Active Pharmaceutical Ingredients

Impurity profiles in APIs: named, unspecified and how to read a COA against a specification

A monograph tells you the floor. The impurity profile tells you what you are actually buying, and it belongs to a route of synthesis rather than to the molecule.

In short

An API impurity profile is the set of impurities that one specific route of synthesis is known to produce. The specification sets the control limits: an individual limit for each specified impurity, a general limit that applies to any unspecified impurity, and a limit for total impurities. The certificate of analysis reports what a single batch actually contained. Reading a COA against a specification means three checks in order: every impurity named in the specification appears on the COA with a result, anything not named sits below the unspecified limit, and the analytical method is capable of detecting the impurities that route can generate. A batch can satisfy a pharmacopoeial monograph and still be unsuitable for your product, because the monograph was never written around the route your supplier uses.

Where impurities come from

Impurities in an active pharmaceutical ingredient fall into families, and each family is controlled by a different test. Organic impurities include unreacted starting materials, intermediates carried through the sequence, by-products of side reactions, reagents and ligands, and degradation products that form during storage. Inorganic impurities include catalyst residues, inorganic salts and filter aids. Residual solvents arrive from the final steps of synthesis and purification, and are controlled by class rather than individually.

A certificate that reports assay, appearance and a single figure for total impurities is not reporting a profile. It is reporting a summary of a profile you have not been shown.

The vocabulary that causes most disputes

Four terms do most of the work, and they are routinely used loosely in commercial correspondence.

Specified impurity

An impurity listed individually in the specification, with its own acceptance limit. Being specified is a decision the manufacturer made because the route produces this impurity predictably.

Identified and unidentified

A specified impurity can be identified, meaning the structure is known and it carries a name or a code, or unidentified, meaning the structure has not been elucidated and the impurity is tracked by its relative retention time. An unidentified impurity with a stable RRT and a consistent limit is a normal, well controlled thing. It is not a red flag by itself.

Unspecified impurity

Any impurity not listed individually. It is controlled by a single general limit that applies to each unspecified impurity separately, not to their sum.

The common purchasing error is to read the unspecified limit as a catch-all that covers everything the manufacturer did not mention. It does not. If a route consistently produces a particular by-product, that by-product belongs in the specification under its own name with its own limit. Finding a recurring peak that nobody specified is a question about the specification, not only about the batch.

Reading the COA against the specification, line by line

Put the two documents side by side and work down.

  • Every specified impurity in the specification should have a matching line on the COA with a result against it. A missing line is not a pass.
  • Prefer numeric results to the word "complies". Complies tells you the batch passed. It does not tell you whether it passed comfortably or by a margin that will disappear on the next campaign, and it destroys your ability to trend across batches.
  • Find the largest unspecified impurity, usually reported by RRT. If the COA does not report it, ask for it.
  • Check that the method reference on the COA matches the method in the specification, including its revision number.
  • Check batch number, manufacturing date, retest date, batch quantity and manufacturing site. The site on the COA should match the site on the GMP certificate and the site described in the DMF you were given.
  • A COA retyped onto a trading company letterhead is not the manufacturer's COA. Ask for the original document, and compare the two.

The profile belongs to the route, not the molecule. Two manufacturers making the same API by different routes will have genuinely different impurity profiles, and both can comply with the same monograph. The same manufacturer moving a step to a different site, or changing a key raw material supplier, can shift the profile as well.

Why a second source is not a drop-in substitution

Before you can treat two suppliers as interchangeable you need comparative impurity data, confirmation that each analytical method can detect the other supplier's impurities, and clarity on what your filing route requires when the source changes. Get a written change notification commitment into the quality agreement rather than relying on an assurance in an email.

A limit is only as good as the method

A specification limit is meaningless if the method cannot see the impurity it governs. Ask for the analytical procedure itself rather than its name, evidence that it is stability indicating (normally forced degradation data showing that degradants resolve from the main peak), the system suitability criteria, and a written description of how unknown peaks are handled.

Reporting, identification and qualification are three separate decisions with three separate thresholds in the applicable guideline, and they are frequently conflated in supplier correspondence. When a supplier quotes a threshold to you as a number, treat it as a claim to verify against the current guideline for your market and dosage form rather than as a settled fact.

Impurities controlled outside the ordinary framework

Some impurities are not discharged by a general impurity limit at all. Mutagenic impurities, elemental impurities and residual solvents each sit under their own control approach with their own justification requirements. If any step in the route carries a structural alert, expect a documented risk assessment and, where required, a dedicated method operating at a far lower level than the ordinary unspecified limit. A supplier who answers a mutagenic impurity question by pointing at the total impurities figure has not answered it.

What to ask for

  1. The full impurity profile for the exact route and the exact site you will be supplied from.
  2. The specification, with specified impurities named, coded or given by RRT.
  3. Certificates of analysis for three consecutive commercial batches, issued by the manufacturer, with numeric results.
  4. The analytical procedures and a validation summary.
  5. Stability data on the API in its shipping pack, showing which impurities grow over time.
  6. A written commitment to notify you before any change to route, site, key raw material supplier or purification step.

A supplier who cannot produce the first two items is not showing you a manufacturer's package. The same discipline applies on the excipient side, where the documentation culture is generally weaker and the distribution chain is longer. If you would like the assessment done independently, that is what our sourcing consultancy work covers, and the document set it rests on is set out in DMF, CEP and COA.

Frequently asked questions

What is the difference between a specified and an unspecified impurity?

A specified impurity is listed individually in the specification and carries its own acceptance limit. It may be identified, meaning the structure is known and it has a name or code, or unidentified, meaning the structure is unknown and it is tracked by relative retention time. An unspecified impurity is any impurity not listed individually, controlled by a single general limit that applies to each one separately. An impurity that the route produces consistently belongs in the specification with its own limit rather than in the unspecified category.

Is "complies" acceptable on a certificate of analysis?

It is acceptable for a qualitative test such as appearance or an identification test. For impurities it is weak. "Complies" confirms the batch passed but hides how close it came to the limit, and it removes the ability to trend results across batches. Ask for numeric results for individual impurities, total impurities and residual solvents.

Does a second supplier with the same pharmacopoeial specification give me the same material?

No. A monograph sets a floor that many materials can meet, and it does not describe the route of synthesis. Two manufacturers using different routes will produce different impurity profiles while both comply. Before treating a second source as interchangeable, compare full impurity profiles, confirm the analytical methods are capable of seeing each other's impurities, and check what your filing route requires for a change.

How often should an impurity profile be re-checked?

Review it whenever the manufacturer changes the route, the site, a key raw material supplier or a purification step, and review it periodically against incoming batch data. A quality agreement that obliges the manufacturer to notify you in advance of such changes is more useful than testing after the fact.

📚
Written by

Dr. Majjid A. Qaria

Dr. Qaria holds two doctorates and a postdoctoral research background in the medical sciences, across molecular microbiology and cancer cell biology. He has personally sourced and supplied active pharmaceutical ingredients and excipients, and leads sourcing and consultancy at NJMC from Nanjing. Research background and publications are on his profile site.

Related pages

Tell us what you need to source

Send the product or molecule, the grade, the quantity and the destination market. You will get a considered answer about what can be sourced and what documentation comes with it.

Send an enquiry