Flowchart Methodology Utilizing Validated Process Models as the Foundation for a Quality-by-Design (QbD)-Based Approach for Pharmaceutical Unit Operations: Example Case Studies

A model-based process development methodology is described which uses a Quality by Design (QbD) approach to establish a manufacturing design space. A flowchart methodology is described that illustrates how to rationally step through process development, using fundamental and engineering models and experimental tools in order to define the relationship between process inputs (input variables and process parameters) and critical-to-quality attributes. A key to this methodology is defining the physical situation and establishing the key control volume that must be understood and recreated from batch-to-batch and upon transfer and scale-up. In addition, the process understanding and experimentally-validated models within this methodology form the basis of a PAT development strategy (including the process control strategy) upon commercialization. This QbD approach can be applied to a number of pharmaceutical unit operations. Three case studies are reviewed – evaporative processes, melt-spray congeal, and liquid mixing operations.

Spray drying to form solid dispersions will be used as a case study to illustrate the QbD process development methodology for evaporative processes. In this process, the control of drying rate and atomization are critical to product performance, stability and manufacturability. The keys to this development approach are to 1) characterize the physicochemical properties of the formulation, 2) define the thermodynamic operating space utilizing a mass and energy balance, and 3) characterize the key kinetic components of the process such atomization and drying rates using experiments and models. Similar logic can be applied to other evaporative processes such as tablet coating and fluid bed coating or granulation.

For a melt-spray-congeal (MSC) process, the control of microsphere size is imperative for attaining a specified and reproducible sustained-release profile. In the case of MSC microspheres created by atomizing a melt stream using a spinning disk and congealing them in a cool gas stream the process variables determining microsphere size are: 1) disk speed, 2) disk diameter, 3) melt feed rate and 4) the state of the melt stream. The state of the melt stream, namely its composition and temperature, determine its viscosity and surface tension and these in combination with disk speed, disk diameter and melt feed rate are the control variables of importance. By determining the combinations of process variables that reach the intended microsphere size one can establish process conditions that result in the desired product.

A liquid mixing process scale-up approach will be described that utilizes computational fluid dynamics and a universal mixer experimental tool. The keys to the approach are to 1) define the physical situation, 2) create models (e.g. CFD) of each appropriate reactor scale/geometry, 3) correlate small scale experiments to model outputs to define key scaling parameters and validate the models 4) utilize a validated correlation between experimental data and model outputs to define conditions for process transfer/scale-up.&'