Fluoride in Tap Water: The Honest Case Against It

What fluoride is — and how fluoridation started

Fluoride is the ionic form of fluorine, a naturally occurring element that shows up in soil, rock, and groundwater everywhere on Earth. Some water supplies carry a lot of it naturally; some carry almost none. It is not an exotic industrial invention — people have been drinking fluoride-bearing water for as long as people have been drinking water. The dose is the entire story, and that’s a phrase worth holding onto for this whole article.

The fluoridation era started with an observation, not a plan. In the early twentieth century, a dentist named Frederick McKay spent years chasing down why people in Colorado Springs had brown-stained teeth — and noticed those same mottled teeth were oddly resistant to decay. The staining turned out to be dental fluorosis from unusually high natural fluoride in the water; the cavity resistance was fluoride hardening the enamel. Public-health researchers put those two facts together and asked the obvious question: is there a level that gives you the cavity protection without the ugly staining? Grand Rapids, Michigan, became the first city to deliberately add fluoride to its water in 1945, and community water fluoridation spread across North America from there. The current US target, set in 2015, is 0.7 milligrams per liter — deliberately at the low end, chosen to hold onto the dental benefit while minimizing fluorosis.

I want to be fair to that history, because fluoridation was not a conspiracy. It was a genuine public-health win in an era when cavities were rampant, dental care was scarce, and fluoride toothpaste did not exist. That last clause is the whole thing — because the world the program was designed for stopped existing around 1975, and the program didn’t notice.

The fact that breaks the whole program: fluoride works on contact

Start here, because everything else follows from it. Fluoride reduces cavities. I’m not going to pretend otherwise — anti-fluoride writing that denies the dental benefit forfeits the argument before it starts. Fluoride genuinely helps enamel resist acid and remineralize after the daily assault of sugar and bacteria. That chemistry is real and I have no interest in disputing it.

The question is where it has to be for that to happen. And the answer, stated flatly in the caries literature, is: on the tooth. Fluoride works primarily through topical mechanisms — inhibiting demineralization at the crystal surface, enhancing remineralization, and inhibiting bacterial enzymes.⁷ It needs to be sitting in saliva and plaque, in contact with enamel. Which raises the obvious question about swallowing it.

Featherstone answered that question in 1999, in Community Dentistry and Oral Epidemiology, and I want to quote the conclusion precisely because people assume this is a fringe claim: the level of fluoride incorporated into dental mineral by systemic ingestion is insufficient to play a significant role in caries prevention, and the effect of systemically ingested fluoride on caries is minimal.⁶ That is not from an activist blog. It is from the caries-prevention literature, written by one of the field’s most-cited researchers. Even fluoride tablets, he noted, are best used by sucking or chewing them — delivering fluoride topically before it’s swallowed.⁶

And this is not one contrarian’s view, which matters, because “a guy wrote a paper” is not an argument. Hellwig and Lennon put it in Caries Research in 2004: a dogma had existed for decades that fluoride has to be ingested, but recent work concluded the caries-preventive effect is almost exclusively posteruptive — after the tooth is already in your mouth — and therefore topical application should be encouraged.¹⁰ Fifteen years later, ten Cate and Buzalaf wrote the field’s history in the Journal of Dental Research and recorded the settlement: it was generally agreed that fluoride’s effect is primarily topical, by fluoride in the oral fluids, rather than systemic by incorporation into enamel crystals.¹¹ Four independent research groups, across twenty years, in three different journals. That is why I grade this STRONG — it is not a fringe position, it is the consensus of the dental research literature.

Now, precision, because there’s a real subtlety here that fluoridation advocates will correctly point out. “Topical” doesn’t only mean the few seconds water washes over your teeth. Swallowed fluoride does return to the mouth at low concentrations in saliva, and Featherstone himself notes that slightly elevated fluoride in saliva and plaque helps prevent and reverse caries.⁶ So the water route isn’t zero. What Featherstone calls minimal is specifically fluoride built into the tooth mineral by swallowing — and that pre-eruptive, systemic incorporation was the entire theory the fluoridation program was designed around. That theory is the one that didn’t survive. Whatever modest topical benefit the water route retains, we don’t have to argue about it in the abstract, because the next section is the empirical answer: Cochrane measured what it’s actually worth.

