The 2025 fluoride IQ study is real, and the US government ran it. It pooled 74 studies. Children with more fluoride in their urine scored lower on IQ tests (Meta-analysis). The gap was about 1.63 IQ points for each extra milligram per liter. But when the team looked at drinking water instead, that link disappeared below 1.5 mg per liter. That is more than twice the level in US taps (Meta-analysis).
So the answer depends on the dose. A lot of fluoride does look like a real risk to children’s brains. The amount in US and UK taps does not, on the evidence we have. There is no need to panic and no need to buy a filter. The useful step is to find out your own water’s number.
What the Fluoride IQ Study Did
A meta-analysis gathers every study on one question and pools the results. A pattern too faint to see in one small study can show up once you stack them.
This one came from the National Toxicology Program, the toxicology arm of the US National Institutes of Health. It was published in JAMA Pediatrics in January 2025 (Meta-analysis).
The team included 74 studies. Sixty-four of them were snapshots. They measured fluoride and IQ at a single moment. Only 10 followed children over time. Every study compared fluoride in drinking water or urine with children’s IQ scores.
The direction was consistent. In 64 of the 74 studies, more fluoride went with lower scores (Meta-analysis).
Thirteen studies measured fluoride in each child’s own urine. Pooled together, IQ was about 1.63 points lower for each extra milligram per liter. In the better-built studies the gap was smaller, about 1.14 points (Meta-analysis).
One line in the paper is easy to miss. The authors say that no study of fluoride and children’s IQ had ever been run in the United States (Meta-analysis).
The Dose Almost Nobody Reports
Where those 74 studies came from matters more than any single number.
Forty-five of them were done in China. India supplied 12, Iran 4 and Mexico 4. Canada added 3 and Pakistan 2. Denmark, New Zealand, Spain and Taiwan had one each (Meta-analysis).
In much of that research, nobody added the fluoride. It was already in the groundwater, dissolved out of the rock. Levels were often several times what a US water plant aims for.
Three numbers are worth remembering. US community water is fluoridated to 0.7 mg per liter. The World Health Organization’s guideline for drinking water is 1.5 mg per liter. The Environmental Protection Agency’s enforceable ceiling is 4.0 mg per liter (Meta-analysis). A softer 2.0 limit is not enforced, but a system above it has to tell the people it serves (Reference).
Most Americans live nowhere near the top of that range. About 62.8% of the population drinks fluoridated water (Reference). Only about 4.5% of community water systems average 1.5 mg per liter or more, serving some 2.9 million people (Study). That is under 2% of everyone on public water.
None of that makes the 74 studies fake. It means most of them looked at water far stronger than what comes out of a US tap.
With most things in the environment, the dose decides whether there is harm. Not everything works that way. Our guide to fine-particle air pollution covers an exposure that seems to do harm at any level. Fluoride, on the evidence so far, does not look like that.
Where the Signal Fades
The best evidence on low doses is inside the same paper.
The team ran the drinking-water numbers again at three cut-offs. Below 4 mg per liter the link held firm. Below 2 it weakened and blurred. Below 1.5 there was no link left at all (Meta-analysis).
That last step rested on 7 studies and 2,832 children. It is a small base for a firm answer.
It is also the version that includes the weakest studies. When the authors kept only the 22 they rated sound, the link stayed negative at every cut-off, below 1.5 included (Meta-analysis). That version leaned on 3 studies and 879 children. With numbers that small, the true answer could still be no effect at all.
So this is not a clean bill of health, and the authors said so. Their conclusion flags limited data and real uncertainty below 1.5 mg per liter, when exposure was judged from drinking water alone (Meta-analysis).
They were also clear about the question they had not answered. The work was not designed to settle whether fluoridating US community water harms anyone.
The same team’s 2024 federal review drew the line in the same place. It reported moderate confidence that fluoride above 1.5 mg per liter goes with lower IQ in children (Review). It then said plainly that there was not enough data to judge 0.7 mg per liter, the US level (Review).
Moderate confidence is a formal grade. Reviewers score their own certainty from very low up to high. Moderate means the finding will probably hold, but new work could move it.
For adults, the same review found only low confidence of any effect on thinking (Review). It had just two snapshot studies to work from.
Why Experts Still Disagree
Two editorials ran beside the meta-analysis in the same issue, arguing opposite ways.
The critics begin with study quality. The federal checklist the team used rated 52 of the 74 studies as seriously flawed, against 22 rated sound (Meta-analysis). Most were single-moment snapshots, which cannot show what came first. A spot urine sample captures one day, not a childhood.
