Construction of a quality index for granules produced by fluidized bed technology and application of the correspondence analysis as a discriminant procedure.

The production of granules by wet granulation in a fluidized bed was assessed after the construction of a quality index based on a file of attributes (relevant factors). These attributes are combined by a methodology relying on Correspondence Analysis, as a discriminant procedure, using two extreme simulated active vectors representing, respectively, the best and the worst cases for the granules quality output ("bad" and "good" pole). From those, a single continuous synthetic variable - the quality index - can be produced referring to a more significant set of samples. As an application of the methodology, the work compares the quality of granules produced at a laboratory scale and a pilot scale. The factors contribution to the bad or good pole allowed the identification of the most relevant factors that affect the quality of the granules. The factors studied, according to a center of gravity design, included formulation (solubility of a drug, different grades of polyvinylpyrrolidone, the polarity of the granulation solution) and processing factors (the rate of administration of the granulation solution, the atomizing air pressure and the fluidizing air rate). Granules were evaluated for production yield, drug content, size, densities (true, bulk and tapped), friability, flowability and compressibility. The study has emphasized the differences between the laboratory and pilot scales and the relative importance of each factor for the quality of the granules produced.