Most carbon and pitcher filters aren’t designed to take fluoride out of water. The ones that do rely on reverse osmosis, activated alumina, distillation, or filter media made specifically for fluoride. And a trustworthy test report shows more than one impressive percentage. It shows how the reduction fades as the filter wears out.
If you’re reading this because you’re using a filter at home and want to know what it’s doing for your family’s water, that’s the short version. The rest of this page helps you work out which kind of filter does what, what a real test looks like, and how to find your own water’s number before you spend anything.
Which filters reduce fluoride?
Fluoride is dissolved in the water, which is why the type of filter matters far more than the brand on the box. Levels are usually written as milligrams per liter (mg/L), about the same as parts per million.
The labels on filters trip up a lot of shoppers, so it helps to know what each one means.
Certified filters are tested against published standards, and the number tells you what the filter was tested for:
- NSF/ANSI 42 covers things you can taste or see, like chlorine and particles;
- NSF/ANSI 53 covers filters that reduce health-related contaminants;
- NSF/ANSI 58 covers reverse osmosis systems.
- NSF/ANSI/CAN 401 covers a group of newer concerns, including microplastics.
- NSF/ANSI/CAN 372 is a materials standard, which means it says the filter’s own parts don’t add lead, not that the filter removes anything.
A test report is a different thing again, a lab’s published results for one product, with no certification attached.
Here’s how the common options stack up.
Table 1. Which filters reduce fluoride
| Filter type | Reduces fluoride? | Depends on the filter? | Evidence |
|---|---|---|---|
| Standard carbon: granular activated carbon, carbon block, or pitcher | Generally no | Yes. A few media-specific carbon products carry an NSF/ANSI 53 listing that names fluoride, so check the listing. | Minnesota Department of Health notes that carbon removal depends on the filter media; WQA and WHO list other fluoride treatments. |
| Reverse osmosis | Yes | Look for an NSF/ANSI 58 listing that names fluoride. | CDC and WQA list reverse osmosis among fluoride treatments. |
| Activated alumina | Yes | A listed system is stronger evidence than a general statement. | WHO says 1 mg/L should be achievable using activated alumina. |
| Strong-base anion exchange | Resin-dependent | Yes. Check for an NSF/ANSI 53 or 58 listing that names fluoride. | WQA lists strong-base anion exchange as a fluoride treatment. |
| Distillation | Yes | Look for a distillation system with a published testing claim. | CDC and WQA list distillation. |
| BOROUX Foundation filter | No. Not tested or claimed for fluoride. | Its certifications are not fluoride standards. | WQA certified to NSF Standards: NSF/ANSI 42, NSF/ANSI/CAN 372 and NSF/ANSI/CAN 401. |
| BOROUX ProActive filter, used with the Foundation | Yes, by test report | Not certified. A fluoride-specific add-on tested at about 0.7 mg/L, tap-water strength. | Test report: 99.9% for the first 200 gallons per pair, falling to 47.4% at 1,000 gallons per pair (Table 2). |
Two of those rows are our filters, and we use them as worked examples because we publish both kinds of evidence: our test reports and our WQA certifications. The habit worth keeping isn’t memorizing the table, though. It’s looking for the exact standard and the exact contaminant on the listing, or for a test report that shows its numbers. So why does ordinary carbon fall short?
Why don’t most carbon filters remove fluoride?
Carbon is very good at some jobs and just isn’t built for this one. Fluoride in tap water is always present as a dissolved ion, whether it comes from natural minerals or from intentional fluoridation, according to the World Health Organization. Carbon holds on to chlorine and many organic chemicals, the things that affect taste and smell, but it doesn’t hold dissolved fluoride well, which is why the pitcher in your fridge usually leaves fluoride in the water. If you’d like the longer story, here’s how activated carbon works on the things it does handle.
One caveat: “carbon filter” isn’t a single material, and a few products are built differently and carry a fluoride listing under NSF/ANSI 53, so the fair sentence isn’t “no carbon filter removes fluoride” but “most ordinary carbon filters don’t, so check the listing.”
So if the Foundation’s answer on fluoride is no, what does a filter that does reduce it actually show?
What does a real test look like?
A number on a box tells you very little until you know how much water went through, how strong the fluoride was, and what happened near the end of the filter’s life, which is why a useful fluoride test report is a curve rather than a single figure. Our ProActive test report, dated January 7, 2026, is built that way, and we publish every result in it so you can check the numbers. The lab adds fluoride to water, runs it through the filter, and measures what’s left, and the report followed one filter to 500 gallons, which counts as 1,000 gallons per pair.
We tested at tap-water strength, about 0.7 mg/L (the report’s water ran 0.65 to 0.76 mg/L), which is the CDC’s recommended level, covered below. The fluoride method in NSF/ANSI 53 uses 8 mg/L instead. We went lower “so customers could better predict filter life in their home environment,” so the curve below shows how the filter performs in a typical home. If your well water runs much higher than 0.7 mg/L, the curve may differ, so it’s worth testing your water first (more on that below).
