Microplastics and your health: what the evidence shows, graded

This is the companion to our plain-English piece, Microplastics are inside you right now. That one tells you what’s been found. This one does the less fun job: going claim by claim and asking how much each one deserves to be believed.

Because here is the uncomfortable truth about microplastics and health. Activists (hello) sometimes talk as if harm is proven. Industry talks as if the absence of proof means safety. Both are getting ahead of the data, in opposite directions. The honest answer is in between, and it varies a lot depending on which question you ask.

How we grade

We use four plain-English grades. They’re our editorial judgement, not a formal scoring system, and we show our reasoning for every one.

• Established. Several independent teams, using different methods, keep finding the same thing in people. We’d be surprised if it turned out wrong.
• Strong signal, not proven. Good, repeated evidence points one way, but it’s mainly from animals, or the human studies are few or have real method problems.
• Early or mixed. One or two studies, small samples, conflicting results, or findings that haven’t been repeated.
• Unknown. Nobody has studied it properly in people yet, or the question can’t be answered with today’s tools.

What moves a grade up: human data over animal data, animal over cells in a dish, replication by independent labs, realistic doses, and methods that guard against contamination. What moves it down: one lab, one study, doses thousands of times higher than real life, and measuring techniques that can be fooled.

The grades at a glance

Question	Grade	Why, in one line
Do microplastics get into the human body?	Established	Found in stool, lungs, blood, placentas and more, by many labs using different methods.
Do they reach blood, placenta, testes, semen, plaque and brain?	Strong signal, not proven	Repeatedly reported, but each organ rests on a handful of small studies and methods doubts are real.
How much is in us, and is it rising?	Early or mixed	The main mass-measuring method can mistake fat for polyethylene. Treat specific numbers as provisional.
Do microplastics harm the gut and other organs in lab animals?	Strong signal, not proven	Consistent in rodents, but mostly one plastic (polystyrene), sometimes at doses above likely human exposure. Relevance to people unclear.
Does plastic in artery plaque raise heart attack and stroke risk?	Early or mixed	One striking observational study. Not yet repeated. Association, not cause.
Do microplastics damage the brain or cause dementia?	Unknown	Higher levels reported in brains of people with dementia, but the disease itself could explain it.
Do they harm human fertility?	Early or mixed	Animal evidence is suggestive; human studies are tiny and inconsistent.
Do they cause gut disease or cancer in people?	Early or mixed (bowel disease); Unknown (cancer)	One study links stool plastic to bowel disease severity. Cancer links come from animal markers only.
Are bisphenols and phthalates (the chemicals in plastic) hormone disruptors?	Established	Decades of animal, human and regulatory evidence. The EU has banned BPA in food contact materials.
Do those chemicals harm people at everyday exposures?	Strong signal, not proven	Robust human associations for several outcomes; causation is hard to prove in people.
What dose of particles is safe?	Unknown	No health-based limit exists, because the data to set one don’t.

Question 1: Are they in us?

Into the body at all: Established

Microplastics enter through food, water and air. They’ve been counted in human stool, in human lung tissue taken during surgery (39 particles in 11 of 13 samples), and in many other places. The lung study matters because it used infrared spectroscopy to identify individual particles you can see and photograph, which is harder to explain away as a measurement artefact than a chemical signal alone. Different labs, different countries, different techniques, same answer. This is as close to settled as this field gets.

Blood: Strong signal, not proven

The landmark study is from Amsterdam in 2022. Researchers measured plastic particles of 700 nanometres and larger in the blood of 22 healthy volunteers, with quantifiable levels in 17 of them and a mean of 1.6 micrograms per millilitre. They found PET (bottle plastic), polyethylene and styrene polymers.

Why not “Established”? Because of the method. That study, like most studies that report the weight of plastic in tissue, used pyrolysis gas chromatography-mass spectrometry (Py-GC-MS): you burn the sample and look for the chemical fragments plastics leave behind. The trouble is that some of those fragments also come from fats and other natural molecules. In 2025 an Australian team tested the method on human blood and concluded that Py-GC-MS is currently not suitable for measuring polyethylene and PVC in biological samples because of interference. Their realistic detection limits were up to 20 times higher than the ones calculated in clean water, and in their own pilot of Australian blood, no polymer other than (interference-prone) polyethylene came out above detection limits.

