Photostability of Bile-Acid APIs: What ICH Q1B Requires
Share
Question eleven in a supplier questionnaire is usually "is the material light sensitive?" The answer that comes back is usually "we use amber drums". Those two sentences do not connect. One asks for a property of the molecule under defined conditions. The other describes a packaging choice, and a packaging choice can be made for reasons that have nothing to do with photostability: customer habit, warehouse convention, the colour of drum that happened to be in stock.
The gap matters because a light-protection claim propagates. It ends up in the customer's regulatory file, in their handling instructions, in the deviation they raise when a drum arrives with a white liner. If the claim was never supported by a study, every downstream document inherits an assumption, and the assumption outlives the person who made it.
Two questions the guideline keeps separate
ICH Q1B does not ask one question. It asks two, answered at different times by different people.
The first is stress testing. Expose the substance to light far beyond anything it will meet in commerce and see what breaks. The purpose is to identify degradation pathways and to prove that the analytical method can see the products. This is method work wearing the clothes of stability work. A method that cannot resolve a photo-degradant will report a flat result no matter how long the lamp stays on.
The second is confirmatory testing. One batch, in the actual marketed pack, at the standard condition, for the standard dose of light. Two batches if the substance is already known to be light sensitive. The purpose is to decide whether the pack is adequate.
Suppliers routinely answer the first and quote it as if it were the second. A forced-degradation result showing eight per cent loss after forty hours of intense UV says nothing about whether a 25 kg drum survives a year under a skylight.
The standard dose of light
| Parameter | Confirmatory condition | What it means in practice |
|---|---|---|
| Visible light | Not less than 1.2 million lux hours | About three months of continuous 500 lux exposure, delivered in days |
| Ultraviolet | Not less than 200 W·h/m² over 320–400 nm | The part of daylight that does chemistry rather than illumination |
| Temperature | 25 °C/60% RH or 30 °C/65% RH | Controlled, because the lamp adds heat of its own |
| Control | Identical sample wrapped in foil | Without it, any change is uninterpretable |
Two practical notes. The exposure is cumulative rather than continuous, so a chamber can deliver the dose inside a fortnight. And the dark control is not bookkeeping. If exposed and wrapped samples drift by the same amount, the cause is heat or moisture, and the study has answered a different question from the one written on its cover.
What light can actually do to a bile acid
Bile acids are not dyes. Ursodeoxycholic acid has no conjugated system: the carboxyl group absorbs below about 210 nm, and the two hydroxyls do not absorb in the terrestrial UV range at all. On paper, direct photodegradation should be slow.
The qualifier is "direct". Two indirect routes remain, and both are real at industrial scale.
The first is photo-oxidation. Trace peroxides, dissolved oxygen and a sensitising impurity can generate radicals under UV, and a radical will abstract a hydrogen from a steroid nucleus without needing a chromophore on the molecule. The products are hydroxylated and keto analogues, which is awkward because they belong to the same family that already appears in the impurity profile. A small increase is easy to miss and easier to dismiss.
The second is thermal. A light cabinet is a warm cabinet. Without a dark control, a rise in 7-keto content gets attributed to photons when it may belong to the lamp.
7-Ketolithocholic acid is the more interesting case, because the ketone gives a weak absorption around 285 nm. Weak absorption means slow direct reaction, not no reaction, and it makes the compound a better candidate for a sensitised pathway than its 7-hydroxy parent. If a customer asks which of the two products in a bile-acid range needs a light claim first, this is the reasoning that answers it.
An illustrative result, and how to read it
The table below shows the shape of a confirmatory dataset. The values are illustrative: they demonstrate the arithmetic, and they do not describe any particular batch.
| Test | Dark control | Exposed | Reading |
|---|---|---|---|
| Assay, dried basis | 99.9% | 99.7% | Inside method variability |
| Total impurities | 0.19% | 0.22% | Inside batch-to-batch range |
| 7-keto analogue | 0.05% | 0.07% | Detectable, not directional |
| Loss on drying | 0.40% | 0.41% | No moisture effect |
| Appearance | White powder | White powder | No colour shift |
The correct reading is not "0.2 per cent loss, acceptable". It is that every change sits inside the variability the method and the process already show, so the study supports neither a sensitivity claim nor a guarantee of stability. That is a legitimate outcome, and it should be written down as one rather than converted into a label line for safety.
Three answers, and what each obliges
Protect from light. Then every pack carries the protection, including the courier sample and the replacement drum. The handling instruction becomes part of the specification, and a white liner leaving the site has breached the claim before the customer opens the lid.
No special precautions. Then the file says so, with the condition and the pack named. Silence is not this answer. Silence gets read as "unknown" by a customer's QA and then re-read as "risky" by their regulatory lead.
Inconclusive. Usually a dosing or control problem. The honest move is to repeat the study, not to convert the uncertainty into a conservative label that nobody can later remove.
