Do refrigerator water filters remove lead?

The short version

• 7 of the 45 best-selling refrigerator filters in our August 2026 sample hold NSF/ANSI 53 — the standard under which lead reduction is certified, for the exact model sold. That is 16%. Holding the standard doesn’t by itself confirm lead is among that model’s certified claims.
• 0 of the 9 Whirlpool-compatible filters in the sample and 0 of the 4 Frigidaire-compatible filters hold one. 4 of 12 LG-compatible, 1 of 9 Samsung-compatible and 2 of 11 GE-compatible filters do.
• NSF/ANSI 372 certifies the lead content of a filter’s materials, not its ability to remove lead from water. It is the badge most often mistaken for a lead-removal certification.
• The only independent lab test of lead removal in refrigerator filters, AHAM/WQRF 2018, found 32 of 32 counterfeit filters failed the NSF/ANSI lead standard at twice rated life. No equivalent test of honestly labelled aftermarket filters exists.
• 22 of the 45 listings (49%) claim certifications the public databases do not show for the model being sold.

Lead reduction is governed by NSF/ANSI 53, and the standard most listings name is not it

A refrigerator filter can be fully certified under NSF/ANSI 42 and remove no lead in practice, because 42 and 53 measure different things. NSF/ANSI 42 is the aesthetic-effects standard: chlorine reduction, taste and odour, plus material and structural integrity. It is the baseline most aftermarket fridge filters carry, and in our dataset it is the most common certification by a wide margin. NSF/ANSI 53 is the health-effects standard, and it is where lead and cyst reduction live. A listing that names NSF/ANSI 42 and nothing else has told you something real about chlorine and nothing at all about lead.

Our 45-filter sample shows the split clearly. Fifteen filters (33%) carry some verifiable certification for the exact model sold. Five of those are certified for 42 alone, and they include well-known names: Glacier Fresh GF-1000P, Mist’s EDR1RXD1-compatible cartridge, HMASAV’s W10295370A replacement and Waterdrop’s WD-F38. Every one of those is a real, listed product. None of them holds a 53 listing.

Then there is NSF/ANSI 372, which is a separate programme about the lead content of the materials in a device, not about what the device removes from water. A 372 listing tells you the plastic and metal in the cartridge are not leaching lead into your glass. It does not tell you the filter captures lead coming in from the supply.

Seven of the 45 filters hold a 53 listing, and five of them trace to one manufacturer

The seven, with the model codes as they appear in the databases:

• Waterdrop WD-DA29-00020B, sold for Samsung fridges taking DA29-00020B / HAF-CIN/EXP. Certified NSF/ANSI 42, 53 and 401, manufacturer Qingdao Ecopure Filter Co., Ltd.
• Waterdrop LT1000PC (element WD-LT1000P), sold for LG fridges taking LT1000P / ADQ74793501. Certified 42, 53 and 401 under Waterdrop EFF-6046S.
• Waterdrop’s LT700P replacement (elements WD-LT700P and WDP-LT700P), sold for LG fridges taking LT700P. Certified 42, 53 and 401 under Waterdrop EFF-6032S.
• Maxblue MB-LT1000P, sold for the same LG LT1000P / ADQ74793501 group. Certified 42, 53 and 401 under Maxblue EFF-6046S.
• Filterlogic FL-RF46, also sold for the LG LT1000P class. Certified 53 and 401, and notably not 42, which is an unusual shape of certification in either direction.
• GE XWFE, an OEM filter, listed under “GE Appliances, A Haier Company” for NSF/ANSI 42, 53, 372, 401 and P473.
• GE RPWFE, the same certification set under the same certificate-holder name.

Five of the seven are element codes filed by a single contract manufacturer, Qingdao Ecopure Filter Co., Ltd. Three of those carry the same EFF-6046S designation but are sold under three different retail brands. That is worth sitting with. The certification belongs to the manufacturer’s element, and the brand on the box is a layer of marketing on top of it. Two brands can share one certification and describe it completely differently.

They do. Two of the seven filters that genuinely hold a 53 certification advertise “NSF 42, 372 certified” on the listing and never mention 53 or 401 at all. Both are Waterdrop LG-compatible SKUs. A buyer reading that box would have no way of knowing that the filter in their hand is one of the few in the entire sample holding NSF/ANSI 53, the standard under which lead reduction is certified. The certification is real and the signposting points the wrong way.

The 53 certification gap is shaped by fridge platform, and it is total for Whirlpool and Frigidaire

Sixty-seven percent of the sample (30 of 45 filters) could not be confirmed certified for the exact model sold. That aggregate number hides a pattern that matters more per fridge than the headline does.

