Someone asked me today what is going on with Cyclospora. Here is where things are as of a few moments ago.
CDC updated its Cyclospora case-data page on September 15, 2026, and said it was the last weekly one. It now reports 19,883 laboratory-confirmed cases of cyclosporiasis acquired in this country since May 1, 1,064 hospitalizations, two deaths and cases in forty-nine states plus the District of Columbia, against 1,180 for the whole of last season. The federal outbreak linked to Taylor Farms de Mexico iceberg lettuce closed on September 11 at 12,883 illnesses, 570 hospitalizations, two deaths and twenty-one states, with a last illness onset of August 17. The fifty state health departments now report 33,556 cases this season.
Cyclosporiasis is an intestinal illness caused by Cyclospora cayetanensis, a microscopic parasite spread through human feces in contaminated food or water. Washing does not reliably remove it, and chlorine does not kill it.
The unattributed pile is now exactly 7,000, and it shrank for a reason that is not an answer.
19,883 confirmed nationally, 12,883 in the outbreak, 7,000 left over. A week ago that number was 8,137. It fell by 1,137, and not one of those 1,137 people was tied to a food anyone has named. CDC’s final outbreak update added 1,425 people to the Taylor Farms count and a twenty-first state, Virginia, on the day it closed the file. The national series grew by 288 in the same stretch. The pile got smaller because CDC finished assigning interviews it already had, not because an investigation found anything.
Roughly two in three confirmed illnesses now carry a federal attribution of any kind, up from three in five a week ago. Measured against everything CDC says it is aware of this season, 36,242 cases, it is about one in three. Measured against the 33,556 the states themselves report, it is closer to two in five.
FDA’s inspections in Mexico are over, and the samples have not been read.
This is the part of the FDA advisory worth reading twice. On August 20, FDA said it had moved from the Taylor Farms de Mexico processing plant to the iceberg lettuce growers in the fields, which is where a waterborne parasite gets onto lettuce. The September 11 update says those onsite inspections and that sample collection, conducted with Mexican officials, have ended, and that the samples are pending analysis. The fieldwork is finished. The laboratory work is not, and nobody outside the agency knows how many samples were taken, from which fields, or when a result will be published.
FDA says its investigation remains open, and it has opened a second page for what comes after: a meeting with senior Mexican officials the week of September 8 on a prevention framework, a letter to industry asking for collaboration, and a promised update to its Cyclospora action plan. Every one of those is worth doing. Not one of them is a finding.
Whatever those samples show, a negative is the expected result.
Taylor Farms is already arguing the answer. Its Cyclospora information hub, last modified September 1, says epidemiologic evidence can point investigators toward a potential source but cannot confirm that a specific product caused illness, that confirmation requires a positive laboratory test of the product, and that the company’s own testing — roughly two thousand samples taken in central Mexico since May — has returned no positives. FDA answered that argument six years ago, in writing, to a different salad company. In the warning letter of October 20, 2020 that followed the Fresh Express bagged salad outbreak, FDA declared the product adulterated on the strength of epidemiologic and traceback evidence, with no positive product sample anywhere in the file. It then told the company exactly what its own negative results were worth: this parasite is not present in high numbers or spread evenly through a lot, and it is not uncommon to see many negatives even when contamination has occurred.
The science under that sentence is not in dispute. When Canadian federal scientists verified FDA’s own detection method, leafy greens spiked with ten oocysts tested positive thirty percent of the time. Seven in ten known-contaminated samples read clean. FDA’s 2013 assessment of this same plant put the infectious dose at ten to a hundred oocysts. A serving capable of putting someone in a hospital sits below the level at which a single sample reliably reports anything. The company’s own trade association tells consumers the same thing, that the absence of a positive sample does not mean the epidemiologic and traceback findings were wrong.
The company did not apply that standard to itself either. It recalled in July 2026 with no positive test in existence, and it suspended shipments from the same plant in 2013 with no positive test in existence. A sample count without a noun attached — two thousand samples of what, taken where, by what method, from which lots — is not a finding. I went through all of it here. The samples FDA carried out of Mexico this summer should be read the same way when they come back: a negative tells you what one 25-gram grab did not catch, and nothing about what 12,883 people ate.
