FDA expectation for sampling in an aseptic filling process

[Eliud Kipchoge]

Greetings,

I am working on a vials filling process, performed after sterilizing filtration, as a preliminary step before lyophilizing the vials to obtain the final drug product.

In this filling process, certain vial weight measurements are taken before and after filling. These measurements constitute the in-process control (IPC), and with this data, process capability parameters are calculated.

I have doubts about what the FDA expectations are regarding sample size in the filling process. Firstly, we have a machine capable of weighing 100% of vials, but due to several line failures (and for optimization and time-saving purposes), it is preferred to weigh a much smaller quantity (around 5%) of the vials in the batch. On other manufacturing lines, the machine's speed simply doesn't allow for more than 5-10% of vials to be weighed before the drug holding-time is surpassed.

I would like to know if anyone has encountered similar situations and has information on the subject. I am particularly concerned about justifying the ability to perform 100% IPC and deciding not to do so (in order to save 'just' a few hours). I have not found anything regarding this in the FDA pharmacopoeia (nor in the European one, if it helps).

Thank you very much for your time. All the best.

 

[Bev D]

Are you using a multi head filler? Since testing lyophilized product is destruct, fill weight is a typical control (it also supports material accountability). High volume product can be sampled if the samples are from each fill nozzle and on a frequency that is done on a regular time basis and when startup shut-down and other process changes or adjustments are made. Of course the severity of an underfilled must be taken into account. You will need to have a reaction plan to clear previous material when An underfilled tube is found.

I’ve not worked with the FDA. I worked with the USDA and they had little real understanding of process control and industrial statistics. I had to work to get them to approve of sampling and control schemes. Someone else here might have more experience with the FDA here.

 

[Eliud Kipchoge]

Are you using a multi head filler? Since testing lyophilized product is destruct, fill weight is a typical control (it also supports material accountability). High volume product can be sampled if the samples are from each fill nozzle and on a frequency that is done on a regular time basis and when startup shut-down and other process changes or adjustments are made. Of course the severity of an underfilled must be taken into account. You will need to have a reaction plan to clear previous material when An underfilled tube is found.

I’ve not worked with the FDA. I worked with the USDA and they had little real understanding of process control and industrial statistics. I had to work to get them to approve of sampling and control schemes. Someone else here might have more experience with the FDA here.

Thank you for your answer. Yes, it is a multi head filler. I'm thinking that the minimum expectation would be as you say, sampling in startup, shut-down and after major adjustments in the machine, as well as a % sampling on a determined frequency. My doubt comes from defining the percentage of samples (5%, 10%...) which would be considered as acceptable, considering as you say the severity of the failure mode (underfilling vials).

 

[Bev D]

There really is no statistical justification for using a %sampling scheme. Nor is there any physics based justification.
You would need an “RQL” plan. I have a spreadsheet in the resources tab that can help you with that. Also recommend reading some of Donald Wheeler’s works on sampling: part 1 and part 2.

If an unfilled or partially filled tube poses a high severity risk then you may need to resort to 100% weighing and/or more robust process controls. Air bubbles and clogged nozzles are not detectable or predictable with continuous data as they are a categorical defects. They can appear suddenly without warning. Some monitoring of input parameters can help with this (changes in flow pressure etc.) and preventive controls…

 

[v9991]

Greetings,

I am working on a vials filling process, performed after sterilizing filtration, as a preliminary step before lyophilizing the vials to obtain the final drug product.

In this filling process, certain vial weight measurements are taken before and after filling. These measurements constitute the in-process control (IPC), and with this data, process capability parameters are calculated.

I have doubts about what the FDA expectations are regarding sample size in the filling process. Firstly, we have a machine capable of weighing 100% of vials, but due to several line failures (and for optimization and time-saving purposes), it is preferred to weigh a much smaller quantity (around 5%) of the vials in the batch. On other manufacturing lines, the machine's speed simply doesn't allow for more than 5-10% of vials to be weighed before the drug holding-time is surpassed.

I would like to know if anyone has encountered similar situations and has information on the subject. I am particularly concerned about justifying the ability to perform 100% IPC and deciding not to do so (in order to save 'just' a few hours). I have not found anything regarding this in the FDA pharmacopoeia (nor in the European one, if it helps).

Thank you very much for your time. All the best.

the emphasis is more towards, a) stratified sampling b) sterility assurance

Refer these documents for more clarity and other useful details..,

Lyophilization of Parenteral (7/93)​

Sterile Drug Products Produced by Aseptic Processing — Current Good Manufacturing Practice

Recommended Best Practices for Lyophilization Validation-2021 Part I: Process Design and Modeling​

Recommended Best Practices for Lyophilization Validation 2021 Part II: Process Qualification and Continued Process Verification​