That doesn’t mean plastic isn’t in blood. It means the blood numbers, especially for polyethylene, need confirming with better-validated methods.

Placenta: Strong signal, not proven

An Italian team first reported 12 microplastic fragments in 4 of 6 placentas in 2021, using Raman spectroscopy that identifies individual particles. In 2024 a New Mexico team using Py-GC-MS reported microplastics in all 62 placentas they tested, at 6.5 to 685 micrograms per gram, with polyethylene making up 54% of the total.

Two methods agree that plastic reaches the placenta. They disagree enormously about how much: a dozen fragments in a few placentas versus hundreds of micrograms per gram in every one. Some of that gap may be about particle size, since Py-GC-MS also picks up particles too small to count under a microscope. Some of it may be the polyethylene interference problem above. Nobody yet knows the true split.

Testes and semen: Strong signal, not proven

Plastic has been reported in human testes and semen by several groups:

• A Chinese study found microplastics in both testis tissue (6 samples) and semen (30 samples), using two different methods.
• An Italian study found 16 fragments in 6 of 10 semen samples from men in a polluted area of Campania.
• A 2025 Chinese study found them in 34 of 45 semen samples from men at a fertility clinic.
• The New Mexico team found plastic in all 23 human and 47 dog testes it examined.

Several countries, several methods, consistent presence. Small samples throughout, so the grade stays one notch below Established.

Artery plaque: Strong signal, not proven

The Italian study we’ll come to in Question 3 detected polyethylene in the carotid plaque of 150 of 257 patients (58.4%), and PVC in 31. Crucially, it didn’t rely on Py-GC-MS alone: electron microscopy showed jagged foreign particles inside plaque immune cells, some containing chlorine (a marker for PVC). A 2025 Chinese study of paired blood and plaque samples from 20 patients also found plastic in plaque, at higher concentrations than in blood, though again by Py-GC-MS.

Brain: Strong signal, not proven (presence) and Early or mixed (amounts)

Two very different studies say plastic reaches the brain.

A Brazilian team examined the olfactory bulb, the part of the brain sitting just above the nose, in 15 people who had died. They found 16 particles and fibres in 8 of the 15, using infrared spectroscopy, with clean procedural blanks. Small numbers of particles, carefully counted. It suggests a plausible route in: through the nose.

The much louder study came from the University of New Mexico in Nature Medicine in 2025. Using Py-GC-MS backed by infrared and electron microscopy, it reported that brain tissue held far more plastic than liver or kidney (7 to 30 times more), mostly polyethylene, largely as nanoscale shards. Concentrations were higher in people who died in 2024 than in 2016, and higher still in 12 people with dementia.

This is where the methods debate gets loud:

• The authors themselves flagged that lipids can interfere with the polyethylene signal. The brain is about the most fat-rich organ you have.
• In 2025 the journal published a formal critique by a group of analytical chemists arguing the study had limited contamination controls and lacked validation steps, which could affect the reported concentrations. The original authors published a reply.
• The paper also carries an author correction from March 2025 for duplicated supplementary images, mis-sized scale bars and a missing sample-prep step. The correction says these don’t affect the spectral data. It’s housekeeping, not a retraction, but it’s worth knowing.
• The Australian blood study above found Py-GC-MS unsuitable for polyethylene in biological samples. Polyethylene is most of what the brain study measured.

And it’s not just these papers. A 2026 review of 218 studies measuring microplastics in human tissue found that few consistently ran blank controls (to catch contamination from air, gloves and lab kit) and spike-recovery tests (to check the method finds plastic that’s known to be there). That’s a field-wide problem, not one lab’s.

Our read: plastic in the brain, yes, probably. “Brains are full of plastic, and it’s getting worse every year”: not proven. The specific amounts and the time trend should be treated as provisional until other labs repeat them with tighter controls.