Sit with the shape of that. Fluoridating water delivers fluoride in the form that does the least for teeth and reaches every other organ you own — which is why every modern caries-prevention guideline is anchored on the topical routes: toothpaste, varnish, rinses. The program was built on a systemic theory of how fluoride works. The theory was revised. The program wasn’t.

Fluoridating water delivers fluoride in the one form that does the least for your teeth — and the only form that reaches every other organ you own.

What fluoridation actually buys you now

If the mechanism argument is theoretical, the outcome data is the receipt. The classic figures everyone still quotes — fluoridation cuts cavities 25 to 40 percent — come overwhelmingly from studies run before 1975, before fluoride toothpaste was in every bathroom in the country. Those numbers are quoted in 2026 as if nothing happened in between.

Something happened in between. When the 2024 update of the Cochrane review restricted itself to contemporary evidence, the effect collapsed: initiating a fluoridation program today was estimated to change decayed-missing-filled primary teeth by about one-quarter of a tooth, and to raise the proportion of cavity-free children by roughly 3 to 4 percentage points — low-certainty findings whose confidence intervals include the possibility of little or no benefit at all.¹ The reviewers said it themselves: the effect sizes in the toothpaste era are smaller than the pre-1975 studies suggested.

I grade that MODERATE — the finding is a Cochrane review, but the certainty rating is genuinely low, and I’m not going to inflate it just because it helps my case. Here’s the honest reading anyway: we are running a permanent, non-consensual, whole-population intervention whose measured benefit is a fraction of a tooth and whose confidence interval touches zero. That is not a scandal. It is something worse and more ordinary — a policy that outlived its evidence and kept running on institutional momentum. The strongest remaining argument for it is equity, that it helps most where dental access is worst, and I take that seriously. But a quarter of a tooth is a thin foundation for mass medication, and there are cheaper, more targeted, consent-respecting ways to get topical fluoride to underserved kids — varnish programs in schools being the obvious one.

What’s actually in the tank: fluorosilicic acid

Here is the part that bothers me most, and the part almost nobody knows, so let me be very careful to state only what is documented. When you picture water fluoridation, you probably picture sodium fluoride — the compound in your toothpaste, the compound used in the original Grand Rapids trial, a manufactured pharmaceutical-grade chemical. That is not what most water utilities use.

The great majority of fluoridated systems in the United States dose fluorosilicic acid (also called hydrofluorosilicic acid, H₂SiF₆) or its salt sodium fluorosilicate. CDC’s own materials describe fluorosilicic acid as the solution used by most water fluoridation programs in the country, with sodium fluoride typically reserved for small systems. And these compounds are not manufactured for the purpose.

I’m going to use the government’s own words here rather than mine, because this is exactly the kind of claim that gets dismissed as conspiracy talk when it’s paraphrased by someone with a position. CDC’s water-fluoridation additive fact sheet lays out the process: phosphorite rock — used primarily to make phosphate fertilizer — is refluxed with sulfuric acid; the heating releases hydrogen fluoride and silicon tetrafluoride gases; those gases are captured and condensed into a water-based solution of roughly 23% fluorosilicic acid. That solution is then sold to municipalities and metered into drinking water.

And CDC puts numbers on it, which closes the obvious escape hatch. Approximately 95% of the fluorosilicic acid used for water fluoridation in the United States comes from this phosphate-fertilizer process. Because the dry additives are themselves manufactured from it, roughly 90% of the sodium fluoride used in the US traces back to fluorosilicic acid too. So “well, some systems use plain sodium fluoride” doesn’t get you out of it — nearly all of that came from the same place. I grade the sourcing claim STRONG, and I want to underline why: every figure in this paragraph comes from the agency that runs the program, not from an advocacy group.

