Enantiomeric composition determination of highly overlapped chromatographic profiles by combination of chromatography and chemometric methods. Application to the analysis of enantiomeric drugs

OSORIO GRISALES, J.; ARANCIBIA, J.; OLIVIERI, A.; CASTELLS, C. B.

Amsterdam

Simposio; HPLC2013; 2013

Chirality is now a major issue in the design and development of new drugs and agrochemicals supported by the understanding of the relevant role of molecular recognition in many biological events. Thus, the analysis of enantiomers became a very important topic in separation science. The development of a chiral chromatographic method often requires expensive and time-consuming testing of different columns and conditions. Furthermore, for checking enantiomeric purity, the enantiomer of analytical interest (the minor peak) has to be quantitatively analyzed at very low levels respect to the main enantiomer. This requires injecting relatively large amounts of sample to enhance detectability and, as a consequence, the major component might become tailed due to mass overloading. Thus, improved accuracy in quantitation of the minor signal close to the major peak in the chromatogram is achievable as enantioresolution factor increases. However, baseline separations are not easily obtained, and partial overlapping between both enantiomers is very often observed. The problem would be even worse if other components in the sample coelute with the target analyte. We demonstrated how chiral liquid chromatography combined with multivariate chemometric techniques, specifically unfolded-partial least-squares regression (U-PLS), provides a powerful analytical methodology. Using U-PLS, strongly overlapped enantiomer profiles in a sample could be successfully processed and enantiomeric purity of ibuprofen could be accurately determined at trace levels (below 0.5 mg/L), with relative prediction error within ±3%, without requiring baseline enantioresolution between peaks. The scope of the algorithm was also explored and excellent precision was found even from almost completely overlapped signal. Finally, we tested the quality of the quantitative information obtained in presence of another component present in the sample that coelutes with the interest signals. The results were successful: we obtained accurate measurements of enantiomeric purity of ibuprofen, although the interferent could not be quantified.