Steven Levy’s editorial made those points and urged caution about the whole evidence base (Editorial).
Bruce Lanphear and two colleagues read the same results the other way. They argued it is time to reassess fluoride during early brain development (Editorial). They noted that the review took nine years and used unusual steps to keep bias out.
The link also proved durable. It held when the analysis used only the studies rated sound, and when fluoride was measured in urine rather than water (Meta-analysis).
The authors have since published a point-by-point reply to the usual criticisms (Response). They note that 19 of the 22 sounder studies reported lower IQ with more fluoride. Seventeen of the 22 also handled rival explanations well.
That reply comes from the authors themselves, not from an outside referee. So the argument stays open. Studies built for this exact question are what can close it.
Testing the Water We Drink
Three recent studies did what the 74 mostly could not. They measured the exposure people in fluoridated countries actually get.
The largest came out in 2026. Researchers followed 10,317 members of the Wisconsin high school class of 1957 (Study). They worked out each person’s childhood fluoride from federal water records, up to age 14. That came before any of the testing. Schools had already given everyone a standard IQ test. Thinking tests followed at 53, 64, 72 and 80.
The result was flat. Of 45 key comparisons, only two showed a difference, about what chance alone would produce (Study). People raised on fluoridated water scored the same as those never exposed, at every age.
The same group had already tested a nationwide sample of 57,960 US students, surveyed in 1980 (Study). Children exposed at recommended levels scored slightly higher in 12th-grade math, reading and vocabulary. The edge was small but steady.
Researchers went back to 13,260 of them in 2021, near age 60. By then the advantage had faded to almost nothing, though it still pointed the same way (Study).
That study has one hard limit. Too few students lived where fluoride ran above 1.2 mg per liter, so it could not test high exposure at all (Study). It tells you about US taps. It says nothing about a village well with naturally high fluoride.
The third study solved the exposure problem another way. An Australian team used dental fluorosis, the faint white flecking that fluoride leaves in enamel. They treated it as a physical record of how much a child took in (Study). They also tracked how much of the first five years each child spent on fluoridated water.
Of those children, 357 sat a full IQ test with a trained psychologist, at ages 16 to 26. The average score was 109. The team had set a fairness line in advance. Scores would count as equal if they differed by less than 3.75 IQ points (Study). Exposed and unexposed children came out equal on both measures, with the exposed group a shade ahead.
A 2023 meta-analysis had already reached the same place. Looking only at 8 comparisons from areas without naturally high fluoride, it found no link to lower IQ (Meta-analysis). The link only showed up in the 23 comparisons from high-fluoride areas. There the exposed group averaged about 3.9 mg per liter, against about 0.7.
All of these studies share a weakness. They watched what happened rather than assigning anyone to drink fluoride, so a hidden third factor could explain the results. What makes them convincing is the agreement. Teams in different countries, using different methods, landed in the same place.
One disagreement is still open. Taylor’s team found a link with fluoride measured in urine, even at low levels. The 2023 meta-analysis found none (Meta-analysis). So the low-dose picture is not fully settled.
What Fluoride Does for Teeth
Nobody argues about whether fluoride protects teeth. The argument is about how much it helps now.
Cochrane reviewers updated the water fluoridation evidence in 2024 and took in 157 studies (Review). None of them could decide by coin flip who got fluoride, because you cannot split a city’s water supply that way. Only a handful actually watched a town start fluoridating.
The reviewers split those studies at 1975 on purpose. That is when fluoride toothpaste became normal in most homes.
Two of them fall after that line. In those, starting fluoridation saved about a quarter of a baby tooth per child, counting teeth that were decayed, missing or filled (Review). About 4 more children in every 100 stayed free of decay. Both numbers rest on weak evidence, and both leave room for no benefit at all.
The older studies looked far better. Before 1975, starting fluoridation was linked to roughly two baby teeth saved per child. Cochrane rates that evidence weaker still (Review).
The familiar figure of about a third fewer cavities comes from Cochrane’s 2015 version, and most of its studies predate the toothpaste tube (Review). The tap is no longer anyone’s only source.
A small benefit per child still adds up across a country. A 2025 study modeled dropping fluoride from every US public water system (Study). It projected 25.4 million more decayed teeth in US children and $9.8 billion in extra costs over five years. Over ten years that rose to 53.8 million teeth and $19.4 billion.
Those costs would not land evenly. The model put the heaviest burden on publicly insured and uninsured children (Study). It also priced only teeth. The authors left brain effects out, because US tap levels are not definitively tied to worse outcomes.
