How to Test for Intestinal Parasites: What Works, What Doesn’t, and What to Ask For

Our review of parasite cleanse protocols ended where most articles on the subject end: with the advice to get appropriately tested instead. That advice is worthless without the next paragraph, because “get tested” is exactly what the cleanse industry also says — right before selling you the test. So this is the next paragraph.

The stool test you are offered is not one test

A comprehensive stool ova and parasite examination (usually written O&P) is three separate procedures performed on one specimen: a direct wet mount, which looks for motile trophozoites; a concentration step, typically formalin-ethyl acetate sedimentation, which brings eggs, larvae and cysts into view; and a permanent stained smear, usually trichrome, which is the part that actually identifies protozoa by morphology. The American Society for Microbiology’s consensus laboratory document notes plainly that these methods “are nonautomated and require extensive bench experience for accurate performance and interpretation.”

This matters more than it sounds, because the container your sample goes into determines which of those procedures can be run at all. CDC’s DPDx laboratory guidance is specific: if a two-vial kit is not available, the specimen “should be divided and stored in two different preservatives, 10% formalin and PVA,” at one volume of stool to three volumes of preservative. Formalin preserves helminth eggs, larvae and cysts well, is compatible with immunoassay kits, but gives “inadequate preservation of morphology of protozoan trophozoites” and “can interfere with PCR, especially after extended fixation time.” PVA does the opposite: excellent trophozoite morphology, but it “cannot be used with immunoassay kits” and is unsuitable for concentration. Refrigerated stool with no preservative at all, CDC says, is “suitable for antigen testing only.”

Several everyday substances also invalidate the test. CDC lists antacids, kaolin, mineral oil and other oily materials, non-absorbable antidiarrhoeal preparations, and gives specific clearance windows: barium or bismuth need 7 to 10 days, antimicrobial agents 2 to 3 weeks, and gallbladder dyes 3 weeks. Bismuth subsalicylate is the active ingredient in Pepto-Bismol. If you have been self-treating a two-week bout of diarrhoea, your sample may not be usable, and no one will necessarily tell you.

How many samples? The rule is contested, and that is worth knowing

CDC’s consumer page states that “CDC recommends that three or more stool samples, collected on separate days, be examined.” Its laboratory guidance is slightly different and more conditional: “Specimen collection may need to be repeated if the first examination is negative. If possible, three specimens passed at intervals of 2-3 days should be examined.”

The reason is intermittent shedding. Protozoan cysts and helminth eggs are not released at a constant rate, and stool consistency changes what is visible — trophozoites in liquid stool, cysts in formed stool. But the size of the effect has been measured, and it is smaller than the folklore suggests.

So a single comprehensive O&P recovers roughly 72 to 76 percent of what three would find; two specimens reach about 92 percent; the third adds around 8 percent. Those are the numbers to hold onto, and they are considerably better than the “stool tests miss half of all parasites” claim that circulates in cleanse marketing. The laboratory literature genuinely disagrees with CDC about whether three specimens are necessary in a low-prevalence population, which is a real scientific disagreement rather than a reason to distrust testing altogether.

Antigen tests: the best-performing option, and the worst, share a name

Antigen tests detect organism-specific proteins rather than looking for the organism itself. For the two most common intestinal protozoa in the US, they have largely displaced microscopy. CDC’s laboratory figures for direct fluorescent antibody (DFA) testing are 99% sensitivity and 100% specificity for Cryptosporidium, and 100% for both measures for Giardia. CDC’s domestic guidance for immigrant and refugee health is blunt about the comparison: “Stool ova and parasite testing has low sensitivity. Stool antigen is the preferred test for identifying Giardia infections.”

But “antigen test” covers two very different technologies, and the independent evaluations are much less flattering than the headline figures. Johnston and colleagues, testing 246 specimens against DFA as reference, found the ImmunoCard STAT rapid strip detected 81% of Giardia and only 68% of Cryptosporidium. Weitzel and colleagues, across 220 samples, found rapid strips ranging from 44% to 88% sensitivity, and concluded that these assays “were less sensitive than conventional microscopical methods” and “might be a useful addition to, but not a substitute for” microscopy. Specificity in both studies was excellent, at 98% or above.

In other words: laboratory immunofluorescence genuinely outperforms a microscope. A lateral-flow strip does not. If you are told you had “an antigen test,” which one is a fair question.

The other limitation is scope. Antigen tests are organism-specific by design. A negative Giardia and Cryptosporidium panel — which is what most US laboratories now run as their front-line parasitology test, per IDSA’s note that the two “are often offered and performed together as the primary parasitology examination” — says nothing whatsoever about Entamoeba, Cyclospora, Strongyloides, or any worm.

