PAT
SUPPLEMENT
IMPLEMENTATION OF MODELLING APPROACHES IN THE QbD FRAMEWORK:
EXAMPLES FROM THE NOVARTIS EXPERIENCE
Dr. Marianna Machin, Dr. Lorenz Liesum and Dr. Antonio Peinado Novartis Pharma AG
The fundamental concepts behind the FDA PAT initiative are driving the pharmaceutical industry to put greater emphasis on the scientific understanding of their manufacturing processes, thus focusing its efforts both on ensuring product quality compliance through end product testing, and on understanding the impact of the manufacturing conditions and process variability on the quality attributes. In this respect, multivariate data analysis (MVDA), used for statistical process control, can be very useful and effective to ensure that a process is under control and, consequently, that it meets the quality specifications. At the same time, MVDA is a valid tool to improve the understanding of the process, to increase its efficiency e.g. in terms of yield and throughput time and consequently, leads to reduce costs. As the setting for this paper, the MVDA principles and tools, benefits and challenges are discussed prior to the review of two examples of application of MVDA at Novartis. Specifically, pharmaceutical and biopharmaceutical processes are discussed.
The use of Multivariate Data Analysis (MVDA) in the framework of Quality by Design and Operational Excellence initiatives to gain increased process understanding and, ultimately, process control is an area of growing interest and under great expansion in the pharma industry. With the implementation of Process
www.europeanpharmaceuticalreview.com
Analytical Technology (PAT) and modern automation infrastructures, numerous process variables are on-line available describing physical and chemical properties of the process, the raw / intermediate materials and the final products. These vast amounts of data are registered with various sensors, either PAT (NIR, Raman and UV- VIS) or classical probes (pH, T). To extract relevant
information out of the primary data retrieved, multivariate statistical process control (MSPC) is used for efficient process control, data trending and early fault detection taking the dynamic and multidimensional nature of these processes into account. Furthermore, MVDA is gaining importance, supporting and enabling real time release by an efficient control of the variability of the process using qualitative MSPC or even applying predictive models for certain critical quality attributes. The main benefits expected to result
from the application of MVDA can be cat - egorised in the areas of process understanding and process control. During process development, MVDA contributes significantly in a structured way to evaluating and visualising data stemming from lab and pilot scale and therefore supports a better understanding and interpretation of the process data. In particular, the main benefits where MVDA can add value to process development are:
European Pharmaceutical Review 5 Volume 16 | Issue 6 | 2011
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17