eduKateSG · Why Science?
Find tiny DNA-reactive impurities, judge their hazard and control exposure before they become a preventable risk
Read FDA’s final July 2023 ICH M7(R2) guidance through hazard assessment, Ames testing, computational toxicology, impurity classes, acceptable intakes, less-than-lifetime exposure, process understanding and control strategies.
A medicine may contain tiny impurities formed from starting materials, reagents, synthesis or degradation. Most are managed through ordinary quality limits, but a DNA-reactive impurity deserves special attention because very low long-term exposure may carry mutagenic and potentially carcinogenic concern.
FDA’s final July 2023 ICH M7(R2) guidance and its addendum and questions-and-answers harmonise how these impurities are assessed and controlled. The framework combines chemistry, toxicology, computation, manufacturing knowledge and exposure science.
Did you know? M7 does not treat every structural alert as proof of danger. Two complementary computational methods can identify alerts, and expert review plus testing can refine the classification.
For students, M7(R2) shows risk-based science at its best: identify a plausible hazard, quantify an acceptable exposure and design the process so patients are protected.
Chapter index
Chapters 1–9: July 2023 final guidance, DNA reactivity, mutagenicity, Ames tests and computational assessment
- 1. M7 focuses on DNA-reactive mutagenic impurities
- 2. The July 2023 M7(R2) package is final
- 3. Scope depends on product and development stage
- 4. Actual and potential impurities both matter
- 5. Structural knowledge begins hazard assessment
- 6. Two complementary computational methods are recommended
- 7. Expert review interprets the prediction
- 8. The bacterial mutagenicity test can clarify hazard
- 9. Impurity classes organise evidence and action
Chapters 10–17: classes, acceptable intake, treatment duration, process purge, specifications and lifecycle knowledge
- 10. Acceptable intake translates hazard into exposure
- 11. Treatment duration changes cumulative exposure
- 12. Some compounds require special consideration
- 13. Control begins with preventing formation
- 14. Purge knowledge can support process controls
- 15. Specifications are one part of a control strategy
- 16. Lifecycle changes can alter the impurity profile
- 17. A complete M7(R2) receipt makes control defensible
CHAPTER 1 OF 17 · Understand the hazard
1. M7 focuses on DNA-reactive mutagenic impurities
The guidance addresses impurities that can directly damage DNA at low levels, cause mutations and potentially increase cancer risk. It provides a practical framework for pharmaceutical development and manufacture rather than assuming every impurity has the same hazard.
The practical value of m7 focuses on dna-reactive mutagenic impurities appears when the scientific claim and its boundary stay together. The guidance addresses impurities that can directly damage DNA at low levels, cause mutations and potentially increase cancer risk. It provides a practical framework for pharmaceutical development and manufacture rather than assuming every impurity has the same hazard. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
Try this
Build an evidence map for m7 focuses on dna-reactive mutagenic impurities. Draw the formation and purge pathway for m7 focuses on dna-reactive mutagenic impurities across every unit operation. Add the owner and decision it supports. Keep this limit: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
A strong science note separates observation from interpretation. Draw the formation and purge pathway for m7 focuses on dna-reactive mutagenic impurities across every unit operation. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Draw the formation and purge pathway for m7 focuses on dna-reactive mutagenic impurities across every unit operation. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
FDA lists M7(R2), its compound-specific acceptable-intake addendum and questions-and-answers as final documents dated July 2023 under docket FDA-2022-D-0055, issued by CBER and CDER.
In the july 2023 m7(r2) package is final, evidence has to match the decision being proposed. FDA lists M7(R2), its compound-specific acceptable-intake addendum and questions-and-answers as final documents dated July 2023 under docket FDA-2022-D-0055, issued by CBER and CDER. Preserve this guardrail: Default acceptable intakes apply only within the framework’s scope and assumptions.
Try this
Test the reasoning for the july 2023 m7(r2) package is final. Reproduce the acceptable-intake calculation for the july 2023 m7(r2) package is final using dose and treatment duration. Write the strongest alternative explanation. Respect this boundary: Default acceptable intakes apply only within the framework’s scope and assumptions.
A strong science note separates observation from interpretation. Reproduce the acceptable-intake calculation for the july 2023 m7(r2) package is final using dose and treatment duration. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Reproduce the acceptable-intake calculation for the july 2023 m7(r2) package is final using dose and treatment duration. Keep this boundary visible: Default acceptable intakes apply only within the framework’s scope and assumptions.
