The newly released draft of the ICH Q3E guideline represents a significant milestone in the harmonisation of global expectations for the assessment and control of extractables and leachables (E&L) in pharmaceutical products. Building on the foundations laid by ICH Q3A, Q3B, Q3C, Q3D, and M7, it distinguishes itself by adopting a more integrated and systematic approach to risk management, firmly aligned with the principles of the ICH Q9 Quality Risk Management Guideline.
ICH Q3E sets out a harmonised, science-driven framework applicable to pharmaceuticals, drug–device combination products, and advanced therapies. While it retains the well-established regulatory paradigm of hazard identification, exposure estimation, and risk characterisation, its most notable innovation lies in the structured classification of E&L and the introduction of route- and duration-specific thresholds. This approach mirrors the tiered classification system of ICH Q3C for residual solvents and incorporates the practical recommendations developed by PQRI, enabling proportionate, patient-focused controls without imposing unnecessary conservatism.
The guideline covers the risk assessment and control of organic E&L in new drug products, and also applies to approved products when changes could affect leachable profiles. Elemental leachables are excluded, as they fall under ICH Q3D. By clarifying scope and lifecycle applicability, it ensures that E&L control is embedded in both product development and post-approval change management, while explicitly excluding contaminants unrelated to the intended manufacturing and packaging materials.
Understanding the Difference Between Leachables and Extractables
Extractables and leachables are both impurities originating from the interaction between a finished product and the materials it contacts, such as packaging, manufacturing components, or delivery systems, but they differ in how they are generated and detected.
Extractables are organic or inorganic chemical species that can be drawn out of these materials under exaggerated laboratory conditions, such as the use of strong solvents, elevated temperatures, or extended surface exposure. These stress conditions are designed to simulate worst-case scenarios, allowing the identification, characterisation, and toxicological evaluation of substances that might migrate into the product.
Leachables, in contrast, are the organic or inorganic substances that actually migrate into the product under normal manufacturing, storage, and use conditions throughout its shelf life. Because leachables represent the real contaminants to which patients may be exposed, they must be monitored against established safety limits to ensure both patient safety and product efficacy.
ICH Q3E Approaches to E&L Assessment
Under ICH Q3E, risk assessment follows the classic regulatory framework of hazard identification, exposure estimation, and integrated risk evaluation, ensuring that both hazard potential and actual exposure are considered to determine if components are acceptable—an approach widely used across regulatory domains for transparent and defensible conclusions.
In the context of ICH guidelines, the most innovative and strategically important aspect of ICH Q3E is not the derivation of Permitted Daily Exposures (PDEs), which follows the well-known formula, albeit with some modifications detailed below, nor the calculation of safety margins, as these are long-established toxicological practices already embedded in other impurity guidelines. Rather, its true novelty lies in the introduction of a structured classification system for extractables and leachables.
The approach, which mirrors the tiered framework seen in ICH Q3C for residual solvents, groups E&L into potency-based classes (Class 1, 2, and 3) according to their toxicological profile and the stringency of control required.
The overall safety assessment process for leachables, illustrated in the guideline’s flowchart, integrates the evaluation of both mutagenicity and general toxicity, ensuring that potential risks are addressed comprehensively. The proposed flow is logical and aligns with good toxicological practice, but there are areas where real-world application will be less tidy.
For example, the decision diamond for “adequate data available” hides a major complexity: what constitutes adequacy? In practice, many leachables are identified tentatively from LC-MS without full structural elucidation, leaving assessors to make calls based on substructure and analogue data. This is where over-interpretation of in silico alerts can push borderline Class 2 substances into Class 1, with a significant regulatory and operational impact.
Let’s examine the main innovations and concepts related to the toxicological risk assessment (TRA).
Route-To-Route differences
The risk assessment is refined by incorporating route-of-administration-specific thresholds, recognising that systemic exposure and local toxicity potential differ significantly between, for example, oral, parenteral, inhalation, and ocular routes. This is an important evolution from earlier guidances, aligning closely with concepts developed by the PQRI, which long advocated differentiated limits based on administration route and duration of patient exposure.
