Excavation · Does Fluoride in Drinking Water Lower Children's IQ? (1945 to the present)
Area: Medicine and Health · Law and Justice · Media and Misinformation · Question: Is a widely shared claim true? · Starts from a widely shared claim
Source notices. Some sources this page discusses are retracted, corrected or unpublished.
The short answer
Partly
Current assessment · as of 2026-10-06
Unsettled at the level added to tap water (0.7 mg/L). At natural levels above 1.5 mg/L, studies consistently find lower IQ (moderate confidence). Pregnancy studies at lower levels are split.
Central claim: Community water fluoridation at the level used in the US and Canada (0.7 mg/L) lowers children's IQ.
Evidence ●●○○○ Public belief ●●●○○ Source checked: primary
What would settle it. For 0.7 mg/L: a large prospective cohort with individual fluoride measures (repeated urine or water-plus-diet intake) in a fluoridated population, with maternal IQ and lead measured, that either reproduces or fails to reproduce the Canadian and Mexican pregnancy results; or a natural experiment comparing cognition in children born before and after a city stopped or restarted fluoridation (Calgary stopped 2011 and restarted 2025; dt-calgary-gap). Not yet read and listed as sub-questions (fl-sq-...): the JAMA Pediatrics editorials of 2019 and 2025, the full Cochrane 2024 review, Health Canada's current position, what has happened in the court case since the 2026 remand, any formal HHS directive to the CDC, and the original source of the "Nazi camps" claim.
The answer depends on the dose, and the two halves of the question have different answers. At high natural levels, mostly in parts of China, India, Iran and Mexico, children exposed to water above 1.5 mg/L score lower on IQ tests: the US National Toxicology Program (2024) rated this "moderate confidence"; of its 19 low-risk-of-bias studies, 18 reported lower IQ with more fluoride (those 19 include two Canadian studies at community levels; the NTP found results below 1.5 mg/L less consistent), and Europe's food safety agency (2025) reached the same conclusion above 1.5 mg/L. At the 0.7 mg/L added to community water in the US and Canada, no effect has been established, and the evidence divides by how exposure was measured. Five studies that compare children by where they lived (New Zealand, Sweden, Australia, two large US cohorts) find no lower IQ, and a pooled analysis of water studies below 1.5 mg/L found no association. One 2026 US cohort that also assigns exposure by home address reports lower fluid cognition above about 0.675 mg/L (abstract only; not yet read in full). Studies that measure fluoride in pregnant women's urine (Canada, Mexico) found lower IQ in children, while similar studies in Denmark and Spain did not, and Spain's pointed the other way. The NTP says its findings do not address whether the sole exposure to fluoride added to drinking water at 0.7 mg/L has a measurable effect, and that they may also be relevant in fluoridated US areas depending on people's other sources of fluoride. A 2024 federal court finding of "unreasonable risk" was a regulatory ruling, not a scientific finding of harm, and it was vacated on procedural grounds in May 2026. What health bodies and governments now do is shown separately, because it is adoption, not evidence: Utah and Florida banned fluoridation in 2025, the FDA restricted fluoride drops and tablets for young children, the EPA is reviewing the standard, and the CDC's own page still recommends fluoridation. The benefit side has also weakened: the 2024 Cochrane review found the effect on tooth decay in modern studies small and uncertain.
Based on: fl-high-exposure-lower-iq fl-water-07-no-lower-iq fl-maternal-urine-signal fl-fluoridation-lowers-iq fl-ntp-silent-on-07 fl-court-ruling-is-not-a-finding fl-small-effect-not-excludable
We give no probability figure. Confidence words describe how sure we are of the basis, and no source supplies a number that would not be invented.
Challenge this assessment A challenge needs a specific claim and evidence we can check. How challenges work.
established confidence moderate · our conclusion
Children exposed to drinking water with more than 1.5 mg/L of fluoride (the WHO guideline value) score lower on IQ tests than children with less exposure, consistently across many studies, most of them from high-fluoride regions of China, India, Iran and Mexico.
Anchor. NTP Monograph 08 (Aug 2024), Abstract p. xix: "with moderate confidence", water fluoride concentrations exceeding 1.5 mg/L "are consistently associated with lower IQ in children"; 72 IQ studies, 19 low risk of bias, 18 of them inverse; more than 7,000 children in the 19; countries China 10, India 3, Mexico 3, Canada 2, Iran 1 [read]. Taylor et al. 2025, Table 2: water-fluoride dose-response over all data, 31 studies, slope -0.15 (-0.20 to -0.11) [read]. Choi et al. 2012: 27 studies (25 China, 2 Iran), SMD -0.45 (-0.56 to -0.35), exposed-group water up to 11.5 mg/L [read]. EFSA 2025 summary: adverse effects on the developing brain at concentrations above 1.5 mg/L [read, summary only].
