Ask a CMC team how they set a release specification, and the honest answer is often: we started from the USP monograph. That is not wrong — it is incomplete. ICH Q6A and its biotechnology counterpart Q6B were built around a different premise: a specification is a set of tests and limits the manufacturer proposes and justifies against its own process and stability data. The monograph is where you start. The decision trees in Q6A's attachments are where you prove you did not stop there.
What a specification actually is
Q6A defines a specification plainly: a list of tests, references to analytical procedures, and appropriate acceptance criteria, proposed and justified by the manufacturer and confirmed by the regulatory authority as a condition of approval. That last clause is the one teams skip past. The specification is not handed down by a pharmacopeia — it is built, defended, and then locked as an approval commitment. Compliance with a compendial monograph does not establish that a product meets every quality attribute that matters for your process; conversely, a compendial test that has nothing to do with how your product is made or degrades may not belong in your specification at all, provided you can justify leaving it out. Either direction requires the same thing: a documented rationale tied to your own CMC regulatory strategy, not a citation.
Universal tests, specific tests, and where the trees come in
Q6A's universal tests are close to non-negotiable: description, identification, assay, and impurities apply to essentially every new drug substance and the products made from it. The judgment work sits in the specific tests — the ones that only apply case by case, decided by dosage form, route of administration, and manufacturing process. This is where the decision trees earn their place in the guideline: they walk a reviewer, and a sponsor, through the logic for whether a given attribute needs a test, which method fits it, and how tight the acceptance criterion should be.
- Dissolution vs. disintegration. The decision tree asks whether dissolution testing is warranted by the dosage form and its release behavior, rather than defaulting to whichever test the monograph happens to specify.
- Degradation products and impurities. New impurities identified in your own stability program get evaluated and limited on their own merits, reconciled against impurity thresholds under ICH Q3A/Q3B — not waived because the monograph is silent on them.
- Polymorphic form and particle size. Tested and limited only where the attribute is shown to affect bioavailability, manufacturability, or stability for your specific product — the trees exist so that decision is documented, not assumed.
- Microbial limits and related attributes. Case-by-case, driven by dosage form, route of administration, and the manufacturing process's actual bioburden risk.
A monograph tells you what a class of products has needed before. It cannot tell a reviewer why your process, your impurity profile, and your stability data support the limits you propose. Only the justification does that. Why the decision-tree logic matters
Where Q6B pulls in a different direction
Q6B governs proteins, polypeptides, and the biologics regulatory work built around them, and it starts from a harder constraint: many biotechnology-derived products cannot be fully characterized by end-product testing alone. A release specification that tried to carry the entire quality-control burden the way a small-molecule spec can would be either impossibly broad or falsely precise. Q6B's answer is to distribute that burden — leaning on manufacturing process consistency, in-process controls, and a characterization package built from clinical and nonclinical lot experience to support acceptance criteria that a specification alone could not justify. Potency and product-related variants get particular attention, because a biologic's activity and its degradation pathways rarely map onto Q6A's chemical-substance assumptions. The practical result: a Q6B specification package leans as much on your process validation and characterization data as on the release tests themselves, and a team that treats Q6B as "Q6A for biologics" under-builds exactly the evidence a reviewer expects to see.
- Split universal from specific tests. Confirm description, identification, assay, and impurities are covered, then decide which specific tests your dosage form and route actually require.
- Walk the decision trees before defaulting to the monograph. For every attribute in question, work the Q6A logic — or Q6B's process-and-characterization approach — rather than copying the compendial default.
- Trace every limit to your own data. Justify acceptance criteria against your process capability and stability data, tightening where warranted and documenting the rationale either way.
- Reconcile upstream. Confirm each test's analytical procedure traces to a defined ATP and that impurity limits align with Q3A/Q3B before the specification locks.
None of this is exotic work. It is a documented, attribute-by-attribute justification exercise — the kind of thing that is easy to shortcut under deadline pressure and expensive to reconstruct after an information request. Teams that treat the specification as a compliance copy-paste from the monograph pass their own internal review and still draw questions in an NDA or BLA cycle. Teams that treat it as a Q6A/Q6B justification exercise, attribute by attribute, are the ones whose specification package survives contact with a reviewer on the first pass.
Frequently asked questions
Does meeting the USP or Ph. Eur. monograph satisfy ICH Q6A?
Not by itself. ICH Q6A treats compendial compliance as a floor, not a ceiling: if your manufacturing process or stability data point to an impurity, degradation product, or tighter limit the monograph doesn't address, you still have to add it. The specification is justified against your own process — the monograph is a reference point, not a substitute.
What is the difference between a universal test and a specific test under Q6A?
Universal tests — description, identification, assay, and impurities — apply to essentially every new drug substance and drug product. Specific tests apply case by case, decided by dosage form, route of administration, and manufacturing process: dissolution, particle size, polymorphic form, microbial limits, and similar attributes that only matter for some products.
Why does ICH Q6B treat specifications differently from Q6A?
Biotechnological and biological products often can't be fully characterized by physicochemical testing alone. Q6B leans more heavily on manufacturing process consistency, in-process controls, and characterization and clinical-lot data to support acceptance criteria — so the release specification carries less of the total quality-control burden than it does under Q6A.
Sources & further reading
- FDA. Q6A Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products: Chemical Substances — Guidance for Industry. fda.gov
- FDA. Q6B Specifications: Test Procedures and Acceptance Criteria for Biotechnological/Biological Products — Guidance for Industry. fda.gov
- Federal Register. International Conference on Harmonisation; Guidance on Q6A Specifications (65 FR 83041, Dec. 29, 2000). federalregister.gov
This article is provided for general informational purposes and reflects the regulatory landscape as of August 2026. It is not legal or regulatory advice. Confirm current ICH Q6A/Q6B expectations with FDA, EMA, or qualified counsel before acting.
