The official sources don't always agree, and some questions they don't answer at all. Each case below shows both sides with their sources, and a risk-based way to decide that you can write down and defend.
When the sources differ, assess the risk to the product and the patient, decide, and record why. A documented, risk-based rationale is what an inspector looks for when the guidance leaves room for judgement.
01
Returns: 24 hours or 10 days?
A customer wants to send back a fridge line they bought from us. How long can it have been away before we can even consider putting it back into stock?
EU GDP gives "for example 10 days" as an acceptable return window from customers without a wholesale authorisation. In GB the MHRA position is tighter: its Inspectorate blog of May 2016, published after the current guidelines, says refrigerated returns from unlicensed sites are completed within 24 hours including transport, and ambient returns within five days.
EU GDP 6.3 (2013)
Returns from customers without a wholesale authorisation can be accepted within an acceptable time limit, for example 10 days. For cold chain items there must be documented evidence that the product stayed in its authorised storage conditions the whole time, and any deviation needs a risk assessment showing integrity.
For non-defective refrigerated products returned from an unlicensed site, the return should be completed within 24 hours including transport, and within five days for ambient products. This is the MHRA position for GB.
Cold chain products may only go back to saleable stock where there is no reasonable possibility the cold chain was compromised, for example through a validated delivery-and-return process or a monitoring device attached to the product.
Treat the time limit as a ceiling, not a test: the deciding question is whether you have evidence of storage throughout.
Check the evidence chain: delivery, receipt, storage at the customer, repacking, collection and return to your fridge.
Review any logger data or attached indicator against the product's labelled range.
Ask the manufacturer or MAH about any deviation before deciding.
Record the RP's decision and rationale, including which time limit your procedure uses and why.
Check with your inspector or RP:Agree it with the GDP Inspectorate in writing before your returns procedure allows anything longer than 24 hours for refrigerated stock or five days for ambient. The COVID-19 10-day allowance was temporary.
Where should the permanent temperature probe go in our fridge or cold room: in the middle, or where mapping found the warmest and coldest spots?
A 2016 MHRA blog says to place probes centrally, between products and not in the door. EU GDP and WHO say to place permanent monitors where mapping shows the extremes, which may not be the centre.
MHRA blog (2016)
Calibrated probes should be placed centrally, preferably between products, and not in the door.
Monitoring sensors go where the greatest temperature variability is expected in the qualified volume, positioned so brief events such as door opening affect them as little as possible.
Worst-case location vs nuisance alarms at the door
Our mapping found the warmest air by the cold room door. With the alarm probe there, it goes off every time someone picks. Are we breaking the rules if we put it somewhere else?
WHO says monitoring and alarm sensors go where temperature varies most, yet positioned to be minimally affected by brief events such as door opening. WHO Supplement 6 says monitors should not sit where door openings generate an abnormally high number of alarms, and the HPRA puts probes in a load simulator so door openings during picking do not trigger alarms. EU GDP and the MHRA mapping blog say only "the extremes of fluctuations" and "the widest temperature fluctuations", which some sites read as the door threshold. That reading produces alarms nobody listens to.
EU GDP 3.2.1 (2013)
"Temperature monitoring equipment should be located according to the results of the mapping exercise, ensuring that monitoring devices are positioned in the areas that experience the extremes of fluctuations." Nothing on doors or alarms.
Lists "heaters, lighting, windows and doors, loading bays" as problem areas to map, then says permanent probe positions "should cover the areas that have the widest temperature fluctuations or indicate areas with any hot or cold spots". It says nothing about door openings or nuisance alarms.
"Probes should not be placed in the door." Probes should be "sited in a central location within the refrigerator and, preferably between the products". An inspector reply (5 February 2016) adds that using a liquid to simulate the load "should be supported by a risk assessment, a mapping exercise".
"Temperature monitoring systems (rather than Min/Max devices) are typically set with a small delay time for temperature alarms based on risk assessment and confirmed via validation and risk packages." A reply to a reader comment, not formal guidance.
Monitoring and alarm sensors go where "greatest variability in temperature is expected to occur", "positioned so as to be minimally affected by transient events such as door opening". Control sensors, by contrast, go at the mapped hot and cold spots "even if affected by door opening" (4.5.1).
