Fluoride, Brain Development and IQ
This is the contested one. It is also the one most often reported badly — by people who say the science is settled that fluoride lowers IQ, and by people who say the science shows no effect at all. Neither is what the evidence says. What follows is what the major reviews actually concluded, in their own terms, including the sentence both sides tend to leave out.
Table of Contents
- The Short Version
- Why Anyone Looked in the First Place
- What the NTP Monograph Actually Concluded
- The 2025 JAMA Pediatrics Meta-Analysis
- The Cohort Studies That Drive the Debate
- The Size of the Claimed Effect
- The Criticisms, Taken Seriously
- How a Mechanism Could Work
- What This Means If You Are Pregnant or Have a Baby
- Key Research Papers
- Connections
- Featured Videos
The Short Version
- At water fluoride above about 1.5 mg/L — roughly twice the U.S. target, and common in parts of India, China, Mexico and Iran — there is moderate-confidence evidence of an association with lower IQ in children. That is the U.S. National Toxicology Program’s own conclusion, not an activist one.
- At the U.S. target of 0.7 mg/L, the NTP said explicitly that it could not determine whether there is an effect. Not “no effect” could not determine.
- The evidence is observational. There is no randomised trial of fluoride and IQ and there never will be, which means confounding is a permanent and legitimate objection.
- The strongest studies are pregnancy cohorts measuring fluoride in maternal urine, not ecological comparisons of towns. Those are the ones worth arguing about.
Why Anyone Looked in the First Place
The question did not begin with the internet. It began in fluorosis-endemic regions — places where groundwater naturally carries several milligrams of fluoride per litre and where children visibly have mottled teeth and, in severe cases, bone disease. Researchers in China and India, working in villages with high and low natural fluoride, repeatedly reported lower test scores in the high-fluoride villages.
Those early studies were mostly small, and many were poorly controlled for the things that obviously matter — parental education, income, iodine status, arsenic and lead in the same groundwater. But there were dozens of them, and they mostly pointed the same direction. In 2012 a Harvard team pooled 27 of them (Choi and colleagues) and found a consistent inverse association. The authors were careful in their own paper: they wrote that “some methodological limitations were also noted” and that their review “cannot be used to derive an exposure limit, because the actual exposures of the individual children are not known.” That caveat has been widely ignored by people quoting the paper.
What the NTP Monograph Actually Concluded
The U.S. National Toxicology Program spent years on a systematic review, published in 2024 after considerable internal and external review. Because it is the most authoritative document in this argument, its conclusion is worth stating precisely rather than paraphrasing:
The NTP reported moderate confidence that fluoride exposure at levels above 1.5 mg/L in drinking water is associated with lower IQ in children. It stated that the evidence was insufficient to determine whether exposure at the concentrations used for U.S. community water fluoridation — 0.7 mg/L — affects children’s IQ.
Both halves of that are load-bearing. The first half is a real finding from the country’s own toxicology programme, not something to wave away. The second half is why the same document cannot be cited as proof that U.S. tap water lowers IQ. “Insufficient evidence to determine” is a statement about the state of the research, not a verdict of safety and not a verdict of harm.
The 2025 JAMA Pediatrics Meta-Analysis
In January 2025 a team including NTP researchers published a meta-analysis in JAMA Pediatrics pooling 74 studies. It found an inverse association between fluoride exposure and children’s IQ, with a dose-response pattern: higher exposure, lower scores. Some of its analyses found associations extending below 1.5 mg/L, including in analyses restricted to what the authors judged the higher-quality studies.
This is the paper most often cited as having settled the question. It did not, and the honest reasons are these:
- The great majority of the pooled studies come from regions with fluoride well above U.S. levels. Pooling them and then extrapolating downward is exactly the step that is contested.
- Many included studies were rated by the authors themselves as high risk of bias.
- Confounding by lead, arsenic, iodine deficiency, socioeconomic status and parental education is difficult to remove and varies by region.
- A published exchange followed — critics arguing the low-exposure signal is an artefact, the authors responding in Annals of Global Health that the criticisms do not overturn the pattern. Both sides of that exchange are published, and together they are the actual state of the science.
The Cohort Studies That Drive the Debate
Ecological studies comparing towns are weak. The studies that genuinely moved expert opinion are prospective pregnancy cohorts that measured fluoride in individual mothers’ urine and followed their children:
- ELEMENT (Early Life Exposures in Mexico to Environmental Toxicants; Mexico City, 2017). Bashash and colleagues reported that higher maternal urinary fluoride during pregnancy was associated with lower cognitive scores in the children at ages 4 and 6–12. Mexico does not fluoridate water but does fluoridate salt, giving a spread of exposures. Follow-up work in 2022 examined which cognitive domains were involved.
