Update — May 2026
On May 1, 2026, the White House Office of Management and Budget cleared two EPA proposed rules that scale back portions of the 2024 PFAS National Primary Drinking Water Regulation. The MCLs for PFOA and PFOS are unchanged at 4 parts per trillion, but the compliance deadline for utilities has been extended from 2029 to 2031. EPA is also reconsidering the regulatory determinations for PFHxS, PFNA, GenX (HFPO-DA), and the Hazard Index for mixtures. The picks below are unchanged — filtering at home does not depend on the federal timeline. See full breakdown of the 2026 rollback.
What PFAS are, and why this guide exists
Per- and polyfluoroalkyl substances are roughly 15,000 synthetic chemicals built around a carbon-fluorine backbone — the strongest single bond in organic chemistry. That bond is what gives PFAS their useful properties (non-stick, water-resistant, grease-proof) and what makes them persistent: they do not break down meaningfully in soil, water, or the human body. Decades of industrial discharge, firefighting foam, and consumer-product waste have left detectable PFAS in drinking water across all 50 US states.
In April 2024 the EPA finalized the first legally enforceable national drinking water standards for six PFAS compounds. The headline numbers:
- PFOA: Maximum Contaminant Level of 4.0 parts per trillion (ppt)
- PFOS: MCL of 4.0 ppt
- PFHxS, PFNA, GenX (HFPO-DA), PFBS: regulated as a mixture via the Hazard Index — sum the ratios of measured concentration to each compound's Health-Based Water Concentration, and a total of 1.0 or higher exceeds the standard
Originally utilities had until 2029 to comply. The May 2026 rollback extended the PFOA/PFOS deadline to 2031 and reopened the determinations for the other four compounds, but the 4.0 ppt MCLs are unchanged. Practically: you cannot assume your tap water is below the regulatory threshold today, and the utility serving you may not be required to fix it for another five years. Filtering at the point of use is the fastest control you have.
How NSF/ANSI P473 and NSF/ANSI 53 work
Most pitcher filters do not target PFAS. Standard activated-carbon cartridges are tested for chlorine taste-and-odor reduction (NSF/ANSI 42) — not PFAS. The certification you actually want is NSF/ANSI P473, the protocol added specifically to address PFOA and PFOS.
The P473 challenge looks like this: NSF (or another accredited lab) takes a production-line filter, runs water spiked to 1.5 parts per billion PFOA and PFOS through it, and measures the outlet concentration. To pass, the filter must drive both compounds below 70 parts per trillion at the outlet — and it has to keep doing it across the rated cartridge life, not just on gallon one.
Two important caveats:
- The 70 ppt threshold predates the 4 ppt MCL. P473 was set when EPA's prior health advisory was 70 ppt. Most filters that pass P473 in independent reports drive PFOA/PFOS into single-digit ppt — well below the 4 ppt MCL — but the certification itself only guarantees ≤70 ppt. Read the actual lab report when the brand publishes one.
- P473 only tests PFOA and PFOS. The other four regulated compounds (PFHxS, PFNA, GenX, PFBS) are not part of the P473 challenge. Filters that hold P473 may or may not reduce them — check the manufacturer's full performance data sheet.
NSF/ANSI 53 is the broader health-effects standard (lead, VOCs, certain pesticides, cysts). A filter can hold NSF 53 without holding P473 — the two are not interchangeable. For PFAS specifically, P473 is the box you want checked.
Categories of removal technology
Per EPA and AWWA technical guidance, three treatment technologies are well-studied for PFAS at the point of use:
- Reverse osmosis (RO). A semi-permeable membrane forced under pressure rejects almost everything larger than a water molecule, including long- and short-chain PFAS. Most thorough single technology. Wastes water (typically 1:3 to 1:4 reject-to-permeate). Removes minerals along with contaminants — the flat taste is real.
- Granular activated carbon (GAC). Adsorbs long-chain PFAS (PFOA, PFOS, PFHxS) reasonably well. Struggles with short-chain compounds like PFBS and GenX, which is why GAC alone is not enough for the full Hazard Index.
- Ion exchange (IX) resin. Anion-exchange resins designed for PFAS perform well across both long-chain and short-chain compounds. Most modern P473 pitchers stack IX resin on top of GAC for this reason.
- KDF media (kinetic degradation fluxion). Useful for chlorine, hydrogen sulfide, and some heavy metals; not a primary PFAS technology. Often paired with carbon in multi-stage systems.
Practically: for PFOA and PFOS only, GAC plus IX (a typical certified pitcher) is enough. For the full Hazard Index of six compounds — including the short-chain ones — RO is the most reliable single answer.
