Fluoride in Tap Water: The Evidence, the Debate and Your Options

Many US cities add fluoride to drinking water to help prevent cavities, and some private wells have it naturally, sometimes a lot of it. Whether the amount added to city water does any harm is now being argued over by scientists, regulators and state legislatures.

Short answer: as of October 18, 2026, the federal rules haven’t changed. The EPA’s legal limit for fluoride in public water is 4.0 mg/L, its non-enforceable secondary standard is 2.0 mg/L, and the US Public Health Service still recommends 0.7 mg/L for systems that add it. The EPA is reviewing the science but hasn’t proposed anything yet. The latest Cochrane review finds that fluoridation today may prevent a small number of cavities in children, on low-certainty evidence, and a federal review found, with moderate confidence, that fluoride above 1.5 mg/L is linked to lower IQ in children. Whether 0.7 mg/L affects IQ is unsettled. If you want less fluoride, find out your level first; a reverse osmosis system certified for fluoride reduction is the main home option.

How much fluoride is in your water?

Fluoride occurs naturally in groundwater, and many US cities add more to reach a target level. On city water, your utility’s yearly water quality report lists the fluoride it measured. The CDC’s My Water’s Fluoride tool lists many systems too, but it says itself that “the best source of information on fluoride levels in community water systems is your local water provider,” and states take part voluntarily.

On a private well, the CDC’s fluoridation FAQ puts it plainly: “The fluoride content of well water can be determined only through laboratory analysis.” Your local health department can tell you where to test, and our well water testing guide covers how to plan one. High natural fluoride is uncommon but real. A US Geological Survey assessment of wells sampled from 1988 to 2017 found more than 85% of drinking-water wells below 0.7 mg/L, but 0.6% of the domestic wells above 4 mg/L. It estimated 172,000 people are served by domestic wells above the 4 mg/L limit and another 522,000 by wells above 2.0 mg/L. Those are estimates from existing well data, so use them as a rough guide, not a forecast for your well.

A quick note on units: 1 mg/L is the same as 1 part per million (ppm).

The rules as of October 2026

There are three numbers to know, and they do different jobs:

Number	What it is	Enforceable?
4.0 mg/L	EPA maximum contaminant level (MCL) and health goal (MCLG)	Yes, for community water systems
2.0 mg/L	EPA secondary standard	No, but it triggers a public notice
0.7 mg/L	Public Health Service recommendation for systems that add fluoride	No, it’s a target

• The legal limit. Federal rules (40 CFR 141) set both the MCL and MCLG for fluoride at 4.0 mg/L. The EPA’s August 2026 assessment protocol explains that this limit, set in 1986, is meant “to protect against crippling (stage III) skeletal fluorosis,” a bone disease from long-term high exposure.
• The secondary standard. The protocol says the 2.0 mg/L level is there “to protect against cosmetically objectionable (i.e., moderate to severe) dental fluorosis in children,” and it isn’t enforceable. But the EPA requires community systems above it to notify customers within 12 months. The CDC’s FAQ adds that the EPA recommends children 8 and younger get water from another source if their main water has more than 2.0 mg/L.
• The fluoridation target. In 2015 the Public Health Service lowered its recommendation to 0.7 mg/L, which it called “the best balance of protection from dental caries while limiting the risk of dental fluorosis.” The CDC’s recommendations page still says “CDC continues to recommend community water fluoridation” at that level. We found no federal notice withdrawing it.

