What is the difference between a retest date and an expiry date, and how do you assign a retest period?

1Retest date vs expiry date.

ICH Q1A(R2) defines the re-test date as “The date after which samples of the drug substance should be examined to ensure that the material is still in compliance with the specification and thus suitable for use in the manufacture of a given drug product.” The expiration date is “The date placed on the container label of a drug product designating the time prior to which a batch of the product is expected to remain within the approved shelf life specification if stored under defined conditions, and after which it must not be used.”¹ A retest date asks a question about a material. An expiry date ends its use.

For drug products in the US, 21 CFR 211.137 requires an expiration date “determined by appropriate stability testing”, with exemptions for homeopathic products, some allergenic extracts and investigational products, and non-enforcement for some stable OTC products.³ It goes on the immediate container and on any outer package.⁴ EU law asks for the expiry date on the outer packaging “in clear terms (month/year)”.⁵ For a date such as 05/2029, Health Canada says “The product should meet approved specifications up to the last day of the specified month.”⁶ WHO puts a number on “immediately”: “e.g. within 30 days”. And a pass does not restart the clock: “If retested and found compliant, the batch does not receive an additional period corresponding to the time established for the retest period.”⁷

2Which date each material gets.

For how incoming lots are sampled and released, see incoming material release.

3How to assign a retest period.

1. Confirm a retest period fits. For APIs known to be labile, WHO says a shelf life is more appropriate; ICH Q1A(R2) says so for most labile biological substances and certain antibiotics.⁷¹
2. Put enough batches on study. At least three primary batches, at least pilot scale, made by the commercial route, with 12 months of long-term data at submission.¹ For an existing active substance, the EU accepts two production-scale or three pilot-scale batches with 6 months of data.¹⁰ For a US drug master file supporting ANDAs, FDA recommends 6 months of accelerated and 6 months of long-term data on three primary batches of at least pilot scale, made under CGMP, for full scientific review, with later long-term data through the proposed retest period filed as amendments. Three production batches can instead be filed with 6 months of accelerated data and long-term data covering the retest period.¹⁶
3. Store them as in practice. Use the container used for storage and distribution, or one that simulates it, at the ICH storage conditions.¹
4. Test on schedule. For a proposed retest period of at least 12 months: every 3 months in year one, every 6 months in year two, then annually.¹
5. Evaluate every attribute. For an attribute that changes, the approach ICH Q1A(R2) gives is the time at which the 95% one-sided confidence limit for the mean crosses the acceptance criterion. Batches can be pooled after statistical tests on their slopes and intercepts, at a significance level of 0.25. Where the data show little change and little variability, a justification can replace the statistics.¹¹⁷
6. Extrapolate within ICH Q1E. For room-temperature storage with no significant change at the accelerated condition: up to twice the long-term period, and no more than 12 months beyond it, where the data show little change and variability, or where a statistical analysis is done and the proposal is “backed by the result of the analysis and relevant supporting data”. Where the data change or vary and no analysis is done: up to one and a half times, and no more than 6 months beyond, and only where “relevant supporting data are provided”. Refrigerated storage, or significant change at the accelerated condition, allows less.¹⁷ Q1E is written for new molecular entities; for existing active substances the EU guideline sets the same ceiling of twice the data and 12 months beyond.¹⁰
7. Commit to finish. If the data do not cover the period granted, continue the studies through it on production batches.¹ Then add at least one batch a year to the ongoing stability programme.²
8. Label it. Put the retest date on the label and/or the certificate of analysis, with the storage statement the data support.²¹

4Example: assigning and using a retest date.