Now the discipline part, because this is exactly where anti-fluoride writing usually goes off a cliff and hands the other side an easy win. Two common claims about this do not hold, and I’m not going to make them. First: it is often said that you are drinking intact, exotic “silicofluoride complexes.” A US EPA chemist, Edward Urbansky, reviewed the chemistry in Chemical Reviews and concluded that at drinking-water pH and dilution these additives dissociate essentially completely into free fluoride ion and silica — chemically, what reaches your glass is not meaningfully different from fluoride delivered any other way.⁹ That argument is dead; I’d rather bury it than repeat it. Second: the additives are certified under NSF/ANSI Standard 60 and tested for contaminants, and measured contributions of arsenic and heavy metals to finished water are typically small fractions of the regulatory limits — most samples show no detectable arsenic at all, and positive ones run around 1% of the EPA allowable amount. Not zero, and adding any arsenic to drinking water in exchange for a quarter of a tooth is a trade I’d decline — but small, and I’m not going to inflate it.

Third, and I’ll hand this one over too, because it’s the strongest thing the other side has here: CDC argues that pharmaceutical grading wouldn’t even help. USP monographs, they point out, set no limits on arsenic or radionuclides — so a nominally “pharma-grade” sodium fluoride could legally carry more contamination than an NSF/ANSI 60-certified product. Fine. I accept that, and it means “use pharmaceutical grade instead” is a bad ask. But notice it answers a question I’m not asking. I’m not asking for a cleaner grade of fluoride in the water. I’m asking why it’s in the water.

So what’s left of my objection once I’ve given away the two weakest versions of it? This: the substance actually dosed into the public supply has never itself been through a drug-style review for chronic safety and efficacy. Fluoride added to water is regulated by the EPA as a drinking-water contaminant — something to keep below a ceiling — not by the Food & Drug Administration (FDA) as a therapeutic agent with an approved indication, dose, and label. No agency has ever evaluated ingested fluoride the way it would evaluate a medication, which is remarkable given that the entire stated purpose is to treat a disease. We took an industrial byproduct, put it in the water of hundreds of millions of people to prevent a condition it barely prevents by that route, and never ran it through the approval process any pill would have to clear. That is the argument. It doesn’t need embellishment.

The dose nobody controls — and the consent nobody gave

Every drug you have ever been prescribed came with a dose adjusted for something — your weight, your kidney function, your age. Water fluoridation adjusts for nothing. It sets a concentration, not a dose, and the actual dose you receive depends entirely on how much water you happen to drink.

Which means the people getting the highest exposure per kilogram of body weight are the ones you would least want to over-dose. A construction worker in July drinks several times what an office worker does. Someone with diabetes insipidus or a kidney condition drinks more still. And an infant fed formula reconstituted with fluoridated tap water gets, by body weight, the highest fluoride intake of any group in the population — while receiving essentially none of the topical benefit, because they have no teeth yet. This is not a fringe worry: the CDC and the American Dental Association (ADA) both advise that parents concerned about fluorosis can use low-fluoride water for reconstituting infant formula. Read it how you like — I read that as the institutions running the program acknowledging the dosing problem in their own guidance.

Then there is consent, which is the part I find hardest to argue past. Cheng, Chalmers and Sheldon laid this out in the BMJ — and Iain Chalmers is a founder of the Cochrane Collaboration, not an activist — framing fluoridation as a case where the evidence base is weaker than the confidence of its advocates, and where the ethical question of adding a substance to everyone’s water to treat a disease deserves genuine public deliberation rather than dismissal.⁸ You can opt out of a vaccine. You can decline a prescription. You cannot easily opt out of your municipal water, and the burden and cost of avoidance falls on the individual rather than on the party doing the adding. For an intervention with a large benefit that trade might be defensible. For a quarter of a tooth, I don’t think it is.

Dental fluorosis: the dose telling on itself

The first hard evidence that the dose is already too high for some people is the thing McKay saw a century ago. Too much fluoride while teeth are still forming — roughly the first eight years of life — disrupts enamel and leaves dental fluorosis: white flecks and streaks in mild cases, brown staining and pitting in severe ones. It is dose-dependent, and it is real.

The Cochrane analysis put numbers on it. At a fluoride level around 0.7 ppm, roughly 12% of people had fluorosis of aesthetic concern, and about 40% had fluorosis of any level under clinical examination.¹ That fluorosis estimate is rated low-certainty evidence — the same certainty level Cochrane assigns the benefit side, which is worth holding onto, because both halves of the ledger rest on the same quality of data. I’ll be straight about the counterargument: most of that is mild and cosmetic — faint white mottling most people would never notice without a dentist pointing it out. The severe staining-and-pitting form belongs to much higher exposures. I grade it MODERATE, and I won’t call cosmetic mottling a health injury, because it isn’t one.

