Temperature Excursions: When a Spike Still Means Release
Share
The container is opened in Rotterdam and a logger is pulled out of the load. The trace shows a flat 22 °C, then nine hours at 47 °C on day nineteen, then a return to 23 °C. The customer's quality team sends the PDF with one line: excursion, please advise.
Two bad answers are available. The first is that the batch is fine because the product melts at 200 °C. The second is that the batch is condemned because the label says 25 °C. Both skip the same step, which is to ask what the excursion actually did to the material inside the pack.
An excursion is a question about degree
A temperature record is not a verdict. It is a description of the environment outside the container, and the material experiences something else: the air inside the pack, modified by the thermal mass of the load, the barrier properties of the liner, and the time constant of the whole assembly.
Four variables decide whether an excursion matters.
Peak. How far above the labelled condition, and whether the peak crosses a physical threshold. For a crystalline solid with a melting range above 200 °C and a loss on drying below one per cent, 47 °C crosses no threshold. For a laminate seal or a polyethylene liner, it can.
Duration. Nine hours is a surface event. Nine days is a bulk event. A pallet of 25 kg drums takes many hours to equilibrate to a new ambient, so a nine-hour spike may barely reach the centre of the load.
Pack. A sealed aluminium laminate with desiccant damps both the moisture and the thermal transient. A fibre drum with a loose liner does neither.
Position in the load. Loggers near the door record the container wall. Loggers inside the stack record the product. The two can differ by ten degrees, and it matters which one produced the PDF.
The mean kinetic temperature arithmetic
Mean kinetic temperature exists because a simple average lies. Warm periods do more damage than cool periods do good, so the calculation weights them accordingly.
Take an illustrative shipment: 80 per cent of the transit at 22 °C, 15 per cent at 30 °C, 5 per cent at 47 °C. The arithmetic mean is about 25.3 °C, which looks compliant. The mean kinetic temperature, calculated with the Arrhenius weighting, lands several degrees higher, in the low thirties. That number is not a limit and it is not a release criterion; it is a way of expressing the thermal load in one figure that can be compared with a stability dataset.
The comparison is only meaningful if the stability dataset covers that thermal load. A 30 °C/65% RH long-term arm covers a higher mean kinetic temperature than a 25 °C/60% arm, which is why the Zone IVb article in this series matters here as well: the excursion assessment borrows its conclusion from the stability file.
Where the real damage happens
Not in the molecule, in most cases. In three places around it.
Condensation. A container that has been at 8 °C in a ship's hold and is then opened in a 30 °C dock develops condensation on every cold surface, including the outer surface of a steel drum and, if the closure is imperfect, the inner one. Water that forms on the outside of a drum finds the seam. This is the single most common cause of a caking complaint that has nothing to do with the material as supplied.
Barrier damage. Heat softens adhesives and stresses seals. A laminate that passed a dye penetration test at 23 °C has not necessarily been tested at 47 °C. The failure is invisible until the customer opens the drum and finds that the inner bag has separated at a fold.
Label and documentation. Thermal paper fades. A batch label that becomes unreadable in transit turns a routine receipt into a traceability investigation, and the material is fine while the paperwork is not.
None of these appear on a certificate of analysis. All of them appear in the complaint file.
A decision framework that survives an audit
| Situation | Assessment | Typical action |
|---|---|---|
| Peak below 35 °C, any duration, sealed pack | Covered by the long-term stability arm | Release, record the trace |
| Peak 35–50 °C, under 24 hours, sealed pack intact | Thermal load within the accelerated arm | Release with a written assessment |
| Peak above 50 °C, or any duration above a week | Outside the studied envelope | Retest key attributes before release |
| Any peak, pack compromised or opened | Moisture route is live | Retest loss on drying and assay, inspect for caking |
The framework is not a substitute for the batch data. It is a way of deciding which batch data to look at, and of writing down why the decision was taken, which is the part an auditor reads first.
What the supplier owes the customer
Three things, in this order. The raw trace, not a summary. The pack description for that specific shipment, because the assessment depends on it. And an assessment letter that states the peak, the duration, the mean kinetic temperature, the stability data relied on, and the conclusion with its limits.
What the supplier cannot do is release the material into the customer's facility. That decision belongs to the receiving quality unit. A supplier who writes "batch is acceptable for use" has overstepped, and a customer who accepts that sentence has outsourced a decision they cannot outsource.
