(746b) Design Space Development for the Dropping Process of a Dripping Pill with Material Attributes Represented As the Scores from Principal Component Analyses of Spectra Data

Authors: 
Gong, X., The State Key Laboratory of Chemical Engineering
Qu, H., Zhejiang University

Design space development for the dropping process
of a dripping pill with material attributes represented as the scores from
principal component analyses of spectra data

Xingchu Gong, Jichen Shen, Haibin Qu*

 (Pharmaceutical Informatics Institute,
College of Pharmaceutical Sciences,

Zhejiang University, Hangzhou 310058, China)

Abstract

Yinxingye Diwan is a
dripping pill preparation treating angina pectoris and cerebral infarction. It
is made from Ginkgo biloba extract and PEG 4000. The pass rate and productivity
of dripping pills were considered as critical quality attributes (CQAs) of
dropping process. Plackett-Burman designed experiments were performed to
investigate the effects of materials and process parameters. The two materials
of Ginkgo biloba extracts and PEG 4000 have significant effects on the process
CQAs. The distance between dropping needle and condensate, and dropping speed
were critical process parameters (CPPs). Split-plot designed experiments then were
carried out to model the dropping process. Material properties were measured by
near infrared spectroscopy (NIR) for Ginkgo biloba extracts and PEG 4000. The
scores from principal component analyses (PCA) of NIR spectra
were used in model development. Models with and without interaction terms
between PCA scores and CPPs were compared. PCA scores with significant effects
on the process CQAs were considered as critical material attributes (CMAs).
After favorable model was selected, design space was developed based on the
probability to attain process CQA limits. The probability was calculated using
a Monte-Carlo simulation method. The design space then was verified using new
materials. The quality levels of Ginkgo biloba extracts and PEG 4000 were also
set according to PCA scores.