"Monitors should not be placed in areas where transient events such as a door opening may affect the monitoring and generate an abnormally high number of alarms." If that cannot be resolved technically or operationally, the area "should not be used to store TTSPPs and should not be monitored".
Recording probes go "in locations with the greatest temperature variability as determined by temperature mapping studies" (3.1). And: "Probes should be situated within an appropriate load simulator so that transient rises in temperature (e.g. when the doors are opened during picking) do not trigger alarms" (10).
A delay may be set "to allow for acceptable small excursions without an alert, e.g. the transient rise in temperature associated with a fridge door opening", but "as short a period as practically possible".
Separate two questions: where does stock get warmest, and where does the air only spike for a minute when the door opens?
Map with door events logged, as WHO Supplement 8 asks when a study includes door openings, so you can see how long each spike lasts and whether stock follows it.
Decide what each probe is for: an alarm on product risk, an early warning, or watching the door zone.
Buffer alarm probes and check their response against the mapping data before setting any delay.
If a door zone alarms constantly, fix the cause (door discipline, strip curtains, a faster door) or stop storing stock there.
Record the reason for each position and each delay.
Check with your inspector or RP:Ask your inspector if your mapped worst case is at a door you can neither buffer against nor keep stock away from, or if you want a delay longer than your door-recovery data support.
How often do we really need to read and record the fridge temperature, and does it depend on which part of the UK we are in?
The Green Book minimum in England is once each working day. Public Health Scotland recommends twice daily, Northern Ireland says at least daily and preferably twice, and WHO model guidance says twice a day, seven days a week.
Green Book, England (2013)
Fridge temperatures must be monitored and recorded at least once each working day, at the same time each day, with each entry signed.
Recommends recording current, maximum and minimum temperatures twice daily, at the start and end of the working session, while acknowledging the Green Book accepts once daily.
Identify which guidance applies to your setting and nation, and treat it as the floor.
Assess how fast your fridge would drift out of range if it failed, using holdover or logger data.
Weigh the value of the stock and how hard it is to replace.
Consider whether continuous monitoring with alerts closes the gap between readings.
Write the chosen frequency and rationale into your SOP.
Check with your inspector or RP:Ask your local screening and immunisation team, ICB or inspector if you operate across nations and want one frequency for every site.
UKHSA's 20-minute allowance and the max/min thermometer
Our max/min thermometer shows the fridge hit 10 °C. How can we tell whether it was under 20 minutes, which is the point where guidance says no further action?
UKHSA and SPS say a one-off rise above 8 °C lasting under 20 minutes is unlikely to matter. The Green Book minimum is a max/min thermometer, which records how far the temperature went but not for how long. The allowance is NHS clinical guidance and does not apply to WDA or MIA holders: under GDP and GMP every excursion is a deviation.
UKHSA incident guidance (2022)
One-off rises above +8 °C lasting less than 20 minutes, for example during restocking, are unlikely to have breached the cold chain and need no further action beyond recording the cause.
With only a min/max recorder, assume the excursion started just after the last reset unless you have information otherwise. Min/max recorders cannot show timing or duration.
Establish how you know the duration: a logger trace, an alarm log or a witnessed event.
Without duration evidence, apply the worst case from the last reset, as SPS describes.
Check the stock was in original packaging and that storage before the event was good.
Look for repeats: the 20-minute allowance covers one-off events only.
Record the cause, the evidence of duration and who decided.
Check with your inspector or RP:Contact your screening and immunisation team or the manufacturer whenever you cannot evidence that an excursion was a one-off under 20 minutes.
How long do we have to keep fridge and room temperature records, and does five years cover us?
Five years is the most common figure, but sources give different periods: at least five years under EU GDP, at least one year plus the stock's storage history under the Green Book, one year after shelf life or at least three years under WHO, ten years after expiry, clinical use or disposal for HTA human-application licensees, and RCPath guidance of 15 years for blood bank fridge charts and 5 years for most pathology fridges and freezers. These are differences between regimes more than contradictions: see [how long to keep records](/guidance/records-and-data-integrity#retention).
EU GDP chapter 4.2 (2013)
Documents should be kept for the period in national legislation but at least five years, and backups of computerised data for at least five years.
Records should be kept for at least one year and cover the full storage history of anything in the fridge. Five years will generally cover that history.
