ICH Q11 Starting Material: How Far Upstream Is Far Enough
Share
Every drug substance file contains a sentence that decides how much of the process the holder will have to describe, control and defend for as long as the file lives. That sentence names the starting material.
Put the line too low and the file describes chemistry the applicant never controlled. Put it too high and the file has to carry steps that belong to somebody else's plant. Both errors are expensive, and the second one is the one regulators most often correct.
The mistake in the intuitive definition
The intuitive definition is "the first thing we buy". That is a purchasing fact, not a regulatory one, and it fails as soon as the first thing bought is a complex intermediate made by a single supplier to a specification nobody else uses.
The regulatory definition is about significance: the point in the process from which the material's contribution to the final impurity profile is understood and controlled. A substance that carries its own impurity burden into the process is a starting material, not a raw material, because what happens to that burden downstream is exactly what the file has to explain.
Four tests that place the line
Test one: does the material carry impurities forward
If the substance contributes impurities that survive into the drug substance, or that require specific purge steps to remove, it belongs inside the described process. A purchased compound whose impurities are irrelevant to the final profile is a reagent. One whose impurities are the reason your purification works is a starting material.
Test two: how many independent sources exist
A starting material available from several qualified suppliers at a stated specification is a commodity with a market. A starting material available from one supplier, to an in-house specification, is a dependency that the file will encode permanently.
This is the test that has a supply chain answer rather than a chemistry answer, and it is the one most often skipped in technical discussions. A defensible starting material should be something the applicant could, in principle, buy from somebody else.
Test three: how much purification follows the line
Steps after the starting material are the applicant's responsibility, and they are also the steps that remove what came before. A short process with one crystallisation at the end has little capacity to purge upstream impurities, which argues for defining the line earlier and describing more.
A long process with several purification points can defend a later line, because the file can show where each upstream impurity goes.
Test four: is the substance described in a pharmacopoeia
A starting material that already has a monograph carries an external specification that a regulator can recognise. That does not automatically make it an acceptable starting material, but it does mean the file is arguing about a defined substance rather than about a proprietary one.
What the three common positions cost
| Where the line sits | What the file must contain | Where it fails |
|---|---|---|
| At a simple purchased reagent | Most of the process, several steps of own development | Over-documentation; slow to write, slower to change |
| At a complex purchased intermediate | Later steps only, with purge data for what arrives | Regulator asks why earlier steps are excluded; line moves |
| At a substance with one supplier and no public specification | Thin process description plus a dependency | Both: scope challenged and supply fragile |
The third row is the position that looks efficient and ages worst. It minimises what the applicant must describe while maximising what the applicant must rely on.
What happens when the line is challenged
A challenge is not usually framed as a rejection. It arrives as a question: please justify why the process begins at this point, and please provide information on the steps before it.
Answering that question late means collecting data from a supplier who may not want to provide it, re-describing chemistry that was never under the applicant's control, and re-writing the impurity story around a longer process. The work is not technically hard. It is slow, and it happens while a customer waits.
Applicants who anticipate the question write the justification into the file at submission: why this point, what comes before, how the incoming impurities are controlled, and what evidence shows the purge is effective. That paragraph prevents a round.
Why this matters to a buyer
The starting material decision is not visible on a certificate of analysis, and it changes what a supplier can honestly promise.
A holder who defined the line early can answer questions about the origin of an impurity, because the steps that create it are inside the described process. A holder who defined it late can only report what arrives. Both can produce material within specification; only one can explain a shift in the impurity profile when something upstream changes.
For substances made by more than one route, the line also shapes the comparison. The impurity consequences of a chemical route against an enzymatic route are not equivalent, and the difference is set by where each process starts. That comparison is set out in how UDCA is made.
For intermediates whose position in the chain is itself the commercial point, see 7-KLCA as the hub intermediate. And when a change of source is being considered, adding a second API source covers what the change costs in documentation.
Primary sources: ICH Q11 for the starting material and process development expectations, the FDA Q7 guidance for the GMP that applies from that point onward, and EudraLex Volume 4 for how the European inspection framework treats the boundary.
Frequently asked questions
Does ICH Q11 name a fixed number of steps before the drug substance?
No. It sets out how to justify the choice rather than a step count, and the justification rests on impurity significance, control and the ability to describe the process. Step counts are a shortcut that regulators do not accept on their own.
Can a purchased intermediate with a pharmacopoeial monograph be a starting material?
It can be proposed, and having a monograph helps because the substance is externally defined. Acceptance still depends on the rest of the argument, particularly on whether its impurities are controlled and purged downstream.
What if the only supplier of my starting material will not share process information?
Then the file cannot describe what happens before the line, which is exactly the position that invites a challenge. The practical options are to move the line upstream, change the source, or accept a file that will be questioned.
Does the starting material definition affect change control later?
Yes, substantially. A change of starting material source or specification is a change to the registered process, and the further upstream the line sits, the more of the process is protected from that kind of change.
Is a longer described process always safer?
Safer in review, more expensive to maintain. Each described step brings controls, validation and change obligations. The aim is a defensible line, not the longest possible one.
How does this apply to a substance used outside medicinal products?
It may not apply at all. Where the material is not used in a regulated medicinal product, the specification and the analytical data carry the weight, and the starting material argument belongs to a different conversation.