But look at what the number is telling you. Fluorosis is the visible record of systemic over-exposure during development — the one fluoride effect you can see with your eyes. At the current target, roughly one in eight people examined carries fluorosis of aesthetic concern. Now let me be precise about what that figure is, because this is where anti-fluoride writing usually overreaches and I’d rather not: 12% is total prevalence at 0.7 ppm, not the excess caused by fluoridation. Some fluorosis happens anyway — from swallowed toothpaste, supplements, and naturally occurring fluoride — so you cannot lay the whole 12% at the program’s door, and anyone who does is inflating their case.

What you can say is that the dose-response runs the right way, and that the share fluoridation does contribute is a permanent cosmetic mark on children’s teeth, bought in exchange for a quarter of a tooth of decay prevention. Read as charitably as I know how, that is still a bad ratio. And it raises the question the next section has to handle honestly: if the systemic dose is high enough to visibly alter enamel formation during development, how much curiosity is warranted about what else it might be doing during development? That’s not an accusation. It’s the question the dose itself puts on the table.

One more hazard belongs here for completeness, and I’m going to handle it carefully rather than deploy it. Skeletal fluorosis — fluoride accumulating in bone, causing stiffness, joint pain, ligament calcification, and in advanced cases crippling changes — is a genuine, well-documented disease. It is also a disease of high natural fluoride regions: parts of India, China, and the East African Rift where groundwater runs several to more than ten milligrams per liter for a lifetime. That is many multiples of 0.7 mg/L, and clinically significant skeletal fluorosis is not an established consequence of North American fluoridation; the Cochrane reviewers found the evidence on skeletal effects at fluoridation levels too sparse to conclude anything.¹ So I’m giving it no Evidence Radar grade and I’m not going to imply your tap water is calcifying your ligaments. It belongs in the ceiling-setting conversation. It does not belong in a scare paragraph, and I’d be doing the same thing I criticize if I put it there.

The IQ question — taken seriously, not weaponized

Now the argument that revived this whole debate, and the one worth the most care. Does fluoride harm the developing brain? For years this was dismissed as fringe. It is no longer fringe, and anyone who tells you it is hasn’t read the recent documents.

Two things happened. First, in 2024, the US National Toxicology Program — a serious federal science body — published a monograph concluding, with moderate confidence, that higher fluoride exposure is associated with lower IQ in children. Second, in 2025, a team led by Kyla Taylor published a systematic review and meta-analysis in JAMA Pediatrics pooling 74 studies, and found an inverse dose-response: as fluoride exposure went up, children’s IQ went down, with the association holding even when they restricted the analysis to the higher-quality, low-risk-of-bias studies.² Their headline individual-level estimate was a drop of roughly 1.6 IQ points per 1 mg/L increase in urinary fluoride (about 1.1 points among the low-bias studies).²

This isn’t only ecological, group-average data either. The ELEMENT birth cohort in Mexico — a prospective study measuring fluoride in individual pregnant women’s urine — found that higher prenatal fluoride was associated with lower cognitive scores in their children years later.³ A later benchmark-dose analysis combining ELEMENT with Canada’s MIREC cohort put the exposure associated with a one-point IQ decrement remarkably low.⁴ Prospective cohorts with individual biomarkers are exactly the study design you want, and they point the same direction as the pooled literature.

So I grade “higher fluoride exposure is associated with lower child IQ” as MODERATE — and I want to earn that grade honestly, in both directions, because this is the claim my own side most wants to inflate. In favor: two independent, rigorous federal-science efforts reached the same conclusion, the dose-response is coherent, and it survived restriction to the better studies. Against: most of the underlying data come from high-exposure regions — 45 of the 74 studies in the meta-analysis were from China, a dozen more from India — where fluoride travels together with arsenic, lead, iodine deficiency, and poverty. Many studies were cross-sectional and rated high risk of bias. “Associated with” is not “caused by.” MODERATE is the honest ceiling: real enough to take seriously, not clean enough to call settled.

