# Multivariate methods in tablet formulation

This dissertation explains the use of multivariate methods in a novel strategy to the formulation of tablets for direct compression. It starts with a short historic overview, accompanied by a simple presentation of important areas of tablet formulation and multivariate data analysis. The majority of the dissertation is involved with the novel approach, where excipients were characterised with regards to multiple physical or (in many instances) spectral variables. By making use of Principal Component Analysis (PCA) the descriptive variables are described into a handful of latent variables, typically named scores or principal properties (PP’s). By doing this the volume of descriptive variables is significantly lowered and the excipients are explained by orthogonal continuous variables. Which indicate that the PP’s may be employed as ordinary variables in a statistical experimental design. The combination of latent variables and experimental design is called multivariate design or experimental design in PP’s. Making use of multivariate design numerous excipients could be incorporated into screening experiments with fairly few experiments.The result of experiments made to measure the outcomes of variations in excipient composition of formulations for direct compression is, obviously, tablets with several properties. When these properties, e.g. disintegration time and tensile strength, are actually identified with standardised tests, quantitative relationships between descriptive variables and tablet properties may be discovered making use of Partial Least Squares Projections to Latent Structures (PLS) analysis….

## Contents: Multivariate methods in tablet formulation

3 Introduction
3.1. Tablet Formulation in a Historical Perspective
3.1.1. Tablet Formulation: Basic considerations
3.2. Tablet Formulation and Multivariate Methods
3.3. Scope of the Thesis
4 4. . Multivariate Methods
4.1. Principal Component Analysis
4.1.1. MSC and SNV
4.1.2. Missing Data
4.2. Multivariate Characterisation
4.2.1. Physical Properties
4.2.2. FT-IR and NIR
4.3. Statistical Experimental Design
4.3.1. Mixture Design
4.4. Multivariate Design
4.5. PLS
4.6. Validity of Models
5 5. . Multivariate Methods Applied d to Table Formulation
5.1. Screening Experiments
5.1.1. Excipient Selection Based on Physical Properties
5.1.2. Excipient Selection Based on Spectroscopic Properties
5.2. Model Interpretation
5.3.1. Strategies for Excipient Selection
5.3.2. Evaluation of Results
5.4. Robustness Testing
6 6. . Concluding Remarks…

Source: Umea University