Theoretical Basis for the EDA Concept - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 151

Fig. 7.3 CI stage collection efficiency curve showing “ideal” step function case ( red dashed line )
and realistic but simplified case for establishing cut point size ( d 50 ); the square root of the ratio of
the sizes corresponding to the 84.1st and 15.9th percentiles for E stage, i is the stage geometric stan-
dard deviation by analogy with the properties of a unimodal, log-normal distribution for this
variable
of the CI at a fixed flow rate, Q . This simplification avoids the need to invoke mass-per-stage
data inversion measures that would require the shape of the response function for
each stage of the CI to be defined mathematically [ 20 ].
In Chap. 2 , the effect of making this assumption on the accuracy of measure-
ments using the ACI and NGI has been investigated further. The main conclusion
from this analysis is that any inaccuracy introduced is sufficiently small to be of no
consequence for full resolution CIs, but the simplification concerning stage collec-
tion efficiency may require consideration for AIM-based apparatuses, particularly
those derived from the ACI.
Figure 7.2 , which is taken from product w9j601 (CFC-suspension MDI) of the
blinded IPAC-RS database of APSDs from marketed OIPs, alludes to the fact that
the mass-weighted APSD from the CI experiment cannot be directly related to a
continuous APSD since there are two different y -axes in the plot. The latter is inher-
ently a frequency distribution that can only be estimated from a mass-weighted
cumulative APSD.
An estimate of the mass-weighted cumulative APSD is directly derived from the
mass-weighted CI results. This is illustrated in Fig. 7.4 . The continuous APSD is
then the derivative of the continuous cumulative APSD (Fig. 7.5 ).