Would change if. well-conducted prospective cohorts in high-fluoride areas, measuring lead, arsenic, iodine and parental IQ, find no association
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
The question is dose-dependent and is filed as two questions that must not be fused: high natural exposure (above 1.5 mg/L), where the evidence is consistent; and community water fluoridation (0.7 mg/L), where it is not. The dig keeps apart (1) what the studies measured, filed as `synthetic`; (2) what the primary documents say (the NTP review, the court rulings, the statutes and agency actions); (3) the popular claims built on top of these; and (4) the institutional record, which in 2025-2026 split sharply in the United States, filed as adoption, never as evidence, in both directions.
Here the usual gap is reversed in places. For decades institutions and the public agreed that fluoridation was safe and the debate was a fringe one; since 2024 the evidence about high exposure has firmed up while the evidence about 0.7 mg/L has not, yet adoption has moved fastest on 0.7 mg/L. Two US states banned fluoridation in 2025, the future US Health Secretary called fluoride an industrial waste associated with IQ loss (Nov 2024), and the FDA cited "possibly decreased IQ" when it began action on children's fluoride drops (May 2025), while the FDA's own scientific evaluation called the US evidence on IQ "equivocal" and the NTP said its findings do not address whether the sole exposure to fluoride added to drinking water has an effect. On the other side, the CDC page and the main dental and paediatric bodies still say fluoridation is safe, a position that rests partly on older dental studies whose benefit the 2024 Cochrane review found much smaller in modern data. Public opinion sits between: in April 2026, 43% of US adults favored fluoridation and 26% opposed it (Annenberg). Stratah records all of this and gives none of it evidential weight. The real open edge is narrow and specific: whether exposure in pregnancy at community levels affects a child's IQ, a question on which four cohorts disagree.
Every point on the timeline links to its source. Open the timeline.
Status: passed with open items · reviewed 2026-10-06. Publish-gate review of PR #15 (head 0f60029). Conformance on the merge with origin/main: 0 errors, 3 warnings for this subject (two established claims on secondary anchors; no taxonomy entry). Sources opened by the reviewer: 26 of the 67 cited in claims.yaml, fetched live and text-searched. The court documents and the NTP monograph were downloaded and their sha256 hashes match the manifest (Ninth Circuit 559ecd6d..., Doc. 445 c6330290..., NTP 4d156ef6...). Wording, dates and numbers that matched: NTP p. ii scope sentence, the moderate-confidence conclusion, 72 studies, 19 low risk of bias (3 cohort, 16 cross-sectional), 18 inverse, the countries, 0.59% (about 1.9 million), "appeared less consistent", "primarily on studies with estimated fluoride exposures higher than", "uncertainty remains", and the state-of-the-science retitling. Doc. 445 (filed 09/24/24): p. 2 "unreasonable risk of reduced IQ in children", "does not conclude with certainty", "does not dictate precisely"; p. 6 4 mg/L / 10 = 0.4 mg/L; paras 121-122 on pp. 79-80. Ninth Circuit No. 25-384 (filed 21 May 2026, not for publication): vacate and remand; "abused its discretion after the first bench trial by commandeering the case"; remand "to rule based solely on the first trial record"; standing remanded. It is a procedural ruling and did not reach the science, as the dig says. The court rulings are stated as what they are and are not. Also matched: the ADA misquote ("with any certainty"); Taylor 2025 Table 2 (all figures cited); Choi 2012 (27 studies, 2 Iran, SMD -0.45, 11.5 mg/L, "in many cases", "cannot be used to derive an exposure limit"); the HSPH release of 25 Jul 2012; EFSA 2025 (>1.5 mg/L); NASEM 2020 ch. 2 and 2021 ch. 2 quotes and FAN as a literature source; FDA 13 May 2025; 90 FR 25329 "not conclusive" (govinfo copy); FDA CDER evaluation "equivocal" (pdf p. 13), "hypothesis-generating" (p. 15); EPA releases of 7 Apr 2025, 22 Jan 2026 and 3 Aug 2026; 91 FR 3722 (citation and date from the FR API; quote from the stored copy); Florida bill history (presented 15 May, approved 20 May 2025, Ch. 2025-22); NSF 2019 (328 samples, arsenic 52%, mean 0.16, max 0.9, "none ... exceeded 1/10th"); CDC 2011 