For HTA human-application licensees, records critical to quality and safety, including raw data such as temperature monitoring records, are kept for at least ten years after expiry, clinical use or disposal, in their original format (paragraph 76).
Blood bank refrigerator and freezer temperature charts: 15 years (section 6.1.5). Most other pathology fridge and freezer temperature records: 5 years. This is professional guidance, not law.
Small premises: how small is "a few square metres"?
Our medicines store is a small room. Can we use a risk assessment instead of a mapping study, and how small does the room have to be?
EU GDP lets small room-temperature premises of "a few square metres" use a risk assessment instead of formal mapping. No source defines the size, and the MHRA mapping blog aimed at small areas still describes mapping with probes.
EU GDP 3.2.1 (2013)
For small premises of a few square metres at room temperature, an assessment of potential risks such as heaters should be done and monitors placed accordingly.
"A few square metres" has no numeric definition in any source reviewed, and the concession applies only to room-temperature areas, not fridges or cold rooms.
The MHRA's COVID page allowing uncontrolled transport below 20 °C is still on GOV.UK. Can we still use it?
The April 2020 MHRA page on exceptional GDP flexibilities has not been formally withdrawn, but it was framed as temporary. Normal MHRA policy on transit storage and transport control sits elsewhere.
MHRA COVID flexibilities (2020)
Temporarily allowed non-temperature-controlled transport when ambient temperature was below 20 °C, storage equipment with limited qualification completed later, and transit hub storage up to 96 hours without a WDA(H), each with a risk assessment.
Normal policy: ambient products held in transit for under 36 hours need not be on a licence; refrigerated products need a licensed site unless they stay overnight in continuously refrigerated vehicles.
Can we use mean kinetic temperature to show that a warm spell did not harm our ambient stock?
The MHRA Inspectorate discourages MKT in wholesale "without robust supporting information and methodology". The HPRA allows it under set conditions. NHS SPS describes MKT as a way to average out short peaks in clinical areas.
MHRA blog (2016)
MKT is appropriate only where the MA holder gives an acceptable MKT value for the product and the record is complete from the manufacturer. Use "in the wholesale environment without robust supporting information and methodology is therefore discouraged".
MKT applies only to controlled room temperature products, never to refrigerated ones, cannot compensate for poor control, and temperatures should not exceed 30 °C if it is used. Consult the MA holder.
MKT can average out peaks, for example a short afternoon spike, if the daily MKT stays within the labelled range. It needs continuous logger data and should not be used routinely to compensate for poor storage.
The fridge went over 8 °C for a few minutes during restocking. Is that a formal deviation we have to investigate, or just something we note?
The HPRA defines an excursion as a departure from labelled conditions for any duration. A 2018 reply on the MHRA Inspectorate blog says very short excursions are typically noted but not raised as individual deviations. SPS and UKHSA use 20 minutes as the point below which a one-off rise is unlikely to matter, in clinical settings only: it does not apply to WDA or MIA holders, for whom every excursion is a deviation under GDP and GMP.
HPRA IA-G0011 (2020)
An excursion is a deviation from labelled storage conditions for any duration of time, to be notified to the RP and investigated.
A reply on the MHRA Inspectorate blog, to a question about door openings on a clinical trial (IMP) fridge, says it is "not typical that very short excursions (e.g. 5 minutes or under) are reported in individual formal deviations, but these should still be noted". This is a reply to a reader comment, not formal guidance.
A rise above 8 °C lasting under 20 minutes, with medicines in original packaging, is unlikely to have warmed them above 8 °C, but every excursion should still be recorded locally.
Record every excursion, however short, so patterns can be seen.
Classify by duration, peak and cause using thresholds written in your own SOP.
Check whether the product was buffered by its packaging and a full fridge.
Escalate repeats or unexplained excursions regardless of duration.
Make sure the threshold you use matches your regulator and setting.
Check with your inspector or RP:Check with your inspector or RP if your SOP uses a minimum duration and you hold an HPRA or MHRA wholesale authorisation.
Our fridge minimum dropped to 1 °C overnight. Is a low reading less of a worry than a high one?
UKHSA says vaccines with a recorded minimum between 0 and +2 °C are unlikely to have been affected. SPS says any excursion below 2 °C is always potentially significant, and the SW ICARS toolkit treats any time below +2 °C as a reportable breach.
