NSF/ANSI 58 is the only NSF standard written specifically for reverse osmosis systems. The other standards in the cert family — 42, 53, 401, P473 — test individual contaminant reduction claims that any point-of-use filter technology can earn. NSF/ANSI 58 tests the system as a whole: how much it reduces total dissolved solids, how many gallons per day it produces, how much water it sends to the drain per gallon of permeate, and which specific contaminants it can claim under its dedicated reduction-claim protocol. It is the system performance standard. For RO, it is the cert that matters most — and it is also the cert most often misread on packaging.
This article unpacks what NSF/ANSI 58 actually tests, where it overlaps with the rest of the cert family, and how to verify a real 58 listing in the NSF Drinking Water Treatment Units database. If you are evaluating a countertop or under-sink RO system, this is the certification language you need to read with precision.
The cert family in 30 seconds
NSF maintains separate standards for separate jobs. The most common reading mistake on RO packaging is treating "NSF certified" as a single concept rather than a family of distinct standards.
- NSF/ANSI 42 — aesthetic effects. Chlorine taste, odor, particulates. No health-effect claims.
- NSF/ANSI 53 — health-effects contaminant reduction for any point-of-use filter. Lead, mercury, VOCs, cysts, hexavalent chromium, asbestos, MTBE, atrazine. Claim-specific.
- NSF/ANSI 58 — reverse osmosis system performance. TDS reduction (required), plus optional contaminant claims for arsenic, fluoride, lead, hexavalent chromium, PFAS, and others.
- NSF/ANSI 401 — emerging contaminants. Pharmaceuticals, pesticide metabolites, BPA, DEET.
- NSF/ANSI P473 — the PFAS-specific protocol that pre-dated the 2022 PFAS additions to 53 and 58.
For the broader contaminant-reduction landscape, see What is NSF/ANSI 53 and What is NSF/ANSI P473. For how 58 fits the underlying technology, see Reverse Osmosis Membrane Chemistry Explained.
What NSF/ANSI 58 actually tests
The ANSI summary of NSF/ANSI 58-2022 lays out the test categories. There are two types: the required test that every certified system must pass, and the optional reduction claims that a manufacturer can pay to add to its cert listing.
Required: total dissolved solids reduction. Every system certified to 58 has to demonstrate at least 75 percent reduction of TDS under the standard's defined test water. The test water is RO/deionized water dosed with 750 mg/L of sodium chloride, which gives the lab a reproducible baseline TDS to push through the membrane. Pass this test, the system can claim TDS reduction. Fail it, the system cannot hold 58 at all. This is the single guard that separates a "certified RO system" from any other under-sink filter that happens to use a membrane.
Reported, not required: daily production rate and recovery rate. This is the part most consumers do not realize. The 58 standard does not mandate a specific gallons-per-day production rate or a specific recovery ratio (the percentage of feed water that becomes permeate, the rest going to drain). It measures both under standardized lab conditions and requires the manufacturer to publish them on the system's performance data sheet. Daily production rate and efficiency rating exist on the cert listing as comparable numbers, not pass/fail thresholds. For a real efficiency floor, look at the EPA WaterSense program, not 58.
Performance over membrane life. The 58 protocol runs the contaminant reduction tests across the manufacturer's rated cartridge and membrane life — typically by treating water at roughly 200 percent of the rated capacity to confirm the membrane still meets the threshold near end-of-life. This is the same "sustained performance" requirement that separates real RO certification from a single-pass laboratory measurement on a fresh membrane.
Optional contaminant reduction claims. This is the part that overlaps with NSF/ANSI 53 in scope but uses a separate test protocol designed for RO. The claims a manufacturer can add to a 58 listing include:
- Heavy metals — arsenic (pentavalent), barium, cadmium, copper, lead, hexavalent and trivalent chromium, selenium
- Inorganics — fluoride, nitrate, nitrite, radium 226/228, perchlorate
- Cysts and particulates — Cryptosporidium, Giardia, asbestos, turbidity
- VOCs — volatile organic compounds via the standard 58 test cocktail
- PFAS — Total PFAS as of the 2022 revision: PFOA, PFOS, PFHxS, PFNA, PFHpA, PFBS, and PFDA
Each claim is paid for separately. A system can hold 58 for the required TDS reduction plus arsenic and fluoride and nothing else. Another system can hold 58 with the full Total PFAS claim, all the heavy metals, and radium. The standard number on the box does not tell you the claim list. The cert listing does.
