Hello, good people of the world! So you have a cGMP warehouse you need to qualify? Well, this post will provide plenty of information to get you started.
First, how do you know your warehouse is a “cGMP” warehouse. Well, it is if:
- The warehouse will house any raw material, excipient, intermediate, or finished product
- The warehouse will house equipment or process consumables, such as filters, bags, disposable tubing, etc.
- Any cGMP operations will occur in the warehouse, such as sampling, dispensing, kitting, track and trace, etc.
Any examples of common or even possible non-cGMP warehouse space in the Biopharma industry? Please leave a comment below.
Now, what are the cGMP requirements of the warehouse? Are you controlling just temperature or temperature and humidity? What about light exposure? The warehouse isn’t supposed to be sterile is it? Is it supposed to be sanitary? Better find out. You can look at the requirements of the materials you’ll be storing, from vendor recommendations and/or stability studies, but typically agencies will expect you to meet specific temperature requirements regardless. For example, EMA (Europe) and MHRA (UK) will expect room-temperature space to control in the range of 15-25 deg C, even if your stability studies show your product is stable at 30 deg C. If you choose to qualify to a wider range, be prepared to have solid data in stability studies to present to inspectors, or risk getting slapped with an observation.
Purpose of Qualification
The purpose of warehouse qualification is to show that your warehouse meets temperature and/or humidity requirements across the space, such that any materials stored would be protected from outside fluctuations of environmental parameters.The qualification will also:
- identify “cold” and “hot” spots or spots of high or low relative humidity – this is important for determining the locations of permanent sensors for ongoing monitoring
- indicate performance of the supporting HVAC systems – particularly important if you’re following the trend to commission-only the HVAC system itself
Note: because of item 1 permanent sensors should not be installed until after initial qualification is complete. Agencies will expect this!
Characterizing your Warehouse
Like environmental mapping of chambers and cold rooms, mapping a warehouse should record data uniformly across the space where materials will be stored. Draw a picture. Your diagram should include all entry/egress points, any windows and skylights, location of air supply and return registers, location of any controlling sensors, RTDs, thermostats, and demarcation of areas where materials will be stored.
Typically warehouses are mapped in the as-found condition, because it is not practical to completely empty or fill a warehouse for qualification studies. The risk this introduces is mitigated by repeating mapping, initially in winter and summer months, and then regularly after that, every year to 5 years (to be determined by the results of summer/winter qualifications), and having ongoing monitoring via permanent sensors.
If finished goods, intermediates, raw materials (or whatever you’re going to store in the warehouse) is not going to be stored directly against the wall (and it should not – OSHA requirement) then don’t place sensors directly against the wall. Get them off 12 inches or so off the wall to avoid excursions due to measuring wall conditions, not room air conditions.
External temperature influencers to keep in mind: major heat gains and losses are from the external environment with the most significant spots being near doors and windows. Gains and losses will be consistent across the space given the structure is uniform. Internal temperature influencers include: lighting, product, equipment, people, HVAC systems, and other mechanical systems operating inside the warehouse (e.g. pallet wrappers, compactors, etc.).
Regulations and guidance to consider
FDA 21CFR211.142 “Warehousing procedures” states quarantine procedures need to be in place, so make sure that’s checked somewhere. It also lists the parameters of temperature, humidity, and light, so make sure you check if your finished product, intermediates, and/or raw materials have any requirements along those lines.
USP <1079> USP 35-NF 25: “Good Storage and Shipping
Practices” has additional requirements:
- Electronic temperature monitors should be calibrated to NIST (or other suitable) standard with an accuracy of ±0.5 deg C (or better)
- Air flow and load patterns should be determined to ensure that all areas are suitable for product storage
- The following factors, which may contribute to temperature variability, should be considered during the process of temperature mapping storage locations: (1) size of the space; (2) location of HVAC equipment, space heaters, and air conditioners; (3) sunfacing walls; (4) low ceilings or roofs; (5) geographic location of the area being mapped; (6) airflow inside the storage location; (7) temperature variability outside the storage location; (8) workflow variation and movement of equipment (weekday vs. weekend); (9) loading or storage patterns of product; (10) equipment capabilities (e.g. defrost mode, cycle mode); and (11) SOPs.
- The recording of temperatures during the thermal mapping of a warehouse or cold room should be sufficient in time frame to capture workflow variation that may impact air flow and the resulting temperature fluctuation (i.e., a period of two weeks is recommended for data collection enabling the capture of two week/weekend workflow cycles).
ISPE Good Practice Guide 8: Cold Chain Management, 2011:
- Stability data is the basis for decisions around storage (and shipping). Stability testing is well-defined in ICH guideline Q1a, Stability Testing of New Drug Substances and Products
Equipment you’ll need to complete a warehouse qualification
You’ll need temperature and humidity sensors, if you’re measuring that. Keep in mind minimum accuracy requirements (±0.5 deg C per USP <1079>).