UKHSA incident guidance (2022)
Vaccines whose recorded minimum was between 0 °C and +2 °C are unlikely to have been affected and can usually be used to expiry where the fridge temperature has been verified.
Excursions below 2 °C are always potentially significant and should be fully explored, because freezing can irreversibly damage biologicals and vaccines.
Check whether any product actually froze, especially items touching the back wall or cooling plate.
Verify the thermometer is accurate before relying on the reading.
Check each product's freeze sensitivity with the manufacturer or SPS tool.
Find out why the fridge ran cold: thermostat fault, overloading or probe position.
Follow the reporting route for your setting and record the decision.
Check with your inspector or RP:Contact your screening and immunisation team or the manufacturer for any low reading involving vaccines or biologicals.
Domestic fridges: never, not recommended, or if tested?
Can we use a household fridge for medicines if it holds 2 to 8 °C?
UK vaccine guidance says domestic fridges must not be used. SPS and CQC say they are not recommended, with a narrow care home exception. WHO accepts household-style units only if independently tested against a pharmaceutical refrigerator standard.
Green Book (2013)
Vaccines and diluents must be stored in specialised pharmaceutical fridges; ordinary domestic fridges must not be used.
Domestic fridges are not recommended. In exceptional circumstances, small services that rarely hold fridge items could use a locked container in the food fridge with full monitoring and a risk assessment.
Refrigerators should be purpose-designed; household-style units are acceptable only if independently tested against a recognised pharmaceutical refrigerator standard.
Re-mapping: every year, every three years, or on triggers?
How often do we have to repeat temperature mapping once the first study is done?
EU GDP, the MHRA, the HPRA and SPS set no fixed re-mapping interval and leave it to risk assessment and change triggers. WHO gives "every three years" as an example.
EU GDP 3.2.1 (2013)
Mapping should be repeated according to a risk assessment or whenever significant modifications are made to the facility or temperature control equipment.
Periodic re-mapping may be needed, for example every three years; where continuous fixed monitors exist, a periodic review of performance may be more appropriate.
Write a list of change triggers that require re-mapping.
Review monitoring data periodically for drift or unexplained variation.
Set an interval based on how critical and costly the stock is, and how stable the space has been.
Re-map after any change to layout, equipment, set point or heating.
Record the rationale and revisit it at self-inspection.
Check with your inspector or RP:Ask your inspector if your chosen interval is longer than a common industry convention and you want to confirm your rationale is acceptable.
We map our ambient warehouse in summer and winter. Do we need to do the same for the cold room inside it?
The European Commission GDP Q&A says mapping should be done in different seasons, with no exception for cold rooms. WHO Supplement 8 says two-season mapping is typically not necessary for cold rooms and freezer rooms.
European Commission GDP Q&A (2014)
Initial mapping should be done in different seasons, when the highest and lowest temperatures are reached.
Warehouses should be mapped empty and full, in summer and winter. Cold rooms should be mapped on installation empty and full, with an ambient probe outside the unit.
Check where the cold room sits: inside a controlled building, or against an external wall or roof.
Look at how much the surrounding ambient temperature swings between seasons.
Review whether the refrigeration plant has headroom on the hottest days.
Use an external ambient probe during mapping to show the conditions tested.
Record why you did or did not map in both seasons.
Check with your inspector or RP:Check with your inspector if your cold room is in an uncontrolled or external building and you plan a single-season study.
How often do our probes need calibrating, and is a single-point check enough?
EU GDP makes storage calibration risk-based but sets at least yearly for transport. UK vaccine and SPS guidance say annually. SPS and the HPRA describe multi-point calibration, while WHO describes a single ice-point check only as a fallback.
EU GDP 3.3 and 9.2 (2013)
Storage monitoring equipment is calibrated at intervals based on risk and reliability; transport monitoring equipment at least once a year.
Sensors should be calibrated annually to ±0.5 °C by an accredited service. 2-point or 3-point calibration is acceptable, with 3-point, in SPS's reading, required in departments under an MHRA licence (the MHRA and EU GDP texts are risk-based and set no number of points).
Data logger fitted: do we still write it down every day?