What NSF/ANSI 58 does NOT do
Three things people commonly assume about 58 that the standard does not actually require.
It does not certify a specific recovery rate. Historically, household RO systems shipped at 4-to-1 or 5-to-1 reject ratios — four to five gallons of waste per gallon of treated water. Modern tankless and pump-assisted RO systems have pushed toward 1-to-1. NSF/ANSI 58 does not mandate any value here. It just measures the recovery rate and reports it on the performance data sheet so you can compare. If you care about water efficiency as a hard threshold, the EPA WaterSense label is the program with teeth — WaterSense-labeled RO systems are required to send 2.3 gallons of reject water or less down the drain per gallon of treated water, against a typical RO system that wastes five gallons or more.
It does not guarantee PFAS coverage by default. The base 58 cert is TDS reduction. PFAS coverage is an optional claim added in the 2022 revision. A system can hold NSF/ANSI 58 — fully real, fully accredited — without holding the Total PFAS claim. Read the listing. The 2022 update is also why some systems still show only PFOA/PFOS in older 58 listings — the seven-compound Total PFAS framing is the post-2022 addition, and not every certification has been updated yet.
It does not certify the entire system if only one stage is the membrane. Most household RO systems are multi-stage: sediment prefilter, carbon block prefilter, RO membrane, post-filter. NSF/ANSI 58 evaluates the assembled system performance, not just the membrane in isolation. But that does not mean every component is independently certified. Specific cartridges in the chain may carry their own NSF/ANSI 42 or 53 certifications for the contaminants they handle (chlorine on the carbon prefilter, for example). The 58 listing reflects the system, not the parts.
Why the contaminant scope is so broad
This is the under-appreciated property of NSF/ANSI 58. The standard's optional claim list is wider than NSF/ANSI 53's by default — not because the manufacturers paid for more tests, but because reverse osmosis works as a physical exclusion mechanism. The polyamide thin-film composite membrane rejects almost anything dissolved in the water that is larger than a water molecule and carries a charge. A peer-reviewed review of TFC polyamide RO membranes walks through the architecture: polyester backing, polysulfone support, and an ultrathin polyamide active layer with effective transport pathways at sub-nanometer scale.
The implication: when an RO system is tested against the 58 contaminant cocktail, the membrane is doing the same physical exclusion work for arsenic, fluoride, lead, hexavalent chromium, and PFAS in parallel. Many household RO systems can earn the full optional claim list because the underlying mechanism rejects all of them. The brand decides which claims to pay for based on what shoppers ask about.
A practical example: the AquaTru Performance Data Sheet documents reduction across PFOA, PFOS, lead, hexavalent chromium, fluoride, microplastics, and dozens more — the same physical mechanism cleaning everything in one pass.
How to verify a 58 cert claim
The verification path is the same as for any NSF certification. The NSF Listings Database is the source of truth.
- Go to info.nsf.org/Certified/dwtu/ — the NSF Drinking Water Treatment Units listing search.
- Filter by Product Standard: NSF/ANSI 58.
- Type the manufacturer or brand name (minimum three letters).
- Optionally select the specific Reduction Claim under 58 — Lead, Arsenic Pentavalent, Fluoride, Total PFAS, Hexavalent Chromium, etc.
- Confirm the specific model appears in the search result with the exact contaminant claim listed.
If the model does not appear in the database for the standard you searched, the brand is making a marketing claim. The product may still be legitimate — many manufacturers publish independent third-party lab reports tested to the 58 protocol by an accredited lab that is not NSF — but the words "NSF/ANSI 58 certified" should appear in the NSF database for that exact model to mean what they sound like.
A separate path: WQA (the Water Quality Association) runs its own Gold Seal certification program that tests against the same NSF/ANSI 58 standard but lists products in WQA's separate database, not NSF's. Both are accredited third-party certifications. Either is meaningful — the key is independent third-party testing by an accredited body, not just claimed conformance.
Examples from our water catalog
A few systems in our water filter index intersect with the NSF/ANSI 58 conversation, and each comes with its own certification footprint that has to be read carefully.
