Multivariate modelling to study the effect of the manufacturing process on the complete tablet dissolution profile

Area of Science:

• Pharmaceutical Sciences
• Process Analytical Technology (PAT)
• Drug Delivery Systems

Background:

• Dissolution is a critical quality attribute for oral solid dosage forms, directly impacting drug absorption and efficacy.
• Traditional methods for dissolution analysis often focus on single responses, limiting comprehensive process understanding.
• Variability in granule properties and compression can significantly influence drug release profiles.

Purpose of the Study:

• To investigate the influence of granule and compression variability on the entire dissolution profile of oral solid dosage forms.
• To apply an innovative multivariate tool, bi-directional projections to orthogonal structures (O2PLS), for enhanced process understanding.
• To compare the O2PLS method with conventional approaches in quantifying dissolution.

Main Methods:

• Utilized a design of experiments approach to introduce variability in granule and compression parameters.
• Applied bi-directional projections to orthogonal structures (O2PLS), a multivariate statistical method, to analyze tablet manufacturing data.
• Examined the relationship between process variability and different phases of the dissolution profile.

Main Results:

• O2PLS analysis demonstrated that the disintegration phase (10-15 min) of dissolution is influenced by granule attributes and tablet hardness.
• The later phase of dissolution (15-30 min) was found to be solely controlled by granule attributes.
• The bidirectional nature of O2PLS enhances exploratory analysis and process understanding, though it may reduce predictive accuracy.

Conclusions:

• O2PLS provides a more holistic understanding of the dissolution process compared to conventional single-response methods.
• This multivariate approach is particularly valuable for exploratory studies in early-phase drug development, requiring no prior knowledge of dissolution mechanisms.
• The findings can guide the selection of critical attributes and parameters for developing predictive models and defining process design spaces.