Elemental Impurities Under ICH Q3D: Where They Come From

Ask for an elemental impurities statement on a bile-acid API and you will usually get one of two documents. One is a heavy metals result, carried over from an old specification template, with a single number on it. The other is a risk assessment that names specific elements and says where each of them comes from. The second document is the useful one, and it is not the same kind of document as the first.

Here is what changed, why the change lands differently on an animal-derived material than on a fully synthetic one, and the question I would ask before trusting either piece of paper.

The heavy metals test is gone, and it is not coming back

The old test dates from 1905. It worked by precipitating sulfides and comparing the colour of the sample solution against a lead standard. One line on the specification, "heavy metals, not more than 20 ppm", was expected to cover everything from arsenic to mercury to whatever else happened to form a coloured sulfide in that tube.

Three problems eventually made it untenable. It did not tell you which element was present. Its recoveries were erratic — published comparisons found mercury coming through at a few percent of the true value. And the limit itself was set by what the test could achieve rather than by toxicology.

So it was withdrawn. The USP omitted its heavy metals chapter on 1 January 2018 and replaced it with element-specific limits and procedures. The European Pharmacopoeia rewrote its general chapter on metal catalyst and reagent residues into a text on elemental impurities, in force from January 2018, added a matching general method for the determination, and deleted the reference to the old heavy metals method from monographs for human use. The Chinese Pharmacopoeia moved in the same direction in its 2020 edition, adding a general chapter on elemental impurity limits and determination built on the ICH Q3D framework. One exception is worth noting: the European heavy metals method was kept for veterinary monographs, so it has not disappeared everywhere.

What replaced it is a risk assessment, not a limit

ICH Q3D sets permitted daily exposures for 24 elements, separately for the oral, parenteral and inhalation routes, and then hands whoever makes the medicine a job: work out which of those elements can plausibly be present in your product, and justify the conclusion. The guideline contains limits. It does not contain the limit for your material, because that only exists once somebody has done the assessment.

The elements are grouped by toxicity and by how likely they are to turn up:

A further group — aluminium, boron, calcium, iron, potassium, magnesium, manganese, sodium, tungsten and zinc — carries no assigned permitted exposure at all. They are covered by other guidance, which is not the same as being ignored.

The guideline then allows four ways of doing the arithmetic: a common concentration limit applied to every component, a component-by-component approach, or direct measurement of the finished product. Any of these can be the right answer for a given product. What the framework does not permit is skipping the reasoning and going straight to a test.

One number inside the framework does more practical work than the rest. If the assessment shows an element consistently below 30% of its permitted exposure, and the data behind that conclusion hold up, no further control is needed. Get above it and a control strategy becomes necessary. That threshold is the line between a document and a testing programme, which is why risk assessments for the same molecule can look so different from one supplier to the next.

Where the elements actually come from

For most synthetic APIs the source analysis is short. Catalysts and reagents in, stainless steel and water around, and that is close to the whole list.

A bile-acid API made from an animal-derived starting material has a second source, and it behaves in a fundamentally different way. The raw material brings elements with it, because the animal took them in from feed, water and the environment it was raised in. That is the pathway by which cadmium, lead, arsenic and mercury — the four Class 1 elements — get into a natural-origin material. No purification step removes a burden that arrives with the starting material and is spread evenly through it, and no amount of process control at the end compensates for it. It has to be managed where it enters, which makes the Class 1 question for a natural-origin bile acid a supply-chain question before it is an analytical one.

The process supplies the rest. Where the 7-position is set by a chromium-based oxidation — the conventional way to install that ketone, as described in our note on what the synthesis route does to the impurity profile — chromium belongs in the assessment, and nickel follows from the reduction system. Where catalytic hydrogenation is used, the platinum-group metals come into play. Equipment contributes chromium and nickel from stainless steel. Water and raw materials contribute whatever the supplier's own specifications allow. Copper, barium and tin are Class 3 elements for the oral route and will usually be excluded, but they still need a line in the assessment explaining why they were excluded rather than a blank.

