Learnaboutgmp Community

Worst case determination in multiproduct API plant

I need suggestion for methodology to find worst case product in multi product API plant. In current situation three different solvent used as cleaning solvent for different products in same plant ie. methanol toluene and 5%sodium hydroxise in water. I would like to determine worst case from all products. What is your suggestion, pl. Additionally can i apply this worst case during dirty hold time study?

The ‘worst’ case target (in your risk analysis) is determined by the ADE (acceptable daily exposure), a measure of toxicity. Do not use USP <467> proscribed residual solvents as your recovery solvent!

There are a few “groupings” which are helpful when determining worst case for cleaning validation of a multi-product facility. And you might actually have a few products which are considered “worst case”.

I would 1st, make a list of all your products. Then I would make a list of all your cleaning practices - this might include the various cleaning agents. So you might call a cleaning practice “manual, sodium hydroxide” or “manual, CIP100”, or “Automated, CIP 100”, etc.

Next you will determine what the worst case product is for EACH of the cleaning practices. NOTE: Your cleaning practices have to be well documented, defined, and repeated (even for manual cleaning). To determine the worst case for each cleaning practice, you might look at solubility, you might do lab “coupon” studies. You might interview people to see what is the hardest to clean. You might do a review to see what product has the highest restults historically.

Also, at this point you might look for what is the hardest to clean, not what is hardest to get to the acceptable limit. For example, product A gets down to 100 ppm historically, but has an acceptance level of 100000 ppm. Product B gets down to 5 ppm historically, but has an acceptance level of 6 ppm. I might (please do your own critical analysis) actually call product A the “worst case” product because it is hard to get clean (but is likely benign because the limit is so high).

From here, you clean product A, and get it down to 5 ppm and validate that process. Thus product B will also get down even more, using the same cleaning method. Repeat this test a few times (historically 3, but that is not set in stone). Then I would do 1 run on all the other products to confirm your cleaning practices are effective for all products.

Overall: (1) Make sure your cleaning methods are well defined and repeatable. (2) Group cleaning methods. (3) Determine worst case product for each/all cleaning methods. (4) Perform cleaning validation on the worst case product for each method. (5) Then verify cleaning for all the other products.

Regarding dirty hold time - this would be great to do during a coupon study, and will help figure out which product is worst-case, before you start cleaning and evaluating full scale operation.