Whirlpool is the hardest case. Nine filters in our sample are sold for Whirlpool fridges taking the everydrop EDR1RXD1 / W10295370 filter. Zero of them holds an NSF/ANSI 53 listing. The OEM cartridge itself produced no listing in any of the three databases, and the aftermarket filters that did certify carry 42 only: Waterdrop WD-F38, Mist’s EDR1RXD1 replacement and HMASAV’s W10295370A replacement. Meanwhile a Waterdrop WD-F38 listing claims “NSF 42 and NSF 372 certifications” when the databases show 42 for the element and no 372.

Frigidaire is similar. Four filters in the sample fit Frigidaire fridges, and none holds a 53 listing for the model sold. The parent company, Electrolux Home Products, genuinely holds five active IAPMO certifications including NSF/ANSI 42 and 53, but the WF3CB / PureSource 3 model itself does not appear in the listing.

Samsung has exactly one: the Waterdrop WD-DA29-00020B. GE has two, and both are the OEM cartridges rather than the compatible alternatives. LG has four, and all four are aftermarket.

The practical version: if your fridge is a Whirlpool and the filter must be an EDR1RXD1 / W10295370 class cartridge, nothing in this 45-filter sample gives you a verifiable NSF/ANSI 53 listing. That is not the same as saying those filters fail to remove lead. It means the certification that would let you check is not there to check.

A “372 certified” badge is the most common way a lead claim goes wrong

ICEPURE is the clearest worked example in our dataset, because the same brand used identical title text for three SKUs and got three different answers.

ICEPURE RWF3600A and RWF1500A are both titled “NSF 42, 372 Certified.” A sequential re-check of IAPMO R&T under the brand’s registrant, TOP PURE USA INC, found NSF/ANSI/CAN 372-2022 active for both models and no NSF/ANSI 42 match for either. For a filter sold for GE fridges, that means the one element of the claim a buyer would read as lead protection is a materials standard, and the removal standard in the same sentence is not confirmed.

ICEPURE RWF5200A carries the same title text and checks out: NSF/ANSI 42 plus NSF/ANSI/CAN 372, both confirmed for the exact model.

The same confusion runs through OEM material. Samsung Electronics holds real IAPMO certifications, including one filed as “Lead Free Plumbing Products” and NSF/ANSI/CAN 61 for drinking water system components. Neither is an NSF/ANSI 42 or 53 filtration-performance certification, and neither names the DA29-00020B / HAF-CIN/EXP cartridge the listing sells. “Lead-free” in a certification file describes the plumbing, not the water.

If you take one rule from this page: on a refrigerator filter listing, 372 is not the lead-removal standard. 53 is.

The independent lead-testing record is one 2018 study of counterfeits, and it has never been repeated

We compiled every findable instance of independent lab testing of refrigerator-filter lead removal. The full inventory, with sources, is in our independent lead-removal testing study. The short version is thinner than any buyer would guess.

The 2018 AHAM study, published as Fake Filters, Real Problems with testing funded partly by the Water Quality Research Foundation, sent 32 counterfeit refrigerator filters to three outside labs. All 32 failed the NSF/ANSI lead standard at twice the rated filter life. A separate batch of 46 counterfeits went through a chemical extraction test, and ten compounds above allowable concentrations leached into water that started clean. The study cost a reported $35,000 and was commissioned by the appliance manufacturers’ trade association, which is a financial interest worth naming alongside the result. Its scope limitation matters more than its age: it sampled filters that were already infringing trademarks, not the wider market of honestly labelled compatible filters.

Nothing has replaced it. Consumer Reports ran what it called the first-ever refrigerator filter test in April 2026, across 29 products and 12-plus brands, and measured flow rate and useful lifespan rather than contaminant removal. Its contaminant claims are pulled from each product’s existing certification paperwork. The 2023 Tang, Lytle, Achtemeier and Tully meta-analysis in Water Research pooled 1,653 filters and found 99% of field-tested units met the lead benchmark against 61% of lab-tested units, but refrigerator filters are not broken out as a category anywhere in its reported breakdown. Stapleton and Knappe’s 2020 study tested PFAS, not lead. CBP’s seizures, 5,202 units in 2019 and 29,056 in 2022 among them, are trademark and IP actions, and CBP does not test water quality.

So the honest answer to the obvious follow-up question, whether cheap aftermarket filters remove lead as well as OEM ones, is that nobody has tested it. That question is the planned lab phase of our own work, and it has not been run yet.