Multiplex PCR panels: more sensitive, harder to interpret

Syndromic molecular panels test one specimen for a fixed list of pathogens at once. In the multicentre evaluation of the BioFire FilmArray gastrointestinal panel across 1,556 prospective specimens, the parasite targets performed extremely well: 100% sensitivity for Cryptosporidium, Cyclospora cayetanensis and Giardia lamblia, with specificity above 99% for each.

The complications are real, and IDSA wrote two of them into formal recommendations. Recommendation 14: “Clinical consideration should be included in the interpretation of results of multiple-pathogen nucleic acid amplification tests because these assays detect DNA and not necessarily viable organisms.” A positive result can mean an active infection, a resolved infection still shedding nucleic acid, or asymptomatic carriage.

Co-detection is the second problem. In the BioFire evaluation, multiple pathogens were found in 31.5% of positive specimens, up to six in one sample, and the authors acknowledged that “the clinical implications of specific pathogen combinations are not well documented or understood.” A review in the same journal was more direct: “insufficient data are available to guide laboratorians and clinicians on how to interpret these findings.”

Two further points rarely make it into consumer coverage. First, these panels are worse than microscopy for worms: in a 2024 Belgian travel-clinic comparison, multiplex PCR detected 59.1% of helminth infections against 100% for the conventional workflow. Second, switching methods inflates apparent prevalence dramatically. A 2025 analysis of 3,495 stools found Blastocystis in 19.25% by qPCR versus 6.55% by microscopy, and Dientamoeba fragilis in 8.86% versus 0.63% — roughly three-fold and fourteen-fold increases with no change in the underlying population. CDC has said the same about its own surveillance data, attributing part of a 47.2% decade-long rise in cryptosporidiosis incidence to “increasing use of diagnostic multiplex PCR panels.”

Reagent cost runs roughly 80 to 155 US dollars per specimen, and the published recommendation is to reserve highly multiplexed panels for immunocompromised patients, the critically ill, or people with prolonged diarrhoea, rather than using them as a first-line screen.

Which test for which parasite

Pinworm: the test almost everyone gets wrong

Pinworm is the one intestinal worm that is genuinely common in high-income countries, and it is the one where a stool test is close to useless. The gravid female migrates out through the anus at night to lay eggs on the perianal skin, so the eggs largely bypass the faecal stream entirely. CDC’s Yellow Book puts it in one sentence: “Examining stool samples is not recommended because pinworm eggs are sparse.”

Published estimates for how often pinworm eggs turn up on a stool exam cluster around 5 percent, with Public Health Ontario giving “under 10%.” CDC publishes no figure of its own.

The correct test costs nothing. Press the sticky side of clear tape against the skin beside the anus first thing in the morning, before washing, using the toilet, or getting dressed, and repeat on three consecutive mornings. Public Health Ontario reports sensitivity below 60% for a single specimen, rising to about 90% with three collected daily, and close to 100% with six, and advises collecting for four to six consecutive days to rule infection out. Adult worms are sometimes visible around the anus two to three hours after the person falls asleep.

Strongyloides: the one where a normal stool test is actively misleading

Strongyloides stercoralis deserves its own section because the stakes are asymmetric. Infection is usually silent for years, and CDC notes that “traditional stool examinations are insensitive and can require up to seven exams to reach a sensitivity of 100%.” A single stool exam misses up to 70% of cases. The best primary data on cumulative yield, from 864 specimens, found 32.4% detection with one specimen, 49.1% with two, 59.3% with three and 66.7% with four.

Worse, the standard concentration step can destroy the evidence: CDC warns that “formalin-ethyl acetate concentration may remove larvae and reduce sensitivity.” A routine O&P sent to a general laboratory can be counterproductive for this organism specifically. The sensitive stool methods — Baermann funnel sedimentation, Koga agar plate culture — are not offered by most US general laboratories.

The right test is a blood antibody test. CDC’s own antigen EIA has 96% sensitivity and 98% specificity, though cross-reactions with filariasis, schistosomiasis and ascariasis occur: in one systematic review, 77% of people with known hookworm infection tested positive for Strongyloides. A negative serology does not fully exclude infection, as a 2026 report of two transplant recipients infected by a donor with a negative ELISA demonstrated.

The two organisms your stool panel probably found, and probably should not have

This is the section that matters most, because it explains most positive results on consumer stool tests.

Blastocystis and Dientamoeba fragilis are the two organisms most likely to appear on a direct-to-consumer stool report, and the evidence that either causes disease is weak enough that several national bodies have stopped looking for them. IDSA’s guideline carries the position in a footnote: “The pathogenicity of Blastocystis hominis and Dientamoeba fragilis remains controversial. In the absence of other pathogens, they may be clinically relevant if symptoms persist.” Neither organism has an IDSA testing or treatment recommendation anywhere in the document. CDC’s clinical page for D. fragilis is plainer still: infections “are often asymptomatic and require no treatment.”