CHAPTER 3 OF 17 · Understand the hazard
3. Scope depends on product and development stage
The guidance primarily concerns new drug substances and products and certain postapproval changes. It identifies exclusions and special circumstances, so a team should establish scope before applying thresholds mechanically.
A trustworthy explanation of scope depends on product and development stage shows both what the evidence supports and what remains open. The guidance primarily concerns new drug substances and products and certain postapproval changes. It identifies exclusions and special circumstances, so a team should establish scope before applying thresholds mechanically. Remember: End-product testing alone may miss process drift or an unrecognised formation pathway.
Try this
Audit the record for scope depends on product and development stage. Review both computational predictions for scope depends on product and development stage and document the expert conclusion. Mark every assumption and missing link. Preserve this caution: End-product testing alone may miss process drift or an unrecognised formation pathway.
A strong science note separates observation from interpretation. Review both computational predictions for scope depends on product and development stage and document the expert conclusion. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review both computational predictions for scope depends on product and development stage and document the expert conclusion. Keep this boundary visible: End-product testing alone may miss process drift or an unrecognised formation pathway.
CHAPTER 4 OF 17 · Understand the hazard
4. Actual and potential impurities both matter
Assessment considers impurities observed in drug substance or product and those reasonably expected from synthesis, reagents, intermediates, degradation and storage. Process understanding narrows a theoretical list to credible risks.
The practical value of actual and potential impurities both matter appears when the scientific claim and its boundary stay together. Assessment considers impurities observed in drug substance or product and those reasonably expected from synthesis, reagents, intermediates, degradation and storage. Process understanding narrows a theoretical list to credible risks. Keep this boundary visible: A negative result is useful only when the method and test conditions are adequate.
Try this
Build an evidence map for actual and potential impurities both matter. Challenge the control strategy for actual and potential impurities both matter with a supplier, scale or storage change. Add the owner and decision it supports. Keep this limit: A negative result is useful only when the method and test conditions are adequate.
A strong science note separates observation from interpretation. Challenge the control strategy for actual and potential impurities both matter with a supplier, scale or storage change. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Challenge the control strategy for actual and potential impurities both matter with a supplier, scale or storage change. Keep this boundary visible: A negative result is useful only when the method and test conditions are adequate.
CHAPTER 5 OF 17 · Classify the impurity
5. Structural knowledge begins hazard assessment
Known carcinogenicity or mutagenicity data take priority. When experimental information is absent, structure-based reasoning can identify DNA-reactive alerts that need further evaluation.
In structural knowledge begins hazard assessment, evidence has to match the decision being proposed. Known carcinogenicity or mutagenicity data take priority. When experimental information is absent, structure-based reasoning can identify DNA-reactive alerts that need further evaluation. Preserve this guardrail: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
Try this
Test the reasoning for structural knowledge begins hazard assessment. Draw the formation and purge pathway for structural knowledge begins hazard assessment across every unit operation. Write the strongest alternative explanation. Respect this boundary: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
A strong science note separates observation from interpretation. Draw the formation and purge pathway for structural knowledge begins hazard assessment across every unit operation. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Draw the formation and purge pathway for structural knowledge begins hazard assessment across every unit operation. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
CHAPTER 6 OF 17 · Classify the impurity
6. Two complementary computational methods are recommended
One expert rule-based and one statistical quantitative structure-activity relationship method can provide independent predictions. Their outputs, model applicability and expert review form a transparent assessment.
A trustworthy explanation of two complementary computational methods are recommended shows both what the evidence supports and what remains open. One expert rule-based and one statistical quantitative structure-activity relationship method can provide independent predictions. Their outputs, model applicability and expert review form a transparent assessment. Remember: Default acceptable intakes apply only within the framework’s scope and assumptions.
Try this
Audit the record for two complementary computational methods are recommended. Reproduce the acceptable-intake calculation for two complementary computational methods are recommended using dose and treatment duration. Mark every assumption and missing link. Preserve this caution: Default acceptable intakes apply only within the framework’s scope and assumptions.
A strong science note separates observation from interpretation. Reproduce the acceptable-intake calculation for two complementary computational methods are recommended using dose and treatment duration. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Reproduce the acceptable-intake calculation for two complementary computational methods are recommended using dose and treatment duration. Keep this boundary visible: Default acceptable intakes apply only within the framework’s scope and assumptions.
A specialist can examine analogues, mechanisms, alert relevance and limitations to resolve or explain computational outputs. Expert judgment should be documented rather than used as an invisible override.
