Are PFAS Dangerous? Are they microplastics? What the Science Actually Shows, and What It Doesn't

If you read this and it gets you worried about PFAS, two things you can do:

1. If you are in a profession or place that has confirmed high levels of PFAS, consider signing up for the pilot of NeutraOat, a modified oat fiber that's designed to remove PFAS from the gut and the blood. If you qualify, you'll get 12 weeks of NeutraOat, plus before and after PFAS tests and blood panels. More info at the link.
2. If you aren't a firefighter or living in a contaminated place, consider reserving a spot in the first commercial batch of NeutraOat, coming out this fall. It requires a $20 refundable deposit, which helps us fund our first batch and prove to our investors that there's commercial demand for this.

There are two easy ways to be wrong about PFAS.

The first is to treat forever chemicals as a horror movie villain: one substance, every disease, in everything, unstoppable. The second is to notice that the first position is overheated, decide the whole field is junk epidemiology, and stop paying attention.

Both are wrong. The second is more expensive, because the evidence is better than the skeptics think. It just isn't evenly good. Some PFAS effects are as well established as anything in environmental health. Some are real associations with unresolved causality. Some are things people assert confidently on the internet that nobody has ever demonstrated.

This post sorts them. I'll tell you what I'd bet money on, what I think is probably true, what's still open, and what I think is overstated, including by people who agree with me.

Disclosure, up front: I'm building NeutraOat, a modified oat fiber designed to bind PFAS in the gut. I have an obvious interest in you concluding that PFAS are dangerous. Read everything below with that in mind. It's also why I'm explicit about the weak parts. If my case only works when I hide the soft spots, it isn't a case.

This is the hedged companion to my longer argument that PFAS are dangerous at low doses. That one makes the affirmative case. This one shows the seams.

Are PFAS microplastics? No, and the difference matters

The popular understanding of the two has fused, and microplastics research has had a rough few years: sensational findings (a credit card of plastic in your brain) followed by quiet corrections about glove contamination and detection artifacts.

PFAS science is in a different state. We know what the molecules are. We know their structure and their chemistry. We can measure them in serum reliably and cheaply, with validated methods. We know roughly how long they stay in you. We have decades of occupational cohorts, several large community cohorts created by contamination events, animal data, and mechanistic work.

None of that makes every PFAS claim true. It does mean most of them can be checked, and many have been.

What PFAS actually do to the human body, sorted by evidence strength

Tier 1: The PFAS health effects that are well established

These are the claims I think a hostile, competent reviewer would concede.

PFAS are in nearly everyone. Detectable in the blood of roughly 97% of Americans. That figure is measured, not modeled. NHANES has been running the assay for two decades.

They persist. Legacy long chain PFAS like PFOA and PFOS have serum half lives measured in years. That follows from the chemistry. The carbon fluorine bond is one of the strongest in organic chemistry, and the molecules resemble fatty acids closely enough that your fat handling machinery keeps picking them up and recirculating them without ever breaking them down.

They raise cholesterol. The National Academies' 2022 clinical guidance found sufficient evidence of an association between PFAS exposure and dyslipidemia in adults and children, alongside decreased antibody response, decreased infant and fetal growth, and kidney cancer in adults. Note who is saying this. NASEM is a conservative body writing guidance for practicing clinicians, not an advocacy group. Treat that list as the floor of the evidence.

They blunt vaccine response. Same tier, same source. Children with higher PFAS levels mount weaker antibody responses to routine vaccines, and the finding has replicated across multiple cohorts in multiple countries.

Regulators treat them as carcinogens. In November 2023, an IARC working group classified PFOA as carcinogenic to humans (Group 1) and PFOS as possibly carcinogenic (Group 2B). Read the basis carefully, because it's instructive. PFOA's Group 1 rating rested on sufficient evidence in experimental animals plus strong mechanistic evidence in exposed humans, with only limited evidence for cancer in humans, specifically renal cell and testicular carcinoma. That's a real classification. It also isn't the same claim as "PFOA causes cancer in people the way tobacco does," and anyone who tells you it is has skipped a step.

Tier 2: PFAS and cancer, thyroid, and liver (real associations, weaker evidence)

NASEM's second tier, limited or suggestive evidence, covers breast cancer, liver enzyme alterations in adults and children, pregnancy induced hypertension, testicular cancer, thyroid disease and dysfunction, and ulcerative colitis.

I believe some of these. I won't build an argument on them, for three reasons that apply to most human PFAS epidemiology.

1. Health histories don't build themselves. Most medical data sits with a doctor or a patient, and where it's aggregated, it's usually aggregated for billing. Billing data has incentives pushing diagnoses up and incentives pushing them down. Using it to establish cause takes great care and a lot of assumptions.