Three lines of evidence explain that caution.

Carriage in healthy people is extraordinarily common. A survey of 2,154 metagenomes found Blastocystis in 14.9% overall and 22.4% in Europe; a later analysis of 56,989 metagenomes across 32 countries found prevalence ranging from 2.46% in Japan to 56.29% in Fiji. In more than 22,000 routine Danish stool samples, D. fragilis was positive in 43%. In a study following 142 Danish day-care children, 68.3% were positive at the first sampling, every one of the 108 children followed was eventually positive, and there was “no statistical association between a recent history of gastrointestinal symptoms and testing positive.”

The two largest population-based case-control studies found both organisms more common in healthy people. A Danish study of 124 people with IBS against 204 controls found that “a greater proportion of controls than cases carried the parasites (50% vs 36%).” A Dutch multicentre study of 1,374 gastroenteritis cases against 1,026 controls found Blastocystis in 25.8% of cases versus 40.0% of controls, and D. fragilis in 25.8% versus 37.6%, concluding that the presence of these organisms “may be considered characteristic of a healthy intestinal microbiome.” A 2024 analysis of nearly 57,000 people went further, associating Blastocystis with more favourable cardiometabolic profiles and inversely with obesity, and finding that improvements in diet quality were followed by increases in Blastocystis prevalence.

And the treatment trials are null. Two double-blind, placebo-controlled randomised trials have tested metronidazole in people carrying these organisms. In 96 children with chronic gastrointestinal symptoms and D. fragilis, symptom improvement was essentially identical between drug and placebo. In 50 adults with Blastocystis, resolution was 47.2% on metronidazole versus 51.4% on placebo. The dissociation is the finding: the drugs clear the organism, and the symptoms do not change.

In fairness, the case is not closed. A 2017 meta-analysis of 17 studies found a pooled association between Blastocystis and IBS with an odds ratio of 2.19, though the same analysis found no association at all for D. fragilis. That evidence is dominated by hospital-based case-control studies in high-prevalence settings, which are vulnerable to reverse causation, and it is contradicted by every large population-based dataset. A 2026 paper also argues that at least one widely used PCR assay for D. fragilis produces false positives in human samples, which would undermine some of the high-carriage findings — but that argument cuts both ways, since the same assays generate the positive results being used to sell treatment.

One US-specific number is worth keeping in mind before applying European prevalence figures to yourself: across 4,804 gastrointestinal parasite PCR tests run in Utah between 2014 and 2024, D. fragilis positivity was 0.6%.

Blood tests, and why “high eosinophils” is not a parasite result

CDC states the position without qualification: “Blood tests look for a specific parasite infection; there is no blood test that will look for all parasitic infections.” Serology is genuinely useful for Strongyloides, for invasive amebiasis, and for tissue-dwelling helminths such as schistosomes — but IDSA recommends against serology for establishing the cause of infectious diarrhoea generally.

Eosinophil count deserves specific attention because it is the number most often presented as evidence of a parasite. Peripheral eosinophilia is defined as 500 or more eosinophils per microlitre. IDSA’s guideline notes that “an increased eosinophil count may occur with parasitic infections that involve a tissue phase” — and that clause is the whole problem, because the intestinal protozoa people actually worry about do not have one. CDC’s Yellow Book explicitly lists Entamoeba histolytica, Cryptosporidium, Cyclospora cayetanensis and Giardia as parasitic infections without peripheral eosinophilia.

The best quantitative test of eosinophilia as a screening tool examined 2,004 US-bound refugees in Thailand against a qPCR reference standard, in a population where roughly 73% were infected with at least one parasite. Sensitivity ranged from 51% to 73%, and specificity from 48% to 65% — barely better than chance on the specificity side. The authors’ conclusion was that “the predictive value of eosinophilia is poor for the most common parasitic infections, and it should not be used alone for screening.” In a low-prevalence US population, positive predictive value falls further still. Atopy, asthma, drug reactions and several non-infectious conditions are far more common explanations for a mildly raised eosinophil count in the United States than worms are.

Who should actually be tested

The clearest threshold in the guidelines is for travellers. IDSA Recommendation 13: “Diagnostic testing is not recommended in most cases of uncomplicated traveler’s diarrhea unless treatment is indicated. Travelers with diarrhea lasting 14 days or longer should be evaluated for intestinal parasitic infections.”

Beyond that, testing is indicated for people with moderate or severe immune deficiency and diarrhoea, where IDSA calls for a broad differential including Cryptosporidium, Cyclospora, Cystoisospora and microsporidia; and for specific exposures, which the guideline tabulates. Swimming in or drinking untreated fresh water, and inadequately chlorinated pools, point to Cryptosporidium and Giardia. Child-care attendance or employment, and healthcare, long-term care or prison settings, point to the same two. Travel to resource-limited countries widens the list to Entamoeba histolytica, Cyclospora and Cystoisospora. Farm and petting-zoo visits point to Cryptosporidium. Anal-genital, oral-anal or digital-anal contact points to E. histolytica, Giardia and Cryptosporidium alongside bacterial and sexually transmitted infections.