- MIREC (Maternal-Infant Research on Environmental Chemicals; Canada, 2019). Green and colleagues, in JAMA Pediatrics, reported an association between maternal fluoride exposure and IQ in offspring — notably in a country whose water levels resemble the U.S. This is the study that made the question impossible to dismiss as a high-exposure-country problem. It also drew heavy methodological criticism, including that the association was significant in boys but not girls, which has no established biological explanation. In numbers: each 1 mg/L more fluoride in the mother’s urine went with 4.49 lower IQ points in boys and no significant link in girls; but each 1 mg/day more fluoride taken in by the mother during pregnancy went with 3.66 lower IQ points in boys and girls together. The boys-only pattern applies to the urine measure, not to the intake measure.
- Infant formula (Canada, 2020). Till and colleagues tested children at 3–4 years of age. Each 0.5 mg/L more fluoride in the water went with 9.3 lower Performance (non-verbal) IQ points in children who had been formula-fed and 6.2 points in breast-fed children — a link more pronounced in the formula-fed group, consistent with formula reconstituted in fluoridated tap water being the highest per-kilogram exposure most people ever get.
These are the serious studies. They are also observational, in three specific populations, with outcomes measured by tests that are themselves sensitive to home environment. They justify concern and further work. They do not, individually or together, establish causation at 0.7 mg/L.
The Size of the Claimed Effect
Reported effects cluster around a few IQ points per unit increase in exposure — often quoted as roughly 1 to 5 points per 1 mg/L. The 2025 JAMA Pediatrics meta-analysis gives the most precise figure: in the 13 studies that measured fluoride in each child’s urine, IQ was 1.63 points lower per 1 mg/L more urinary fluoride, and 1.14 points lower among the studies at low risk of bias. Of its 74 studies, 52 were rated at high risk of bias. Where exposure was measured only as fluoride in drinking water, the association across all studies was null below 1.5 mg/L, but stayed inverse below 1.5 mg/L in the low risk-of-bias studies. Two honest observations about a number that size:
It is small for an individual and large for a population. A handful of IQ points is invisible in any one child and swamped by genetics, schooling and home life. But shift an entire population’s distribution a few points to the left and you measurably increase the number of people at the low tail and reduce the number at the high one. This is the same logic that drove lead out of petrol, and it is a legitimate argument.
It is also small enough to be manufactured by confounding. An effect of two or three points is exactly the size that residual confounding produces routinely in observational research. That is not a dismissal — it is why the debate has not resolved.
The Criticisms, Taken Seriously
| Criticism | How much weight it deserves |
|---|---|
| Most studies come from high-fluoride regions | Substantial. This is the central problem with extrapolating to 0.7 mg/L, and the NTP conceded it directly. |
| Co-exposure to lead, arsenic, iodine deficiency | Substantial. High-fluoride groundwater frequently carries other neurotoxicants. Better studies adjust; not all do. |
| Spot urine samples are noisy | Real but two-edged. Measurement error of this kind usually biases toward finding no effect, not toward inventing one. |
| Sex-specific findings with no mechanism | Real. A result that appears in boys but not girls, with no biological account, is a warning sign about robustness. |
| “Decades of fluoridation and no visible drop in IQ” | Weak. Population IQ scores rose through the fluoridation era for unrelated reasons; that trend cannot rule out a small drag. |
| “The researchers are activists” | Not an argument. Address the methods. This applies symmetrically to people defending fluoridation. |
Two U.S. studies published since the meta-analysis point the other way at ordinary water levels. A 2025 analysis of the nationally representative High School and Beyond cohort found that children exposed to the recommended level of fluoride in drinking water scored modestly better on cognitive tests in secondary school, an advantage that was smaller and no longer statistically significant at about age 60. A 2026 analysis of the Wisconsin Longitudinal Study, in PNAS, used historical water-fluoridation records and state school IQ tests and found no evidence that community water fluoridation was linked to lower adolescent IQ or adult cognition; its authors set this against the studies cited in recent decisions to end community water fluoridation in Utah and Florida. Both studies are observational, like the ones they answer.
How a Mechanism Could Work
A mechanism is plausible, not demonstrated. Fluoride crosses the placenta, so fetal exposure tracks maternal exposure. Animal work has reported effects on the hippocampus, on oxidative stress and on thyroid hormone signalling, which matters because maternal thyroid hormone is genuinely required for normal fetal brain development. That last route is the most biologically coherent story on offer: not fluoride poisoning neurons directly, but fluoride nudging maternal thyroid function in women who are already iodine-marginal.
Plausibility is not proof. A mechanism that could work is a reason to keep studying something, not a reason to treat the outcome as established.
What This Means If You Are Pregnant or Have a Baby
The practical upshot is narrower and less dramatic than the argument around it, and it does not require you to decide who is right.