The guideline adopts a risk-based, scientific framework for E&L safety assessment, focusing on substances above the Safety Concern Threshold (SCT), the lower of the mutagenicity TTC (ICH M7) or the non-mutagenic Qualification Threshold (QT), both adjusted for route and exposure duration. QT values, derived from ~330 PDEs, differ between oral and parenteral use.
While systemic toxicity evaluation is generally sufficient, local toxicity must be considered for sensitive tissues. For ophthalmic products, a compound-specific assessment is advised above 20 ppm for topical use, and qualitative evaluation is required for injected products. Neurotoxicity concerns apply to intracerebral, intrathecal, and epidural routes, even at very low concentrations.
For dermal products, sensitisation is the most sensitive endpoint, with a Dermal Sensitisation Threshold (DST) of 1 µg/cm²/day (~500 ppm); exposures below this generally require no further local toxicity review. Transdermal systems follow the same principle. Sensitisation potential is route-dependent: dermal exposure is well-characterised through in silico, in chemico, and in vitro models; respiratory sensitisation lacks predictive tools and needs case-by-case review; sub-cutaneous and intradermal routes use a 50 ppm threshold due to higher immune cell proximity, while intravenous and intramuscular routes are unlikely to pose sensitisation risks at typical E&L levels.
E&L Classification
The classification scheme in ICH Q3E is a genuine practical innovation, grouping leachables into three potency classes.
Class 1 represents the high-concern category, covering mutagenic carcinogens in the ICH M7 “cohort of concern,” ICH M7 Class 1 compounds with acceptable intakes below 1.5 µg/day, and certain highly potent non-mutagens such as bisphenol A or benzo(a)pyrene, where default thresholds offer insufficient protection. These should be avoided wherever possible, or otherwise controlled through a compound-specific PDE supported by robust scientific justification. A real-world challenge arises with borderline compounds that do not meet Class 1 criteria yet raise concerns in silico or via partial structural alerts; while the default tendency is to treat them as Class 1, high-quality toxicological data can often support a less conservative classification.
Class 2 E&L form the default category, where standard mutagenicity TTC or non-mutagenic QT values are deemed protective. These thresholds are route- and duration-specific, derived from a large dataset of leachable PDEs, such as a chronic parenteral QT of 12 µg/day compared to 48 µg/day for oral exposure, reflecting the higher concern associated with parenteral delivery.
Class 3 E&L, at the opposite end of the spectrum, are those with very low toxic potential and PDE values well above typical E&L findings. Where exposure remains below 1 mg/day, no further qualification is required, allowing focus to remain on substances of genuine clinical concern.
This stratification is sensible, but the assignment of a compound to a class depends on having enough toxicological information to make that call. Many real-world cases fall into the “non-listed” category, with no published PDE or sufficient toxicity data.
Here, ICH Q3E points to Appendix 5, which sets out how to derive PDEs from the toxicology literature, selecting a point of departure (PoD) such as a NOAEL or LOAEL, applying appropriate assessment factors, and considering route-to-route extrapolation. This is where toxicological skill becomes critical: inappropriate factor selection or over-conservative defaults can drive PDEs, and hence allowable leachable limits, down to impractical levels, potentially forcing unnecessary reformulation or repackaging.
Safety limits determination
When deriving a PDE for a leachable, the assessment should take into account a broad range of relevant information. This includes any known biological activity that could influence safety, toxicokinetic data such as absorption, bioavailability, route-to-route extrapolation, and potential for bioaccumulation, and systemic toxicity findings from acute, subchronic, and chronic studies. It should also consider the compound’s sensitisation potential and any evidence of local irritation, developmental and reproductive toxicity, including possible endocrine-disrupting effects, and available genotoxicity and carcinogenicity data, supported where appropriate by in silico predictions. Finally, other pertinent toxicological or epidemiological information—such as existing health-based exposure limits, clinical observations, and data from structurally related compounds—should be incorporated to ensure that the derived PDE is scientifically robust and defensible.