These are lab results, not an NSF/ANSI 53 certification. Table 2 is the whole report, row for row.
Table 2. ProActive test report, every row (fluoride, single filter tested)
| Volume tested, per filter (as printed) | Gallons per pair | Influent (mg/L) | Effluent (mg/L) | Reduction |
|---|---|---|---|---|
| 10 UV | not stated | 0.74 | <0.001 | 99.9% |
| 20 gal | 40 | 0.70 | <0.001 | 99.9% |
| 40 gal | 80 | 0.70 | <0.001 | 99.9% |
| 60 gal | 120 | 0.69 | <0.001 | 99.9% |
| 80 gal | 160 | 0.69 | <0.001 | 99.9% |
| 100 gal | 200 | 0.71 | <0.001 | 99.9% |
| 120 gal | 240 | 0.65 | 0.002 | 99.7% |
| 140 gal | 280 | 0.68 | 0.006 | 99.1% |
| 160 gal | 320 | 0.67 | 0.050 | 92.5% |
| 180 gal | 360 | 0.73 | 0.077 | 89.5% |
| 200 gal | 400 | 0.71 | 0.072 | 89.9% |
| 220 gal | 440 | 0.71 | 0.108 | 84.8% |
| 240 gal | 480 | 0.73 | 0.129 | 82.3% |
| 260 gal | 520 | 0.70 | 0.116 | 83.4% |
| 280 gal | 560 | 0.74 | 0.203 | 72.6% |
| 300 gal | 600 | 0.72 | 0.181 | 74.9% |
| 320 gal | 640 | 0.73 | 0.192 | 73.7% |
| 340 gal | 680 | 0.73 | 0.258 | 64.7% |
| 360 gal | 720 | 0.76 | 0.298 | 60.8% |
| 380 gal | 760 | 0.71 | 0.222 | 68.7% |
| 400 gal | 800 | 0.70 | 0.325 | 53.6% |
| 420 gal | 840 | 0.73 | 0.263 | 64.0% |
| 440 gal | 880 | 0.72 | 0.367 | 49.0% |
| 460 gal | 920 | 0.72 | 0.342 | 52.5% |
| 480 gal | 960 | 0.71 | 0.351 | 50.6% |
| 500 gal | 1,000 | 0.69 | 0.363 | 47.4% |
The per-pair column doubles the tested volume, as the report explains for a two-filter set. The report prints its first volume as “10 UV” without defining the unit, so that row has no per-pair value here. Influent and effluent are as printed, in mg/L, and the report’s cover reads “Version 1.0.”
Reading down the curve, reduction was 99.9% through the first 200 gallons per pair, still 99.1% at 280 gallons per pair, and 47.4% at 1,000 gallons per pair, and our product page states a 78.9% average up to 1,000 gallons. So the filter does the most at the start and less as it wears out, which is why we recommend replacing a pair at 1,000 gallons, about every 6 to 12 months depending on use. That’s the last row of the table, 47.4%, and now you have the full curve in front of you.
Why is fluoride a separate filter?
It’s the question we hear most often, usually when someone sees a second filter in the cart, and the short answer is that fluoride needs different media, not everyone wants it removed, and it wears out on a different schedule.
The first reason is the media: our Foundation is a carbon block tested for a long list of contaminants, and its published reports contain no fluoride test, which is why our FAQ answers the question directly: “Does the BOROUX Foundation Filter reduce fluoride? No.” The ProActive is a dual-layer nano-carbon filter that we built and tested for fluoride.
The second is that not everyone has fluoridated water or wants it gone. In 2022, 62.8% of the US population received fluoridated community water, according to the CDC, which leaves roughly a third of Americans without it, and some people on fluoridated systems want the dental benefit left in. A separate filter lets a household decide for itself.
The third is that the two wear out on different clocks. We recommend replacing a pair of Foundation filters every 12 months, and our filter-life guide gives tested gallons by contaminant. A ProActive pair is replaced at 1,000 gallons, which a heavy-use household could reach in about six months, and a combined filter would make you swap the long-life part on the short-life schedule.
The design does have a cost: the add-on is a second purchase with its own replacement schedule, and only households that want fluoride reduced need it. If you want one unit that handles fluoride by itself, reverse osmosis with a fluoride listing does it in one (see Table 1). Our position is, “We believe in giving people options. Some people don’t have fluoride in their water, or don’t want to remove it.” And for anyone starting from a gravity filter, the practical question is what fits.
Does it fit a Berkey?
If you already own a Berkey-style gravity system, yes. Our ProActive filters are compatible with Berkey systems, and they attach to compatible upper-chamber filters, including Black Berkey® Elements, and sit in the lower chamber.