How selections were made
VettedClean is an information aggregator — this site does not test, install, or use the filters on this page. Selections are based on what credible third parties disclose: NSF certification listings, published independent third-party lab reports, EPA water-quality data, and the published material literature on each filtration category. Brand claims are cross-checked against the NSF certified products database where applicable. See How we vet for the full methodology.
Pitchers — the entry point
Clearly Filtered is the pitcher we feature for most households. Independently tested per NSF/ANSI P473 protocol for PFOA and PFOS reduction, plus dozens of other contaminants per the brand's published test reports. Manufacturer-specified cartridge life is ~100 gallons. Slower flow rate is the documented trade-off for the contaminant-removal class.
Epic Pure is the budget-leaning sibling with similar testing-backed PFAS reduction. ~150-gallon cartridge life, lower unit cost, slightly slower flow. The right call when the price difference matters.
Brita Elite is the mass-market option that actually holds NSF/ANSI P473 certification — formerly the Longlast+ line. Roughly 120 gallons per filter, replacements available everywhere, lowest filtered-water cost-per-gallon in our list. Plastic pitcher and plastic-housed cartridge are the trade-offs.
ZeroWater is the five-stage budget pitcher with a TDS meter included and published P473-protocol testing. Filter life is short (15-25 gallons in hard water) and the filtered water tastes flat because dissolved minerals are removed alongside contaminants. Best for renters who want the lowest upfront cost on a tested pitcher.
Countertop — most thorough no-plumbing option
AquaTru is countertop reverse osmosis with no plumbing. Front tank fills with tap, four-stage filtering pushes purified water into the back tank. Tested to reduce PFOA, PFOS, PFNA, and 80+ other contaminants. Bigger upfront cost, biggest counter footprint, most thorough filtration in the category that does not require an under-sink install. The full review is in our AquaTru deep-dive.
Crystal Quest Countertop sits between a pitcher and full RO — multi-stage carbon and KDF media that screws onto the faucet via a diverter. Manufacturer-published contaminant list rather than P473 certification. The trade-off pick when you want longer filter life than a pitcher but cannot install under-sink.
Under-sink — out of the way, longest filter life on a single change
Aquasana AQ-5300+ is the easiest path to NSF/ANSI P473 certification under the sink without going to RO. Three-stage filtration, dedicated faucet, ~600 gallons per filter set (about a year for a household of two). Holds P473 specifically for PFOA and PFOS — the regulatory checkbox most people actually want.
Hydroviv ships a filter cartridge tuned to your zip code's known contaminants based on EPA water-quality data. Independent testing covers PFOA, PFOS, lead, and chromium-6. DIY install on most kitchens. Replacement cartridges cost more than generic options because they are zip-tuned.
Gravity — long filter life, no electricity
Big Berkey is the household-sized gravity system, 2.25-gallon lower chamber, two Black Berkey elements rated to 6,000 gallons per pair. Berkey Travel is the same engine in a smaller stainless body. Both publish their own independent lab reports rather than holding NSF P473 — that is a real editorial caveat we cite explicitly. Read our Big Berkey review for the long version. Throughput is slow; capacity buffers it.
The "treated water still tests positive" problem
A point most articles miss: even at the new 4 ppt MCL, parts-per-trillion is the practical detection floor for most commercial water-quality labs. A "non-detect" result does not mean zero — it means below the lab's reporting limit, which can be 2 ppt, 4 ppt, or higher depending on the method. When you read a manufacturer's test report, look at the method detection limit and whether the post-filter result is genuinely "non-detect" or just "below MCL." Both are good outcomes; only the first is unambiguous.
This is also why lead testing and PFAS testing belong on the same kitchen-water audit — both are sub-ppb contaminants where the detection method matters as much as the result.
Comparing the technologies
For a deeper comparison of how RO, carbon, and gravity systems differ on contaminants beyond PFAS, see our RO vs carbon vs gravity primer. For the certification mark itself, What is NSF/ANSI P473 walks through verification.
Which one is right for you
| If you... | Pick | | --- | --- | | Want the simplest entry point | Clearly Filtered pitcher | | Want the cheapest tested PFAS pitcher | Epic Pure or Brita Elite | | Want the most thorough no-plumbing filter | AquaTru countertop | | Want NSF P473 under the sink without RO | Aquasana AQ-5300+ | | Want zip-code-tuned under-sink filtering | Hydroviv | | Want long filter life with no electricity | Big Berkey or Berkey Travel | | Want a TDS meter and short cartridges | ZeroWater |
Frequently asked questions
(See structured FAQ at the bottom of this page for full answers.)
