The ICH M7 classification into five classes is the first — and most delicate — step of any mutagenic impurity assessment. The assigned class decides everything that follows: whether a compound-specific limit is needed, whether the TTC threshold is enough, whether an Ames test is worth running, or whether the impurity can be handled as an ordinary impurity under ICH Q3A/Q3B. The current version of the guideline, ICH M7(R2), was adopted in April 2023 and consolidates the assessment framework and the acceptable intakes for compounds with carcinogenicity data into a single text. In this article you will find the five classes explained with the official table, the applicable limits and the mistakes inspectors challenge most often.
ICH M7 classification: what it is for and when it applies
ICH M7 applies to DNA-reactive (mutagenic) impurities in new pharmaceuticals and, in defined cases, to marketed products — for example after synthesis changes that introduce new impurities or raise the levels of existing ones. The goal is not to remove every theoretical risk, but to limit the potential carcinogenic risk to a negligible level (roughly one additional case in 100,000 for lifetime exposure). The logical flow is always the same: identify actual and potential impurities (raw materials, intermediates, reagents, degradation products), evaluate their mutagenic potential through literature data and (Q)SAR, assign the class and derive the limit and the control strategy from it. Without a documented classification, any limit claimed in the dossier is indefensible.
The 5 classes of ICH M7 Table 1
Table 1 of the guideline defines the classes based on two bodies of evidence: mutagenicity (typically from the Ames test) and carcinogenicity (rodent studies). Here is the complete scheme:
| Class | Definition | Control action |
|---|---|---|
| 1 | Known mutagenic carcinogens | Control at or below a compound-specific limit calculated from carcinogenicity data |
| 2 | Known mutagens (Ames positive) with unknown carcinogenic potential | Control at or below TTC-based acceptable intakes |
| 3 | Alerting structure, unrelated to the drug substance; no mutagenicity data | Control within the TTC, or run an Ames test: if negative it becomes Class 5, if positive Class 2 |
| 4 | Alerting structure, same alert in the drug substance or related compounds (e.g. intermediates) that have been tested and are non-mutagenic | Treat as a non-mutagenic impurity |
| 5 | No structural alerts, or alerting structure with sufficient data to demonstrate lack of mutagenicity/carcinogenicity | Treat as a non-mutagenic impurity |
Classes 4 and 5 therefore leave the M7 perimeter and fall back under the ordinary qualification thresholds of ICH Q3A/Q3B; Classes 1, 2 and 3 require specific limits and a dedicated control strategy.
ICH guidelines, nitrosamines and requirements on limits and controls keep changing. To get a practical weekly update on GMP, impurities and audit readiness, subscribe to The Pragmatic GMP, our free newsletter: ten minutes of reading to stay one step ahead of the next inspection.
How the class is assigned: two complementary (Q)SAR systems
When experimental data are missing, classification relies on (Q)SAR systems. ICH M7 is explicit: two complementary methodologies must be applied, one expert rule-based and one statistical. The absence of structural alerts from both is sufficient to conclude that the impurity is of no mutagenic concern (Class 5), with no further testing. A positive or equivocal outcome can instead be assessed through expert review, which must document the rationale for the final conclusion. In audits, the points checked are predictable: version and validity of the software used, handling of discordant outcomes between the two systems, qualification of the expert review and consistency between the (Q)SAR output and the class claimed in the dossier.
From class to limit: TTC and less-than-lifetime
For Class 2 and 3 impurities the reference is the threshold of toxicological concern (TTC): 1.5 µg/day for lifetime exposure, corresponding to a theoretical excess cancer risk below 1 in 100,000. For shorter treatment durations, ICH M7 allows higher acceptable intakes (less-than-lifetime approach, Table 2):
- treatment up to 1 month: 120 µg/day;
- 1 to 12 months: 20 µg/day;
- 1 to 10 years: 10 µg/day;
- over 10 years / lifetime: 1.5 µg/day.
When three or more Class 2 or 3 impurities appear on the drug substance specification, a limit on the total also applies: 5 µg/day lifetime, rising to 30, 60 and 120 µg/day for shorter durations. Class 1 impurities are excluded from the total and keep their own compound-specific limit. Watch out for a frequent mistake: less-than-lifetime limits apply to the total intended treatment duration of the product, not to the length of a single course of therapy.
Cohort of concern: when the TTC is not enough
Three structural families — aflatoxin-like, N-nitroso and alkyl-azoxy compounds — make up the so-called cohort of concern: high-potency mutagenic carcinogens for which TTC-based intakes are not considered adequate and a case-by-case approach is needed, with typically much lower limits. This is nitrosamine territory: for NDMA, NDEA and the drug-substance-related nitrosamines (NDSRIs), EMA and FDA have published dedicated limits and methodologies (including the Carcinogenic Potency Categorization Approach, CPCA) that complement — not replace — the M7 framework.
The four control options
Once the class is assigned and the limit set, the remaining decision is where to control the impurity. ICH M7 provides four options:
- Option 1: test the API against the specification, at or below the acceptable limit;
- Option 2: test a raw material, starting material or intermediate, or use an in-process control, at the API limit;
- Option 3: test upstream with a limit above the API limit, justified by demonstrated understanding of fate and purge across the process;
- Option 4: no analytical testing, based on process understanding alone, when purge guarantees levels below the limit.
Options 3 and 4 are the most efficient but also the most challenged during assessment: they require conservatively calculated purge factors and, where appropriate, confirmation through spike-and-purge experimental data.
The most common mistakes in audits and assessments
- Class claimed without evidence: no archived (Q)SAR output, or a single system used instead of two complementary ones;
- Class 4 used as a shortcut: the impurity's alert is not actually the same as the tested drug substance alert;
- Less-than-lifetime applied backwards: limit calculated on the course of therapy instead of cumulative exposure;
- Class 2/3 total ignored when three or more impurities sit on the specification;
- Option 4 without a documented purge assessment or with non-conservative purge factors.
GuideGxP recommendation
Treat the classification as a controlled deliverable, not a one-off exercise: for each impurity, archive the (Q)SAR output, expert review, class, limit and control option in a single living document, and reassess it at every synthesis change, new supplier or update of the cohort-of-concern compound lists. A traceable M7 assessment defends itself, in audits as in regulatory assessment.
If you want a ready-made system, our operational guide Elemental and mutagenic impurities and nitrosamines: risk assessment per ICH Q3D and M7 covers the whole flow — classification, TTC and CPCA limits, purge assessment and ready-to-use Excel and Word templates — to reach the dossier with complete evidence.
