Table of Contents:
In the scope of this work the applicability of reflectance infrared spectroscopy combined with multivariate data analysis as a process analytical method for in-line and at-line process control of critical quality attributes (CQAs) during the coating process of laminates for oral thin films was evaluated. To obtain a homogeneous drug containing laminate after the coating and drying process with regard to the active pharmaceutical ingredient (API) content, it is important, that the laminate is as homogeneous as possible in terms of coat weight, moisture and residual solvents and API content to fulfill the requirements of the European Pharmacopoeia in the monograph "Uniformity of dosage units". With the current in-process control (IPC) the coat weight is determined gravimetrically at the beginning and at the end of the coating process of each master roll. It is in general possible to take more in-process control samples with the current gravimetric determination, however sampling means destruction of the material and requires interruption of the process. Since starting of the coating process after stopping bears a certain risk towards the quality, stopping the process should be avoided whenever possible. CQAs like API content or residual moisture content are controlled by the surrogate parameter coat weight of the laminate. The real CQA content is tested on the single oral thin film after fabrication. Even after drying of the water and/or solvent based liquid mass the laminate contains a significant amount of residual water and/or residual solvents which contributes to the coat weight. Therefore, the correct potency which corresponds to a coat weight of the drug containing laminate is not easily determined.
Objective of the present work was to develop a process analytical method for monitoring in-line and at-line CQAs during the coating and drying process of laminates for oral thin films independent of the current gravimetrically in-process control by using reflectance infrared spectrometry. There are studies about using near infrared spectroscopy in the field of the pharmaceutical application of oral thin films but no studies were reported using reflectance infrared spectroscopy for in-line monitoring of a CQA during the coating and drying process of an oral thin film. In two independent test parts two different sensors/spectrometers with different spectral ranges were used for in-line and at-line process control trials.
By using a reflectance infrared sensor from Honeywell (RIS 3-4810) in the first trial part a process analytical method was developed which enables, without destroying the laminate, the monitoring and determination of the residual moisture content during the coating and drying process of a three-layer laminate containing Fentanyl citrate as API independent of the currently performed gravimetrically in-process control and the coulometric Karl-Fischer titration. A quantitative model could be developed for the two-layer protective layer and the three-layer laminate. With continuous gathering of real-time data of the residual moisture content combined with the results from the gravimetrically in-process control of the coat weight it is now possible to calculate contemporary the dried coat weight of the laminate to allow real-time process adjustments to keep the coat weight within predefined limits and hence decrease loss of material and batch. In a feasibility study it could be shown that the Fentanyl citrate content based on reflectance infrared spectral data cannot be monitored in-line and in real-time. New trials should be conducted to clarify if another spectrometric method like Raman spectroscopy can quantify the API content during the coating process.
In another trial for the first time reflectance infrared spectroscopy was applied for in-line monitoring of the CQA nicotine content during the coating process for an oral thin film.
With a feasibility study under lab scale conditions the general applicability of reflectance infrared spectroscopy for the quantification of nicotine in a methacrylic acid-ethyl acrylate copolymer matrix was successfully demonstrated. Using real time data gained from a coating run on a production coating line different quantitative models were evaluated. With an external validation run and an independent laminate the prediction accuracy of the generated model with the best predictive parameters was evaluated. According the goals of the PAT framework a process analytical method was developed which enables real-time monitoring and control of a critical quality attribute during a manufacturing process. Until now there was no possibility to monitor and control the nicotine content during the whole coating and drying process apart from the currently used gravimetric in-process control. The nicotine content was up to now controlled by the surrogate parameter coat weight of the laminate. The real nicotine content was up to now tested on the single oral thin film after fabrication. With the implementation of the new process analytical method the continuous collection of data during the coating and drying process is possible which enables real-time process adjustments to keep the nicotine content within predefined limits and to fulfil the requirements of homogeneity.
In the second test trial part of the present work the applicability of reflectance infrared spectroscopy for at-line process controls of oral thin films during the coating process was evaluated. With a polarization Fourier transformation near-infrared spectrometer from Büchi (NIRFlex N-500) process analytical methods could be developed for oral thin films containing Dexamethason and Rizatriptan benzoate as API. Based on laminates with pre-defined concentration ranges of API and polymer manufactured under lab scale conditions quantitative models could be generated for nondestructive monitoring of the API independent of the gravimetrically coat weight in-process control. The second trial part showed that at-line process controls based on reflectance infrared spectroscopy are possible if in-line process controls are not feasible due to technical or constructional problems.
To summarize the results of the present work it could be shown that according the goals of the PAT framework process analytical methods based on reflectance infrared spectroscopy could be developed for in-line and at-line monitoring of CQAs during the coating process of oral thin films. Until now there was no possibility to monitor and control the API content or residual moisture content during the whole coating and drying process apart from the currently used gravimetric in-process control. The API content and the residual moisture content were up to now controlled by the surrogate parameter coat weight of the laminate. With the implementation of the new process analytical method the continuous collection of data during the coating and drying process is possible which enables real-time process adjustments to keep the CQA within predefined limits and an improvement of process reliability. To implement the real time release idea from the PAT initiative to evaluate and ensure the acceptable quality of a final product based on process data, the developed in-line and at-line process controls have to be validated according ICH guideline Q2.
In continuative trials it should be clarified if real-time date from the in-line process control can be used to change automatically the drying profile, the coating web speed or the gap size of the coating head to affect the coating and drying process if CQA data are monitored which are out of pre-defined ranges to obtain a homogeneous laminate after the coating and drying process with regard to the API content or residual moisture content.
The present work also showed, that reflectance infrared spectroscopy is a rapid in-line process analytical method once it has been validated but the requisite calibration and validation runs are time-consuming. It requires the production of a range of samples under production conditions that differ in concentration. Model-building also requires an optimisation step including the choice of a regression method and spectral pre-treatments strongly informed by the used data set. A final point to bear in mind is that calibration models include the errors of the reference method mainly due to sample preparation and handling.