Reverse osmosis and two-stage under-sink filters removed 90% or more in field and lab studies, often to below detection. Pitcher and fridge filters averaged about half, with brands ranging from about 20% to 99%. Short-chain PFAS are harder for carbon. Certification and keeping to the cartridge life matter as much as the filter type.
| Source | What it found (quoted) and its limits | Evidence |
|---|---|---|
| Mulhern R, Bynum ND, Liyanapatirana C, DeStefano N, Knappe D, Gibson JM. Longitudinal assessment of point-of-use carbon filters for removal of PFAS from private well water. AWWA Water Sci 3(6):e1262 (2021) · new field longitudinal study · private wells, North Carolina, monthly x 8 months · n = 18 well users |
influent sum PFAS 4.7-131 ng/L; filters removed 97%-99% of all influent PFAS incl. short-chain PFEAs for full rated lifetime; prevalence above reporting limits reduced 99.5%; above detection limits reduced 92% (abstract) Limits: abstract only; low influent; one product type |
Evidence: strong Access unknown · source · added 2026-10-30 |
| Herkert NJ, Merrill J, Peters C, Bollinger D, Zhang S, Hoffman K, Ferguson PL, Knappe DRU, Stapleton HM. Assessing the effectiveness of point-of-use residential drinking water filters for perfluoroalkyl substances (PFASs). Environ Sci Technol Lett 7:178-184 (2020) · new field study in homes · Central + southeastern North Carolina · n = 73 homes; 89 filtered samples (76 POU, 13 POE) |
RO and dual-stage: >=90% PFCA/PFSA removal; pitcher ~50% (long-chain ~61%, short-chain ~46%); refrigerator ~50%; faucet (n=2) >75%; single-stage under-sink >84% PFSAs but ~half PFCAs; AC filters short-chain avg ~41%; whole-house GAC: 4 of 8 systems showed increased PFAS after filtration Limits: Convenience sample; small n per type (faucet n=2, countertop n=1); no link found to filter age (7.5 months avg); detection-limit capped removal % |
Evidence: strong Paywalled · source · added 2026-10-30 |
| Teymoorian T, Dinh QT, Barbeau B, Sauvé S. Performance of pitcher-type POU filters for the removal of 75 PFAS from drinking water: comparing different water sources. Front Environ Chem 5 (2024) · new lab + home test, real tap water · Saint-Donat and Montreal (Quebec) tap water · n = 5 pitcher products |
background sum75 PFAS 13 and 56 ng/L; volume-weighted removal after 160 L: ZeroWater 99% / ~100%; ClearlyFiltered 99% / 96%; Aquagear 77% / 60%; Brita Elite 20% / 48%; Brita Standard 38% (Montreal); ZeroWater desorbed TDS/TOC after 80 L in Saint-Donat; Aquagear sulfate rose after 20 L Limits: 2 waters only; low influent PFAS; Canadian waters |
Evidence: moderate Free to read · source · added 2026-10-30 |
| MacKeown H, Magi E, Di Carro M, Benedetti B. Removal of PFAS from tap water by means of point-of-use treatment: A review. Sci Total Environ 954:176764 (2024) · new review (narrative, structured) · global literature to 2024 · n = 22 peer-reviewed papers (10 US; 5 real-life) |
only 5 of 22 studies in real-life settings; summarises Herkert (RO/dual-stage >=90%, GAC short-chain 41%), Teymoorian pitchers 38-60% (3 filters) vs ~100% (2), EWG 22-100%; NSF challenge 2160 ng/L total requires >=99%; abstract: 'improperly maintained activated carbon systems might even raise PFAS levels' Limits: Not a formal systematic review/meta-analysis; cites EWG (grey) |
Evidence: moderate Free to read · source · added 2026-10-30 |
| Choudhary K, Lanzarini-Lopes M, Parzen H, Perez Medina X, Reckhow D, Rogers R, Tobiason J, Wittbold P (UMass Amherst / MassDEP). PFAS Point-of-Use Treatment System Study, Final Report (2023) · new government-funded on-site test with real well water · Townsend MA well water, mobile lab, spring 2022 · n = 17 devices |
