Why retention samples matter

Every result on a certificate was produced at one moment from material that has since been used, shipped or consumed. If a question about that batch comes up a year later — a complaint, an unexpected observation, a regulator’s request — the original test cannot be repeated on the original material unless some of it was kept. Retention samples exist so that it can be.

Two kinds of kept sample

EU GMP Annex 19, Reference and Retention Samples, separates two purposes that are often combined under one name.

A reference sample is a quantity of a batch kept for analysis. It is there so that the material can be tested again if needed. It must be large enough for the full set of specification tests to be repeated, and the annex asks for enough to carry them out on at least two occasions.

A retention sample is a fully packaged unit kept for identification. It shows what the batch looked like as supplied: its container, closure, label, lot number and printed details. It answers questions about presentation and packaging, such as whether a label was misprinted, rather than questions about content.

In the United States, 21 CFR Part 211 uses the term reserve sample and sets a similar minimum: at least twice the quantity needed for all the required tests. Many organizations keep both a reference and a retention sample from each batch, and in some cases one set of units serves both purposes.

What gets kept

Retention practice usually covers more than finished batches:

• Starting materials, so that a raw material lot can be tested again if a batch made from it is questioned.
• Packaging components that carry printed information, so that a label can be checked against what was approved.
• Finished product from every batch released.
• Intermediates, where they are stored or supplied before further processing.

How samples are stored

Kept samples are stored under the conditions on the label, in the same container and closure as the material supplied, in a secure area with controlled access. Each is logged with the material, the lot, the date it was set aside, the quantity, the storage location and the date it may be destroyed. Withdrawals are recorded, because a sample that has been partly used without a record no longer contains a known quantity of anything.

Storage under the labeled conditions is what gives a later test its meaning. A reference sample tested two years after release shows how the batch behaved under proper storage. If a complaint concerns material that degraded faster somewhere else, the comparison between the kept sample and the complaint sample is what separates a manufacturing problem from a storage problem.

How long they are kept

Retention periods are tied to the material’s dating. For finished products, both EU and US rules generally require samples to be kept for at least one year after the batch’s expiry date. ICH Q7, the GMP guide for active pharmaceutical ingredients, sets periods linked to the expiry or retest date and to the date the batch was distributed. Starting materials are generally kept for a set period after release of the product they were used in, unless their own stability is shorter. A laboratory’s written procedure turns these rules into a specific destruction date for each sample.

What retention samples are used for

Most kept samples are never opened. They are destroyed at the end of their retention period, and that is a sign the system worked. When one is used, it is usually for one of a short list of reasons: investigating a complaint, supporting an out-of-specification or out-of-trend investigation, confirming the identity of material found elsewhere, answering a regulator’s question, or checking a batch against a later change in method.

Because the samples are used so rarely, the discipline around them tends to slip. Logs go unreconciled, samples are stored with general stock, or quantities are taken for routine work and never replaced. An audit of a retention program usually checks a handful of batches at random: whether each sample can be found, whether the quantity matches the log, and whether storage conditions were recorded over the whole period.

In each case the kept sample provides the one thing no document can: the material itself, stored under known conditions, from a known batch, available to be measured again. A certificate states what a laboratory found. A retention sample makes it possible to check.