The practical value of expert review interprets the prediction appears when the scientific claim and its boundary stay together. A specialist can examine analogues, mechanisms, alert relevance and limitations to resolve or explain computational outputs. Expert judgment should be documented rather than used as an invisible override. Keep this boundary visible: End-product testing alone may miss process drift or an unrecognised formation pathway.
Try this
Build an evidence map for expert review interprets the prediction. Review both computational predictions for expert review interprets the prediction and document the expert conclusion. Add the owner and decision it supports. Keep this limit: End-product testing alone may miss process drift or an unrecognised formation pathway.
A strong science note separates observation from interpretation. Review both computational predictions for expert review interprets the prediction and document the expert conclusion. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review both computational predictions for expert review interprets the prediction and document the expert conclusion. Keep this boundary visible: End-product testing alone may miss process drift or an unrecognised formation pathway.
CHAPTER 8 OF 17 · Classify the impurity
8. The bacterial mutagenicity test can clarify hazard
An appropriate Ames test can assess mutagenic potential. A negative result under adequate conditions can support treatment as a non-mutagenic impurity, while a positive result leads to risk-control considerations.
In the bacterial mutagenicity test can clarify hazard, evidence has to match the decision being proposed. An appropriate Ames test can assess mutagenic potential. A negative result under adequate conditions can support treatment as a non-mutagenic impurity, while a positive result leads to risk-control considerations. Preserve this guardrail: A negative result is useful only when the method and test conditions are adequate.
Try this
Test the reasoning for the bacterial mutagenicity test can clarify hazard. Challenge the control strategy for the bacterial mutagenicity test can clarify hazard with a supplier, scale or storage change. Write the strongest alternative explanation. Respect this boundary: A negative result is useful only when the method and test conditions are adequate.
A strong science note separates observation from interpretation. Challenge the control strategy for the bacterial mutagenicity test can clarify hazard with a supplier, scale or storage change. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Challenge the control strategy for the bacterial mutagenicity test can clarify hazard with a supplier, scale or storage change. Keep this boundary visible: A negative result is useful only when the method and test conditions are adequate.
M7 classifies impurities using known carcinogenic and mutagenic data, structural alerts and testing. The class helps determine whether compound-specific limits, the threshold of toxicological concern or ordinary impurity controls apply.
A trustworthy explanation of impurity classes organise evidence and action shows both what the evidence supports and what remains open. M7 classifies impurities using known carcinogenic and mutagenic data, structural alerts and testing. The class helps determine whether compound-specific limits, the threshold of toxicological concern or ordinary impurity controls apply. Remember: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
Try this
Audit the record for impurity classes organise evidence and action. Draw the formation and purge pathway for impurity classes organise evidence and action across every unit operation. Mark every assumption and missing link. Preserve this caution: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
A strong science note separates observation from interpretation. Draw the formation and purge pathway for impurity classes organise evidence and action across every unit operation. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Draw the formation and purge pathway for impurity classes organise evidence and action across every unit operation. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
CHAPTER 10 OF 17 · Set the intake
10. Acceptable intake translates hazard into exposure
An acceptable intake is a daily exposure considered to pose negligible carcinogenic risk under the framework. Compound-specific data can justify a value, while the general threshold of toxicological concern can apply in appropriate cases.
The practical value of acceptable intake translates hazard into exposure appears when the scientific claim and its boundary stay together. An acceptable intake is a daily exposure considered to pose negligible carcinogenic risk under the framework. Compound-specific data can justify a value, while the general threshold of toxicological concern can apply in appropriate cases. Keep this boundary visible: Default acceptable intakes apply only within the framework’s scope and assumptions.
Try this
Build an evidence map for acceptable intake translates hazard into exposure. Reproduce the acceptable-intake calculation for acceptable intake translates hazard into exposure using dose and treatment duration. Add the owner and decision it supports. Keep this limit: Default acceptable intakes apply only within the framework’s scope and assumptions.
A strong science note separates observation from interpretation. Reproduce the acceptable-intake calculation for acceptable intake translates hazard into exposure using dose and treatment duration. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Reproduce the acceptable-intake calculation for acceptable intake translates hazard into exposure using dose and treatment duration. Keep this boundary visible: Default acceptable intakes apply only within the framework’s scope and assumptions.
Less-than-lifetime approaches allow adjusted daily intakes for shorter treatment periods while managing cumulative risk. The intended duration and maximum daily dose therefore belong in the calculation.
