Multivariate approaches for assessing dissolution behaviour of spironolactone. Dissolution-driven determination of polymorphic purity limits

Spironolactone (SPN) is an aldosterone antagonist effective for preventing and controlling cardiovascular diseases. Form I and Form II exhibit different behaviour in their dissolution properties, which can compromise the quality of the pharmaceutical formulation. Hence, in the present work, we propose the combination of two multivariate approaches to estimate the polymorphic purity of SPN in commercial tablets and predict the dissolution behaviour of these formulations considering their polymorphic composition. Multiple linear regression (MLR) approach was applied to model the dissolution behaviour of formulations depending on their polymorphic composition, showing excellent agreement among actual vs. predicted dissolution curves. This model was applied to establish a design space were the risks of dissolution failure due to the polymorphic purity was avoided. Next, to evaluate the purity of Form II of SPN in formulated products, two quantitative partial least squares (PLS) models were developed based on near (NIR) and mid (MIR) infrared spectroscopies. Training (n = 14) and validation (n = 8) sets were prepared by mixing both polymorphs and the excipient matrix to simulate commercial tablets. Methodologies showed recoveries non-statistical different from 100 %, but MIR-PLS model showed a higher dispersion (100.1 +/- 1.82 % and 99.5 +/- 6.94 % for NIR and MIR, respectively). Similar values of Form II content (1.02 +/- 0.01 w/w with NIR and 1.03 +/- 0.03 w/w with MIR) were obtained during the analysis of commercial and spiked SPN tablets. Finally, NIR-PLS was coupled to MLR model for the prediction of dissolution in entire tablets based on polymorphic composition to act as Quality by Design (QbD) tool. The full coincidence among actual and NIR-MLR predicted curves validated the final purpose of this approach.