USP <2232> vs. <232>/<233>: Which Elemental Impurity Standard Actually Applies to Your Botanical Raw Materials

Every few months, a quality manager at a Midwest botanical brand sends us a supplier COA that lists ICP-MS results for lead, arsenic, cadmium, and mercury — all under the stated limits — with a footnote that reads “tested per USP <232>/<233>.” On paper, it looks solid. In practice, there’s a problem hidden in plain sight.

USP <232> and <233> were not written for dietary supplement ingredients. They were written for pharmaceutical drug products, and they apply limits derived from a risk model that doesn’t account for the naturally occurring elemental chemistry of botanical raw materials. When those chapters appear on a botanical supplement COA, it usually means one of two things: the supplier’s analytical testing lab used the right technique but the wrong reference framework, or your incoming specification was drafted by someone more familiar with pharma GMP than with DSHEA. Either way, the result is a specification that’s either needlessly restrictive or quietly inadequate — and it won’t be obvious which until something goes wrong.

How USP <232> and <233> Actually Work

USP <232> — “Elemental Impurities—Limits” — establishes permitted daily exposures (PDEs) for 24 elements, grouped by toxicological concern. Class 1 elements (arsenic, cadmium, lead, mercury) carry the highest concern and the strictest oral PDEs. Class 2 and Class 3 elements are monitored based on route-of-exposure risk and manufacturing likelihood.

These PDEs were derived through ICH Q3D, a guideline developed jointly by regulators in the US, Europe, and Japan specifically for pharmaceutical drug products. The analytical procedures in USP <233> — primarily ICP-MS, though ICP-OES is also recognized — are the validated methods for making those measurements.

None of that is controversial. The analytical science is rigorous, and ICP-MS is the appropriate tool for this work regardless of which framework you’re operating within. What changes is what you do with the numbers once you have them.

For a finished oral drug product, the oral PDE for lead under ICH Q3D and USP <232> is 5 μg/day. Work that backward through the expected daily dose of your drug product, and you get a concentration limit in ppm for each ingredient at its intended contribution level. That’s a defensible model for a pharmaceutical product where synthetic or semi-synthetic ingredients have relatively predictable, narrow elemental profiles.

Botanical ingredients are not predictable in that way — and that’s the entire problem.

What USP <2232> Was Built For

USP <2232> — “Elemental Contaminants in Dietary Supplements” — is a general information chapter. The ≥2000 numbering in the USP system signals this: chapters in that range are informational guidance, not compendial requirements. But “informational” doesn’t mean optional in a practical quality sense. FDA’s 2019 draft guidance on elemental impurities in dietary supplements explicitly references USP <2232> as the appropriate scientific framework for risk-based evaluation, and that guidance reflects the agency’s clear direction for manufacturers building defensible incoming specifications.

The chapter was developed to address something USP <232>/<233> structurally can’t: the reality that botanical ingredients accumulate elements from soil, irrigation water, and agricultural inputs, and that “contamination” looks fundamentally different when the source is geochemistry rather than manufacturing process.

A few distinctions that matter day-to-day:

Risk assessment starts with dose, not concentration. USP <2232> frames limits as daily intake values tied to the actual serving size of the finished product. If an ashwagandha extract is dosed at 600 mg/day in one formula and 1,200 mg/day in another, the maximum allowable elemental concentration in the raw material is different in each case. An analytical testing laboratory working within the <2232> framework needs the finished product’s intended use before it can evaluate whether a raw material clears spec — not just an ICP-MS result in isolation.

Naturally occurring elements are treated differently. Marine botanicals — spirulina, chlorella, kelp — routinely accumulate total arsenic at 10 to 50 mg/kg. But the vast majority of that arsenic is arsenobetaine and other organic species that carry negligible human toxicity risk. USP <2232> explicitly calls for arsenic speciation when the ingredient type makes that distinction meaningful. A COA reporting “total arsenic: 2.4 ppm, pass” against a threshold back-calculated from a USP <232> PDE may look fine for a pharmaceutical excipient. For a kelp ingredient, it’s the wrong question answered correctly.

Daily intake values differ from pharmaceutical PDEs. For Class 1 elements, the daily intake targets under USP <2232> are established through a separate risk assessment process — one that accounts for typical supplement use patterns, background dietary exposure, and the reality that supplements are consumed voluntarily and intermittently rather than dosed precisely like drugs. These values are not identical to the ICH Q3D-derived PDEs in USP <232>. Using the wrong limit set can flag clean material as a failure or, in the other direction, allow marginal material to pass unexamined.

