Real-time vs accelerated stability for registration, which one your dossier needs

In an export registration timeline, one item is usually the longest and least compressible, and it is the stability data.

Everything else in a dossier can be assembled, bought, expedited or paid for. Real-time stability data accrues at one month per month, and no amount of money changes that.

So the question “will accelerated data be accepted?” is often the single most consequential question in an export plan, and brands routinely ask it after committing to a launch date.

This article sets the two methods side by side, says what each one can and cannot support, and gives the order in which to find out which you need.

This is orientation, not legal, regulatory or trade advice. Acceptance of study types differs by destination, product category and route, and changes. Confirm the requirement with the importing authority or a competent regulatory consultant before commissioning any study.

The two methods, briefly

Real-time stability. The finished product, in its final pack, stored at the conditions it will actually experience, tested at intervals until the claimed shelf life is reached. Typical pull points at 0, 3, 6, 9, 12, 18 and 24 months.

Accelerated stability. The same product and pack, stored at elevated temperature and humidity, tested over a shorter period, commonly six months, with the result used to support a provisional shelf life by extrapolation.

Both test the same parameters: assay of the declared actives, appearance, moisture, disintegration or dissolution, microbiology, pH for liquids, and peroxide or acid value for oil-based products.

We have written separately about what stability testing is and what it measures. This article is about choosing between the two.

What accelerated data can support

Its real strengths, and they are genuine.

Speed. A provisional shelf life in months rather than years, which is often what makes a launch possible at all.

Comparison between options. Two packs, two formulations or two excipient systems can be screened against each other quickly. This is a good use of the method and an underused one.

Early warning. A product that fails accelerated conditions badly will usually have a real-world problem too, and finding that in four months rather than eighteen is worth the study on its own.

And a reasonable basis for a provisional claim in markets that accept it, with real-time running in parallel.

What accelerated data cannot do

Four limits, and they are why the question of acceptance exists at all.

Extrapolation assumes the degradation mechanism is the same at both temperatures. For botanicals it frequently is not. Heat can trigger reactions that do not occur at ambient conditions, producing failure that overstates real risk; and slow processes, moisture migration through a seal, microbial growth, physical settling, can be understated by a short hot exposure.

Physical changes do not extrapolate at all. Caking, natural pigment colour shift, oil separation, tablet hardening and capsule shell changes follow their own kinetics. A smooth chemical extrapolation says nothing about them.

Microbiological outcomes do not extrapolate reliably either, and in some products microbiology is the limiting parameter rather than assay.

And it is a prediction, which a regulator may accept provisionally and may not accept as final substantiation.

Which one your dossier needs, the determining factors

Four, and you have to establish all of them.

The destination. Some regimes require real-time data for the full claimed shelf life. Some accept accelerated for a provisional period with real-time to follow. Some, for low-risk food supplement categories, accept a well-reasoned justification supported by accelerated data.

The product category and route. Within one country, a food supplement and a traditional medicine registration are different requirements. The same product can face either depending on how it is classified, which is itself a decision worth making deliberately.

The claimed shelf life. A 36-month claim is harder to support by extrapolation than an 18-month one, and shortening the claim is a legitimate way to make a dossier work sooner.

And the climatic zone the destination sits in. Storage conditions for real-time studies are specified by zone, and a study run to the wrong zone’s conditions may not be accepted. This is a specific, common and entirely avoidable error.

The practical strategy for most brands

Three things in parallel, not in sequence, and the parallelism is the whole point.

Commission accelerated stability now, to support a provisional claim and to catch problems early.

Start real-time on the same batch at the same time. The marginal cost of running real-time alongside is modest compared with discovering in eighteen months that you have to start.

And confirm the destination’s requirement before either, because it may make one of them unnecessary or insufficient.

Then, when real-time data arrives, reconcile it. If real-time contradicts the extrapolation at 12 or 18 months, shorten the claim on subsequent artwork and tell your buyer. That is an uncomfortable conversation and a much smaller problem than the alternative.

The things that invalidate a study you already paid for

Five, and each one means starting again.

A formulation change, any ingredient, any quantity, a different extract supplier, a different standardisation grade.

An excipient change, including something as apparently minor as a different flow agent.

A pack change, bottle material, closure, seal, film, desiccant presence or absence. The pack is part of the result.

A capsule shell change, gelatin to HPMC or between suppliers.

And a manufacturing process change that affects the product, such as a different drying or granulation step.

Which is why the sequence matters: finalise the formulation and the pack, then run stability. Running stability on a product you are still developing is a study you will pay for twice.

The timeline conversation to have with your buyer

Because this is where the commercial damage actually happens.

Ask, at enquiry stage: which stability data does your authority require for this product category, and for what duration?

Then say what you have, honestly. “We have six months accelerated and twelve months real-time in progress” is a usable answer. “We can provide stability data” is not an answer, and it will be found out at exactly the wrong moment.

And build the project timeline backwards from this item rather than from your manufacturing capacity. Four weeks of production does not shorten a requirement for 24-month real-time data.

And our position

We tell clients when their existing product has no real-time data and the destination requires it, which is an unwelcome conversation and a cheaper one than the alternative.

We will not confirm an export delivery date before the data requirement is established.

And we do not describe accelerated data as equivalent to real-time, because it is not, and a dossier built on that equivalence fails at submission rather than at manufacture, by which point the client has paid for everything.

The short version

1. Stability data is usually the longest, least compressible item in a dossier
2. Real-time accrues at one month per month and money does not change that
3. Real-time: final pack, actual conditions, pull points to the claimed life
4. Accelerated: elevated temperature and humidity, short duration, extrapolated
5. Both measure assay, appearance, moisture, disintegration, microbiology
6. Accelerated gives speed, and a provisional claim where accepted
7. It is excellent for comparing two packs or two formulations quickly
8. And for early warning, a bad result at four months is worth knowing
9. But extrapolation assumes the same mechanism at both temperatures
10. For botanicals it frequently is not
11. Heat can trigger reactions that ambient conditions never produce
12. And slow processes can be understated by a short hot test
13. Physical changes do not extrapolate at all
14. Caking, pigment shift, oil separation, tablet hardening follow own kinetics
15. Microbiological outcomes do not extrapolate reliably either
16. And a prediction may be accepted provisionally, not finally
17. Four determining factors for what your dossier needs
18. The destination’s regime, requirements genuinely differ
19. The product category and route within that country
20. The claimed shelf life, 36 months is harder to extrapolate than 18
21. Shortening the claim is a legitimate way to submit sooner
22. And the climatic zone, which sets the storage conditions
23. A study run to the wrong zone may be rejected, avoidable and common
24. Strategy: accelerated and real-time in parallel on the same batch
25. The marginal cost of running real-time alongside is modest
26. Confirm the requirement before commissioning either
27. If real-time contradicts the extrapolation, shorten the claim and tell the buyer
28. Five changes invalidate a completed study
29. Formulation, excipients, pack, capsule shell, process
30. The pack is part of the result, not packaging around it
31. So finalise formulation and pack first, then run stability
32. Ask the buyer which data their authority requires, at enquiry stage
33. “We can provide stability data” is not an answer
34. And build the timeline backwards from this item, not from capacity

Related guides: how to choose a contract manufacturer, supplement costing and margins. Related reading: stability testing and shelf life explained, export documentation, what buyers actually ask for, ingredients that will get your consignment held, and packaging choices that cause returns.