Matrix effect screening for cloud-point extraction combined with liquid chromatography coupled to mass spectrometry: Bioanalysis of pharmaceuticals

Abstract

Matrix effects are one of the most challenging issues in the analysis of complex samples using liquid chromatography coupled to mass spectrometry (LC-MS). Apart from the instrumental origin, these effects are also related to sample preparation. Cloud-point extraction (CPE) is rarely combined with LC-MS as it requires the use of surfactants which might interfere with droplet evaporation. Thus, they are suspected to cause a significant matrix effect (signal suppression). In this paper, 73 model drugs with different physicochemical properties were screened to analyse how susceptible LC-MS is to the absolute and relative matrix effect (ME) when coupled with CPE for measurements in human plasma. Three combinations of the surfactant Triton X-114 concentration (1.5% or 6%) and extraction temperature (40 or 55 °C) in six pH values gave over 1300 analyte-sample preparation condition pairs. A new term - surfactant effect (calculated for the standard solution) - allowed us to distinguish between the surfactant effect and that related to interferences from human plasma. The screening revealed that CPE combined with LC-MS is not related to a significant ME in the optimal pH of extraction. A significant absolute ME (<85% or>115%) was observed only for 25% of the analytes. Data processing (principal component analysis, classification trees, partial least squares-discriminant analysis) based on the extraction conditions and molecular descriptors helped to identify compounds prone to the matrix effect and speed up method development. A low surfactant concentration and low temperature decreased both the absolute and relative ME. pH of the extraction influenced only the relative ME. Low retention time reduced the risk of relative ME, whereas high polarity and the possibility of hydrogen bond formation minimized the occurrence of the surfactant effect and absolute ME. A significant relative ME (>15%) was observed only for 11% of the compounds, thus CPE merged with LC-MS allowed to measure drug concentrations in a reliable manner for majority of compounds. The presented approach may be further applied to other analytes and matrices.

Keywords: Electrospray ionisation; Human plasma; Ionisation enhancement; Ionisation suppression; Molecular descriptors; Surfactant-mediated extraction.