Here's an interesting case. Intermediate X is required for making drug substance or API.

We can make intermediate X by two different processes. Process 1 has less yield. It was superseded with process 2 which improved yield. Validation data exists for both the processes.

Question is, can we mix intermediate X derived from two different processes to make final API? Any response with supporting guidance documents references is highly appreciated.

If yes, what will be the additional studies required for submission to regulatory bodies?

Hi Reddys

Within the Social Insterest groups in Elsmar there is one which groups folks with background in Drugs (The "PharmaVigilant" group). I would suggest you to send these folks a direct message (even individualy) to see if they can get you the answer or guidance you need.

The idea you propose sounds really insteresting so I am also curious to hear what the answers could be. Please keep me posted. In my opinion, if the intermidiate X from both alternate processes is equivalent and meets its defined USP criteria or other as applicable, I will consider the approach valid, however as you said is better to have sound reference documentation to take the next step so it can be documented in your QMS. Probably my only other concern would be the additional controls you may need to maintain to demonstrate traceability, as well as the possible need to validate the mix.

I wish I were of more help, however for the time being I can only point you to bring your question up to the team members of the Pharmavigilan group.

Take care

Alberto

Greetings!
Alberto Rios

Hi!



It is not the GMP to mix the output from the two different processes to make the API.

Moreover there is no guideline reference for mixing the intermediate with two different processes to make the API, as guideline discusses only about blending of API.
As per your statement, Process 1 is superseded with process 2. Hope this would be happened through change control system (Have you evaluate the changes between the two processes, e.g solvent changes, changes in critical process parameters or steps etc.,). If that is the case means, you cannot mix the process 1 with process 2, since as per the current situation process 1 is obsoleted.
Suppose if the validation of both process 1 and process 2 is alive you can manage as follows:


Complete train till the final API, with process 1 as input should be validated.
Complete train till the final API, with process 2 as input should be validated.
Its equivalency in terms of impurity profile has to be proved.
It is better to have the stability data in hand with these two different processes.

If the equivalency in terms of both impurity profile and stability is proved then you can mix the Process 1 output with Process 2 output to make an API.
It is better to validate the process again with the mixed input.


Regards / Raghu