The short version: NSF/ANSI P473 is the Special Protocol that tests a drinking water filter for reduction of a specific list of long-chain PFAS compounds — most importantly PFOA and PFOS. If a filter does not hold P473 — or publish equivalent independent third-party testing against the same protocol — its PFAS claims are not third-party verified.
This matters because most pitcher filters do not hold it. The Brita on your counter, the cheap pitchers at the grocery store, the activated-carbon faucet attachments — almost none are P473 certified. They are usually tested for chlorine taste-and-odor reduction (NSF/ANSI 42) and many are tested for some health-effects contaminants like lead (NSF/ANSI 53), but PFAS is a separate test with separate pass criteria. For a deeper protocol-level walkthrough including how P473 interacts with NSF/ANSI 53 and 58 and the broader 12,000-compound PFAS family, see our companion piece on the P473 protocol and what it misses.
P473 is a Special Protocol, not a full ANSI standard
P473 is technically a Special Protocol rather than a full NSF/ANSI standard. The "P" signals Protocol — a method developed faster than the multi-year ANSI consensus process to address an urgent contaminant. NSF has been folding the P473 methodology into the parent NSF/ANSI 53 and NSF/ANSI 58 standards as part of the regular standards-update cycle, which is why some products are labeled "NSF/ANSI 53 + P473" and others show a PFAS reduction claim under updated 53 or 58 listings. The underlying test is the same; the wrapper is shifting.
NSF International is the independent, accredited public-health organization that writes these standards and runs the certification labs. NSF does not sell water filters. When a filter carries the P473 mark, it means NSF or another accredited lab took a production-line unit, ran defined test water through it, and measured contaminants at the outlet against the standard. The certifications are public records, searchable by anyone.
The contaminant set P473 actually targets
P473 is not a generic "PFAS" certification. It tests a specific group of long-chain perfluoroalkyl acids:
- PFOA (perfluorooctanoic acid) — the legacy Teflon manufacturing compound
- PFOS (perfluorooctane sulfonate) — the legacy Scotchgard compound
- Six additional long-chain PFAS that the protocol's analytical method captures: PFNA, PFHxS, PFHpA, PFBS, PFDA, and PFUnDA
The headline pair is PFOA and PFOS. These two are the most studied PFAS in the CDC ATSDR toxicology literature and the two that drove the original certification work. The protocol's pass/fail criteria are written explicitly around them.
What P473 does not cover: GenX (HFPO-DA), PFBA, PFHxA, and the thousands of shorter-chain replacement compounds the chemical industry shifted to after PFOA and PFOS were phased out. For the difference between these compound families and why it matters, see our PFOA vs PFOS vs GenX explainer.
The pass criteria — what 70 ppt actually means
Per the NSF Contaminant Reduction Claims Guide, to pass P473 the filter must:
- Take test water spiked with PFOA or PFOS at 0.5 parts per billion (500 parts per trillion) influent
- Reduce the outlet concentration to below 0.07 parts per billion (70 parts per trillion)
- Maintain that reduction across the entire rated cartridge life — not just the first gallon
The structure is the same as other NSF reduction tests: spiked influent in, single-digit-percentage residual out, sustained over the cartridge life.
Why 70 ppt — and why it is now dated
The 70 ppt outlet target is not arbitrary. It was set to match EPA's 2016 lifetime Health Advisory Level of 70 ppt for PFOA and PFOS individually or combined. The Health Advisory was non-enforceable guidance, but it was the public-health benchmark of the day, so the certification target tracked it.
That benchmark has since moved twice. EPA issued interim updated advisories in 2022 dropping the recommended levels into the parts-per-quadrillion range, and in April 2024 EPA finalized the first National Primary Drinking Water Regulation for six PFAS compounds, setting legally enforceable Maximum Contaminant Levels of 4 parts per trillion for PFOA and PFOS individually. The certification threshold did not move with the regulation. A filter passing P473 today is meeting a target roughly 17 times higher than what utilities must meet under the federal rule.
In practice this is less alarming than it sounds. Most P473-certified filters in published independent lab reports show outlet concentrations well below 70 ppt — often into single-digit parts per trillion. The protocol is written so 70 ppt is the worst acceptable outlet, not the typical one. The cert mark alone does not tell you the actual reduction; read the lab report.
P473 vs the NSF/ANSI 53 PFOA/PFOS challenge
A common source of confusion is the relationship between P473 and the PFOA/PFOS challenge built into NSF/ANSI 53 itself. The two overlap but are not identical:
- NSF/ANSI 53 PFOA/PFOS challenge typically covers a narrower compound set and historically used a lower influent challenge concentration.
- NSF/ANSI P473 covers a broader set of long-chain PFAS species with a more aggressive influent spike, designed to demonstrate sustained performance against real-world hot-spot tap water.
This is why most pitcher and reverse osmosis systems targeting PFAS now carry both the NSF/ANSI 53 listing and the P473 protocol notation, or the newer NSF/ANSI 58 PFAS reduction claim for RO systems. The 53 listing tells you the filter handles standard health-effects contaminants like lead and VOCs; the P473 notation tells you the same filter has demonstrated PFAS reduction under the more demanding spiked-influent test.
How to verify a P473 claim
The 60-second verification:
- Go to the NSF Listings database at info.nsf.org/Certified/DWTU/
- Search by manufacturer, brand, or model number
- Confirm the specific product appears under NSF/ANSI 53, P473, or NSF/ANSI 58 with the PFOA/PFOS or Total PFAS reduction claim
- Cross-check the cert mark on the product packaging or specification sheet — the listing must be a product, not a generic brand mention
If a brand advertises "NSF certified for PFAS removal" but the product does not appear in the listings database, that is a marketing claim, not a certification. The Water Quality Association certification program operates a parallel accreditation that mirrors NSF/ANSI standards, and a WQA Gold Seal listing for the equivalent claim is a legitimate cross-reference if you cannot find an NSF entry.
What P473 does not mean
Three common misreadings to avoid:
- It does not mean "removes all PFAS." P473 covers a defined long-chain subset. Newer compounds like GenX are not in scope unless the product is separately tested under the updated NSF/ANSI 58 Total PFAS framework.
- It does not guarantee the 4 ppt MCL. The certification target is 70 ppt at the outlet. Most certified filters perform far below that, but the mark alone does not prove MCL-level reduction. Read the manufacturer's published lab report for the specific product.
- It does not transfer between products in a brand line. Certification is per-model. A brand's flagship pitcher may hold P473 while a budget model does not. Verify the exact model number.
For our current picks across pitcher, countertop reverse osmosis, and under-sink systems with verified P473 or equivalent third-party PFAS data, see Best PFAS Water Filters and the AquaTru countertop product entry for the specs and certification summary. For a related contaminant where certification is more mature than for PFAS, see Lead in Tap Water.
The takeaway
If PFAS is your concern, the certification to ask about is NSF/ANSI P473 — or the equivalent PFAS reduction claim listed under updated NSF/ANSI 53 or 58. If the brand cannot point to a listing on info.nsf.org, ask for the independent third-party test report. If they cannot point to either, the PFAS claim is unverified, however confident the marketing copy sounds.
Frequently asked questions
(See structured FAQ at the bottom of this page for full answers.)