This example is invented. Three pilot-scale batches of API-311 have 24 months of data at 25 °C/60% RH and 6 months at 40 °C/75% RH with no significant change. Assay, impurities and water show little change and little variability. ICH Q1E allows up to twice 24 months (48) but no more than 12 months beyond 24 (36), so the applicant proposes 36 months. Because no production batch is on study yet, it commits to place the first three production batches on long-term study through 36 months.¹⁷¹

If the API maker wants a later date for B-26-0507 itself, ICH Q7’s Q&A allows it on long-term stability results for the API plus testing of that batch stored as labelled; some regions require approval. A longer period for future batches needs stability data and a regulatory filing.¹¹

The drug product made from it gets an expiry date, and the start of shelf life depends on where it is sold. For a tablet batch with a 24-month shelf life, granulated (the first step combining API and excipients) on 3 March 2026:

In the US, neither 211.137 nor FDA’s expiration-date Q&A says when the period starts. FDA checks each application’s proposed expiry date against its stability studies, so follow what your approved application states.³¹⁹ Biologics follow their own rule: the EU note excludes them, and the draft ICH Q1 starts shelf life at manufacture, which for a liquid filled over several days is the first day of filling. A batch filled from 8 to 10 June 2026 with a 36-month shelf life expires on 8 June 2029.¹⁸⁹

5What inspectors check.

In fiscal year 2025, FDA’s 713 system-generated drug Form 483s cited 21 CFR 211.137(a) “Expiration date lacking” 12 times, 211.166(b) “Adequate number of batches on stability” 10 times, 211.166(a) “Results not used for expiration dates, storage cond.” 7 times and 211.160(b)(1) “Retesting” once. The counts leave out 483s prepared by hand.²⁰ When FDA finds weak component controls, it asks for a review of whether materials “are assigned appropriate expiration or retest dates”, and whether incoming controls prevent use of unsuitable components.²¹

“Additionally, your firm approved the use of expired raw materials during the production of your ChigaRid External Analgesic drug product.” Colgin Inc., FDA warning letter, 25 January 2024²¹

FDA asked Colgin for a risk assessment of “distributing drug products manufactured with expired components”, including results from testing retain samples of each affected batch.²¹

“You lacked stability data to support your firm’s (b)(4) expiry date for your (b)(4) API. This is also a repeat deviation from a previous inspection.” Seema International, FDA warning letter 320-26-19, 13 November 2025²²

At Lexia, which lacked stability data for its labelled expiry period, the firm told investigators its product’s expiry was “anecdotal”.²³

6Edge cases the guidance answers.

• Retest periods over 3 years. Records and reserve samples for retest-dated APIs are kept 3 years after distribution, but for a 5-year retest date the Q&A says “it was never intended that the reserve sample be destroyed before the 5 year retest date was reached.”²¹¹
• Expired APIs. The Q&A says expired API should not be used in drug product formulation; for when it may be reworked, see reprocessing vs reworking.¹¹
• Early clinical APIs. Q7’s expiry and retest dating section applies to existing APIs used in clinical trials, but not normally to new APIs in early trials.²
• Frozen biologics. The draft ICH Q1 limits extrapolation for a well-characterised frozen biological drug substance to one and a half times the long-term data, at most 12 months beyond.⁹

7What is changing.

The draft ICH Q1 combines Q1A to Q1F and Q5C into one stability guideline. It reached Step 2b on 11 April 2025, and ICH lists it at Step 3 on 8 November 2026.²⁵ Beyond the biological retest periods and the production-date default shown above, its glossary defines shelf life for “a drug substance or drug product”, where ICH Q1A(R2) defines it for a drug product.⁹¹ Until each regulator adopts the final text, the current guidelines apply. See what the new ICH Q1 changes.

8How it runs in Seal.

In Seal, a material’s definition carries its shelf life or retest rule, and inventory applies expiry, retest, open-life and use-by rules to each container’s eligibility automatically. A container past any of them is blocked from normal selection, and a new version of the rule applies from a defined date without rewriting earlier receipts. Raw material management carries the authorised status, expiry or retest condition and remaining quantity into storage and dispensing.

In stability, a shelf-life estimate is a reviewed record linked to the results it used, with its population, data cutoff and model, so a descriptive chart is not mistaken for a shelf-life model.