- AquaTru Countertop Reverse Osmosis — the four-stage countertop RO unit publishes a comprehensive Performance Data Sheet tested under NSF protocols, with reduction documented for PFOA, PFOS, lead, hexavalent chromium, fluoride, and dozens of other contaminants. AquaTru's marketing references NSF/ANSI 58 testing protocols. As with all such claims, verify the current model and claim list directly in the NSF Listings Database before relying on the cert mark for a specific contaminant. Best fit for renters and apartments where under-sink installation is not an option; the AquaTru countertop write-up covers the published testing summary and certification status.
- Hydroviv Under-Sink Filter — Hydroviv is not a reverse osmosis system. It is a multi-stage carbon and ion-exchange under-sink filter tuned to local EPA water-quality data by zip code. Its testing covers PFOA, PFOS, lead, hexavalent chromium, chlorine, and chloramine, and Hydroviv publishes its own performance data tested to NSF protocols. It is included here because it is the comparable non-RO answer for the same contaminant problems — relevant to readers weighing 58-certified RO against a non-RO under-sink.
A note on testing transparency. Some RO systems hold the formal NSF certification (you find them in the NSF Listings Database). Others publish independent reports from accredited third-party labs that test to the 58 protocol but did not pay NSF specifically to issue the cert mark. Both are real third-party testing — neither is the same as a manufacturer-internal assertion of conformance. We cite the lab report URL on each product page so the underlying claim is traceable to a document, not just a marketing line.
The 58, 53, and P473 question for RO buyers
For a household RO system, the cleanest framing of the cert stack:
- NSF/ANSI 58 is the cert that matters most for the system itself. It validates the RO performance and adds the contaminant claims the manufacturer paid for.
- NSF/ANSI 53 can still appear on an RO system for specific contaminant claims that fall outside the 58 default scope or that the manufacturer wants to validate under the 53 protocol independently.
- NSF/ANSI P473 was the PFAS-specific protocol that pre-dated the 2022 PFAS additions to 53 and 58. Many countertop ROs from the mid-2010s still carry P473 listings alongside 58. Post-2022, the Total PFAS claim under 58 (or 53) covers the same ground with a broader compound list — see our P473 explainer for the timeline.
A real-world example: a countertop RO unit might list NSF/ANSI 58 for arsenic pentavalent reduction, NSF/ANSI 53 for cyst reduction, and NSF/ANSI P473 for PFOA and PFOS — three separate listings, each verifiable in the NSF database. The full stack is more complete than any single cert.
The cost implication
RO systems cost more than carbon block filters for three reasons that all show up in the 58 cert process: the polyamide membrane chemistry is more expensive to manufacture than activated carbon, the multi-stage architecture (sediment prefilter, carbon prefilter, membrane, post-filter) requires more cartridge SKUs, and the reservoir-tank or pump components add hardware. A countertop AquaTru retails around $449. An under-sink RO system from APEC or iSpring sits in the $200 to $400 range. A pitcher-style filter is $40 to $100.
For a broader filter-mechanism comparison, see RO vs Carbon vs Gravity. The cost picture is one part of the decision; the contaminant scope on the actual cert listing is the other.
What to do with this
The practical workflow for shoppers evaluating RO systems against an NSF/ANSI 58 claim:
- Identify the contaminant. Pull your utility's Consumer Confidence Report or the EWG Tap Water Database for your zip code. The contaminants that matter most for RO purchases are arsenic, fluoride, lead, hexavalent chromium, nitrate, and PFAS.
- Find the system in the NSF database. Search by model in info.nsf.org/Certified/dwtu/ under Standard 58. Confirm the listing shows the specific contaminant claims you care about — not just the standard number.
- Cross-check daily production and recovery rate. Both are reported on the performance data sheet. If water efficiency matters for your situation (high water cost, drought-restricted region, well water), look for the EPA WaterSense label as a separate guard.
- Read the manufacturer's lab report. For specific contaminant claims, the underlying lab report has the actual percent-reduction numbers and the influent/effluent concentrations the test ran. The cert listing tells you a system passed; the lab report tells you by how much.
The cert badge alone does not tell you the answer. The contaminant list under the badge does — and for RO, NSF/ANSI 58 is the cert that determines which contaminant list the system was tested against.
For broader background, What is NSF/ANSI 53 covers the general health-effects standard, What is NSF/ANSI P473 covers the PFAS-specific protocol, Reverse Osmosis Membrane Chemistry Explained walks through the mechanism, and Reverse Osmosis vs Carbon vs Gravity is the buyer-facing comparison across the three main filter classes.
