Iceberg lettuce from central Mexico has been linked to one major cluster in an extraordinary summer surge of cyclosporiasis. The parasite is difficult to detect, difficult to wash away and slow to trace. At the same time, parts of the federal surveillance system designed to spot foodborne disease have been scaled back.
Most foodborne outbreaks begin with a familiar warning.
Do not eat this product.
Check your refrigerator.
Throw away anything with this lot number.
The 2026 Cyclospora outbreak is more complicated.
As of Aug. 4, the Centers for Disease Control and Prevention had received reports of 10,468 laboratory-confirmed cases of domestically acquired cyclosporiasis since May 1. At least 517 people had been hospitalized and two had died.
And that is not necessarily the full count.
CDC says it is aware of more than 12,255 additional cases that have not yet been laboratory-confirmed or still require investigation, including determining whether patients were exposed while traveling overseas. Cases have been reported across 47 states.
That puts the number of confirmed and still-under-review illnesses above 22,000.
One major portion of the outbreak has been linked through epidemiological investigation and food traceback to iceberg lettuce sourced from central Mexico and distributed by Taylor Farms de Mexico.
But that outbreak is itself only part of the national increase.
And that distinction matters.
The story is not simply that one batch of lettuce made thousands of people sick.
It is that the United States is confronting an unusually large surge of a parasite that is intrinsically difficult to track at the same time that parts of the country’s foodborne-disease surveillance infrastructure have become less comprehensive.
The question is not whether federal cuts created Cyclospora.
They did not.
The question is whether we made an already difficult pathogen harder to find.
First: What exactly is Cyclospora?
Cyclospora cayetanensis is a microscopic parasite that infects the human intestine and causes an illness called cyclosporiasis. Humans are believed to be its only natural host.
The most common symptom is frequent watery diarrhea.
Patients may also experience loss of appetite, weight loss, abdominal cramping, bloating, gas, nausea, fatigue and sometimes fever, headache, body aches or vomiting. Symptoms commonly begin about a week after exposure, although the incubation period can range from roughly two days to two weeks or longer.
Untreated illness can last for weeks.
It can also appear to improve and then return.
For healthy people, cyclosporiasis is usually not life-threatening. But prolonged diarrhea can cause serious dehydration, and the illness can be more severe in older adults, young children and people with weakened immune systems.
The first-line treatment is trimethoprim-sulfamethoxazole, commonly known as TMP-SMX or Bactrim. The original research brief correctly identifies it as the standard treatment, though some of the more exotic complications described in the brief are uncommon enough that they should not be presented as routine consequences of infection.
One strange feature makes Cyclospora particularly difficult
A person infected with Cyclospora sheds the parasite in feces.
But unlike some other gastrointestinal pathogens, newly shed Cyclospora is not immediately infectious.
It must spend time in the environment developing before it can infect another person.
That means direct person-to-person transmission is considered unlikely.
Instead, outbreaks generally occur when food or water becomes contaminated with human fecal material and is later consumed.
Fresh produce is an especially troublesome vehicle.
Raspberries, basil, cilantro and bagged salad products have all been implicated in previous outbreaks.
And Cyclospora presents another problem for consumers:
Washing produce does not reliably remove it.
The parasite’s oocyst can persist on fruits and vegetables, making prevention much more dependent on sanitation where produce is grown, harvested and processed than on what a consumer does at the kitchen sink.
That is one reason contamination early in the agricultural supply chain can become difficult to stop once food has entered national distribution.
Then came the lettuce
By mid-July, investigators had identified a significant outbreak among people who reported eating at Taco Bell restaurants.
FDA traceback work converged on shredded iceberg lettuce supplied by Taylor Farms de Mexico and grown in central Mexico.
Taylor Farms announced a voluntary recall on July 17 and removed iceberg lettuce from central Mexico from the U.S. market. Recalled products included certain Marketside iceberg salad and shredded-lettuce packages sold at Walmart as well as products distributed to restaurants and other food-service customers.
The affected food-service products had been distributed widely—including into Texas.
By July 24, FDA said 1,947 people in nine states who were infected with Cyclospora had reported exposure to Taco Bell. At least 98 had been hospitalized.
But there is an important distinction:
Those 1,947 cases were not the entire national outbreak.
CDC is investigating multiple clusters.
The iceberg-lettuce outbreak has contributed significantly to the national increase, but federal officials have not attributed every one of the more than 10,000 confirmed domestic cases to Taylor Farms, Taco Bell or even iceberg lettuce.
That is an important correction to some of the more sweeping accounts circulating around the outbreak.
And no, FDA did not find Cyclospora on the lettuce
This episode produced another unusual complication.
On July 18, FDA initially reported a positive Cyclospora result from a lettuce sample.
The following day, agency laboratory experts reversed that finding.
After reviewing the test, FDA concluded that the apparent result was a false positive and said that, as of July 19, it had no confirmed positive Cyclospora product samples.
That did not exonerate the lettuce.
FDA specifically emphasized that the recall remained supported by epidemiological evidence and traceback data showing convergence on Taylor Farms iceberg lettuce from central Mexico.