Notably, although the formula used for deriving PDEs remains the same as in ICH Q3C and Q3D, ICH Q3E introduces two additional modifying factors—F6 and F7—specifically for leachables. These expand upon the traditional uncertainty factors F1–F5 described in existing toxicological frameworks, allowing route-to-route extrapolation and the application of read-across approaches to be more transparently and consistently addressed.
F6 is applied when extrapolating toxicity data from one route of exposure to another (e.g., oral to parenteral) to account for differences in bioavailability:
- F6 = 100 if oral bioavailability is <1%
- F6 = 10 if oral bioavailability is ≥1% and <50%
- F6 = 2 if oral bioavailability is ≥50% and <90%
- F6 = 1 if oral bioavailability is ≥90%
When direct in vivo data are lacking, alternative approaches such as New Approach Methodologies (NAMs), combining in vitro absorption and clearance studies with PBPK modelling, can be used, or a conservative default F6 of 100 applied. Surrogate bioavailability data from structurally similar compounds may be applied through a well-justified read-across approach. Route-specific defaults are provided, such as absorption estimates for dermal exposure (70%, 50%, or 10% depending on solvent type, molecular weight, and logPow), with adjustments for factors like enhanced formulations or compromised skin. Inhalation assessments should also consider deposition, absorption rate, and metabolism.
F7 applies when a read-across approach is used to substitute data from a similar compound. Depending on the degree of structural and toxicological similarity, an additional factor of up to 5 can be applied, with F7 = 1 reserved for surrogates that meet the similarity criteria in the guideline.
Notably, the E&L classification and PDE derivation together form a coherent and logically structured system, but one that requires significant toxicological expertise to apply. SCT values, defined as the lower of the TTC or QT, are intentionally conservative, with the parenteral QT of 12 µg/day frequently becoming the limiting factor, even for substances with decades of demonstrated safe human use. Although the guideline allows replacement of the default QT with a compound-specific PDE, generating one that will satisfy regulatory authorities necessitates an in-depth literature review, thorough appraisal of study quality, and a transparent, well-justified application of all modifying factors.
Can Extractables and Leachables Be Assessed Without Animal Testing?
One of the more progressive aspects of Appendix 5 is its encouragement of read-across: using data from structurally similar substances to inform the safety of a leachable with limited direct data. This, together with in silico tools and in vitro methods, offers a route to scientifically sound PDEs without resorting to new animal studies, provided the analogues are chosen with chemical and toxicological relevance in mind.
The guideline’s nod to NAMs is welcome, but it stops short of providing criteria for their acceptance, leaving it to the assessor and the reviewing authority to agree on their validity. Finally, the guideline distils the safety assessment into a decision tree linking mutagenicity and general toxicity endpoints. Leachables above the Analytical Evaluation Threshold (AET) are classified, and thresholds are applied. Class 1 compounds require compound-specific limits, while others are assessed against the mutagenicity TTC and non-mutagenic QT. Exceedances must be managed by reducing exposure, qualifying toxicologically, applying read-across, or generating new data. For mutagens, genotoxicity testing may de-risk them; for non-mutagens, subchronic or chronic studies can refine the PDE.
Along with the guideline, ICH Q3E published supporting documentation that should accompany Class 3 leachable monographs.
Final Thoughts
In conclusion, ICH Q3E offers a harmonised, science-based framework for extractables and leachables assessment, but its effective application requires expert toxicological judgement to interpret classifications, derive robust PDEs, and apply route- and duration-specific thresholds.
At ToxHub, we specialise in navigating these complexities, combining regulatory insight and toxicological expertise to deliver defensible, regulator-ready assessments that go beyond compliance to set high industry standards.
Issued by Carla Landolfi
References:
Q3A(R2) GuidelineQ3B(R2) Guideline – ICH Q3E Draft Guideline – Guideline Supporting Documentation – Q9(R1) Guideline – Q3C(R9) Guideline – Q3D(R2) Guideline – M7(R2) Guideline