Wow, I have to say, ChatGPT is amazing after using it for the first time, just now. I didn’t write this with AI, but I did play around with it to see what it could produce. I typed in the command, “Write an 800 word article comparing retest dating vs expiration dating for commercial pharmaceutical stability programs. First define retest dating and expiration dating. Note any cost savings opportunities from using retest dating. Note cost savings by extending expiration dates. What if products fail on stability before the end of the retest period?” It actually produced a relatively good article, but a bit basic and sterile for my liking. I thought I’d open with that, though, since it is impressive how powerful the AI tool was. This tells me that we need to commit to strengthening our stability community, otherwise this really could replace most of us in our stability roles.

It so happens that retest dating and efforts I’ve seen to extend it are topics that I’m passionate about, and ChatGPT was just not able to communicate that adequately. Maybe if I’d included a command to include passionate language…I’ll try that in a poem to my wife. Just kidding of course!

So, on to the actual topic of this month’s blog – two aspects of retest dating: 1. Dating extensions and 2. Use after testing. To ensure we’re all on the same page, a retest date is the date after which a material must be tested prior to use. This form of shelf-life may be assigned to raw material and most API drug substances, but not usually to antibiotics and biologics drug substances and not to the final drug product. But, just like an expiry date, a retest date is assigned based on stability data and in the case of API drug substances at least the retest date period will be supported by annual (routine) batches.

This annual support of retest period assignments is where things begin to get interesting. One might quickly assume that any time you can extend a retest dating period it would be beneficial to the company holding the material in inventory—they would not have to retest the material as much. But the hidden stability costs need to be taken into account. You might save retesting costs for a few lots in inventory, but at the same time you’ve increased the amount of stability testing required to support the extended retest period.

Here’s an example:

Let’s say you have eight batches of a drub substance in inventory produced over the last three years. You hastily extend the retest period from 2 years to 4 years to avoid having to retest some of these batches prior to production in the next 6 months. That extension might seem like a victory, until you realize that the testing costs after extension will actually exceed testing if the retest period had been left at 2 years.

Here’s why. Besides the many man hours to implement the extension, you will likely still have to retest at least a few of these lots, unless production picks up more than anticipated—lets say at least 3 of the 8 batches will still end up being retested. But the new unforeseen cost is that now you’ve added at least two additional stability intervals (36MO and 48MO) to your stability program and over the last three years of production you must have at least three batches on stability, but probably more. Let’s conservatively assume 4 batches. That means you’ve added 8 additional stability test intervals. So, with the 3 retests and 8 new stability intervals you now have 11 total testing events for these batches when without the retest extension you would not have had more than 8 testing events.

Of more concern is the future years when production planning is more robust, and excess batch production is avoided so the company manages to prevent any batch inventory aging to retest, but you still end up with two additional non-value-added stability test intervals for each of these batches that were actually used up back in the first or second year. Oh well…

So, retest dating extensions might not be a good idea, especially for low running products. With just a few batches in inventory, it might be cheaper to just take the hit and retest the material prior to use than to keep testing it on stability for years to come. Just something to consider, along with the added risk introduced simply by adding that testing.

One other thing about retesting: Although not there anymore, an older version of ICH Q1A had verbiage implying that to “use immediately” meant within 30 days. With that thought in place, that 30-day period times 2 is can be reasonably applied to the use of retested materials—meaning 30 days for testing then an additional 30 days for use or 60 days total from initiation of retesting. The 30 days for testing is necessary especially for some of the time consuming tests for biologics. If not used in production within 60 days from initiation of retesting, then retesting would have to be initiated again. Hopefully planners are all on board to ensure that doesn’t happen!

In procedures describing these timing requirements it would be best to include justification for why the additional 60 days is not impacting the product such as:

  • Have a statistically justified calculation arguing the lack of impact of 60 days. This could be product specific and based on the 95% confidence interval utilized to justify that 24-month retest period assignment, but ideally is could be even more generic across all materials.
  • Utilizing retest specifications limits that are slightly elevated above the shelf-life limit to statistically account for the additional 60 days of degradation.
  • Calculating individually for each batch to confirm that based on retest results and the product’s degradation rate it would remain within specification for at least 60 days. The product’s degradation could be defined as linear from release to the lower shelf-life limit.

These two aspects of retesting can help any company make better informed quality decisions regarding their retesting program. Even better, it is very likely that an individual could spend some time brainstorming and come up with some notable cost savings proposals related to retesting. It might cause a headache, though, so I could recommend that you just connect to some AI system and ask, “How can I save money on retesting of pharmaceutical drug substances?” Uh…I just entered that and was given 6 interesting ideas. They weren’t earth shattering, so it might not be there just yet, but someday…

Meanwhile, back to brainstorming.

Share This Article with the Stability Community!

Share your questions and experiences

A stabilitarian encounters new situations every day. StabilityHub’s discussion forums give Stabilitarians an opportunity to ask questions and offer solutions to specific scenarios. Join in the conversations with other Stabilitiarians and share your knowledge!

A stabilitarian encounters new situations every day. StabilityHub’s discussion forums give Stabilitarians an opportunity to ask questions and offer solutions to specific scenarios. Join in the conversations with other Stabilitiarians and share your knowledge!