In treatment duration changes cumulative exposure, evidence has to match the decision being proposed. Less-than-lifetime approaches allow adjusted daily intakes for shorter treatment periods while managing cumulative risk. The intended duration and maximum daily dose therefore belong in the calculation. Preserve this guardrail: End-product testing alone may miss process drift or an unrecognised formation pathway.
Try this
Test the reasoning for treatment duration changes cumulative exposure. Review both computational predictions for treatment duration changes cumulative exposure and document the expert conclusion. Write the strongest alternative explanation. Respect this boundary: End-product testing alone may miss process drift or an unrecognised formation pathway.
A strong science note separates observation from interpretation. Review both computational predictions for treatment duration changes cumulative exposure and document the expert conclusion. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review both computational predictions for treatment duration changes cumulative exposure and document the expert conclusion. Keep this boundary visible: End-product testing alone may miss process drift or an unrecognised formation pathway.
Highly potent mutagenic carcinogens and compounds such as certain N-nitrosamines may need compound-specific evaluation outside a simple default threshold. Current scientific and regulatory recommendations should be checked.
A trustworthy explanation of some compounds require special consideration shows both what the evidence supports and what remains open. Highly potent mutagenic carcinogens and compounds such as certain N-nitrosamines may need compound-specific evaluation outside a simple default threshold. Current scientific and regulatory recommendations should be checked. Remember: A negative result is useful only when the method and test conditions are adequate.
Try this
Audit the record for some compounds require special consideration. Challenge the control strategy for some compounds require special consideration with a supplier, scale or storage change. Mark every assumption and missing link. Preserve this caution: A negative result is useful only when the method and test conditions are adequate.
A strong science note separates observation from interpretation. Challenge the control strategy for some compounds require special consideration with a supplier, scale or storage change. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Challenge the control strategy for some compounds require special consideration with a supplier, scale or storage change. Keep this boundary visible: A negative result is useful only when the method and test conditions are adequate.
Route selection, reagent choice, reaction conditions, quenching, purification and packaging can avoid or reduce impurity creation. Prevention is often more robust than relying only on end-product testing.
The practical value of control begins with preventing formation appears when the scientific claim and its boundary stay together. Route selection, reagent choice, reaction conditions, quenching, purification and packaging can avoid or reduce impurity creation. Prevention is often more robust than relying only on end-product testing. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
Try this
Build an evidence map for control begins with preventing formation. Draw the formation and purge pathway for control begins with preventing formation across every unit operation. Add the owner and decision it supports. Keep this limit: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
A strong science note separates observation from interpretation. Draw the formation and purge pathway for control begins with preventing formation across every unit operation. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Draw the formation and purge pathway for control begins with preventing formation across every unit operation. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
CHAPTER 14 OF 17 · Control the process
14. Purge knowledge can support process controls
Solubility, volatility, reactivity and unit-operation removal can show why an impurity is cleared from the process. Experimental data and conservative calculations should match the process and variability.
In purge knowledge can support process controls, evidence has to match the decision being proposed. Solubility, volatility, reactivity and unit-operation removal can show why an impurity is cleared from the process. Experimental data and conservative calculations should match the process and variability. Preserve this guardrail: Default acceptable intakes apply only within the framework’s scope and assumptions.
Try this
Test the reasoning for purge knowledge can support process controls. Reproduce the acceptable-intake calculation for purge knowledge can support process controls using dose and treatment duration. Write the strongest alternative explanation. Respect this boundary: Default acceptable intakes apply only within the framework’s scope and assumptions.
A strong science note separates observation from interpretation. Reproduce the acceptable-intake calculation for purge knowledge can support process controls using dose and treatment duration. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Reproduce the acceptable-intake calculation for purge knowledge can support process controls using dose and treatment duration. Keep this boundary visible: Default acceptable intakes apply only within the framework’s scope and assumptions.
CHAPTER 15 OF 17 · Control the process
15. Specifications are one part of a control strategy
Testing of starting material, intermediate, drug substance or product may be appropriate depending on where the impurity arises and how reliably it is removed. Detection capability and sampling need to support the proposed limit.
A trustworthy explanation of specifications are one part of a control strategy shows both what the evidence supports and what remains open. Testing of starting material, intermediate, drug substance or product may be appropriate depending on where the impurity arises and how reliably it is removed. Detection capability and sampling need to support the proposed limit. Remember: End-product testing alone may miss process drift or an unrecognised formation pathway.
