Linear Support Vector Regression and Partial Least-Squares for Determination of Dapoxetine Hydrochloride and Tadalafil in Binary Pharmaceutical Mixtures

Abstract

Abstract Background: Dapoxetine (DAP) is a serotonin-norepinephrine reuptake inhibitor, and Tadalafil (TAD) is a phosphodiesterase type-5 inhibitor. Both are coformulated as tablets called Erectafil® for treatment of erectile ejaculation. Objective: DAP and TAD were analyzed in their binary mixtures and pharmaceutical formulations using two multivariate calibration chemometric models. Methods: Partial least-squares (PLS) and linear support vector regression (SVR) models were applied using two factor-four level experimental design and UV-spectrophotometric data. They were compared to each other, and their advantages and disadvantages were discussed. Results: The developed methods succeeded to determine DAP and TAD in different ratios with good results regarding International Conference on Harmonization guidelines. Linearity ranges were 2–15 μg/mL and 3–30 μg/mL for DAP and TAD, respectively, with good accuracy of 100 ± 0.37 for DAP and 100 ± 0.8 for TAD regarding PLS model and 100.04 ± 0.32 for DAP and 99.89 ± 0.77 for TAD regarding SVR model. Good precision values of 0.787 for DAP and 0.793 for TAD regarding PLS model and 1.105 for DAP and 0.930 for TAD regarding SVR model were obtained. The two models were applied on the dosage forms and statistically compared with the published HPLC method with no significant difference regarding accuracy and precision. Conclusions: The two models can be utilized for routine analysis and QC of DAP and TAD in their bulk and pharmaceutical formulations. The SVR model gives better results and generalization ability than those of the PLS model regarding accuracy and prediction error, while the latter is better for being simpler and faster.