That split — brought in by the material, brought in by the process — is the most useful thing to carry out of this whole area. It tells you which question goes to a supplier's quality department and which goes to their process engineers. It also tells you what a change of supplier will move and what it will leave alone.

Chromium: the loosest limit that still needs attention

Chromium looks like a contradiction, which is why it is worth understanding.

Its oral permitted daily exposure is 11,000 micrograms per day, the highest of any classified element and more than two thousand times the limit for cadmium. It also sits in Class 3, which for an oral product normally means it is not evaluated at all.

But if the synthesis uses a chromium-based oxidant, chromium was added deliberately. That single fact pulls it into the assessment. A generous limit changes the conclusion, not the obligation to reach one.

There is a second layer underneath. The permitted exposure was derived from trivalent chromium. Hexavalent chromium is a different toxicological proposition, and a total-chromium result from an ICP instrument does not distinguish between the two. The guideline's own position is that speciation is not required and total element content is what gets compared against the permitted exposure — which is a defensible position, and also the reason a supplier whose route runs through a chromium oxidant should be asked whether anyone has thought about oxidation state. The answer says something about the file behind the certificate.

Nickel, and why equipment counts as a source in its own right

Nickel belongs to Class 2A, with a permitted daily exposure of 200 micrograms per day. It arrives from two directions at once: as a component of metal-catalysed reduction systems, and as a constituent of the stainless steel that every reactor, centrifuge, filter and dryer in the plant is made from.

That double source makes nickel a good test of whether an assessment was written or copied. A real one names the alloy grades and the contact conditions. A template says "manufacturing equipment: no contribution" and moves on.

The four Class 1 elements, in an animal-derived material

Cadmium, lead, arsenic and mercury have to be considered for every route of administration and every potential source. For a purely synthetic API that requirement is often discharged with a single sentence: no mined excipient, no metal-bearing reagent, therefore no plausible source.

That sentence cannot honestly be written for a material derived from porcine bile. The animal concentrates what it takes in, and what it takes in depends on feed, water and husbandry. The Class 1 risk for a natural-origin bile acid therefore has to be argued from the source upward — which origin, what surveillance of that origin, what the batch history looks like. Our note on dried pig bile powder origin and traceability deals with the document side of that conversation.

None of this means a natural-origin material carries more cadmium than a synthetic one. It means the pathway is different, the control point is upstream rather than downstream, and a supplier who cannot describe that control point has not finished the assessment.

A permitted exposure is not yet a specification

The most common misunderstanding, on both sides of a purchase order, is what these numbers apply to.

The limits are set for the medicinal product. An API is one component of it. That is why the European framework states plainly that individual monographs on substances for pharmaceutical use do not carry elemental impurity specifications unless the monograph says so itself, and why the Chinese guidance makes the same practical point — the limits are not directly applicable to APIs and excipients, but an API supplier's testing data and risk assessment are what the company making the finished product uses to demonstrate compliance.

Two things follow. When a supplier says they "meet Q3D", the accurate version of that claim is that they hold an assessment whose conclusions support the finished product's own. And the expectation on an API supplier is not zero: the European general monograph on substances for pharmaceutical use requires that elements deliberately added during production be controlled through risk management. So the obligation is real. It simply does not arrive as a single figure on a certificate.

The arithmetic behind a concentration limit is easy once the daily dose is known. Divide the permitted daily exposure by the grams per day. At the guideline's default assumption of ten grams per day or less, the numbers come out as follows:

Values from the guideline's oral permitted exposures and its Option 1 concentration table for products with a daily intake of not more than 10 grams. Confirm against the current revision before relying on them — the 2022 revision corrected the figures for gold, silver and nickel, and documents written earlier often carry the superseded values.

If the daily dose is higher than ten grams, the division is simply repeated with the real dose and the permitted concentrations tighten in proportion. If it is lower, they relax. This is also the point at which a procurement conversation clarifies itself: the same API can be perfectly acceptable in a low-dose product and a problem in a high-dose one, without anything about the API having changed.