Two Scenarios Where the Wrong Chapter Creates Real Problems

Scenario 1: Rejecting material that’s actually acceptable. A supplier ships ashwagandha root powder with a lead result of 0.8 ppm. Your incoming specification, drafted from a pharma reference and using the USP <232> oral PDE of 5 μg/day divided against a conservatively assumed 10 g/day dose, calculates a passing threshold of 0.5 ppm. The material is rejected. But if the finished product uses that ingredient at 600 mg/day — which is typical for a standardized ashwagandha extract — and USP <2232> is applied correctly with that dose, the allowable lead concentration in the raw material works out to approximately 8 ppm. The material you rejected at 0.8 ppm passed at a margin of more than ten to one. You’ve created a supply disruption, a supplier relationship problem, and a qualification delay, all for a specification error that originated at your desk, not their factory.

Scenario 2: Approving material that warrants closer scrutiny. A supplier’s COA shows total arsenic at 1.2 ppm for a kelp powder, listed as passing against a <2232>-referenced daily limit. Your quality team accepts the material for a product used at 3 g per serving, two servings per day. At that dose, 1.2 ppm total arsenic translates to 7.2 μg/day total arsenic exposure — seemingly under most reference values. But nobody requested arsenic speciation. If that 1.2 ppm reflects a higher-than-typical proportion of inorganic arsenic — possible in kelp from certain Pacific growing regions — the inorganic arsenic exposure could meaningfully exceed risk thresholds, and the organic/inorganic split is the only way to know. The total number passed; the right number was never measured.

Both scenarios stem from the same place: the limit framework and the analytical scope weren’t matched to the ingredient’s chemistry and the product’s intended use.

What a Properly Structured Raw Material COA Should Include

When botanical raw material samples arrive at our Chicago receiving facility and are routed to the ISO 17025-accredited analytical testing laboratory for ICP-MS analysis, the reports we generate specify four fields that supplier COAs frequently omit:

The analytical method in detail. ICP-MS per USP <233> or a validated equivalent, with internal standard recoveries and instrument performance documented. Recovery data isn’t optional — it’s how you confirm the matrix didn’t suppress or enhance the signal.

The limit framework explicitly stated. Which chapter or guidance was used to set the acceptance criteria — USP <2232>, a manufacturer-specified limit derived from dose calculation, or a regulatory threshold such as California’s Proposition 65 safe harbor levels. “Tested per USP <232>/<233>” is a method citation, not a limit citation. These are two different things.

The dose assumption used in any back-calculation. If limits were derived from a daily intake value, the serving size and daily serving frequency assumed must be stated in the COA. A limit without a dose assumption is a number without context.

Speciation data where the ingredient type warrants it. For arsenic in marine botanicals and high-accumulating terrestrial species. For chromium in high-dose botanical proteins. For selenium in Brazil nut extracts or selenium-enriched ingredients, where the difference between selenomethionine and inorganic selenium matters enormously for both safety and efficacy claims.

Without all four fields, a COA is a data sheet, not a quality release document.

What This Means for Your Incoming Specification Program

If you’re building or auditing incoming raw material specifications for botanical ingredients, three questions are worth working through before your next supplier review:

Does your specification reference the right framework? If botanical ingredients are currently spec’d against USP <232>/<233> without a dose-adjusted back-calculation, those specifications need review. For most botanical raw materials used in finished dietary supplements, USP <2232> — applied with the actual serving size and daily serving frequency of the finished product — is the appropriate scientific reference.

Have you defined which elements require speciation testing? Total arsenic in a marine botanical is not the same toxicological event as inorganic arsenic. Your raw material spec should explicitly require speciation for any ingredient where the organic/inorganic ratio meaningfully changes the risk picture. Marine botanicals are the obvious case, but it applies to others depending on your supplier geography and ingredient type.

Does your analytical testing lab understand the distinction operationally? Not every contract lab that offers ICP-MS for dietary supplement materials routinely applies USP <2232>. Labs that primarily serve pharmaceutical clients often default to <232>/<233> because that’s the framework their pharma customers require — and they may not flag it unless you ask directly. Confirm that your lab can apply dose-specific limit calculations and document the dose assumptions used, rather than just reporting a concentration against a fixed threshold.

The underlying ICP-MS science is the same regardless of which chapter you reference. What changes is the interpretive layer — and that layer is what makes the difference between a specification that’s scientifically defensible and one that looks rigorous until a regulatory auditor asks how the limits were calculated.

A COA with four clean ICP-MS numbers is a starting point for a conversation, not a release decision. Make sure you know which chapter governed the limits before you sign the release.

---

Written by Nour Abochama, VP Operations, Qalitex | Quality Consultant, Ayah Labs. Learn more about our team

Ship your sample to our Chicago facility — get a Qalitex CoA in 5–7 days. Contact us

Related from our network

• ISO 17025-Accredited Supplement Ingredient Testing at Qalitex Laboratories — Full-spectrum elemental impurity and contaminant testing for finished supplements and raw materials, with ICP-MS results backed by ISO 17025 accreditation.