2. Isolating PFAS is hard. The most exposed populations, meaning chemical workers, firefighters, and people living near military bases, are exposed to a whole stew of nasty chemicals. Firefighters inhale combustion products by the lungful. Pinning a specific cancer in that population on PFAS specifically is a genuinely difficult inference, and I don't think the field has solved it.

3. The ground truth wobbles. Thyroid cancer is the cleanest example. It gets badly under-diagnosed when nobody is looking and badly over-diagnosed when they are, because a suspicious nodule often gets a thyroid removed and "thyroid cancer" entered in the chart with no confirmatory biopsy. South Korea ran the natural experiment: aggressive screening produced an enormous surge in diagnoses with essentially no effect on thyroid cancer mortality. A PFAS-exposed community also gets more medical attention than a control community, almost by definition. You will find more thyroid cancer there whether or not PFAS caused any of it.

So when I see suggestive studies on kidney and testicular cancer, I take them seriously and I still don't lean on them. I'd rather lean on the boring endpoints, cholesterol and immune response, because they're continuous, objectively measured, predicted in advance by the mechanism, and independently tied to hard clinical outcomes.

Tier 3: PFAS and fertility, neurodevelopment, and diabetes (genuinely unresolved)

I don't know the answer to these, and I don't think anyone else does either.

• Neurodevelopment. Plausible mechanism, messy literature, enormous confounding by socioeconomic status.
• Fertility and reproductive outcomes. Some signal, a lot of noise, hard to separate from everything else that correlates with living near industry.
• Metabolic disease, meaning obesity and type 2 diabetes. Mechanistically tempting given the PPAR-alpha story. Empirically inconsistent.
• Cardiovascular outcomes outside the cholesterol pathway. NASEM put these in the inadequate evidence bin and I think that's right. My own harm estimate deliberately routes through cholesterol rather than claiming a direct vascular effect.
• Whether short chain replacements are safer. The big one, and unresolved in both directions. The case for: they're less fatty acid like, bind albumin less avidly, and clear in days to weeks rather than years. The case against: smaller molecules tend to cross barriers more easily, including placental and organ barriers, and we have a fraction of the toxicology. We phased out the compounds we studied and replaced them with compounds we haven't.
• Mixture effects. "PFAS" covers thousands of compounds. Almost all the health evidence concerns about five of them.

Tier 4: The PFAS claims that are overstated

Things I think my own side gets wrong.

• Treating PFAS as one substance. A finding about PFOS is not a finding about GenX. The umbrella term is a regulatory convenience that leaked into science communication, and it does real damage to credibility.
• Causal language from cross-sectional data. A single NHANES slice showing PFAS correlates with condition X is a hypothesis, not a result. Reverse causation is a live threat here specifically, because PFAS and bile acids share the enterohepatic recirculation pathway. Anything that changes cholesterol clearance also changes PFAS clearance. That cuts against my own cholesterol argument, which is why I flag it instead of burying it.
• Body burden theater. "There is X amount of forever chemical in your blood" is not by itself a health claim. The health claim needs a dose response relationship, which is the actual question.

How PFAS raise cholesterol, and what that costs in lives

Why I keep coming back to cholesterol instead of cancer:

The relationship between serum LDL and cardiovascular risk is one of the most thoroughly established findings in all of epidemiology. Log linear, replicated across observational cohorts, randomized drug trials, and Mendelian randomization. If you can establish that PFAS moves LDL by some amount, you can convert that into expected events using machinery that already exists and that nobody seriously disputes.

When I run that conversion, my central estimate is roughly 3,400 excess cardiovascular deaths per year in the US, about 12,000 additional non-fatal heart attacks, and around 50,000 QALYs lost annually. The full methodology is open, so you can check it or break it.

It's a central estimate with wide uncertainty, and it rests on an assumption I'll defend but won't pretend is settled: that the dose response has no safe floor.

Is there a safe level of PFAS? The threshold vs linear debate

This is the question that actually matters for most readers, and it's the most contested part of the field.

The two models are threshold, meaning fine below some level and harmful above it, and linear, or more precisely additive, meaning each additional molecule buys a little more risk. The theoretical argument for additive comes from Crump, Hoel, Langley and Peto (1976). When a toxin plugs into a biological process that's already running, the first dose already buys some extra risk, which makes a genuine threshold hard to justify.

PFAS plug into a process that is very much already running. Fatty acids signal through PPAR-alpha, ride on albumin and FABP1, and get cleared by renal transporters, thousands of times a day. PFAS run the same pathways and don't get degraded at the end of the trip. The receptor stays occupied longer than it should.

