Nitrosamine Risk Assessment: An ICH M7 Workflow

You get a nitrosamine signal on an API. QA wants an ICH M7 classification, the structural basis written down, and a second, independent computational opinion: this week, not next quarter. And if the compound is an NDSRI (nitrosamine drug substance related impurity), there is no read-across standard sitting in a database waiting for you.

This article walks the assessment the way a practitioner actually runs it, with real results from CovaSyn's mutagenicity assessment. It also shows the part most vendor material hides: where the statistical model tells you it does not know.

Why NDSRIs broke the old workflow

Classic nitrosamines such as NDMA and NDEA are small, well studied, and have carcinogenicity data you can cite. NDSRIs are different. They are formed when the API's own secondary or tertiary amine nitrosates, so the impurity is a drug-sized molecule for which there are usually no Ames or carcinogenicity data and no structural analogue with a published TD50.

That combination is what forces you into a computational assessment. ICH M7(R2) already expects one, and both FDA and EMA now expect a documented structure-based rationale for NDSRI limits (the agencies' Carcinogenic Potency Categorization Approach is built on exactly this kind of structural reasoning). The practical problem: the QSAR models you would normally lean on were trained on Ames data that contains very few drug-sized nitrosamines.

The ICH M7 dual assessment, concretely

ICH M7 asks for two complementary (Q)SAR methodologies: one expert-rule-based, one statistical. Two methods, one conclusion, both documented.

Here is the assessment of NDMA (CN(C)N=O):

• Computed assessment: alert and statistical model positive (class 3 without data; NDMA has data and is class 1)
• structural alert: nitrosamine, evidence high
• mechanism: CYP2E1 activation to the diazonium ion, then DNA alkylation; reference: IARC Monographs
• statistical Ames score: 0.6905
• confidence: 0.628

Expert rule positive, statistical method positive, both agree: both indicate mutagenic potential. A prediction alone without data gives class 3 at most under ICH M7, because classes 1 and 2 require data. NDMA has those data: it is a known mutagenic carcinogen, ICH M7 class 1, with a published acceptable intake of 96 ng/day. The TTC does not apply here. N-nitroso compounds sit in the ICH M7 cohort of concern, so the generic 1.5 µg/day TTC is off the table and you need a compound-specific acceptable intake. The assessment should state this explicitly rather than quietly print a number: TTC not applicable, control or elimination required.

The contrast case: an alert the statistics contradict

Assess aniline (Nc1ccccc1) the same way:

• Computed assessment: alert positive, statistics negative (class 3 without data; Ames negative, therefore class 5)
• structural alert: aromatic amine, evidence high, mechanism: metabolic activation to the nitrenium ion, then DNA adduct; reference Benigni & Bossa, 2006
• statistical Ames score: 0.1677 (negative)
• confidence: 0.848

The alert fires. The statistical model disagrees. Without data this would be class 3: alerting structure, unrelated to the drug substance, no mutagenicity data; aniline, however, has negative Ames data, which makes it class 5. This matters more than it looks: a system that maps "alert present" straight to a mutagen class is a substructure matcher with a compliance label on it, and reviewers spot that quickly. The two assessments have to be allowed to disagree, and the disagreement has to be visible in the record.

Worked example: an NDSRI on a secondary-amine API

Propranolol is a secondary amine, so the plausible NDSRI is N-nitroso-propranolol (CC(C)N(N=O)CC(O)COc1cccc2ccccc12). Assessing the impurities together with the API structure gives per-impurity classification plus Tanimoto similarity to the parent, which is what you need to argue that an impurity is drug-related rather than a process contaminant.

Five impurities against propranolol:

Summary: 5 assessed, three agreeing positive findings (all three nitrosamines), one conflicting, one without alert, highest-risk impurity NDMA, TTC not applicable. The assessment column follows from existing data and regulatory values, not from the computation.

Two things are worth reading off this table. First, the similarity column separates the NDSRI (0.563 to the API) from the small-molecule nitrosamines (0.02 to 0.07) without you having to argue it by hand. Second, the Class 5 assignment for benzoic acid comes with a low confidence of 0.438: the model is telling you the negative call is weakly supported, which is the correct posture for a "no alert, no concern" verdict.

The honest part: what the domain check says about nitrosamines

This is where many nitrosamine assessments stop talking. The applicability-domain check for the Ames endpoint on the same structures:

NDMA is itself in the training data, so its nearest similarity is 1.0. It is still out of domain because feature coverage stays below the threshold: a molecule with five heavy atoms carries too few structural features for the model to place it reliably. For NDMA the existing Ames data and the established limit decide, not the score. N-nitroso-propranolol is in domain, but its nearest neighbour reaches only 0.56. That is the actual NDSRI problem: drug-sized nitrosamines are rare in public Ames data, and a close structural relative is missing.

So the Ames score of 0.7206 for N-nitroso-propranolol is not evidence on its own. It is a directional signal that happens to agree with the expert rule. The defensible conclusion rests on the structural alert, its mechanism and the IARC reference: the statistical method is there because ICH M7 asks for a complementary method and because agreement between two independent methods is itself informative.

For calibration on where the underlying model is solid: on a held-out test set split by scaffold, our Ames classifier reaches 0.736 balanced accuracy (n_test = 1089; 0.848 on a random split). That is a respectable general mutagenicity model. For NDSRIs without a close neighbour in training, treat the figure as an upper bound, not a promise. Quote the benchmark, then quote the domain result next to it.

A workflow that survives review

1. Enumerate. List every nitrosatable amine in the drug substance, the route, and the formulation (including nitrite in excipients). This step is chemistry, not software. 2. Classify. Assess the enumerated impurities together with the API structure, so you get class, alert, mechanism, reference and API similarity in one record. 3. Record the domain check. Check the domain for every compound where the statistical score is load-bearing. Attach the result and the nearest training-set similarity to the assessment. 4. Resolve disagreements explicitly. Where the expert rule and the statistical method conflict, ICH M7 expects expert judgement and a documented rationale. Write it down; do not average the two. 5. Set the limit off the right framework. For cohort-of-concern nitrosamines, that is a compound-specific acceptable intake (CPCA or a read-across/TD50 argument), not the 1.5 µg/day TTC. 6. Confirm analytically. A confirmed or assumed mutagenic potential drives an LC-MS/MS method at the AI-derived limit and a control strategy. The in silico work targets the analytical effort; it does not replace it.

What this does not tell you

• It is not a potency prediction. Agreeing methods indicate mutagenic potential; an acceptable intake in ng per day needs carcinogenicity data, read-across or a CPCA category.
• The statistical Ames score is weakly supported for nitrosamines: NDMA is out of domain, and NDSRIs rarely have close neighbours in training. Treat it as corroboration of the alert, never as a standalone negative.
• A negative result is weaker than a positive one. Absence of an alert plus a low QSAR score, at confidence 0.438 as in the benzoic acid case, is a screening outcome, not a clearance.
• An alert written for N-nitrosamines does not necessarily cover N-nitrosoamides. Know what your alert set covers before you claim completeness.
• Nothing here substitutes for an Ames test, an enhanced Ames protocol where required, or the analytical confirmation the agency will ask for.

Frequently asked questions

Related reading

• Forced degradation and impurity identification with LC-MS
• ICH M7 Classification: Class 1 to 5, by Example
• PBT Assessment for Pharmaceuticals: Very Toxic Is Not PBT