreak |
| Michigan | 4,846 |
| Ohio | 2,679 |
| Missouri | 1,010 |
| Indiana | 823 |
| Kansas | 331 |
| Kentucky | 329 |
| Illinois | 278 |
| Pennsylvania | 244 |
| Oklahoma | 241 |
| Nebraska | 199 |
| West Virginia | 160 |
| Iowa | 130 |
| Arkansas | 82 |
| New Hampshire | 42 |
| Texas | 20 |
| Massachusetts | 18 |
| Maine | 16 |
| Tennessee | 5 |
| Georgia | 3 |
| North Carolina | 2 |
| Total | 11,458 |
Illnesses attributed to the multistate Cyclospora outbreak linked to Taylor Farms de Mexico iceberg lettuce, by state. Source: CDC outbreak map data table, downloaded August 27, 2026. Twenty rows, summing to 11,458, which matches the total CDC and FDA published the same day.
Michigan and Ohio together carry 7,525 of the 11,458. Two states account for roughly two-thirds of the entire federal outbreak.
The three states added this week account for twenty-eight illnesses between them.
Texas is in for 20 people, Tennessee for 5, Georgia for 3. All three had been on FDA’s confirmed distribution list since the recall notice went out, and all three are on the Walmart retail list as well. The lettuce did not arrive in Georgia this week. It arrived in June. What arrived this week were the interviews.
Where the lettuce went, from FDA’s Stores Affected section. Foodservice distribution ran June 29 through July 16, 2026. Hatching marks the fifteen states where Marketside-brand product also reached select Walmart stores. FDA says product may have travelled beyond the confirmed list. This is where the product went, not where illness occurred.
Set the two maps beside each other and the mismatch is the whole problem. Thirty-one states got the lettuce. Twenty have illnesses assigned to it. Eleven states received the recalled product and appear nowhere in the outbreak count, and Louisiana, Mississippi, Alabama, South Carolina, Wisconsin, New York, New Jersey, Connecticut, Maryland, Florida and Virginia are the eleven. I am not saying every one of them has hidden outbreak cases. I am saying nobody has published a reason to think they do not.
Michigan’s own count is roughly three times what the federal outbreak assigns it.
CDC puts 4,846 Michigan illnesses in this outbreak. MDHHS reported 14,277 cases as of August 20, its most recent figure at the time I am writing. Ohio is 2,679 against a state count above 5,400. Texas is 20 against a state count of 315. North Carolina appears in the federal outbreak with two people, in a state that was reporting more than three hundred of its own back in July.
The honest caveat is that these are different series. CDC counts only laboratory-confirmed illnesses it has affirmatively tied to the recalled lettuce through an interview. The states count every case reported during the investigation period. Neither is wrong. But the gap is the whole argument, and it is the one I made when CDC told a television station that a thousand sick Missourians were not part of this. A twenty-state outbreak is not a description of who got sick. It is a description of who got interviewed.
The illness CDC has not tied to the lettuce grew faster this week than the outbreak did.
CDC stood up a new Cyclospora case data page on August 25. With data through August 24, it counts 17,180 laboratory-confirmed cases of cyclosporiasis acquired in this country since May 1, 922 hospitalizations, two deaths and cases in forty-eight states plus the District of Columbia, against 1,180 for the whole of last season. It is aware of at least 11,844 more that are not laboratory confirmed or need further investigation, and it separately counts 1,805 travel-associated cases.
Now do the subtraction. A week ago the same series stood at 15,716 confirmed, with 10,930 in the outbreak. Today it is 17,180 confirmed, with 11,458 in the outbreak. The confirmed national count grew by 1,464. The outbreak grew by 528. The illness this country has confirmed and has not attributed to the recalled lettuce grew by 936 in seven days, and now stands at 5,722 people. Hospitalizations tell the same story: 495 of the national 922 are in this outbreak, which leaves 427 that are not.
There is something in those two columns worth an epidemiologist’s afternoon. Inside the outbreak, about four in every hundred people were hospitalized. Outside it, in the confirmed cases with no attributed source, it is about seven in a hundred. The comparison mixes pull dates three days apart and hospitalization is captured differently in different states, so I would not call that a finding. I would call it a question, and it is one nobody has asked out loud.
One person in five in this outbreak has no illness onset date on file.
The curve is built on 9,256 people. The outbreak has 11,458. That leaves 2,202 people counted in the total and invisible on the chart. That is a real improvement over where this stood two weeks ago, when the curve rested on 5,934 of 9,481 cases, barely three in five, and the improvement deserves to be said out loud. Somebody has been doing the unglamorous work of chasing down onset dates.
One small thing for anyone reading the chart closely. The data as published runs through August 14, while both agencies give August 15 as the last onset in the outbreak and CDC’s national page shows domestic onsets as late as August 19. The chart trails the text by a few days. That is normal and it is not a scandal. It is worth knowing before you quote a date off the graph.
FDA is still in the fields in Mexico, and the season closes on Monday.
FDA’s update says its onsite inspections and sampling are running at iceberg lettuce growers and the processing facility in Mexico, in coordination with Mexican officials. Growers is the word that matters. On July 17 this was an investigation of a plant. It is now also an investigation of fields and the water that runs through them, which is where a waterborne parasite actually gets onto lettuce. The cyclosporiasis season CDC recognizes runs May 1 through August 31. That is Monday.
Three questions I am getting this week.
Is the Cyclospora outbreak over?
The exposure appears to have ended in the middle of July. Illness onsets collapse after July 12 and there have been only fifteen in August. The investigation is still open, the counting will run into the fall, and CDC and FDA both still describe the outbreak as ongoing.
Is the recalled lettuce still out there?
No. The best-by dates on the recalled Taylor Farms de Mexico product ran through August 3 and have long passed. FDA says it is confident the product is off the market. Taco Bell stopped using lettuce from that supplier on July 17.
I was sick in July and never got tested. Does it still matter?
Yes. Call your doctor and ask specifically about Cyclospora. A routine stool test does not look for it, symptoms can relapse for a month or more, and it is treatable. A case that gets confirmed now still counts, and the count is the only thing that makes anybody act.
Under control is a claim about a food. It is not a claim about a system.
The lettuce is off the market and the curve has come down. Both of those are real and I will say so. Neither of them is the same as knowing how the parasite got onto that lettuce, and neither is the same as being able to do this faster the next time. The lot-level traceability records that would have let investigators walk a head of lettuce back to a field in days rather than weeks are not required until July 2028, because Congress and this industry agreed to push them there.
Two people are dead. At least 495 are out of the hospital and paying for it. More than five thousand confirmed illnesses this summer have no attributed source at all, and that pile grew again this week. The season ends Monday. The counting does not.
Recent Comments