The usual finding, running backwards

Most of this site reports the same shape of problem: a contaminant EPA regulates, and no certified product to match it. Thirty-three regulated entries have no NSF claim at all.

NSF/ANSI 401 is the mirror image.

Fifteen named contaminants. Not one of them has a federal drinking water limit. NSF certifies against every one.

The fifteen, and how many hold each

NSF’s full title for the standard is Drinking Water Treatment Units — Emerging Compounds/Incidental Contaminants.

Claim Rows Distinct models
Bisphenol A 608 411
Estrone 608 411
Nonylphenol 608 411
Naproxen 596 400
Phenytoin 596 400
Ibuprofen 498 361
Linuron 400 253
Trimethoprim 399 253
Atenolol 272 183
Carbamazepine 269 180
DEET 250 173
Metolachlor 235 161
Meprobamate 231 159
TCEP 197 133
TCPP 195 132

Every one of those fifteen extractions matched NSF’s own reported count exactly.

Seven of them are medicines

Atenolol, carbamazepine, ibuprofen, meprobamate, naproxen, phenytoin and trimethoprim.

438 distinct models hold at least one of those seven claims. 112 hold all seven.

For scale, PFAS has 72 certified models and nitrate has 84. The certified field for medicines nobody regulates is several times larger than either. → The 72 filters certified for PFAS

A correction to an earlier page

This site previously stated, on its unregulated contaminants page, that pharmaceuticals have no certified claim.

That was wrong. Seven of the fifteen named NSF/ANSI 401 claims are for medicines, and the register already recorded that sixteen NSF/ANSI 401 claims are selectable. The page prose contradicted the site’s own claim register.

It has been corrected, and this page exists partly because of it. → Unregulated contaminants: the candidate list and the monitoring rule

The claims travel in packages, with three exceptions

Bisphenol A, estrone and nonylphenol are held by the identical 411 models. Naproxen and phenytoin by the identical 400.

Order all fifteen by size and each set sits inside the next larger at eleven of fourteen steps. Then it breaks three times: linuron is not inside ibuprofen, carbamazepine is not inside atenolol, and TCEP is not inside meprobamate.

So this is not the clean ladder NSF/ANSI 58 turned out to be. A model holding a wider claim cannot be assumed to hold a narrower one, and the only reliable check is the specific claim against the specific model. → How to check a certification claim yourself

Pitchers do well here

Worth noting, because they do badly almost everywhere else.

Of the 411 models holding the bisphenol A, estrone and nonylphenol claims, 231 are refrigerator filters and 57 are pour-through pitchers.

At the narrow end that collapses. Of the 132 holding TCPP, one is a pitcher.

Compare lead, where pitchers hold 4 of 615 certified models. → Which claims exist in which product form

What a claim here does not tell you

The limitation is structural, and it is the reverse of the usual one.

There is no limit to compare a result against. No monitoring requirement, so no national occurrence data for these fifteen. No line on your annual report, because nothing is measured.

NSF describes the standard as addressing compounds identified in published studies as occurring in drinking water. That is a basis for testing products. It is not a measurement of your supply.

So a claim here reduces something nobody has told you is present, at a concentration nobody has published for your system. → How to read your water quality report

The decision framework

1. Start by accepting there is no number

Every other page here tells you to match a product to a figure on your report. For these fifteen there is no figure to match.

2. Decide whether breadth is what you are buying

Many models carry these claims alongside lead, cyst and microplastics. The emerging-compound claims often come with a product bought for something else.

3. If you want the widest coverage, name the claim

Bisphenol A, estrone and nonylphenol share one set of 411 models, the widest of the fifteen claims in this standard. That is the broadest single point of entry into it.

4. Do not assume the package

Three breaks in the ordering mean a wider claim does not imply a narrower one. Check the model against the claim you actually want.

5. Weigh it against a claim that does have a number behind it

If lead, nitrate or PFAS appears on your report, that is a measured problem with a certified answer. This is not. → Which claims exist at all, and how populated each is

The decision path

Your situation What decides it Where that lands you
A measured contaminant on your report There is a figure to treat to Treat that first
Concerned about medicines generally No limit, no monitoring, no figure Claims exist; evidence for your tap does not
Want the broadest 401 coverage 411 models share the widest set Bisphenol A, estrone or nonylphenol
Want a specific medicine claim Counts range 159 to 400 models Search that claim by name
Want a pitcher 57 of 411 at the wide end Available, and check the claim
Want TCEP or TCPP 133 and 132 models, 1 pitcher Under-sink or counter-top
Already buying for lead or cyst Many models carry both Check whether 401 claims come with it
Wanting proof it is in your water No monitoring exists No such figure is published

Choose a 401-certified product if

  • You are already buying a filter for a measured contaminant and the 401 claims come with it at no extra cost.
  • You want the coverage for your own reasons and accept that no measurement for your supply exists.

Choose to prioritise differently if

  • Your report shows a regulated contaminant above a limit. That is a number, and a number beats a category.

Choose to do nothing if

Mistakes that cost the most money

Reading a certified claim as evidence of contamination. NSF tests products against compounds found in published studies, not in your supply.

Assuming the widest claim covers the narrow ones. The ordering breaks three times, so check the claim you want against the model you want.

Buying for this before buying for a measured contaminant. One has a figure behind it and one does not.

Assuming pitchers cannot hold these claims. 57 of the 411 widest-claim models are pitchers, which is unusual on this site.

Treating NSF/ANSI 401 as a health standard with limits behind it. No contaminant in it has a federal limit at all.

Trusting a page that says no claim exists — including one of ours. That error is corrected above, and the register was right when the prose was wrong.

Three situations, and what changes

Someone who read that medicines are in tap water. The certified field is genuinely large, and there is still no figure for any particular supply. A 401 claim is available and buying one is a judgement call rather than a response to a measurement.

A household already buying a lead filter. Worth checking the model’s other claims. Many lead-certified refrigerator filters and pitchers carry 401 claims as well, so the coverage may arrive with a purchase already justified by a number on the report.

Someone specifically concerned about flame retardants. TCEP and TCPP are the narrowest claims in the standard, at 133 and 132 models, and almost none is a pitcher. The search has to be specific, and the form choice narrows with it.

The objections worth answering

“If nobody regulates them, why does certification exist?” Certification follows what manufacturers choose to test, not what EPA chooses to regulate. This standard is the clearest case of the two diverging.

“So the certification is meaningless.” It records tested reduction against a named compound, which is more than an uncertified product offers. What it cannot do is tell you the compound is present.

“Surely this means our water is full of drugs.” NSF’s basis is compounds identified in published studies as occurring in drinking water. No occurrence figure for any particular system follows from a product claim.

“A reverse osmosis system handles all of this anyway.” Possibly, but NSF/ANSI 58 carries none of these fifteen claims, so a reverse osmosis system would not be certified against them.

“You said elsewhere there was no claim for pharmaceuticals.” We did, and it was wrong. The correction is above and the page that carried it has been fixed.

Who this is for

Anyone who has read that medicines, hormones or plasticisers turn up in drinking water and wants to know what is regulated, what is measured and what can actually be bought against.

It is also for readers who found the earlier error on this site and deserve to see it corrected rather than quietly edited away.

What to settle first

Is there a measured problem on my report that a certified claim could answer?

If there is, treat it first. These fifteen have claims without numbers, which is the opposite of most of this site and a weaker basis for spending.

Which claims exist at all, and how populated each isUnregulated contaminants: the candidate list and the monitoring ruleMicroplastics: what the NSF/ANSI 401 claim coversThe regulated contaminants with no certified filterWhich claims exist in which product form