Real-time prediction of dissolution profiles of coated oral dosage forms

Abstract

Optical coherence tomography (OCT) has emerged as an in-line monitoring technique for pharmaceutical coating processes based on a representative number of samples. In this study, an approach was developed to correlate the coating thickness measured in-line via OCT with the resultant tablet dissolution profile. This strategy enables prediction of the dissolution profile of coated oral dosage forms for each individual state of the coating process in real-time. Correlation models were developed for a tablet pan coating process and for a pellet fluid-bed coating process. The feasibility of the correlation models was tested using different process parameters and types of coating formulations. This work demonstrated that using the OCT data to predict dissolution could possibly form a unique way of assuring drug product quality and establishing a control strategy within the real-time release testing (RTRT) concept.

Introduction

For tablets and pellets with functional coatings, the coating quality attributes have a major impact on the dissolution profiles (Seo et al., 2020). Both thickness and quality of a coating layer are defined in the coating process (Suzzi et al., 2010). However, the conventional quality control is limited to manual measurements of a limited number of samples during the process and to in-vitro dissolution testing of the final drug product. Those conventional methods are labour intensive and time consuming. Therefore, the development of fast and reliable methods to predict the active pharmaceutical ingredient (API) release during the coating process has been generating much interest.

The internationally harmonized ICH Q13 Guideline (ICH, 2023) on continuous manufacturing highlights the role of process analytical technology (PAT) and process monitoring and control in continuous pharmaceutical manufacturing. PAT tools have demonstrated impact in continuous manufacturing where rapid measurements directly on the products being made in the process and utilizing these data for real-time process and quality control. Another major benefit of PAT is the possibility of real-time release testing (RTRT) for drug products. Concepts for RTRT in tablet manufacturing have been developed by monitoring all critical quality attributes (CQA) at all relevant stages of a processing line (Markl et al., 2020).

For solid dosage manufacturing, a large portfolio of established PAT tools are available (Fonteyne et al., 2015, Laske et al., 2017). However, the coating process, which defines the drug release profile for functional coatings, is commonly monitored via in-process controls (IPC). Typically, the weight of a small number of tablets or diameter gain of particles/pellets/beads drawn from the process is manually measured, mainly for the purposes of end-point determination at the final process stage. In addition to the established methods, some new technologies have been tested. The coating thickness was measured via acoustic microscopy (Bikiaris et al., 2012), air-coupled acoustics (Akseli and Cetinkaya, 2008) and broadband acoustic resonance dissolution spectroscopy (BARDS) (Alfarsi et al., 2018). In recent years, substantial progress has been made in the development and implementation of more accurate and representative methods of automated monitoring of the coating thickness. These include near infrared spectroscopy (NIRS) (Möltgen et al., 2013, Hattori et al., 2018), Raman spectroscopy (Kim and Woo, 2018, Radtke and Kleinebudde, 2020), terahertz pulsed imaging (TPI) (Alves-Lima et al., 2020) and optical coherence tomography (OCT) (Lin et al., 2018). Especially, TPI and OCT have shown promising results. In addition to offering high resolution, TPI and OCT can directly measure the coating thickness of single dosage units (Lin et al., 2017). OCT is suitable for real-time acquisition of coating properties during pan coating (Sacher et al., 2019) and fluid-bed coating (Markl et al., 2015) processes because of the fast acquisition rate with many high-resolution cross-sectional images acquired per second (up to 100). With real-time coating thickness data acquired via OCT, how to utilize those data to develop a control strategy via an integrated quantitative modelling approach has been a legitimate task for the scientific community.

Conventional dissolution testing has repetitive steps and overall is a time-consuming analytical procedure. Efforts have been made to predict dissolution results using the available process and material information and to include such inputs in an RTRT concept (Zaborenko et al., 2019). For example, approaches to predict dissolution profiles based on NIR spectra via multivariate regression models were presented for sustained-release tablets manufactured on a continuous direct compaction line (Pawar et al., 2016) and for immediate release tablets produced in the batch mode (Zhao et al., 2019). NIR data can account for sources of variability such as composition, blender speed and compaction force. Recent studies have used machine-learning approaches to predict the dissolution profile of sustained-release tablets using NIR data, the compression force and the particle size distribution of the main excipient (Galata et al., 2021). Further, the porosity measured via TPI has been shown to be a suitable parameter to predict disintegration and dissolution time for immediate release tablets (Bawuah et al., 2021).

Predicting the dissolution profile based on the coating layer thickness of tablets has been carried out via most of the available at-line and off-line technologies. Linear correlations between the coating thickness predicted via NIRS and the drug dissolution rate were derived (Wu et al., 2015, Ariyasu et al., 2017). Raman spectral data were correlated with the mean dissolution time (MDT) and with the coating thickness measured off-line via TPI (Müller et al., 2012). The MDT of sustained release tablets was also correlated with terahertz waveforms via multivariate linear regression (Ho et al., 2008). In addition, Raman chemical imaging was used to predict the dissolution profile of uncoated sustained-release tablets containing a matrix polymer for various contents and particle size distributions of the polymer (Galata et al., 2022).

For fluid-bed coating of pellets in-line NIRS was applied to predict the dissolution performance. NIR spectra combined with the process parameters (PPs) were correlated with 1- and 2-hour dissolution data via a PLS model for sustained release coated pellets (Ibrahim et al., 2019). A PLS correlation between the NIR spectra and the parameters of a kinetic release profile model was developed (Pomerantsev et al., 2011).
With regard to coated tablets, most efforts to predict the dissolution behaviour have involved off-line or at-line measurements. To date, no direct coating thickness measurements have been correlated with dissolution data. In this study, for the first time, a model for predicting the dissolution profile based on OCT real-time coating thickness data was developed.