What’s changing

• The EPA review. In April 2025 the EPA committed to a faster review of fluoride under the Safe Drinking Water Act. In January 2026 it released a preliminary assessment plan, noting that on the normal schedule the next full review “would not be due until 2030.” On August 3, 2026, it published a protocol for a new health assessment that will look at effects including lower IQ in children and dental fluorosis. A draft “will be available for public comment,” and only the final version “will inform potential revisions” to the standard. No date is given for the draft. The protocol also notes that in an earlier review the EPA worked out a possible revised health goal of 0.9 mg/L but didn’t change the rule. For now, nothing has been proposed and the 4.0 mg/L limit stands.
• The lawsuit. In a long-running case brought under a different law (the Toxic Substances Control Act), a federal judge in California ordered the EPA to act on fluoride in 2024. On May 21, 2026, the Ninth Circuit vacated that ruling and sent it back to the district court to decide on the first trial’s record only.
• States. Utah’s H.B. 81 says “a person may not add fluoride to water” in a public water system, effective May 7, 2025. Florida’s 2025 law, in effect since July 1, 2025, bans additives in public water unless they are used to meet drinking water standards, reduce contaminants or improve water quality; it doesn’t mention fluoride by name, and preventing tooth decay isn’t among the listed purposes. Wherever you live, your utility’s report shows what your water has now.
• Fluoride supplements. The FDA now recommends that prescription fluoride drops and tablets “be limited to children aged three years and older who are at high risk for tooth decay,” and in October 2025 it warned four companies selling them for younger or lower-risk children. That’s about supplements, not tap water, but it’s part of the same debate. If your child was prescribed them, ask the prescriber.

Does fluoride in water prevent cavities?

Evidence rating: weak for a benefit today, and the benefit it suggests is small. The latest Cochrane review finds a smaller effect than older studies from before fluoride toothpaste was widespread, and none of the studies behind it are randomized.

The 2024 Cochrane review included 157 studies, and “all used non-randomised designs.” For studies done after 1975 (just two, with 2,908 children), it found that starting fluoridation may lead to a slightly bigger drop in decayed, missing or filled baby teeth: “approximately one-quarter of a tooth in favour of CWF,” an estimate that “includes the possibility of benefit and no benefit.” The certainty was low. The authors conclude that contemporary studies show benefit “but with smaller effect sizes than pre-1975 studies.” They found “insufficient evidence” on what happens when a city stops fluoridating, and no eligible studies on cavities in adults. One co-author co-directs a dental health unit funded by Colgate-Palmolive; the review says that co-author didn’t take part in assessing the included study they co-wrote.

Is fluoride in water harmful?

Teeth: dental fluorosis

Too much fluoride while teeth are forming can cause dental fluorosis, faint white streaks or spots on the enamel. The CDC’s fluorosis page says that “in the U.S., dental fluorosis is mostly mild and cosmetic,” that moderate and severe forms “are rare,” and that permanent tooth enamel is fully formed by about age 8. Cochrane’s estimate, from older studies and also low certainty, is that at 0.7 ppm about 12% of participants have fluorosis “of aesthetic concern” and about 40% have some fluorosis of any level.

Infant formula is where this matters most. The CDC’s FAQ says powdered or concentrated formula mixed with fluoridated water “may increase the chance of dental fluorosis if it is a child’s main source of food,” that ready-to-feed formula “contains little fluoride and does not cause dental fluorosis,” and that parents “can use low-fluoride bottled water some of the time” to mix formula. Ask your pediatrician if you’re unsure.

Bones: skeletal fluorosis

This is the effect the 4.0 mg/L limit is built around, and it takes high, long-term exposure. A World Health Organization background document says skeletal fluorosis “may be observed when drinking-water contains 3–6 mg of fluoride per litre,” and that the crippling form “usually develops only where drinking-water contains over 10 mg of fluoride per litre.” That’s a concern for a few high-fluoride wells, not for fluoridated city water. The WHO’s own guideline value is 1.5 mg/L.

Children’s IQ: the center of the debate

Evidence rating: moderate that levels above 1.5 mg/L are linked to lower IQ; weak at 0.7 mg/L, where the studies conflict. All the human studies are observational, so they show associations, not cause.

