How do you run a hold-time study for bulk and in-process material, and how long can material be held?

1Hold time study guidelines: FDA, EU, WHO and ICH.

Any step where material waits, and the wait could change its quality, needs a time limit supported by data. In the US the binding rule is 21 CFR 211.111: “When appropriate, time limits for the completion of each phase of production shall be established to assure the quality of the drug product.”¹ FDA answers the hold question directly in question 14 of its laboratory controls Q&A: a risk- and science-based assessment can support holds of stable intermediates, sometimes with testing at the hold stage; unstable materials, and holds longer than the assessed period, need stability studies under an approved protocol.² For APIs, time limits “may be inappropriate when processing to a target value (e.g., pH adjustment, hydrogenation, drying to predetermined specification)”.³

WHO’s guideline is the most detailed, but it says: “These guidelines do not cover aspects for hold times in cleaning validation, or the manufacturing of active pharmaceutical ingredients (APIs) or biologicals.”⁴ The PIC/S GMP Guide, revised as PE 009-18 in September 2026, carries the same Annex 1 8.18 and Annex 15 10.8 requirements as EU GMP.¹⁰ In the UK, regulation C17 of the Human Medicines Regulations 2012 keeps the EU GMP guidelines as they applied on 31 December 2020, Chapter 4 and Annex 15 among them, unless the UK licensing authority replaces them.¹¹¹²

2Hold-time study protocol and report template.

WHO lists what the protocol and report contain; the two rows citing Cosmax answer what FDA asked one firm to add.⁴¹³

One common split: validation or technical services writes the protocol, production holds the material and pulls the samples, QC tests them, and the quality unit approves the protocol and report. In the US, the production procedures that carry hold limits, and any change to them, are approved by the units concerned and then by the quality control unit, so a new limit reaches the master batch record through change control.¹⁴ With WHO’s time points for cores, the last pull is on day 90 and the report waits for its results, so plan on about four months from protocol approval to an approved limit.⁴

3How long can material be held?

As long as your data support, under the conditions you tested, and no longer. None of these texts puts a number on the maximum hold; each asks for a limit supported by data.¹⁴⁶ The numbers that do appear are dossier thresholds. EMA treats “more than 30 days for solid oral dosage forms and more than 24 hours for sterile products” as prolonged, needing stability data on at least two pilot-scale batches “Where relevant”.⁵ The ICH Q1 draft, still at Step 3 and not final in November 2026,¹⁵ gives as examples of a prolonged intermediate hold “more than 30 days for solid dosage forms for the entire manufacturing process or more than 24 hours for non-solid dosage forms or sterile products”, with evidence “included in the regulatory submission, when requested”.⁸

You do not have to test until failure: “Studies may extend beyond the chosen maximum but it is not necessary to extend testing to determine the extreme limits at which failure occurs.”⁴ Once set, the limit binds: “Intermediates should not be used beyond the established holding times.”⁸

4How to design the study.

• Batches: “Typically one or more batches”, the number chosen by risk.⁴ For biologics, physicochemical data may come from small scale, but microbial control is studied at production scale.⁸
• Container: the production container or, if that is exceptionally large, an equivalent of the same material and closure; justify a smaller one.⁴
• Headspace: where oxidation matters, a headspace-to-contents ratio at least as great as the worst possible in production, part-filled containers included.⁴
• Conditions: those of the quarantine area or manufacturing stage, and transport between sites where it happens.⁴⁸
• Methods: validated and stability-indicating.⁴
• Criteria: “typically more stringent than registered specifications”, compared with the time-zero result.⁴
• Timing: hold times are normally set before marketing; pilot-scale results are confirmed in process validation at commercial scale.⁴

For a liquid or other non-solid product, the ICH draft’s example of a prolonged hold is one over 24 hours.⁸

5Example: bulk tablet cores.

Invented for illustration: Exemplol 10 mg film-coated tablets. Uncoated cores from PPQ Batch 26-021 are held in their production drums, double polyethylene liners inside closed HDPE drums at the routine fill level, in the quarantine store. The study runs to 90 days to show the margin.

The other WHO core tests ran too. Description, hardness, thickness, friability and disintegration met their limits at every point, and the microbial test, run at day 0 and day 90 because the cores are dry, met its limit at both. Uniformity of dosage units was tested at day 0 only, because holding cores cannot move drug between tablets; the report gives that reason, as FDA asked Cosmax to do for any attribute not tested at the maximum hold.⁴¹³

Day 90 still meets the specification but not the tighter study limits, so the report sets the core hold at 60 days and checks the margin with the trend analysis WHO asks for where appropriate.⁴ A least-squares line through the water results rises 0.01 points a day from 2.03% and reaches 2.8% at day 77; the dissolution line falls 0.1 points a day from 93.3% and reaches 85% at day 83. Those mean lines overstate the margin. The approach ICH Q1E describes for shelf life, the one-sided 95% confidence limit for the mean, crosses at about day 67 for water and day 65 for dissolution on these five points (3 degrees of freedom, t of 2.353).¹⁶ Water also rises faster after day 60, from 2.6% to 3.0%, so a straight line is optimistic. Sixty days stands, with about five days to spare rather than 17. Dissolution falls as water rises, so the conclusion keeps the storage conditions and drum type with the limit.