Five questions, one message
| Question | Why it separates real work from a sentence |
|---|---|
| Was the study run on the substance or on a drug product? | Product data does not transfer to a different formulation or pack |
| Was the marketed pack used? | A study in a petri dish does not describe a drum |
| What dose and which light source? | The option used determines whether the result is comparable with the customer's own study |
| Was there a dark control? | Without one, photochemical and thermal effects are inseparable |
| Where is the conclusion written? | A conclusion that exists only in an email cannot be filed |
Each has a short factual answer. A supplier who can produce them has done the work. A supplier who answers with a drum colour has not, and the useful inference is about documentation discipline rather than about photochemistry.
The pack is the other half of the study
A confirmatory study tests a pack, not a molecule. Change the liner from aluminium to polyethylene and the study no longer describes what ships. This is why the pack description belongs in the stability section and not only in the packaging section, and why a packaging change is a change-notification event rather than a purchasing decision made by whoever finds a cheaper drum.
The same logic runs through moisture, which is why hygroscopicity and photostability are usually answered together by anyone who has actually run a stability programme, and separately by everyone else.
Where this leaves a supply-chain partner
We do not present a light-protection claim as a default. Where a study has been run on a pack we ship, the conclusion travels with the pack description and the customer can file it. Where it has not, we say the data does not exist yet instead of borrowing a sentence from a pharmacopoeia. That position costs a little in the first email and saves a great deal at the customer's audit.
If your specification carries a photostability statement, ask what it rests on. If it carries none, ask whether that is a finding or a gap. They are different answers, and only one of them survives being questioned by somebody who was not in the room when it was written.
What the customer's own file may already contain
Before commissioning anything, it is worth checking what the customer already holds. A formulator who has run a photostability study on the finished product has data on the substance as formulated, in that container, under ICH Q1B conditions. It is not the same study as one on the substance alone, and it is frequently enough to answer the question for the customer's own file.
The supplier's obligation is different: to describe the material and the pack as supplied. Where the customer's product study exists, the two documents sit beside each other rather than replacing one another. Where only the product study exists and the customer intends to qualify a second source, the substance-level data becomes necessary again, because the product study was generated with the first source in the drum and cannot be assumed to describe a different impurity profile.
The EDQM publishes the monographs and general methods that most European customers will compare against, and the ICH Q1B text itself is short enough to read before accepting anybody's summary of it. Neither document will tell you what a particular supplier did, but both tell you what a complete answer looks like, and the EudraLex volume 4 annexes show where the resulting data sits in a European filing.
A note on the light source
Two options are permitted and they are not interchangeable in a dataset. Option 1 uses a combination of daylight and ultraviolet lamps chosen to simulate outdoor daylight. Option 2 uses cool white fluorescent lamps with a defined ultraviolet component. Both can deliver the required dose, and their spectral distributions differ, which matters when the absorption of the material sits near the edge of a lamp's output.
For a comparison across two suppliers, ask which option was used. If the answer differs on each side, both datasets can still be valid and cannot be plotted on one graph without a note. That is a small point that costs a sentence at the time and a week at the customer's technical review, where somebody will ask why the numbers do not line up and nobody will remember why.
Handling and storage instructions derived from the study belong with the pack description. The moisture half of the same question is covered in the article on hygroscopicity and packaging, and where the assay method has not been shown to detect photodegradants, a negative result means less than it appears to, which is the argument in stability-indicating assay methods. Where the data is needed for an ongoing programme rather than a single filing, retention samples and ongoing stability covers what has to be held and for how long.
Frequently asked questions
Does every API need a photostability study?
ICH Q1B applies to new active substances and to new drug products. For an established substance already described in a pharmacopoeia, the position depends on the jurisdiction and on what the customer's own file already contains. The practical answer is that a buyer who intends to file the material needs the study or a documented justification for its absence, and a supplier who cannot say which of the two applies has not settled the question.
How much light is 1.2 million lux hours?
It is an exposure dose, not a duration. A cabinet delivering 6,000 lux reaches it in about 200 hours, roughly nine days. Under 500 lux of ordinary office lighting the same dose takes about 100 days of continuous exposure. Because it is cumulative, the study compresses months of warehouse life into a fortnight.
What is the dark control for?
It separates photochemical change from thermal change. A light cabinet is also a warm cabinet, so an exposed sample that loses 0.2 per cent could be reacting to photons or to the lamp's heat. If the wrapped control loses the same amount, the cause is not light, and the conclusion written into the file has to change accordingly.
Can a forced-degradation study replace confirmatory testing?
No. Forced degradation establishes degradation pathways and proves the method can detect the products. Confirmatory testing establishes whether the marketed pack protects the material at the standard dose. They answer different questions, and a supplier who offers the first as evidence for the second has misunderstood the guideline rather than evaded it.
Our supplier packs in amber drums. Is that enough?
An amber drum is a control measure, not a result. It may well be the right choice, but the question a customer's file has to answer is whether the pack as shipped was tested or whether the colour was inherited from convention. Ask for the pack description used in the study and compare it with the pack on the purchase order.
Does a light-protection claim change the specification?
It changes handling, not the numeric limits. The specification still states assay, impurities and the rest. What the claim adds is an obligation on packaging and storage that has to be met by every batch, including samples shipped by courier, which is where the claim is most often broken without anyone noticing.