additives (95% FSA, phosphorite); the RFK Jr. post ("industrial waste", "IQ loss"; no arsenic); Green 2019 (512 pairs, 0.69 vs 0.40, boys -4.49); Do 2024 (357, 1.07, margin 3.75); Warren 2025 (57,960, about 7% of an SD higher); Aggeborn & Ohman (0.14 SD bound, 81,776, mean .3533); Simon 2026 abstract. Not opened by the reviewer: the CDC recommendations page (403), Grandjean 2024 Odense (Europe PMC 500), Utah H.B. 81 live (TLS failure; the stored enrolled text has "A person may not add fluoride to water" and effective 7 May 2025), eCFR 141.62 (blocked), the Choi-Grandjean statement repost (403), the Florida chapter-law text, the FDA 31 Oct 2025 release ("4 companies"), and the remaining 41 cited sources. Refuted claims: Rule 9 is met for fl-court-proved-harm, fl-harvard-study-applies-to-us and fl-additives-toxic-waste (each names a target and rests on primary text or test data). The narrative_drift class fits all three. Absence claims are marked and capped (fl-nazi-camps and the searched gaps). Nothing borrows weight from adoption, and threads.yaml sets confers_weight: false. Two blocking findings are in round 1: the NTP's scope is paraphrased more narrowly than the monograph states it (R1-1), and the surface text and fl-water-07-no-lower-iq say residence-based studies at community levels find no lower IQ, while the dig itself holds a 2026 US study (ECHO, exposure assigned by residential address) that reports lower fluid cognition above about 0.675 mg/L, which the assessment never mentions (R1-2). Not added to build/site.yaml. Round 2 (second reviewer, head bb939f4; diff 61ef6d1..bb939f4; edited only this file). Conformance on the merge with origin/main: 0 errors, the same 3 warnings for this subject. Sources opened by the round-2 reviewer: the NTP monograph (downloaded live, sha256 4d156ef6... matches; p. ii and printed p. 84 read), the Simon 2026 ECHO abstract (Europe PMC REST; no open full text), the Florida SB 700 enrolled text (downloaded live, sha256 ff5be8e5... matches; ss. 31-32 on p. 51; Chapter 2025-22 on the bill page) and the stored Utah H.B. 81 enrolled text (live le.utah.gov failed TLS). R1-1, R1-2 and R1-3 are verified fixed and resolved. R1-4 and R1-5 (non-blocking) were not answered and stay open items. The revised surface lines (headline, search summary, assessment, key points, short answer, divergence note) say no more than the claims: "unsettled" / "not established, evidence divided" matches fl-fluoridation-lowers-iq (contested, low), and ECHO is labelled abstract-only everywhere it appears. Adoption still carries no evidential weight (the ADA/AAP note on fl-water-07 is adoption_note only). Log: L-10 added; L-01 to L-09 unchanged (the diff has no removed lines). Two new non-blocking requests (R2-1: ECHO's spline result above 1107 micrograms/L is omitted; R2-2: three places still state the residence-based result without the ECHO exception). Every blocking request is answered and resolved by a reviewer who did not write the work, so the status is passed_with_open_items. review.yaml is owner-only: the verdict is posted on PR #15 and escalated to the owner. Not merged; not added to build/site.yaml. The review record.
refuted: the claim below is false confidence high · our conclusion
A US federal court found that fluoridated water lowers children's IQ.
Anchor. The 2024 opinion itself says the finding "does not conclude with certainty that fluoridated water is injurious to public health" (Doc. 445 p. 2) [read]; it was vacated on 21 May 2026 (9th Cir. No. 25-384) [read].
Would change if. a court issues a finding of causation on a full record
refuted: the claim below is false confidence high · our conclusion
The 2012 "Harvard study" showed that fluoridated tap water in the US lowers children's IQ.
Anchor. Choi 2012: 27 studies, 25 from China and 2 from Iran; exposed-group water up to 11.5 mg/L, "in many cases" above US levels; "Our review cannot be used to derive an exposure limit" [read]. The authors' statement of 11 Sep 2012 says the results do not allow a judgment about risk at exposure levels typical of US fluoridation [read in a repost]. The "Harvard" label comes from the school's press release of 25 Jul 2012 [read].
Would change if. not applicable (the paper's own text)
refuted: the claim below is false confidence moderate · our conclusion
Fluoridation chemicals, as industrial co-products, put harmful levels of contaminants into tap water.
