Retest dates and expiry dates

Three kinds of date turn up on laboratory reagent labels and the paperwork that travels with them: a retest date, an expiry date, and something softer, usually worded as "recommended use by" or "best used within". They are not synonyms. They rest on different evidence, they commit the supplier to different things, and they ask different actions of the lab holding the container.

Most receiving systems flatten them. A single field called Expiry takes whatever date was printed, and the distinction is lost at the moment of booking-in — the one moment when someone was actually looking at the label. This note covers what each date means, what data stands behind it, what a temperature excursion in transit does to it, and what to write down. It does not cover use of the material in any biological system, and it is not a stability protocol.

Three dates, three different claims

The vocabulary is borrowed from pharmaceutical stability practice, which is where it was defined. It is worth using the definitions as written rather than the ones people assume.

• Retest date. ICH Q1A(R2) defines it as the date after which samples of the substance should be examined to ensure the material is still in compliance with the specification. The instruction is to test, not to discard.
• Expiry (expiration) date. ICH Q1A(R2) defines this, for a drug product, as the date before which a batch is expected to remain within its approved shelf-life specification. It is terminal in the regulated sense: past it, the material is not to be used as that product.
• "Recommended use by", "best used within", and similar. No standards body defines these. They are the supplier's opinion, offered without committing to either of the above. Their meaning is whatever the supplier's own documentation says it is, which is often nothing.

The difference between the first two is the difference between a checkpoint and a cliff. A retest date says the evidence supporting the material's specification runs out on that date and can be renewed by measurement. An expiry date says the evidence runs out and cannot. Treating a retest date as an expiry discards usable material. Treating an expiry as a retest date keeps material in service that the supplier has stopped standing behind. Both errors start in the same place: a form with one date field.

What the date was derived from

A date is a conclusion. The question worth asking is what it was concluded from.

ICH Q7, the guideline on good manufacturing practice for active pharmaceutical ingredients, states in section 11.6 that an API expiry or retest date should be based on an evaluation of data derived from stability studies, and observes that common practice is to use a retest date rather than an expiration date. The same section says that where material moves outside the manufacturer's own material-management system with a date attached, supporting stability information should be available — published data or test results. That is the standard to hold a date to: something exists behind it, and someone can say what.

Research-use reagents are not APIs and are not sold under that framework. The questions it implies still transfer intact.

1. Is the date based on real-time data on this material, on accelerated data, on extrapolation, or on a general assumption applied across a family of similar compounds?
2. Was that data generated at the storage condition printed on this label, or at a different one?
3. Is the date lot-specific, or a standing figure applied to every lot of the item?
4. Does the certificate of analysis for this lot carry the same date as the container, and if the two disagree, which one governs?

A useful comparison sits in ISO 17034:2016, which sets requirements for the competence of reference material producers. Producers working to it are expected to establish and monitor the period of validity of a reference material on the basis of a stability assessment, and to state it. Most research reagents are not certified reference materials and are not offered as such. The standard is still a fair yardstick for what "the supplier assigned a date" ought to mean.

What transit does to a date

Every date on a label is conditional on the storage condition printed beside it. The condition is not a suggestion attached to the date; it is part of the date. Material held outside that condition has been outside the scope of the data the date came from.

Transit is where this usually happens. USP General Chapter <1079>, Risks and Mitigation Strategies for the Storage and Transportation of Finished Drug Products, describes the failure modes: handoffs between organisations, gaps in monitoring, inconsistent documentation along a route. USP <1079.2> deals specifically with mean kinetic temperature in evaluating temperature excursions during storage and transportation. Mean kinetic temperature is a single weighted value summarising a thermal history. It is an input to a judgement, not the judgement. It can also flatten a short severe excursion into an unremarkable average, which is why the underlying trace matters and not only the summary figure.

The absence of monitoring is its own finding. A shipment that arrived with no logger, or with a logger nobody downloaded, has no thermal history. That is not the same as a clean one and should not be recorded as though it were.

What the receiving lab should record

Booking-in is the only point at which the label, the certificate, the shipping documents and the physical state of the package are all in front of the same person. Anything not captured there gets reconstructed later, badly.

For labs operating to ISO/IEC 17025:2017, clause 6.6 covers externally provided products and services and expects the laboratory to have a procedure for verifying that what it receives meets stated requirements before use. A supplier's date is one of the things being verified against. It is not a substitute for the verification.

1. The date exactly as printed, with its wording intact. "Retest 2027-03" is a different record from "Expires 2027-03", and paraphrase destroys the distinction.
2. Where each date appeared — container label, certificate of analysis, packing list — and whether they agree. Disagreement is a supplier query, not a rounding problem.
3. The lot or batch identifier, tied to the specific container rather than to the order. Lot traceability is treated separately in this library.
4. Date and time of receipt, and the storage location the material went into.
5. Condition on arrival: outer and inner packaging, seal integrity, state of any coolant. Recorded as observation, not as a pass mark.
6. Logger or indicator data: present or absent, downloaded or not, reviewed by whom, and the conclusion reached.
7. The accept, quarantine or reject decision, with the name of whoever made it.
8. The lab's own next-review date, which may fall earlier than the supplier's date and is the lab's to set.

One structural point about systems. If the inventory record has a single date field, the name of that field determines how everyone reads its contents, and a retest date entered into a field called Expiry becomes an expiry date the moment it is saved. Two fields, or one field plus a date type, costs very little at configuration and settles the question permanently.

Reading the standards without over-reading them

The documents named above define the vocabulary. Naming them here says where the terms come from. It is not a statement that any particular material was manufactured, tested, dated or shipped under them, and it should not be read as one. ICH Q1A(R2) and ICH Q7 address drug substances and products. USP <1079> and <1079.2> address finished drug products. ISO 17034 addresses reference material producers. Research-use reagents sit outside all four in scope while borrowing their language.

That gap is the practical reason to record the wording of a date rather than the date alone. A retest date on a research reagent may carry the ICH meaning exactly, or it may be a supplier reaching for a familiar phrase. The label will not tell you which. The supplier's documentation might, and the question is cheaper to ask before purchase than after receipt. Supplier qualification is covered elsewhere in this library; date convention belongs on that list of questions.

None of this makes an older container suspect or a monitored one safe. It makes the state of the evidence legible to the next person who opens the record — which, on the day someone queries a result, is the only thing the record is for.