FDA has open files on 312 of those 7,000 people.
FDA’s active investigation table, updated September 16, carries four Cyclospora investigations besides the lettuce outbreak. Together they account for 312 people.
| Ref # | Posted | Cases | Outbreak status | Product linked | FDA steps taken |
| 1392 | Jul 22 | 234 | Ongoing | Not yet identified | Traceback, sampling |
| 1385 | Jul 8 | 38 | Ongoing | Not yet identified | Traceback, sampling |
| 1398 | Aug 5 | 22 | Ongoing | Not yet identified | Traceback, sampling |
| 1384 | Jul 8 | 18 | Ended | Not yet identified | Traceback only |
| All four | 312 | None |
FDA active investigation table, as posted September 16, 2026. Case counts are FDA’s. Reference 1390, the iceberg lettuce outbreak, is listed separately as ended with the investigation still open.
Against 7,000 confirmed illnesses with no attributed source, 312 is about one in twenty-two. Reference 1392 is 234 people, posted July 22, with traceback and sampling initiated and the inspection column still blank eight weeks on. Reference 1384 has ended, and it ended without a food ever being named. All four read Not Yet Identified.
Seven counters have gone quiet in three weeks.
CDC says it will stop updating weekly because recent infections have significantly declined, and will post periodically instead. Michigan moved to the first Thursday of each month and stands at 14,718 cases and 366 hospitalizations until October 1. Indiana called September 2 its final update until the official close of the outbreak, at 2,104. Virginia ended its weekly table on September 1 at 242. North Carolina moved to every other week on September 8 and does not report again until September 29. West Virginia closed its outbreak investigation outright on September 11, at 315 cases and 21 hospitalizations. Ohio has published no statewide number since 5,452 on August 13, thirty-five days ago.
CDC’s own page says illnesses from this season will keep arriving as they are identified, and that because of the unprecedented number of cases this year, illnesses from earlier in the season may keep coming in as well. The counting continues. The publishing has stopped.
What moved this week, and what did not.
North Carolina added eight cases, to 963 for the year and 32 hospitalizations, and its dashboard still says the increase there does not appear to be associated with the outbreak in other states involving Taco Bell, and that its patients have reported eating parsley and cilantro more often than a population-wide baseline would predict. Wake County holds 527 of them. Virginia’s last column changed for a different reason: the state says it is now in the multistate outbreak after verification of traceback data from CDC and FDA. Nothing else in the table moved, because almost nobody published.
Underneath the confirmed 19,883, CDC has 9,765 laboratory-confirmed cases still being sorted into domestic or travel-acquired and at least 6,594 more that are not laboratory confirmed or are awaiting an interview, which is how it reaches 36,242. Travel-associated cases, counted separately, are 2,078 with 74 hospitalizations. The median onset is still July 8, the latest still August 30.
Here is where all fifty states stand.
The Lettuce column is whether FDA has confirmed the recalled product reached that state. The last column is whether anybody has tied that state’s illnesses to Taylor Farms at all. Thirty-one states received the lettuce. Twenty-one are in the federal outbreak. Every count is a floor.
