ICH Q3A(R2): Impurities in New Drug Substances, Thresholds and Qualification

ICH Q3A(R2) is guidance for registration applications on the content and qualification of impurities in chemically synthesised new drug substances not previously registered in a region or member state. ICH Q3A(R2), section 1 frames the subject through two complementary perspectives: chemistry and safety.

The chemistry perspective covers impurity classification and identification, reporting, listing impurities in specifications, and analytical procedures. It connects the description of an impurity with how it is measured, presented in the application and addressed in the drug substance specification.

The safety perspective addresses qualification of impurities that were absent, or present at substantially lower levels, in the drug substance batches used for safety and clinical studies. The comparison is therefore with the impurity content of those study batches, not simply with an impurity’s name or analytical identity.

Read together, these perspectives connect impurity chemistry, specification listing and safety evaluation within the registration application. For broader orientation, see the ICH Q3A/B overview.

ICH Q3A(R2) is guidance for a defined registration context, not a general impurity standard for every substance or development stage. The scope boundaries in section 1 distinguish the intended application from clinical research and expressly excluded substance types.

Classification describes the impurity category

Section 2 classifies impurities into three categories:

• Organic impurities (process- and drug-related)
• Inorganic impurities
• Residual solvents

Section 2 then lists what each category contains. Organic impurities can arise during manufacture and/or storage and include starting materials, by-products, intermediates, degradation products, and reagents, ligands and catalysts. Inorganic impurities result from the manufacturing process, are normally known and identified, and include reagents, ligands and catalysts, heavy metals or other residual metals, inorganic salts and other materials such as filter aids and charcoal. Excluded from the guideline are extraneous contaminants that are more appropriately addressed as GMP issues, polymorphic forms, and enantiomeric impurities.

Identification and specification status answer different questions

The definitions in section 8 separate structural characterisation from how an impurity is controlled in the specification. “Unidentified” and “unspecified” are therefore not interchangeable labels.

An unidentified impurity can thus have a specific specification entry: lack of structural characterisation does not itself mean that the impurity is unspecified. Keep the category, identification status and specification status distinct when reading an impurity profile. For related terminology, see the ICH Q3A/B glossary.

Attachment 1 of ICH Q3A(R2) gives the following guidance thresholds. Select the maximum daily dose band first, then read the reporting, identification and qualification columns separately.

Attachment 1 footnotes: ¹ “The amount of drug substance administered per day”; ² “Higher reporting thresholds should be scientifically justified”; ³ “Lower thresholds can be appropriate if the impurity is unusually toxic”. Footnote 2 qualifies the reporting column; footnote 3 applies to all three threshold columns.

Dose basis and the lower-of comparison

Exactly 2 g/day belongs in the first row; only doses greater than 2 g/day belong in the second. The dose basis is the daily amount of drug substance, not tablet mass or total formulation mass. Keep that basis explicit when selecting the row.

For the first row, evaluate identification and qualification independently. Compare the impurity intake corresponding to 0.10% of the daily drug-substance dose with 1.0 mg/day for identification. Separately compare the intake corresponding to 0.15% with 1.0 mg/day for qualification. In each comparison, the lower amount controls. Neither percentage is an alternative that can be chosen instead of the lower daily-intake value.

Report the result, then compare like units

Section 5, Reporting Impurity Content of Batches, recommends two decimal places below 1.0%—for example, 0.06% or 0.13%—and one decimal place at and above 1.0%, for example, 1.3%. It also specifies conventional rounding. These are numerical presentation rules; the subsequent threshold comparison follows its own sequence.

1. For a percentage threshold, round the reported result to the same decimal place as the threshold, then compare directly.
2. For a total daily intake (TDI) threshold, convert the reported percentage to impurity intake: TDI (mg/day) = reported impurity (%) ÷ 100 × drug-substance dose (mg/day).
3. Round that TDI to the same decimal place as the daily-intake threshold, then compare. Do not compare a percentage directly with a value in milligrams.

For broader orientation beyond this threshold reference, see the ICH Q3A/B overview.