This is a useful example of how outbreak investigations actually work.
A contaminated product does not always yield the pathogen when investigators finally test it.
The implicated batch may already be gone.
Contamination may be uneven.
And Cyclospora is technically difficult to detect.
Public-health investigators therefore combine laboratory results with interviews, supply-chain records, geographic clustering and genetic information from patients.
It is less satisfying than finding a smoking gun in a bag of lettuce.
But epidemiology rarely provides television-style evidence.
Modern outbreak control increasingly depends on determining whether patients hundreds of miles apart were infected by the same source.
For bacteria such as Salmonella or Listeria, whole-genome sequencing has revolutionized that process.
Cyclospora is more difficult.
CDC says whole-genome sequencing is not currently available for the parasite in the same way it is for many bacterial pathogens. Instead, laboratories use genetic testing targeting particular markers in patient stool samples to help group potentially related cases.
You can go much deeper into this technology, targeted amplicon sequencing and the multilocus tools scientists have developed to distinguish clusters. That portion of the research is broadly supported by the scientific literature and illustrates just how technically sophisticated modern foodborne surveillance has become.
But sophisticated laboratory technology only matters if someone is collecting samples, processing them and connecting the results to patient interviews quickly enough to matter.
That brings us to the uncomfortable part.
The surveillance system changed before this outbreak
For nearly three decades, the Foodborne Diseases Active Surveillance Network—FoodNet—has operated as a collaboration among CDC, FDA, USDA and health departments in 10 states.
Its surveillance area covers roughly 16 percent of the U.S. population.
Historically, FoodNet actively tracked eight major foodborne pathogens, including Cyclospora.
Then the system changed.
Beginning July 1, 2025, FoodNet reporting became optional for every pathogen except Salmonella and Shiga toxin-producing E. coli.
Cyclospora was among the pathogens moved into that optional category.
That does not mean America stopped monitoring Cyclospora.
This distinction is critical.
Cyclosporiasis remains a nationally notifiable disease. Cases continue to reach CDC through the National Notifiable Diseases Surveillance System and other outbreak-reporting channels.
So it would be inaccurate to say federal officials simply turned off Cyclospora surveillance.
What changed was one particular layer of active surveillance.
And in public health, losing a layer can matter.
Then there were personnel cuts
There is stronger evidence of another reduction.
Joel Barratt, a molecular parasitologist who previously led CDC’s Cyclospora laboratory, said that the team responsible for helping respond to Cyclospora outbreaks was reduced from 11 people to three amid federal staffing cuts.
Barratt argued that a team that small would necessarily have less capacity to respond quickly to large outbreaks.
That is consequential.
But it should not be stretched into something the evidence cannot establish.
There is presently no basis for saying that reducing the laboratory staff caused the contaminated lettuce.
Obviously it did not.
There also is not enough public evidence to say precisely how many infections would have been prevented if FoodNet had retained its previous structure or the CDC laboratory had retained all of its staff.
Outbreaks are messy.
There is an incubation period.
People may not seek medical care.
Doctors may not initially order the correct test.
Laboratory confirmation takes time.
Food histories must be reconstructed.
Supply chains must be traced backward.
CDC itself estimates that in this outbreak there can be roughly a six-week lag between illness onset and a case being reported federally.
Any serious analysis has to acknowledge all of that.
But the inverse is also true.
A complicated system is not an argument for weakening the system.
It is an argument for asking what capacity was lost.
The $11.4 billion number requires context
Another claim circulating around the outbreak deserves clarification.
In March 2025, the Trump administration announced the termination of approximately $11.4 billion in pandemic-era public-health funding that had been going to state and local governments and other recipients.
Those funds had originally been allocated for the COVID-19 response, although health departments had subsequently used parts of the money for laboratory capacity, epidemiology, staffing and other public-health functions. The administration argued the emergency had ended and continued COVID-era spending was unnecessary.
That is different from saying DOGE specifically removed $11.4 billion from Cyclospora surveillance.
It did not.
And linking the entire amount directly to this outbreak would be misleading.
The more defensible question is whether broad reductions in public-health staffing and laboratory capacity left state and federal agencies with less flexibility when an unrelated outbreak arrived.
That is a question worth investigating.
It is not yet a proven causal conclusion.
Michigan shows how extraordinary this summer became
Michigan provides the clearest illustration of the scale.
The state typically records roughly 40 to 50 cases of cyclosporiasis in an entire year.
On June 30, Michigan had identified more than 170 cases associated with the developing outbreak.
Four days later, the count had reached 572.
By Aug. 6, Michigan officials reported 12,485 cases, 279 hospitalizations and two deaths associated with cyclosporiasis. Both people who died had significant underlying medical conditions.
The state has since relaxed its broad warning about lettuce as reports declined and recalled products moved out of circulation.
That does not mean the national investigation is finished.
CDC warns that confirmed counts will continue rising as delayed reports are processed.
Is this really bigger than every previous Cyclospora outbreak combined?