Real stoppages point the same way. Calgary stopped fluoridating in 2011. Seven years later, about 65 in every 100 seven-year-olds there had decay in their baby teeth. In Edmonton, which kept fluoridating, it was about 55 in 100 (Study). The gap between insured and uninsured children tended to widen in the city that stopped (Study).
These dental numbers do not answer the IQ question. They sit on the other side of the same decision, and both sets of numbers are real.
What Parents Can Actually Do
Three things are worth doing.
Know your number. If you are on mains water, your utility publishes its fluoride level in an annual quality report. Look it up once. Private wells are the real exception. They sit outside monitored municipal fluoridation, and natural levels vary a lot (Study). A well is worth a lab test.
Supervise the toothpaste. The American Dental Association advises a smear the size of a grain of rice until age 3, then a pea from 3 to 6 (Guidance). The rice smear holds about 0.1 mg of fluoride and the pea about 0.25 mg. Teach your child to spit rather than swallow. Children under 6 may swallow around 0.3 g of paste per brushing (Meta-analysis). Swallowed paste is the main cause of dental fluorosis, and a lower-strength paste before age 6 probably lowers that risk (Review).
Do not add fluoride drops or tablets yourself. Supplements are dosed against the fluoride already in your water. That makes them a dentist’s call rather than a shopping decision.
Two popular moves have no support in the evidence. Skipping fluoride toothpaste is one (Meta-analysis). Filtering fluoride out of tap water at typical levels is the other (Study). Both cost your teeth something real, in exchange for a risk nobody has measured at those levels.
Frequently Asked Questions
Does the fluoride IQ study mean tap water lowers my child’s IQ?
Not at the level US and UK taps run at, on current evidence. The 2025 fluoride IQ study found the drinking-water link disappeared below 1.5 mg per liter. That is more than twice the US target of 0.7 (Meta-analysis). Large studies at US levels found no cost to IQ in childhood or in old age (Study). The concern is real where groundwater runs naturally high.
How much fluoride is in US tap water?
Community systems aim for 0.7 mg per liter (Meta-analysis). About 62.8% of Americans get fluoridated water (Reference). The enforceable federal ceiling is 4.0 mg per liter, which mostly matters for naturally high groundwater (Reference). Your utility’s annual report gives your own number.
Should I switch to fluoride-free toothpaste?
The evidence does not support it. Fluoride toothpaste cut decay compared with fluoride-free paste across 85 trials in 48,804 children and teenagers (Meta-analysis). For young children, use a lower-strength paste and watch the amount. A lower-fluoride paste before age 6 may lower the chance of white enamel marks (Review). Ask a dentist before you switch.
What are the white spots on my child’s teeth?
Faint white flecks or lines usually mean mild dental fluorosis. It comes from swallowing fluoride while the adult teeth were forming (Review). Reviewers most often found the mild, cosmetic version. Swallowed toothpaste is the usual source, so watch the amount and teach spitting (Guidance). Ask a dentist if the marks are brown, pitted or spreading.
Key Takeaways
- The study is real and government-run: the National Toxicology Program pooled 74 studies. More fluoride went with lower IQ, about 1.63 points per extra milligram per liter in urine (Meta-analysis).
- Most of its evidence is not about your tap: 45 studies came from China and 12 from India. That groundwater often carries natural fluoride far above the US target of 0.7 mg per liter (Meta-analysis).
- The link fades at the level we drink: it held below 4 and below 2 mg per liter, then vanished below 1.5. The federal review found too little data to judge 0.7 (Review).
- Studies at US levels find no cost: 10,317 people were tracked from high school to age 80. No link showed up to IQ or to later thinking (Study).
- A biological marker agrees: 357 young Australians with fluorosis from childhood intake scored the same as those without (Study).
- The dental benefit is real but smaller: modern studies point to about a quarter of a tooth saved. Before fluoride toothpaste it was roughly two teeth (Review).
Know Your Number, Not the Headline
Both sets of headlines flattened this. One said fluoride lowers your child’s IQ. The other said the study was junk. Neither is true.
The 74 studies did find a consistent link, and the federal reviewers took it seriously enough to grade it. But that evidence mostly describes water two to five times stronger than a US tap. At the levels most readers drink, three large studies built for that question found nothing lost. The benefit to teeth still shows up, smaller than it once was.
One fact is worth having, and it is about your own home rather than the news. Find your utility’s annual water quality report and read the fluoride line. If you are on a well, get it tested. Then watch how much toothpaste a small child swallows, and let the rest go.
The evidence is still building, and it is worth watching.
This article is for educational purposes and is not medical advice. Talk to a qualified clinician before changing your health regimen.
Leave a comment