development of a statistical tolerance-based fluid bed drying design space

the pharmaceutical industry has invested a substantial amount of resources in recent years to develop manufacturing systems that offer improved product quality while limiting costs. the us food and drug administration (fda) has encouraged the use of the guidelines put forth by the international conference on harmonization [ich-q8(r2)] that allow for operational flexibility within a validated design space to allow automated systems that incorporate real-time data management. use of these systems requires substantial process understanding that is typically gained through appropriately developed design of experiments. appropriate analyses of the data are required to optimize the understanding gained from experiments. furthermore, these analyses must be combined with existing product specifications to set justifiable operating limits for the process.

material and methods

granules containing gabapentin, and hydroxypropyl cellulose were prepared in a high-shear granulator and dried in a fluid bed processing system. the fluid bed dryer was outfitted with near infrared, pressure, temperature, and flow sensors, which were connected to a distributed control system that was used to exercise control of the system. the dried granules were blended with extragranular excipients and then compressed on a rotary tablet press. designed experimental variations to processing were utilized at each of the manufacturing steps to produce a design space. the outputs of the design (quality attributes) that were considered in this work were the quantity of gabapentin lactam formed during processing, tablet crushing strength, final blend flow properties, and median particle size of dried granules.

results

a tolerance-based design space was constructed for each of the quality attributes considered, and an overall design space was calculated based on the individual unit operations. the overall design space was a combination of process models and was calculated to include the uncertainty (variance) from each of the unit operations and measurements. a tolerance design space was calculated that reflects the probability that a given combination of processing parameters will produce critical quality attributes within specifications.