• The federal review. In August 2024 the National Toxicology Program concluded, “with moderate confidence,” that higher fluoride exposures, “such as drinking water containing more than 1.5 milligrams of fluoride per liter, are associated with lower IQ in children.” It also said “there were insufficient data to determine if the low fluoride level of 0.7 mg/L currently recommended for U.S. community water supplies has a negative effect on children’s IQ.” Its evidence came mainly from studies in countries such as Canada, China, India, Iran, Pakistan and Mexico, and it stresses that an association “does not prove a cause and effect.”
• The meta-analysis. The same team’s 2025 review in JAMA Pediatrics pooled 74 studies, 45 of them from China; 52 were rated high risk of bias. Higher fluoride was linked to lower IQ overall, but by drinking water level “the association was null at less than 1.5 mg/L,” although it stayed inverse below 1.5 mg/L in the studies at low risk of bias. In studies with individual urine measurements, IQ was 1.63 points lower per 1 mg/L more fluoride in urine. The authors note limited data and uncertainty below 1.5 mg/L. Other researchers, including a member of the American Dental Association’s National Fluoridation Advisory Committee, challenged its methods in letters to the journal, and the authors defended them in a published reply.
• Studies near 0.7 mg/L in North America and Australia point different ways:
  • A Canadian pregnancy study of 512 mother-child pairs (2019) linked each 1 mg/L more fluoride in mothers’ urine to 4.49 points lower IQ in boys, but found no significant association in girls. One author served as an unpaid expert witness in the federal fluoride lawsuit.
  • A US study of 2,514 children in 23 states (2026), with prenatal fluoride exposure estimated from public water records by census tract, found fluid-cognition scores 0.67 points lower for each extra 0.1 mg/L above a best-fit change point of about 0.68 mg/L; a second model found significant drops only above about 1.1 mg/L. It reported an association for one part of cognition, measured with a short test battery, not full IQ.
  • A nationally representative US cohort (2025) found that children exposed to recommended fluoride levels had “modestly better cognition in secondary school,” an advantage that faded by about age 60.
  • A Wisconsin study (2026) found “no evidence” that fluoridation was negatively associated with teenage IQ or adult cognition. It shares lead authors with the study above.
  • In Australia, a study of 357 young people (2025) found full-scale IQ “equivalent” in those exposed and not exposed to fluoridated water as young children.

What this means for you. At the levels found in some high-fluoride wells, the evidence (moderate confidence, per the federal review) is strong enough to act on for children. At 0.7 mg/L, careful studies disagree, and nobody can tell you today whether it affects a particular child. The EPA’s new assessment is meant to weigh exactly this. If you’re pregnant or have young children and want to limit fluoride while the science settles, the options below are reasonable; talk to your dentist or pediatrician about fluoride for your children’s teeth.

How to reduce fluoride at home, if you want to

• Know your number first. Check your utility’s report or test your well. Many readers on city water will see a level around 0.7 mg/L; on a well, it could be far lower or higher.
• Reverse osmosis. The Virginia Department of Health says “reverse osmosis systems and filters containing activated aluminum can remove fluoride” and suggests systems “meeting NSF/ANSI 58 and certified for fluoride reduction.” Fluoride is a separate claim on each listing, so check that it’s there. Our reverse osmosis comparison shows which models are certified for it, and how reverse osmosis works covers wastewater and maintenance.
• Other methods. The Minnesota Department of Health lists activated alumina and distillation as options that remove fluoride.
• Don’t assume a carbon filter reduces it. Minnesota’s health department says only some carbon filters are certified to remove fluoride, so it’s a claim to check model by model; see what carbon filters remove. Look for fluoride on the certification listing, not just the box. Our pitcher comparison found only one pitcher certified for it.
• On a well above 2 mg/L, Wisconsin’s health department advises that children 8 and younger use a different water source, and that everyone does above 4 mg/L.

The debate above is about fluoride in drinking water. The CDC says “fluoridated water and fluoride toothpaste work together to prevent cavities”; ask your dentist what’s right for your family’s teeth.

The bottom line

• The federal rules haven’t changed. 4.0 mg/L is the legal limit, 2.0 mg/L a notice level and 0.7 mg/L the fluoridation target, as of October 18, 2026. The EPA review has no draft yet.
• The cavity benefit today looks small, and the evidence for it is low certainty and non-randomized.
• Above 1.5 mg/L, the IQ concern is credible. At 0.7 mg/L, the evidence is mixed and weak.
• Check your level before you buy anything. If you want less fluoride, reverse osmosis certified for fluoride reduction is the main home option; don’t count on a carbon filter unless it’s certified for fluoride.
• On a high-fluoride well, act, especially for young children.