Present: Sylvia Gantt, Garth Boehm, Gerry Mergen, Jerry Planchard, Tom Garcia, Jim Prescott, Neeru Takiar, Al Nyhius, Joep Timmermans, John Dietrick, Mike Gavini
Discussion of the Stratified Sampling Proposal
Tom Garcia related discussions with the DPTC. We must be sure that statistical calculations are validated. Jerry Planchard thought we were in good shape since cross checks of both the computer programs and probability results gave consistent results. However, he is concerned that a full formal validation would be a major undertaking.
Tom also said that Ajaz wants to see more statistical calculations using skewed (non-Gausian) distributions in addition to the normal distributions used to date. This challenge does not have to occur prior to submitting the proposal to the DPTC and SC.
Action: Jerry Planchard thought this would be possible and is to perform the computer simulations.
Since PQRI is aiming to reduce regulatory burden, we need to separate the late stage proposal from this document, and to focus on meeting the requirements of the current draft of the ANDA Draft Guidance document with our proposal. We must also bear in mind that creating additional testing requirements over those already existing may not meet the direction of PQRI.
Decision: It was agreed to separate the late phase testing proposal from this draft document.
Finally, there is urgency to get the proposal posted as soon as possible for public comment. We must finalize the draft quickly. Rich Poska and Rag Uppoor will review the proposal on behalf of the DPTC.
Action: The proposal should be finalized in time (preferred date, Feb 2nd, last possible date, Feb 5th) to allow it to be presented to SC at their February 6th meeting.
The proposal was then discussed in sections.
No comments
Garth Boehm felt that the phrase “tablet cores or capsules” on page 2 should be broadened to encompass other solid dose forms such as prefilled sachets. Also the phase “content uniformity” in the first bullet point on page 2 should be changed, perhaps to “product homogeneity”.
No comments
Gerry Mergen felt that based on feedback he had received, 25 positional samples would be too many, especially for small batches. There was discussion of what constitutes a small batch, either time, size, or one container of blend. Jim Prescott thought time was not appropriate as it can vary with machine speed. Time between positional samples was discussed, for example one sample every 10 minutes. However, it was generally felt that these criteria were somewhat artificial. Jerry Planchard expressed concern that the statistical calculations are only valid if the positional samples are truly representative of the whole batch, and that the number of positional samples was really something that should be decided based on physical considerations of the process. Jim Prescott thought that based on his experience, systematic non-uniformities would be detected by 15 positional samples if evenly spread through the batch.
After further discussion it was decided that Jerry Planchard would make calculations for sampling plans as follows.
Stage 1 - 3, 4, and 5 units tested for 15, 20, and 25 positional samples
Stage 2 - total 7 or 10 for 15, 20, and 25 positional samples.
Action: Jerry agreed to make representative calculations, although probably not for every combination.
Al Nyhuis asked if this proposal applied to all products or whether it applied only to those less than 50% active or less than 50 mg (as per the ANDA draft guidance). It was felt that for validation, it applied to all products, but for the early routine manufacture, only to those under the “50/50” rule. There is a suggestion that ICH and USP may move from 50/50 to 25/25. Garth Boehm suggested that we say that early routine testing applies only to products that require chemical assay for USP Content Uniformity. [Subsequent to the conference, G Boehm realized that this would cause trouble for coated tablets and respectfully withdraws the suggestion.] It was agreed that the 50/50 intent be clearly stated.
Garth Boehm pointed out that with weight correction, the requirement that individuals be within 75% to 125% was no longer appropriate. There was discussion of whether we should weight correct, but on balance (with some dissent) it was felt to be appropriate.
Action: Jerry Planchard suggested footnoting the individuals to indicate that the “as is” assay of units must meet this criteria for individuals. This was felt to be appropriate and was accepted.
Several members felt that the requirement to perform this testing for 6 months or 10 batches was excessive. In the case of a “blockbuster” this could result in testing 100’s of batches. Skip lot testing was discussed but rejected because if a later batch failed, there would not be testing support for skipped batches. There was discussion of the intent of early routine testing. While some felt that 10 or perhaps 20 batches would be a good test of differing lots of raw material, others felt that some way of testing robustness of the process over time was important. The possibility of using a phase such as “significant body of evidence” was discussed, but acknowledged as being too vague and not providing guidance. [The group needs to balance the vagueness of this statement with the consequences of offering a specific number of batches, which may be inappropriate or not supported by either FDA or industry.] This point was acknowledged by all as an issue, but we need a proposal to address this.
Action: All - Unless further recommendations are forwarded, the number of batches will not be stated and the phrase “significant body of evidence will be used.”
This will be removed to a separate document.
There was agreement with Jim Prescott’s early e-mail that this should be removed together with the list. It was agreed that reference to high risk situations will be made in the body of the text with clear indication that these practices may require testing beyond that contained in this proposal.
Action: Tom Garcia to draft a statement to be added to the text.
Next Steps:
Forward all comments to Laura for incorporation into document.
Forward proposal to Raj/Rich (DPTC) and SC for their review and comment. Consider their feedback and incorporate pertinent comments into proposal.
Post on PQRI website for public review and comment.