How much, and is it rising? Early or mixed

Everything above points the same way. Presence: increasingly solid. Quantities: shaky. Trends over time: based on very few studies. If you see a headline with a precise number of grams or “spoons” of plastic in an organ, that number is the least reliable part of the story.

Question 2: Do they cause harm?

This is where we have to separate three kinds of evidence, because they say different things.

Cells in a dish

Many lab studies expose human or animal cells to micro- and nanoplastics and report oxidative stress, inflammation and cell death. They’re useful for working out how harm could happen. They’re poor at telling you whether it happens in a living person, because cells in a dish have no gut lining, liver, immune system or clearance, and the doses are often very high. On their own, cell studies never get past Early or mixed for human harm.

Animals: Strong signal, not proven

Animal studies are more convincing, and they’re fairly consistent. The most thorough review of them so far, a 2024 systematic review led by UCSF’s Program on Reproductive Health and the Environment, pooled 28 animal studies and three human studies on digestive, reproductive and respiratory health. It used a formal hazard-rating method with four possible conclusions: “known”, “presumed”, “suspected” or “not classifiable” as a hazard to humans.

Its conclusion: microplastics are “suspected” to harm human reproductive, digestive and respiratory health, with a suggested link to colon and lung cancer. Note the word. “Suspected” is the lowest of the three tiers that indicate harm. The evidence on sperm quality and gut immune effects in animals was rated high quality, but the reviewers downgraded from “presumed” because of the small number of studies and sample sizes. For birth outcomes in humans, the evidence was “not classifiable”.

The review is also candid about its limits, which is a good sign:

• 22 of the 28 animal studies (79%) used polystyrene, mostly as uniform beads. Real-world microplastics are a messy mix dominated by polyethylene, polypropylene and PET, with irregular shapes and chemical additives.
• For the smallest particles, the doses could be higher than estimated human exposure.
• It was a “rapid” review, with one person screening studies and another verifying, rather than two working independently.
• It was funded by the University of California and the JPB Foundation, with NIEHS centre support. One co-author is affiliated with the environmental group NRDC. No industry funding.

A good example of the more realistic animal work: a 2025 study fed mice food chopped on plastic or wooden cutting boards for up to 12 weeks. Mice eating food cut on polypropylene boards had higher blood markers of inflammation (including CRP and TNF-α), more gut inflammation, and lower expression of the proteins that keep the gut lining sealed, compared with mice whose food was cut on wood. Food cut on polyethylene boards shifted gut bacteria but didn’t cause obvious inflammation. That’s a real-world source of plastic, at a dose that came from actually chopping food, which is better than beads in a tube. It’s still mice.

Why “Strong signal, not proven” and not higher: the animal signal is consistent and biologically plausible. But mice aren’t people, the plastics used rarely match what we’re actually exposed to, and the dose question is unresolved.

People: mostly Unknown or Early

There are very few studies that measure microplastics in people and then look at their health. That’s the single biggest gap in this field. The handful that exist are the subject of Question 3.

Question 3: Which organs and outcomes?

Heart and blood vessels: Early or mixed

This is the strongest human evidence anywhere in the microplastics literature, so it’s worth getting exactly right.

In a 2024 study in the New England Journal of Medicine, Italian researchers analysed artery plaque removed from the necks of 304 patients having surgery to clear a narrowed carotid artery. 257 were followed for an average of 33.7 months. Polyethylene was found in the plaque of 150 of them (58.4%), and PVC in 31 (12.1%). People with detectable plastic in their plaque were 4.53 times as likely to have a heart attack, a stroke or die of any cause during follow-up (95% confidence interval 2.00 to 10.27).

What makes it strong:

• It’s in humans, it’s prospective (people were followed forward in time), and the effect is large.
• It didn’t rely only on Py-GC-MS. Electron microscopy found jagged particles inside plaque immune cells.

What keeps it at Early or mixed:

• It’s observational. People with more plastic in their arteries may differ in other ways (where they live, what they eat, what they work with, how sick they already were) that also raise heart risk. An observational study can’t fully rule that out.
• It’s one study, in one specific group: people already sick enough to need artery surgery. The result hasn’t been repeated in another cohort.
• Polyethylene is the polymer most prone to measurement interference in Py-GC-MS. The microscopy helps, but doesn’t fully settle how much plastic was there.
• Plaque samples go through an operating theatre and a lab, both full of plastic. Contamination controls matter enormously.