Anchor. CDC engineering fact sheet (2011): most additives come from phosphorite rock used mainly in phosphate fertilizer manufacture; about 95% of fluorosilicic acid comes from captured gases in that process [read, mirror copy]. NSF fact sheet (2019), 2012-2017 samples (328): arsenic detected in 52%, mean 0.16 ppb, maximum 0.9 ppb, against the NSF limit of 1 ppb and the EPA maximum of 10 ppb; "none of the samples exceeded 1/10th the U.S. EPA MCL" [read].
Would change if. independent testing finds contaminants above the NSF/ANSI 60 limits in product as dosed
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence moderate · our conclusion
Children exposed to drinking water with more than 1.5 mg/L of fluoride (the WHO guideline value) score lower on IQ tests than children with less exposure, consistently across many studies, most of them from high-fluoride regions of China, India, Iran and Mexico.
Anchor. NTP Monograph 08 (Aug 2024), Abstract p. xix: "with moderate confidence", water fluoride concentrations exceeding 1.5 mg/L "are consistently associated with lower IQ in children"; 72 IQ studies, 19 low risk of bias, 18 of them inverse; more than 7,000 children in the 19; countries China 10, India 3, Mexico 3, Canada 2, Iran 1 [read]. Taylor et al. 2025, Table 2: water-fluoride dose-response over all data, 31 studies, slope -0.15 (-0.20 to -0.11) [read]. Choi et al. 2012: 27 studies (25 China, 2 Iran), SMD -0.45 (-0.56 to -0.35), exposed-group water up to 11.5 mg/L [read]. EFSA 2025 summary: adverse effects on the developing brain at concentrations above 1.5 mg/L [read, summary only].
Would change if. well-conducted prospective cohorts in high-fluoride areas, measuring lead, arsenic, iodine and parental IQ, find no association
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence high · what a source says
The NTP monograph states that its findings do not address whether the sole exposure to fluoride added to drinking water at 0.7 mg/L, the US and Canadian level, is associated with a measurable effect on IQ; it adds that the moderate-confidence conclusions may also be relevant in optimally fluoridated US areas depending on people's additional exposure from other sources, and that targeted US research is needed. It gives no estimate of IQ points lost.
Anchor. "About This Monograph", p. ii: the findings "do not address whether the sole exposure to fluoride added to drinking water" at 0.7 mg/L in the United States and Canada is associated with a measurable effect on IQ [read]. p. 84, after the 1.9 million figure: the conclusions "may also be relevant to people living in optimally fluoridated areas of the United States depending on the extent of their additional exposures to fluoride from sources other than drinking water", and targeted US research is needed [read, reviewer round 1]. Addendum, p. 102: the rating rests "primarily on studies with estimated fluoride exposures higher than" those of optimally fluoridated US water, and "uncertainty remains" at the lower range [read]. p. 47: below 1.5 mg/L the association "appeared less consistent" [read].
Would change if. a revised NTP edition states a conclusion about 0.7 mg/L
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence high · what a source says
The National Academies reviewed two drafts of the NTP monograph. In 2020 they found the draft's conclusion not adequately supported; in 2021 they found that the added meta-analysis substantially increased support for it, but that no conclusions could be drawn about low exposures such as water fluoridation.
Anchor. NASEM 2020, Summary: "the committee does not find that NTP has adequately supported its conclusion. That finding does not mean that the conclusion is incorrect"; criticisms include inconsistent risk-of-bias ratings, use of an advocacy group's literature list, and no meta-analysis [read]. NASEM 2021 letter: the meta-analysis "substantially increases the support"; NTP should make clear that the monograph "cannot be used to draw any conclusions regarding low fluoride exposure concentrations, including those typically associated with drinking-water fluoridation" [read]. After these reviews the NTP dropped its hazard conclusion and retitled the document a state-of-the-science review (monograph p. xi, pp. 3-5) [read].
Would change if. not applicable (a record of what the reviews say)
established confidence moderate · our conclusion
Five cohorts that compare people by the fluoride level of the community water they grew up on (fluoridated water of about 0.7 to 1.0 mg/L, or in Sweden natural fluoride averaging 0.35 mg/L) find no lower IQ, cognitive or achievement test scores. A sixth, the 2026 ECHO analysis (fl-echo-2026-change-point, abstract only), assigns exposure the same way and reports lower scores above about 0.675 mg/L.
