Heavy metals are the least glamorous topic in this industry and the one that ends brands fastest. A single failing lot, discovered by a retailer or a plaintiffs' firm rather than by you, is enough to trigger a recall, a delisting and a settlement that dwarfs the profit on the entire product.
It is also one of the few quality topics a brand can actually control from the outside, because the data is objective and the testing is cheap relative to the risk.
Here is what is being measured, where it comes from, and how to read the number on the page.
What "heavy metals" means on a specification
In a supplement context, four elements matter almost every time:
- Lead (Pb) - the most litigated, the most common finding in botanical ingredients
- Cadmium (Cd) - accumulates in cocoa, rice, some leafy botanicals, and in shellfish-derived ingredients
- Arsenic (As) - inorganic arsenic is the concern; found in rice-derived ingredients, kelp and some clays
- Mercury (Hg) - mainly a fish oil and marine ingredient issue
Aluminium and tin appear occasionally, mostly in relation to packaging or processing equipment. Some buyers add nickel and chromium. The four above are the core panel.
Where they come from
Almost never from the factory adding them. The sources are upstream:
- Soil uptake. Plants take up elements from the ground they grew in. This is why the same botanical from two regions can differ by an order of magnitude.
- Processing equipment. Old stainless steel, worn milling surfaces, lead-containing alloys in valves and fittings.
- Excipients and carriers. Some clays, talcs and mineral-based excipients carry a load of their own.
- Water and utilities. Less common now, but a real variable in older facilities.
- Cross-contamination. Shared lines, shared drying trays, shared storage.
- Deliberate adulteration. Rare but not unknown in powdered botanicals sold on price alone.
Note what is missing from that list: the capsule shell, the bottle, the label. Contamination is a raw material problem about nine times out of ten, and it is the reason the cheapest ingredient in a formula is rarely the cheapest decision.
Which ingredients carry the most risk
This is the table worth memorising before you choose a formula.
| Ingredient group | Main risk element | Why |
|---|---|---|
| Turmeric and curcumin | Lead | Soil, and historically adulteration with lead chromate for colour |
| Ashwagandha and other root botanicals | Lead, cadmium | Soil uptake, root concentration |
| Cocoa and dark chocolate flavoured powders | Cadmium | Uptake by the cacao plant |
| Rice protein, rice starch, rice syrup solids | Inorganic arsenic | Paddy field uptake |
| Kelp, spirulina, chlorella, seaweeds | Arsenic, mercury | Marine accumulation |
| Fish oil, krill oil | Mercury, PCBs | Marine food chain |
| Shellfish-derived glucosamine and chondroitin | Lead, arsenic | Shell and tissue |
| Bentonite and other clays | Lead, arsenic | Naturally occurring |
| Calcium from oyster shell or bone | Lead | Source material |
| Mineral premixes | Lead, arsenic | Concentrated geology |
The pattern is consistent: if the ingredient came out of the ground or out of the sea, test it. If it came out of a chemical reactor to a pharmacopoeial spec, the risk is lower.
What the limits actually are
There is no single global number, and this is where most brands get caught.
- Pharmacopoeial standards, such as USP general chapter on elemental impurities in dietary supplements, set limits by element and by daily dose. They are the reference most contract manufacturers work to.
- California Prop 65 sets a maximum allowable dose level for lead that is far stricter than most pharmacopoeial limits, and it is enforced by private litigation rather than by an agency.
- The EU regulates contaminants through maximum levels in food and applies them to food supplements by extension.
- Your customer's specification can be stricter than all of the above, and often is.
That means "our product passes heavy metals" is not an answer. The answer is: passes which limit, for which element, at what serving size, using which method.
The per-kilogram trap
Results are usually reported in mg/kg, which is the same as parts per million. Your exposure is per serving. These are not the same thing, and the difference is where brands get confused in both directions.
| Reported result | Serving size | Per serving | Versus a strict lead limit |
|---|---|---|---|
| 0.5 mg/kg | 500 mg | 0.25 micrograms | Comfortably inside |
| 0.5 mg/kg | 5,000 mg | 2.5 micrograms | Inside pharmacopoeial, near the tighter state limit |
| 2.0 mg/kg | 500 mg | 1.0 micrograms | Needs a closer look |
| 2.0 mg/kg | 5,000 mg | 10 micrograms | Likely a problem |
The same powder is compliant in a 500 mg capsule and non-compliant in a 5 g scoop. This is why a test result cannot be judged without the dose, and why reformulating the serving size is sometimes the cheapest fix available.
