Application of Process Analytical Technology for Investigation of Fluid Bed Granulation and Active Coating during Process Development and Scale-up

Abstract

This thesis discusses the potential applications of process analytical techniques (PAT) and associated benefits of exploiting such in-line and at-line measurement tools during process development of solid dosage forms using case studies of two different development projects. The most critical unit operations of the manufacturing process affecting downstream processability and Critical Quality Attributes of the final drug product to a great extent were fluid bed granulation for the first project (Project A) and active coating of tablets for the second project (Project B).<br /> In Project A, fluid bed granulation processes were investigated by in-line particle size measurements using spatial filtering velocimetry and in-line moisture determination by near infrared spectroscopy (NIRS). Implementation of these in-line tools significantly enhanced process understanding during development. In-line particle sizing provided insight into the dynamics of particle growth during different phases of fluid bed granulation process: wetting and nucleation, agglomeration and finally a steady-state phase, when agglomeration and attrition are at par. Furthermore, different technologies for particle sizing were compared. The effect of varying process parameters on resulting granulate particle size was systematically investigated by means of a Design of Experiments (DoE) study, performed as several fractional factorial designs. The developed statistical model obtained from this study could accurately predict the median granulate size (D50) in dependence on the process parameters inlet air temperature and product temperature in the investigated ranges. Higher inlet air temperature has a positive correlation with particle size resulting in increased D50 values, while lower product temperatures yield larger granules. Besides, a NIRS calibration model applying a fibre optic probe and diffuse reflection measuring mode was developed for in-line moisture monitoring during the granulation process. The calibration is built upon Partial Least Squares regression of NIR spectra and reference values generated by loss on drying on a halogen dryer. It is possible to predict the water content in the calibrated range with a precision comparable to the one of the reference method. Finally, in-line moisture monitoring by NIRS turned out to effectively support the scale-up activities to commercial scale. The added value of application of PAT during process development in this project did not save development time, but increased the required understanding of the (inter-) relationships between process parameters and attributes of the produced granulates. Based on this knowledge, reasonable specifications for routine manufacturing can be selected and process windows or design spaces for single unit operations can be implemented. <br /> Project B involves an active coating process of tablet cores in a drum coater, where tablets are layered by a film-coating liquid containing the Active Pharmaceutical Ingredient (API). Critical process parameters affecting content uniformity of the API in the final drug product were assessed by means of a DoE study. It was found that content uniformity could be improved by reducing the spray rate, reducing the solid content of the coating suspension by adding water or by increasing the rotation speed of the drum coater. One of the objectives of this study was to perform a feasibility assessment regarding the development of a NIRS method for end-point determination of the coating process in real time. Currently, the end-point is estimated by indirect measurements. An appropriate at-line measurement procedure and a NIR calibration model were established for determination of the actual API content. The calibration is based on time-consuming HPLC reference values, which is the current standard release test. Two distinct models were developed: one focusing on measurements of individual tablets and one using a rotating sample cup as a measuring device, which allows the simultaneous measurement of a larger surface area of several tablets. Principal Component Analysis proves the specificity of both final models for the API. A complete validation approach including further validation parameters i.e. linearity, precision, accuracy as well as robustness was exemplarily realised for the NIR method considering individual tablets. In conclusion, near infrared spectroscopy is found to be an appropriate tool for end-point determination of the active coating process, but further optimisation of the manufacturing process itself might be necessary for reducing variability in content uniformity of the film-coated tablets. The benefit of applying NIRS as a PAT tool in active coating processes compared to conventional HPLC analysis is mainly related to saving of time because the bulk product can directly be processed further without delay between completion of a unit operation and sampling, analysis and release. <br /> In the last section, important regulatory requirements concerning the submission of PAT data and specifics for NIR data are summarised.