A study of mapping studies: Open-door versus defrost cycles

refrigerator qualification open doors and defrost cycles
Paul Daniel, Vaisala
Senior GxP Regulatory Compliance Expert
Published:
Life Science
We receive many questions from customers on mapping studies and monitoring applications. Often the questions generate interesting exchanges that dig deep into the application details, methods used, and rationales. In this blog we share one such exchange between Paul Daniel, Vaisala's Senior GxP Regulatory expert, and a customer who qualifies refrigerators and regularly performs mapping studies as an Installation/Operational/Performance Qualification service provider.
 

C wrote:

Dear Paul,
 
I was reading the blog: “Using MKT to Justify Normal Temperature Excursions during Qualification: Should you?” and a few questions occurred to me. I’m curious about freezer defrost cycles and wanted to know what approach you recommend for fridges and freezers that have built-in defrost cycles that create temperature excursions from the acceptance range.
 
I have three questions: 
 
  1. Do you feel that a sufficient impact on the products can be shown by measuring just one buffered probe inside the chamber throughout all mapping studies? (i.e. 10 or more air probes and one buffered probe during several studies; 24-hour, empty/loaded, open door, etc.)
     
  2. Should the buffered probe be positioned at the center of the chamber? Or, would you position it at a predetermined worst-case location for each study?
     
  3. What about using the open-door cycle as a comparison against defrost cycles? 
We rely heavily on Vaisala products to do much of this work and you seem to be well versed in these topics, so I'd love to hear your thoughts.
Thank you!
 
- C
 

Paul wrote:

Hi C,

In regards to your first question about several air probes and one buffered probe, this sounds like a novel approach.  In my experience with mapping studies, I’ve used either air probes or buffered probes, but never a mix. To me, your approach sounds reasonable, though if I were to take this approach, I would always place an air probe immediately beside the buffered probe to allow for comparison against all other air probes.
 
I am under the assumption that the reason you wanted to map with a buffered sensor was to show that the monitoring probe (likely in glycol buffer) was a good representation of the entire chamber. But perhaps your motivation was to show that the worst-case locations weren’t actually all that bad.  I guess that depends on where you place that buffered probe.  For me, the question comes down to this: what do you intend to do with the data from the single buffered mapping location?  The answer to this question can inform where a probe should be placed. And this is why I would tend to have either all buffered probes, or all air probes;  I want to use the data to say the chamber is good for any storage, or if there are limits necessary on what can be stored due to the volumes necessary to buffer the temperature swings.
 
For your second question about placing the buffered probe at the center, versus a worst-case location, we have several blogs (linked below) on these topics.
 
But for the purposes of this discussion, I will simplify my stance to say that I think worst-case locations are artifacts of the way we analyze data.
 
The maximum and minimum values only matter because we can quickly compare those to acceptance criteria.  I have never seen a company place a monitoring sensor in a “hot spot” that had been identified by the mapping study. At least not when dealing with a free-standing refrigerator or freezer because these are typically too small inside to use up storage space with a sensor. The monitoring sensors in a small chamber typically end up on the wall opposite the door hinges. So, I would anticipate that and then place the buffered probe wherever you intend to put the monitoring probe (with an air probe right beside it).
 
Finally, you asked about using the open-door cycle as a comparison against the defrost cycles.  While it's an interesting approach, I would caution against it because there is no way of knowing if the open-door state and the defrost state are comparable in temperature, or in duration.
 
Let’s look at this a little more closely. An open door brings warm air into the chamber only from the front. By contrast, the defrost cycle typically involves heating coils, which brings short-term warm air from the back (usually). An open door brings warmed air from the front whereas a defrost cycle conducts heat from the back of the chamber. I don’t think those are directly comparable.
 
I suggest that the next time you are mapping, do a preliminary study to determine the defrost frequency. Once you know that, you should be able to catch a defrost cycle during your mapping and see how bad it really is.
 

