Fluoride is one of the most-asked questions in tap water right now, and the conversation has shifted in the past two years. The CDC reports that in 2022, 72.3 percent of the US population on community water systems received fluoridated drinking water — roughly 209 million people. Some of that fluoride is naturally occurring in the source water, and some is added at the treatment plant, mostly in the form of fluorosilicic acid or sodium fluoride.
This article does two things: it summarizes what current US authorities say about fluoride in drinking water, and it explains which home filter mechanisms actually remove it. It does not take a position on whether fluoride should be added to municipal water — that is a policy question communities are working through right now, and the science the policy debate hinges on has genuine open questions that this article covers without editorializing.
What fluoride does and where it comes from
Fluoride is a naturally occurring compound found in soil, water, and many foods. Its concentration in untreated groundwater varies dramatically by region — some US groundwater sources contain less than 0.1 mg/L, others contain more than 4 mg/L. Community water fluoridation, the practice of adjusting the fluoride level at the treatment plant, has been standard in much of the US since the 1940s. The CDC's stated public-health rationale is reduction in dental cavities.
The federal regulatory framework treats high-concentration fluoride as a health concern. The EPA's enforceable Maximum Contaminant Level is 4.0 mg/L. There is also a non-enforceable secondary standard of 2.0 mg/L, described by EPA as a guideline upper boundary for areas with high naturally-occurring fluoride. The WHO guideline value is 1.5 mg/L based on dental fluorosis risk. These thresholds are about upper bounds, not the smaller doses used in community fluoridation.
The 2015 HHS recommendation and the 2024 NTP monograph
Two pieces of recent context are essential to understanding the current conversation.
In April 2015, HHS released a final Public Health Service recommendation setting the optimal fluoride level for community water systems at 0.7 mg/L. The previous recommendation, in place since 1962, was a range of 0.7 to 1.2 mg/L that varied based on local climate. HHS settled on the bottom of the old range as a single national recommendation, citing dental fluorosis risk as the reason for moving to the lower end.
In August 2024, the NIH National Toxicology Program published a monograph on fluoride and neurodevelopmental health effects. The headline conclusion, in NTP's exact language:
"The NTP monograph concluded, with moderate confidence, that higher levels of fluoride exposure, such as drinking water containing more than 1.5 milligrams of fluoride per liter, are associated with lower IQ in children."
NTP also explicitly noted: "It is important to note that 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."
That is the precise wording. "Moderate confidence" is a defined evidentiary tier in the NTP framework — between "low" and "high." The 1.5 mg/L threshold matches the WHO guideline value, and is roughly twice the 0.7 mg/L US community fluoridation level. Whether that gap matters depends on the toxicology framework you bring to it, which is exactly the question the current debate turns on.
The current debate, neutrally framed
Following the 2024 NTP monograph, several US municipalities have voted to discontinue fluoridation; many continue. The science the debate turns on has genuine open questions that the NTP monograph itself acknowledges. Reasonable people read the monograph and arrive at different answers.
This article does not pick a side. The remainder is about a narrower technical question: if you decide you want to remove fluoride from your tap water, which mechanisms actually work?
Which filter mechanisms actually remove fluoride
Fluoride is small, charged, and fully dissolved — that combination makes it harder to remove than chlorine, lead, or PFAS. The mechanisms that work are specific.
Reverse osmosis (RO) membrane. RO physically excludes fluoride along with most other dissolved solids. Independent NSF/ANSI 58 certification testing requires at least 95 percent fluoride reduction for systems carrying that listing. RO is the most reliable consumer-grade option and removes a wide range of other contaminants in the same pass. A countertop unit like AquaTru carries NSF/ANSI 58 certification and does not require under-sink plumbing. An under-sink option like Hydroviv uses different filter media tuned to your specific zip code.
Activated alumina or bone-char carbon. These are specialty media engineered specifically for fluoride. They work via adsorption, are commonly used in fluoride-targeted gravity-filter add-ons, and require periodic replacement based on water volume and starting fluoride concentration.
Gravity systems with fluoride-specific elements. Big Berkey base filters do not remove fluoride; the brand sells optional PF-2 fluoride elements that bolt onto the base filters and add activated alumina chemistry. The combination removes fluoride; the base filter alone does not. This is a common point of confusion when comparing gravity systems.
Distillation. Distilling tap water removes fluoride completely along with virtually everything else. Distillers are slow, energy-intensive, and produce water with no minerals — some people find the taste flat. As a household primary water source, distillation is rarely the right pick, but it is the most thorough option for a single batch.
What does not work for fluoride. Standard activated-carbon pitchers and faucet filters (Brita, PUR Classic, refrigerator filters) do not remove fluoride at any meaningful rate. A handful of carbon-based pitchers list fluoride as a certified-reduction claim — these use specialty composite media beyond standard granular activated carbon, and the certification number is the only reliable way to verify the claim. Boiling water concentrates fluoride rather than removing it, since water evaporates while fluoride stays in the pot.
What a reader can actually do
If you want to know the fluoride level in your current tap water, your annual Consumer Confidence Report lists it. Search for your utility name plus "Consumer Confidence Report" — every public water system is required to publish one annually. The CCR shows total fluoride concentration without distinguishing natural from added fluoride; for the breakdown, contact your utility directly.
From there, the decision is yours:
- If you read the NTP monograph as broadly applicable, an RO or fluoride-targeted gravity system removes essentially all fluoride.
- If you read it as specific to >1.5 mg/L exposures, your decision depends on your local CCR. A reading at or below 0.7 mg/L is well below the NTP threshold and the WHO guideline. A reading above 1.5 mg/L is unusual and more typically reflects naturally-occurring fluoride than community fluoridation.
- If you have private well water, EPA limits do not apply — wells are not regulated federally. A water test will tell you your starting fluoride concentration, which determines which filter mechanism makes sense.
The certification labels to look for are NSF/ANSI 58 for reverse osmosis (which always covers fluoride) and any NSF/ANSI 53 listing that explicitly includes fluoride as a tested claim. A filter without one of those certifications has not been independently verified to reduce fluoride, regardless of marketing copy.
Frequently asked questions
(See structured FAQ at the bottom of this page for full answers.)