| State | 2026 cases | As of | CDC Aug 27 | All of 2025 | Lettuce | Taylor Farms |
| Michigan | 14,718 | Sep 3 | 4,846 | 50 | Yes | Linked |
| Ohio | 5,452 | Aug 13 | 2,679 | 76 | Yes | Linked |
| Missouri | 2,152 | Aug 23 | 1,010 | none reported | Yes | Linked |
| Indiana | 2,104 | Sep 2 | 823 | 27 | Yes | Linked |
| Illinois | 1,437 | Aug 18 | 278 | 301 | Yes | Linked |
| North Carolina | 963 | Sep 15 | 2 | 300 | Yes | Linked |
| New York | 939 | Aug 17 | 0 | 694 | Yes | Unknown |
| Kansas | 829 | Aug 26 | 331 | 29 | Yes | Linked |
| Kentucky | 819 | Aug 12 | 329 | 45 | Yes | Linked |
| Oklahoma | 556 | Aug 25 | 241 | 41 | Yes | Linked |
| Florida | 450 | Aug 15 | 0 | 210 | Yes | Unknown |
| Iowa | 327 | Aug 20 | 130 | 64 | Yes | Linked |
| Texas | 315 | Aug 11 | 20 | 512 | Yes | Linked |
| West Virginia | 315 | Sep 11 | 160 | 8 | Yes | Linked |
| Arkansas | 287 | Aug 25 | 82 | 13 | Yes | Linked |
| Wisconsin | 284 | Aug 26 | 0 | 65 | Yes | Unknown |
| Virginia | 242 | Aug 31 | 0 | 115 | Yes | Linked |
| Nebraska | 218 | Aug 6 | 199 | 51 | Yes | Linked |
| Pennsylvania | 214 | Aug 11 | 244 | not notifiable | Yes | Linked |
| Colorado | 150 | Jul 17 | 0 | 205 | No | Unknown |
| Massachusetts | 147 | Aug 5 | 18 | 74 | Yes | Linked |
| New Hampshire | 94 | Aug 25 | 42 | 7 | Yes | Linked |
| Alabama | 85 | Aug 3 | 0 | 19 | Yes | Unknown |
| Washington | 76 | Aug 28 | 0 | not available | May have | Travel link |
| Maryland | 69 | Jul 17 | 0 | 123 | Yes | No link found |
| New Jersey | 46 | Jul 11 | 0 | 180 | Yes | Unknown |
| Tennessee | 44 | Jul 16 | 5 | 38 | Yes | Linked |
| California | 41 | Jul 14 | 0 | not available | May have | Unknown |
| Minnesota | 41 | Jul 17 | 0 | none reported | No | Unknown |
| Connecticut | 35 | Jul 17 | 0 | 41 | Yes | Unknown |
| Maine | 27 | Aug 7 | 16 | 4 | Yes | Linked |
| Oregon | 23 | Jul 24 | 0 | not available | No | Travel link |
| Arizona | 19 | Jul 17 | 0 | 49 | No | Unknown |
| Georgia | 11 | Jul 17 | 3 | 116 | Yes | Linked |
| Mississippi | 7 | Jul 22 | 0 | not notifiable | Yes | Unknown |
| Hawaii | 6 | Aug 26 | 0 | not available | No | No link found |
| Alaska | 5 | Jul 16 | 0 | 6 | No | Unknown |
| Rhode Island | 4 | Jul 17 | 0 | 6 | May have | No link found |
| South Dakota | 3 | Jul 16 | 0 | 3 | May have | Unknown |
| Louisiana | 1 | Jul 17 | 0 | 99 | Yes | Unknown |
| Utah | 1 | Jul 17 | 0 | 37 | No | No link found |
| Delaware | none published | — | 0 | 1 | May have | Unknown |
| Idaho | none published | — | 0 | not notifiable | No | Unknown |
| Montana | none published | — | 0 | 8 | No | Unknown |
| Nevada | none published | — | 0 | not notifiable | No | Unknown |
| New Mexico | none published | — | 0 | 9 | No | No link found |
| North Dakota | none published | — | 0 | 4 | May have | Unknown |
| South Carolina | none published | — | 0 | 45 | Yes | Unknown |
| Vermont | none published | — | 0 | 3 | May have | Unknown |
| Wyoming | none published | — | 0 | 0 | No | Unknown |
| All fifty states | 33,556 | 11,458 | 31 | 21 |
Counts as published by each state health department, compiled from the linked sources, as of September 17, 2026. Forty-one counted rows sum to 33,556; nine states publish no number. The CDC column is the last per-state outbreak split CDC published, on August 27, and sums to 11,458; CDC’s final September 11 update raised the outbreak to 12,883 without publishing a new state-by-state breakdown. State totals and CDC totals count different things and are not directly comparable.
What the season leaves behind.
Two people are dead. 1,064 have been hospitalized. The states willing to print a number printed 33,556 of them. The largest Cyclospora outbreak in American history is closed at 12,883, and 7,000 confirmed illnesses this summer still have no attributed source, with an open federal file on 312 of them. FDA’s inspectors have left the fields in Mexico and the samples they carried out have not been read. CDC has stopped publishing weekly. The lot-level traceability records that would let anyone walk this lettuce back to a field in days rather than weeks are not required until July 2028. Every count above is a floor.