Science commentary · 2026-11-11 · Updated 2026-11-11
How fresh should a protein powder COA be? Lot testing explained
FDA dietary-supplement CGMPs require contamination specifications and finished-batch verification every batch or via a sound statistical sampling plan — not a single forever COA. How to match lots, why results drift, and a buyer checklist.



The short answer
A useful protein-powder certificate of analysis (COA) is lot-matched, quantitative for lead, cadmium, arsenic and mercury, and recent enough to describe current production — not a three-year-old PDF recycled on a marketing page [1][2]. U.S. dietary-supplement CGMPs require manufacturers to set contamination specifications and to verify finished batches every batch or through a sound statistical sampling plan [1][2][3]. There is no federal calendar that says “retest heavy metals every 6 months” for every powder. Prefer brands that publish µg per serving for the code on your tub — the same evidence our rankings score [4]. Label literacy, not medical advice.
Why “we test for heavy metals” is not a date
Front-of-pack seals and a solitary “heavy metals tested” badge answer a binary question. Shoppers need a time-stamped, lot-coded number. Ingredient geography, cocoa lots, plant-protein harvests and supplier changes can move results without changing the brand name on the lid — category screens already show wide product scatter for lead [5][6]. A COA without a lot, a date or units you can convert is a brochure, not a control chart.
This page covers freshness and retesting cadence. For ND vs LOQ, ppm→µg maths and red-flag layouts, use the companion how to read a COA. For why the printed best-by date is mostly about quality, see shelf life and storage.
What 21 CFR 111 actually requires
| Rule | Shopper translation |
|---|---|
| 21 CFR 111.70 | Brands must establish specifications, including limits on contamination that may adulterate the finished supplement [1] |
| 21 CFR 111.75(c) | For finished batches, verify product specs — including those contamination limits — on every finished batch or on a subset chosen through a sound statistical sampling plan [2] |
| Federal Register final rule (21 CFR part 111) | Establishes the DS CGMP framework manufacturers must follow for quality and contamination controls [3] |
| Prop 65 MADLs (our scoring yardstick) | Oral lead 0.5 µg/day, cadmium 4.1 µg/day — warning safe harbors we use on cards, not a federal “pass” stamp [4] |
Read carefully: the CGMP text does not invent a universal “test metals every Monday” schedule, and it does not name lead µg limits for every protein powder. It demands written specs plus verification that an investigator can audit [1][2][3]. Voluntary industry guidance (for example AHPA’s elemental impurity policy for herbal products) offers optional daily µg ceilings some brands adopt — useful context, not a substitute for your tub’s lot sheet [7].
Why lot-to-lot metals can move
Agricultural and mineral inputs are volatile; synthetic add-ins usually are not. Plant proteins and cocoa track soil and processing dust stories that show up in category averages — CR’s plant powders averaged about nine times dairy lead; CLP’s protein study reported higher over-Prop-65 rates for plant and organic powders [5][6]. Dairy whey still needs lot evidence: CR’s better daily dairy options sat under 0.5 µg lead per serving in named panels, but that does not freeze every future lot [5][8].
Change-control triggers that justify fresh testing (from CGMP logic, not from invented brand schedules):
- New supplier or growing region for pea/rice/cocoa
- Formula or flavour change (chocolate ≠ vanilla)
- Process change at the co-packer
- Out-of-trend or failing result on a prior lot
- Large batch that becomes a big share of annual volume
Seals such as NSF Certified for Sport answer banned-substance and program-limit questions; they are not the same as a Prop 65–style 0.5 µg lead bar (third-party seals) [5][10].
How fresh is “fresh enough” for a published COA?
Match the lot first; then look at the test date. Ideal order:
- Lot/batch code on the COA equals the code on your tub (or the brand publishes a clear lot index you can look up) — see how to read a COA.
- Test date sits close to that lot’s production — not a multi-year orphan PDF.
- Four metals with numbers or ND/<LOQ plus the detection/quantitation limits.
- Units convertible to µg per labeled serving.
- Named lab / method (ICP-MS / AOAC-style elemental methods are the usual bar; CR’s powder work used AOAC 2015.01) [5][9].
If the brand only posts one evergreen COA with no lot, treat it as a sample of capability, not proof for the tub in your cart — same rule we use when store brands lack publishable panels (store-brand lead).
We do not invent a legal maximum age such as “90 days.” Prefer every-lot publication when metals are your main purchase filter; a recurring multi-month cadence is better than silence, worse than lot-matched sheets.
