Aplicação do método de regressão iPLS em conjunto com a espectroscopia Raman para a análise qualitativa e quantitativa de polimorfos do fármaco cloridrato de sertralina

Abstract

Active Pharmaceutical Ingredients (APIs) present in solid-state drugs may exist in more than one crystalline phase, forming a phenomenon known as polymorphism. Polymorphs may exhibit different values in physico-chemical and pharmacokinetic properties such as solubility, dissolution rate, bioavailability, and may cause serious problems in the production process and affect the intended therapeutic effect of the drug. Several techniques of analysis have been used to identify this phenomenon, however there is still no standard method by which the regulatory bodies are based for the quantification of polymorphs. Raman spectroscopy has many advantages for this purpose, since it is not destructive, it is not invasive, it does not require preparation of samples and the results are obtained in minutes, besides having characteristics of high selectivity. This, among other characteristics, favors the use of this technique for the simultaneous qualification/quantification of APIs and excipients in medicines. In this work the use of the multivariate calibration method of the iPLS type was evaluated for qualitative and quantitative analysis of Raman spectroscopy data obtained from samples of polymorphic mixtures in the presence and absence of excipients of the drug sertraline hydrochloride. In this sense, the first step was to synthesize and confirm the polymorphic structures by the DRX, DSC, TGA and Raman spectroscopy techniques. Then, the construction of the models was started based on the planning of binary mixing between the API commercially available on the drug (SRTL-I) and its hydrate (SRTL-H) synthesized by recrystallization of the ethanol solvate in water. In order to improve the performance of the models, different pretreatment algorithms were used to reduce the effects of the systematic variations, being the normalization by area, standard normal variation (SNV) and orthogonal signal correction (OSC) of data. The result of planning binary mixtures of the 50 mg samples (SRTL-I and SRTL-H mixture) provided an iPLS model with prediction errors of 2.6 mg for SRTL-I and 2.4 mg for the SRTL-H. For planning binary mixing with the excipients commonly used in the formulation, despite the presence of these interferents, the results were even more promising. Considering the simulation of the polymorphs in the presence of excipients of a sertraline hydrochloride tablet of approximately 300 mg, 250 mg of excipients and 50 mg of sertraline hydrochloride (mixture of polymorphs), the prediction errors were 1.42 mg for the SRTL-I and 1.02 mg for the SRTL-H. The results obtained show that Raman spectroscopy in combination with the iPLS regression method are suitable to quantify binary mixture of sertraline hydrochloride polymorphs.