5.1Cumulative hold.

The same report sets 30 days for the blend and 60 days for the coated bulk. In this direct-compression process the blend is the first step that mixes the active ingredient with other ingredients. A batch using every limit is blended on day 0, compressed by day 30, coated by day 90 and packed by day 150. ICH’s draft says cumulative holds are generally assessed in process validation, and a cumulative study may be needed if a stability risk is found.⁸

Both dossier thresholds are crossed. The 60-day coated-bulk hold is prolonged under EMA’s 30 days for solid orals, and the 150 days are prolonged under the ICH draft’s 30 days “for the entire manufacturing process”. So an EU dossier needs stability data on at least two pilot-scale batches where relevant, and the ICH draft expects evidence in the submission when requested, with data not generated at production scale verified on production batches after approval.⁵⁸

The 150 days also cost shelf life. WHO counts shelf life from when the active ingredient is mixed with other ingredients, “irrespective of hold times”.⁴ In the EU, shelf life starts at release if release is within 30 days of that date, and otherwise at that date.¹⁷ So a 24-month shelf life leaves about 19 months at packing on day 150. To count shelf life from any other date, the ICH draft asks for the holds to be declared, justified and covered by stability batches “that represent the cumulative maximum holding times of drug product intermediates”.⁸ Where a hold needs a stability study, FDA asks for expiry to be calculated from the intermediate’s manufacture or release date until stability data on product made from held material exist.²

6Putting hold limits into the batch record.

The master instructions carry each limit. ICH Q7 lists “time limits for completion of individual processing steps and/or the total process, where appropriate”.³ EU GMP asks the processing instructions for “the requirements for bulk storage of the products; including the container, labeling and special storage conditions where applicable”.¹⁸

The batch record then proves the hold was kept. In the US it documents the dates of each significant step of manufacture, processing, packing or holding.¹⁹ EU GMP, and the PIC/S guide in the same words, asks for “Dates and times of commencement, of significant intermediate stages and of completion of production”.¹⁸²⁰ Times, not only dates, are what make an hours-long hold auditable: record when each hold starts and ends, and compare the elapsed time, and the running total, with the limit before the next step.

An exceeded limit is a deviation. Under 211.111 it may be acceptable if quality is not compromised: “Such deviation shall be justified and documented.”¹ ICH Q7 says the same for APIs: “Deviations should be documented and evaluated.”³ For sterile products, Annex 1 also limits the time from the start of preparation to the end of filling, with “a maximum permissible time for each product”.⁶

7What inspectors cite.

“Your firm lacked bulk hold time studies for multiple drug products that may be held at the in-process stage for up to six months prior to final fill. However, you assigned (b)(4) as the maximum bulk hold time to your drug products without an approved study, procedure, or protocol. Furthermore, you exceeded your maximum bulk hold time without an appropriate investigation.”

Cosmax USA, Inc., warning letter, 15 October 2020.¹³

A 2026 letter to a maker of an over-the-counter psoriasis cream makes the same point after a commercial scale-up:

“Furthermore, you did not conduct adequate in-process hold time studies and stability studies to support shelf life.”

BioMylz Pvt. Ltd., warning letter 320-26-103, 13 July 2026.²¹

FDA’s 2026 programme for CDER-regulated biologics checks that hold times for process intermediates are not longer than those in the application, and directs investigators to “Evaluate whether, in production, the predefined hold times for buffers and media are supported by microbiological and chemical stability data.”²² For a licensed biologic, an exceeded hold on product that is later distributed can also need a BPDR, under FDA’s code PC-24-07, “Stored for an excessive hold time”.²³

8How it runs in Seal.

In Seal’s electronic batch records, operators work through steps resolved from the effective master version, with limits and holds applied as configured, and each controlled value has a configured response, such as requiring verification, placing the batch on hold or opening a deviation. The deviation example follows a 52-minute post-mix hold against a 45-minute limit; its corrective change makes the master batch record calculate elapsed time from the start and transfer timestamps, and a missing timestamp reports neither a duration nor a pass.

For equipment, cleaning validation in Seal starts the dirty-hold clock at the last product contact, keeps clean, dirty and expired states distinct, and blocks normal use once a hold is exceeded. Stability studies keep each pull’s nominal timepoint, allowed window and actual removal as separate events.