ICH Q3DICH Q3D(R2) Elemental Impurities

ICH Q3D, finalised in 2014 and revised as Q3D(R2) in 2022, harmonised the regulatory treatment of elemental impurities (formerly 'heavy metals') across FDA, EMA, PMDA and other ICH regions. It replaces the legacy colorimetric heavy-metals test in USP <231> / Ph. Eur. 2.4.8 with a risk-based, toxicology-driven framework: identify which of 24 named elements could enter the drug product through any route (API, excipient, water, container/closure, manufacturing equipment, leaching), assess the risk against the route-of-administration-specific Permitted Daily Exposure (PDE), and document the conclusion.

USP <232> mirrors the Q3D PDEs in US compendial law; USP <233> defines the test procedures (typically ICP-OES or ICP-MS). EMA expects Q3D risk assessment in all marketing-authorisation applications and variations; FDA's expectation, set out in Q3D and reinforced in Form FDA 483 observations, is identical.

Q3D groups 24 elements into four classes based on toxicity and probability of occurrence. Each class has a different default risk posture, but every element with a 'reasonable likelihood of occurrence' must be assessed regardless of class.

PDEs are expressed in µg/day and differ by route of administration — oral (least conservative), parenteral (more conservative because bioavailability ≈ 100 %) and inhalation (most conservative because of local lung toxicity). PDEs are derived from the most relevant toxicology endpoint (NOAEL, LOAEL or cancer slope factor for genotoxic elements) using standard modifying factors (interspecies, intraspecies, severity of effect).

ICH Q3D §5 sets out three options for demonstrating that a drug product complies with PDEs:

1. Option 1 — common permitted concentration. For each element, apply a single concentration limit (PDE ÷ maximum daily dose for a 10 g/day product) to every component. Conservative and simple; useful for low-dose products.
2. Option 2a — common permitted concentration scaled to actual daily intake. The concentration limit per element is scaled to the product's actual maximum daily intake (PDE ÷ actual MDD). Tighter than Option 1 only when MDD ≠ 10 g/day.
3. Option 2b — risk-based concentration per component, based on component intake. Allow different element limits per component, so a small-mass colouring agent can carry a higher concentration than the bulk API.
4. Option 3 — finished-product analysis. Measure the final drug product against the PDE-derived concentration limit. Most flexible but most analytically demanding; useful where upstream data are missing.

• API synthesis — residual catalysts (Pd, Pt, Ru, Rh, Ni) from coupling and hydrogenation; Lewis-acid catalysts (Ti, Zn, Al); chiral catalysts. The synthetic route is often the dominant Class 2A/2B source.
• Excipients — natural-origin excipients (mined minerals like talc, kaolin, mineral oils) tend to carry Class 1 (As, Cd, Pb) and Class 3 (Ba, Cr) at variable levels; synthetic excipients typically carry only trace amounts.
• Water — purified water and WFI are normally below detection but should be verified per the site's water-system risk assessment.
• Container/closure — leachables from glass (As, Cd in old formulations; Si, Al from borosilicate), elastomers (Zn, Pb in some carbon-black sources), plastics (Sb from PET catalysts).
• Manufacturing equipment — stainless-steel surfaces (Cr, Ni, Mo) under aggressive cleaning or acidic process conditions; reactor erosion.
• Intentionally added — Co (vitamin B12), Cu / Zn / Mo (trace minerals in supplements and parenteral nutrition), Pt (cisplatin and related anti-cancer APIs).

USP <233> describes the standard test procedures: sample preparation by closed-vessel microwave digestion (the only practical option for organic matrices to avoid loss of volatile elements like Hg), separation/dissolution as required, and instrumental analysis by ICP-OES (sufficient for most Class 1 and Class 3 elements at typical PDE levels) or ICP-MS (required for Class 2A elements at parenteral PDEs and for ultra-low-level work). The method must be validated per <233> — system suitability, accuracy, precision, specificity, limit of quantitation — and is treated as a Q2(R2) analytical procedure.

1. Risk assessment is a one-off document with no review trigger when the API supplier or excipient grade changes.
2. Risk assessment relies on 'no addition' arguments without supporting data for raw materials with variable mineral content (mined excipients, colourants).
3. Parenteral or inhalation product treated under oral PDEs (much less conservative than required).
4. USP <232> claimed but Cd, As, Hg or Pb actually exceed Class 1 PDEs in a specific batch — and there's no investigation.
5. Test method validated for Class 1 elements only — Class 2A/2B/3 omitted with no scientific justification.
6. Catalyst-residue data on the API CoA but no cross-check against Q3D PDEs — the data are there, the gate isn't.
7. Container/closure leaching study omits elemental impurities — only organic leachables were considered.

• Risk-assessment register per product holds the Q3D conclusions per element — class, source(s), measured/estimated contribution per component, PDE-derived limit at the chosen Option (1, 2a, 2b, 3), and the conclusion (no further control / spec on raw material / finished-product test).
• Supplier CoA module captures element-by-element data when present; the platform compares against the per-element specification and flags any drift over time, not just an outright fail.
• Change-control workflow auto-routes API-source changes, excipient-grade changes, container-closure changes and synthesis-route changes to the Q3D risk-assessment owner — the file is re-opened, not silently invalidated.
• Finished-product testing (Option 3) integrates with the LIMS like any other release attribute; the ICP-MS method is held as a validated Q2(R2) procedure with its own change-control history.
• Audit trail covers every revision of the risk assessment — what changed, who approved, what triggered the review — so the regulator sees the file's evolution, not a single static PDF.