Aplicação de métodos de quimiométricos de calibração multivariada - PLS e iPLS em conjunto com a espectroscopia Raman para análise qualitativa e quantitativa de polimorfos de valsartana

Abstract

Law 9,787 / 1999, also known as the law of generics, allowed the national pharmaceutical industries to produce drugs with expired patents. Valsartan, substance developed by Ciba-Geigy and marketed by Novartis Pharmaceutical Company under the trade name Diovan®, has become attractive to the national industry after the law of generics for its important application in the control of hypertension. This drug is an inhibitor of angiotensin and its receptors, class II biopharmaceutical classification. During the production of substances, including drugs, there is a possibility of polymorphism, which is the change in the crystalline habit of the material. In pharmaceuticals, this phenomenon impacts physicochemical properties with a direct consequence in bioavailability, besides the judicial processes, once the polymorphs are patented. Various conditions are those that may be required for the energy transition, the factors that affect the production line, transport and storage of the drug. In this way the FDA has established a series of analytical techniques to identify polymorphism in drugs. In this work the technique of Raman spectroscopy associated with chemometric tools, such as multivariate regression (PLS and iPLS), was used for the qualification and quantification of Valsartan polymorphs. Amorphous polymorphic forms (commercially used) and E forms (produced in the presence of water / ethanol) were used, which had their structures proved by X-ray diffraction and by thermal analysis (TG and DSC). The samples were qualified and quantified in the absence and presence of the excipients by Raman spectroscopy associated to the multivariate calibration chemometric tool - PLS and iPLS. The calibration model obtained through the iPLS regression with the pre-treatments of the first derivative Savitzky Golay and standard normal variation presented high predictive capacity for new samples, proved by the low values of RMSECV and RMSEP, for the active principle without excipient. The calibration model for the active ingredient together with excipients also presented low values of RMSECV and RMSEP, evidencing its ability to predict new samples even in the presence of the excipients. Therefore, the method presents high reliability in the analysis of drugs, low cost and versatility, and can be used by pharmaceutical companies and regulatory agencies in drug quality control.