Sort of—but this is where statistical language matters.
Compare the 2026 numbers with CDC historical outbreak data, and you conclude that this summer surpassed more than three decades of previous outbreaks combined.
That comparison is attention-grabbing but needs qualification.
Historical outbreak databases count illnesses epidemiologically associated with recognized outbreaks.
The current CDC surveillance figure of 10,468 counts laboratory-confirmed domestically acquired cases, including multiple clusters and cases that may not all belong to the same outbreak.
Those are not identical denominators.
So I would not say:
One 2026 outbreak produced more cases than every Cyclospora infection in the previous 33 years combined.
That is not what the data show.
What we can safely say is:
The 2026 surge is extraordinary by modern U.S. standards and the largest cyclosporiasis event officials have confronted in recent memory.
That is plenty dramatic without abusing the statistics.
This is also a globalization story
Cyclospora did not suddenly appear in America in 2026.
U.S. outbreaks associated with fresh produce stretch back decades.
Raspberries from Guatemala were repeatedly implicated in the 1990s.
Other outbreaks have involved basil, cilantro, snow peas and mixed salads.
In the 2010s, FDA investigations repeatedly identified contamination concerns involving cilantro from Puebla, Mexico.
The broad pattern matters.
Americans increasingly expect strawberries in winter, herbs year-round and fresh salads assembled through enormously complex agricultural networks.
That system provides remarkable abundance.
It also means contamination in one agricultural region can reach restaurants and grocery stores thousands of miles away before anyone realizes people are becoming sick.
That is not an argument against imports.
It is an argument for recognizing that international food supply chains require international sanitation standards, traceability and surveillance capable of moving at the same speed as the food.
Consumers cannot inspect their way out of this
Cyclospora creates an uncomfortable lesson for people accustomed to being told to wash their vegetables.
You should wash fresh produce.
That remains sensible food-safety practice.
But washing does not reliably eliminate Cyclospora.
Consumers therefore occupy the last—and weakest—point in the prevention chain.
The better defenses exist upstream:
Clean agricultural water.
Adequate toilets and sanitation for farmworkers.
Protection against fecal contamination in fields and packing facilities.
Rapid traceback.
Reliable product records.
Laboratory testing.
Disease surveillance.
And enough epidemiologists to notice when 50 cases become 500.
The FDA’s Produce Safety Rule includes agricultural-water requirements intended to reduce microbial contamination, and the agency has developed specialized methods for detecting Cyclospora in agricultural water.
That is where prevention has to begin.
So did government cuts make people sick?
That is the politically irresistible question.
It is also the question we should resist answering before the evidence exists.
The contaminated food caused infections.
Failures somewhere in agricultural sanitation allowed human fecal contamination to enter the food chain.
Those are the immediate mechanisms.
We do not yet know whether federal staffing or surveillance changes allowed substantially more people to become sick before the source was identified.
But we know enough to ask.
FoodNet reduced mandatory reporting requirements for Cyclospora.
CDC’s specialized laboratory reportedly lost most of its staff.
States were simultaneously dealing with the loss of billions in pandemic-era public-health funding.
And this summer the country encountered an unusually large, technically difficult outbreak requiring precisely the epidemiologists, laboratories and information systems those programs support.
Those facts do not prove causation.
They justify oversight.
Congress should ask how long samples waited for processing.
How many cases were delayed between state and federal reporting systems.
Whether staffing shortages slowed genomic clustering.
Whether traceback capacity was sufficient.
Whether optional FoodNet reporting produced meaningful blind spots.
Whether CDC requested additional personnel.
And whether the July 17 recall would have occurred sooner under the previous staffing and surveillance structure.
Those answers would tell us something.
Partisan slogans will not.
The outbreak is declining. The lesson should not.
The recalled lettuce is aging out of the food supply.
Michigan has begun relaxing its broad recommendations.
New cases eventually will decline.
The headlines will move on.
That is exactly when the institutional questions become important.
Public-health surveillance is an unusual government service because success often looks like nothing happening.
A cluster is noticed.
A food is traced.
A recall happens.
An outbreak stops.
There is no ribbon cutting.
No visible construction project.
No obvious constituency celebrating an epidemiologist who discovered that 37 people in three states had eaten the same lettuce.
That makes surveillance easy to describe as bureaucracy when budgets are being cut.
Until the system is needed.
Cyclospora itself does not care which administration funded the laboratory.
It does not care whether a surveillance program was created by Democrats or reduced by Republicans.
It does not care whether lettuce crosses a state line or an international border.
It exploits the distance between contamination and detection.
And the 2026 outbreak has demonstrated how large that distance can become.
More than 10,000 Americans have already had laboratory-confirmed infections this season.
Thousands more cases remain under review.
Five hundred people have been hospitalized.
Two have died.
The most important question is therefore not simply what was on the lettuce.
It is what we learned about the system watching it.
We cannot prevent every contaminated crop from entering a supply chain. But if we choose to make outbreaks harder to see, we should not be surprised when we discover them only after thousands of people are already sick.