Anchor. New Zealand (Dunedin, born 1972-73): childhood IQ, adjusted difference -0.14 (-3.49 to 3.20), 992 children, of whom 99 unexposed; authors call the exposure measure "an ecological one" [read]. Sweden (Aggeborn & Öhman 2021, natural fluoride, mean 0.353 mg/L): conscription test scores of 81,776 men, no effect; losses larger than 0.14 SD per 1 mg/L ruled out [read, third-party copy]. Australia (Do 2024): 357 young adults, 100% vs 0% early-life fluoridated water, 1.07 (-2.86 to 5.01) IQ points, inside a 3.75-point equivalence margin [read]. US High School and Beyond (Warren 2025): 57,960 adolescents, exposed scored about 7% of a standard deviation higher on achievement tests (math, reading, vocabulary) [read]. US Wisconsin Longitudinal Study (Warren 2026): 10,317 people, only 2 of 45 key coefficients significant, no negative association [read].
Would change if. a residence-based cohort with prenatal exposure and individual measures finds lower scores, or the existing cohorts are shown to misclassify exposure enough to hide an effect
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence high · our conclusion
The studies read cannot rule out a small effect of community-level fluoride on IQ; the largest null study rules out only effects larger than about 0.14 standard deviations per 1 mg/L.
Anchor. Aggeborn & Öhman 2021: negative effects larger than 0.14 SD per 1 mg/L ruled out [read]. Do 2024: equivalence margin of 3.75 IQ points; the estimate fell inside it [read]. Confidence intervals in the other null studies include losses of up to about 3.5 points (Dunedin, -3.49) [read].
Would change if. a much larger study with individual exposure measures narrows the interval
established confidence high · what a source says
On 24 September 2024 a federal court ruled that fluoridation at 0.7 mg/L poses an "unreasonable risk" of reduced IQ under the Toxic Substances Control Act and ordered the EPA to start rulemaking; the same ruling says it does not conclude with certainty that fluoridated water harms public health and does not order fluoridation to stop.
Anchor. Doc. 445, p. 2: fluoridation at 0.7 mg/L "poses an unreasonable risk of reduced IQ in children"; "this finding does not conclude with certainty that fluoridated water is injurious to public health"; "This order does not dictate precisely what that response must be" [read]. p. 6: a margin-of-exposure reasoning in which a hazard level of 4 mg/L divided by 10 gives 0.4 mg/L [read]. paras 121-122, pp. 79-80: unreasonable risk under 15 U.S.C. 2620(b)(4)(B)(ii); EPA to initiate rulemaking under section 6(a) [read].
Would change if. not applicable (a record of what the ruling says); see fl-court-ruling-vacated
established confidence high · what a source says
On 21 May 2026 the Ninth Circuit vacated the 2024 ruling and sent the case back, holding that the district court abused its discretion in how it handled the case after the first trial; it did not rule on the science.
Anchor. 9th Cir. No. 25-384, memorandum filed 21 May 2026 (not for publication): "we vacate the district court's decision and remand"; the court "abused its discretion after the first bench trial by commandeering the case"; remand to rule on the first trial record (the first trial, held before the final NTP monograph); standing returned to the district court [read].
Would change if. the district court rules again on remand, or the vacatur is reheard
established confidence high · what a source says
Utah (signed 27 March 2025, in force 7 May 2025) and Florida (approved 20 May 2025, in force 1 July 2025) each enacted a prohibition on adding fluoride to public water (whether any state did so earlier is not checked: fl-sq-other-state-bans); Florida's law does so without using the word fluoride.
Anchor. Utah H.B. 81, enrolled, section 19-4-111(2): "A person may not add fluoride to water" in a public water system; effective 7 May 2025; signed 27 Mar 2025 per the bill status record [read]. Florida SB 700, Chapter 2025-22, sections 31-32 (p. 51): defines "water quality additive" and bars any additive that does not meet the definition; presented to the Governor 15 May 2025, approved 20 May 2025, effective 1 Jul 2025 [read].
Would change if. not applicable (a record of enacted law)
established confidence high · what a source says
In October 2025 the FDA moved against ingestible fluoride drops and tablets for young children (labelling for children under 3, or not at high risk of decay) and not against fluoridated water; its own scientific evaluation called the evidence on IQ in the US population "equivocal".
Anchor. 13 May 2025 release: "initiating action to remove" ingestible fluoride prescription products for children; studies "suggested an association" with "possibly decreased IQ" [read]. 90 FR 25329 (16 Jun 2025): "These safety findings are not conclusive" [read]. 31 Oct 2025: notices to 4 companies of intended enforcement against labelling for children under 3 or not at high risk; the CDER evaluation, p. 13, calls the evidence "equivocal in the U.S. population"; p. 15, neurocognitive concerns "more hypothesis-generating" [read].
Would change if. the FDA takes a further or different action
established confidence high · what a source says
The EPA is reviewing the fluoride drinking-water standard under the Safe Drinking Water Act; as of October 2026 it has published a plan and a protocol but no conclusions, and the legal maximum remains 4.0 mg/L.