Assume maximum daily dose means drug-substance mass, results are percentages of that substance, and the usual Attachment 1 thresholds apply. These are illustrative calculations, not release determinations. Keep the dose, reported percentage and comparison units explicit throughout.

Impurity intake in mg/day = dose in g/day × 1000 mg/g × reported percentage ÷ 100.

Case 1: 0.5 g/day and a reported 0.12%

The dose is 500 mg/day. Impurity intake is 500 × 0.12 ÷ 100 = 0.60 mg/day. The identification comparison is 0.10%, equivalent to 0.50 mg/day; the qualification comparison is 0.15%, equivalent to 0.75 mg/day. Both percentage-based amounts are lower than the respective 1.0 mg/day alternatives. The reported 0.12% exceeds 0.10% but not 0.15%: identification is indicated, while the usual qualification threshold is not exceeded.

Case 2: 0.8 g/day and rounding at the comparison step

Following Attachment 2, example 2 and note 2, a raw result of 0.124% becomes a reported 0.12%. At 800 mg/day, intake is 800 × 0.12 ÷ 100 = 0.96 mg/day. Round this to 1.0 mg/day for comparison with the qualification threshold of 1.0 mg/day: the threshold is not exceeded.

Contrast a reported 0.14%: 800 × 0.14 ÷ 100 = 1.12 mg/day, rounded to 1.1 mg/day. That exceeds the qualification threshold. The comparison uses the reported percentage converted to daily intake, then rounded to the threshold’s decimal place—not an independently chosen precision.

Case 3: 3 g/day and a reported 0.04%

The dose is 3000 mg/day. Impurity intake is 3000 × 0.04 ÷ 100 = 1.20 mg/day. Use the greater-than-2-g/day row: reporting is 0.03%, identification is 0.05%, and qualification is 0.05%. Thus, the reporting threshold is exceeded, but identification and qualification thresholds are not. Do not import the lower-dose row’s 1.0 mg/day alternative: the calculated intake does not change which dose band applies.

For practical review, retain the calculation alongside the batch execution history. Record the assessment trigger separately from any specification comparison or out-of-specification handling; authorised reviewers make the batch decision.

All three outcomes rest on the quantitation assumption, usually the drug substance’s own response factor. Attachment 2, note 1 adds that after identification, if the response factor is determined to differ significantly from the original assumptions, it may be appropriate to re-measure the actual amount of the impurity present and re-evaluate against the qualification threshold. A borderline case is therefore provisional until that check is made.

An impurity specification is not simply a copy of the threshold table. ICH Q3A(R2) distinguishes the impurities listed for control from the rationale supporting their acceptance criteria. The recommendations below are guidance for new drug substances within Q3A’s scope, not universal requirements for every manufactured material.

What belongs in the specification

Section 6, Listing of Impurities in Specifications, recommends the following organic-impurity entries, where applicable:

• Each specified identified impurity
• Each specified unidentified impurity
• Any unspecified impurity with an acceptance criterion of not more than () the identification threshold
• Total impurities

The same section’s summary also includes residual solvents and inorganic impurities, where applicable. No separate numerical limits for those categories are assigned here. Keep the four organic-impurity entries distinct in the specification layout: individual specified entries, the general unspecified-impurity criterion and the total-impurities entry serve different review purposes.

Support the proposed criterion, not just the screening result

Section 6 states that stability studies, chemical development studies and routine batch analyses can predict impurities likely to occur in the commercial product. Together, these provide a basis for considering which impurities belong in the control strategy; the list should not be assembled from threshold values alone.

A threshold-triggered assessment is not itself the justification for a proposed acceptance criterion. Under section 7, Qualification of Impurities, the applicant should provide a rationale that includes safety considerations. Where qualification data are unavailable, studies can be appropriate when the usual qualification thresholds are exceeded; that conditional recommendation is not a ready-made specification limit.

An illustrative dossier cross-reference map

Use this organisational aid to make the reasoning reviewable. It is not a prescribed CTD numbering scheme or an additional regulatory requirement.

For related terminology, see ICH Q6 specifications and ICH Q2 analytical validation. Keep those supporting discussions linked to the relevant entry rather than treating a general dossier reference as its complete justification.