Here’s how that lands for me, though, and it’s a different move than the one both camps make. The usual fight is over whether the IQ evidence is strong enough to convict fluoridation. I don’t think it is — and I don’t think it has to be. When the benefit you’re defending is a quarter of a tooth, an unresolved neurodevelopmental signal doesn’t need to be proven to change the answer. It just needs to be plausible. A large benefit can absorb a lot of uncertainty. A tiny one cannot. That asymmetry is the actual argument, and it survives even if every criticism of the IQ literature turns out to be correct.

Where I won’t follow my own side

Now the part where I break with people who agree with my conclusion, because a stance is only worth anything if it can say what would falsify it. The claim that 0.7 mg/L fluoridation is proven to lower your child’s IQ is not established, and I’m not going to say it.

Look at what the 2025 meta-analysis actually reported when it sliced the drinking-water data by concentration. The inverse association held below 4 mg/L and below 2 mg/L. But at concentrations below 1.5 mg/L — measured in drinking water — the association was null.² The authors said so directly: there were limited data and real uncertainty about the dose-response when exposure was estimated by drinking water alone below 1.5 mg/L. And 1.5 mg/L is already more than double the 0.7 mg/L US fluoridation target. Here I have to be careful in both directions, because this is the one place my own case is genuinely exposed: when the authors restricted the analysis to the low-risk-of-bias studies, the inverse association persisted below 1.5 mg/L — in drinking-water measurements as well as urinary ones.² The null result comes from the all-studies pool; the better-quality subset didn’t go null. Anyone quoting only the first half of that, in either direction, is cherry-picking. Just as important, the 2024 NTP monograph explicitly did not assess whether the 0.7 mg/L level lowers IQ — its “moderate confidence” conclusion was about exposures above 1.5 mg/L, and it made a point of saying the low-level question was unresolved.

That is why I grade “standard 0.7 mg/L fluoridation is proven to lower IQ” as WEAK. If you have shared a post claiming the science shows your tap water is lowering your kid’s IQ, that post went further than the evidence does, and every time our side does that we hand the other side a free win and lose the credibility we need for the arguments that actually hold.

Two things keep this from being a full retreat, though. First, “not proven harmful” is emphatically not “proven safe.” The authors’ own 2025 response reiterates the point I conceded above and turns it around: among the high-quality evidence, inverse associations were still observed at exposure levels below 1.5 mg/L, on both urinary and drinking-water measurements.⁵ That is not proof of harm at 0.7 — 1.5 is still more than double it, and the NTP still declined to assess the fluoridation level. But it does mean the door is genuinely open, and “the association vanishes at low doses” is not the clean rebuttal it gets used as. Second, and more important: the data thin out exactly where the policy lives. After eighty years of fluoridating water for hundreds of millions of people, we do not have the well-designed low-exposure studies that would answer the question. That absence is not reassurance. It is a measure of how little curiosity the program has had about itself — and when the entity with the burden of proof has failed for eight decades to generate the evidence that would discharge it, “you can’t prove harm” is not the defence its users think it is.

The NTP report and the EPA court ruling

The policy machinery moved in 2024, and it’s worth understanding what actually happened rather than the headline version. After years of delay and internal wrangling, the NTP released its fluoride monograph with that “moderate confidence” high-exposure conclusion. Around the same time, in Food & Water Watch v. EPA, a US federal district court ruled that fluoridation at current US levels poses an unreasonable risk under the Toxic Substances Control Act and ordered the EPA to respond — to conduct a regulatory review of the risk.

Read that ruling carefully, because it gets misreported by people on my side of this. The court did not find fluoridation proven to harm anyone, did not order fluoridation stopped, and did not conclude the 0.7 level lowers IQ. Under TSCA, “unreasonable risk” is a regulatory trigger with a relatively low bar — enough hazard and uncertainty to warrant formal review, not a verdict. Anyone passing this around as “the courts confirmed fluoride is poisoning children” is overstating it, and I’d rather say so than let it stand.