Try this
Audit the record for specifications are one part of a control strategy. Review both computational predictions for specifications are one part of a control strategy and document the expert conclusion. Mark every assumption and missing link. Preserve this caution: End-product testing alone may miss process drift or an unrecognised formation pathway.
A strong science note separates observation from interpretation. Review both computational predictions for specifications are one part of a control strategy and document the expert conclusion. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review both computational predictions for specifications are one part of a control strategy and document the expert conclusion. Keep this boundary visible: End-product testing alone may miss process drift or an unrecognised formation pathway.
CHAPTER 16 OF 17 · Control the process
16. Lifecycle changes can alter the impurity profile
Scale, site, supplier, route, equipment, formulation, packaging and storage changes may create new formation or purge conditions. Change control should revisit the risk assessment instead of inheriting old assumptions.
The practical value of lifecycle changes can alter the impurity profile appears when the scientific claim and its boundary stay together. Scale, site, supplier, route, equipment, formulation, packaging and storage changes may create new formation or purge conditions. Change control should revisit the risk assessment instead of inheriting old assumptions. Keep this boundary visible: A negative result is useful only when the method and test conditions are adequate.
Try this
Build an evidence map for lifecycle changes can alter the impurity profile. Challenge the control strategy for lifecycle changes can alter the impurity profile with a supplier, scale or storage change. Add the owner and decision it supports. Keep this limit: A negative result is useful only when the method and test conditions are adequate.
A strong science note separates observation from interpretation. Challenge the control strategy for lifecycle changes can alter the impurity profile with a supplier, scale or storage change. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Challenge the control strategy for lifecycle changes can alter the impurity profile with a supplier, scale or storage change. Keep this boundary visible: A negative result is useful only when the method and test conditions are adequate.
CHAPTER 17 OF 17 · Control the process
17. A complete M7(R2) receipt makes control defensible
The receipt records scope, impurity origin, structure and data search, computational models, expert review, Ames result, class, acceptable-intake basis, dose and duration, analytical method, process controls, purge evidence, specifications, deviations and lifecycle triggers. It connects hazard to patient exposure and manufacturing action.
In a complete m7(r2) receipt makes control defensible, evidence has to match the decision being proposed. The receipt records scope, impurity origin, structure and data search, computational models, expert review, Ames result, class, acceptable-intake basis, dose and duration, analytical method, process controls, purge evidence, specifications, deviations and lifecycle triggers. It connects hazard to patient exposure and manufacturing action. Preserve this guardrail: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
Try this
Test the reasoning for a complete m7(r2) receipt makes control defensible. Draw the formation and purge pathway for a complete m7(r2) receipt makes control defensible across every unit operation. Write the strongest alternative explanation. Respect this boundary: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
A strong science note separates observation from interpretation. Draw the formation and purge pathway for a complete m7(r2) receipt makes control defensible across every unit operation. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Draw the formation and purge pathway for a complete m7(r2) receipt makes control defensible across every unit operation. Keep this boundary visible: A structural alert is a screening signal, not proof that an impurity is carcinogenic.
M7(R2) mutagenic-impurity control map
| Question | Evidence to capture | Boundary to retain |
|---|---|---|
| Could it form? | Route, reagents, degradation and process knowledge | Assess credible possibilities |
| Is it mutagenic? | Known data, two predictions, expert review and Ames testing | Alerts are not proof |
| What exposure is acceptable? | Compound data, TTC, dose and duration | Check special compounds |
| How is it controlled? | Prevention, purge, analytical testing and lifecycle change | Control the process, not only the result |
Science learning becomes useful when a student can show where a claim came from, why it applies and where it stops. Continue with the eduKate Science Learning Hub, compare choices in the Education Hub, and connect these habits to career pathways built from subject capabilities.
M7(R2) is a reassuring example of chemistry serving public health. It takes a risk that may exist at trace levels and turns it into a transparent sequence of questions—could it form, is it mutagenic, what exposure is acceptable and how will the process keep it controlled?
Primary sources and verification date
Source status checked 26 October 2026. FDA guidance describes the agency’s current thinking and does not replace applicable statutes, regulations or product-specific communication with FDA.
- FDA: ICH M7(R2) mutagenic impurities — Final guidance package, July 2023; docket FDA-2022-D-0055.
- FDA: Full ICH M7(R2) guidance PDF — Official hazard-assessment, classification, intake and control principles.
- FDA: M7(R2) Questions and Answers — Final implementation clarifications and examples.