Classification of PVC for pharmaceutical blister packaging using pattern recognition techniques.

摘要

The objective of this study is to establish methods for the determination of the interchangeability of pharmaceutical PVC (polyvinyl chloride) packaging materials using computerized pattern recognition techniques. PVC is a primary material used in blister packaging of pharmaceutical products with about 90 percent of blister packaging being made of PVC alone. The popularity is due to its high chemical stability, low cost and aesthetic appearance. However, PVC analysis and characterization are difficult because a variety of additives are used in its manufacturing. When a pharmaceutical company has to use a new PVC material to replace a FDA approved PVC packaging material, a new, costly and time-consuming stability study is required. Currently, there are no established USP tests and specifications for PVC interchangeability. In the dissertation, a new method of analytical testing coupled with data analysis algorithms is presented. The collected data from each test are analyzed using computerized pattern recognition software. Data analysis and classification methods are used to differentiate and classify the PVC materials. The combination of analytical testing and statistical data analysis clearly separates and correctly classifies the polymer samples according to their chemical identities and even the different lots. The combination described in this dissertation provides a promising method to justify PVC interchangeability. USP regulations and current researches on packaging material interchangeability are reviewed. Also, PVC background, PVC general characterization methods, interchangeability study methodology, experimental procedures and results are presented and discussed.