Certification guide
NSF P473 explained, and why PFAS certification moved to NSF/ANSI 53 and 58
If you have been told P473 is the certification to look for, that advice is out of date. The test method still exists. The standalone protocol does not. Here is where it went and what to look for instead.
Evidence last checked . Spot something out of date?
The short answer
NSF moved the P473 PFAS test method into NSF/ANSI 53 (for carbon and anion exchange filters) and NSF/ANSI 58 (for reverse osmosis), then removed the standalone P473 protocol. EPA now tells shoppers to look for NSF/ANSI 53 or 58 with a PFAS reduction claim. The claim is the part that matters. Plenty of filters hold 53 and reduce no PFAS at all.
- No paid rankings
- Primary sources checked
- Costs include replacements
- Affiliate-supported
- Evidence last checked
- How we check
What P473 was
NSF P473 was a protocol for one job: measuring how well a home filter reduces two PFAS compounds, PFOA and PFOS. It came into use around 2016 and 2017.
A protocol is not the same thing as a full standard. NSF publishes protocols when a need appears faster than the standards process can move. P473 was exactly that. PFAS became a public concern quickly, and there was no way for a maker to prove a PFAS claim, so NSF built one.
Why it existed at all
Before P473 a filter could say "reduces PFOA" on the box and nobody could check it. There was no agreed test, no agreed starting concentration and no agreed pass mark. P473 supplied all three. That made claims comparable for the first time.
How it moved into NSF/ANSI 53 and 58
In 2017 the NSF Joint Committee on Drinking Water Treatment Units set up a task group to fold the protocol into the main standards. The method went into NSF/ANSI 53 for systems using activated carbon or anion exchange resin, and into NSF/ANSI 58 for reverse osmosis systems.
Once that was done, products and references to P473 were removed and replaced by 53 or 58. The work was complete around 2019. Nothing was weakened. The test simply moved house.
You will still see P473 printed on boxes and in maker documentation. That is legacy labelling, not a current certification. It usually means the product was certified under the old protocol and the packaging never caught up. It is not a reason to rule a product out, and it is not proof of anything either.
What to look for today
EPA's guidance to consumers is short and worth quoting: "check the product packaging for certification to NSF/ANSI 53 or NSF/ANSI 58 for PFAS reduction."
Read that carefully. It is two conditions, not one. The standard number and the PFAS reduction claim. Here is what that looks like in practice:
"Certified to NSF/ANSI 53 for PFOA and PFOS reduction"
Standard named, claim named. This is what you want.
"Certified to NSF/ANSI 58 for PFOA and PFOS reduction"
Same thing for a reverse osmosis system.
"Certified to NSF/ANSI 53"
Certified for something, but the box does not say PFAS. Could be lead. Could be cysts. Check the maker's performance data sheet before you assume.
"Tested to NSF/ANSI 53 standards"
A lab ran the method. Nobody certified the result. Weaker than a certification, and the wording is deliberately similar.
"Meets NSF standards"
This means nothing. It names no standard, no claim and no certifier.
PFOA and PFOS versus Total PFAS
There are two levels of PFAS claim and they are not equal.
The older claim covers PFOA and PFOS, the two most studied compounds. The newer Total PFAS claim covers seven: PFOA, PFOS, PFHpA, PFHxS, PFNA, PFBS and PFDA. It adds shorter-chain compounds, which matter because they are what many manufacturers switched to when PFOA and PFOS were phased out, and because carbon filters generally hold on to them less well.
Total PFAS is the stronger claim. It is also newer, so far fewer products carry it. If a filter only claims PFOA and PFOS, that is not a scandal. It is just a narrower promise than the label might suggest.
"Certified" is not "tested to the standard"
This distinction is doing a lot of work on filter packaging, so it is worth being blunt about.
Certified
An accredited body tested the product, audited the factory, and keeps listing it publicly. The listing can be pulled if the product changes. You can look it up.
Tested to comply with
A lab ran the same test method once. There is no ongoing audit and no public listing. It may be perfectly good data. It is not a certification, and there is nothing to look up.
Both can appear on good products. One of the filters we recommend uses the second wording, and we say so on its card. What matters is that you know which one you are being shown.
