The 1.5 Microgram Number Is a Lifetime Default. Most Programs Aren't Lifetime.

Ask a CMC team what ICH M7 requires and most will answer with one number: 1.5 micrograms per day. It is the guideline's most quoted figure, and it is also the most commonly misapplied one. The 1.5 mcg/day Threshold of Toxicological Concern is a lifetime-exposure default — calibrated to a patient taking the drug daily for 70 years. Most real development programs, from early-phase trials through many approved short-course and orphan therapies, are not lifetime exposure. A risk assessment that reaches for the headline number without first working through classification and duration is not being conservative. It is very often applying the wrong limit.

Classification comes before the number

ICH M7 applies specifically to DNA-reactive, or mutagenic, impurities — a narrower category than ICH Q3D's elemental impurities, which is a different guideline addressing a different hazard entirely. Before a duration-based limit means anything, an impurity has to be classified. The process starts with a database and literature search for existing carcinogenicity and bacterial mutagenicity data. Where that data classifies the impurity outright, it lands in Class 1 (a known mutagenic carcinogen, including the cohort-of-concern chemicals such as many nitrosamines) or Class 2 (a known mutagen of unknown carcinogenic potential). Where data is unavailable, a computational (Q)SAR assessment predicting bacterial mutagenicity routes the impurity toward Class 3, 4, or 5 instead.

What each class actually means for control

• Class 1 — known mutagenic carcinogens. Controlled at or below a compound-specific acceptable limit, not the generic TTC.
• Class 2 — known mutagens, carcinogenicity unknown. Controlled on a generic basis using the TTC-based default intake limits.
• Class 3 — alerting structure, no mutagenicity data. Testable: a negative Ames result reclassifies to Class 5; a positive result reclassifies to Class 2 and TTC control applies.
• Class 4 — alerting structure shared with a drug substance already shown non-mutagenic. Treated as non-mutagenic once that read-across is documented.
• Class 5 — no alerting structure, or negative data. Standard impurity-limit controls only — no TTC-based calculation needed.

The TTC is not the assessment. It is what a Class 2 or 3 impurity defaults to once classification says TTC-based control is the right tool — and duration decides which TTC. Why classification comes first

The adjustment most assessments skip: less-than-lifetime

The 1.5 mcg/day figure assumes daily dosing across a 70-year lifetime. ICH M7's less-than-lifetime (LTL) provisions recognize that most real exposure isn't that, and allow a higher permitted daily intake the shorter the treatment duration, because the acceptable cumulative lifetime dose gets distributed over far fewer actual dosing days. A single mutagenic impurity taken for under a month can be controlled to a substantially higher daily limit than one taken for a decade or more, and the guideline sets out the duration tiers to calculate it. Where multiple mutagenic impurities co-occur in the same product, the cumulative allowable limit across all of them is lower than any single-impurity tier, but it is still duration-adjusted — not the flat lifetime TTC applied to the sum. A risk assessment that skips this step and controls a Phase 1 or short-course product to the lifetime number is not wrong on safety; it is leaving a guideline-sanctioned, scientifically defensible limit on the table, and often driving CMC teams to chase specification tightening the program's actual exposure duration never required.

None of this is exotic regulatory CMC work — it is a bounded, well-precedented analysis most quality and toxicology teams already have the underlying data for. The gap is usually sequencing: reaching for the lifetime TTC as a reflex instead of working classification and duration in order. Teams that route the assessment through their nonclinical development function early, before the CMC specification is drafted, are the ones who end up with a limit the guideline actually supports — not a stricter one nobody asked for, and not a looser one the classification never justified.