- The exposure that the research actually points at is pregnancy and the first year of life, not adulthood.
- The single largest lever is what you reconstitute infant formula with. Using low-fluoride bottled or filtered water is already a mainstream recommendation for reducing dental fluorosis, and it happens to address the exposure window the IQ research focuses on. That makes it a low-cost decision that is defensible whichever way the science lands.
- Iodine sufficiency in pregnancy matters for fetal brain development on its own, and iodine status is the modifier that keeps appearing in this literature. See Iodine.
- Breastmilk is very low in fluoride. Nursing infants get almost none regardless of the water supply.
- Tea is a further source in pregnancy. In the Canadian MIREC cohort (485 mother-child pairs, published 2026), 61% of mothers drank green or black tea during pregnancy. Tea drinkers had a higher average total fluoride intake (0.76 mg/day) than non-drinkers (0.41 mg/day), and each 0.5 mg/day more fluoride from tea went with 2.24 lower IQ points at age 3–4. The finding is observational; the authors suggested that pregnant women may consider limiting tea and called for clearer labelling of fluoride content.
Not medical advice. In short: formula made up with low-fluoride water lowers an infant’s fluoride intake, breast milk is very low in fluoride, and the U.S. community water target is 0.7 mg/L.
Key Research Papers
Every PMID below was resolved against PubMed while writing this page.
- National Toxicology Program. NTP monograph on the state of the science concerning fluoride exposure and neurodevelopment and cognition: a systematic review. NTP Monogr. 2024. — PubMed PMID: 39172715
- Taylor KW, Eftim SE, Sibrizzi CA, et al. Fluoride exposure and children's IQ scores: a systematic review and meta-analysis. JAMA Pediatr. 2025. — PubMed PMID: 39761023
- Taylor KW, Eftim SE, Sibrizzi CA, Blain RB, Magnuson K, Hartman PA, Bucher JR, Rooney AA. Addressing critiques of the evidence linking fluoride and children's IQ. Ann Glob Health. 2025;91(1):83. — PubMed PMID: 41393310
- Choi AL, Sun G, Zhang Y, Grandjean P. Developmental fluoride neurotoxicity: a systematic review and meta-analysis. Environ Health Perspect. 2012. — PubMed PMID: 22820538
- Grandjean P. Developmental fluoride neurotoxicity: an updated review. Environ Health. 2019. — PubMed PMID: 31856837
- Bashash M, Thomas D, Hu H, et al. Prenatal fluoride exposure and cognitive outcomes in children at 4 and 6-12 years of age in Mexico. Environ Health Perspect. 2017. — PubMed PMID: 28937959
- Goodman CV, Bashash M, Green R, et al. Domain-specific effects of prenatal fluoride exposure on child IQ at 4, 5, and 6-12 years in the ELEMENT cohort. Environ Res. 2022. — PubMed PMID: 35276192
- Green R, Lanphear B, Hornung R, et al. Association between maternal fluoride exposure during pregnancy and IQ scores in offspring in Canada. JAMA Pediatr. 2019. — PubMed PMID: 31424532
- Till C, Green R, Flora D, et al. Fluoride exposure from infant formula and child IQ in a Canadian birth cohort. Environ Int. 2020. — PubMed PMID: 31743803
- Farmus L, Till C, Green R, et al. Critical windows of fluoride neurotoxicity in Canadian children. Environ Res. 2021. — PubMed PMID: 34051202
- Warren JR, Rumore G, Kim S, et al. Childhood fluoride exposure and cognition across the life course. Sci Adv. 2025. — PubMed PMID: 41259530
- Warren JR, Rumore G, Sicinski K, Herd P, Engelman M. Municipal water fluoridation, adolescent IQ, and cognition across the life course: evidence from the Wisconsin Longitudinal Study. Proc Natl Acad Sci U S A. 2026. — PubMed PMID: 41973917
- Khan D, Hall M, Ayoub K, Hornung R, Flora D, Lanphear B, Till C. Prenatal fluoride exposure and child IQ: is tea a key source of fluoride? Environ Pollut. 2026. — PubMed PMID: 42508594
- Grandjean P, Hu H, Till C, et al. A benchmark dose analysis for maternal pregnancy urine-fluoride and IQ in children. Risk Anal. 2022. — PubMed PMID: 34101876
PubMed Topic Searches
Connections
- Fluoride Chelation — natural ways people remove fluoride: boron, iodine displacement, calcium, herbs, amino acids, binders
- Fluoride Toxicity — the overview
- Thyroid Function — the most coherent proposed mechanism
- Dental Fluorosis — the undisputed marker of the same childhood exposure window
- Fluoride in Water
- Water Fluoridation and Informed Consent
- Iodine
- Lead — the confounder, and the precedent for population-level IQ shifts
- How Your Body Reads DNA — animation