Section 6 also explains how to justify the list. The selection should be based on the impurities found in batches manufactured by the proposed commercial process, and a rationale for inclusion or exclusion should discuss the impurity profiles observed in the safety and clinical development batches together with those of commercial-process batches. Acceptance criteria should be set no higher than the level that can be justified by safety data, and should be consistent with the level achievable by the manufacturing process and the analytical capability. Where there is no safety concern, criteria should be based on commercial-process batch data, allowing sufficient latitude to deal with normal manufacturing and analytical variation and the stability characteristics of the new drug substance.

Impurity results need an analytical context, especially when development spans more than one procedure. A useful review question is: which procedure generated this result, and where is its supporting validation information?

What Q3A recommends

Section 3.1 asks the applicant to summarise the actual and potential impurities most likely to arise during synthesis, purification and storage, based on sound scientific appraisal of the reactions, raw-material impurities and possible degradation products. The laboratory summary should include results of development batches and batches from the proposed commercial process, as well as stress testing (see ICH Q1A) used to identify potential impurities arising during storage, and the impurity profile of batches intended for marketing should be compared with those used in development, and any differences discussed.

Section 4 requires documented evidence that the analytical procedures are validated and suitable for the detection and quantification of impurities, and states that the quantitation limit for the analytical procedure should be not more than the reporting threshold. It also notes that the use of two decimal places for thresholds does not necessarily reflect the precision of the routine procedure, so lower precision techniques such as thin-layer chromatography can be acceptable where justified and appropriately validated. Differences between development and commercial procedures should be discussed in the application.

Illustrative method-change handoff

As an implementation practice, suppose development results were generated under procedure A and later results under procedure B. Retain the following links rather than presenting both sets under a single current-method label:

• Procedure identifier: preserve the identifier and revision actually used for each result.
• Validation reference: connect each procedure to its appropriate validation information; see the related glossary entry on ICH Q2.
• Affected result links: identify the batches and reported results generated under each procedure, with links to their chromatograms.

This suggested handoff lets reviewers distinguish results generated under different procedures without reconstructing the relationship from filenames or dates alone. It is a record-linking approach, not an additional validation requirement. Connecting it to the batch execution history can keep the analytical context accessible during review. Reviewers can then examine the relevant procedure and validation information before interpreting apparent differences between results.

Under Q3A(R2) section 7, qualification means acquiring and evaluating data establishing the biological safety of an individual impurity or a given impurity profile at specified levels. It is therefore a level-specific safety assessment, not simply structural identification. This distinction supports the broader ICH Q3A/B impurity-control framework.

Start with existing evidence and reduction options

Section 7 considers the level of an impurity present in drug substance adequately tested in safety and/or clinical studies qualified. The relevant evidence is the impurity level represented by that material—not merely the existence of studies on the drug substance.

Reducing the impurity to not more than the threshold can be simpler than providing safety data. Alternatively, adequate scientific literature may qualify the impurity. If neither route resolves the question, additional safety testing should be considered; threshold exceedance does not itself prescribe a fixed study package.

Read the decision tree by branch

The following is a simplified reading companion to Attachment 3, Decision Tree for Identification and Qualification, not a replacement for its branches and notes:

1. Identification: begin with whether the identification threshold is exceeded. The tree then addresses structural identification and, where structure is not identified, reduction to not more than the identification threshold.
2. Known risks: for an identified structure, assess known human-relevant risks. Note d asks whether safety data for the impurity or its structural class preclude human exposure at the concentration present; the risk branch leads to reduction to a safe level.
3. Qualification: follow the applicable qualification-threshold and reduction branches. Note c allows lower thresholds for unusually toxic impurities.
4. Studies and outcomes: where the tree reaches study consideration, consider patient population and duration of use. The final branch distinguishes clinically relevant adverse effects, leading to reduction to a safe level, from the no-adverse-effect branch marked qualified.