The right reaction to this paper is not “plastic causes heart attacks.” It’s “this is the kind of signal that should trigger large, well-funded follow-up studies.” Two and a half years on, we could not find a study that has repeated the health follow-up in another group of patients.

Brain and dementia: Unknown

The New Mexico brain study found more plastic in the brains of 12 people with dementia. The authors themselves said “no causality is assumed”, and for good reason. Dementia damages the blood-brain barrier and the brain’s waste-clearance systems, so the disease could let more plastic in, or stop it getting out. Add the methods dispute, and the honest grade is Unknown. Not “no effect”. Unknown.

Reproduction and fertility: Early or mixed

In animals, the UCSF review rated the evidence that microplastics harm sperm quality as high quality, and the harm as “suspected”. In humans, it’s much thinner:

• In the New Mexico testis study, sperm counts were measured only in the dogs, not the men. In dogs, some polymers (polycarbonate and ABS) correlated with lower counts.
• In the 2025 study of 45 Chinese men, those with PET in their semen had lower progressive sperm motility (20.6% vs 34.9%), but that difference fell just short of statistical significance (p = 0.056), and there was no link with sperm concentration or total count.

Plausible, worth watching, not shown in people.

Gut and bowel disease: Early or mixed (IBD) and Unknown (cancer)

A Chinese study found that people with inflammatory bowel disease had more microplastics in their stool than healthy people (41.8 vs 28.0 items per gram of dry stool), and that higher levels went with more severe disease. The authors were upfront that this could run either way: plastic might worsen the disease, or the disease might make the gut hold on to more plastic. The paper drew a published comment and a reply from the authors. One study, a direction of cause that can’t be determined, and the result has not been confirmed by others.

The “suggested link to colon cancer” in the UCSF review comes from animal studies showing biological changes that are characteristic of carcinogens, such as chronic inflammation and changes in cell proliferation. That is a reason to study cancer risk, not evidence that microplastics cause cancer in people. Grade: Unknown.

Lungs: Early or mixed

Plastic is clearly in lung tissue. The UCSF review rated animal evidence for lung injury, inflammation and reduced lung function as moderate quality, harm “suspected”. For the general public, the human evidence is thin.

Pregnancy and babies: Unknown

Plastic reaches the placenta. The UCSF review found the human evidence on birth weight and gestational age “not classifiable” because of its low quality. That’s a gap, not a reassurance.

The chemicals are a separate problem, and a better-studied one

Here’s the most important distinction in this whole article, and the one most coverage blurs.

A microplastic is two things at once: a physical particle, and a carrier of chemicals added to the plastic to make it soft, stiff, clear, coloured or flame-resistant. The particle question is new and unresolved. The chemical question is decades old, and much further along.

Phthalates: Established as hazards; Strong signal, not proven at everyday doses

Phthalates are plasticisers, the chemicals that make PVC soft. A 2022 overview of systematic reviews of human studies found robust evidence linking phthalate exposure to lower semen quality, effects on children’s neurodevelopment and childhood asthma, and moderate to robust evidence for changes in genital development in baby boys. It found moderate evidence for links with low birth weight, endometriosis, lower testosterone, ADHD, type 2 diabetes, and breast and uterine cancer. These are associations in people, backed by animal studies showing how phthalates interfere with hormones. That combination is why several phthalates are already restricted in toys and childcare articles in the US and EU (see our guide to plastic toys and teethers).

Bisphenols: Established as hazards; Strong signal, not proven at everyday doses

In 2023 the European Food Safety Authority re-evaluated bisphenol A and cut its tolerable daily intake from 4 micrograms to 0.2 nanograms per kilogram of body weight, a 20,000-fold reduction. The critical effect was on immune cells in mice, supported by human observational studies. EFSA concluded that average and high dietary exposures in every age group exceeded the new limit by two to three orders of magnitude, and that there was a health concern.

