Future Directions for the AIM and EDA Concepts - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 397

17. Can you comment on the pharmacopeial discussion?
[Dr. Marjolein Weda]: A full-resolution impactor is needed in the Ph. Eur. for
characterization purposes. For quality control, you could remove some stages.
Introducing changes to Ph. Eur. quality control methods is allowed and, for exam-
ple, already happens frequently in HPLC methods (using different columns).

18. If EDA provides better discriminating power, what is the expected sample size?
[Dr. Prasad Peri]: We fi rst need to see the data that [confi rms] it indeed provides
better discrimination. If this is demonstrated, and if it is validated, FDA would
support the use of EDA. The data we have seen so far looks promising, and
we'd like to move forward.
19. Would APSD be needed as a release test with the Quality by Design (QbD)?
[Dr. Prasad Peri]: If you can correlate APSD to some “in-process” measure-
ment, I personally would support that approach instead of doing end-product
testing. The more knowledge and process control, the better assurance that the
fi nal product will meet quality requirements.
20. On the statistical issue, and your comment that with two parameters there might
be a loss of information: Since these two metrics are capable of discriminating
and detecting changes while collecting fewer numbers on a routine basis—that is
the whole point. If we could miraculously fi nd a single metric that could control
APSD, then that would be better still. You need all possibly relevant information
in development, but once you have characterized the product and established con-
trols, you shouldn't need to accumulate as many numbers as possible.
[Dr. Prasad Peri]: The only two metrics in the EDA approach are ISM and
LPM / SPM . If you can show that these give better assurance, discrimination,
and detection that your product is changing over time, then we would consider
it favorably. Right now, we do not have the complete data. Papers are being
published, presentations are made at meetings, and this knowledge base should
continue to build up. We want to encourage better analytical methods and better
methodologies. We don't want to hamper progress and innovation.
21. Would FDA require demonstration of the same performance between EDA and
the current approaches or better performance?
[Dr. Prasad Peri]: It can be the same statistically but better in terms of labor,
environmental impact, etc.
22. What upcoming guidance documents should we expect to see soon?
[Dr. Marjolein Weda]: Currently, there are no concrete plans to update the
EMA guidance. If the AIM and EDA topic moves forward, however, EMA/
CHMP could decide to reopen the guidance or to provide clarifi cation and
additional information via a Q and A.
[Mr. Bill Doub and Dr Prasad Peri]: Work on the new guidance is ongoing.
We have another internal FDA teleconference coming up regarding the
guidance. But even the current draft MDI/DPI guidance from 1998 does allow
alternate approaches. So even if the guidance did not change, you could pro-
pose to use AIM and EDA in your application. We want a less prescriptive
guidance, but you don't have to wait for it. We might update the guidance with
alternative methods. Now the document is in the process of being updated.