Buyer checklist before you trust a metals claim
- Open the COA PDF — not a blog summary.
- Match product + flavour + lot.
- Convert to µg/serving and compare with Prop 65 MADLs if that is your yardstick [4].
- Confirm the brand’s public sheet is the same evidence class we score (brand-published quantitative metals) on rankings.
- Keep a photo of the lot code; storage does not add lead, but recalls and COAs need the code (shelf life).
- Remember amino-spiking literacy is a different checklist (amino spiking labels).
| Trust signal | Weak signal |
|---|---|
| Lot-coded four-metal ICP-MS panel | “Third-party tested” with no µg table |
| Dated within the production window of your lot | Single PDF reused across years/flavours |
| µg/serving or convertible ppm/ppb + serving size | “Pass/fail” only |
| Brand publishes failures/retests policy | Metals mentioned only in ads |
Examples on our board where lot matching already matters: Puori PW1 (metals 10.0), Momentous Essential Whey (10.0), Natural Force Organic Whey Unflavored (scored unflavored lot 2605212 — flavours differ), Legion Whey+ (9.4).
More guides in the journal. As an Amazon Associate I earn from qualifying purchases.
Bottom line
CGMPs require contamination specifications and finished-batch verification every batch or via a documented statistical plan — not a forever homepage badge [1][2][3]. Freshness means lot match + usable date + µg maths, scored the same way as our rankings [4]. Category lead scatter in CR and CLP is why one old PDF is a weak shopping tool [5][6].
FAQ
Does FDA require every protein powder lot to be tested for lead?
Finished dietary-supplement batches must be verified against product specifications — including contamination limits the firm established — every batch or on a statistically justified subset [2]. The 2007 DS CGMP final rule is the Federal Register source for that part 111 framework [3]. That is not identical to a public lead COA on every Amazon listing.
Is a six-month-old COA automatically invalid?
Not automatically. First ask whether it matches your lot. A recent PDF for the wrong flavour still fails the usefulness test.
Can I rely on NSF or Informed Sport instead of a metals COA?
Those programmes are valuable for their own scopes; they are not a drop-in replacement for Prop 65–style µg-per-serving shopping (seals explainer) [5][10].
Do metals get worse as the tub sits in my pantry?
No — elemental metals do not accumulate from dry storage time (shelf life).
What if a brand refuses to share lot data?
Treat metals as unverified for scoring purposes (metals 0 on our board until a usable brand panel exists) [4].
Is this legal or medical advice?
No. It summarises cited CGMP text and shopper COA practice.
References
- eCFR. 21 CFR § 111.70 — What specifications must you establish?
- eCFR. 21 CFR § 111.75 — What must you do to determine whether specifications are met?
- Federal Register. Current Good Manufacturing Practice in Manufacturing, Packaging, Labeling, or Holding Operations for Dietary Supplements (final rule PDF, 25 Jun 2007)
- OEHHA. Current Proposition 65 NSRLs and MADLs
- Consumer Reports. Protein powders and shakes contain high levels of lead (2025)
- Clean Label Project. Protein Study (January 2025)
- AHPA. Guidance policy — heavy metals (elemental impurities) PDF
- Consumer Reports. These five protein powders had low lead levels (January 2026)
- Consumer Reports. Protein powders and shakes methodology / test results PDF (2025)
- NSF. Sport supplement safety white paper (NSF/ANSI 173 overview PDF)
Related articles
Protein powder without heavy metals: what “zero metals” actually means
2026-11-15
No lab can prove a scoop is metaphysically metal-free. ND and below-LOQ results still beat vague “clean” claims — here are the powders that publish the tightest brand panels against Prop 65 MADLs.
Protein bars vs powder: which format is cleaner on heavy metals?
2026-11-11
Clean Label Project’s 2025 snack/nutrition bar study found detectable heavy metals in every bar tested, with about 22% over Prop 65 lead. Consumer Reports’ powder/shake round is a different format. When a scoop with a lot COA is the cleaner daily default.
UK and EU heavy metal limits for supplements: what applies to protein powder
2026-11-08
EU Regulation 2023/915 and GB assimilated 1881/2006 set lead, cadmium and mercury ceilings for food supplements in mg/kg. How those convert to µg per scoop, how they compare with Prop 65, and a UK buyer checklist.
As an Amazon Associate I earn from qualifying purchases. Affiliate status never changes rank. 0 paid placements.