(M1430-06-34) Application of a Population Balance Equation Based Wet Bead Milling Model to Process Design and Scale-Up of a Long Acting Injectable Formulation

Purpose: The practical application of a population balance equation (PBE) based wet bead milling model for process design and scale-up of a long acting injectable product will be presented. PBE is widely used in many scientific disciplines. The in-house developed PBE milling model was capable of simulating particle size distribution evolution with time in a mechanistic manner. This was very useful for process design and scale-up as particle size was the primary quality attribute for milling and was the source of change for additional critical physical properties.

Methods: Like many models, some existing data are needed to build the model. Models with the same predictive capability, but requiring a smaller number of experimental data, are preferred from a cost consideration, which is especially true for larger scale commercial batches. The minimum number of milling batches needed for building this mechanistically derived model will be discussed. The purpose of the work was to develop a model for simulating batches on a larger mill (4 L) for scale-up, but the equipment for making a larger scale batch was not ready at the time of the simulations. Batches made in a smaller mill (0.15 L) were used to mimic the larger scale mills for validating the model technically. As anticipated, adjusting a model parameter was needed for achieving good simulation for the smaller mill (0.15 L) batches. This result indicated that the same parameter adjustment should be needed for the prediction of larger mill batches, which was supported later by results from actual batches.

Results: In nano-scale milling, particle size change closer to the milling end time can be very small. Particle size sampling and measurement typically have some intrinsic errors, which can make the determination of the end point specification difficult and expensive experimentally. In this case, PBE model simulations can reveal the end milling time change more precisely. Optimization of the total milling time was desired during the scale-up process design. Various combinations of discrete pass and recirculation modes were proposed. Testing all these combinations experimentally would have taken a significant amount of time and effort. Simulations using the PBE based model allowed for rapid data availability, which lead to faster design decisions. The simulations also revealed an outcome that was against intuition but true if thought through carefully. Finally, the model was used to fine-tune process parameters, such as water addition time and flow rate change time to make the total milling process as optimized for cycle time as possible. An example of using the model to interpret an unexpected batch result in commercial scale will also be discussed.

Conclusion: This model application has been completed successfully. Modeling results have been fully obtained and will be described in detail in the poster.