What the ruling does establish is still significant: a federal court examined this record and concluded the risk is serious enough that the EPA must formally respond. Combine that with a Cochrane review shrinking the benefit to a fraction of a tooth and a federal monograph finding a moderate-confidence neurodevelopmental association, all inside about eighteen months, and the direction of institutional travel is not subtle. Both the NTP monograph and the ruling are government and legal documents rather than peer-reviewed studies, and I’m citing them as such rather than dressing them up as trial data.

Europe mostly didn’t fluoridate. Their cavities fell anyway.

Here is the natural experiment that gets remarkably little airtime in North America. The great majority of continental Western Europe does not fluoridate its public water. Germany, Sweden, Denmark, the Netherlands, Norway, Belgium, Austria, Italy, France, Switzerland — most either never started or discontinued it, some on explicit consent grounds rather than safety ones.

And tooth decay in those countries fell over the same decades, on broadly the same trajectory as in fluoridated countries. The DMFT data (decayed, missing and filled teeth — the standard caries index) compiled by the World Health Organization Collaborating Centre for oral health — the CAPP database housed at Malmö University — shows the decline in child caries across the developed world is not neatly sorted by fluoridation status. Denmark and the Netherlands are not experiencing a dental catastrophe. What changed everywhere, fluoridated or not, was fluoride toothpaste — the topical route — along with better dental access and hygiene.

I’ll be careful about what this proves. It is ecological, country-level data, and countries differ in sugar consumption, dental care systems, salt fluoridation, and reporting standards, so this is not a controlled comparison and I’m not grading it. But it is a serious problem for the strong version of the pro-fluoridation case. If adding fluoride to water were doing heavy lifting, the countries that skipped it should look meaningfully worse. They don’t. That is precisely what you’d predict if the benefit is topical and toothpaste is doing the work — which is exactly what the mechanism literature says.⁶

Why “settled science” is the biggest overclaim here

I reserve my one HYPE grade for the claim that keeps this policy alive: that fluoridation is settled science and the question is closed.

Look at what has to be ignored to say that. A 2024 Cochrane review finding the contemporary benefit is roughly a quarter of a tooth with confidence intervals including no benefit.¹ A 2024 federal monograph concluding with moderate confidence that higher fluoride exposure is associated with lower child IQ. A 2025 JAMA Pediatrics meta-analysis finding an inverse dose-response that held in low-bias studies.² Prospective birth-cohort data with individual biomarkers.³⁴ A federal court ordering the EPA to conduct a risk review. A mechanism literature stating that swallowed fluoride does little for teeth.⁶ A Cochrane co-founder publishing in the BMJ that this deserves genuine public deliberation.⁸

That is not the profile of a closed question. It is the profile of a policy whose evidentiary base has been quietly eroding for twenty years while its public defence stayed frozen at “one of the ten great public health achievements of the twentieth century.” It was one of those achievements. Something can be a genuine achievement in 1955 and an unjustified imposition in 2026, and treating people who notice the difference as anti-science is both wrong and strategically stupid — it drives them toward the actual cranks. Asking hard questions about a whole-population intervention is not an attack on public health. It is public health. We’ve made the same argument about the seed-oil consensus — “settled” is a claim about the state of a literature, and it has to be re-earned every time that literature moves.

What I actually do — and what I’d suggest

Policy arguments are cheap. Here’s the practical version, and it’s built on the single best thing about this whole topic: the mechanism lets you keep all of the benefit and drop most of the exposure. Cavity protection is topical. The concern is systemic. You can simply decouple them, and it costs you nothing dentally.

Keep using fluoride toothpaste. I want to be unambiguous about this, because “fluoride-free toothpaste” is where a lot of people land after reading a piece like this, and it’s the wrong conclusion drawn from the right premise. Toothpaste is the delivery route with by far the best benefit-to-exposure ratio — fluoride exactly where it works, spat out rather than swallowed. Throwing it out to avoid systemic fluoride is trading away the only part that actually protects your teeth. Same for varnish at the dentist.