The EU acted. Regulation (EU) 2024/3190 bans BPA in food contact materials, including can linings and polycarbonate. It entered into force in January 2025, with most transition periods ending in July 2026, some can coatings and professional food equipment allowed until 2028, and repeat-use items already on the market allowed to stay on sale until January 2029. It also covers other hazardous bisphenols, noting that BPS, a common BPA replacement, is itself classified as toxic to reproduction.

Why this matters for the particle debate

Two reasons. First, when someone says “there’s no proof microplastics harm people,” that’s narrowly true of particles and badly misleading about plastic. The chemicals in plastic are hazards regulators already act on. Second, it’s genuinely unclear how much of any harm from microplastics comes from the particle itself and how much from the chemicals it carries. That’s one of the field’s open questions, and the answer may differ by plastic type.

What the big reviews say

WHO, 2019 and 2022

In 2019, launching its report on microplastics in drinking water, WHO’s Maria Neira said: “Based on the limited information we have, microplastics in drinking water don’t appear to pose a health risk at current levels. But we need to find out more.” The same release noted that particles larger than 150 micrometres are unlikely to be absorbed by the body. That was seven years ago, before nearly all the tissue studies in this article existed. It’s often quoted as if it were current. It isn’t.

WHO’s broader 2022 report on dietary and inhalation exposure reviewed data up to December 2021 and its main output was a list of research needs to address “current uncertainties”. In other words: not enough data to assess the risk. That’s still the most accurate summary for particles.

The UCSF systematic review, 2024

Covered above: microplastics “suspected” to harm digestive, reproductive and respiratory health, mainly on animal evidence. It’s the most rigorous attempt so far to grade the hazard, and its authors call for governments to act to reduce exposure now rather than wait.

The Lancet Countdown on health and plastics, 2025

Launched in The Lancet in 2025, the Countdown is a new monitoring project that will track indicators of plastics’ health harms over time. Its launch paper says plastics cause disease and death from infancy to old age and are responsible for health-related economic losses exceeding US$1.5 trillion a year, and that production has grown from 2 million tonnes in 1950 to 475 million tonnes in 2022. Read that figure carefully: the paper’s body explains it is an estimate of the 2015 health costs of just three plastic chemicals (the flame retardant PBDE, BPA and the phthalate DEHP) in 38 countries, assuming the reported links are causal. It says nothing about microplastic particles in your body. It’s a strong case that plastic chemicals harm health. It isn’t a measurement of what particles do.

What’s still unknown

• The dose. How much plastic is in a typical person, measured reliably. Until that’s known, no one can say whether animal doses are realistic.
• Whether particles get out. Whether, and how fast, the body clears micro- and nanoplastics from blood and organs.
• Which plastics matter. Most animal work uses polystyrene; most human tissue contains polyethylene. That mismatch is a problem.
• Particle versus chemical. How much harm comes from the physical particle, and how much from additives and pollutants riding along.
• Causation in people. No study has yet shown that microplastics cause a disease in humans. The Marfella result is the closest anyone has come.
• Nanoplastics. The smallest particles are the hardest to measure and the most likely to cross into cells. They’re the least studied.

What to do with all this

Uncertainty is not the same as safety. You don’t need proof of harm to decide you’d rather not eat polyethylene, and the steps that cut particle exposure tend to cut chemical exposure too, which is the better-evidenced risk. In rough order of how good the evidence is that they reduce exposure:

1. Don’t heat food or drinks in plastic. Heat speeds up the release of both chemicals and particles. See microwaving plastic containers.
2. Swap the plastic you cut and cook on. Cutting boards, spatulas and kettles. See our kitchen swap guide.
3. Drink tap water, not bottled, and don’t count on a filter. Treatment plants already remove most microplastics, while household filters are poorly tested against them and some carbon pitchers made things worse. If your water is hard, boiling and decanting has better evidence. See what filters and boiling actually remove.
4. Deal with dust. Indoor air and dust are a major, underrated route. See microplastics in household dust.

And because individual swaps can only do so much against 475 million tonnes a year, push for the thing that actually works at scale: making less plastic in the first place.

We’ll update the grades in this review as new studies land. When one moves, we’ll say which, and why.