- Paul

C wrote:

Thank you Paul! Here is some background on our mapping study process: We map the temperature in an empty chamber for 24 hours. This typically captures one to three defrost cycles. We repeat the study under loaded conditions and get similar results. We then perform an open-door study, where the door is opened for one minute to simulate typical usage of the equipment.
 
Our thinking is that the data collected from the open-door study could easily be compared directly against a defrost cycle. We would be able to analyze the data and see, if in fact, the duration and impact of the defrost cycle is more or less significant on all mapping locations than the impact of the open door study.
 
For example, the maximum temperature of a typical defrost cycle is 10°C, which generates a temperature excursion for three minutes. The maximum temperature of an open door could be 15°C, generating a temperature excursion duration of five minutes. Of course, if the impact of defrost is more significant we would need to consider other means to justify the use of the chamber (e.g. Glycol probes as per above).
 
I completely agree that the source of heat is not the same, but if the impact is always less than that of an open door, would we not be able to claim that the impact was less than that of typical open door use?
 
Keep in mind that we are a service company who is providing the IOPQ services. We do not own these freezers/fridges. So, in discussing the probe locations, I am strictly referring to the qualification mapping probes and not the monitoring device that the equipment owner will use to track temperature.
- C
 

Paul wrote:

Okay, I think I see your rationale.  If the general effect of a typical defrost cycle is less than the impact of a typical door opening, then we can accept the defrost cycles as “normal use” because we accept the door openings as “normal use”.  I agree that the comparison might be interesting, but my concern is that it could create a situation where you have data that both the defrost cycle and the door opening have unacceptable effects and need to mitigate both.  Another way of saying this is an overly liberal door opening policy could then hide the negative effects of an out-of-control defrost cycle.  (Although your defrost cycle seems pretty normal.)
 
Please remember that a probe buffered in glycol doesn’t mitigate problems, it conceals them.  Things that actually mitigate risk to product include: procedural controls like not storing small volumes in the chamber, or limits on how long or how often doors can be opened.

I hope this is helpful!
 
See also our recently revised application note: “5 Rules of Sensor Placement in Validation/Mapping Applications
Or, our recorded webinar: “Mapping made easy: where to place sensors & why

Join the discussion

Do you have further thoughts or questions on this topic? Please place your comments in the fields below.
 

Mean Kinetic Temperature Webinar

Vaisala's Senior Regulatory Expert Paul Daniel provides you with 6 clear Do's and Don'ts for using the calculation, including where it is useful in decision-making for regulated industries, and where it is not.

You will leave the webinar with an understanding of MKT as a non-linear weighted average that shows the effects of temperature variations over time. Included is a Q&A section where live webinar participants asked about best practices for MKT in their own controlled environments. 

Key Learning Objectives

  1. References and resources for MKT calculation
  2. The Arrhenius equation and the non-linear effect of temperature excursions
  3. How regulatory bodies (like the FDA and EC) recommend MKT
  4. What data to use when MKT is not appropriate

Comment

Sachin Bhandari

Dec 13, 2021
Which guideline should be followed for MKT calculation

Paul Daniel

Dec 14, 2021
The guidelines you choose to follow, for any GxP purpose, should be selected based on the regulatory agencies to which you report. This in turn, should be more or less dictated by the location of your facilities, and the countries into which you wish to import your products.

The FDA is widely accepted as a reliable source and the FDA relies on the USP for specifics of MKT. As a start, you could look at USP chapter <1079.2> “Mean Kinetic Temperature In The Evaluation Of Temperature Excursions During Storage And Transportation Of Drug Products”. This has a nice explanation of the application of MKT. If you want something more general, you can always refer to our Vaisala whitepaper on the topic:
https://www.vaisala.com/sites/default/files/documents/CEN-G-Mean-Kinetic-Temperature-Application-Note-B211534EN-A-FINAL.pdf

Before you begin using MKT for GMP decision, please verify that the relevant regulatory authorities support the use of MKT for your application. For instance, it is not universally accepted to use MKT to resolve deviation encountered during shipping.

Add new comment