Anchor. EPA release 7 Apr 2025 ("Without prejudging any outcomes") [read]. 91 FR 3722 (28 Jan 2026): the plan "does not contain conclusions regarding harmful human health effects" [read]. Release of 3 Aug 2026 and protocol EPA-822-D-26-001: the assessment will cover neurodevelopment and dental fluorosis; draft not released [read]. 40 CFR 141.62: MCL 4.0 mg/L, as of 1 Sep 2026 [read].
Would change if. the EPA releases a draft toxicity assessment or proposes a new limit
established confidence moderate · what a source says
As fetched on 3 October 2026, the CDC's fluoridation recommendations page still says the CDC continues to recommend community water fluoridation, although the US Health Secretary said in April 2025 that he would tell the CDC to stop.
Anchor. CDC page, last reviewed 15 May 2024, fetched 3 Oct 2026: "CDC continues to recommend community water fluoridation" [read]. AP via PBS (published 8 Apr 2025, reporting remarks of 7 Apr): Kennedy "plans to tell" the CDC to stop recommending fluoridation [read; a report, not the directive].
Would change if. a formal CDC or HHS change of recommendation is found
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence moderate · what a source says
Fluoridating systems in the US mostly run near 0.7 mg/L; about 63% of the US population and 39% of Canadians receive fluoridated water; about 1.9 million Americans drink water with 1.5 mg/L or more of natural fluoride.
Anchor. CDC MMWR 2023 (2016-2021): 0.7 mg/L the most common level; 16.3% of population-weighted months below 0.6 mg/L and 0.01% above 2.0 mg/L [read]. CDC 2022 statistics: 209,135,866 people, 62.8% of the US population [tool]. PHAC 2022: about 14.4 million Canadians, 38.8% [tool]. NTP monograph (April 2020 data): 0.59% of the US population, about 1.9 million, at 1.5 mg/L or more natural fluoride [read]. PHS 2015: 0.7 mg/L replaced the 1962 range of 0.7-1.2 mg/L [read].
Would change if. newer surveillance data
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence moderate · our conclusion
Fluoridation's measured benefit for tooth decay is much smaller in studies after 1975 than in the older studies on which it was founded, and in the 2024 Cochrane review the post-1975 effect is uncertain.
Anchor. Cochrane 2015: 35% reduction in decayed, missing and filled baby teeth; 71% of studies before 1975; over 97% at high risk of bias [abstract]. Cochrane 2024: starting fluoridation, post-1975 studies, difference in change of dmft 0.24 (-0.03 to 0.52), 2 studies, 2,908 children, low certainty, "approximately one-quarter of a tooth" [abstract]. CATFISH (Cumbria, 2024): caries in the birth cohort 17.4% vs 21.4%, adjusted OR 0.74 (0.55 to 0.98); older cohort not significant [abstract].
Would change if. new controlled comparisons in modern populations with fluoride toothpaste
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
established confidence moderate · our conclusion
In Calgary, which stopped fluoridating in 2011, decay in children's baby teeth rose more than in Edmonton, which continued; across all cessation studies the results are mixed.
Anchor. McLaren 2022: in 2018/19, baby-tooth decay prevalence 64.8% in Calgary vs 55.1% in Edmonton; results for permanent teeth inconsistent [read]. Juneau (Meyer 2018): more caries procedures in Medicaid children after cessation in 2007, before-after design without a comparison city [read]. Hung 2026 scoping review of 19 studies: "Findings were heterogeneous" [read]. Calgary's council directed reinstatement on 15 Nov 2021 after a plebiscite (62% in favour) and fluoride returned on 30 Jun 2025 [read].
Would change if. a controlled follow-up after Calgary's 2025 restart
established confidence moderate · our conclusion
The known, accepted cost of fluoride at community levels is dental fluorosis (white marks on teeth, mostly very mild), and it was the reason the US lowered its recommended level to 0.7 mg/L in 2015; severe fluorosis and skeletal effects arise at far higher levels.
Anchor. PHS 2015: 0.7 mg/L gives the "best balance of protection from dental caries while limiting the risk of dental fluorosis"; NHANES 1999-2004: 23% of ages 6-49 with fluorosis, about 2% moderate, under 1% severe [read]. Cochrane 2015 at 0.7 ppm: fluorosis of aesthetic concern about 12%, any fluorosis 40% [abstract]. Wiener 2018, ages 16-17: 29.7% (2001-02) to 61.3% (2011-12) [read]. Neurath 2019, ages 12-15: any fluorosis 65% in 2011-12 [abstract]. NRC 2006: the 4 mg/L goal should be lowered; severe enamel fluorosis about 10% of children at or near 4 mg/L [read].