Who is allowed to certify
EPA names five ANSI-accredited third-party certification bodies for drinking water filters:
- NSF
- WQA
- IAPMO R&T
- CSA Group
- UL
All five carry equal weight. A filter certified by IAPMO is not worth less than one certified by NSF. If a box shows a logo that is not one of these five, it is not a third-party certification, whatever it looks like.
Why certification does not promise you 4 parts per trillion
This is the most important thing on this page and almost nobody explains it.
EPA set the drinking water limit for PFOA and PFOS at 4 parts per trillion. The certification tests do not use that number as their pass mark.
Under the older PFOA and PFOS claim, a filter was challenged with roughly 1,500 ppt of PFAS and had to bring the water down to 70 ppt to pass. Under the newer Total PFAS claim, the challenge is about 2,160 ppt across seven compounds and the pass threshold is 20 ppt, with tighter individual limits for some compounds.
Twenty is better than seventy. Both are above four. EPA says it directly: the current certification standards for PFAS filters do not yet indicate that a filter will remove PFAS down to the levels EPA has now set for a drinking water standard. EPA has said it is working with the standards bodies to close that gap.
What this means for you. A certified filter is still the best evidence available, and a certified filter starting from a low PFAS level will very likely land you well under 4 ppt. But certification is a pass mark on a hard test, not a guarantee about your particular tap. If you want to know your number, test your water after the filter is installed.
How to check a certification yourself
It takes about two minutes and it is the single most useful thing you can do before buying.
- Find the exact model number. Not the product family, the model.
- Open the listing database for the body named on the box. All five are linked below.
- Search the maker or the model.
- Open the listing and read the claims. You are looking for PFOA, PFOS or Total PFAS by name.
- If the claim is not listed, the product is not certified for PFAS, no matter what the box implies.
Where the filters we cover actually stand
Applying all of the above to the products on this site gives a short list and a longer one.
Certified with a PFAS claim
AquaTru Classic Countertop Reverse Osmosis Purifier
IAPMO R&T, NSF/ANSI 42, NSF/ANSI 53, NSF/ANSI 58, NSF/ANSI 401, NSF P473 (legacy label)
Good products, but the PFAS claim is not certified
Clearly Filtered Water Pitcher
Holds NSF/ANSI 42, NSF/ANSI 372, PFAS not on the certification
Clearly Filtered 3-Stage Under the Sink System
No certification published
Waterdrop G3P800 Tankless Reverse Osmosis System
Holds NSF/ANSI 42, NSF/ANSI 53, NSF/ANSI 58, NSF/ANSI 372, PFAS not on the certification
SpringWell PFAS Water Filter System
No certification published
None of these are bad filters. Two of them are ones we recommend. The point is that you should know which kind of evidence you are buying.
Common questions
Does NSF P473 still exist?
Not as a standalone certification. NSF's joint committee moved the P473 test method into NSF/ANSI 53 for activated carbon and anion exchange systems and NSF/ANSI 58 for reverse osmosis, then removed the references to P473. Some makers still print P473 on packaging, so you will still see the label on shelves.
Is NSF/ANSI 53 the same as a PFAS certification?
No, and this is the mistake that costs people money. NSF/ANSI 53 covers many health-related contaminants. A filter can be certified to NSF/ANSI 53 for lead and carry no PFAS claim whatsoever. You have to check that the PFAS reduction claim is on that specific product's certification.
What should I look for on a filter box today?
EPA tells consumers to check the packaging for certification to NSF/ANSI 53 or NSF/ANSI 58 for PFAS reduction. Look for the standard number and a stated PFAS, PFOA or PFOS reduction claim together. One without the other does not tell you much.
What is the difference between a PFOA/PFOS claim and a Total PFAS claim?
A PFOA and PFOS claim covers two compounds. A Total PFAS claim covers seven, adding shorter-chain compounds that some carbon filters struggle with. Total PFAS is the stronger claim, and it is newer, so fewer products carry it.
Does a certified filter guarantee my water is under 4 ppt?
No. EPA states plainly that current filter certification standards do not yet indicate that a filter will remove PFAS down to the levels EPA has set for drinking water. Certification proves the filter met a defined laboratory test, and the pass threshold in that test sits above the 4 ppt limit.
Which organisations can certify a PFAS filter?
EPA names five ANSI-accredited third-party certification bodies: CSA Group, IAPMO R&T, NSF, UL and WQA. A certification from any of them counts. A logo that is not from one of these is not a third-party certification.