GxP Temperature Mapping: Acceptance Criteria & Reporting

In this blog, Vaisala Senior GxP Regulatory Expert Paul Daniel and validation/commissioning expert Nathan Roman answer questions on validation reports and acceptance criteria. These were questions we didn't have time to answer at our webinar: Controlled temperature chambers: Qualification, mapping & monitoring.

All webinar questions and answers from this webinar are available in this PDF. 

Transcript: 

Question: In a thermal mapping report, what data should be included? Do we include data such as minimum, maximum temperature? And what about mean kinetic temperature?

[00:01:08]

Paul Daniel:  Nate you do more mapping than I do lately. Why don't you take this one?

Nathan Roman: Sure thing, Paul. I think the final report that we're speaking of is the thermal mapping report. This report should always be a comprehensive overview of the entire qualification process. It should identify the equipment or the system undergoing qualification. It should list all performed test cases, including test results. It should summarize the key findings of the study and document any deviations, their impact and how they were resolved.

Ultimately, the report describes whether the equipment or system met the pre-defined requirements and states whether it can be taken into use. And as for the data to summarize, it's standard to provide the minimum maximum and the average temperatures. That gives us a clear and direct representation of the temperature range and average within your equipment, which then can be compared to your operating range and your acceptance criteria. Not necessarily the operating range, but your process range.

[00:02:36]

Paul Daniel: I love that answer, Nate, because you answered like a validation guy. And I've been working for a vendor for so long that I was literally just thinking about the report that gets printed out from, say, the viewLinc system showing the thermal data. And on that report I want to see the raw data. I want to see the maximum and minimum so I know if it was in specification and I don't really care about anything else.

I have to touch on this mean kinetic temperature part of the question. In my opinion, mean kinetic temperature has no place in mapping. It really has no place in Good Distribution Practices. I've written a whole white paper on it that we can put a link to in the transcript (below), but it's been adapted for use in explaining excursions. I don't see any reason why you would want to use a tool that's used to explain excursions while you're trying to qualify a system to prove that it doesn't have excursions. So think you're going into dangerous territory if you're considering reporting mean kinetic temperatures in a mapping report.

Webinar: Mean Kinetic Temperature: History, Math, and Applications

Application Note: Mean Kinetic Temperature in GxP Environments

Question:  Our acceptance criteria are typically based on the operational range. Should we have the same acceptance requirement for a 2 to 8°C fridge versus -80°C freezer with an operating range that goes from -90°C to -70°C? 

[00:04:37] Paul Daniel: I think this question is too much about the equipment and not enough about the process. The acceptance criteria are based on what you're storing in the chamber, the product. If the product requires, 1 to 10°C as you would for specialized blood storage, then that range is your acceptance criteria. A range of 1 to 10°C is a drug that requires refrigerated 2 to 8°C storage. Then your acceptance criteria are 2 to 8°C.

Don't get confused by thinking you're mapping a refrigerator. You're mapping a storage process for a product that needs to be controlled at temperatures between 2 to 8°C to protect its quality. The mechanism being used for this process, the equipment being used for this process, just happens to be a refrigerator. What do you think about this question, Nate?

[00:05:27] Nathan Roman:  I agree. However, to take it a step further, but basically repeating what you said, acceptance criteria for temperature mapping studies should be defined based on the  specific process requirements of the equipment being validated. It's crucial to consider the intended use of the equipment and the criticality of the products being stored or processed. So that's the process range. I know the question you have asked is about the acceptance criteria, typically based on the operational range, the process range, the unit's operational range.

...Acceptance criteria for temperature mapping studies should be defined based on the  specific process requirements of the equipment being validated.

I think this is where you were you were headed Paul; there are two distinct aspects when discussing acceptance criteria for qualification. Process range is the specific temperature range or other parameters within a CTU intended to operate during its normal use. This is determined by the requirements and specifications of the processes that it supports. The range defines the acceptable limits within the CTU and how it must perform to ensure the desired outcomes of the process. The unit's operational range, on the other hand, represents a broader range of operating conditions in which the CTU is capable of functioning correctly and reliably. It encompasses the full range of temperatures and the other environmental conditions in which the CTU can fulfill its intended function.

[00:07:25] Paul Daniel: There's a really interesting application that we see occasionally where customers have an incubator wherein they want multiple set points, sometimes at a lower set point for one product and a higher set point for another. And they ask questions, like: "How do I map it? Should I map the same unit at different set points?"

Changing the set point affects a chamber's ability to perform at one of the lower set points. The acceptance criteria can change for the same unit depending on how it's being used. I think we should also understand that acceptance criteria is not just about temperature range, but also can include acceptable excursions from that range. We should know how long a chamber can be out of specification and still be considered to be operating in an acceptable manner.

Question: How can you establish the number of sensors to remove from a thermal mapping due to technical failures without affecting the study?

[00:08:35] Paul Daniel: Back in the old days when I started mapping, we used thermocouples, which are prone to failure and accuracy drift, especially when you're bending them and trying to position them. All these things contribute to a loss of measurement stability and eventual failure. If you got through a mapping study with all your sensors still functioning and all of them passing post calibration, you got very lucky. So there was a logistic requirement to allow some of them to fail in order to not have to repeat your study every time you lost one thermocouple.

We used 10% allowable failure as common sense. By analogy, you could ask the same thing of a vehicle. How many wheels can it lose and still function? You're out of luck. If you lose one wheel on a four wheeled car, you're out of luck. But when you have a six-wheeled vehicle, you might get away without losing one wheel. Mapping requires a lot of sensors... 

But, Nate, I know you use more modern sensors than I did in my mapping day, and you may have something much more technical to add to this rather than just my common sense of just trying to make sure you don't have to repeat your study due to a bad thermocouple.

[00:10:05] Nathan Roman: So I think your point is absolutely correct and in all the validation protocols that I've written and executed, a 90% confidence level is often used when it comes to execution or sensor post calibration or post verification in our industry. If this wasn't specified in your protocol, then you would need to treat it as a deviation.

The point is that the failure rate is specified ahead of time in your protocol so that you have a guideline to follow for the minimum number of sensors required for a study. This should be established during the document development phase and based on the size and complexity of the area that is being mapped. If sensors fail during the study, the remaining sensors should still provide sufficient coverage. It's like Paul said about the car and the wheels. Note that, if more than the established number of sensors fail, I would say the study should be repeated with properly functioning sensors that are calibrated and cover the range that you need.

Please leave comments in the field below...