The various GMP regulations around the world require that environmental conditions (generally temperature and humidity) be controlled, monitored, and recorded in areas where it can affect product quality. There is no sampling frequency specified in the regulations.
In practice, the stability testing of a drug product produces data that will define the storage requirements of that product. A drug's stability data gives us an idea of how tolerant the product is to excursions of environmental parameters like temperature and humidity. Manufacturers will then produce their own internal "Storage Specifications," which guide the storage of that drug product, i.e., 2 - 8 degrees C." The Specifications will require the creation of a storage and monitoring procedure and that procedure should state the storage limits and sensor sampling frequency.
There is a regulatory requirement to have written procedures on how drugs are handled and stored. There is also a regulatory requirement to create a deviation report whenever there is a fail to follow procedure. However, it's never a good idea to create a situation where you will have to deal with documenting a lot of deviations, so storage specifications usually have some "wiggle room" in terms of allowable excursions. In practice, storage specifications can look like this: "2-8 degrees C, with allowable excursions of Not More Than (NMT) 30 minutes." That means temperatures can be out of specification for up to 30 minutes without necessitating a deviation report.
What does this mean for sampling frequency?
- Basically, it means you want the slowest sampling time that your specifications will allow (for actual GMP data points).
- However, since we also want to protect the product from catastrophic events, we also want to know immediately if there is an alarm condition.
With regard to #1, if I have an allowable excursion of 30 minutes, and I use a sampling frequency (for my logger and GMP data points) of 15 minutes, I can have one data point out-of-spec, and still, know that I haven't had an excursion requiring a deviation report. If I sample at 10 minutes, I can have two data points out-of-spec, and so forth. Generally, the practice is to choose the slowest sampling rate they can and still be able to show that they were within their excursion limits. In this scenario, choosing a 5-minute interval only means I have to store 3 times as much data, and my reports are three times longer. It doesn't offer me meaningful resolution with regards to my storage specifications.
With regard to #2, this is why the Vaisala continuous monitoring system samples in approximately 1-minute intervals to get "real-time" data. However, this is for alarming purposes and it's often not considered GMP data because we defined our GMP sampling frequency in our monitoring procedure.
In my estimation, it would be hard to justify a sampling frequency of greater than 15 minutes simply because it's such a common practice; most companies choose sampling frequencies at 10 or 15 minutes. That being said, I have seen distributors that use a 1-hour interval. This is likely because the USP definition of "Controlled Room Temperature" allows 24-hour excursions. Also, this only applies to storage. If you are monitoring a dynamic process with temperature changes, then a faster sampling rate is appropriate.
Don't miss the upcoming webinar: How to Validate your Monitoring System Software. For this webinar, we will send all attendees a brand new infographic on validating the three most common types of monitoring systems - so don't miss out!