Select studies conditionally

Attachment 3 lists genotoxicity studies (point mutation and chromosomal aberration), general toxicity studies (one species, usually 14–90 days), and other specific toxicity endpoints as appropriate. Note a makes the minimum screen conditional and describes two in vitro studies. Note b bases duration on relevant information and species selection on detecting impurity toxicity; single-dose studies may be appropriate case by case, especially for single-dose drugs. Section 7 also identifies patient population, daily dose, route and duration of administration as study-selection factors.

As an implementation practice, keep the scientific qualification conclusion separate from out-of-specification handling and batch disposition. Responsible reviewers should document the evidence, supported level and unresolved safety questions.

ICH Q3A(R2), section 5 recommends presenting quantitative impurity results numerically, rather than only stating “complies” or “meets limit”. This preserves the measured impurity profile for comparison between batches, rather than reducing it to a pass/fail conclusion. For reported total impurities, sum all impurities at levels greater than (>) the reporting threshold; the total is not simply an unrestricted sum of every detected signal.

For each new drug substance batch, section 5 lists these seven report elements:

• Batch identity and size
• Date of manufacture
• Site of manufacture
• Manufacturing process
• Impurity content, individual and total
• Use of batches
• Reference to analytical procedure used

Section 5 also recommends a tabulation linking the specific substance batch to each safety study and each clinical study in which it was used. A general label such as “clinical use” does not provide that study-by-study connection.

Implementation practice: linked evidence records

The following is an illustrative record design, not a Q3A-prescribed database schema. Its purpose is to keep the batch, its measurements and the scientific interpretation connected without forcing them into one document.

1. Process and batch context: connect the batch identifier to its manufacture details and process version. Use the relevant batch execution history as the context record, rather than copying an ambiguous process name.
2. Analytical results: connect each individual result and reported total to the batch and analytical procedure reference. Keep the numeric value accessible alongside any assessment label.
3. Specification rationale: link the result set to the rationale under which it was assessed. Preserve the distinction between the observation and the interpretation discussed in the specification section.
4. Qualification review: connect the impurity assessment to the relevant batch-to-study entries and review conclusion. Let reviewers follow the relationship in either direction: from a batch to its studies, or from a study to the batches used.

An electronic batch record can be one linked record in this design; it need not contain the entire evidence chain. Record relationships support scientific review, but do not replace the reviewer’s judgement.

Q3A section 6 directly refers to ICH Q6A Decision Tree #1 for establishing an acceptance criterion for a specified impurity in a new drug substance. This is the explicit connection to Q6 specifications, rather than a reason to treat every impurity guideline as interchangeable.

Within Q3A itself, two boundaries are explicit. Section 3.2 says inorganic impurities are normally detected and quantified using pharmacopoeial or other appropriate procedures, that carry-over of catalysts should be evaluated during development, and that acceptance criteria should be based on pharmacopoeial standards or known safety data; Attachment 1 thresholds address organic impurities. Section 3.3 sends residual solvents to ICH Q3C: their control should be discussed and presented according to the ICH Q3C Guideline for Residual Solvents. Impurities in the finished new drug product fall under Q3B rather than Q3A.

For regional reference points, consult EMA's ICH Q3A(R2) page and EMA's ICH Q6A page. For jurisdiction-specific applicability, check the current regional text rather than assuming identical application across markets. Retain the product-type boundaries established above when selecting the relevant framework.

The failure modes below apply the linked records described in the batch-reporting section; they are practice, not additional Q3A requirements.

Failure modes and suggested controls

Exception recovery and decision ownership

1. Preserve the original record.
2. Identify affected assessments.
3. Correct the underlying linkage or value.
4. Rerun the affected calculation.
5. Document human review of the correction and its consequences.

Use an audit trail review workflow to make the correction history reviewable. Authorised people retain scientific approval and batch release decisions; a successful import or recalculation does not replace their judgement.

Implementation review checklist

Use this editorial review aid to check the assembled assessment. It is separate from the batch-report list in Q3A section 5, not a source-prescribed list of review elements.

• Scope assessment and dose basis are recorded.
• Results link to the applicable analytical method.
• Threshold evaluation is reproducible from recorded inputs.
• Specification rationale and qualification evidence are linked.
• Unresolved exceptions are visible to the authorised reviewer.