Current Approaches to APSD Measurements of OIPs Based on Inertial Impaction - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 20

D m D m

Equal mass in
each group

mass transfer from
group 3 to 2

mass transfer from
group 3 to 4

1 2 3 4 1 2 3 4 1 2 3 4

Group number Group number Group number

Fig. 2.4 API mass transfer from one stage grouping to its neighbor results in an equal and oppo-
site change to both groupings

2.3 Capability of the Cascade Impaction Method

The multistage CI is fundamentally a particle size-fractionating instrument, making
use of differences in particle inertia in (ideally) a laminar flow field to result in
either collection on a solid substrate or impingement into a liquid. CIs are designed
to operate at a constant flow rate. This limitation therefore becomes an issue in con-
nection with DPI testing, in which it is necessary to approximate an idealized inspi-
ratory maneuver in order to aerosolize and disperse the aliquot of dry powder either
from a reservoir or from a punctured retaining capsule/blister. A compromise is
reached in the compendial DPI testing procedures, in that although the CI is oper-
ated from zero flow at start of sample, sufficient volume of air is drawn through the
system (typically 4 L), so that the finest suspended particles have time to pass
through the entire apparatus, enabling the CI to properly size fractionate the emitted
dose. This requirement places a limit on the lowest sample volume it is reasonable
to use, because below this volume there will be insufficient time for the aerosol to
pass entirely through the CI system to be fractionated. This minimum volume is
based on the magnitude of the internal dead volume of the CI together with the
preseparator (PS), if used, and induction port (IP). Recent work undertaken by the
European Pharmaceutical Aerosol Group (EPAG), in which this sample volume has
been intentionally reduced to be close to the dead volume of the system (ca. 2 L),
has confirmed that the APSD is significantly affected in the case of the NGI under
such circumstances [ 13 ]. In the NGI, operated at 60 L/min, FPF 0.94-4.46μm was shown
to be systematically reduced with decreasing sample volume, reflecting the fact that
the finer particles have had insufficient time to move through the CI to the stages at
which they would normally be collected. In contrast, the ACI appeared not to be as
sensitive as the NGI, with FPF 0.76-6.18μm remaining nearly independent of sample
volume. The underlying reason for this anomalous behavior is not obvious, but it is
believed to be likely related to flow maldistribution in the ACI. Given this situation,
it is good sampling practice to ensure that the sample volume exceeds the dead
volume by at least a factor of two.