We have a continuous logger with alerts on the vaccine fridge. Do we still need the daily paper reading?
CQC says GP practices must still read, record and reset each working day even with a logger. SPS points to continuous monitoring with remote alerts as the answer to closed periods. The MHRA treats stored electronic data as the raw data, with a paper chart as a secondary record.
CQC GP mythbuster 17 (2024)
Even with a data logger, practices must still read and record minimum, maximum and current temperatures and reset each working day. It does not address whether an alarmed remote system could replace this.
Monitor continuously, check, record and reset min/max at least daily, with an independent daily check. Networked alerts are "the most robust means of alerting users" out of hours.
Room temperature: 25 °C ceiling or a tolerance in heatwaves?
Our treatment room went over 25 °C during a heatwave. Do we have to quarantine everything labelled "store below 25 °C"?
UKHSA says most medicines should be kept below 25 °C. SPS says a single excursion of less than 5 °C over the limit for under seven days is unlikely to matter. The 2015 NHS QA document set a stricter quarantine trigger and noted its Q10 method is not accepted under GDP.
UKHSA hot weather guidance (2024)
Most medicines should be kept below 25 °C, somewhere cool and dry, away from direct sunlight.
Ideally all storage areas would be controlled between 15 and 25 °C; storing outside stated conditions takes a medicine outside its licence and needs governance review.
A single, isolated excursion of less than 5 °C above the labelled maximum, lasting under 7 days, is unlikely to harm medicines. It does not cover rooms that run hot every summer, or WDA holders.
Example trigger: quarantine ambient stock if the limit was exceeded by more than 5 °C or for more than a day. Its Q10 expiry-reduction method is for clinical areas and is not accepted by the MHRA under GDP.
We operate in GB and Ireland. How much experience does our Responsible Person need?
MHRA Guidance Note 6 says at least one year's practical experience of the licensed activities. The HPRA GDP guide treats six months' full-time direct GDP experience as the minimum.
MHRA Guidance Note 6 (2023)
The RP must have at least one year's practical experience of the licensed activities, with knowledge of storage conditions and temperature control.
We are closed at weekends and bank holidays. How do we know the fridge was fine, and is a Monday reading enough?
GP guidance and contracts require readings on working days, so weekends are covered only by the next max/min reading. CQC asks care homes for at least once a day. WHO expects checks seven days a week, and SPS points to continuous monitoring with remote alerts.
NHS England GP enhanced service (2024–26, now expired)
Readings must be recorded on all working days, with action when out of range.
Work out the longest unattended period, for example a bank holiday weekend.
Estimate how much stock would be at risk if the fridge failed at the start of it.
Decide how you would detect a failure: logger download, remote alert or a call-out.
Name who responds to an alert and test the route.
Add a closing check before weekends and holidays to your routine.
Check with your inspector or RP:Talk to your commissioner or inspector if your stock value or patient impact makes a Monday-only check hard to justify.
How often should we test that our temperature alarms actually work?
EU GDP says alarms should be regularly tested but gives no frequency. WHO says at least every six months. SPS and the HPRA give no interval, leaving it to local policy.
EU GDP 3.3 (2013)
Alarm levels should be set appropriately and alarms regularly tested; no test frequency is defined.
Paper and electronic together: control it or replace it?
We keep a paper log and an electronic logger. Is that acceptable, or should we move to one system?
The MHRA accepts hybrid systems if the full data set is defined, reviewed and kept. WHO says replacing hybrids should be a priority with a documented CAPA plan.
MHRA data integrity guidance (2018)
Hybrid systems are allowed if it is clearly documented what makes up the whole data set and all of it is reviewed and kept, with more data review because changes may not be attributable.
Where hybrid systems are used, even for an interim period, risks should be assessed with documented controls, and replacing them should be a priority with a documented CAPA plan.
Is there a maximum we can fill the medicines fridge to?
The Green Book and the MHRA give no percentage and ask for air circulation and evidence. Public Health Scotland says no more than two-thirds, and Northern Ireland no more than half.
Green Book (2013)
Leave enough space for air to circulate and keep vaccines away from side and back walls; no percentage is given.
Our courier holds stock at a depot overnight. When does that depot need a wholesale authorisation?
