Heavy metals in supplements: set the limit, choose a suitable method and record who decided
Write down which elements you limit, the reference behind each number, the method's quantitation limit and who decided.
In Short
- Part 111: you set the limits.
- Per-gram limits need the daily dose.
- Method must suit its purpose.
The Dietary Supplement Operations Guide
What the label says. What the batch proves.
- The rules and who decides
- Specification to finished batch
- Test, review, investigate
We ask for your name and email so we can handle this request. The full guide then opens on this site, ready to translate, print or save. Privacy policy. Practical orientation, not legal advice; your quality unit decides which rules apply.
UltimateFree guide
The Dietary Supplement Operations Guide
What the label says. What the batch proves.
Quality and laboratory leads at supplement manufacturers must decide which contaminants to limit, at what level and how to test. Opened (current eCFR, up to date 1 October 2026): 21 CFR 111.70(b) and (c), 111.75(c) and 111.320. The sections read set no numeric heavy-metal limits; that does not mean no other limit applies to your product. Not opened: USP chapters on elemental contaminants, ICH Q3D, FDA's Closer to Zero work, California Proposition 65, the rest of Part 111 and any retailer or certifier limit. It is orientation, not legal advice. Sources checked 5 Oct 2026.
Where the duty comes from
21 CFR 111.70(b) requires component specifications: an identity specification, specifications needed to meet the purity, strength and composition of the finished supplement, and, in 111.70(b)(3), limits on those types of contamination that may adulterate or may lead to adulteration of the finished batch. The sections read do not name lead, arsenic, cadmium or mercury.
More on where the duty comes from
Many teams treat heavy metals as one contamination type to assess; deciding which elements to limit, for which components, is your quality team's call and should be written down with its reasoning.
Choosing the Numbers
The sections read set no numbers, so the values come from references your team selects: pharmacopeial chapters, intake-based approaches, customer specifications, certifier limits or state law. None of these was opened here.
Worked example: per-day exposure for capsule SP-221 (1.5 g/day)
| Element | Result (µg/g) | Per day (µg) |
|---|---|---|
| Lead | 0.20 | 0.30 |
| Cadmium | 0.10 | 0.15 |
| Arsenic (total) | 0.40 | 0.60 |
| Mercury | 0.02 | 0.03 |
- Lead
- Result (µg/g)
- 0.20
- Per day (µg)
- 0.30
- Cadmium
- Result (µg/g)
- 0.10
- Per day (µg)
- 0.15
- Arsenic (total)
- Result (µg/g)
- 0.40
- Per day (µg)
- 0.60
- Mercury
- Result (µg/g)
- 0.02
- Per day (µg)
- 0.03
More on choosing the Numbers
Record the reference, its version and date, and how a per-gram result converts to per-day exposure at the labelled serving. Component origin, ingredient type and daily dose all affect the outcome. When references disagree, record which one governs and who chose it.
Finished-Batch Verification
111.75(c) requires you to verify that finished batches meet product specifications, including limits on contamination, for a subset of batches identified through a sound statistical sampling plan, or for every batch. You select one or more established specifications that, if tested or examined on finished batches, would verify that your production and process control system is producing product that meets all specifications.
Which finished batches to test? (111.75(c), simplified)
- Test selected specifications on each batch.
More on finished-Batch Verification
Write down why the heavy-metal specifications you test were chosen, and how the sampling plan was justified.
Method Suitability
111.320(a) requires you to verify that laboratory examination and testing methodologies are appropriate for their intended use; 111.320(b) requires an appropriate, scientifically valid method for each specification that needs testing. ICP-MS is one technique laboratories use for elements; the sections read do not name any technique.
Method record (21 CFR 111.320)
More on method Suitability
Record the method, its limit of quantification against your limit, sample preparation for the matrix, the laboratory, and the verification evidence. A limit of quantification above the limit cannot show the limit is met.
When a result is over the limit
A component or batch result above a limit needs a recorded decision by quality under your written procedures. The Part 111 sections on rejection, reprocessing and material review were not opened for this guide, so they are not summarised here.
More on when a result is over the limit
Link each result to its lot, supplier, method and reviewer, so trends by supplier or ingredient are visible. Review limits when a formula, serving size, supplier or growing region changes, because the per-day exposure changes with them.
Questions
Does 21 CFR 111 set numeric heavy-metal limits?
The sections read here, 111.70, 111.75(c) and 111.320, set none. They require you to establish contamination limits; other laws or customers may set numbers.
Do I have to test every finished batch for heavy metals?
111.75(c) allows verifying a subset identified through a sound statistical sampling plan, or every batch, against specifications you select and justify.
Is ICP-MS required?
The sections read require an appropriate, scientifically valid method verified for its intended use. They do not name a technique.
Does V5 set limits or release batches?
No. V5 records specifications, results against lots and review decisions. Your quality team sets limits and decides disposition.
Practical orientation, not legal advice. Rules differ by jurisdiction; your quality and compliance team decides what applies.
Keep reading
Supplement Allergen Control: Validated Cleaning and Labelling
ASEAN Supplements Readiness: Singapore HSA and the Regional Hub
Australian Listed Medicines: What the Sponsor Certifies Under Section 26A
Bakery Label and Date Code Control: a Practical US Guide
Bakery Pack Weight Control: Net Weight, Tare, and Sampling
Bakery Recipe Control: Versions, Scaling, and Batch Evidence