(233g) Modelling Approaches to Spray Drying Process Development, Scale-up and Troubleshooting

Modelling approaches to
spray drying process development, scale-up and troubleshooting

P.C. Valente, Í.
Duarte, F.Neves & M.N. Temtem

¹ R&D
Drug Product Development, Hovione Farmaciência S.A, Sete Casas, 2674-506
Loures, Portugal;

Initiatives such as
Quality by Design in the pharmaceutical industry have greatly increased the
emphasis on physical and mathematical modelling of the various processes and
the focus on science-based development [1]. The shift from the more traditional
approach of the Quality by Testing is driven by the rational that better
understanding of the underlying physical mechanisms leading up to the final dosage
form leads to improved control of the quality of the final product and to
accelerated product and process development. Spray drying of drug product
intermediates is one of the pharmaceutical processes being shifted from Quality
by Testing to Quality by Design and it is also illustrative of the various
models that need to be developed in order to reach that goal. For example, when
a spray drying process is employed to obtain an amorphous solid dispersion, an
adequate understanding of the miscibility/phase separation between polymer and
API can accelerate the screening of polymer candidates and optimized process
conditions to ensure a homogeneous molecular dispersion between polymer and
API. However, due to the complexity of the underlying physical mechanisms
together with the fact that multiple physical phenomena are simultaneously at
play, the task of creating an adequate model can be quite daunting. To overcome
these difficulties it is crucial to have a pragmatic approach to modelling and
to adequately formulate the problem statement so that no more complexity than
absolutely necessary is fed into the model.

In this work we discuss
five case studies where different levels of modelling are used, from
statistical to first principles, to address product and process development
needs as well as troubleshooting. In the first and second case-studies two
models based on first principles are applied to the formulation of amorphous
solid dispersions and to forecast the impact of particle design on the dissolution
behavior. In the third and fourth case-studies two different levels of
modelling, statistical and mechanistic, respectively, are applied to predict
particle size of a spray dried powder based on the nozzle geometry, atomization
and process parameters, showing that there is an unavoidable compromise between
model accuracy and model generality. This case-study is also used to emphasize that
the choice of model closely depends on the problem statement. A fifth case-study
of troubleshooting a pharmaceutical process using computer fluid dynamics is also
presented (Fig. 1).

Finally, the
adequateness of the model complexity for tackling a given problem is discussed
and two illustrative examples of overkilling or oversimplifying a model are
given.

Figure 1: Troubleshooting example of a spray drying
process using computer fluid dynamics. (Left) Gas velocity contours, (center)
droplet spatial distribution within the drying chamber and (right) gas
streamlines in the gas disperser and drying chamber of a spray drier.

References

[1] Lawrence X. Y. “Pharmaceutical
Quality by Design: Product and Process Development, Understanding and Control”,
Pharm. Res., 25 (4) 2008.