Performance Characterization of EDA and Its Potential to Improve Decision Making in Product Batch Release - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 198

Table 8.6 Assumed limits used to construct operating characteristic curves comparing the
performance of the ratio metric LPM / SPM to “grouped stages”
Goodness-of-fit
criterion ( R 2 ) Assumed limits for MMAD (μm)

w9j601 w9j801 w9j901 w9jk01

0.9 2.03-3.04 1.14-1.83 2.25-3.00 2.11-2.97

0.8 1.81-3.24 1.02-1.98 2.09-3.15 1.94-3.14

0.7 1.62-3.40 0.93-2.08 1.96-3.27 1.81-3.27

0.6 1.47-3.54 0.85-2.17 1.86-3.36 1.71-3.38

w9k001 w9k201 w9k901 w9kw01

0.9 3.17-4.04 3.49-4.31 2.26-3.10 2.56-3.36

0.8 2.99-4.23 3.33-4.47 2.08-3.27 2.39-3.52

0.7 2.85-4.37 3.19-4.59 1.95-3.40 2.26-3.66

0.6 2.73-4.49 3.08-4.69 1.84-3.51 2.15-3.77

8.5.2.2 Overview of “Christopher-Dey” Strategy

Using the same stage groupings as those chosen by Tougas (Table 8.1 ), Christopher-
Dey established, for the purpose of this assessment, a range considered to represent
acceptable values of MMAD and determined the relationship between each of the
grouped stages and values of MMAD over that range. From this, acceptance limits
for each of the grouped stages were determined, which were confirmed to be
consistent with typical FDA limits. Applying the same technique, the relationship
between the metric LPM / SPM ratio and MMAD was established, along with corre-
sponding acceptance limits for LPM / SPM , thus providing a basis for an “apples-to-
apples” comparison of the grouped-stage and EDA approaches. It was necessary to
expand the range of CI results while maintaining characteristics of the actual CI data
in order to obtain robust estimates of performance characteristics at and beyond the
range of acceptable values for MMAD . To accomplish this objective, the original
data were shifted toward either higher or lower values of MMAD in a way that is
consistent with physical mechanisms that govern and constrain APSD shifts in OIPs.

8.5.2.3 Results from “Tougas” Strategy

The generation of OCCs by this approach started with defining a target APSD for
the specific product under consideration. Figure 8.21 illustrates the cascade impac-
tor results from the IPAC-RS database for one of the products studied, a CFC sus-
pension MDI identified as w9j601 . This figure is a superimposition of all individual
CI determinations ( n = 43) and presents individual stage data as a function of
aerodynamic particle size. Note that the process of defining a target APSD in a real
OIP development situation would depend on gathering sufficient representative
results that adequately characterize the product in question. In this case, all avail-
able data from the IPAC-RS database were utilized. The target APSD was selected
from these data, by first averaging the individual mass API-per-stage data on a
stage-by-stage basis, then converting to cumulative weighting, and finally fitting a