Would change if. new national fluorosis surveys with reliability checks
Why this confidence. Starts at high ().
These reasons were drafted by the project and have not yet been reviewed.
contested confidence low · our conclusion
Community water fluoridation at the level used in the US and Canada (0.7 mg/L) lowers children's IQ.
Anchor. Taylor 2025, Table 2: water fluoride below 1.5 mg/L, 7 studies, n=2,832, slope 0.05 (-0.36 to 0.45), p=.82; restricted to low risk of bias, 3 studies, -0.32 (-0.91 to 0.26), p=.28, inverse but not significant; urinary fluoride below 1.5 mg/L, 5 studies, -0.08 (-0.15 to -0.003), p=.04; Canada (2 studies) group comparison SMD 0.01 (-0.14 to 0.16) [read]. For the individual cohorts see fl-water-07-no-lower-iq and fl-maternal-urine-signal.
Would change if. a large cohort in a fluoridated population with repeated individual exposure measures reproduces the pregnancy-urine result, or several such cohorts find none
Why this confidence. Starts at moderate ().
These reasons were drafted by the project and have not yet been reviewed.
Disputed by.
contested confidence low · our conclusion
In pregnancy cohorts at community-level exposure, higher fluoride in the mother's urine is associated with lower IQ in her child.
Anchor. Canada, MIREC (Green 2019): 512 mother-child pairs; mean maternal urinary fluoride 0.69 vs 0.40 mg/L in fluoridated vs non-fluoridated cities; boys -4.49 IQ points per 1 mg/L (-8.38 to -0.60), girls 2.40 (-2.53 to 7.33), sex interaction p=.02; fluoride intake, both sexes, -3.66 per 1 mg/day; adjusting for lead and other metals did not substantially change estimates; no maternal IQ data [read]. Mexico, ELEMENT (Bashash 2017): 299 pairs, 211 with IQ data, -2.50 IQ points per 0.5 mg/L creatinine-adjusted maternal urinary fluoride (-4.12 to -0.59); exposure from fluoridated salt [read]. Denmark, Odense (Grandjean 2024): 837 children, 0.08 (-1.14 to 1.30) per doubling, null, no sex difference [read]. Spain, INMA (Ibarluzea 2022): positive association in boys (higher scores with more fluoride) [abstract].
Would change if. a further cohort with repeated urine samples and maternal IQ reproduces or fails to reproduce the association; or the MIREC exposure measurements are shown to be invalid
Why this confidence. Starts at moderate ().
These reasons were drafted by the project and have not yet been reviewed.
Disputed by.
proposed confidence provisional · what a source says
In the Canadian MIREC cohort, infants fed formula made with fluoridated water had lower later IQ scores than those with less fluoride exposure, but the full-scale IQ result lost significance after adjusting for prenatal exposure.
Anchor. Till et al. 2020: 398 children (200 breastfed, 198 formula-fed); per 0.5 mg/L water fluoride, full-scale IQ in formula-fed -4.4 (-8.34 to -0.46); after adjusting for prenatal exposure -3.58 (-7.83 to 0.66); performance IQ lower in both feeding groups; verbal IQ null; authors: results "must be interpreted with caution" [read].
Would change if. a second cohort with recorded formula preparation reproduces or contradicts it
proposed confidence provisional · what a source says
A 2026 analysis of US ECHO cohorts reports that children's fluid cognition begins to fall above a water fluoride level of about 0.675 mg/L.
Anchor. Simon et al. 2026 (abstract): 2,514 children; mean water fluoride 396.9 micrograms/L; above a change point of 675 micrograms/L, -0.67 points per 100 micrograms/L (-0.92 to -0.42) [abstract]. Full text not read.
Would change if. the full text shows a different model or a fragile change-point estimate
proposed confidence provisional · our conclusion
The Nazis added fluoride to concentration-camp water to make prisoners docile.
Anchor. PolitiFact (2014) reports that a US Holocaust Memorial Museum historian knew of no experiments involving fluoride; that the author of a book critical of fluoridation said he never came across any documentation; and that the head of an anti-fluoridation group called the historical evidence extremely weak (interviews of 2011, quoted in 2014) [read]. The earliest written form found is a later reproduction (2000) of a letter attributed to Charles Perkins (1954), which does mention Nazis; whether the original did is unknown (fl-sq-nazi-origin) [read].