The empirical picture is messier, and I want to be honest about the fight. The strongest support for a low dose linear effect on vaccine response comes from the Faroe Islands cohort, which found a log linear reduction in tetanus and diphtheria antibodies down to very low PFOA. Against it, a German study found that a threshold model fit best, and that's the study Europe based its limits on. Against that, a reanalysis argued the German group's statistics were wrong and a linear model fits better. The reanalysis is coauthored by the Faroe Islands researcher.

I side with the linear reading, mostly because the German analysis threw away an entire vaccine's worth of data for failing significance at low doses. Tiny effects at the lowest doses are exactly what a linear model predicts. Discard everything below a given effect size and you've assumed your conclusion.

You should also know that these studies are unusually political. They set regulatory limits and they decide liability in litigation worth billions. "Safe below a threshold" versus "harmful at low doses" is a fight over money as much as a fight over statistics, and that's true of the researchers I agree with as well as the ones I don't.

What would change my mind about PFAS

A reasonable thing to ask of anyone arguing a position they profit from.

• A trial where serum PFAS drops substantially and lipids don't follow. If you meaningfully reduce someone's PFAS burden and their cholesterol doesn't move, the causal story I've built is in real trouble. Nobody has run that experiment cleanly yet. Cholestyramine is a poor tool for it, since it lowers LDL by itself, so a clean test needs a binder that doesn't already treat the endpoint.
• The Faroe findings failing in a new, pre-registered cohort. The vaccine response evidence is load bearing for the low dose argument.
• Short chain replacements clearing fast enough not to accumulate. That would make this a legacy cleanup problem rather than an ongoing exposure problem. A much better world, and a much smaller one for what I'm building.

What the DuPont and 3M settlements tell us about the science

One thing that raises my confidence: unlike most scientific claims, these got adversarially tested by parties with enormous incentive to destroy them.

DuPont paid roughly $670M over PFAS dumping, and that settlement funded the C8 Science Panel, the first large PFAS epidemiology program. 3M paid $10.3B over water system contamination. Both companies had billions of reasons to litigate the science into the ground, and their own internal research to do it with. They settled.

Companies settle for all sorts of reasons, so this proves nothing on its own. It's still a kind of stress testing that most published findings never get.

So, are PFAS dangerous? The short answer

Yes, at the population level, and I'd say with high confidence. But the confidence attaches to specific mechanisms rather than to the whole cloud of claims.

PFAS reliably raise cholesterol and blunt immune response at levels common in the general population. Those two effects on their own are enough to impute real mortality, and that's where my estimate comes from. The cancer evidence is real and weaker than the advocacy suggests. The neurological, metabolic and reproductive claims are mostly unresolved. The newer compounds we replaced the old ones with are largely unstudied.

For any individual, it depends on your exposure, and most people don't know theirs. If you've lived near an airbase, a firefighting foam site, or industrial contamination, or you're a firefighter, your level is probably well above the general population's. NASEM's clinical guidance changes what screening it recommends above 2 ng/mL and again above 20 ng/mL. A blood test is the only way to find out where you sit.

Meanwhile the regulatory floor is moving. In its May 2026 proposals, EPA upheld the 4 ppt drinking water limits for PFOA and PFOS while offering water systems the option of two additional years, to 2031, to comply. It separately proposed rescinding the limits for PFHxS, PFNA, GenX and the hazard index mixture. Several states are keeping their own standards regardless. Whatever you make of the legal reasoning, the practical effect is that fewer compounds are federally regulated and the deadline for the two that remain is further away.

How to reduce your PFAS exposure

• Filter your water. Get reverse osmosis and stop researching filters. If you're on well water or your utility has flagged PFAS, this is the largest lever you control.
• Watch your animal product sources. PFAS accumulate up the food chain, so animal products from unregulated or contaminated sources are the higher risk category.
• Find out where you stand if you have any plausible high exposure. A serum panel is cheap now, and it turns a vague worry into a number you can act on.

Those three address what's coming in. They do nothing about what's already stored in you, which keeps recirculating between your gut and your bloodstream for years. That loop is the problem I work on.

---

Where this goes next

We're trying to be the rigorous voice in this space while building a real product. That means publishing the uncertainty alongside the argument, and telling you when our own data comes back badly.

NeutraOat is a modified oat fiber designed to bind PFAS in the gut so they leave the body instead of recirculating. It's in a human pilot in highly exposed people right now, with first results in October 2026. I'm not going to tell you it works until my own data says so.

If you want to see where it goes, you can reserve a spot for a fully refundable $20 deposit, credited toward your first bottle, refunded on request, no commitment. You'll get the October pilot results before anyone else, whatever they say.

Reserve my place →

Reserving holds your place for a product in active testing. It isn't a treatment, and it isn't a drug approved to treat or prevent any disease.