How the test is done, and why "not detected" is a phrase to distrust
The standard method is inductively coupled plasma mass spectrometry, usually written ICP-MS. It is sensitive, fast and now cheap enough that there is no excuse for skipping it.
Three details decide whether the number means anything:
- Sample preparation. A closed-vessel acid digestion done properly is not the same as a quick open digestion. Ask which method was used.
- Limit of detection and limit of quantification. "Not detected" often means "below the detection limit of the method we used", which may be far above the level your customer cares about. Ask for the LOQ in mg/kg.
- Which lab. An accredited third-party lab, not the factory's own bench, not a screenshot from a supplier.
A result of "none detected" with no stated LOQ is not a result. It is a sentence.
Reading a certificate of analysis properly
A heavy metals line on a certificate of analysis should give you five things: element, method, result with units, specification limit, and a pass or fail against that limit. If any of those is missing, the line is decoration.
Two more checks that catch most problems:
- Is the lot number on the certificate the lot number on the drums? Certificates get recycled. Match them.
- Is the date of analysis plausible? A certificate dated before the material was manufactured is a copy from another batch.
There is a full walkthrough of the rest of the document in our guide to the certificate of analysis, including the sections most buyers never read.
Batch to batch: why one passing lot proves very little
Botanical raw materials vary by harvest, region and season. A supplier who passed in March can fail in September without changing anything, because the crop changed.
What actually reduces risk:
- Qualify the supplier, not just the lot. Ask for three consecutive recent certificates for the same material, not one.
- Test every lot of high-risk ingredients yourself, at least for lead and cadmium. It costs a few hundred dollars per lot and it is the cheapest insurance in the business.
- Keep retained samples. If a customer complains in eighteen months, you need the actual material, not a certificate.
- Track the trend, not the pass. A supplier drifting from 0.2 to 1.4 mg/kg over four lots is telling you something before they fail.
If a lot fails
In order of preference:
- Reject it. This is why the purchase order should make acceptance conditional on your own testing, not on the supplier's certificate.
- Ask for a retest at a second accredited lab, if the result is close to the limit and the sampling was questionable. Sampling error is real, especially in powders.
- Do not blend it down. Diluting a failing lot with a clean one to hit an average is not a compliant strategy, and in most markets it is itself a violation.
- Change the source, or change the dose. A different origin for the same botanical can be an order of magnitude cleaner, and a smaller serving size reduces exposure proportionally.
- Reformulate. If the ingredient cannot be made clean at a commercial price, it should not be in the product.
Adding a Proposition 65 warning label is a legal option in California and a commercial mistake almost everywhere else. Retailers treat it as a flag, and once it is on the label, it is hard to remove.
Eight questions for any supplier
- Which elements do you test for, and to which limit?
- Which laboratory, and is it accredited to which standard?
- What is the limit of quantification for lead and cadmium in this matrix?
- Do you test every lot of botanical raw material, or only on qualification?
- Can you send the last three certificates for this exact material?
- Is the certificate lot-specific, and will it match the drums I receive?
- What is the specification for the finished product, per serving?
- What happens if my own test disagrees with yours?
A manufacturer who can answer all eight without hesitation has a quality system. One who answers five and changes the subject has a sales team. More of the questions buyers ask us most often are answered in our FAQ.
Frequently asked questions
Is heavy metal testing required by law?
Finished product testing is not universally mandated in every market, but importers, retailers and state-level rules effectively require it, and the liability sits with the brand whose name is on the label either way.
How much does a heavy metals panel cost?
As a rough guide, a four-element ICP-MS panel on a finished product runs from a few hundred dollars per sample at a contract lab. Compared with a recall, the number is not material.
Can a supplement be completely free of lead?
No. Trace levels occur naturally in soil and water, which is why the standards are limits rather than zero, and why "not detected" almost always means "below the method's detection limit".
Which is riskier, a capsule or a gummy?
The format is not the issue; the ingredient load is. A gummy with a high dose of a root botanical carries more risk than a capsule with a synthetic vitamin, simply because of what is in it.
Do organic certifications remove the risk?
No. Organic certification restricts agricultural inputs, it does not test for elemental contaminants in the finished product.
The takeaway
Heavy metals are a raw material problem, measured per serving, judged against limits that differ by market. Test every risky lot, demand lot-specific data with a stated method and limit of quantification, and treat a passing certificate as the beginning of your due diligence rather than the end of it.
Tell us your formula and target market and we will tell you which ingredients in it need lot-by-lot testing, and what specification to write into the purchase order before the first batch is made. If you are still choosing between formats, start with our private label range, where the specification is written before the price is quoted.
