Dear All,
I am trying to prepare a protocol for the validation of packaging. I am in the field of process validation of generic tablet & capsule products. I would also appreciate your opinion on approaches you follow:
How do you validate the packaging stage (primary and secondary packaging)? Does validation have to be for each tooling and each packing material ? what does process validation have to do for the packing stage?
Thank you in advance,
Carlos
Packaging is part of process validation, particularly since related substances (reaction product between the API and the package) can decrease the API potency below the minimum specification and thus shorten the shelf-life.
Thus, I would test the API potency, moisture, microbes (if applicable), and related substances. The acceptance criteria would be your specifications.
Thank you very much for your response.
In this case, if the stability is a must in order to validate the packaging process?
Do we need to validate the packaging process for a set of critical parameters?
For example, for a specific blistering machine, specific tooling, specific packaging material etc.
Yes. These are the parameters as listed in the process validation protocol section on packaging. Of course, upon manufacturing the process validation lot(s) are put on stability.
ok, thank you for your help
Yes you will need to validate the various tooling.
Packaging can impact product quality/identity. Changes is primary packaging can impact stability and quality. Changes in secondary packaging can impact labeling/identity.
The trick to packaging validation is to determine what are critical parameters and which are not critical parameters, since there is so much that can change in a packaging line.
For example, I would not validate the settings for infeed - these can change but often don’t impact product quality- but I would document them to assist with set-up. I would validate temperature and speed and dwell (for blister packaging). I would also validate print settings (heat, dwell, alignment, eetc.)
Dear Jared, Thank you very much for your reply.
As process validation, and based on Annex 15, are we supposed to challenge the upper and lower limit of the temperature to be used OR should the limits of the machine checked during qualification?
I am very confused about at what stage we (PV) should be involved?
Should each operating range be challenged for each blistering process during process validation? Or process validation should be carried out during routine production? thanks again.
My opinion on this is fairly untapped in the industry. Also, Validation should help guide engineering studies to come up with documented and defended settings.
I think packaging engineering should follow a three step approach.
1 - Using a Design of Experiment (DoE), come up with a performance profile.
2 - Pick a result that you want (seal strength, speed, etc.).
3 - Run qualification at a few settings, and confirm that the results align with your performance profile (step 1).
Your DoE supports the limits of operation. Your qualification confirms the DoE. You have also tested “ranges” more than just high/low of each setting, you have also looked for attribute interaction, curvature, etc using WAY more robust statistical analysis than traditional. Overall you have shown that you understand your process and that it is in control, which is what validation is intended to do.
Alternatively. You could do the old school approach (I don’t love this method, but it also works).
1 - List all your parameters
2 - Lock down some parameters which WILL NOT change - so these are fixed at one set point, and don’t need to be challenged during qualification.
3 - Determine (thought engineering studies) appropriate ranges for all the remaining settings (not fixed in step 2)
4 - Determine which settings are “Critical” and will be qualified and which settings are “suggested” and can be changed based on equipment efficiency.
4a - for example rail position might be an efficiency setting and might be adjusted after some wear.
4b - for example temperature and speed (dwell time) might be a critical setting, and which needs to be controlled tightly.
NOTE: Packaging qualification can change due to differences in stock material. Somehow make allowances for that. A qualified setting for one lot of materials might not be optimal for another lot of materials.
Overall, your qualification DOES need to test the ranges of operation. But which ranges? That is up to you. Temperature and speed are some of the settings which require qualification of ranges. You asked “should EACH range be tested?” The answer is, how is that possible? There are so many ranges, you could not possible test all the combinations of each range - thus using statistical DoE is the better method to get the most information possible.
Do you guys have a statistician who can help set this up?
Dear Jared,
Thank you for your recommendations. Highly appreciated.
From what you have written, I understand that process validation should give directions to engineering in terms of whats critical for the product stability or not and for those critical process parameters, engineering should establish a range based on DoE. Right?
I dont have much experience with DoE. Would you be kind enough to suggest some literature on this except for the FDA guidelines (ICHQ8 etc).
And to answer your question, no we dont have a statistician.
thanks a lot
Validation is primarily concerned with stability/quality (you are correct). But…FDA also believes that a process that is slow, with lots of waste or is inefficient means that the process is not in control or not well understood. So even though efficiency might not affect “Quality” it can be an indicator of a process which is not well controlled - so it is kind of an indicator of problems.
Regarding DoE and statistics. I leaned heavily upon a statistician for help each time I did this. There is software that can do this (set up the DoE, and also interpret the results to help optimize the process).
I can recommend a good individual who did this well (he really “got” the idea of integrating statistics, with engineering, and validation). I would prefer to send his contact via private message, rather than on a discussion board. I could contact him and see if he has some recommendations.
Thank you very much for your input. Highly appreciated
We will not need a statistician at this point.
Good luck!