Why the method belongs in the same conversation

Where the assessment concludes that testing is needed, the determination is ICP-MS or ICP-OES after closed-vessel digestion, because the quantities involved are nanograms per gram. That is a different order of measurement from the parts-per-million world the old colour test lived in, and it brings its own rules.

Two details show how unforgiving it gets. Mercury is volatile and adhesive: it can be lost during sample preparation and it lingers in the instrument, which is why mercury determinations use gold in the rinse solutions to stop it carrying over between samples. And several of these elements interfere with one another inside the plasma — an argon-carbon ion sits on top of the mass used for chromium, and an argon-chloride ion sits on the mass used for arsenic unless the instrument is configured to separate them.

This matters beyond the laboratory. An elemental impurity result is only as good as the method behind it, and a certificate line without a method reference or a validated procedure is worth less than it looks.

What to ask a supplier

• Ask for the assessment, not a statement about it. A one-page risk assessment naming elements and sources is the document. "Complies with ICH Q3D" is a conclusion with the reasoning removed.
• Ask which elements were excluded and on what evidence. The exclusions carry more information than the inclusions, because they are where a copied template shows through.
• Ask which conclusions depend on the manufacturing route. Chromium and the platinum-group metals fall in or out of the assessment depending on how the molecule is made, not on what the molecule is.
• Ask what daily dose the assessment was written against. An assessment that assumes ten grams a day is not valid for a product dosed at two grams, and vice versa.
• Ask what happens if the origin of the raw material changes. For a natural-origin material this is the question that decides whether the assessment is a living document or a filing.

Frequently asked questions

Does ICH Q3D apply to an API, or only to the finished product?

The limits are set for the medicinal product, and the API is one component of it. In practice an API supplier supports the finished product's assessment with testing data and a risk assessment of their own, and the European general monograph on substances for pharmaceutical use expects deliberately added elements to be controlled through risk management.

What replaced the heavy metals test?

Element-specific limits based on toxicology, together with a requirement to assess which elements can plausibly be present. The USP withdrew the colour-comparison heavy metals chapter on 1 January 2018; the European Pharmacopoeia replaced its metal-residue chapter with a text on elemental impurities in the same period and retains the old method only for veterinary monographs.

Why does chromium need attention when its limit is the highest?

Because a chromium-based oxidant means chromium was deliberately added during manufacture, which brings it into the assessment regardless of how generous its permitted exposure is. The permitted exposure also reflects trivalent chromium, while a total-chromium result does not distinguish oxidation states.

Is a natural-origin bile acid more likely to carry Class 1 elements?

The pathway is different rather than the level. Cadmium, lead, arsenic and mercury enter an animal-derived material through feed, water and environment, so they are controlled at the source of supply rather than by a purification step downstream.

What is the 30% threshold in ICH Q3D?

If an element is consistently below 30% of its permitted daily exposure, and the supporting data are sound, no further control is needed. Above that level a control strategy has to be established.

Which analytical method is used for elemental impurities?

ICP-MS or ICP-OES following closed-vessel digestion, with a method validated for the specific elements being controlled. Matrix and interference effects differ by element, and mercury in particular requires attention to carry-over between samples.

The question that separates an assessment from a template

Ask which elements were excluded, and what evidence justified each exclusion. A supplier with a written assessment will walk through it element by element, including the ones they left out. A supplier working from a template will answer with the certificate — and the certificate is not the document you asked for.

Our chenodeoxycholic acid oxide grade and the other bile-acid materials we handle are described with the parameters buyers ask about most, and you can request a quotation against your own acceptance criteria. The parallel question for the reagent tail is covered in our note on residual solvents under ICH Q3C.

The documents I would point you to are the ICH quality guidelines, the ICH Q3D(R2) guideline text and the European Pharmacopoeia. Versions differ, and the details matter here.

This article describes regulatory frameworks and analytical interpretation for elemental impurities, written for quality, regulatory and procurement professionals. It states no therapeutic use, dose or indication for any substance and no order quantity or lead time, and makes no claim about the compliance status of any particular material or facility. Limit values and element classifications change between revisions of the guideline; confirm them against the current version of ICH Q3D and the applicable pharmacopoeial text before relying on them.