Chemistry Reference
In-Depth Information
Figure 11.7.
Example design space for feed pressure and outlet temperature for the telaprevir
spray-dried dispersion manufacturing process.
range is shown in Figure 11.7c, which shows that the design space is an irregular shape,
incorporating all possible combinations of feed pressure and outlet temperature that
result in both particle size and bulk density being acceptable.
When design space complexity precludes de
nition of a straightforward NOR,
alternative approaches are used to maintain simplicity during manufacturing. The SDD
spray drying model relates the impact of various adjustable process parameters on PS and
BD. The process parameter NORs (i.e., combinations of process parameters that will
result in SDD within the particle size and bulk density NOR range) vary depending on
other
fixed and variable inputs. Therefore, model calculations are made in real time, and a
continuously updated NOR is presented to the operator in a
“
dashboard
”
format. The
dashboard identi
es updated NOR and PAR (proven acceptable range) limits and
the location of current operation, and suggests what changes can be made to adjust-
able parameters to maintain both PS and BD within their NOR ranges.
The intended use of the spray drying design equations is to (i) establish the initial
spray drying conditions (set points) and (ii) estimate the adjustment in spray drying
parameters required to recenter the response following in-process particle size and bulk
density measurement.
SDD material attributes (particle size and bulk density) in combination with the
tablet compression process parameters would be expected to affect tablet CQA of
dissolution. For example, small PS SDD compressed to the same tablet hardness as large
PS SDD would be expected to provide faster dissolution profile. Therefore, DoEs linking
spray drying and compression processes have to be executed to study the interaction
between these steps. This is illustrated in Figure 11.8. The spray drying DoE produces
SDD with a range of particle size and bulk density, which are then used as inputs to a
downstream DoE on tablet compression.
A regression model was developed to predict tablet dissolution as a function of SDD
particle size and bulk density and average tablet hardness. A total of 241 tablet sublots
were used to produce the dissolution models. The project team used mixed stepwise
linear regression to determine the parameters affecting the mean cumulative release of
telaprevir API from the tablet after 15min, normalized by the 90min time point. This
early time point was selected as this part of the dissolution curve was considered to be
more sensitive to processing and SDD particle size and bulk density effects on