Reduce what you drink, if you want to. A reverse-osmosis system removes fluoride effectively. Standard carbon pitcher and faucet filters mostly do not — this is the single most common mistake, and people spend years believing they've removed something they haven't. Read the certification, specifically for fluoride reduction. Distillation and certain activated-alumina filters also work. The certifications to look for are NSF/ANSI 58 (reverse osmosis), 62 (distillation), and 53 (activated alumina), each with an explicit fluoride-reduction claim on the data sheet.

If you’re mixing infant formula, this is the one place I’d be actively deliberate — use low-fluoride, distilled, or RO water for reconstitution. That’s the highest per-kilogram fluoride exposure in the entire population, delivered to a developing brain, in exchange for topical benefit to teeth that haven’t erupted yet. And you don’t have to take my word for it, which is the point: the CDC and ADA both list it as a reasonable step. If you’re pregnant, it’s worth raising with your clinician given the prenatal-cohort data.³ For a related look at what water-treatment marketing does and doesn’t deliver, see our read on hydrogen water.

Where I land

Line the claims up. Fluoride works topically, and swallowed fluoride does little for teeth — STRONG. The marginal benefit of fluoridating water today is about a quarter of a tooth, with confidence intervals touching zero — MODERATE. The chemical used is a phosphate-fertilizer byproduct that never faced a drug-style review — STRONG. Fluorosis marks one in eight children at aesthetic-concern level — MODERATE. The high-exposure IQ association is real — MODERATE. That 0.7 mg/L is proven to lower IQ — WEAK, not established. That the question is settled — HYPE.

So: I don’t think fluoride belongs in the water. Not because I think your tap is poisoning your children — I’ve spent several sections refusing to say that, and I’d rather lose readers than earn them with a claim I can’t defend. I think it doesn’t belong there because the justification has quietly evaporated. It’s the wrong delivery route for how the drug actually works. The measured benefit is now a rounding error. The compound is an industrial byproduct that no agency ever evaluated as a medicine. The dose can’t be controlled, scales wrong with body weight, and lands hardest on formula-fed infants. Nobody consented. And there’s an unresolved neurodevelopmental question that eighty years of the program failed to resolve.

Any one of those alone might be tolerable. Together they describe an intervention that would never, ever be approved if you proposed it today from scratch — and that survives purely on the inertia of having already been done. My position isn’t “ban fluoride.” It’s narrower and, I think, harder to argue with: take fluoride out of the water and put it on the teeth, where it works and where people can choose it. Keep the toothpaste. Fund school varnish programs for the kids fluoridation was meant to protect. Run the low-exposure studies we should have run decades ago. That’s not the fringe position. It’s roughly what most of Western Europe already does.

What this article is not saying

This is not “fluoride is poison, throw out your toothpaste.” That is the mirror-image overreach and I want no part of it. The topical benefit is real, cavities are a real disease with real consequences, and stripping fluoride from your family’s dental care would be trading a demonstrated benefit for a hypothetical harm — while doing nothing about the systemic exposure I actually object to. If you take one practical thing from this piece, make it “topical yes, swallowed no,” not “fluoride bad.”

This is not a claim that your tap water is measurably harming your child. I graded that WEAK on purpose. The IQ evidence is real at high exposure and unresolved at 0.7 mg/L, and I’d rather state that plainly than borrow certainty I haven’t earned. My case against fluoridation doesn’t rest on it — it rests on mechanism, marginal benefit, sourcing, dosing, and consent, each of which stands without the neurotoxicity argument.

This is not an accusation of bad faith against dentists or public-health workers. Fluoridation was a real achievement, implemented by people trying to help poor children keep their teeth, and it worked in the world it was designed for. Policies can outlive their evidence without anyone having lied. That’s the ordinary and much more common failure mode, and it’s what I think happened here — the same pattern we walked through with the dairy guidance, where an old recommendation kept its authority long after the data underneath it had shifted.

And this is not personalized medical or dental advice. Every figure here describes what published studies and government reports actually reported; every place the author extended past the data into judgment is flagged as a stance. If you’re pregnant, caring for a young child, or making decisions about your family’s water or dental care, weigh it with a clinician or dentist who knows your full picture — not an article with a point of view, however cited. The goal is to give you a position you can actually defend at the dinner table, in either direction. For the wider evidence-first view, browse the lifestyle hub and the full Evidence Radar.