Vendors to look at for sensors: Sensitech, Vaisala, Onset, Ellab. I’ve used Sensitech’s TempTales, Onset’s Hobos, and Ellab’s Tracksense Pro system with success. Vaisala looks like a good choice as well – I’m especially interested in their device that measures three points with one sensor. Ellab has very nice sensors but you’ll pay a premium for them as of this writing (Jan ’13) Hopefully they’ll come down in price in the future.
One thing to thing about is how you’re going to control of your data and 21CFR11 compliance. While I’ve never see it or heard of it happening, it is possible that an inspector could ask for evidence that data recorded by sensors and presented in a qualification final report is 21CFR11 compliant. Many (most) vendors supplying the biopharma industry have software that claims to be 21CFR11 compliant, so just think about how you would answer this question if it came up.
Who does what (roles and responsibilities)
You’ll need the following folks on your team:
- Someone to write study protocols (Validation)
- Someone to review and approve study protocols (Validation, system owner, Engineering, Quality Assurance)
- Someone to execute study protocols (Validation)
- Someone to review executed protocols for compliance with protocol instructions, GDPs, and for completeness (Validation)
- Someone to review and approve execution deviations (Validation, system owner, Engineering, Quality Assurance)
- Someone to write the study reports (Validation)
- Someone to review and approve study reports(Validation, system owner, Engineering, Quality Assurance)
Test Duration – initial qualifications of an empty, unused warehouse can probably be mapped for 3 days (72 hours). If you have dynamic operations or a warehouse in-use, you may want to extend the mapping period from 7 days to two weeks in order to capture the normal differences in schedule through the workweek (per USP 35-NF 25 <1079>). 24 hours seems like too short for a large warehouse space to me – what if you have a 2 hour stabilization period in that 24 hours? Then almost 10% of your study time is taken up be stabilization.
Data Sample Rate – my opinion is get as much data as possible without overloading the sensors. Data sample rate is then determined by the duration of the study (± some contingency factor) and the storage capacity of the test instruments. For instance, if your device can store 5,000 data points and you want to do a 72-hour test, you can easily sample every 5 minutes (72*60/5=2,160) but maybe you want to sample every minute for better resolution in your data (72*60/1=4,320)?
Number and placement of sensors (how to record actual placement) – sensors are placed in areas considered for mapped where product will be stored, with the exception of at controlling sensors/RTD/Thermostat. Actual locations of sensors should be recorded on a protocol form at the time of placement so that the test is repeatable and troubleshooting is possible if a problem should occur.
Outside sensors – it’s industry standard to measure outside temperature at the same sample rate as internal sensors. While it’s interesting to see what the external temperature profile looks like I’ve never actually used the data to support and conclusion or troubleshoot any problem, but you definitely do want to have it just in case. One watch out: make sure outside sensors are secured or hidden if outside the warehouse is common or public – I’ve had sensors disappear during qualification!
Door Entry Logs – door entry logs should be posted at all doors (on both sides) so that opening of doors can be tracked in the case they influence the temperature profile. This is less likely with warehouse that with smaller cold room and walk-in freezers.
Data Analysis, what to do with your data – you can calculate maximums, minimums, averages, mean kinetic temperatures, “hot” and “cold” spots – from absolute min, max and/or average min, max and/or min, max mean kinetic temps.
You may want to set your permanent sensor monitoring alarm delay based on the data you collected during PQ. Keep in mind that the temperature requirements of what you are storing are the main driver of your alarms, but if your requirements are not too strict you may want to use alarm delays to avoid nuisance alarms. Note that temperature fluctuations recorded during PQ will be dampened if permanent sensors are in solution (e.g. water or glycol).
Post-study test instrument verification – I’d recommend performing a post-study verification of test instrument calibration to ensure all instruments were operating in the calibrated range during the study. This closes the loop on calibrated instruments and is better than finding out 9 months later when the calibration comes up that the instrument was somehow out of tolerance.
Acceptance criteria should, as always, be concise and clear. Think through what it would mean if any acceptance criterion was not met – better to know before than scramble for a solution after. Typical acceptance criteria could include: all readings meet requirements, all sensors pass post-verification.
Be sure to think about how to handle sensors that fail during mapping and how to handle sensors that fail post-verification. This is more common than you might think.
Prerequisites to qualification
There’s some things you want to make sure are in place before starting a warehouse mapping study. Typically these would be explicitly verified in a prerequisite section of the study protocol:
- HVAC related to the space is commissioned and/or qualified with no outstanding issues
- Controlling sensors/RTDs are calibrated
- If performing some kind of dynamic testing, operating SOPs should be approved for qualification use, at a minimum
- Test instruments are calibrated
Because I’m so good to you, here’s a template you can use to create your very own warehouse qualification protocol:
Please leave any questions or comments below or email me directly: Michael@dikuw.com