What no authority recommends is testing asymptomatic people in the general population. IDSA advises against follow-up testing after diarrhoea has resolved in most cases, and against empiric treatment of asymptomatic contacts. Choosing Wisely Australia’s infectious-diseases recommendation is a single line: “Do not investigate or treat for faecal pathogens in the absence of diarrhoea or other gastro-intestinal symptoms.” The American College of Gastroenterology recommends against routine stool testing for enteric pathogens in irritable bowel syndrome, while endorsing targeted Giardia testing in people at high risk — a sensible exception, given that the relative risk of developing IBS after a documented Giardia infection is around 3.4.

Tests to avoid, and one sentence that explains why

Direct-to-consumer “comprehensive stool analysis” panels — the category that includes GI-MAP, GI Effects and similar products — are laboratory-developed tests. They run under CLIA certification rather than FDA clearance, and the vendors’ own reports say so: one states that its assays “have not been cleared by the U.S. Food and Drug Administration.” Following a federal court decision in March 2025 and a subsequent FDA rule reverting the regulation, FDA does not currently exercise premarket review over laboratory-developed tests. FDA-cleared stool parasite testing does exist, for contrast — the BioFire gastrointestinal panel holds 510(k) clearance K230404.

The sentence that matters most comes from CMS, which administers CLIA: “CMS’ CLIA program does not address the clinical validity of any test,” and the routine survey “does not include a review of the clinical validation of a LDT — that is, the accuracy with which the test identifies, measures, or predicts the presence or absence of a clinical condition.” So “CLIA-certified laboratory” is a real statement about laboratory process, and not a statement that the result means what the report says it means.

Two specific features of these reports are worth naming. One panel reports seven organisms in its parasitology section, four of which are textbook non-pathogenic commensals, and prints numeric “reference ranges” for Blastocystis and D. fragilis — implying a validated pathogenic threshold that does not exist in the peer-reviewed literature. The same panel is not available for ordering in New York State, which independently reviews laboratory-developed tests. We could find no peer-reviewed validation study of any of these panels for parasite detection, which is an absence of evidence rather than a demonstration of inaccuracy, and worth stating as exactly that.

The closest thing to independent evidence on this category is a 2026 study from the National Institute of Standards and Technology, which sent identical stool material in triplicate to seven direct-to-consumer services. It found “major discrepancies, both within and across the different service providers,” with between-provider variability on the same scale as biological variability between different donors — three companies detecting C. difficile in material where four reported it absent, and one company grading replicates of a single sample as both healthy and unhealthy. Two honest caveats: that study tested microbiome profiling rather than parasite panels, and it is a single study.

The methods with no diagnostic role at all are easier to summarise. Live blood analysis, iridology, applied kinesiology or muscle testing, and electrodermal, bioresonance or “zapper” devices have no validated ability to detect parasites; iridology in particular has been tested in blinded controlled studies and a systematic review, all negative.

As for symptom quizzes, the honest position is narrower than a debunking. The two published studies of symptom-based screening, both from high-prevalence settings, found that no single symptom performed adequately — for example, abdominal pain at 54% sensitivity, and anal itching at 29.8%. But the symptoms those studies examined were abdominal pain, diarrhoea, loss of appetite, nausea and perianal itching. For the items that dominate consumer quizzes — sugar cravings, fatigue, bloating, teeth grinding — we found no published diagnostic-accuracy assessment at all. They have never been evaluated as diagnostic indicators, which is different from having been refuted, and it means nobody selling a quiz based on them can know that it works.

How common are intestinal parasites in the US, honestly?

There has never been a nationally representative US stool-parasitology survey, so any confident percentage should be treated with suspicion — including the widely repeated claim that a fifth of Americans harbour parasites, which does not trace to a US population study. The verifiable picture is that giardiasis and cryptosporidiosis are nationally notifiable and tracked in the thousands to low tens of thousands of cases a year; that the CDC no longer publishes the “40 million” pinworm figure that circulates online, and has removed the page it came from; that a 2017 finding of 34.5% hookworm prevalence in rural Alabama did not replicate, with a later 704-stool study in the same county finding none; and that Toxocara seroprevalence has been revised down from 14% to about 5%, and is not an intestinal infection in humans anyway.

None of this means parasitic infection does not happen in the United States. It means the base rate for an average person with bloating is low, which is precisely why the pre-test question — what specific exposure or symptom pattern makes this plausible? — matters more than the test itself.

Frequently asked questions