(492f) The Use of T-PLS to Investigate the IMPACT of Raw Material Properties on Granule Quality Attributes Obtained after Continuous Twin-SCREW WET Granulation

Since 15 years, the interest for continuous manufacturing has steadily increased in the pharmaceutical industry. The advantages offered in terms of process efficiency, costs reduction, flexibility and environmental footprint reduction have been explained in multiple articles and reviews. Among the different continuous drug product manufacturing processes currently available, twin-screw granulation (TSG) is the most evaluated continuous wet granulation process. However, a predictive platform linking the influence of raw material properties, the ratio’s in which the raw materials are mixed, and the applied process settings with the granule quality attributes is not yet available.

PURPOSE

In this study, T-shaped partial least squares (T-PLS) was used to study the combined influence of raw materials properties, formulation composition and process settings on granule quality attributes.

MATERIALS AND METHODS

A set of 25 powders consisting of 13 active product ingredients (APIs), 9 fillers and 3 binders for immediate release was characterized using 15 different techniques covering dry powder characteristics (particle size, compressibility, flow properties, ...) and wet powder characteristics (solubility, contact angle, dissolution rate, ...). Principal Component Analysis (PCA) was then performed on the API and filler raw material characteristics subsets to elucidate correlations between the powders and the raw material properties. Based on this analysis, formulation blends were selected, ensuring that a maximal area of the material variability space was covered. Each formulation consisted of an API, a filler and a binder. Herewith, three different blend ratios were used for each formulation, corresponding with an API content of 10, 40 or 70%. As the binders were intended for immediate release, their content was varied from 2 to 5%. Each formulation blend was granulated using the twin-screw wet granulation unit of the ConsiGmaâ„¢ 25 system (GEA Pharma systems, Wommelgem, Belgium). A 2-level full factorial Design of Experiments (DoE) was performed on each formulation blend to investigate the influence of the granulation process parameters mass flow rate (MFR) and liquid to solid ratio (L/S ratio) on the resulting granule quality attributes. The MFR was varied from 10 to 20 kg/h, while the applicable L/S ratio ranges were experimentally determined for each formulation blend, because this depended on the properties of the selected raw materials. Granule size and shape (PartAn^3D, Microtrac, Montgomeryville, United States), friability (Friabilator, PTFE Pharma Test, Hainburg, Germany) and flow properties (GranuDrum, GranuTools, Awans, Belgium) were measured for the obtained granules after each granulation experiment.

RESULTS AND DISCUSSION

The T-PLS model was able to indicate which raw material properties affected the most the granulation process. It was seen that dissolution rate, solubility, porosity, specific surface area and water binding capacity had the highest impact on the variability in granule quality attributes. Furthermore, it was also possible to highlight the most influencing raw materials for each granule quality attribute. For example, the granule size fraction that remained ungranulated (i.e. granule size fraction <150 µm) was correlated with water binding capacity, while it was anti-correlated with solubility, dissolution rate and compressibility.
The T-PLS model indicated as well that L/S ratio affected more the granulation process than MFR. But, in general, the variation in granule quality was more influenced by raw material properties than by process parameters.

CONCLUSION

Besides an enhanced process understanding of the impact of raw material properties on granule quality attributes, the developed T-PLS model has also the potential to predict a reasonable starting point for formulation and process parameters for new API’s. This approach can reduce the amount of API and the time needed to find optimal process parameters during product development.