Systematic evaluation of matrix effects in supercritical fluid chromatography versus liquid chromatography coupled to...

Matrix effects (ME) is acknowledged as being one of the major drawbacks of quantitative bioanalytical methods, involving the use of liquid chromatography coupled to mass spectrometry (LC-MS). In the present study, the incidence of ME in SFC-MS/MS and LC-MS/MS in the positive mode electrospray ionization (ESI+) was systematically compared for the analysis of urine and plasma samples using two representative sets of 40 doping agents and 38 pharmaceutical compounds, respectively. Three different SFC stationary phase chemistries were employed, to highlight the importance of the column in terms of selectivity. Biological samples were prepared using two different sample treatments, including a non-selective sample clean-up procedure (dilute and shoot (DS) and protein precipitation (PP) for urine and plasma samples, respectively) and a selective sample preparation, namely solid phase extraction (SPE) for both matrices. The lower susceptibility to ME in SFC vs. reversed phase LC (RPLC) was verified in all the experiments performed on urine, and especially when a simple DS procedure was applied. Also, with the latter, the performance strongly varied according to the selected SFC stationary phase, whereas the results were quite similar with the three SFC columns, in the case of SPE clean-up. The same trend was observed with plasma samples. Indeed, with the PP procedure, the occurrence of ME was different on the three SFC columns, and only the 2-picolylamine stationary phase chemistry displayed lower incidence of ME compared to LC-MS/MS. On the contrary, when a SPE clean-up was carried out, the results were similar to the urine samples, with higher performance of SFC vs. LC and limited discrepancies between the three SFC columns. The type of ME observed in LC-MS/MS was generally a signal enhancement and an ion suppression for urine and plasma samples, respectively. In the case of SFC-MS/MS, the type of ME randomly varied according to the analyzed matrix, selected column and sample treatment.