Robustness Testing of Microemulsion Liquid Chromatographic Separation of Simvastatin and its Impurities

Abstract

Liquid chromatography employing the microemulsion as eluent was applied for the analysis of simvastatin and its six impurities. Previously, the optimized and validated method was tested to prove the method's capability to perform during the small changes of important parameters. In this paper, influence of parameters tipical for microemulsion liquid chromatography (MELC) on the method's robustness was presented.

For the results evaluation, and presentation, two statistical methods were applied. In order to designate the most convinient one, multiple regression (MR) and artificial neural networks (ANN) were applied and compared. A set of experiments was defined by the central composite design (CCD). Indepedent variables (SDS, n-buthanol and diisopropyl ether content) and dependent variables (retention factors of investigated substances) were the framework for both, MR and ANN. This study demonstrated that MR and radial basis function ANN are useful tools in understanding the effects of the investigated factors on the chromatographic system and definition of the robustness limits.

Keywords:

• Artificial neural networks
• MELC
• Multiple regression
• Robustness testing
• Simvastatin

ACKNOWLEDGMENT

The authors thank the Ministry of Science of Republic of Serbia for supporting these investigations in Project 142077 G.

Manuscript 6416

Notes

x₁ – content of diizopropylether (% w/v); x₂ – content of n-butanol (% w/v);x₃ – content of SDS (% w/v).

k ₁ – retention factor of hydroxy acid; k ₂ – retention factor of lovastatin; k ₃ – retention factor of simvastatin; k ₄ – retention factor of methyl simvastatin; k ₅ – retention factor of acetete ester; k ₆ – retention factor of anhydro simvastatin;k ₇ – retention factor of simvastaitin dimer.

SS_pe – Sum of squares due to purely experimental uncertainty.

SS_t – Total sum of squares.

SS_mean – Sum of squares due to the mean.

SS_corr – Sum of squares corrected to the mean (corrected sum of squares).

SS_fact – Sum of squares due to the factors (due to regression).

SS_r – Sum of squares of residuals.

SS_lof – Sum of squares due to lack of fit.

F_tab – 4.494 (p = 0.05).

Tr: Training; Vr: Varification.

r – coefficient of correlation.