Lead failures cluster past rated capacity, which makes the replacement schedule a safety issue

The strongest pattern in the existing data is not about brand. It is about how long a filter has been in service. The AHAM counterfeits were tested at 200% of rated capacity and every one failed. In the 2023 meta-analysis, failures were concentrated where filters had been pushed beyond their certified conditions, which is the gap between the 99% field figure and the 61% lab figure.

Both findings point the same way: a cartridge that removes lead on day one is not guaranteed to remove lead at month twelve. If you have a 53-certified filter, the certification describes performance inside its rated capacity, and the rated capacity is the number on the box.

We cannot tell you how many of the 45 filters in our sample are being run past spec, because our dataset records certification status and listing claims, not rated capacity or replacement intervals. That is a real limit on what this study can say, and it is on the list below.

Checking whether a specific listing is certified for lead takes about a minute

Match the exact model, not the brand. Search the model code as printed, including the alternate codes that appear on the same cartridge: DA29-00020B and HAF-CIN/EXP, LT1000P and ADQ74793501, EDR1RXD1 and W10295370. Then search the same code in NSF’s listings at info.nsf.org, IAPMO R&T’s product directory at pld.iapmo.org and WQA’s product finder at find.wqa.org. Look for the number 53 attached to your model, then open the listing and check whether lead specifically is named among its reduction claims — holding 53 does not by itself guarantee lead is one of them.

Search the certificate-holder name as well as the retail brand, because certification databases index by whoever filed the paperwork. GE’s OEM cartridges are listed under “GE Appliances, A Haier Company.” Frigidaire’s parent files as Electrolux Home Products. Icepure’s filings sit under TOP PURE USA INC. Searching only the brand on the box is the single most common way to reach a wrong “no listing” conclusion.

Brand-level certification is not model-level certification, and this is the most common halo in the sample: 14 of 45 filters (31%) came from brands with genuine certifications covering other models. AquaCrest holds NSF listings for exactly five models, and none of the six best-selling AquaCrest filters in our set is one of them. If a listing names a standard without naming a certifier, or names a certification the brand holds for a different cartridge, you have found the pattern.

A refrigerator filter is the wrong tool for a problem bigger than the dispenser

A fridge cartridge treats water dispensed through the refrigerator, which is a small share of household water use and none of the water used for cooking at the sink, showering or laundry.

What our files support is the certification question, not the device question. The EPA publishes consumer guidance on identifying filters certified to reduce lead in drinking water, most recently a June 2024 document, and that guidance turns on the same NSF/ANSI 53 certification we are checking here. Authoritative guidance on when a point-of-use refrigerator filter is the wrong device for a lead problem, for example where a service line or a whole-home condition calls for point-of-entry treatment, is not in our files, so we state no verdict on it. That gap is recorded below rather than filled with an assertion.

What we could not determine

• Whether LG’s two OEM cartridges are certified under a name we have not found. LT1000P / ADQ74793501 and LT700P are absent from NSF, WQA and IAPMO under both “LG” and “LG Electronics.” Only third-party compatible filters using LG’s model numbers appear. We treat this as a certificate-holder naming gap, not as evidence of anything about the filters.
• Whether a model’s NSF/ANSI 53 listing names lead specifically, and whether that certified claim matches a measured result. Every 53 finding here records the standard held, not the model’s full reduction-claims list or a lab result.
• How many of the 45 filters are being run past rated capacity. No capacity or replacement-interval column exists in our dataset.
• When a refrigerator filter is the wrong device for a lead problem. Our files carry the certification record, not authoritative guidance on service lines or whole-home conditions that call for point-of-entry treatment.
• Who manufactures Walmart’s Great Value house-brand fridge cartridges. The contract manufacturer could not be identified from the public listing, so the “NSF/ANSI 42 & 53 Listed” title could be neither confirmed nor refuted.
• Any named enforcement action against a refrigerator filter model. NSF operates a public notices process and all three bodies run surveillance programmes, but no specific, dated, sourceable de-listing of a fridge filter was found in this pass.
• The residual seven filters (16%) marked could-not-verify. After a clean sequential re-check these remained genuinely unresolved, which is not the same as uncertified. We keep those categories separate throughout the dataset.

Where to go next

If your next question is whether anyone has actually measured lead removal, our independent lead-testing study answers it and lists the gaps. If it is how many filters fail on their claims generally, the 45-filter counterfeiting and certification study has every row, every source URL and the analysis script. If it is which filter fits your fridge, the fitment index covers 50 models. And our method and affiliate disclosure explains how the verification was run and how the site is funded.