raw PFAS6 122.6 ng/L; at 100% rated throughput 5 of 17 below 20 ng/L MCL, 12 of 17 exceeded; PUR faucet 13.8 ng/L at 100% -> 32 ng/L at 200%; 3M plumbed-in, A.O. Smith plumbed-in, ZeroWater non-detect at 100% and 200%; annual cost $46-$176 Limits: One water source; only 2 sampling points per device; device count inconsistent (17 vs 20) in report |
Evidence: moderate Free to read · source · added 2026-10-30 |
| Patterson C, Burkhardt J, Schupp D, Krishnan ER, Dyment S, Merritt S, Zintek L, Kleinmaier D. Effectiveness of point-of-use/point-of-entry systems to remove PFAS from drinking water. AWWA Water Sci (2019) · new lab treatability study (US EPA) with Widefield-aquifer-type test water · Fountain/Security/Widefield CO context · n = 3 RO units; 2 GAC media (RSSCT) |
test water PFOS 1,600 ng/L, PFHxS 1,000, PFOA 800 ng/L; RO#1 and RO#3 all PFAS below detection; RO#2 6 of 42 samples detected (e.g. PFOS 77 ng/L, 94.3%; PFNA 49 ng/L, 76.7%) after shutdown/restart; GAC exceeded 70 ng/L PFOS+PFOA after ~4,800 BV (coconut, ~33 days at max-day conc) and ~2,900 BV (coal, ~20 days); EPA summary: RO non-detect '<3 ppt' Limits: Short-duration tests; high challenge conc.; POE GAC modelled not field-tested |
Evidence: moderate Free to read · source · added 2026-10-30 |
| Zhang X, Wang D, Wang K, Yang Z, Ai M, Zhang Y. Legacy and novel PFAS in tap water from East China: impact from water sources and risk mitigation by household purifiers. Environ Sci Technol (2025) · new monitoring + household purifier sampling · East China tap water · n = not stated in abstract |
tap water up to 8970.3 ng/L; RO purifiers 'demonstrated near-complete removal for most PFAS' (no % in abstract) Limits: abstract only; no numeric removal in abstract |
Evidence: weak Paywalled · source · added 2026-10-30 |
| US EPA Office of Water. Identifying Drinking Water Filters Certified to Reduce PFAS (web guidance) (2024) · new government guidance · US · n = n/a |
GAC/IX cost $20-$1,000, RO $150-$1,000; RO wastes ~1 gallon per gallon treated; as of April 2024 certification standards 'do not yet show removal to the levels of EPA's new drinking water standard' Limits: Guidance, no primary data |
Evidence: weak Free to read · source · added 2026-10-30 |
| He A, Lu Y, Chen F, et al. Exploring the origin of efficient adsorption of PFAS in household point-of-use water purifiers: joint experimental and computational study. Sci Total Environ (2022) · new in-field + lab mechanistic · Chinese tap water (homes) · n = not stated in abstract |
in-field removal 21%-99% of 14 PFAS (influent/effluent ratio); hydrophobicity key factor Limits: abstract only; n not given; China |
Evidence: weak Access unknown · source · added 2026-10-30 |
| NC Dept of Environmental Quality, Div. of Waste Management. GAC System Pilot Study handout (Chemours/GenX private wells) (2018) · new government field pilot · Cape Fear region private wells, NC · n = 6 homes |
raw GenX 159-1,910 ng/L; use 450-2,500 gal/week; no effluent numbers in handout Limits: Interim handout only; results posted elsewhere; not peer reviewed |
Evidence: weak Free to read · source · added 2026-10-30 |
| Minnesota Department of Health (Water Science & Marketing / WQA). MDH Evaluation of Point-of-Use Water Treatment Devices for Perfluorochemical Removal, Final Report Summary (2008) · new government lab + field test · Oakdale and St. Paul Park MN well water · n = 14 selected; 11 field-tested (4 AC, 7 RO) |
all RO removed PFCs below reporting limit 0.2 µg/L; AC field: all 4 below 0.2 µg/L; lab: PFBA breakthrough Kenmore at 500 gal (0.2-0.7 µg/L), Kinetico at 250 gal (1.4-2.2 µg/L at 500 gal) Limits: Old (2008); reporting limit 0.2 µg/L = 200 ng/L, far above current 4 ng/L MCLs; only 3 PFAS |
Evidence: weak Free to read · source · added 2026-10-30 |
Evidence grade: strong = trials, large field studies or official datasets; moderate = smaller or lab studies with real water; weak = single small studies, makers' specs or secondary reports. "Free to read" means the source itself is free to read online.