Berkey does offer a fluoride filter option (the PF-2 Elements), but the main takeaway is that they provide no public test report or independent certification to back up their claims. While their product page states the filters reduce fluoride, there is no lab test data or standard certification cited to verify how well they perform or how long that reduction actually lasts. That’s why checking for published test results is so important before choosing a filter.
How much fluoride is in your water?
Before you decide anything, get your own number, which is usually easy. On city water, your utility’s annual water-quality report lists the fluoride level it measured, and the CDC’s My Water’s Fluoride tool shows it too, for participating states. On a private well there’s no utility report, so use a certified lab test. We have a guide to testing your own water, and our well-water article covers well owners.
Once you have that number, the next question is what it means, and what the health agencies say about it.
What do health agencies say about fluoride at US levels?
Three numbers do most of the work here. In mg/L, the CDC recommends 0.7 for community water systems that fluoridate, the World Health Organization’s drinking-water guideline value is 1.5 (based on the risk of dental fluorosis above that level and of skeletal fluorosis at progressively higher ones), and the EPA’s enforceable maximum is 4.0, set in 1986 and reviewed in 2024.
| Level | Source and meaning |
|---|---|
| 0.7 mg/L | CDC’s recommended level for US community water fluoridation |
| 1.5 mg/L | WHO drinking-water guideline value, based on the risk of dental and skeletal fluorosis; also the level above which the National Toxicology Program (NTP) had moderate confidence of an IQ association |
| 4.0 mg/L | EPA’s enforceable maximum contaminant level (MCL) and its goal (MCLG), set in 1986 and reviewed in 2024 |
The IQ question is the one parents scroll for, so it deserves exact wording. The National Toxicology Program monograph, published August 21, 2024, found “moderate confidence” that drinking water with more than 1.5 mg/L of fluoride is associated with lower IQ in children, and it said the data are “insufficient” to judge the 0.7 mg/L level. A JAMA Pediatrics meta-analysis published January 6, 2025 reported an average IQ decrease of 1.63 points per 1 mg/L increase in urinary fluoride, which reflects what’s in the body, not what’s in the tap, and the association was null below 1.5 mg/L of fluoride in water. Of the 74 studies, 45 came from China, none came from the US, and many were rated at high risk of bias.
On teeth, CDC describes dental fluorosis as a condition of the tooth-forming years, so only children aged 8 and younger are at risk, and most US cases are very mild to mild. The EPA’s drinking water regulations list “bone disease (pain and tenderness of the bones)” and “children may get mottled teeth” as potential effects of long-term exposure above the 4.0 mg/L MCL.
There’s a benefit side too, since CDC continues to recommend 0.7 mg/L community fluoridation for cavity prevention, and a 2024 Cochrane review found that in studies of programs begun after 1975, fluoridated water was associated with 0.24 fewer decayed teeth on average, a difference whose 95% confidence interval (−0.03 to 0.52) includes zero, compared with 2.10 fewer in earlier studies, likely reflecting how common fluoride toothpaste has become. The rules and recommendations behind all of this have been moving lately, so the dates matter.
What changed in 2025 and 2026?
Fluoride policy shifted in 2025 and is still under review. Here’s where things stood as of September 30, 2026.
- Utah ended community water fluoridation with H.B. 81, effective May 7, 2025.
- Florida followed with S.B. 700, effective July 1, 2025. The statute bars public water systems from using additives other than water-quality additives and doesn’t name fluoride directly.
- No additional statewide bans had been enacted in 2026 as of CareQuest’s September 25, 2026 tracker, though bills were introduced in several states.
- EPA opened an expedited review under the Safe Drinking Water Act, released a preliminary assessment plan in January 2026, and published its toxicity-assessment protocol on August 3, 2026. The MCL remained 4.0 mg/L.
- On May 21, 2026, the Ninth Circuit vacated a September 2024 district-court order requiring EPA to regulate fluoride under the Toxic Substances Control Act, without ruling on whether fluoride at 0.7 mg/L is safe.
- In April 2025, CNN reported that HHS planned to direct CDC to stop recommending fluoridation. CDC’s published recommendation was unchanged as of September 30, 2026.
For most families, the practical takeaway is to keep an eye on your own utility’s report. If you do decide on an add-on, the day-to-day part is simple.
Using a fluoride add-on: taste, priming, replacement
Nothing here is complicated, but it’s easier to know before you fill the first bottle. A subtle earthy or mineral-like taste is normal at first, until the media is fully flushed or primed, and our priming guide walks you through it. Replacement is tied to use, so we recommend replacing a pair at 1,000 gallons, about 6 to 12 months depending on use, and one ProActive attaches to each Foundation filter.
If you’ve decided
Start with your own water number: if it’s low and you want to keep the fluoride, our Foundation filters are the base system and aren’t claimed for fluoride. If your water report shows fluoride you want less of, our ProActive filter is the answer we’d point you to, installed in the lower chamber under a Foundation filter. If you’re still deciding, read the test report linked above first, because the full curve is there.