Would change if. a wartime document showing fluoride added to camp water is found
proposed confidence provisional · our conclusion
Fluoride calcifies the pineal gland.
Anchor. Luke 2001 measured fluoride in the pineal glands of 11 elderly cadavers: high fluoride that correlated with the gland's calcium (r=0.73) [abstract]. It shows fluoride collects in calcified tissue; it does not show fluoride causes the calcification.
Would change if. a study shows fluoride exposure causing pineal calcification
proposed confidence provisional · what a source says
States with more fluoridation have more ADHD diagnoses.
Anchor. Malin & Till 2015: state-level association; the authors list "an ecological design" as a limitation [read]. Perrott 2018: after adding altitude, the association was no longer statistically significant [abstract].
Would change if. an individual-level study finds the association
searched gap confidence provisional · result of a search
Sub-question: what has happened in Food & Water Watch v. EPA since the Ninth Circuit's remand of 21 May 2026?
Anchor. Searched 3 Oct 2026: GovInfo package files -0 to -16 (latest is the fee order of 5 Jan 2026); a web search for remand coverage found reports of the March 2026 argument and the May vacatur only. CourtListener refused this environment.
Would change if. a new district-court ruling on the first trial record, or a rehearing petition
Next step. Read the N.D. Cal. docket for 3:17-cv-02162 after May 2026 (PACER or GovInfo's package files beyond -16), and the Ninth Circuit docket for 25-384 (rehearing).
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
searched gap confidence provisional · result of a search
Sub-question: did HHS issue a formal directive to the CDC to stop recommending fluoridation, and has the CDC changed its recommendation anywhere other than the page read?
Anchor. Tried 3 Oct 2026: hhs.gov (refused); cdc.gov fluoridation pages (read; no withdrawal found).
Would change if. a dated HHS or CDC document is found
Next step. Read hhs.gov press releases for April 2025 onward and the CDC oral-health pages (hhs.gov refused this environment).
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
searched gap confidence provisional · result of a search
Sub-question: what is Health Canada's current position on fluoride and neurodevelopment after the 2024 NTP review, and what is the basis of its 1.5 mg/L maximum?
Anchor. Tried 3 Oct 2026: canada.ca (connection errors), publications.gc.ca (script gate).
Would change if. the Health Canada guideline technical document or a 2024-2026 statement is read
Next step. Read Health Canada's Guidelines for Canadian Drinking Water Quality technical document on fluoride and any 2024-2026 statement (canada.ca dropped every direct connection; try from a local browser).
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
searched gap confidence provisional · result of a search
Sub-question: what do the JAMA Pediatrics editor's note on the 2019 Canadian study and the two editorials published with the 2025 meta-analysis actually say?
Anchor. jamanetwork.com refused this environment (403); Europe PMC has metadata only.
Would change if. the texts are read
Next step. Read Christakis 2019 (PMID 31424484), Levy 2025 (PMID 39761058) and Lanphear 2025 (PMID 39761028) at jamanetwork.com from a browser.
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
searched gap confidence provisional · result of a search
Sub-question: what do the full Cochrane 2024 review and its certainty tables say about the size of the modern dental benefit and about fluorosis?
Anchor. cochranelibrary.com and the PMC body returned a captcha; abstract read.
Would change if. the body changes the reading of the abstract
Next step. Read CD010856.pub3 in full at cochranelibrary.com from a browser.
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
searched gap confidence provisional · result of a search
Sub-question: have any states other than Utah and Florida enacted bans or restrictions on fluoridation in 2025-2026?
Anchor. NCSL tracker refused (challenge page); one advocacy summary seen but not used.
Would change if. an enacted bill is found
Next step. Read 2025-2026 bill status pages for Louisiana, Kentucky, Tennessee and other states with filed bills.
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
searched gap confidence provisional · result of a search
Sub-question: what is the earliest dated source for the claim that Nazis fluoridated camp water, and does the original Perkins letter (1954) mention Nazis?
Anchor. Chronicling America and loc.gov refused (Cloudflare); only later reproductions found.
Would change if. an original letter or a 1950s newspaper printing is read
Next step. Search 1950s newspapers (Chronicling America, Canadian newspaper archives) for the Perkins letter, and the Lee Foundation's records; one secondary source names an 18 Oct 1956 printing without Nazi wording, unverified and from an unreliable outlet.
Start this step → (signs you in with GitHub; trusted contributors can send it to the dig agent)
Publication status is recorded only for sources someone looked up. A source with no status shown has not been checked for a retraction or correction; that is not the same as clean.
These pages are generated from plain YAML: claims.yaml · sources manifest. If you can show a finding is wrong, challenge it.