The sentence that saves the argument
Put the excursion protocol into the quality agreement before the first shipment. Define what counts as an excursion, who is notified within how many hours, what data travels with the notification, and which party performs the assessment. Four lines, agreed in a quiet week, remove most of the argument that otherwise happens in a loud one.
Our own position is that a trace is supplied with every temperature-controlled shipment whether or not anybody asks, because the cost of the logger is trivial next to the cost of reconstructing a three-week journey from a forwarder's email.
Where the logger sits decides what the record means
A data logger taped inside the container door records the container wall. That surface heats and cools faster than the load, it takes the sun directly when the doors are opened on a dock, and it is not where the product is. A logger buried in the middle of a pallet, four drums deep, records something much closer to what the material experienced, and it will usually show a flatter, lower trace.
Both records are true. They answer different questions. The door logger answers "what did the container see", which is useful for a carrier dispute. The in-load logger answers "what did the goods see", which is what a quality assessment needs. Sending the first to a customer's quality team and calling it the shipment record is a common and avoidable argument.
Placement is also a sampling decision. One logger per shipment describes one position. For a full container of high-value material, one per pallet row is defensible, and the cost is trivial against the cost of reconstructing a three-week journey from a forwarder's email chain. The general chapter on good storage and distribution in the USP is a useful reference for what a distribution record is expected to contain, and the ICH stability framework supplies the conditions against which the record is judged.
Writing the assessment letter
An assessment letter is a short document with a fixed shape, and the shape is what makes it usable. Six parts, in this order.
The record. Logger position, calibration date, start and end of the recording, and the raw file attached rather than summarised.
The excursion. Peak, duration above the labelled range, number of excursions if more than one, and the mean kinetic temperature for the whole journey.
The pack. What the material was in, whether the barrier was intact on arrival, and whether the container had been opened at any point in the chain.
The data relied on. Which stability arm covers the observed thermal load, at which condition, and for how long. A conclusion without this section is an opinion.
The conclusion. Release, release with retest, or refer to the receiving quality unit, with the reasoning stated in one paragraph.
The limits. What the assessment does not cover, which is usually the period before the logger started and any leg of the journey without a record.
The EudraLex volume 4 distribution requirements and the assessment practice described in the Zone IVb article point the same way: the assessment is a document about evidence, not a reassurance. Where the pack turns out to be the weak point rather than the molecule, the packaging article covers what a barrier is actually doing, and where the thermal load has to be judged against a study, the ongoing stability programme is the place the numbers come from.
Frequently asked questions
What counts as a temperature excursion?
Any departure from the labelled storage condition, outside the tolerance the label or the quality agreement allows. A label reading 15–25 °C with a two-degree tolerance is breached at 28 °C. The important part is that the definition is documentary: if nobody wrote down what counts, the argument after arrival will be about definitions rather than about the material.
What is mean kinetic temperature?
A single calculated temperature that represents the effect of a varying temperature history, weighted by the Arrhenius equation so that warmer periods count for more than cooler ones. It is not an average. A shipment at 25 °C for most of the journey with a short period at 45 °C has an arithmetic mean of about 26 °C and a mean kinetic temperature several degrees higher.
Is a short spike worse than a long mild deviation?
For most solid APIs, the long mild deviation is worse, because it drives moisture uptake and slow oxidation. A spike matters when it crosses a threshold that changes state: a phase transition, a melting point, the glass transition of a laminate, or the dew point inside a container. Nine hours at 47 °C is usually a pack question; nine days at 38 °C is a product question.
Who decides whether the batch is still usable?
The qualified person or quality unit responsible for release of the receiving site, using the stability data and the excursion record. The supplier can supply the data and an assessment, but cannot release material into a customer's facility. An arrangement in which the supplier decides is a documentation gap that surfaces at the next audit.
Should every pallet carry a logger?
At minimum one per shipment, placed inside the load rather than taped to the door, because the door position measures the container wall and not the product. For high-value or regulatory shipments, one per pallet is defensible. A logger that only records the air next to the door will show excursions that never reached the material and miss the ones that did.
Does a sealed laminate make the excursion irrelevant?
It removes the moisture route and leaves the thermal one. If the material inside is dry and the laminate is intact, a short thermal excursion usually produces no measurable change. If the laminate has a pinhole, or the drum was opened at a customs inspection, the moisture route is back and the assessment has to start again.