MHRA policy uses 36 hours for ambient products in transit and requires licensing for refrigerated storage unless it stays in refrigerated vehicles. The HPRA uses 48 hours and treats any storage of refrigerated product as needing a WDA.
MHRA blog (2016)
A site need not be named on a licence if ambient products are held for under 36 hours in transit and ownership has not transferred. Refrigerated products need a licensed site unless they stay overnight in continuously refrigerated vehicles.
GMP and GDP: map on risk, or requalify on a schedule?
We manufacture and distribute from the same site. Do our stores follow the GDP mapping rule or the GMP requalification rule?
GDP asks for mapping before use, then a repeat on risk assessment or after significant change. GMP Annex 15 asks for qualification and evaluation at an appropriate frequency, with any fixed period justified, and never uses the word mapping. Neither sets an interval for storage; Annex 1 sets fixed intervals only for cleanrooms. These are differences of approach rather than a conflict.
EU GDP 3.2.1 (2013)
Map each storage area before use under representative conditions, place monitoring at the extremes, and repeat the mapping according to a risk assessment or after significant modifications to the facility or temperature control equipment.
Evaluate facilities, equipment and systems at an appropriate frequency to confirm they remain in control. Where requalification is done at a specific time period, justify the period and define the evaluation criteria.
Requalify grade A and B cleanrooms at least every 6 months and grade C and D at least every 12 months. This applies to classified cleanrooms, not storage.
Periodic re-mapping may be needed, for example every three years; where continuous fixed monitors exist, a periodic re-evaluation may be more appropriate.
List every space and the regime each one serves: GMP storage, GDP storage, or both.
Qualify GMP spaces through Annex 15, using mapping as the evidence, and map GDP spaces before use.
Run one change control list of remapping triggers for every space.
Review continuous monitoring data every year as the periodic evaluation.
Set and justify an interval for each space and record it in the validation master plan.
Check with your inspector or RP:Ask your GMP inspector if a shared store relies on GDP-style trigger-based remapping alone, with no periodic evaluation.
GDP says at least five years. Our GMP batches expire in three years. Which retention period applies to storage temperature records?
GDP keeps all documents at least five years. GMP ties retention to the batch: one year after expiry or five years after QP certification, whichever is longer, and keeps raw data supporting the marketing authorisation, such as stability data, while the authorisation is in force. API records run one year after expiry, or three years after complete distribution for retest-dated APIs. Chapter 4 does not name temperature records, so which period applies depends on what the record supports.
EU GDP 4.2 (2013)
Keep documents for the period in national legislation but at least five years; keep backups of computerised data at least five years (3.3.1).
Keep batch documentation one year after batch expiry or at least five years after QP certification, whichever is longer. Keep critical documentation, including raw data relating to validation or stability that supports the marketing authorisation, while the authorisation remains in force.
Decide what each temperature record supports: a batch release, a stability study, a qualification or distribution.
Apply the period of the record it supports, and never less than five years where GDP also applies.
Keep stability chamber records with the stability data for the life of the marketing authorisation.
Make sure the monitoring system, or your archive, keeps the data and audit trail readable that long.
Write one retention schedule with the clause behind each period.
Check with your inspector or RP:Ask your inspector if you plan to keep storage records for less than five years on the strength of a short batch expiry.
GMP and GDP: alarm testing and calibration intervals
GDP says alarms must be regularly tested and transport loggers calibrated yearly. GMP says neither. What should a manufacturer do?
Only GDP asks for alarms to be regularly tested (3.3) and gives a calibration minimum, at least once a year for transport monitoring equipment (9.2). GMP asks for calibration at defined intervals (3.41) and, in sterile areas, for alarms to be acknowledged and trended and delays justified (Annex 1), but says nothing about testing alarms. An MHRA Inspectorate blog reply (Sara Berry, 26 February 2018) says "the general expectation is that equipment should be calibrated annually".
EU GDP 3.3 and 9.2 (2013)
Set alarm levels appropriately and test alarms regularly. Calibrate transport monitoring equipment at regular intervals, at least once a year.
Warn operators instantly of critical pressure failures, justify any alarm delay in the contamination control strategy, and acknowledge and trend alarm events. No alarm test interval is given.
In 20 minutes we go through how you decide and record excursions, mapping and monitoring, and where the gaps are. You get them in writing. No product tour.