Two numbers, two different kinds of number
Most fluoride coverage quotes one limit. There are two, and the difference between them is the whole of this page.
| Figure | Value | Type | What EPA lists it against |
|---|---|---|---|
| Maximum Contaminant Level | 4.0 mg/L | Enforceable | Bone disease (pain and tenderness of the bones); children may get mottled teeth |
| Secondary MCL | 2.0 mg/L | Not enforced | Tooth discoloration |
The 4.0 mg/L figure is a health-based limit. Exceed it and the system is in violation.
The 2.0 mg/L figure is a Secondary Maximum Contaminant Level. EPA does not enforce secondary standards — they are “guidelines established for aesthetic considerations.”
So the cosmetic guideline sits at exactly half the enforceable limit. Water can be fully compliant with the health standard while sitting well above the level at which EPA notes tooth discoloration, and nothing in a compliance statement will flag that.
That is the same structure as lead, where the health goal is zero and the action level is not, and the same reading problem the annual water report creates: compliance is measured against one number, and the other one goes unmentioned.
Where it comes from, which the number does not tell you
EPA lists three sources in the same row: “Water additive which promotes strong teeth; erosion of natural deposits; discharge from fertilizer and aluminum factories.”
A concentration cannot distinguish them. Naturally occurring fluoride can exceed both federal figures with nothing added at all, which is common in parts of the country with particular geology — and it is why a high reading is not evidence about a fluoridation policy.
Your annual report is where the local answer lives. Fluoride is a regulated contaminant, so if it was detected it will be listed.
One number this page does not give you
The US Public Health Service publishes a recommended concentration for community water fluoridation. It is the figure most people actually want, and it is not stated here.
We could not retrieve it from a primary source: cdc.gov returns an error to automated requests and is blocked from the browser tooling on this machine, and the HHS landing page carries the title without the number. The value appears readily in search summaries, which is exactly the kind of unattributable repetition this site refuses.
It is logged as an open gap and will appear here when it can be quoted from source. That is a deliberate omission, not an oversight.
If you want to reduce it
Start from what fluoride is: a dissolved ion. That single fact eliminates most of the filter aisle.
- Sediment filters and micron ratings — no effect. Pore size stops particles, and fluoride is in solution. → What micron ratings can and cannot reach
- Standard activated carbon — not a fluoride treatment. Carbon addresses taste, odour, chlorine and some organics.
- Reverse osmosis — the usual answer, with a condition attached.
That condition is the one this site keeps returning to. Under NSF/ANSI 58, TDS reduction is the only required claim. Fluoride reduction is optional.
So “certified to NSF/ANSI 58” is not evidence a system reduces fluoride. Check whether that specific model lists the fluoride claim.
→ What each NSF standard actually certifies
The decision path
| Your figure | What it means | What to do |
|---|---|---|
| Above 4.0 mg/L | System in violation of the enforceable limit | Read the required health language and restoration actions in your report |
| Between 2.0 and 4.0 mg/L | Compliant, above the cosmetic guideline | A legitimate reason to reduce it if you choose; nothing obliges the utility |
| Below 2.0 mg/L | Below both federal figures | No federal basis for action |
| Not listed in your report | Not detected, or not reported | Ask the utility; fluoride is regulated so it should appear if detected |
| Private well | No monitoring and no fluoridation | Test for it; natural fluoride can be high in some geology |
Choose reverse osmosis if
- Your figure sits above the 2.0 mg/L cosmetic guideline and you want it lower.
- The specific model lists a fluoride reduction claim, not just NSF/ANSI 58 certification.
- You are treating drinking and cooking water rather than the whole house.
Choose to test first if
- You are on a private well. Nobody monitors it, and natural fluoride varies with geology.
- Your report does not list fluoride and you want to know rather than assume.
Choose to do nothing if
- Your figure is below both federal numbers.
- You were about to buy a carbon filter for fluoride, in which case do nothing yet and buy the right technology instead.
Mistakes that cost the most money
Quoting one limit as “the” limit. There are two, they differ by a factor of two, and only one is enforceable.
Buying a carbon filter or pitcher for fluoride. Fluoride is dissolved. Standard carbon does not address it.
Assuming NSF/ANSI 58 covers fluoride. TDS reduction is the only required claim under it; fluoride is optional and model-specific.
Reading “meets federal standards” as “below 2.0”. Compliance is measured against 4.0. The cosmetic guideline is not enforced and will not be flagged.
Treating a high natural reading as evidence about fluoridation. EPA lists erosion of natural deposits as a source. Geology and policy are different things.
Testing the whole house. If the concern is ingestion, the drinking tap is the exposure route and the cheaper place to treat.
Three situations, and what changes
Municipal supply, figure listed below 2.0 mg/L. Both federal numbers are satisfied. There is no technical case for treatment, whatever your view of fluoridation.
Municipal supply, figure between 2.0 and 4.0. Compliant, above the cosmetic guideline. This is the band where reverse osmosis with a fluoride claim is a reasonable purchase, and where nothing will be done for you.
Private well. No monitoring, no fluoridation, and natural fluoride entirely dependent on local geology. Testing is the only route to a number, as it is with arsenic.
The objections worth answering
“Is fluoride in water safe?” EPA regulates it at 4.0 mg/L on the basis of bone disease and mottled teeth, and flags tooth discoloration at 2.0. This page reports those limits and where your figure sits against them. It does not adjudicate the wider debate, and anyone claiming a citation settles that debate is overreading their source.
“My filter jug says it improves water quality.” Against what claim? Carbon pitchers typically carry taste, odour and chlorine claims. Fluoride is a separate, optional claim and most do not hold it.
“Can I just boil it off?” No. Boiling removes water as steam and leaves dissolved fluoride behind, concentrating it — the same error people make with nitrate.
Who this is for
Anyone who has found a fluoride figure in their annual water report and wants to know which limit it is being judged against, and what would actually reduce it if they choose to.
What to check before buying anything
Which of the two numbers is my figure above, and does this specific model list a fluoride reduction claim?
The first is in your water report. The second is in NSF’s listing, not on the box.
→ Reading your annual water quality report → What each NSF standard covers, and why claims vary by product → City water treatment: the three gaps a utility cannot close → Why a fluoride claim exists only under NSF/ANSI 58 → The 127 filters certified for fluoride, and why they are all one format → The one secondary standard that carries a notice duty