Storage condition labels explained: 2–8 °C, 15–25 °C, 'do not freeze'

ICH Q1A(R2) defines the storage conditions under which stability is studied: long-term at 25 °C ± 2 °C / 60 % RH ± 5 % RH or 30 °C ± 2 °C / 65 % RH ± 5 % RH, intermediate at 30 °C / 65 % RH, accelerated at 40 °C / 75 % RH, refrigerated at 5 °C ± 3 °C, and frozen at −20 °C ± 5 °C. Label statements are derived from the study results.

Reading the label

• 'Store at 2–8 °C': refrigerated; both limits matter, and freezing is an excursion.
• 'Store below 25 °C': an upper limit only; the lower limit comes from any 'do not freeze' or product-specific statement.
• 'Store at 15–25 °C': controlled room temperature with both limits.
• 'Do not freeze': a separate lower-limit requirement often applied to refrigerated products.

Mapping acceptance criteria and monitoring alarm limits should be derived from these statements for the products actually stored, not from a generic site default.

From study condition to label statement

The chain runs one way: the stability data establish what the product tolerates, the label states it in words, and your mapping and monitoring limits follow from the label. It does not run backwards. A warehouse cannot decide that 27 °C is acceptable for a product labelled store below 25 °C because the building runs warm in August; that is a decision only the stability data can support, and only the marketing authorisation holder can make.

Note also that a label condition is not the same as a study condition. A product studied at 25 °C ± 2 °C is not labelled store at 23–27 °C. The tolerance on the study is control of the chamber; the label is the claim about the product.

Excursions, and where mean kinetic temperature fits

For products labelled with an upper limit only — store below 25 °C, store below 30 °C — the relevant question over a long period is usually cumulative thermal exposure rather than any single reading, and mean kinetic temperature is the accepted way to express it. MKT weights higher temperatures more heavily than an arithmetic mean, because degradation kinetics do, so a week at 30 °C is not cancelled by a week at 20 °C.

MKT does not apply to a 2–8 °C product in the same way. A refrigerated label has a lower limit as well, and a freezing event is a discrete failure that no averaging can absorb — which is why do not freeze appears as a separate statement. Assessing a cold-chain excursion means comparing the actual time-temperature profile against the stability data, not computing an average.

Setting your own limits from a label

• Take the tightest label among the products actually stored in that space, and derive the mapping acceptance criteria from it. A site default of 15–25 °C applied to a room holding a 2–8 °C product is a common and serious finding.
• Pull the alarm limits inside the label limits by at least the monitoring sensor's uncertainty, so a ± 0.5 °C sensor is not asked to police 8.0 °C at 8.0 °C.
• Write down what happens when a limit is crossed, before it is crossed: who assesses, against what data, and who may release the stock.
• Re-derive the limits when the product mix changes. A new refrigerated line in an ambient warehouse changes every number above.

Regional wording differs — the USP uses controlled room temperature with its own MKT-based definition, and the WHO conditions differ for Zone IVb climates — but the logic is the same everywhere: the label sets the limit, and the limit sets the study.